Labiyi F 2016 PHD Thesis

Electrochemical Kinetics and Mechanisms of
Iron Oxidation in CO2-containing Aqueous

Monoethanolamine
Folasade Labiyi
June 2016
Department of Chemical Engineering and Chemical Technology

Imperial College London
South Kensington Campus
London SW7 2AZ
A thesis submitted for the degree of Doctor of Philosophy of Imperial College

London and the Diploma of Membership of Imperial College London
I
Declaration of Originality
I, Folasade Labiyi, herby certify that this thesis has been written by me, based on the original
work conducted by me, under the guidance and advice of my academic supervisors in the
Department of Chemical Engineering at Imperial College London. The information used or
derived from other published or unpublished sources have been clearly sited and appropriately
acknowledged in the list of references.
Copyright Notice
Imperial College of Science, Technology and Medicine
Department of Chemical Engineering
Electrochemical Kinetics and Mechanisms of Iron Oxidation in CO2-containing Aqueous

Monoethanolamine
The copyright of this thesis rests with the author and is made available under a Creative
Commons Attribution Non-Commercial No Derivatives licence. Researchers are free to copy,
distribute or transmit the thesis on the condition that they attribute it, that they do not use it for
commercial purposes and that they do not alter, transform or build upon it. For any reuse or
redistribution, researchers must make clear to others the licence terms of this work.
2016 Folasade Labiyi

II
Abstract
Concerns about the impact of anthropogenic CO2 emissions on global climate change are
necessitating the development of low-carbon technologies. Despite its energy requirements,
carbon capture and sequestration (CCS) provides a promising approach to limiting carbon
dioxide emissions, in addition to renewable energy and nuclear power. Due to the depth of
technological experience and retrofitting capabilities associated with the process, post-
combustion CO2 capture based on absorption in alkanolamine solutions is presently the most
feasible technology available for mitigating the soaring levels of atmospheric CO2 and
achieving the ambitious targets set by the Intergovernmental Panel for Climate Change (IPCC).
Whilst the process of CO2 capture by alkanolamines involves a technologically mature process,
it is still faced with numerous challenges, such as high energy requirements and operating costs,
as well as operational difficulties. Monoethanolamine (MEA), the benchmark amine for the
process, frequently becomes contaminated with degradation products and, when CO2-loaded,
is corrosive to the carbon steel process equipment. Hence, this thesis aimed to determine the
oxidation and reduction reaction kinetics and mechanisms of iron in aqueous MEA-CO2
systems as function of experimental variables, aiming to predict iron corrosion rates under
process conditions.
The behaviour of iron in aqueous MEA solutions was characterized by voltammetry with a
rotating disc electrode (RDE) and an electrochemical quartz crystal microbalance (EQCM), as
functions of temperature (25-80oC), CO2 loading (0-0.6 mol CO2 (mol amine)-1), pH (8.10-
12.55), MEA concentration (5-60 wt%) and oxygen concentration. Electrode potential-pH and
activity-pH diagrams of iron-water–CO2 systems were used to assist with reaction assignments.
III
The passive electrochemical behaviour of Fe in MEA at pH ca. 12 switched to active dissolution
on loading the MEA with CO2, causing the pH to decrease to ca. 8. Analysis of the resulting
kinetic data enabled corrosion rates to be predicted as functions of the experimental variables.
Based on the proposed corrosion mechanisms from the voltammetric results, a mechanistic
model was developed for the uniform corrosion of iron in CO2-loaded MEA systems, taking
into account the CO2 absorption equilibria reactions and the electrochemical reactions at the
iron | solution interface. Equilibrium concentrations of the amine species (RNH2, RNH3+,
RNHCOO-), carbon(IV) species (HCO3-, CO32-) and hydrogen ions (H+) were calculated with
a Kent-Eisenberg type model. The electrochemical reactions incorporated in the model were
the anodic dissolution of iron and the cathodic reduction of H+, direct water reduction and the
reduction of oxygen. The corrosion model was developed by simulating polarization curves
based on the species concentrations and the transport limited current densities of the iron RDE
defined by the Levich equation.
In order to measure dissolved iron concentrations under typical CO2 absorption conditions, an
electrochemical flow reactor was designed and fabricated from PTFE with an iron/steel anode,
platinised titanium cathode and a cation-permeable Nafion membrane. Inductively coupled
plasma optical emission spectrophotometry (ICP-OES) was used to determine FeII
concentrations in the amine solution after potentiostatic electrolyses as a function of solution
flow-rate, temperature, CO2 loading, pH, MEA concentration, oxygen content, steel type and
amine type, enabling partial current densities leading to iron dissolution to be deconvoluted
from measured current densities.
Results from the voltammetric data from the RDE were used to propose a mechanism for the
oxidation and reduction reactions occurring at the iron surface. During positive-going potential
sweeps, the voltammograms were characterised by 3 anodic peaks corresponding to the anodic
IV
dissolution of Fe to FeII at low potentials, leading to iron carbonate formation if its solubility
product was exceeded, and a passive region from the formation of adsorbed products such as
iron(III) (hydr-)oxide at higher potentials. The cathodic reactions at the iron surface included
reductions of H+, H2O and O2.
Based on the thermodynamic predictions, the anodic dissolution of iron was assigned to the
formation of a soluble FeII species. Using a flow reactor operated in batch recycle mode,
constant potential electrolyses at a potential -0.6 V, within the anodic dissolution potential
range, resulted in a dissolution charge yield of less than unity, which varied depending on the
experimental conditions, implying the formation of insoluble iron species such as Fe(OH)2 or
FeCO3. The mass transport behaviour of the flow reactor was characterised as a function of
solution flow rate using the transport controlled reduction of hexacyanoferrate(III) ions at a
platinised titanium electrode, resulting in a mass transport correlation for the reactor. A value
of the diffusion coefficient for FeII from the literature enabled prediction of the conditions
required for FeCO3 and Fe(OH)2 formation, based on measured fluxes and dissolved FeII
concentrations in solution, from which the rate of FeCO3 formation could also be estimated.
Kinetic analysis of the data from the RDE, EQCM, flow reactor and the corrosion model
resulted in a complementary set of results. The corrosion behaviour of iron in aqueous MEA-
CO2 solutions was sensitive to changes in the operating parameters; iron dissolution rates were
enhanced by increasing temperature, CO2 loading, solution velocity and oxygen content.
However, at high temperatures, high CO2 loading and high FeII concentration and lower
solution velocity provided favourable conditions under which a protective FeCO3 layer could
precipitate. Dissolution rates increased with concentration at lower concentrations of MEA (5-
40 wt%) and decreased with increasing concentration at higher concentrations (40-60 wt%),
due to viscosity effects. Significantly lower corrosion rates were measured in the other
V
commercially available solvents tested namely methyldiethanolamine (MDEA), 2-amino-2-
methyl-1propanol (AMP) and aminoethylpiperazine (AEP). Carbon steel exhibited similar
dissolution rates to those of iron, whereas those for stainless steel were significantly lower.
From the comprehensive results on the oxidation and reduction kinetics of iron in benchmark
MEA-CO2 systems, the effect of MEA concentrations on predicted iron corrosion rates brings
into question the effectiveness of the concentration of MEA used in amine scrubbing being
limited to 30 wt%, specifically to limit corrosion related issues. The results also indicated that
MEA was the most ‘aggressive’ in corrosion behaviour, as other amines such as MDEA, AMP
and AEP provided more promising alternatives, based on both their predicted corrosion
behaviour and their reported efficiencies in the absorption process.

VI
Acknowledgements
First and foremost, I would like to give thanks to God, the Almighty for providing me with this
opportunity and granting me the capability to finish.
Pursuing a PhD degree has been an extraordinary experience, which will forever be a pivotal
chapter in my life. It stands to reason that I would not have made it this far without the technical
and moral support from the following people.
For the most contribution to this thesis, I would like to express my sincerest appreciation to my
academic supervisors, Prof. Geoff Kelsall and Prof. Paul Fennell, who are a great source of
knowledge and ideas. It has been a pleasure working with them and their meticulous and
punctilious approach to research has made a significant impact on my work. Without their
words of encouragement, patience and immense support, it would have been an impossible
task to maintain momentum and navigate through the numerous personal and work-related
obstacles that I faced throughout my PhD.
I would like to thank and acknowledge the Engineering and Physical Science Research Council
(EPSRC) for sponsoring my research.
A very warm thank you to my colleagues and friends from the electrochemical engineering
research group, Anna, Lisa, Franky and Nick for being a constant source of humour,
encouragement and support. My time working on this project would not have been as enjoyable
without them. I would also like to thank Dr. Kyra Campbell, who was always very encouraging
and kindly offered to proofread a few sections of my thesis, in spite of her very busy schedule.
I thank my fellow colleagues in Paul’s group, for the inspiring discussions on CCS-related
issues as well as fun social activities that provided a much-needed distraction.

VII

I am extremely grateful to my mother Abosede for her unconditional love, unwavering support
and endless sacrifices and my aunts Enitan Osibodu and Aishatu Usman, who always
encouraged me from afar. I must extend my thanks to my dearest friends Mikki, MJ and Olly,
all of whom have been there for me during this entire journey and have been exceptionally kind
and tolerant throughout.
My most personal thank you is to my husband Kunle, who I owe an enormous debt of gratitude
for being a constant source of motivation and support. Finally, to my daughter Tumi, without
whom, this thesis would have been completed a long time ago.
VI
Contents
Declaration of Originality .......................................................................................................... I
Abstract ..................................................................................................................................... II
Acknowledgements ................................................................................................................... V
Contents .................................................................................................................................. VI
List of Figures ........................................................................................................................ XII
List of Tables ..................................................................................................................... XXIX
Nomenclature .................................................................................................................. XXXIII
1. Chapter 1 Introduction ................................................................................................... 1
1.1. Motivation for carbon capture and sequestration........................................................ 1
1.2. Operational difficulties in CO2 absorption process .................................................... 6
1.3. Corrosion in Amine Scrubbing Plants ........................................................................ 7
1.4. Existing Corrosion Control ......................................................................................... 8
1.5. Limitation to Available Corrosion Data ..................................................................... 9
1.6. Project Aims.............................................................................................................. 10
1.7. Project Scope and Objectives.................................................................................... 10
1.7.1. Project Tasks ......................................................................................................... 11
1.8. Thesis outline: ........................................................................................................... 12
1.9. References ................................................................................................................. 14
2. Chapter 2 Principles of Electrochemical Systems for Corrosion Studies ................. 16
2.1. Thermodynamics of Electrochemical reactions ........................................................ 18
2.1.1. Nernst Equation .................................................................................................... 19
2.1.2. Pourbaix Diagrams................................................................................................ 20

VII
2.2. Kinetics of Electrochemical Reactions ..................................................................... 26
2.2.1. Rate of reaction: Faraday’s Laws of Electrolysis ................................................. 26
2.2.2. Rate limiting Mechanisms .................................................................................... 28
2.2.3. Charge Transfer Controlled Reactions.................................................................. 28
2.2.4. Mixed Potential Theory ........................................................................................ 33
2.3. Mass transport ........................................................................................................... 36
2.3.1. Transport Processes in Electrochemical Systems ................................................. 36
2.3.2. Electrochemical Reaction Coupled to Mass Transport ......................................... 43
2.4. References ................................................................................................................. 46
3. Chapter 3 Literature Review ........................................................................................ 47
3.1. The CO2 Absorption Process ................................................................................... 48
3.2. Solvents for CO2 Absorption ................................................................................... 53
3.2.1. CO2 Absorption Reactions ................................................................................... 53
3.3. CO2 Induced Corrosion in Amine Plants ................................................................. 59
3.3.1. Plant Experience .................................................................................................. 60
3.4. Corrosion Mechanisms in Amine Plants.................................................................. 68
3.4.1. Wet Acid Gas Corrosion ...................................................................................... 69
3.4.2. Amine Solution Carbon Steel Corrosion ............................................................. 78
3.5. Factors Affecting CO2 Corrosion in Amine Plants .................................................. 80
3.5.1. Temperature ......................................................................................................... 80
3.5.2. CO2 Loading ........................................................................................................ 83
3.5.3. Amine Type ......................................................................................................... 85
3.5.4. Amine Concentration ........................................................................................... 88
3.5.5. Effect of Oxygen .................................................................................................. 88
3.5.6. Effect of Surface Films ........................................................................................ 89
3.5.7. Amine Degradation Products ............................................................................... 91

VIII
3.6. Review of Corrosion Product Films ........................................................................ 94
3.6.1. Iron Carbide Films ............................................................................................... 95
3.7. Iron carbonate Film .................................................................................................. 97
3.7.1. Mixed Iron Carbonate and Iron Carbide Film ................................................... 106
3.7.2. Iron Carbonate and Iron Oxide Film .................................................................. 108
3.8. Corrosion Inhibitors for Amine Absorption Plants................................................ 110
3.9. Equilibrium Models for Aqueous Amine CO2 Systems ........................................ 115
3.9.1. Solution Chemistry ............................................................................................ 115
3.9.2. Thermodynamic Framework .............................................................................. 116
3.10. Review of the mechanisms of iron dissolution and hydrogen evolution ............... 120
3.10.1. Mechanism of the dissolution of iron ................................................................ 120
3.10.2. The influence of halide ions on dissolution kinetics .......................................... 129
3.10.3. Mechanisms for Dissolution of Iron in Solutions in the Mid-pH range ............ 131
3.10.4. Mechanisms of Iron Dissolution in Alkaline Solutions ..................................... 132
3.11. The Mechanism of Hydrogen Evolution Reaction (HER) on Iron ........................ 135
3.12. Experimental work in CO2-containing solutions ................................................... 138
3.13. References .............................................................................................................. 141
4. Chapter 4 Experimental Methods and Materials ..................................................... 155
4.1. Materials ................................................................................................................ 157
4.1.1. Chemicals........................................................................................................... 157
4.1.2. Standard reagents ............................................................................................... 157
4.1.3. Electrode Materials ............................................................................................ 158
4.1.4. Cation Exchange Membrane .............................................................................. 160
4.1.5. Gases .................................................................................................................. 160
4.2. Electrode and Solution Preparation ....................................................................... 160
4.2.1. Electrodes ........................................................................................................... 160

IX
4.2.2. Aqueous Amine Solutions ................................................................................. 161
4.3. Instrumentation for Electrochemical Measurements ............................................. 162
4.4. Experimental Methods for Micro-kinetic studies on Iron...................................... 162
4.4.1. Cyclic Voltammetry ........................................................................................... 162
4.4.2. Kinetic Studies Using Rotating Disc Electrodes ............................................... 163
4.4.3. Tafel Slopes Extrapolation Method for Corrosion Rate Determination ............ 169
4.4.4. Electrochemical Quartz Crystal Microbalance .................................................. 171
4.5. Electrochemical Flow Reactor for Time-dependent Analysis of Iron Corrosion in

Flow Reactor Mode............................................................................................................ 178
4.5.1. Reactor design, configuration and materials ...................................................... 178
4.5.2. Experiment Procedure ........................................................................................ 180
4.6. References .............................................................................................................. 184
5. Chapter 5 Electrochemical Study of Iron Electrodes in MEA-H2O-CO2 Systems 188
5.1. Fe in MEA+ H2O (absence of CO2)................................................................... 188
5.2. Fe in MEA + H2O + CO2 ................................................................................... 194
5.3. Effect of electrochemical parameters................................................................. 200
5.3.1. Effect of electrode rotation rate ..................................................................... 200
5.3.2. Effect of positive potential limit .................................................................... 203
5.3.3. Effect of potential scan rate ........................................................................... 206
5.4. Effect of Process Parameters ............................................................................. 210
5.4.1. Effect of temperature ..................................................................................... 210
5.4.2. Effects of CO2 loading and pH ...................................................................... 214
5.4.3. Effect of MEA concentration ......................................................................... 231
5.4.4. Effect of Oxygen ............................................................................................ 237
5.5. Interpretation and Discussion ............................................................................ 239
5.6. Kinetic analysis .................................................................................................. 245

X
5.6.1. Effect of CO2 loading..................................................................................... 246
5.6.2. Temperature effect ......................................................................................... 248
5.6.1. Effect of MEA Concentration ........................................................................ 253
5.6.2. Effect of Mass Transport ............................................................................... 260
5.6.2. Effect of Oxygen ............................................................................................ 262
5.7. Summary & Conclusions ................................................................................... 266
6. Chapter 6 Mechanistic Model for the Prediction of Corrosion of Iron in MEA + H20
+ CO2 Solutions .................................................................................................................... 270
6.1. Integrated Model for Fe corrosion in MEA+ H2O+CO2.................................... 271
6.1.1. Model scheme ................................................................................................ 274
6.1.2. Model Structure and Development ................................................................ 276
6.1.3. Thermodynamic Modelling of Aqueous MEA-CO2 system .......................... 280
6.1.4. Electrochemical model for Iron in MEA-CO2-H2O....................................... 282
6.2. Model Verification by Comparison with Experiments ...................................... 290
6.3. Effect of CO2 loading......................................................................................... 292
6.4. Effect of Temperature ........................................................................................ 295
6.5. Effect of MEA Concentration ............................................................................ 300
6.6. Discussion and Conclusions .............................................................................. 305
6.7. References .......................................................................................................... 308
7. Chapter 7 Analysis of Time-dependent Corrosion of Iron in MEA- H2O- CO2

Systems ................................................................................................................................. 310
7.1. Electrochemical Behaviour of Iron in Large Scale Electrode ............................... 312
7.2. Effect of Electrode Potential .................................................................................. 314
7.3. Effect of Solution Velocity .................................................................................... 319
7.3.1. Reactor Characterisation with Mass Transport Correlations ......................... 319
7.3.2. Reactor Characterisation by the Reduction of Hexacyanoferrate(III) ........... 321
7.3.3. Effect of solution flow rate on iron corrosion in MEA+ CO2 solutions ........ 325
XI
7.4. Temperature Effect - Full CO2 saturation .............................................................. 330
7.5. Effect of CO2 loading............................................................................................. 337
7.6. Effect of Oxygen Content ...................................................................................... 346
7.7. Effect of MEA concentration ................................................................................. 350
7.8. Effect of Metal Type .............................................................................................. 361
7.9. Effect of Amine Type ............................................................................................ 367
7.10. Longer Term Corrosion Measurements ................................................................. 376
7.11. References .............................................................................................................. 376
8. Chapter 8 Conclusions and Future Work ................................................................. 382
8.1. Summary of this project ......................................................................................... 383
8.2. Rotating Disc Electrode and Electrochemical Quartz Crystal Results .................. 383
8.3. Mechanistic Modelling .......................................................................................... 387
8.4. Electrochemical Flow reactor ................................................................................ 388
8.5. Future Work Proposals .......................................................................................... 391
9. Chapter 9 Appendices ................................................................................................. 393
I. CO2 Solubility Data .................................................................................................... 394
II. Cyclic Voltammetry Data ........................................................................................... 396
III. Effect of MEA Concentration ................................................................................. 397
IV. The influence of ionic composition on the electrochemical behaviour of iron in

sodium carbonate-hydrogen carbonate solutions at pH 8.01 ............................................. 401
V. Kinetic Data obtained from Polarization curves ......................................................... 402
VI. Thermodynamic Data .............................................................................................. 406

XII
List of Figures
Figure 1-1: Post-combustion, pre-combustion and oxy-fuel combustion CO2 capture systems
(Figueroa et al., 2008). ............................................................................................................... 5
Figure 2-1: Pourbaix diagram of water; 298 K, 0.1 MPa. ........................................................ 22
Figure 2-2: Potential-pH diagram for Fe-H2O systems with dissolved iron activity of 10-4, 298 K
and 0.1 MPa (Kelsall 2011). ..................................................................................................... 26
Figure 2-3: Schematic representation of the relationship between current and potential for a
simple electrochemical reaction under kinetic control. ............................................................ 29
Figure 2-4: Plot of Butler-Volmer Equation for charge transfer limited reaction. ................... 31
Figure 2-5: Schematic of Tafel plot .......................................................................................... 32
Figure 2-6: Anodic and cathodic half-cell reactions present simultaneously on a corroding iron
surface in acidic media.............................................................................................................. 34
Figure 2-7: Relationship of measured polarization curve to the Evans diagram for corroding
metal in acidic media. ............................................................................................................... 35
Figure 2-8: Schematic of an electrochemical system showing the transport processes and one
dimensional distribution of electrical potential......................................................................... 38
Figure 2-9: Regions of Cathodic Current density-Overpotential Relationship ........................ 44
Figure 3-1: Amine CO2 capture system flow diagram............................................................. 49
Figure 3-2: Reaction Scheme of Alkanolamines with CO2 ..................................................... 54
Figure 3-3: Flow diagram of a typical amine plant showing principal areas of corrosion and
suggested materials of construction (Kohl & Nielsen, 1997). ................................................. 68
Figure 3-4: Effects of temperature and CO2 partial pressures on carbon steel corrosion rates (de
Waard and Lotz, 1993) ............................................................................................................ 82
Figure 3-5: Nomograph for corrosion rates of carbon steel in aqueous CO2 environments (de
Waard and Lotz, 1993). ........................................................................................................... 82
XIII
Figure 3-6: Calculated growth rate of iron carbonate as a function of temperature and
supersaturation (Jonson and Tomson, 1991). .......................................................................... 99
Figure 3-7: FeCO3 embedded in a Fe3C Film; 80oC, 10-100 supersaturation ratio (Dugstad,
1998). ..................................................................................................................................... 102
Figure 3-8: FeCO3 embedded in a Fe3C film; 60oC, 10-100 supersaturation ratio (Dugstad,
1998). ..................................................................................................................................... 103
Figure 3-9: FeCO3 embedded in a Fe3C film; 60oC, 1-3 supersaturation ratio (Dugstad, 1998).
................................................................................................................................................ 103
Figure 3-10: The composition of protective and non-protective iron carbide and iron carbonate
films (Crolet et al., 1996). ...................................................................................................... 108
Figure 3-11: Potential-pH diagram for Fe-CO2-H2O System, showing the lines for dissolved
iron activity of 10-4 and (0.1 MPa CO2, 298 K) (Kelsall 2013). ............................................ 109
Figure 3-12: Potential-PH diagram for the V-H2O system at 298K. Activity of dissolved species
= 0.01 (Kelsall et al., 1992). .................................................................................................. 114
Figure 3-13: General reaction scheme for iron dissolution (Drazic, 1989) ........................... 135
Figure 4-1: Schematic of the side and end view of the RDE and the flow pattern induced by
rotation of an RDE system ..................................................................................................... 164
Figure 4-2: Rotating disc electrode system. ........................................................................... 167
Figure 4-3: Tafel plot for the determination of corrosion current density and hence rate. .... 170
Figure 4-4: Acoustic waves propagating through a quartz crystal, electrodes and deposit ... 172
Figure 4-5: Schematic diagram of membrane-divided electrochemical flow reactor............ 178
Figure 4-6: Schematic diagram of the experimental set-up of the electrochemical flow reactor
................................................................................................................................................ 179
Figure 4-7: (a) The assembled electrochemical flow reactor in action (b) The anolyte
compartment (c) The catholyte compartment (d) A Fe plate (anode) ................................... 181
Figure 5-1: Cyclic voltammogram of a stationary iron RDE in 30 wt% MEA, at 10 mV s-1
(T = 25oC, pH 12.55). ............................................................................................................ 189
Figure 5-2: Influence of electrode rotation rate on voltammogram of iron RDE in 30 wt% MEA,
at 10 mV s-1 (T = 25 °C, pH 12.55) ....................................................................................... 191
XIV
Figure 5-3: Influence of rotation rate on the voltammogram obtained on an iron RDE in 30 wt%
MEA, at 10 mV s-1 (T = 40 °C, pH 12.27) ............................................................................ 192
Figure 5-4: Influence of electrode rotation rate on voltammograms of iron RDE in 30 wt%
MEA, at 10 mV s-1 (T = 60oC, pH 12.10) .............................................................................. 192
Figure 5-5: Stable (black) and meta-stable (red) potential-pH diagrams for Fe-H2O system for
dissolved iron activity of 10-6, 298 K and 0.1 MPa (Kelsall 2011); red lines correspond to meta-
stable behaviour of iron hydroxides rather than the more stable oxides................................ 193
Figure 5-6: Liquid-phase composition of 30wt% MEA solution loaded with CO2 at 25°C.
Compositions were calculated with the Kent-Eisenberg Model (Kent and Eisenberg, 1976).
................................................................................................................................................ 195
Figure 5-7: Cyclic voltammogram of stationary iron RDE in 30 wt% MEA + H2O + CO2;
10 mV s-1 (T = 25 °C, pH 8.01). ............................................................................................ 196
Figure 5-8: Cyclic voltammogram of a stationary iron RDE in 30 wt% MEA + H2O + CO2, at
10 mV s-1 (T = 25 °C, pH 8.01) with reactions assigned. ...................................................... 196
Figure 5-9: Potential-pH diagram for Fe-CO2-H2O System, showing the lines for dissolved iron
activity of 10-4 and (0.1 MPa CO2, 298 K). ........................................................................... 199
Figure 5-10: Effect of rotation rate on cyclic voltammograms of an iron RDE in 30wt% MEA
+ H2O + CO2, at 10 mV s-1 (T = 25oC, pH 8.01). .................................................................. 201
Figure 5-11: Effect of rotation rate on jP for Fe RDE in 30wt% MEA + H2O + CO2, at 10 mV s-
1
(T = 25 °C, pH 8.01) in the a) positive-going and b) negative-going potential scan. ......... 203
Figure 5-12: Cyclic voltammograms for Fe | 30wt% MEA + H2O + CO2, at 10 mV s-1
(T = 25 °C, pH 8.01) with positive potential limits: (a) -0.40 V, (b) -0.20 V, (C) +0.20 V (d)
+0.60 V (e) +1.0 V. ................................................................................................................ 205
Figure 5-13: Effect of potential scan rate on voltammograms for a Fe RDE in 30wt% MEA +
H2O + CO2 (T = 25oC, pH 8.01, f = 1500 rpm,) (a) 100 ≤ v ≤10 mV s-1 (b) 10 ≤ v ≤1 mV s-1.
................................................................................................................................................ 208
Figure 5-14: Effect of potential scan rate on current density for peak I of stationary Fe electrode
in 30wt% MEA + H2O + CO2 (T = 25 °C, pH 8.01). ............................................................ 208
XV
Figure 5-15: Effects of rotation and potential scan rates on current density of peak(II), jPII, on
Fe in 30wt% MEA + H2O + CO2, at scan rates 100 ≤ v ≤10 mV s-1 ..................................... 209
Figure 5-16: Effects of rotation and potential scan rates on peak (II) current density jPII of Fe in
30wt% MEA + H2O + CO2, at various scan rates 10 ≤ v ≤ 1 mV s-1 .................................... 209
Figure 5-17: Effect of aqueous solution temperature on cyclic voltammograms for iron RDE in
30 wt% MEA + H2O + CO2 at pH ~8.01; 10 mV s-1, f = 1500 rpm. ..................................... 210
Figure 5-18: Effect of rotation rate and temperature (18 oC ≤ T ≤ 80 oC) on jPII for Fe RDE in
30 wt% MEA + H2O + CO2, at 10 mV s-1 (pH 8.01)............................................................. 212
Figure 5-19: Arrhenius plot for jpII 30 wt% MEA + H2O + CO2, at 10 mV s-1 (pH 8.01) for
varying aqueous solution temperatures (18 oC ≤ T ≤ 80 oC) and rotation speeds (60 rpm ≤ f ≤
1500 rpm). .............................................................................................................................. 213
Figure 5-20: Cyclic voltammogram and corresponding massogram for Au quartz crystal in
300 g dm-3 aqueous FeSO4 at pH 3.2; at 10 mV s-1, T = 25 C). ............................................ 215
°
Figure 5-21: Potentiostatic deposition of Fe on Au quartz crystal in 300 g dm-3 FeSO4 at pH

3.2, - 1.0 V (AgCl|Ag) and T = 25 °C.................................................................................... 216
Figure 5-22: Cyclic voltammogram obtained on the deposited Fe quartz crystal in 30 wt% MEA
+ H2O, at 100 mV s-1 (T = 25 °C, pH 12.55) with equilibrium reaction potentials assigned.217
Figure 5-23: Mass change corresponding to voltammogram () of electrodeposited Fe quartz

crystal in 30 wt% MEA + H2O, at 100 mV s-1 (T = 25 °C, pH 12.55) with reactions assigned.
................................................................................................................................................ 218
Figure 5-24: Cyclic voltammogram of electrodeposited Fe quartz crystal in 30 wt% MEA +

H2O + CO2 at pH 8.01, 100 mV s-1, T = 25 °C; equilibrium reaction potentials assigned. ... 219
Figure 5-25: Mass change corresponding to voltammogram (Figure 5-23) of electrodeposited

Fe quartz crystal in 30 wt% MEA + H2O + CO2, at 100 mV s-1 (T=25 °C, pH 8.01). .......... 220
Figure 5-26: Effect of rotation rate on cyclic voltammograms of iron RDE in 30 wt% MEA at
pH 12.55 and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned. ...... 221
Figure 5-27: Cyclic voltammogram of stationary iron RDE in 30 wt% MEA+ CO2 at pH 11.00
and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned. ...................... 221
XVI
Figure 5-28: Cyclic voltammogram of stationary iron RDE in 30wt% MEA+ CO2 at pH 10.01
and T = 25 C; 10 mV s-1, with equilibrium potentials of reactions assigned......................... 222
and T = 25 °C; 10mVs-1, with equilibrium potentials of reactions assigned. ........................ 223
Figure 5-33: Cyclic Voltammogram obtained on a stationary iron RDE in 30 wt% MEA+ CO2
at pH 8.17 and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned. .... 224
Figure 5-34: Cyclic Voltammogram obtained on a stationary iron RDE in 30wt% MEA+ CO2
at pH 8.01 and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned. .... 225
Figure 5-35: Effect of pH on voltammograms of a stationary iron RDE in 30wt% MEA+ CO2
at T = 25 °C; 10 mV s-1, f = 1500 rpm. .................................................................................. 225
Figure 5-36: Effects of rotation rate and pH on peak current density jPII of Fe in 30 wt% MEA
+ H2O + CO2; potential scan rate: 10 mV s-1. ........................................................................ 226
Figure 5-37: Effect of aqueous solution pH on peak (II) electrode potential EPII. ................ 226
Figure 5-38: Effects of CO2: – Comparisons between cyclic voltammogram for iron RDE in
30 wt% MEA+ CO2 and 30 wt% MEA+ 1 M H2SO4; 10 mV s-1 (T = 25 °C, f = 0- 1500 rpm).
................................................................................................................................................ 229

H2SO4 at pH 8.01 and T = 25oC; 100 mV s-1, with equilibrium potentials of reactions assigned.
................................................................................................................................................ 230
Figure 5-40: Mass change corresponding to voltammograms (Figure 5-39) of electrodeposited

Fe quartz crystal in 30 wt% MEA + H2O + CO2 at pH 8.01 and T = 25 C; 100 mV s-1. ...... 230 °
Figure 5-41: Effect of (5-30 wt%) MEA concentration on cyclic voltammograms of iron RDE
rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 25 °C; 10 mV s-1..................... 232
XVII
Figure 5-42: Effect of (30-60 wt%) MEA concentration on cyclic voltammograms of an iron
RDE rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 25 °C; 10 mV s-1............ 233
Figure 5-43: Effects on jPII of rotation rate, temperature (25-80 °C) and aqueous MEA
concentration (10-60 wt%) for Fe RDE in MEA + H2O + CO2 at pH ̴ 8.01; 10 mV s-1........ 234
Figure 5-44: Cyclic voltammograms and corresponding mass change of electrodeposited Fe

quartz crystal in (5-30 wt%) MEA + H2O + CO2 at pH ̴ 8.01 and T = 25 C; 100 mV s-1. .... 235
Figure 5-45: Cyclic voltammogram of electrodeposited Fe on quartz crystal in (5-30) wt% MEA
+ H2O at pH = 8.01 and T = 25 °C; 100 mV s-1. ................................................................... 236
Figure 5-46: Mass changes corresponding to voltammograms (Figure 5-45) of electrodeposited

Fe quartz crystal in (5-30 wt%) MEA + H2O + CO2 at pH 8.01 and T = 25 °C; 100 mV s-1.236
Figure 5-47: Effect of oxygen concentration and temperature on cyclic voltammograms of iron
RDE rotating at f = 1500 rpm in 30 wt% MEA+ CO2 at pH (7.24-8.01) ; 10 mV s-1. ......... 238
Figure 5-48: Activity-pH diagram of Fe-H2O-CO2 at 0.1 MPa CO2; 298 K (Kelsall, 2015).241
Figure 5-49: Effect of pH on polarisation behaviour of an iron RDE in 30 wt% MEA+ CO2, at
10 mV s-1; T = 25 oC, f = 1500 rpm. ...................................................................................... 247
Figure 5-50: Effect of pH on oxidation / corrosion rates of an iron RDE in 30 wt% MEA+ CO2,
at 10 mV s-1 (T = 25 oC, f = 1500 rpm). ................................................................................ 248
Figure 5-51: Effect of CO2 loading on hydrogen ion concentration based on the Kent-Eisenberg
model for 30 wt% MEA solution and 25 oC. ......................................................................... 248
Figure 5-52: Effect of solution temperature on polarisation curves of iron RDE in 30 wt% MEA
+ H2O + CO2 at pH 7.24 - 8.01; 10 mV s-1, f = 1500 rpm. .................................................... 250
Figure 5-53: Effect of solution temperature on the corrosion rate obtained on iron RDE in 5-60
wt% MEA + H2O + CO2 at pH 7.24 - 8.01; 10 mV s-1, f = 1500 rpm. .................................. 250
Figure 5-54: Effect of temperature on hydrogen ion concentration [H+] based on the Kent-
Eisenberg model for 30 wt% MEA solution and 25oC. ......................................................... 251
Figure 5-55: Effect of solution temperature on polarisation curves of iron RDE in 5-60wt%
MEA + H2O + CO2 at pH 7.24-8.01; 10 mV s-1, f = 1500 rpm. ............................................ 252
Figure 5-56: Effect of MEA concentration on the polarisation curves obtained on iron RDE in
MEA + H2O + CO2 at pH 8.01 and T = 25oC; 10 mVs-1, f = 1500 rpm. ............................... 255
XVIII
Figure 5-57: Effect of MEA concentration on polarisation curves of iron RDE in MEA + H2O
+ CO2 at pH 7.83 and T = 40 oC; 10 mV s-1, f = 1500 rpm. .................................................. 256
+ CO2, at 10 mV s-1 ; pH 7.49, f = 1500 rpm, T = 60oC. ...................................................... 256
+ CO2 at pH 7.24 and T = 80 °C; 10 mV s-1, f = 1500 rpm. .................................................. 257
Figure 5-60: Effect of MEA concentration on the corrosion rate obtained on iron RDE in MEA
+ H2O + CO2 at pH 7.24 – 8.01 and T = 25-80oC; 10 mV s-1, f = 0-1500 rpm. .................... 258
Figure 5-61: Effect of MEA concentration on the species concentrations of the aqueous bulk
solution at chosen CO2 loading of 0.6 CO2 loading / mol CO2 (mol amine)-1, T = 25 oC,
calculated by the Kent-Eisenberg Model (Kent and Eisenberg, 1976). ................................. 259
Figure 5-62: Effect of rotation rate on polarisation curves of iron RDE in 30 wt% MEA + H2O
+ CO2, at 10 mV s-1; pH 8.01, f = 0-1500 rpm, T = 25 oC. .................................................... 261
Figure 5-63: Effect of rotation rate on corrosion rates of iron RDE in 5-30 wt% MEA + H2O +
CO2 at pH 8.01; 10 mV s-1, f = 0-1500 rpm, T = 25 oC. ........................................................ 261
Figure 5-64: Effect of oxygen on the polarisation curves obtained on iron RDE in 30wt% MEA
+ H2O + CO2 at pH 7.24-8.01 and T = 25-80oC; 10 mV s-1, f = 1500 rpm. .......................... 265
Figure 6-1: Schematic representation of the corrosion process in aqueous MEA-CO2 solutions,
including the chemical reactions occurring at the iron surface.............................................. 275
Figure 6-2: Flow diagram showing the model structure ........................................................ 279
Figure 6-3: Comparison between predicted and measured voltammetric sweep for iron RDE in
30 wt% MEA+ CO2, at pH=8.1 and T = 25 oC; 10 mV s-1, f = 1500 rpm. Black dotted lines
represent the Tafel extrapolation of the anodic and cathodic linear sections to find the
intersection at Ecorr and icorr. ................................................................................................... 290
30 wt% MEA+ CO2, at pH=8.1 and T = 25 oC; 10 mV s-1, f = 960 rpm, 540 rpm and 240 rpm.
................................................................................................................................................ 291
Figure 6-5: Comparisons of experimental and model-predicted corrosion rates as a function of

rotation rate for iron RDE in 30 wt% MEA+ CO2, at pH=8.1 and T = 25 oC; 10 mV s-1. .... 292
XIX
Figure 6-6: Predicted effect of CO2 loading on liquid phase compositions of 30 wt% aqueous
MEA solution at 25 °C, as predicted by the Kent-Eisenberg Model (Kent and Eisenberg, 1976).
................................................................................................................................................ 292
30 wt% MEA+ CO2,at T = 25 oC; 10 mV s-1, f = 1500 rpm as a function of CO2 loading(α).
................................................................................................................................................ 293
Figure 6-8: Comparisons of experimental and model-predicted corrosion rates for iron RDE in
................................................................................................................................................ 294
Figure 6-9: Comparison of model-predicted and experimental current density – electrode

potential data for iron RDE in 30 wt% MEA+ CO2 at pH=8.1 and T = 25 oC; 10 mV s-1,
f = 1500 rpm. Direct reduction of HCO3- included................................................................ 294
Figure 6-10: Figure 6-11: Comparison between predicted and measured voltammetric sweep
for iron RDE in 30 wt% MEA+ CO2, at pH=8.1 10 mV s-1, f = 1500 rpm as a function of
temperature. ........................................................................................................................... 296
Figure 6-11: Comparison of experimental and model-predicted Fe corrosion rates in

30 wt% MEA solution at different temperatures. .................................................................. 296
Figure 6-12: Effect of temperature and CO2 loading on hydrogen ion concentration [H+]
predicted by Kent-Eisenberg model for 30 wt% aqueous MEA solution. ............................. 297
Figure 6-13: Comparison of experimental and model-predicted corrosion rates at different MEA
concentrations, T = 25oC, CO2 loading = 0.5 mol CO2 (mol amine)-1. ................................. 300
Figure 6-14: Dynamic viscosity of 30 wt% MEA at 25 °C and CO2 loading (α =0.5), as a
function of mass fraction (Amundsen et al., 2009). ............................................................... 301
Figure 7-1: Cyclic voltammograms of an iron electrode in 30 wt% MEA + H2O + CO2 at
10 mV s-1 (T = 25oC, pH 8.03) with a linear velocity of 0.5 mm s-1. .................................... 313
Figure 7-2: Effect of potential on chronoamperograms obtained on iron electrode at -

0.80 V ≤ E ≤ -0.20 V in contact with 30 wt% MEA + H2O + CO2 (T = 25oC, pH 8.03) with a
linear velocity of 0.5 mm s-1. ................................................................................................. 315
XX
Figure 7-3: Effect of potential on the charge density-time data for iron electrode at -0.80 V ≤ E
≤ -0.20 V in 30 wt% MEA + H2O + CO2 (T = 25oC, pH 8.03) with a linear velocity of 0.5 mm s-
1
. ............................................................................................................................................. 316
Figure 7-4: Effect of potential on FeII concentration-time data for iron electrode at -
0.80 V ≤ E ≤ -0.20 V in contact with 30 wt% MEA + H2O + CO2 (T = 25 C, pH 8.03) with a °
linear velocity of 0.5 mm s-1. ................................................................................................. 317
Figure 7-5: Images of iron electrode in 30 wt% MEA + H2O + CO2 at pH 8.03 and T = 25oC,)
with a solution velocity of 0.5 mm s-1 (Re = 1.69) after constant electrolyses at -0.70 V, -0.60 V,
-0.50 V & -0.40 V. ................................................................................................................. 318
Figure 7-6: Dependences of Sherwood number on Reynold number and distance from inlet of
the flow reactor based on the reactor geometry, solution flowrate 0.01 - 0.1 m s-1 and MEA
solution viscosity (3.22×10-6 m2 s-1). ..................................................................................... 320
Figure 7-7: Cyclic voltammograms of an iron electrode in contact with 50 mol m-3 of
K3Fe(CN)6 / 100 mol m-3 of K4Fe(CN)6 and 1000 mol m-3 of Na2CO3 at pH 11.3 with a flow
rate of 0.1 mm s-1. .................................................................................................................. 322
Figure 7-8: Mass transfer coefficient versus linear velocity for Fe(CN)63- measured during
hexacyanoferrate(III) reduction and the corresponding values calculated for FeII with
comparisons to the calculations based on mass transport correlations. ................................. 324
Figure 7-9: Effect of solution velocity on the diffusion layer thickness for FeII measured by the
reduction of hexacyanoferrate(III) with comparisons to the calculations based on mass transport
correlations............................................................................................................................. 324
Figure 7-10: Effect of solution velocity on chronoamperometric data for iron electrode in 30
wt% MEA + H2O + CO2 at pH 8.03 and T = 25 °C,); -0.60 V (AgCl|Ag)............................ 325
Figure 7-11: Idealised Nernst diffusion layer and concentration profile for FeII ions required for
FeCO3 and Fe(OH)2 precipitation. ......................................................................................... 326
Figure 7-12: Effect of solution velocity on charge density-time data for iron electrode in 30
wt% MEA + H2O + CO2 at pH 8.03 and T = 25 oC, at -0.60 V (AgCl|Ag). ......................... 328
Figure 7-13: Effect of solution velocity on time dependence of mass of FeII dissolved from iron
electrode in 30 wt% MEA + H2O + CO2 at pH 8.03 and T = 25 oC; -0.60 V (AgCl|Ag). .... 329
XXI
Figure 7-14: Effect of temperature on current density – time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 oC, pH 8.03) with a velocity of 0.5 mm
s-1. ........................................................................................................................................... 331
Figure 7-15: Effect of temperature on the charge density-time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25–80 oC, pH 8.03) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 332
Figure 7-16: Temperature effect on FeII concentration-time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 C, pH 8.03) with a solution velocity of 0.5
°
mm s-1. ................................................................................................................................... 334
Figure 7-17: Effect of temperature on the FeCO3 precipitation rate at iron electrode in contact
with 30 wt% MEA + H2O + CO2 (T = 25 C, pH 8.03) with a solution velocity of 0.5 mm s-1
°
after constant potential electrolyses at -0.60 V (AgCl|Ag). ................................................... 335
Figure 7-18: Images of iron electrode in contact with 30 wt% MEA + H2O + CO2 (T = 25 – 80
°
C, pH 8.03) with a solution velocity of 0. 5 mm s-1 after constant electrolyses at -0.60 V
(AgCl|Ag). ............................................................................................................................. 335
Figure 7-19: Chronoamperometric data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA
+ H2O (T = 25-80 °C, pH 12.31, α = 0) with a solution velocity of 0.5 mm s-1. ................... 339
Figure 7-20: Charge density-time data for iron electrode at -0.60 V (AgCl|Ag) in contact with
30 wt% MEA + H2O (T = 25-80 °C, pH 12.31, α = 0) with a solution velocity of 0.5 mm s-1.
................................................................................................................................................ 339
Figure 7-21: Corrosion-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA +
H2O (T = 25-80 °C, pH 12.31, α = 0) with a solution velocity of 0.5 mm s-1. ...................... 339
Figure 7-22: Chronoamperometric data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA
+ H2O (T = 25-80 °C, pH 10.70, α = 0.2) with a solution velocity of 0.5 mm s-1. ................ 340
Figure 7-23: Charge density-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA
+ H2O (T = 25-80 °C, pH 10.70, α = 0.2) with a solution velocity of 0.5 mm s-1. ................ 340
Figure 7-24: FeII concentration-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt%
MEA + H2O (T = 25-80 °C, pH 10.70, α = 0.2) with a solution velocity of 0. 5 mm s-1. ..... 340
XXII
Figure 7-25: Chronoamperogram for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA +
H2O (T = 25-80 °C, pH 9.90, α = 0) with a solution velocity of 0.5 mm s-1. ........................ 341
Figure 7-26: Charge density-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA
+ H2O (T = 25-80°C, pH 9.90, α = 0.4) with a solution velocity of 0.5 mm s-1. ................... 341
MEA + H2O (T = 25-80 °C, pH 9.90, α = 0.4) with a solution velocity of 0.5 mm s-1. ........ 341
Figure 7-28: Effect of CO2 loading on the FeII concentration-time data for iron electrode at -
0.60 V (AgCl|Ag) in contact with 30 wt% MEA + H2O (pH 8.03-12.3, α = 0-0.6) with a solution
velocity of 0.5 mm s-1 for a) 25 oC b) 40 oC c) 60 oC d) 80 oC. ............................................. 342
Figure 7-29: Effect of CO2 loading on current density - time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O (pH 8.03-12.3, α = 0-0.6) with a solution velocity of 0.5
mm s-1 for a) 25 oC b) 40 oC c) 60 oC d) 80 oC. ..................................................................... 343
Figure 7-30: Effect of CO2 loading on charge density-time data for iron electrode at -0.60
mm s-1 for a) 25 oC b) 40 oC c) 60 oC d) 80 oC ...................................................................... 344
Figure 7-31: Effect of CO2 loading and temperature on total dissolved [Fe2+] from iron electrode
in 30 wt% MEA + H2O (T = 25-80 oC, α = 0-0.6) at -0.60 V (AgCl|Ag) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 345
Figure 7-32: Effect of CO2 loading and temperature on total charge density for iron electrode
at -0.60 V (AgCl|Ag) in 30 wt% MEA + H2O (T= 25-80 oC, α = 0-0.6) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 345
Figure 7-33 Effect of CO2 loading on the FeCO3 precipitation rate at the iron electrode in 30
wt% MEA + H2O + CO2 (T = 80 °C, pH 8.03) with a solution velocity of 0.5 mm s-1 after
constant potential electrolyses at -0.60 V (AgCl|Ag). ........................................................... 346
Figure 7-34: Effect of oxygen on current density – time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O (T = 40oC, pH 8.03, α = 0.6) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 347
Figure 7-35: Hydrogen abstraction mechanism for the oxidative degradation of MEA (Petryaev
et al., 1984) ............................................................................................................................ 348
XXIII
Figure 7-36: Effect of CO2 loading on charge-time data for iron electrode at -0.60 V (AgCl|Ag)
in 30 wt% MEA + H2O (T= 40 oC, pH 8.03, α = 0.6) with a solution velocity of 0.5 mm s-1
................................................................................................................................................ 349
Figure 7-37: Effect of CO2 loading on FeII concentration - time data for iron electrode at -0.60
V in 30 wt% MEA + H2O (T = 40 oC, pH 8.03, α = 0.6) with a solution velocity of 0.5 mm s-1
................................................................................................................................................ 349
Figure 7-38: Effect of temperature on current density –time data for iron electrode at -
0.60 V (AgCl|Ag) in 5 wt% MEA + H2O + CO2 (T = 25 - 80oC, pH 8.03) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 352
Figure 7-39: Effect of temperature on charge density-time data for iron electrode at -
of 0.5 mm s-1. ......................................................................................................................... 352
Figure 7-40: Temperature effect on FeII concentration-time data for iron electrode at -
of 0.5 mm s-1. ......................................................................................................................... 352
Figure 7-41: Effect of temperature on current density - time for iron electrode at -
0.60 V (AgCl|Ag) in 10 wt% MEA + H2O + CO2 (T = 25 - 80oC, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 353
Figure 7-42: Effect of temperature on the charge density-time data for iron electrode at -
0.6 V (AgCl|Ag) in 10wt% MEA + H2O + CO2 (T = 25-80oC, pH 8.03) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 353
0.6 V (AgCl|Ag) in 10 wt% MEA + H2O + CO2 (T = 25 – 80 °C, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 353
Figure 7-44: Effect of temperature on current density-time data for iron electrode at -
velocity of 0.5 mm s-1. ........................................................................................................... 354
0.6 V (AgCl|Ag) in 20 wt% MEA + H2O + CO2 (T = 25 - 80 °C, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 354
XXIV
velocity of 0.5 mm s-1. ........................................................................................................... 354
0.6 V (AgCl|Ag) in 40 wt% MEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 355
Figure 7-48: Effect of temperature on charge density-time data for iron electrode -
0.6 V (AgCl|Ag) in 40 wt% MEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 355
Figure 7-49: Temperature effect on the FeII concentration-time data for iron electrode at -
0.6 V (AgCl|Ag) in 40 wt% MEA + H2O + CO2 (T =25-80 C, pH 8.03) with a solution velocity °
of 0.5 mm s-1. ......................................................................................................................... 355
Figure 7-50: Effect of temperature on current density - time data for iron electrode at -
0.6 V (AgCl|Ag) in 50 wt% MEA + H2O + CO2 (T = 25-80 C, pH 8.03) with a solution velocity °
of 0.5 mm s-1. ......................................................................................................................... 356
of 0.5 mm s-1. ......................................................................................................................... 356
of 0.5 mm s-1. ......................................................................................................................... 357
Figure 7-53: Effect of [MEA] on FeII concentration-time data for iron electrode at -
0.6 V (AgCl|Ag) in 5-50wt% MEA + H2O + CO2 (pH 8.03) with a solution velocity of 0.5 mm
s-1 . .......................................................................................................................................... 358
Figure 7-54: Effect of [MEA] on current density - time data for iron electrode at -
0.6 V (AgCl|Ag) in 5 – 50 wt% MEA + H2O + CO2 (pH 8.03) with a solution velocity of 0.5
mm s-1. ................................................................................................................................... 359
Figure 7-55: Effect of [MEA] on the charge density-time data for iron electrode at -0.6 V
(AgCl|Ag) in 5 - 50wt% MEA + H2O + CO2 (pH 8.03) with a solution velocity of 0.5 mm s-1.
................................................................................................................................................ 360
XXV
Figure 7-56: Potential-pH diagram for Fe-Cr-Ni-H2O system at 298 K, 0.1 MPa and dissolved
activities = 10-4....................................................................................................................... 362
Figure 7-57: Effect of temperature on current density – time data for carbon steel S355 electrode
at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 - 80oC, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 363
Figure 7-58: Effect of temperature on charge density – time data for carbon steel S355 electrode
at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 oC, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 363
Figure 7-59: Temperature effect on the dissolved FeII concentration-time data for carbon steel
S355 electrode at -0.6 V (AgCl|Ag) in 30wt% MEA + H2O + CO2 (T = 25-80 C, pH 8.03) with °
a solution velocity of 0.5 mm s-1............................................................................................ 363
Figure 7-60: Effect of temperature on current density – time data for carbon steel EN8 electrode
velocity of 0.5 mm s-1. ........................................................................................................... 364
Figure 7-61: Effect of temperature on charge density – time data for carbon steel EN8 electrode
velocity of 0. 5 mm s-1. .......................................................................................................... 364
Figure 7-62: Temperature effect on FeII concentration-time data for carbon steel EN8 electrode
at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 364
Figure 7-63: Effect of temperature on current density - time data for stainless steel electrode at
-0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 - 80oC, pH 8.03) with a solution
velocity of 0.5 mm s-1. ........................................................................................................... 365
Figure 7-64: Effect of temperature on charge density – time data for stainless steel electrode at
-0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 C, pH 8.03) with a solution velocity
°
of 0.5 mm s-1. ......................................................................................................................... 365
Figure 7-65: Temperature effect on dissolved FeII concentration - time data for stainless steel
electrode at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 - 80oC, pH 8.03) with a
solution velocity of 0.5 mm s-1. ............................................................................................. 365
XXVI
Figure 7-66: Effect of steel composition on the corrosion-time curve obtained on iron electrode
at -0.6 V (AgCl|Ag) in contact with 30 wt% MEA + H2O + CO2 (pH 8.03) with a solution
velocity of 0.5 mm s-1. ............................................................................................................ 366
Figure 7-67: Effect of temperature on current density – time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% MDEA + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity
of 0.5 mm s-1. ......................................................................................................................... 370
Figure 7-68: Effect of temperature on charge density - time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% MDEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 370
Figure 7-69: Temperature effect on FeII concentration - time data for iron electrode at -0.6 V
(AgCl|Ag) in 30wt% MDEA + H2O + CO2 (T = 25oC, pH 8.03) with a solution velocity of 0.5
mm s-1. ................................................................................................................................... 370
Figure 7-70: Effect of temperature on current density-time data of iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% AEP + H2O + CO2 (T = 25 - 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 371
Figure 7-71: Effect of temperature on the charge density-time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% AEP + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 371
(AgCl|Ag) in 30 wt% AEP + H2O + CO2 (T = 25 oC, pH 8.03) with a solution velocity of 0.5
mm s-1. ................................................................................................................................... 371
(AgCl|Ag) in 30wt% AMP + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 372
(AgCl|Ag) in 30 wt% AMP + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 372
(AgCl|Ag) in 30 wt% AMP + H2O + CO2 (T = 25 oC, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 372
XXVII
Figure 7-76: Effect of amine type on current density – time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% [Amine] (T = 25-80 oC, pH 8.03) with a solution velocity of 0.5 mm s-1.
................................................................................................................................................ 373
Figure 7-77: Effect of amine type on FeII concentration-time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% [Amine] (T = 25 - 80oC, pH 8.03) with a solution velocity of 0.5 mm s-1.
................................................................................................................................................ 374
Figure 7-78: : Effect of amine type on charge density-time data for iron electrode at -0.6 V
(AgCl|Ag) in contact with 30 wt% [Amine] (T =25 - 80oC, pH 8.03) with a solution velocity of
0.5 mm s-1. ............................................................................................................................. 375
Figure 7-79: Time dependence of open circuit potentials of iron and steel electrodes at -0.6 V
(AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 oC, pH 8.03) with a solution velocity of 0.5
mm s-1. ................................................................................................................................... 376
Figure 7-80: Time-dependence of iron(II) concentrations dissolved from iron and steel S355
electrodes at -0.6 V (AgCl|Ag) in contact with 30 wt% MEA + H2O + CO2 (T = 25 oC, pH 8.03)
with a solution velocity of 0.5 mm s-1.................................................................................... 377
Figure 7-81: Images of the flow reactor system for the longer term studies at open circuit
potential for a) 7 days b) 14 days c) 28 days ......................................................................... 378
Figure 7-82: Images of the iron and carbon steel S355 electrodes after immersion in 30 wt%
MEA + H2O + CO2 (T = 25oC, pH 8.03) with a solution velocity of 0.5 mm s-1 at open circuit
potential for 28 days............................................................................................................... 379
rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 60 °C; 10 mV s-1. .................... 399
XXVIII
Figure 9-9: Effect of NaHCO3 + Na2CO3 concentration on cyclic voltammograms of an iron

Figure 9-10: Effect of NaHCO3 concentration on current density for peak II of stationary Fe
electrode in NaHCO3 + Na2CO3 solutions (T = 25 °C, pH 8.01). ......................................... 401

XXIX

List of tables
Table 1: List of Symbols ................................................................................................. XXXIII
Table 2: List of Abbreviations ........................................................................................ XXXIV
Table 3: List of subscripts and superscripts .................................................................... XXXIV
Table 2-1: Modes of mass transport (Walsh, 1993) ................................................................. 37
Table 3-1: Chemical Structures of the amines commonly used in CO2 capture ...................... 56
Table 3-2: Summary of Amine Scrubbing Plant Experience on Corrosion............................. 66
Table 3-3: Summary of Amine Scrubbing Plant Experience on Corrosion (Continued) ........ 67
Table 3-4: growth kinetics of FeCO3 , 2σ = 95% confidence level ....................................... 101
Table 3-5: Standard State Formation Properties of FeCO3 at 298.15 K and 0.1 MPa........... 104
Table 3-6: Types of corrosion films formed from 30-150oC in steel-CO2-H2O systems ...... 110
Table 3-7: Equilibrium constants used in the Kent-Eisenberg model (Kent and Eisenberg,
1976). ..................................................................................................................................... 119
Table 3-8: Compilation of diagnostic criteria for Fe dissolution and deposition reactions
obtained under different experimental conditions (Drazic, 1989). ........................................ 122
Table 4-1: Common reagents ................................................................................................. 157
Table 4-2: Standard reagents ................................................................................................. 158
Table 4-3: Electrode Materials............................................................................................... 159
Table 4-4: Cation exchange membrane ................................................................................. 160
Table 5-1: Effects of rotation and potential scan rates on anodic charge / C m-2 passed with
Fe RDE in 30 wt% MEA + H2O + CO2 (T = 25°C, pH 8.01). 207
Table 5-2: Effect of temperature on the peak (II) electrode potential after taking into account
the shift in electrode potential of AgCl/Ag with respect to standard hydrogen electrode (SHE)
................................................................................................................................................ 211
XXX

Table 5-3: Activation energy associated with jpII for Fe RDE in 30 wt% MEA + H2O + CO2,
at 10 mV s-1 (pH 8.01) for varying aqueous solution temperatures (18 oC ≤ T ≤ 80 oC) and
rotation speeds (60 rpm ≤ f ≤ 1500 rpm) ............................................................................... 213
Table 5-4: Effect of solution temperature on the pH for 30 wt% MEA at full CO2 saturation.
................................................................................................................................................ 251
Table 6-1: Chemical and electrochemical reactions represented in the mechanistic corrosion
model...................................................................................................................................... 276
Table 6-2: Equilibrium expressions for chemical reactions ................................................. 281
Table 6-3: Temperature dependent equilibrium constants for chemical reactions 1-5 .......... 282
Table 6-4: Diffusion Coefficients for species in water at 25oC ............................................. 284
Table 6-5: Liquid properties as a function of temperature (Haynes and Lide, 2011) ............ 284
Table 6-6: Density ρ/ kg m3 for MEA +H2O from (T= 25- 80oC) as a function of MEA mass
fraction w (Amundsen et al., 2009). ....................................................................................... 285
Table 6-7: Density ρ / kg m3 for 30 wt% MEA +H2O from (T= 25- 80oC) as a function of
temperature and CO2 loading (α =0.1 - 0.5) (Amundsen et al., 2009). .................................. 285
Table 6-8: Dynamic viscosity µ / mPa s-1 for MEA +H2O from (T= 25- 80oC) as a function of
MEA mass fraction w (Amundsen et al., 2009). ................................................................... 285
Table 6-9: Dynamic viscosity µ / mPa s-1 for 30 wt% MEA +H2O from (T= 25- 80oC) as a
function of temperature and CO2 loading (α =0.1 - 0.5) (Amundsen et al., 2009) ................ 286
Table 6-10: Kinetic parameters of the electrochemical reactions .......................................... 289
Table 6-11: Predicted species concentration / kmol m-3 for 30 wt% MEA-H2O-CO2 systems
at 25oC.................................................................................................................................... 298
at 40oC.................................................................................................................................... 298
at 60oC.................................................................................................................................... 299
at 80oC.................................................................................................................................... 299
XXXI

Table 6-15: Predicted species concentration / M for 20 wt% MEA-H2O-CO2 systems at 25oC
................................................................................................................................................ 302
................................................................................................................................................ 302
................................................................................................................................................ 303
Table 7-1: Effect of solution velocity on the conditions required for FeCO3 and Fe(OH)2
precipitation by determining the [FeII] at the iron electrode surface, equilibrium [CO32- ] from
the Kent Eisenberg model and the results from iron electrode in 30 wt% MEA + H2O + CO2
(T = 25 oC, pH 8.03) at -0.60 V (AgCl|Ag). .......................................................................... 329
Table 7-2: Effect of temperature on the conditions required for FeCO3 precipitation by
determining the [FeII] at the iron electrode surface, equilibrium [CO32- ] from the Kent
Eisenberg model and the results from iron electrode in contact with 30 wt% MEA + H2O +
CO2 (T = 25 °C, pH 8.03) at -0.60 V. .................................................................................... 336
Table 7-3: Effect of temperature on the solubility product, Ksp (calculated from
thermodynamic data in (Benezeth et al., 2009)) .................................................................... 336
Table 9-1: Solubility data in a solution of MEA (30 mass %) and water (70 mass %) at T =
25oC from (Jou et al., 1995) ................................................................................................... 394
40oC from (Jou et al., 1995) and (Tong et al., 2012) ............................................................. 394
60oC from (Jou et al., 1995) ................................................................................................... 395
80oC from (Jou et al., 1995) ................................................................................................... 395
Table 9-5: Effects on jPII of rotation rate, temperature (25-80 °C) and aqueous MEA
concentration (10-60 wt%) for Fe RDE in MEA + H2O + CO2 at pH ̴ 8.01; 10 mV s-1........ 396
Table 9-6: Summary of kinetic parameters obtained from the polarization curve of iron RDE
rotating at f = 1500 rpm in 10 wt% MEA+ CO2 at pH ̴ 8.01; 10 mV s-1 ............................... 402
XXXII

Table 9-10: Summary of kinetic parameters obtained from the polarization curve of iron
RDE rotating at f = 1500 rpm in 50 wt% MEA+ CO2 at pH ̴ 8.01; 10 mV s-1 ...................... 405
RDE rotating at f = 1500 rpm in 60 wt% MEA+ CO2 at pH ̴ 8.01; 10 mV s-1 ...................... 405
Table 9-12: Thermodynamic properties of species in Fe-CO2-H2O systems at 298K (Bard et
al., 1985). ............................................................................................................................... 406
Table 9-13: Specific Heat Capacity Coefficients(HSC Software Database, 2016) ............... 407

XLI
Nomenclature
Figure 1: List of Symbols

Symbols Definition Units
ai Activity of species i
A Area m2
ci Concentration of species i mol m-3
Di Diffusion coefficient of species i m2 s-1
de Equivalent hydraulic diameter m
E Electrode potential V
Eº Standard electrode potential V
EO/R Reversible potential V
f Frequency Hz
F Faraday constant C mol-1
Gº Standard Gibbs free energy of formation J mol-1
ΔG Gibbs free energy of change J mol-1
ΔH Enthalpy change J mol-1
I Current A
IL Mass transport limited current A
j Current density A m-2
j0 Exchange current density A m-2
jL Mass transport limited current density A m-2
kb Boltzmann constant J K-1
km Mass transport rate coefficient m s-1
K Equilibrium constant
l Length m
m Mass kg
M Molar mass g mol-1
N Flux mol m-2 s-1
NA Avogadro’s constant mol-1
p Partial pressure of gas mol m-2 s-1
Q Electric charge C
r Rate of electrochemical reaction
R Universal gas constant J mol-1 K-1
∆S Entropy Change J mol-1 K-1
t Time s
T Temperature K
u Linear flow rate m s-1
U Ionic mobility m2 V-1 s-1
U Cell voltage V
V Volume m3
! Volumetric flow rate m3 s-1
w Width m
zi Electromical charge of species i
α Electrochemical reaction transfer coefficient
XLII
β Tafel coefficient V-1

β Tafel slope mV-1 dec-1
ẟN Nernst diffusion layer thickness m
" Conductivity S m-1
µ Dynamic viscosity Pa s
η Overpotential V
# Interfacial potential drop V
ρ Density kg m-3
ν Kinematic viscosity m2 s-1
νe Electron stoichiometry
Φe The charge yield
Figure 2: List of Abbreviations

Abbreviation Definition
O Oxidised species
R Reduced species
e- Electron
CE Counter electrode
WE Working electrode
RE Reference electrode
C Cathode
A Anode
SHE Standard hydrogen electrode
SCE Standard Calomel electrode
OCP Open circuit potential
RDS Rate determining step
CV Cyclic voltammogram
RDE Rotating disc electrode
HER Hydrogen evolution reaction
ICP-OES Inductively coupled plasma optical emission spectroscopy
AEP Aminoethylpiperazine (AEP)
AMP 2- amino- 2- methyl-1-propanol
DEA Diethanolamine
DMMEA 2-Dimethylaminoethanol
MEA Monoethanolamine
MDEA N-Methyldiethanolamine
Figure 3: List of Subscripts and Superscripts

Subscript/ Superscript Definition
a Anodic
c Cathodic
Ads Adsorbed state
Sol Solute state
b Bulk value
s Surface value
x= 0 Value at the electrode surface
Page |1
1. Chapter 1
Introduction
1.1. Motivation for carbon capture and sequestration
The debate over the possibly devastating phenomenon of climate change, the classification of
CO2 as the primary greenhouse gas (GHG) and a major trigger for rising global average
temperatures is considered to have reached a consensus (IPCC, 2001). The increase in global
average temperature since the mid-20th century has been attributed to the observed increase in
anthropogenic green house gasses, in particular rising CO2 concentrations (IPCC, 2007). The
consequences of such changes could be dire and lead to the collapse of ecosystems, adverse
weather conditions and the extinction of 15 to 40 percent of all species (Figueroa et al., 2008).
Fossil fuels currently supply 85% of the world’s energy needs due to their low cost,
availability, and most importantly the existing reliable technology for electrical power
production (Strazisar et al., 2003); however, the combustion of carbonaceous fuels leads to
the production of CO2. In fact, it is estimated that 83% of GHG emissions are produced from
Page |2
the combustion of fossil fuels, with CO2 having the most negative impact (66%). The
consumption of fossil fuels in the U.S for example, is expected to increase by 27% over the
next 20 years, thereby increasing the CO2 emissions from the current 6 Gt a-1 to 8 Gt a-1 by
2030 (Figueroa et al., 2008). In particular, coal-fired power stations emit twice as much CO2
per kW h in comparison to natural gas plants. However, coal will remain the primary source
of energy until at least 2030, due to its abundance and wide geological distribution (IEA,
2008), especially in rapidly growing economies such as China and India.
However, it has been considered that the emissions of CO2 should be vastly reduced and the
Intergovernmental Panel on Climate Change (IPCC) has recommended a 50 to 85 percent
reduction of global GHG emissions from 2000 to 2050 (IPCC, 2001). In the Energy
Technology Perspectives publication by the International Energy Agency (IEA), (IEA, 2008),
a baseline scenario and a BLUE map scenario for the world’s energy system from 2012 to
2050 were created. The baseline scenario which makes an assumption of no new energy or
climate change policies beyond the existing policies results in a doubling of CO2 emissions in
2050 compared to 2007. On the other hand, in the BLUE map scenario when effective
policies and technology advancements were predicted and taken into consideration, the
anthropogenic emissions of CO2 in 2050 could be cut to half of the 2007 levels. Achieving
these optimistic goals would limit the global temperature increase to between 2oC and 3oC
above pre-industrial levels by 2100 (IPCC 2001; Michaels and Balling, 2009).
Taking these facts into consideration, it has become a matter of extreme urgency to develop
technologies to tackle the issue of carbon emissions and decarbonise the global energy
system. Effecting a more than 50% reduction in the 2050 CO2 emissions from the baseline
scenario of 57 Gt to the BLUE Map scenario requires substantial efforts to be made globally
across almost every energy related sector.

Page |3
Carbon Capture and Sequestration (CCS) is one of the many technologies being developed
currently to achieve these ambitious emission reduction targets and is expected to contribute
19% of the total predicted emission cuts, 10% resulting from CCS in power generation and
9% from CCS in industry and transformation (IPCC, 2001). CCS involves capturing CO2
from large point source such as power stations, compressing and transporting by pipeline or
tankers to an injection site and sequestrating for long-term storage in underground geological
formations. It is currently the most probable approach for combating CO2 emissions, while
still allowing the continued use of fossil fuels. CCS could reduce CO2 emissions from power
stations and industrial processes by more than 85% (Michaels and Balling, 2009; IPCC,
2001). Without the implementation of CCS, the overall costs to reduce CO2 emissions could
increase by up to 70% (IEA, 2008). However, the BLUE Map scenario is particularly
ambitious and requires over 3000 projects with a predicted expenditure of US$ 2.5 - 3 trillion
by 2050 (IEA, 2008).
The capture of CO2 could be beneficial not only to the environment but economically, as it
could potentially be used for by-product end use in the industrial and energy production
sectors. While totally avoiding the emission of carbon has been considered via the use of
renewable energy technologies, it will be extremely difficult to achieve the targeted
atmospheric CO2 of below 450ppm without employing CCS technologies (IEA, 2008).
On an individual basis, each of the stages of CCS namely capture, compression, transport and
storage are technically viable; however, there have been several challenges associated with
the integration and scale-up of these technologies. According to the Global CCS Institute in
its 2015 update report, there are at present 15 large scale integrated CCS projects operating
globally, 7 that are under currently under construction and a further 22 at various stages of
development planning (Global CCS Institute, 2015). The total CO2 capture capacity of the 22
projects in operation or under construction is more than 40 Mt a-1. According to the report 2
Page |4
of the large scale CCS projects, The Quest Project, Alberta, Canada and The Uthmaniyah
CO2-EOR Demonstration Project, Saudi Arabia became operational in 2015 and 2 more are
expected to begin operation in 2016. The almost doubling in large scale CCS projects since
the start of 2010 is a strong indication in the determination of governments and energy-related
industries to make significant advancements in CCS technology at a commercial scale.
CO2 capture can be classified broadly into three main technologies: post-combustion, pre-
combustion capture and oxy-fuel combustion as shown schematically Figure 1-1. Post
combustion capture entails the CO2 being separated from the flue gas produced by
combustion in power plants. Power plants currently in existence use air for combustion and
produce a flue gas which has a CO2 concentration of less than 15% (Michaels and Balling,
2009; IPCC, 2001). Pre-combustion capture involves the separation of CO2 from fuel prior to
combustion. The fuel is first gasified to produce H2 and CO2 before the H2 is separated and
combusted. This involves higher concentrations of CO2 than the post combustion method
facilitating the separation. Oxy-fuel combustion, which is in demonstration phase, involves
the combustion of the fuel in pure oxygen instead of air, causing the flue gas to contain
mostly CO2 and steam which can be separated easily.
In the case of diluted and low CO2 pressure streams such as flue gases from power plants, the
current technology of choice for CO2 capture is the post-combustion capture method, based
on the use of amine solvents to absorb CO2 reversibly. This method of separating CO2 from
flue gas streams was first developed in the 1970s, not to tackle the greenhouse effect but for
the potential economic benefits such as enhanced oil recovery and improving the quality of
natural gas. It is already well established in the natural gas ‘sweetening’ (i.e. H2S removal),
ammonia production and hydrogen manufacture industries. An additional benefit of this
method is that it can be retrofitted to existing power plants. In fact, it is currently the only
Page |5
proven technology that is suited to the huge investment that has been made in fossil fuel
infrastructure.
Figure 1-1: Post-combustion, pre-combustion and oxy-fuel combustion CO2 capture systems
(Figueroa et al., 2008).
This method of CO2 capture is installed downstream of the power plant, so if the capture
system encounters any issues, the power plant can continue to operate uninterruptedly.
Employing amine-based carbon capture techniques will not decrease CO2 emissions only by
85 to 90 percent, but also remove considerable amounts of other pollutants such as NOx and
SO2 due to prior treatment of the flue gas (Michaels and Balling, 2009; IPCC, 2001). This
method involves a reversible reaction between the amine and CO2 which, under ideal
conditions, would allow for the recycling of the amine once the CO2 has been released.
Page |6
1.2. Operational difficulties in CO2 absorption process
Although the concept of CO2 capture by amine scrubbing involves proven technology, which
has been used in industry for over 50 years, the process is still plagued with technology risks
and technical barriers. There are numerous challenges hindering the deployment and
successful operation of CCS technology, notably the high energy requirements (parasitic
needs) and costs, which must be overcome before it can become commercially viable on a
large scale. Currently, the cost of utilizing this process has been estimated to run as high as
66-75 % of the total power plant operating cost (McKinsey & Company, 2008). Furthermore,
the process faces several operational difficulties such as degradation of the amines, pre-
precipitation, corrosion of process equipment, solvent loss, fouling, increased viscosity of
amine and foaming.
The amine solutions regularly become contaminated with degradation products, heat stable
salts, heavy hydrocarbons and particulates (Abdi and Meisen, 2000). These degradation
products are a result of irreversible side reactions with CO2 and other flue gas components. As
the amine process usually involves cycles of absorption and desorption which permit the
recycling of the absorbent, the closed loop nature of this process leads to an accumulation of
the degradation products which can yield a significant decrease in the absorption capacity of
the amines. These large volumes of spent solution have to be disposed of by incineration or
deep well injection. For instance, on a plant which uses monoethanolamine (MEA), it has
been estimated that 2.2 kg of MEA must be replaced per tonne of CO2 captured (Strazisar et
al., 2003). However, this practice is undesirable due to the degradation products being toxic
and represent several environmental and health risks, while it also amounts to a loss of
valuable amine, as spent solutions contain about 85 % of virgin amine (Abdi and Meisen,
2000).
Page |7
Amongst the numerous difficulties facing the use of alkaonolamines for CO2 absorption,
corrosion is the most serious operating problem, so it requires and has received the most
attention. Several inconclusive theories have been proposed to explain the corrosion
mechanisms, patents have emerged to alleviate the issue and a few papers have been
published on the subject.
1.3. Corrosion in Amine Scrubbing Plants
Based on substantial information from a large number of amine treating plants, corrosion has
been found to occur in several sections of the plant, such as the bottom of the absorber, the
heat exchanger, the entire stripper and the overhead condenser. Corrosion of the carbon steel
equipment occurs in both the form of uniform corrosion and localised attack such as pitting,
galvanic, stress cracking and inter-granular corrosion (Veawab et al., 1999). While uniform
corrosion results in a more significant loss of the carbon steel, it is less likely to result in
catastrophic process failures, due to the relative ease in detection and monitoring. On the
contrary, localised corrosion consumes less material but results in more calamitous failures as
it can easily go undetected. Chronic corrosion can lead to significant economic losses as it
results in unplanned downtime, production losses, reduced life of the equipment and even
injury or death (Veawab et al., 1999); Hawkes & Mago, (1971) estimated that the economic
consequence of unplanned downtime of a typical plant can cost up to $30,000 per day in
terms of production losses. In addition to losses due to unplanned downtime, there is also the
added expenditure of restoring the corroded units and treatments used to mitigate corrosion,
the costs of which can run up to millions of pounds yearly (Veawab et al., 1999). Leaking
units, tanks and pipelines result in significant loss of products, which are not only associated
with a high cost, but could also be hazardous to the surrounding environment. An
accumulation of corrosion products in units such as the heat exchanger tubing and pipelines
could decrease efficiencies and reduce pumping capacity (Jones, 1996). The most devastating
Page |8
consequence of corrosion is the threat to the safety of plant personnel. According to Craig and
McLaughlin (1996), seventeen workers on a refinery plant in Lemont, Illinois were killed in
an explosion in an MEA absorber caused by a leakage of flammable gasses from a non-stress-
relieved weld repair to the vessel.
In addition to the economic penalties discussed above, corrosion can also affect the economy
of the plant in a more indirect manner by limiting the operating conditions of the process. It is
a generally accepted practice that low concentrations of amines have to be used to keep
corrosion rates to acceptable levels and operating beyond a typical concentration (10-20 wt%
monoethanolamine) can cause a significant increase in the system’s corrosiveness (Kohl &
Riesenfeld, 1985; DuPart et al., 1993; Veawab et al., 1999). This decreases the effectiveness
of the solvent, bringing about the need for bigger equipment, faster circulation rates and larger
energy requirements (Strazisar et al., 2003, Strazisar et al., 2002). The extra energy
requirement would in turn increase the parasitic load of the plant, leading to higher fuel
consumption, costs and ironically CO2 emissions. Employing this method almost completely
negates the benefits of sequestration (Strazisar et al., 2002). However, operating the process
beyond these conditions may also increase the system’s corrosiveness; hence, the capacity of
existing plants may not be easily increased.
1.4. Existing Corrosion Control
There are several approaches which have been considered strongly for corrosion control on
amine scrubbing plants: i) appropriate equipment design and operating conditions, ii) use of
more corrosion-resistant alloys of steel instead of carbon steel, iii) the frequent and
intermittent removal of solid contaminants from the amine solutions, and iv) use of corrosion
inhibitors (Kohl & Riesenfeld, 1985; Veawab et al., 1999). While the use of corrosion
inhibitors has been recommended and even patented as the most economical and easy to
Page |9
implement technique for corrosion control, it has not been employed successfully in amine
units (Kohl & Riesenfeld, 1985). This could be due to the most effective corrosion inhibitors
currently available containing toxic inorganic substances such as vanadium compounds,
which make for difficult disposal and pose serious environmental concerns.
1.5. Limitation to Available Corrosion Data
While it is apparent that CO2 capture based on the use of amines can be improved by reducing
or completely eliminating the excessive corrosion issues associated with the process, it has
been difficult to establish the mechanisms under which it occurs and in turn impossible to
recommend a long-lasting solution. This is in part, due to the limitations in the current
knowledge of the corrosion phenomena. There is a limited amount of research data on
corrosion in amine plants available in literature. Corrosion rates were measured using weight
loss or rudimentary electrochemical techniques to establish the effect of operating conditions
such as amine concentration, CO2 loading and temperature on corrosion rate; those rates vary
significantly depending on the source and the testing environments under which they were
determined. The available data are sparse and in some cases are uncorrelated because they
were measured by different sources under either laboratory or field conditions using as
aforementioned different corrosion measuring techniques. As such, the information available
in open literature has been useful only for describing the corrosion behaviour qualitatively and
establishing some common trends between the operating parameters and corrosion rates. Most
of the data collected so far have also been reported with an incomplete list of crucial operating
conditions, making it difficult to derive a complete understanding of the role of process
parameters on the corrosion phenomenon. Therefore, to be able to establish further
understanding of the corrosion behaviour, it was essential for a comprehensive set of data to
be generated under well-controlled conditions, similar to those found in amine absorption
plants.
P a g e | 10
A number of corrosion mechanisms have been proposed to explain the corrosion in amine
plants (DuPart et al., 1993, Tomoe et al., 1996) and it has been hypothesized that the
oxidising agents driving the corrosion reactions are the same as those commonly present in
CO2-water systems, i.e. hydrogen ion (H+),) and bicarbonate ion (HCO3-). However, it has yet
to be established whether this is indeed the case in amine-H2O-CO2 systems, as such the exact
mechanisms of corrosion in such systems are little understood and remain undefined.
1.6. Project Aims
An ideal amine plant would be able to achieve a high degree of CO2 capture at a minimal
financial and environmental cost. These concerns about efficiency, parasitic needs and the
economics of running amine scrubbing plants have made decreasing solvent degradation and
corrosion rates a matter of high priority. As the knowledge of corrosion in amine-H2O-CO2
systems is currently sparse and inconclusive, hitherto corrosion prevention and mitigation has
been difficult to achieve. A thorough understanding of the mechanisms of the corrosion of
carbon steel in the solvent would constitute significant progress towards achieving this goal.
Thus, this thesis aimed to extend and bridge the gaps in the current knowledge by determining
the oxidation and reduction kinetics and mechanisms of iron in aqueous MEA-CO2-H2O
systems as functions of experimental variables, with the view to predicting iron corrosion
rates under typical CO2 absorption process conditions.
1.7. Project Scope and Objectives
This project focused on iron as the principal component of carbon steel used as construction
material in the amine scrubbing technology, concentrating on the electrochemical behaviour
and characterisation, both experimentally and via thermodynamic and electrochemical
modelling. The project had three objectives:

P a g e | 11
1. Quantification of the electrochemical kinetics and mechanisms of iron oxidation and
hence corrosion rates, in aqueous MEA-CO2-H2O systems, as a function of experimental
variables.
2. A mechanistic model for iron oxidation kinetics in aqueous MEA-H2O-CO2.
3. A description of the dependence of iron(II) dissolution rates on experimental variables,
using a specially-designed electrochemical flow reactor with subsequent analysis of
dissolved iron(II) concentrations, enabling deconvolution of total oxidation current
densities in the small scale electrochemical experiments, into rates of reactions leading to
adsorbed and dissolved products.
1.7.1. Project Tasks
The following tasks were defined in order to achieve the aims and objectives of this project:
1. Characterisation of the electrochemical behaviour of iron in MEA-CO2-H2O by cyclic
voltammetry with the use of a rotating disk electrode (RDE) and electrochemical
quartz crystal microbalance (EQCM), thus permitting the following:
a) Determination of the reaction mechanisms for iron oxidation in amine systems
in the absence and presence of CO2.
b) Investigation of the effects of process parameters, MEA concentration,
temperature, CO2 loading, pH, mass transport rates and dissolved oxygen
concentration on the electrochemical behaviour of iron.
c) Kinetic analysis and identification of quantitative trends in corrosion rates as
functions of the experimental parameters.
2. Identification of primary solution species in CO2 loaded MEA systems, permitting
analysis of the chemical speciation and equilibrium concentrations of the amine

P a g e | 12
species (RNH2, RNH3+, RNHCOO-) and carbon (IV) species (HCO3-, CO32-) and
hydrogen ions (H+) via a Kent-Eisenberg type model.
3. Determination of unknown electrochemical parameters (e.g. exchange current
densities, Tafel slopes) and their dependence on operating conditions from reliable
data published in the literature.
4. Development of a predictive micro-kinetic model for the uniform corrosion of Fe in
MEA-H2O-CO2 systems based on the proposed corrosion mechanisms from the
voltammetric results taking into account the CO2 absorption equilibria reactions and
the kinetic equations that describe the electrochemical reactions at the iron surface.
5. Design , construction and characterisation of a large scale electrochemical flow reactor
fabricated from PTFE with an iron/steel anode, platinised titanium cathode and a
cation-permeable Nafion membrane.
6. Operation of the flow reactor in order to :
(a) Measure dissolved iron (II) concentrations under typical CO2 absorption
conditions.
(b) Determine the charge yield for dissolution, so enabling discrimination between
current densities leading to dissolution from those resulting in adsorbed
products.
1.8. Thesis outline:
This thesis is organised into 8 main sections. The general introduction and background theory
on electrochemical systems with a focus on corrosion is detailed in Chapter 2. Chapter 3
presents a comprehensive literature survey covering the carbon capture technology and the
types of amines employed in the carbon capture process. The literature review also evaluates
previous experimental work on corrosion in CO2-loaded amine systems and advancements in
iron and steel corrosion in CO2- H2O systems within other technical fields, such as the oil and
P a g e | 13
gas industry. In Chapter 4, the experimental apparatus and procedures for the various
electrochemical and analytical techniques used are described. Chapters 5 to 7 detail the
results from the experimental and electrochemical model based on the objectives outlined
above. Chapter 5 details the electrochemical reaction mechanisms and kinetics analysis of Fe
in MEA solutions based on cyclic voltammetry experiments conducted with the RDE and
EQCM. Based on the proposed corrosion mechanisms, a mechanistic model for the uniform
corrosion of iron in CO2-loaded MEA systems was developed and detailed in Chapter 6,
taking into account the solution speciation and the electrochemical reactions at the iron
interface. Chapter 7 is dedicated to the experimental results and discussion from the flow-
through reactor from which, dissolved iron concentrations were measured after constant
potential electrolyses. Finally, conclusions are summarised and recommendations for future
work are suggested in Chapter 8.

P a g e | 14
1.9. References
Abdi, M. A. & Meisen, A. 2000. Amine degradation: problems, review of research

achievements, recovery techniques. Proceedings of the 2nd International Oil, Gas and
Petrochemical Conference Tehran, Iran.
Craig, H. L. & Mclaughlin, B. D. 1997. Corrosive Amine Characterization. NACE
Corrosion/96, 96394.
Dupart, M. S., Bacon, T. R. & Edwards, D. J. 1993. Understanding corrosion in alkanolamine
gas treating plants Part 1: proper mechanism diagnosis optimizes amine operations.
Hydrocarbon Processing.
gas treating plants Parts 2: case histories show actual plant problems and their
solutions. Hydrocarbon Processing.
Figueroa, J. D., Fout, T., Plasynki, S., Mcilvried, H. & Srivastava, R. D. 2008. Advances in
CO2 capture technology--The U.S. department of energy's carbon sequestration
program. International Journal of Greenhouse Gas Control, 2, 9-20.
Global CCS Insititute. 2015. The Global Status of CCS. December 2015 Update. Onlime.
Australia. Available from http://www.globalccsinstitute.com/publications/global-
status-large-scale-integrated-ccs-projects-december-2015-update
Hawkes, E. N., Mago, F. F., Butwell, K. F. 1973. Corrosion control in CO2 removal systems.
Chem. Eng. Prog., 69(2), 57-61
International Energy Agency. 2009. Technology roadmaps- carbon capture and storage.
France.
International Energy Agency. 2010. World energy outlook 2010 edition. London,
Organisation for Economic Co-operation and Developmemt.
International Energy Agency. 2010. Energy technology perspectives 2010 Edition. London,
Organisation for Economic Co-operation and Developmemt.
IPCC. 2001. Working Group I: The Scientific Basis. Switzerland.
IPCC. 2007. Fourth Assessment Report (AR4) Climate Change 2007: The Physical Science
Basis. UK
Jones, D. A. 1996. Principles and Prevention of Corrosion, Nevada, Prentice-Hall, Inc.
Kohl, A. L. & Riesenfeld, F. C. 1985. Gas Purification (4th edition). Gulf Publishing
Compay, Houston, Texas.
Michaels, P.J., Balling Jr., R.C., 2009. Climate of Extremes: Global Warming Science They
Don’t Want You to Know. Washington, DC, Cato Institute.
Strazisar, B. R., Anderson, R. R. & White, C. M. 2002. Degradation of monoethanolamine
used in CO2 capture from flue gas of a coal-fired electric power generating station.
Abstracts of Papers of the American Chemical Society, 223, 054-FUEL.
Strazisar, B. R., Anderson, R. R. & White, C. M. 2003. Degradation pathways for
monoethanolamine in a CO2 capture facility. Energy & Fuels, 17, 1034-1039.
Tomoe, Y. Shimizu, M. Kaneta, H. Active dissolution and natural passivation of carbon steel
in carbon dioxide loaded alkanolamine solutions. NACE Corrosion, 96, 395.
P a g e | 15
VEAWAB, A., TONTIWACHWUTHIKUL, P. & CHAKMA, A. 1999. Corrosion behavior

of carbon steel in the CO2 absorption process using aqueous amine solutions.
Industrial & Engineering Chemistry Research, 38, 3917-3924.
P a g e | 16
2. Chapter 2
Principles of Electrochemical Systems

for Corrosion Studies
Studies of metallic corrosion can be traced back at least to the end of the 18th century during
the industrial revolution, when references to corrosion phenomena in literature first occurred.
The emergence of modern corrosion science occurred in the twentieth century when (Evans,
1937) proposed a local cell model and (Wagner and Traud, 1938) introduced the mixed
potential corrosion model. These two models have served as the basis for the electrochemical
theory of corrosion, which describes corrosion as coupled electrochemical reactions consisting
of anodic metal oxidation and cathodic oxidant reduction(s) leading to the metal’s disintegration
into its constituent ions.
This chapter addresses the fundamental principles of electrochemistry which provides
background theory on the experimental techniques employed and the analysis of the results
reported in subsequent chapters. This includes the fundamental definitions within
electrochemical systems, thermodynamics and kinetics of electrochemical reactions and
transport processes.
P a g e | 17
Electrochemistry is the study of chemical reactions that occur via the transfer of charge across
the interface of an electrode, which is a metallic or semiconducting solid and an electrolyte
which is a liquid or solid ionic conductor (Walsh, 1993). Two types of charge transfer reactions
occur at the interface: ion transfer reactions from electrode to the electrolyte, or vice versa;
electron transfer reactions between ions or molecules in the electrolyte and an electrode surface.
The general form of electrochemical reactions is expressed in equation (2. 1) where the redox
couple O and R denote the oxidised and reduced species, respectively.
cathodic reduction
O + ne− ←⎯⎯⎯⎯⎯
⎯⎯⎯⎯⎯⎯ ⎯ →R
anodic oxidation (2. 1)
When a metal is immersed into a solution, electrochemical reactions may occur at the surface
of the metal causing the metal to corrode. Two or more reactions simultaneously occur: the
oxidation of the metal (anodic reaction) and the reduction of the oxidizing agent (cathodic
reaction). Oxidation involves the loss of electrons, while reduction occurs when electrons are
gained.
A typical example of corrosion may be a piece of metal which becomes wet leading to the
occurrence of corrosion. Ions are exchanged through the surface water layer (electrolyte)
between the anode and cathode while electronic exchange will take place through the bulk
metal. Corrosion often involves oxidation of metal atoms to form ionic species with higher
oxidation states and the liberation of electrons.
For a generic metal M:
! ⇆ !#$ + &' ( (2. 2)
In the case of iron:
)' → )' ,$ + 2' ( (2. 3)

These are called half-cell reactions because the electrons liberated by the oxidation reaction
must be consumed by a reduction reaction occurring on the same electrode. A common
reduction reaction is the hydrogen evolution reaction:

P a g e | 18
2.$ + 2' ( → ., (2. 4)
This would gives an overall reaction of
)' + 2.$ → )' ,$ + ., (2. 5)
Anodes and cathodes can be spatially separated at fixed locations associated with
heterogeneities on the electrode surface. Alternatively, the locations of the anodic and cathodic
reactions can fluctuate randomly across the sample surface. The former case results in a
localised form of corrosion, such as pitting, crevice corrosion, inter-granular corrosion, or
galvanic corrosion and the latter case results in nominally uniform corrosion.
2.1. Thermodynamics of Electrochemical reactions
Thermodynamic calculations can be used to predict the tendency for a metal to corrode in a
particular environment. The driving force for a spontaneous reaction is a negative value of
Gibbs free energy change (∆G < 0) at a fixed temperature and pressure. For non-spontaneous
reaction such as an electrolytic cell, the Gibbs free energy change is positive, so an input of
energy is required for the reaction to proceed. The Gibbs energy change for a reaction is defined
by equation (2. 6). Where ∆H is the enthalpy change, T is the temperature and ∆S, the entropy
change associated with the cell reaction.
∆0 = ∆. − 3∆4 (2. 6)
The Gibbs energy for any species, G is related to the standard Gibbs energy, G0 (the free energy
for the species in its standard state) by equation (2. 7), where R is the gas constant, and a denotes
the activity of the species.
0 = 0 5 + 63ln9 (2. 7)
For a given reaction, the Gibbs energy change is given by the difference between the summation
of values for the products and reactants:
∆0 = ∆0:;<=>?@A − ∆0;BC?@C#@A (2. 8)

P a g e | 19
The Gibbs energy change for an electrochemical cell reaction is directly related to the reversible
potential, EO/R via equation (2. 9). In this expression F is Faraday constant, the product of
Avogadro’s constant NA and the charge on a single electron Qe (F= NAQe, 96 485 C mol-1).
∆0 = −&)DE/G (2. 9)
Half-cell potentials are established with reference to the standard hydrogen electrode (SHE)
with the electrode potential of hydrogen arbitrarilly assigned a value of zero volts. The SHE
(Pt, H2 / H+) represents the electro-active species in the electrolyte in their standard states: H2
gas at 0.1 MPa pressure, 298.15 K and the activity of H+ protons is 1. Essentially, the electrode
whose standard electrode potential is being measured is configured as the cathode and the anode
is the reference electrode.
2.1.1. Nernst Equation
The standard state requires all reactants and products to be at unit activity. The Nernst equation
was established to calculate the effect of deviation from standard state on the reversible
potential. Equation (2. 10) represents the free energy change for a generalised reaction in which
the reactants and products are not in their standard states.
(CIJKL )N
∆0 = ∆0° + 63ln (2. 10)
(CJOPQR )S
If the reaction is a reduction reaction (the reactants and products are the oxidised and reduced
forms of the species respectively) substituting ∆0 = −&)DE/G and ∆0° = −&)DE/G ° into
equation (2. 10) results in:
GU (CKW )N
DE/G = DE/G ° + ln (2. 11)
#V (CJOL )S
Equation (2. 11) is called the Nernst Equation and it describes the equilibrium potential of an
electrochemical half reaction in which the oxidised or reduced species are not necessarily in
their standard states.

P a g e | 20
2.1.2. Pourbaix Diagrams
As mentioned previously, an electrode reaction is in a state of dynamic equilibrium at its
equilibrium potential at which no net reaction current flows. The anodic reaction (metal
corrosion) can only occur at potentials more positive than the equilibrium potential and the
cathodic reaction (oxidant reduction) proceeds at potentials more negative than its equilibrium
potential. Thermodynamically it is known that the corrosion of metals in solutions is not only
dependent on the electrode potential but also on the pH of the solution. Thus, when several
reactions are possible, as is the case with Fe oxidation, the electrochemical behaviour in
aqueous systems can be illustrated by Pourbaix diagrams in the form of phase stability plots
which are created by relating the reversible potential to pH and total dissolved cation activity.
These visual representations of the equilibrium conditions in the potential/pH space were
pioneered by Professor Marcel Pourbaix (Bard, 2003). They are based purely on the
thermodynamics and indicate the stability of predominant species under given conditions but
give no indication of the rates of reactions from one phase to another.
2.1.2.1. Pourbaix Diagrams for water
In corrosion mechanisms, the most important cathodic reactions involve water. There are two
primary cathodic reactions, each of which takes a different form in acids or bases. The first is
the hydrogen evolution reaction (HER), in which hydrogen gas is evolved.
In acids:
2.$ + 2' ( ↔ ., (Y) (2. 12)
5.5_`
E([ \ /[] ) = D°([ \ /[] ) + log .$ ,
(2. 13)
,
= −0.059f.
In bases:
P a g e | 21
2., g + 2' ( ↔ ., Y + 2g.( (2. 14)
The Nernst equation for the base form of the HER equation given is also given by D([] E/[] ) =
−0.059f..
When dissolved oxygen gas is present in the aqueous solution the following oxygen reduction
reactions are also possible:
In acids:
g, + 4.$ + 4' ( ↔ 2., g (2. 15)
5.5_`
D(E] /[] E) = D°(E] /[] E) − log .$ (i
(2. 16)
i
= 1.229 − 0.059f.
In bases:
g, + 4., g + 4' ( ↔ 4g.( (2. 17)
The Nernst equation in equation (2. 16) also applies for the oxygen reduction reaction in bases.
The Nernst equation for the HER (D([ \ /[] ) = −0.059f.) is represented on the Pourbaix
diagram for water shown above (Figure 2-1) with an intersection at (0, 0) and a slope of -0.059
pH-1 at 298 K. The hydrogen evolution reaction is in equilibrium along this line. At potentials
above this line in E/pH space, water (H+) is the stable phase and any dissolved H2 gas that might
be present in solution is oxidised. Below this line, H2 gas is the stable phase, and water (or H+)
is reduced.
The Nernst equation for the oxygen reduction reactions is also plotted in Figure 2-2 with a line
parallel to the HER line, which intersects at 1.23V. Above the oxygen line, O2 is the stable
phase and water (OH-) is oxidised. Below the oxygen line, water is stable and O2 is reduced.
P a g e | 22
The two plots combined give the region of water stability with water being stable in between
both lines. It is oxidised to O2 above the oxygen line and reduced to H2 below the hydrogen
line.
It can be seen from the Nernst equations that as the activity of the oxidizer H+ increases, pH
decreases and D([ \ /[] ) increases. This is the case for all oxidizers according to the Nernst
equation.
2.0
Electrode potential (SHE) / V
1.6
1.2
0.8
0.4
0.0 Water
-0.4
-0.8
-1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
Figure 2-1: Pourbaix diagram of water; 298 K, 0.1 MPa.
2.1.2.2. Pourbaix Diagram for Iron
Pourbaix diagrams for metals in aqueous solutions are created in order to visualise the regions
of stability for the metal and its various corrosion products. To generate a Pourbaix diagram,
the possible reaction products in an aqueous solution must be known. Commonly, a metal will
oxidize to form a soluble cation, a soluble anion and a metal oxide or hydroxide.
In the case of iron, it will dissolve to form ferrous cations via reaction:
)' → )' ,$ + 2' ( (2. 18)
With the Nernst Equation

P a g e | 23
5 5.5_`
At 298 K: D(VB ]\ /VB) = D(VB ]\ VB) + log Fe,$ (2. 19)
,
= −0.44 + 0.0295 log )' ,$
This line is dependent only on the dissolved FeII activity and is independent of the pH. It is
characterized by horizontal lines in the E/pH space depending on the value of [FeII]. Horizontal
lines on Pourbaix diagrams represent reactions which are purely electrochemical. Above this
line the oxidation reaction will proceed and Fe2+ is the stable phase while Fe is stable below it.
Iron can react with water to form ferrous hydroxide:
)' + 2., g → )' g. , + 2.$ + 2' ( (2. 20)
5.5_`
D(VB E[ ] /VB) = D 5 (mn op ] /VB)
+ log . , (2. 21)
,
= −0.047 − 0.059f.
This is represented by a line parallel to the HER and oxygen reduction lines with a slope of -
0.059 pH-1 and intersection of -0.047 V. Sloping lines on Pourbaix diagrams indicate a reaction
which involves both electrons and protons.
Ferrous iron reacts to form ferrous hydroxide
)' ,$ + 2., g ↔ )' g. , + 2.$ (2. 22)
The equilibrium constant for this reaction describes the combinations of Fe2+ concentration
and pH for which this reaction is in equilibrium
[\ ]
r = = 10(st.,` (2. 23)
VB ]\
f. = 6.645 − 0.5log )' ,$
This relationship is independent of potential and is displayed in the Pourbaix diagram as a
vertical line at a given pH that depends on [Fe2+] and separates the Fe2+ and Fe(OH)2 phases.
Vertical lines on Pourbaix diagrams represent purely chemical reactions which involve
protons but not electrons.

P a g e | 24
Fe can react to form an oxy-anion such as HFeO2- according to:
)' + 2., g → .)'g, ( + 3.$ + 2' ( (2. 24)
5.5_`,
D([VBE] w /VB) = D 5 ([VBE] w /VB) + log ( .$ t
.)'g, ( (2. 25)
,
0.0592 3
= D 5 ([VBE] w /VB) + log .)'g, ( − 0.059 f.
2 2
= 0.493 + 0.295 log .)'g, ( − 0. 0886 f.
Ferrous hydroxide (Fe(OH)2) dissolves chemically to form HFeO2-:
)'(g.), ↔ .)'g, ( + .$ (2. 26)
r = .$ .)'g,( = 10(sy.t (2. 27)
f. = 18.3 + log .)'g,(
Other possible reactions and their accompanying Nernst (or equilibrium) equations are listed
below.
)' ,$ + 3., g ↔ )'(g.)t + 3.$ + ' ( (2. 28)
D(VB E[ z /VB ] ) = 1.057 − 0.1773f. − 0.0591log )' ,$ (2. 29)
)' t$ + 3., g ↔ )'(g.)t + 3.$ (2. 30)
s
f. = 1.613 − log )' t$ (2. 31)
t
)'(g.), + ., g = )'(g.)t + .$ + ' ( (2. 32)
D ;B{ = 0.271 − 0.0591f. (2. 33)
The full Pourbaix diagram for Fe-H2O systems showing the equilibrium lines for dissolved iron
activity of 10-4 is displayed in Figure 2-2. Examples of points that can be taken from the
potential-pH diagram are as follows:

P a g e | 25
Any point on the diagram will give the most thermodynamically stable (abundant) form of Fe
at that potential and pH. For example at E = +0.80 V and pH 6, Iron (III) oxide, Fe2O3 is
predominant and at E = +0.20 V and pH 2, ferrous ions, FeII are most stable.
In general, there are three phase fields in E-pH diagrams: immune, active and passive.
Elemental Fe is most stable at low potentials, in the region of immunity. In this region corrosion
is thermodynamically unfavourable. The tendency will be for any Fe ions in solutions to plate
out. The active phase field in which dissolved ions such as FeII and FeIII are predominant and
regions where the tendency would be for corrosion to occur. Solid oxides and hydroxides are
most stable in the regions of passivity, the condition of corrosion resistance occurring when
certain metals form thin, oxidized protective films on the metal surface.
The diagram also visualises the oxidising and reducing abilities of the stable forms of Fe.
Oxidising agents are at the top of the diagram while reducing agents are at the bottom. For
example HFeO4- and FeO42- are the most oxidising form of iron while the elemental Fe is the
strongest reducing agent.
Pourbaix diagrams for metal-water systems always include the E-pH lines for the cathodic
reactions in water (HER and oxygen reduction). The position of these lines relative to the stable
phases indicate the availability of the cathodic reactions for corrosion.

P a g e | 26
2.2
H3FeO4+ H FeO
2.0 2 4 HFeO4-
1.8
1.6 FeO42-
1.4
1.2
1.0
Fe3+ Fe3+
0.8
0.6
Fe2O3 (c) O2
0.4 Fe(OH)3 (ppt) H O
2
Fe(OH)4-
0.2
H+ Fe2+
0.0
-0.2 H2 Fe(OH)4-
-0.4
-0.6 Fe(OH)2 Fe O
3 4
-0.8
-1.0 Fe Fe(OH)2 Fe(OH)42-
-1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Figure 2-2: Potential-pH diagram for Fe-H2O systems with dissolved iron activity of 10-4, 298 K
and 0.1 MPa (Kelsall 2011).
2.2. Kinetics of Electrochemical Reactions
While the information about the tendency for corrosion to occur can be obtained from
thermodynamic calculations, this does not necessarily mean that it will actually occur nor does
it give any information about corrosion rates, which are determined by reaction kinetics. An
understanding of the fundamental laws of corrosion kinetics is essential in developing a more
thorough understanding of the electrochemical system being studied.
2.2.1. Rate of reaction: Faraday’s Laws of Electrolysis
As electrochemical reactions result in the production or consumption of electrons, the rate of
electron flow to or from a reacting surface gives a measure of the reaction rate. The overall rate
of the electrochemical reaction may be expressed by Faraday’s laws of electrolysis. The amount
P a g e | 27
of reactant or product material undergoing electrochemical change m (mols) is proportional to
the amount of electrical energy, Q, involved:
}
| = (2. 34)
#V
The charge, Q is defined as the integral of cell current with respect to time, t.
~= ÄÅ (2. 35)
If the current is constant:
~ = Å (2. 36)
Equation (2. 34) becomes:
Ç@
|= (2. 37)
#V
Differentiating with respect to time gives the expression for the rate of reaction (mols-1):
=É Ç
= (2. 38)
=@ #V
Ä|/ÄÅ represents the loss of reactant (O) which is equal and opposite to the rate of formation
of the product (R).
If the expression is divided by the area (A), it can be expressed in terms of the current density
j, Ñ = /Ö
s =É à
Ü= = (2. 39)
á =@ #V
The units on both sides of the equation are in mol m-2 s-1.
It is also possible to expresses the amount of material transformed per unit reactor volume, VR
per unit time with the units mol m-3 s-1
Ç =É à
Ü = = (2. 40)
âä =@ #Vâä
P a g e | 28
2.2.2. Rate limiting Mechanisms
From equations (2. 39) and (2. 40) it can be seen that the rate of reaction is proportional to the
current density. However prior to the charge transfer reaction, the reactant must be transported
from the bulk solution to the electrode interface, and the reaction products, once formed will
diffuse away from the electrode. A particular reaction may be controlled by the kinetics of the
electron transfer or by the rate at which material is brought to (or from) the electrode through
mass transport. The overall reaction rate will be limited by the slowest step. Purely electron
transfer controlled reactions will be discussed before taking into account mass transport
considerations.
2.2.3. Charge Transfer Controlled Reactions
2.2.3.1. Exchange Current density
Consider again reaction (2. 1), the forward reaction representing the cathodic reduction of O to
form R, and the reverse reaction is the anodic oxidation of R to O. Electrochemical equilibrium
is a dynamic process in which electrons are continuously being passed back and forth at the
metal-solution interface. The exchange current density (jo) for an electrochemical half cell
reaction occurs at equilibrium, when the rate of the forward and reverse reactions are equal.
àK
Üã = Ü; = (2. 41)
#V
The exchange current density (jo), a fundamental parameter of electrode kinetics can be
determined using the following equation:
é
Ñ< = &)å°9[ ∗ (2. 42)
Where ko is a specific rate constant, a is the activity of the associated component and w is the
reaction order. However, in practice it is difficult to calculate the exchange current density
directly from equation (2. 42) above due to the unavailability of the specific rate consant (ko)
P a g e | 29
value for half-cell reactions, as such the exchange current density (jo) can only be determined
experimentally.
2.2.3.2. Electrochemical Polarization
The darkened dot on Figure 2-3 (a) below representing equilibrium (EO/R, jo) signifies a finite
exchange current density. At this point, both the forward and reverse current densities are equal
and the net current density is zero. Electrochemical polarization is a phenomenon that changes
the equilibrium state of an electrochemical reaction and causes the potential to differ from the
equilibrium potential EO/R. Polarization, η = E- EO/R, is defined as the change in potential from
the reversible potential of a given half-cell electrode reaction (Bard, 2003).
(a) (b)
Log |j| Log |j|
Mn+ + ne- → M M → Mn+ + ne-

jo jo
EO/R E EO/R E
Figure 2-3: Schematic representation of the relationship between current and potential for a
simple electrochemical reaction under kinetic control.
This deviation from EO/R in Figure 2-3 (b) causes the half cell reaction to proceed preferentially
in one direction over the other resulting in a net current. For anodic polarization the potential is
changed positively to a value higher than the equilibrium potential due to the removal of
electrons from the metal surface. The reverse is the case for cathodic polarization, electrons are
P a g e | 30
supplied to the surface and a build-up in the metal causes the surface potential to become
negative to EO/R.
In electrochemistry, potential and current are interdependent. It is common practice to control
one and measure the effect on the other
2.2.3.3. Butler- Volmer Equation
The overpotential that results from a change in potential is called activation overpotential if the
concentrations of the reactants and products at the electrode surface are the same as in the bulk
solution and the rate of charge transfer in the electrochemical reaction is governed only by an
electron transfer step of the half-cell reaction at the metal-solution interface. This occurs when
the rate of mass transport is fast compared to the rate of charge transfer. For such a system, the
rate of reaction is limited by activation overpotential or charge transfer rate. Equation (2. 43) is
called the Butler-Volmer equation which relates the rate of reaction (current density) and the
driving force for the reaction (overpotential).
ëí V ëî V
Ñ = Ñ< exp ì − exp − ì (2. 43)
GU GU
Where αA and αC represent the transfer coefficients for the anodic and cathodic reactions
respectively. They indicate the fraction of the applied overpotential, η used to provide a driving
force for the reaction.
ì = D − DE/G (2. 44)
The first term of the Butler-Volmer Equation represents the anodic partial current density ja and
is taken as positive while the second term, jc, the cathodic partial current density is taken as
negative. The net current (the sum of the two terms) is positive when the electrode is polarised
anodically and negative when the electrode is polarized cathodically.

P a g e | 31
The Butler-Volmer equation illustrated in Figure 2-4, represents the kinetics for charge-transfer
controlled electrochemical reactions. When plotted on a current density versus electrode
potential axes, the curve crosses the x-axis at the reversible potential.
20
Current Density, j / Exchange current density, j0
óá ) óô )ì g + &' ↔ 6
Ñ = Ñ< ïexp ñ ìò − exp ñ− òö
15 63 63
10 óá )
Ñá = Ñ< exp ñ ìò
63
5
Oxidation
0
Reduction
-5
-10
óô )
-15 −Ñô = Ñ< exp ñ−
63
ìò EO/R
-20
-0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20
Overpotential , η / V
Figure 2-4: Plot of Butler-Volmer Equation for charge transfer limited reaction.
2.2.3.4. Tafel Equation
For particularly large values of anodic polarization from the equilibrium potential (ηa ≥ 10 mV),
the first term on the of Butler-Volmer equation (anodic part) dominates over the cathodic
contribution. Therefore, at large anodic overpotentials, the Butler-Volmer equation simplifies
to the Tafel equation (2. 47):
ëí V
Ñ = ÑC = Ñ< exp ì (2. 45)
GU
ëí V
log (ÑC ) = log (Ñ< ) + ì (2. 46)
,.t5t GU
P a g e | 32
à
η = úC log (2. 47)
àK
Where:
,.t5tùû
úC = (2. 48)
ëí V
Cathodic polarization results in a similar equation:
à
ì? = −ú? log (2. 49)
àK
As corrosion reactions typically occur at relatively high overpotentials, the Tafel equation is
applicable in describing the corrosion kinetics under negligible mass transfer considerations.
log |j|
Tafel region Tafel region
jo
Ecorr E
Figure 2-5: Schematic of Tafel plot
An Evans diagram (a semi-logarithmic plot of log |j| versus E) is displayed in Figure 2-5. As
polarisation proceeds in either the cathodic or anodic direction, the current density (j) deviates
linearly from the equilibrium current density (jo). The linear deviation of the potential current
density with activation polarisation can be described by the Tafel equations. These slopes
P a g e | 33
referred to as Tafel constants (β), give the magnitude of polarization per decade of current
density and may be used to determine the transfer coefficient α. The value of Tafel constants
usually ranges from a value of 0.03 to 0.20 V depending on the conditions such as pH and
temperature the activity of the interacting species (Jones 1992). Once the values of the
electrochemical parameters are known for a particular system, the charge-transfer controlled
current densities can be predicted as a function of the electrode overpotential. Values such as jo
and β can sometimes be obtained from open literature, otherwise they have to be determined
experimentally.
2.2.4. Mixed Potential Theory
The Mixed Potential principle of charge conservation states that the sum of all the anodic
oxidation currents must be equal to the sum of all the oxidant reduction currents, i.e. the total
rate of oxidation must be equal to the total rate of reduction. When a metal is corroding in an
aqueous solution both the anodic and cathodic half-cell reactions occur simultaneously on the
surface. Each of these half-cell reactions will have its own half-cell electrode potential and
exchange current density. As the half cell reactions (anodic and cathodic) polarize on the same
surface, the half-cell electrode potentials change respectively, according to the Tafel equations
until they become equal at Ecorr, the corrosion potential at which:
∑C + ∑? = 0 (2. 50)
The corrosion potential is also referred to as the open-circuit potential or rest potential. This is
a mixed potential as the value is determined by both the rate of the anodic and cathodic reactions
and it lies somewhere in between the equilibrium potentials of the anodic metal dissolution and
the cathodic oxidant reduction. The net measured current for a corroding metal in an aqueous
solution typically take a similar form to the Evans diagram (Figure 2-6) below. For an
electrochemical reaction in which HER is the main cathodic reaction, the Evans diagram shows
the Tafel lines for the anodic and cathodic parts of the hydrogen evolution reaction along with
P a g e | 34
the Tafel lines for the dissolution and plating of the metal with a reversible potential below that
of the HER. D†\ /† and D[ \ /[] are the reversible potentials while Ñ<,†\ /† 9&¢ Ñ<,[ \ /[] are the
current exchange densities for the metal dissolution and HER respectively.
log | j|
jco
jo,H
j
o,
EEM+/M E Ecorr EEH+/H E E
Figure 2-6: Anodic and cathodic half-cell reactions present simultaneously on a corroding
iron surface in acidic media.
As the two reactions then polarize on the conducting surface to Ecorr, the half-cell electrode
potentials change according to the following relationship.
à£\ /£
ì†\ /† = ú†\ /† log (2. 51)
àK ,£\ /£
à§\ /§
ì[ \ /[] = −ú[ \ /[] log ]
(2. 52)
àK ,§\ /§
]
At the potential Ecorr, the rates of the anodic and cathodic reactions are equal, which is the rate
of corrosion.
Ñ†\ /† = Ñ[ \ /[] = Ñ?<;; (2. 53)

P a g e | 35
Arranging equations (2. 51) & (2. 52) into (2. 53) gives an equation for calculating jcorr:
• \ • \
£ /£ § /§]
¶ \ ¶ \
Ñ?<;; = Ñ<,†\ /† . 10 £ /£ = Ñ<,[ \ /[] . 10 § /§] (2. 54)
Where ì†\ /† = D?<;; − D†\ /† and ì[ \ /[] = D?<;; − D[ \ /[] .
The net current goes to zero at the corrosion potential and extrapolation to the corrosion
potential of the Tafel regions (the linear regions of the semi-logarithmic plot) gives the
corrosion rate. The corrosion potential is predicted by the point at which the extrapolated
anodic line for the metal dissolution intersects that of the cathodic line for the HER. At this
potential, the oxidation and reduction current densities are equal. The rate of the anodic
dissolution at the corrosion potential is the corrosion rate (corrosion current density). This
corrosion potential always takes a value between the reversible potential for the two half
reactions. The electrode is not in equilibrium at the corrosion potential since net changes are
occurring: metal is oxidised and water is reduced.
EM/M+
jcorr
0
EH+/H E
Ecorr
Figure 2-7: Relationship of measured polarization curve to the Evans diagram for corroding
metal in acidic media.
P a g e | 36
The concept of corrosion potential is further illustrated in Figure 2-7. The potential at which
the rate of metal dissolution is equal to the HER rate is the potential corresponding to the
corrosion current density / corrosion rate.
While the corrosion potential clearly depends on both the anodic and cathodic reactions, the
rate of corrosion may be controlled by either one of these reactions. In most cases of metallic
corrosion in acidic media, HER controls the rate of corrosion while oxygen reduction controls
the rate in neutral media (Bard, 2003).
2.3. Mass transport
The term ‘mass transport’ in electrochemical systems refers to the mode by which reactant and
product species are supplied to and from the solution | electrode interface. For kinetically
controlled reactions, the rate determining step is the rate of the electrochemical reaction at the
electrode surface rather than the rate of diffusion of species to the surface. Thus, the kinetics
explained in Section 2.3 only applies in cases where the surface concentrations of all the species
are equal to the bulk solution concentrations. At larger overpotentials and high current densities,
mass transport effects have to be taken into consideration as the rate of transport to (or from)
the surface can also affect or control the overall reaction rate. In such cases, the reaction at the
electrode surface occurs faster than the reactants can diffuse from the bulk to the
solution | electrode interface. Therefore, the reactant species are depleted at the reaction surface
resulting in a mass transport limited current discussed in more detail in the following section.
2.3.1. Transport Processes in Electrochemical Systems
As discussed above, the rate of an electrochemical reaction is dependent on the applied potential
difference, the kinetics and the transport rate of reactant species to and from the electrode
surface. Ionic motion in the electrolyte will respond to differences in density, concentration
gradients and electric fields as well as bulk motion of the fluid. In general, there are three
mechanisms of ion transport to and from the electrode/electrolyte interface: convection,

P a g e | 37
diffusion and migration. From Table 2-1, it can be seen that the three modes of ionic transport
are driven by a form of gradient near the electrode surface. This gradient gradually changes the
particular property with distance from the electrode surface and results in boundary layers.
Table 2-1: Modes of mass transport (Walsh, 1993)
Mode Driving Force Nature of Force

Convection Velocity gradient dv/dx Density differences (if
natural convection) or
external mechanical
forces (in forced
convection)
Diffusion Concentration gradient dc/dx Physical
Migration Potential Gradient dß/dx Electrostatic
Thus, the layers of electrolyte near the electrode surface have a different velocity, concentration
or potential than the electrode surface and the bulk electrolyte. The forms of mass transport in
electrochemical reactions are illustrated schematically in Figure 2-8.
Anode Cathode
ßA
Bulk
EA+ ηA electrolyte
Ohmic Migration
potential -
drop, IR +
Convection EC+ ηC
ßC
Diffusion layer
Figure 2-8: Schematic of an electrochemical system showing the transport processes and one
dimensional distribution of electrical potential
P a g e | 38
Diffusion is the movement of a species down a concentration gradient and it is induced by a
chemical change at the electrode solution, such as the conversion from the oxidised species (O)
to the reduced species (R). This results in a lower concentration of O at the electrode surface
than in the bulk solution and a higher concentration of R at the electrode surface.
The rate of material transport by diffusion is governed by Fick’s first law. The law states that
the diffusion flux of a dissolved species i to a surface depends on the diffusion coefficient of
the species in the electrolyte Di and is proportional to the concentration / activity gradient.
®=©ãã,© = −™© ∇¨© (2. 55)
Where the Del operator for gradients in three dimensions is:
≠ ≠ ≠
∇= + + (2. 56)
≠Æ ≠Ø ≠∞
Where D is the diffusion coefficient (m2 s-1) and x, y z represent the distance from the electrode
(m).
Under non-steady state conditions, Fick’s second law can be used to predict how the diffusion
causes the concentration to change with time. It states that the change of concentration with
time is equal to the difference of the diffusive fluxes in and out of a given volume element
(Bard, 2003)
=?±
= ™© ∇, ¨© (2. 57)
=@
In the case of a dissolved species reacting chemically, Fick’s second law is modified to take
into account homogeneous chemical reactions. For example, if the species is being consumed
by an irreversible first-order chemical reaction with a rate constant ki, the resulting equation is:
=?±
= ™© ∇, ¨© − å© ¨© (2. 58)
=@
Convection is the movement of a species down a velocity gradient due to mechanical forces
such as pressure, density, or temperature gradient (Jones, 1996, Bard, 2003). There are two
P a g e | 39
types of convection, natural and forced convection. Natural convection results in small
differences in solution density due to chemical changes at the electrode surface or by localised
temperature fluctuations. (Walsh, 1993) Forced convection is usually the main mode of mass
transport and may be brought on by electrode or electrolyte movement.
The flux of a species i due to convection Nconv,i (mol m-2 s-1) is the product of the bulk
concentration of that species (ci) and the velocity of the solution (vsol):
®?<#{,© = ¨© ≤A<≥ (2. 59)
Migration is the movement of a charged species in response to a potential gradient caused by
the presence of an electric field which results in ionic currents passing through the electrolyte.
It is the preferential drift of ions toward the electrode of the opposite sign. This type of ionic
transport is driven by electrostatic forces and it takes the ion size and charge into account. Ions
will move in a field with a net drift velocity ≤= (m s-1) that is proportional to the potential
gradient ¥µ (V m-1).
≤= = −∂B ∇ß (2. 60)
Where ue (m2 V-1 s-1) is the electric mobility and it is a measure of the velocity the ion would
acquire in an electric field of 1 V m-1. The induced flux of species i, due to migration is:
® É©∑;,© = ∏© )∂© ¨© ∇ß (2. 61)
In this expression, mobility ui is the velocity that one mole of species i would acquire under the
effect of a force of 1 N. ziF (C mol-1) is the charge per mole of the ionic species. zi is positive
for cations and negative for anions.
Therefore, the total flux for a charged electro-active species, Ni (mol m-2 s-1) to an electrode
surface is the sum of the three different modes of mass transport:
®© = ®=©ãã,© + ®É©∑;,© + ®?<#{,© (2. 62)
π© = −™© ∇¨© − ∏© )∂© ¨© ∇ß + ¨© ≤A<≥ (2. 63)

P a g e | 40
The diffusion coefficient (Di) can be expressed in terms of mobility by the Nernst-Einstein
equation:
GU
Dª = ∂© (2. 64)
∞V
From the Faraday’s law, the current density can be calculated from equation (2. 65).
Ñ=) © ∏© ®© (2. 65)
Where the current density at an electrode with area Ae:
Ñ = /ÖB (2. 66)
In the absence of significant concentration gradients, substitution of (2. 65) equation (2. 64)
results in ohm’s law:
Ñ = −º∇ß (2. 67)
º=) © ∏© ∂© ¨© (2. 68)
Where, º is the conductivity of the solution and with the assumption of electroneutrality in the
bulk solution:
© ∏© ¨© =0 (2. 69)
The movement of ions due to an electric field may cause re-distribution of the charge in the
solution adjacent to the electrode interface which will cause the potential drop at the interface
to be altered. This could result in fluctuations in the localised driving force for electrochemical
reactions which is particularly detrimental during kinetic measurements during which η should
be measured or controlled as accurately as possible (Fisher, 1996). However, if there is an
abundant concentration of charged species also known as a ‘supporting electrolyte’ in addition
to the primary reactant species, then the field fluctuations can be stabilised as they carry the
most of the charge in the bulk solution (Pletcher, 1984). Mass transport limiting current
densities (jlim) are also lower in the presence of a supporting electrolyte, because increasing the
electrode potential has no effect on the migrational contribution to the transport rate of the
P a g e | 41
electroactive species, the majority of the current being carried in the bulk solution by the
supporting electrolyte. Thus, it is common practice in electrochemical systems to add a
supporting electrolyte to increase the conductivity of the solution and decrease the ohmic
potential drop between the electrodes. A supporting electrolyte also serves to improve
uniformity in the current density distribution across electrochemical reactors with larger
electrodes and in some cases can be used to adjust the solution pH.
In the presence of excess supporting electrolyte, the migrational contribution in equation (2. 62)
is negligible. For the following discussion, it is assumed that near the electrode surface there is
a stagnant layer of electrolyte of thickness δN, called the Nernst diffusion layer. In reality, this
stagnant layer near the electrode surface of a finite thickness does not exist as the solution
velocity vsol decreases gradually closer to the electrode surface becoming zero at the electrode
surface. However, the concept of a Nernst diffusion layer is important when considering the
convective mass transport at electrodes and allows for distinction between the contributions due
to convection and diffusion. Within the stagnant Nernst diffusion layer, the net flux of the
species is due only to diffusion, while beyond the diffusion layer, ohm’s law applies in a well-
stirred bulk electrolyte in which there are essentially no concentration gradients, so convection
is the only relevant mass transport mechanism.
The thickness of the Nernst diffusion layer is dependent on the convection rate as convection
has the effect of decreasing the hydrodynamic boundary layer at the electrode | solution
interface. Rotating disk electrodes are often employed in electrochemical laboratory
experiments to provide a well-characterised system for controlling mass transport rates due to
convection, as will be discussed further in Chapter 4.
According to Fick’s law for one dimensional steady state diffusion, the flux to a plane is
proportional to the concentration gradient
=?±
® = −™© (2. 70)
=Æ
P a g e | 42
where the diffusion coefficient Di (ca. 10-9 m2 s-1 for species in aqueous solution at 298 K) is
assumed to be independent of the concentration. Integration of equation (2. 70) over the Nernst
diffusion layer gives the following equation for the current density.
(ô± (ô± ,W¿¡ )

Ñ = −&)™© (2. 71)
¬√
Where the current density is linearly related to the flux via Faraday’s law, equation (2. 72). ¨©
and ¨© ,Æƒ5 are the bulk and surface concentrations respectively and δN is the Nernst diffusion
layer thickness.
Ñ = −&)®© (2. 72)
Solving equation (2. 71) for the absolute value of the surface concentration of species i results
in:
à ¬√
≈© ,Æƒ5 = ≈© − (2. 73)
#V∆«
The surface concentration of the species ≈© ,Æƒ5 continues to decrease with increasing current
density until it reaches a value of ≈© ,Æƒ5 = 0 at the maximum possible rate. The current density
at this point is called the mass transport limiting current density, Ñ≥©É .
(#V∆± ô±
Ñ≥©É = = &)åÉ ¨© (2. 74)
¬
Where km is the mass transport rate coefficient. Dividing by (2. 71):
à ô± ,W¿¡
=1− (2. 75)
à»±N ô±
Thus the concentration polarization which arises due to the changes in concentration of the
electroactive species in the electrode | electrolyte interface can be expressed as (Jones 1996)
à
ì?<#? = úlog 1 − (2. 76)
à»±N
With increasing electrode potential from its equilibrium value, the current density increases
until it approaches the limiting current density Ñ≥©É , which is independent of potential, unless
P a g e | 43
another reaction such as solvent decomposition occurs. As the current reaches the limiting
current, the concentration polarisation goes to negative infinity. The limiting current density,
Ñ≥©É increases with solution velocity, temperature, and bulk concentration of the species.
Mass transport considerations are particularly significant in corrosion rate measurements in
where there is a limited concentration of oxidant species. For example, in a case of iron
corrosion occurring in a neutral solution with oxygen as the primary oxidant, then the maximum
corrosion rate will be determined by the limiting current density of the cathodic reactant to the
iron surface.
2.3.2. Electrochemical Reaction Coupled to Mass Transport
Electrode reactions typically involve a series of steps: the transport of the reactant species to
the solution | electrode interface, the transfer of charge at the electrode and the transport of the
product species formed away for the surface. The charge transfer and mass transport processes
are interlinked in series, i.e. the charge transfer can occur only after transport of reactants
species to the electrode surface. At low overpotentials, the mass transport rates are fast
compared to the electrochemical reaction kinetics i.e., electron transfer is the rate determining
step. The extended Butler-Volmer takes into account both the kinetically controlled
overpotential and the mass transfer controlled overpotential, which arise due to the changes in
the concentration of the electroactive species in the electrode-electrolyte interface. (Bard and
Faulkner, 2001). The activation and concentration polarisations combine in the case of the
cathodic branch by substituting equation (2. 76) into equation (2. 43) which results in:
à ( s(ë V( ( JOÀ )
Ñ = ÑE (1 − )× exp (2. 77)
à»±N GU
This can be rearranged to give the equation which describes the j-E relationship for the cathodic
reaction controlled by charge transfer kinetics as well as the mass transport kinetics.
w Œwœ –
àK nÃÕ ( ( JOÀ )
ä—
Ñ= “ w Œwœ – (2. 78)
s( K nÃÕ ( ( JOÀ )
“»±N ä—
P a g e | 44
20
18
16 <Æ©=C@©<#
14 6 ◊⎯⎯⎯⎯⎯⎯Ÿ g + ≤B ' (
Current density, j / A m-2
12
10
8 ÑC = Ñ< exp{β› ì}
6
4 π⁄,E = −≤B )åÉ [g]
2
0
-2
-4 π⁄,G = −≤B )åÉ [6]
-6
-8 Ñô = Ñ< exp{β’ ì}
-10
-12
-14 ;B=>?@©<#
-16 g + ≤B ' ( ◊⎯⎯⎯⎯⎯⎯Ÿ 6
-18 EO/R
-20
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Overpotential, η / V
Figure 2-9: Regions of Cathodic Current density-Overpotential Relationship

P a g e | 45
In the kinetically controlled potential region, the activation overpotential dominates over
( s(ë V
concentration overpotential as Ñ≥©É > Ñ< exp (D − D ;B{ ) so the denominator of
GU
equation (2. 78) goes to 1 and the equation simplifies to the Tafel equation. However, larger
overpotentials result in an exponential increase in current density according to the Bulter-
Volmer Equation, the reactant species being progressively depleted at the interface, resulting in
a concentration gradient in the boundary layer (δ). In this potential region, the rate the reaction
is influenced by both charge transfer and mass transport and the reaction occurs under mixed
control.
In the mass transport controlled region, further increase in the overpotential results in the
( s(ë V
surface concentration of the reactant species tending to zero, Ñ≥©É < Ñ< exp (D −
GU
D ;B{ ) the equation becomes Ñ = Ñ≥©É and the reaction rate becomes independent of potential.
The influence of charge transfer and mass transport can also be seen by rearranging equation
(2. 78):
GU à GU à
D − D ;B{ = ì = − ln − ln 1 − (2. 79)
s(ë V à« s(ë V à»±N
The total polarisation is a sum of the charge transfer polarisation (the first term) and the
concentration polarisation (second term).
ì U = ìC?@ + ì?<#? (2. 80)
The whole Butler-Volmer equation, equation (2. 43), can be extended to take all the regions in
the polarization curve (Figure 2-9) into account.
œ–·‚ w(Œ w œ)–·‚

à¡ nÃÕ ( àK nÃÕ
ä— ä—
‡Ñ#B@ = “¡ œ–·‚ “¡ w(Œ w œ)–·‚
(2. 81)
s$ nÃÕ ( nÃÕ
“„,« ä— “„,ä ä—
P a g e | 46
2.4. References
Bard, A. J. & Faulkner, L. R. 2001. Electrochemical Methods: Fundamentals and Applications,

Wiley, New York.
Bard, A. J., Stratmann, M., Frankel, G. S. 2003. Encyclopaedia of Electrochemistry, Volume 4,
Corrosion and Oxide Films. Wiley-VCH Verlag GmbH, Weinheim, Germany.
Evans, U. R. 1937. Metallic Corrosion, Protection and Protection. Edward Arnold & Co,
London.
Fisher, A. 1996. Electrode Dynamics, Oxford University Press, New York.
Jones, D. A. 1996. Principles and Prevention of Corrosion. Prentice-Hall, Inc, Upper Saddle
River, New Jersey.
Kelsall, G.H. 2011. Lecture notes of Electrochemical Engineering elective course, Imperial
College London.
Newman, J. & Thomas-Alyea, K. E. 2004. Electrochemical Systems. 3rd Edition, New York
Pletcher, D. 1984. Industrial Electrochemistry. Chapman and Hall, London
Wagner, C. & Traud, W. 1938. On the interpretation of corrosion processes through the
superposition of electrochemical partial processes and on the potential of mixed
electrodes. Elektrochem. 44, 391.
Walsh, F. C. 1993. A First Course in Electrochemical Engineering. The Electrochemical
Consultancy, Romsey, Hant.
P a g e | 47
3. Chapter 3
Literature Review
This chapter presents a general overview of the amine scrubbing process, including a discussion
on the different types of amines currently used for the CO2 absorption process, their chemical
properties and chemical reactions with CO2. This is followed by a discussion on the operation
difficulties commonly encountered on amine plants, with a particular focus on corrosion. This
chapter also gives an in-depth literature survey of the current views on the corrosion
mechanisms in aqueous CO2 systems and a description of the various forms that such corrosion
damage can take. The important parameters influencing or exacerbating the rate of CO2
corrosion in amine scrubbing systems will then be discussed. Next, a review on the occurrence
and characteristics of corrosion product films formed on the steel surfaces in CO2-H2O
environments, such as protective iron carbonate, will be presented. Finally, a review of the
mechanisms of iron dissolution and hydrogen evolution on iron is presented, including data for
exchange current densities and Tafel slopes at various pHs and temperatures.
P a g e | 48
3.1. The CO2 Absorption Process
As discussed briefly in Chapter 1, existing coal-fired power plants provide about 30% of the
total power generated and represent more than 25% of CO2 emissions in the United Kingdom.
In the United States, coal-fired power plants have a power capacity of more than 300,000 MW,
providing about 50% of the total power generated nationally resulting in more than 30% of CO2
emmisions (Rochelle, 2009). A sensible strategy for mitigating anthropogenic climate change
must address the issue of CO2 emissions without shutting down these plants. At present, CO2
capture by amine scrubbing is the only technology for post combustion capture that is available
to tackle this issue and will be discussed in detail in the following section.
CO2 removal by amine absorption is a robust, well understood and widely used technology.
The first amine scrubbing process was invented and patented by (Bottoms, 1930). The
absorption plant typically consists of an absorber and a stripper in which the absorbent is
regenerated at a high temperature. The continuous amine scrubbing process is based on a
temperature-dependent-reversible reaction between a weak base (amine) and a weak acid (CO2)
to form a soluble carbonate salt. In the absorber, which is operated at low temperatures, (ca.
40 oC), CO2 reacts with the amine exothermically to form protonated amine (RNH3+), hydrogen
carbonate (HCO3-), and carbamate (RNHCO2-) (Isaacs et al., 1980). Impurities such as sulfur
and nitrogen oxides are first removed from the flue gas, prior to its entering the absorber, to
prevent their preferential reaction with the amines, which could reduce the capacity for CO2 or
irreversibly contaminate the solvent. Acid gases such as SO2, SO3 and NOx react with amines
to form heat-stable salts, which would reduce the CO2 absorption capacity of the solvent, so
extremely low concentrations (less than 10 ppm in total) of these contaminant gases are required
to prevent excessive loss of solvent (Department of Trade Industry, 2005). The flue gas enters
a packed bed absorber from the bottom and is contacted counter-currently with the aqueous
amine solution. The amine proceeds to selectively absorb the CO2. When the amine solution
has been loaded with CO2, it is passed to a stripper and is regenerated with water vapour at
P a g e | 49
100oC to 120oC to reverse the chemical reaction and release CO2. After regeneration, the CO2-
lean solvent is then recycled back to the absorber to absorb more CO2. The pure CO2 captured
from the stripper is then compressed to 10 to 15 MPa for transportation and geological
sequestration. The schematic (Figure 3-1) below shows a more detailed description of the
absorption/desorption process.
Figure 3-1: Amine CO2 capture system flow diagram
1. The flue gas from the power plant enters through the bottom of the absorption column
and flows upward.
2. The absorber operates at a pressure of 0.1 MPa and a temperature range of 40-60oC. A
mixture of water and amine-based solvent flows down the absorber and CO2 reacts with the
amine to form carbamate salt. The absorber is filled with packing to maximise the contact
between both fluids and ensure adequate residence time for the reaction to occur.
3. CO2-free gas leaves at the top of the absorber.
4. The ‘rich’ amine solvent, with a CO2 loading (α) of around 0.40 leaves through the bottom
of the absorber and is preheated by a hot stream of ‘lean’ regenerated solvent (from the stripper
/ reboiler) in a heat exchanger.

P a g e | 50
5. The now heated ‘rich’ solvent is passed to the top of the stripping column. The stripper
also operates at a pressure of 0.1 MPa but at a higher temperature range of 100-120oC. As the
solvent flows down the column, CO2 is released gradually until it is further heated to saturation
in the reboiler, which completes the CO2 removal process.
6. The hot ‘lean’ amine formed, α ≈ 0.15, then flows back to the heat exchanger, in which it
is cooled before being recycled into the absorption column.
7. The CO2 leaves through the top of the stripping column at about 80oC, saturated with
water vapour. The mixture is then cooled and condensed before compressing the CO2 to prepare
for transportation.
For over 30 years, amine scrubbing with aqueous monoethanolamine solutions has been used
to capture CO2 from flue gas streams resulting from various processes ranging from fossil-
fuelled boilers to petrochemical process heaters. The first large scale CO2 capture plant in the
United States was the Searles Valley Minerals Plant, Trona, California which has been in
continuous operation since 1978, capturing 400 ton day-1 of CO2 from flue gas resulting from
a coal-fired boiler. Presently, there are hundreds of other plants employing this technology to
remove CO2 from natural gas, hydrogen and other gases with low oxygen concentrations
(Rochelle, 2009). There are over 20 pilot-scale coal-fired plants with power outputs of 6 to
40 MW using 20% -30% monoethanolamine to separate CO2 from flue gas with substantial
oxygen content (Rochelle, 2009). In addition to MEA, other solvents such as aqueous ammonia
and proprietary hindered amines such as KS-1, have also been used successfully to capture CO2
from dilute gas streams (Global CCS Institute, 2012). According to the (Global CCS Institute,
2012) there are several near-term solvent-based technologies involving either ammonia or
proprietary amines which include Fluor’s Econamine FG PLUS, Mitsubishi Heavy Industries
KS1 solvent, Cansolv Technologies, Aker Clean Carbon and Alstom’s Chilled Ammonia
Process (ACAP).
P a g e | 51
The Fluor Econamine FG PlusSM technology is a monethanolamine-based process but also
employs proprietary inhibitors to control corrosion, oxidative degradation and thermal
degradation (Fluor, 2016). The addition of inhibitors permits higher concentrations of MEA to
be used (30-35 wt% versus 18-20 wt% used previously) which, in combination with process
optimisation such as heat integration, absorber intercooling and split flow configuration, is said
to increase the CO2 absorption rate and reduce the steam consumption of Fluor Econamine FG
PlusSM by over 30% in comparison to ‘generic’ MEA technology. Thus, Econamine FG PlusSM
is frequently used as a representative technology for economic evaluation of post-combustion
CCS technologies.
Another near-term technology for coal-fired power plants is the Mitsubishi Heavy Industries
(MHI) KS-COR process that uses KS-1, a proprietary sterically-hindered solvent (Mitchell,
2008). KS-1, developed by MHI and Kansai Electric Power Company, is claimed to be
advantageous in comparison to conventional MEA for a number of reasons, including higher
CO2 loading, lower circulation rate (60% of MEA), lower stripping energy requirements (68%
of MEA) and lower solvent loss and degradation (10% of MEA). However, the unit cost of KS-
1 is higher than that of MEA by a factor of 5. The KS-1 solvent also requires significantly lower
levels of SOx and NOx (typically 1 ppm), so has higher pre-treatment and upstream polishing
requirements.
The Alstom’s Chilled Ammonia Process (ACAP) is a more recent technology based on
exploiting the changes in chemical equilibrium of ammonia, ammonium carbonate and
ammonium hydrogen carbonate with changes in temperature (Alstom, 2016). Being a low-cost
solvent and resistant to degradation, ammonia allows for stripping at higher temperatures and
exposure to NOx and SOx. The AEP’s Mountaineer presently employs this technology and is
able to achieve CO2 capture with 80-90% efficiency and at 99.9+ % purity (AEP, 2016).
However, the ACAP process typically requires a significantly larger absorber, due to the
relatively low absorption rate and CO2 loading capacity. Another issue commonly encountered
P a g e | 52
with the use of ammonia for CO2 capture is ammonia slip, which results from the high volatility
of ammonia leading to large amounts of ammonia evaporating and leaving the absorber with
the flue gases.
The Powerspan ECO2TM process was reformulated to use a proprietary solvent, which
incorporates a mixture of amines (Powerspan Clean Technology, 2016). With this solvent,
Powerspan announced in May 2010 that the technology was ready for scale up with a cost of
less than $40 per ton of CO2 captured and compressed (Powerspan Clean Technology, 2016).
Cansolv, owned by Shell Global Solutions International B.V, developed aqueous diamine
solvents for the selective removal of CO2, SOx, and NOx and mercury. With this technology,
90% CO2 capture rates have been achieved and it has since been chosen for the retrofit of a
150 MWe unit at SaskPower’s Boundary Dam.
In addition to the technologies mentioned above, Just Catch Technology was developed by
Aker Clean Carbon and had undergone several refinements throughout the slipstream testing
phase in a flue gas plant at Scottish Power’s Longannet station. In 2009, Scottish Power
announced that Just Catch could achieve a one-third reduction in energy requirements in
comparison to reference plants by using of low-energy solvents. Alstom and Dow Chemical
Company are also collaborating on the development of DOW UCARSOLTM FGC 3000, a
proprietary Dow amine solvent for removal of CO2 from flue gas streams from a coal-fired
boiler at a DOW chemical plant in South Charleston, West Virginia. The Dow amine is also
claimed to have a higher CO2 capture efficiency, higher tolerance to oxygen and other trace
contaminants and to be less susceptible to degradation. Doosan Babcock Energy and HTC
Purenergy are working jointly on the commercialization of a mixed amine solvent process that
was originally developed and field tested at the University of Regina, Saskatchewan (Global
CCS Institute, 2012). HTC also claims a reduced regeneration energy of less than 1.0 ton steam
per ton CO2 which is about 30% less than conventional 30 wt% MEA.
P a g e | 53
Apart from amine technologies, Siemens Energy and TNO are also working on developing an
amine acid salt based technology (Alston, 2016). Amino acid salts are considered to offer
several advantages over amine solvents, such as rapid CO2 absorption kinetics, higher CO2
loading capacity, lower solvent flow rate, lower regeneration requirements, degradation
resistance, lower volatility and less environmental impact (Brouwer, 2005). The different types
of amines and their respective properties are discussed below.
3.2. Solvents for CO2 Absorption
The most extensively used group of amine solvents for CO2 capture are alkanolamines, which
are compounds that contain both hydroxyl (-OH) and amino functional (-NH2, -NHR and -NR2)
groups on an alkane backbone. The attractive advantage of alkanolamines is that, while the
amino group provides the alkalinity to absorb CO2, the hydroxyl group reduces their vapour
pressure and increases their solubility in aqueous solutions (Strazisar et al., 2002).
Alkanolamines can be classified structurally as primary, secondary or tertiary, based on the
number of carbon-containing groups attached to the nitrogen atom or the number of hydrogen
atoms that have been replaced by the organic functional groups. Amines such as
monoethanolamine (MEA) and diglycolamine (DGA) with only one carbon containing group,
are classified as primary amines, while secondary amines such diethanol amine (DEA) and
diisopropyl amine (DIPA) have two carbon containing groups and methyldiethanol amine
(MDEA) with three carbon containing groups falls into the tertiary amine category. Some of the
most commonly used amines in CO2 capture and the relevant chemical structures are listed in
Table 3-1.
3.2.1. CO2 Absorption Reactions
Alkanolamines have been found to exhibit varying reaction rates, equilibrium absorption
characteristics as well as different degradation and corrosion rates (Global Climate & Energy
Project, 2005). The reactivity of amines to CO2 has been found to have a strong dependence on
P a g e | 54
the structure, with primary amines displaying the fastest reaction kinetics, followed by
secondary, then tertiary amines. For example, the reaction constants for MEA, DEA and MDEA
with CO2 are 7000, 1200 and 3.5 mol m-3 s-1 kmol-1 respectively at 25oC (Sada et al., 1976;
Hikita et al., 1979; Bishnoi and Rochelle, 2000). MEA, developed over 70 years ago, is at
present the most extensively used amine and acts as the benchmark for the CO2 absorption
process (Isaacs, 1980). It has several notable advantages such as fast reaction kinetics, achieved
due to low steric hindrances around the nitrogen atom. MEA is also readily available at a
relatively low cost, has a high solubility in water and a low absorption of hydrocarbons when
used for natural gas cleaning processes (Isaacs, 1980; Lepaumier et al., 2009). However, it also
suffers from the severe drawback of a high enthalpy of solution leading to high energy
requirements in the stripper, and a susceptibility to degradation, as well as being corrosive in
nature. The secondary amine DEA has greater steric hindrance, which ultimately slow down the
reaction kinetics. This in turn implies that the energy requirements for the release of CO2 in the
stripper are much less than that required with the use of MEA, so is less parasitic.
RNHCOO- + RNH3+ + H2O
1: Formation of carbamate 2. Hydrolysis of carbamate
2RNH2 + CO2 + H2O 2RNH3+ + CO32-

3.
RNH3+ + HCO3- + RNH2
Figure 3-2: Reaction Scheme of Alkanolamines with CO2

P a g e | 55
While the chemistry of the reactions of amines with CO2 is not fully understood, a series of
reactions were identified by (Danckwerts and McNeil, 1967) and have been generally accepted.
Alkanolamine-CO2-H2O solutions at equilibrium undergo a series of reactions to form the
carbamate anion (RNHCOO-), protonated cation (RNH3+), hydrogen carbonate ion (HCO3-) and
carbonate ion (CO32-) (where R stands for hydrogen or an organic group such as CH3 or
CH2CH2OH).
(Isaacs et al., 1980) found that the most commonly used primary and secondary amines MEA,
DEA, DIPA and DGA exhibit limited thermodynamic capacities to absorb CO2. The inability
of these common alkanolamines to absorb CO2 much beyond 0.5 mol CO2 (mol amine)-1 is due
to the stability of the carbamates formed with amines in which the amino nitrogen is attached
to a primary alkyl group (Park et al., 2003). In the case that formation of carbamate is the only
reaction (reaction path 1), the reaction stoichiometry indicates that 2 moles of amine are required
for each mol of CO2. However, the formation of hydrogen carbonates by hydrolysis (reaction
path 3) creates free amines, which can then further react with the CO2 molecules and enhance
the loading capacity. The reaction stoichiometry for the reaction involving hydrogen carbonate
ions shows a one to one ratio.

P a g e | 56
Table 3-1: Chemical Structures of the amines commonly used in CO2 capture
Class of Amine Alkanolamine Structural Formula

Primary Amines Monoethanolamine MEA HO-CH2-CH2-NH2
Diglycolamine DGA HO-CH2-CH2-O-CH2-CH2-NH2
Secondary Amines Diethanolamine DEA HO-CH2-CH2
NH
HO-CH2-CH2
Diisopropanolamine DIPA OH
CH3-CH-CH2
NH
CH3-CH-CH2
OH
Tetiary Amines Methyldiethanolamine MDEA HO-CH2-CH2
N CH3
HO-CH2-CH2
Sterically Hindered 2-amino-2-methyl-1- AMP CH3

Amines propanol
HO-CH2-C-NH2
CH3
Piperazine PZ NH
CH2 CH2
C-OH CH2
NH
P a g e | 57
While the reaction of primary and secondary amines with CO2 leads to the formation of
carbamates, tertiary amines such as MDEA exhibit a different limiting behaviour as they are able
to form hydrogen carbonate ions, but unable to form carbamates due to the absence of available
hydrogen atoms. (Jou et al., 1995) found that the absorption of CO2 in MDEA can reach 1.0 mol
of CO2 (mol of amine)-1. Although this quality of high CO2 loading capacity is highly desired,
the low absorption rate due to steric hindrance limits the attractiveness of MDEA. On a positive
note, they require much less heat in the reverse reaction, due to their inclination to release the
absorbed CO2.
In more recent years, sterically hindered amines such as AMP (2-amino-2-methyl-1-propanol)
and piperazine (a cyclic diamine with the molecular formula C4H10N2) have been developed due
to their high CO2 loading capacity and ease of regeneration. The stability of the carbamate formed
from AMP is decreased, due to steric hindrance of the bulky substituent group adjacent to the
amino nitrogen (Park et al., 2003). The unstable carbamate can undergo hydrolysis to form
hydrogen carbonate ions via reaction scheme 3 while still maintaining a relatively high
absorption rate for CO2. Despite advantages such as AMP having an 80% higher CO2 release rate
than MEA (Yeh and Pennline, 2001), it has been found that sterically hindered amines do not
always have a higher CO2 loading capacity than the more commonly used amines such as MEA.
The energy savings that are achievable from using sterically hindered amines are balanced out
by the higher capital costs associated with the larger scrubbing equipment required, due to slower
absorption rates. Experimental results show that under certain conditions such as low CO2 partial
pressures, the solubility of CO2 is lower in hindered amines than in the MEA solutions (Park et
al., 2003). The absorption of CO2 usually takes place at atmospheric pressure or at low CO2
partial pressures, so it is particularly important to select an absorbent with a high solubility of
CO2 even at low pressures.

P a g e | 58
It has proven to be extremely difficult to identify any single amine that has all the desirable
properties, such as rapid absorption kinetics, low energy requirements for desorption, high
chemical stability, low vapour pressure, resistance to degradation and low corrosion rates etc.
This issue has led newer technology to involve the use of a mixture of a number of amines. It
has been suggested that the combination of primary or secondary amines with tertiary amines
might actually synergise their individual advantages (Park et al., 2003). The aim is to have a
mixture of amines retaining the reactivity of the primary or secondary amines at comparable or
lower circulation rates, but reducing regeneration costs in a similar manner to tertiary amines.
Several blends have been studied experimentally such as MEA/MDEA, DEA/MDEA,
MEA/AMP, DEA/AMP, MDEA/AMP, MDEA/DEA/AMP, PZ/MDEA, DIPA/MDEA, and
DIPA/PZ (Park et al., 2003). There have been significant reductions in energy requirements, as
well as circulation rates that have been reported for amine blends in comparison to a single
amine system (Strazisar et al., 2003).
In addition, traditionally used amines have been blended with amines that are less corrosive and
require less steam to regenerate or involve the use of the additive piperazine (PZ) (Global
Climate & Energy Project, 2005). Examples of such include MEA-PZ blends, MDEA-PZ
blends, blends of Nmethyldiethanolamine and triethylene tetramine (Shao and Stangeland,
2009). While PZ is less soluble in water and more volatile than MEA, its presence increases
CO2 absorption and allows the use of smaller concentrations of MEA (Dang and Rochelle,
2003). Research has also shown that solvents based on PZ-promoted K2CO3 can have reaction
rates similar to that of MEA, but their capacity has been found to be significantly lower
(Cullinane and Rochelle, 2003).
Degradation of alkanolamine-based solvents on contact with oxygen is a great disadvantage.
Typical flue gas streams contain oxygen in the order of 3-12% which can bring about the
oxidative degradation of the amines (Goff and Rochelle, 2004). This will cause a reduction in
P a g e | 59
their CO2 absorption capacity and lead to the production of unwanted degradation products,
leading eventually to the amine solution having to be discarded. These degradation products
often lead to severe corrosion (Goff and Rochelle, 2004). The operational cost savings that have
been reported with blended amines would only be possible if the loss of amine due to
degradation and the corrosion of the equipment is less than or comparable with losses from
single amine systems. Presently, there is very little information on the degradation of mixed
amines and steel corrosion behaviour. (Dawodu and Meisen, 1996) studied the absorption of
CO2 with MEA/ MDEA and DEA/MDEA mixtures and found that higher circulation rates were
required for DEA in the MDEA/DEA blend than the MEA in the MDEA/MEA blends. They
also reported that the blends affected the rate of degradation, but were unable to establish a firm
relationship between the blends and the degradation rates (Strazisar et al., 2003).
3.3. CO2 Induced Corrosion in Amine Plants
Amine absorption-desorption is the process of choice for CO2 capture, because coal power
plants can be retrofitted easily and have been reported to be comparatively free from operational
difficulties. However, there are several issues apart from solvent degradation that can lead to
additional costs and result in a significant decrease in capacity for CO2 capture. From an
economic perspective, the major sources of such operational difficulties are corrosion, amine
loss, foaming and plugging of equipment.
Corrosion can be defined as an irreversible reaction of a material with the environment, which
usually results in the degradation of the material or its properties and is thought to be the most
severe of the operational difficulties facing the equipment used in alkanolamine treating plants
(Kohl and Nielsen, 1997; Veawab and Aroonwilas, 2002). It is a natural potential hazard
associated with amine scrubbing plants, due to CO2 being dissolved in the aqueous solution
phase, the inherent corrosivity of which, is dependent on the concentration of CO2 and the water
P a g e | 60
chemistry. CO2 corrosion or ‘sweet corrosion’ is also the predominant form of attack and is a
serious cause of concern in many other applications using carbon and low alloy steels, such as
oil and gas production and transportation facilities. The occurrence of such corrosion can lead
to unscheduled downtime, loss of products, reduced equipment life and in extreme cases injury
or death.
Despite extensive previous experimental research and the development of predictive models,
the corrosion process is still not fully understood and there are still ambiguities with respect to
the parameters that affect it. Whilst it would be impractical to eliminate corrosion completely,
it is possible to minimise or mitigate its occurrence; to achieve this, a good understanding of
the corrosion chemistry is required, involving the kinetics, product identification,
stoichiometry, and mechanisms.
3.3.1. Plant Experience
Being the most serious operational setback presently faced on alkanolamine gas purification
plants, corrosion has been given the most consideration and research. There has been a
considerable amount of information and reports in the open literature about amine scrubbing
plants and the experiences with corrosion issues under a wide range of conditions. Table 3-2
lists a summary of the plants, amine type, details about the type of corrosion and the ensuing
issues. Corrosion occurs in various units within the plant, such as the bottom of the absorber,
the heat exchanger, the stripper, the overhead condenser, pumps and valves where the acid gas
loading and temperatures are high (Soosaiprakasam and Veawab, 2008).
Several factors have been attributed to aggravating corrosion issues, such as poor plant design
and operation, i.e. high amine solution velocity in pipelines, high operating temperatures in the
stripper and reboiler, as well as the presence of contaminants and degradation products. The
construction material is also an important factor in determining the susceptibility to corrosion.

P a g e | 61
Whilst it is less costly in terms of the initial up-front capital investment to construct the plant
entirely from carbon steel, it is common practice to use stainless steel or other corrosion-
resistant alloys in certain sections, which allows for plant operation under more ‘aggressive’
operating conditions, such as high amine concentrations and CO2 loadings.
Corrosion in amine scrubbing plants has been reported as uniform thinning or as a localised
attack, such as pitting, erosion, intergranular and stress corrosion cracking. As listed in Table
3-2, general or uniform corrosion has been observed to occur at various parts of the plant such
as the bottom of the absorber or in the regenerator trays, but the occurrence has been noted with
other localised attacks such as pitting and galvanic corrosion (Heisler and Weiss 1975; DuPart
et al., 1993). Although uniform corrosion of metals in aqueous CO2 environments has received
more attention by researchers, localised attacks are actually of greater concern as they are more
likely to result in unexpected failures. Localised forms of corrosion are characterised by loss of
metal at distinct areas of the metal surface, while other surrounding areas are seemingly
unaffected or undergo general corrosion. These forms of localised corrosion may appear on the
metal surface in various geometrical shapes. Pitting corrosion presents itself as circular
depressions, usually with tapered or smooth sides. Messa type corrosion attack is characterized
by a deep and sometimes flat-bottomed local attack which covers a large surface area with very
sharp edges. Localised corrosion can also take other geometrical shapes such as slits, also
known as knifeline attack. Flow induced localised corrosion can be identified by localised
attacks in the form of parallel or shallow grooves related to the direction of flow.
Localised forms of corrosion are connected strongly to the formation of passive layers on the
metal surface. The presence of the passive layer acts to protect the metal surface against
degradation and dissolution; however, the presence of aggressive anions can lead to local
breakdown in the passive layer and the formation of corrosion pits. Pitting formation has long
been recognized in CO2 systems from field observations (American Petroleum Institute, 1958;
P a g e | 62
Hobler and Schonewolf, 1974; Hackerman and Schmidt, 1949) with a relatively small number
of experimental investigations on pitting phenomena and localized attack (Schmitt and
Rothman, 1978; DeBerry et al., 1980). This may be due to the long induction period, sometimes
on the order of several months, required for pit initiation. For pitting corrosion to occur, the
passive layers formed on the surface of the metal must have local defects. These imperfections
may be as a result of a non-uniform growth of passive layers and/or the mechanical disruption
by hydromechanic attack (high fluid shear stress, erosion, cavitation) or by scratches on the
coating, uneven surface coverage of the corrosion inhibitor, scale formation, dust and other
impurities. The non-uniform growth of the passive layers is influenced by alloy effects,
metallurgical factors, the presence of certain chemicals, CO2 partial pressure, pH and
temperature. The pit then grows due to differences in the chemical compositions of the aqueous
solution below the pit and the directly surrounding steel surface. The bottom of the pit acts as
the anode and the nearby steel acts as the cathode. It is also known that all steels may undergo
pitting corrosion in CO2 environments under the enabling conditions (American Petroleum
Institute, 1958; Schmitt and Rothman, 1978; DeBerry et al., 1980; Hackerman and Schmidt,
1949). Whilst stainless steels become susceptible to pitting corrosion only in the presence of
aggressive anions like chlorides and small quantities of oxygen, medium and especially low
alloy carbon steels need only the CO2 aqueous system. Generally aggressive operating
conditions, such as high temperature have been found to exacerbate the formation of pitting
corrosion and to decrease the time it takes to occur (Hausler and Goddard, 1984). Pit formation
was observed frequently in the dead zones and hot regions, such as the bottom of the absorber,
the regenerator valve trays and the tube bundles of the reboiler (DuPart et al., 1993;
Litschewski, 1996). So far, the existence of a critical CO2 partial pressure below which pitting
formation does not occur, has not yet been identified. However, in practice, it is the rule of
P a g e | 63
thumb that corrosion damage from pitting is unlikely below 50 kPa CO2 (American Petroleum
Institute, 1958).
Mesa type corrosion occurs mainly when protective passive layers such as iron carbonate form,
but are not very stable and do not offer complete protection. As film formation generally occurs
at elevated temperatures, typically above 60oC, mesa type corrosion is less likely to occur at
temperatures below this. According to (Nyborg,1998), the attack starts as several small pits
beneath a porous passive film, which continue to grow beneath the film until the top of the
passive film is torn away by mechanical forces such as turbulent flow. As the attack progresses,
the pit continues to grow both laterally and in depth and the original passive film is removed
stepwise by flow and with dissolution of the underlying metal. As several pits are initiated, they
start to grow together into a wider mesa attack of nearly uniform depth. As long as the protective
films are not reformed, the mesa attack continues to grow with a high corrosion rate both
laterally and in depth. However, on reaching a certain scale, initiation of new pits ceases, due
to galvanic effect between the underlying corroding metal at the bottom of the mesa attack and
the film-covered metal surrounding the mesa attack. Whilst the initiation of pits is independent
of the flow, the growth from small pits into a larger mesa attack is dependent on the flow
conditions.
Galvanic corrosion or dissimilar metal corrosion occurs when two dissimilar metals are in
electrical and ionic contact. This type of corrosion typically occurs in the junction between two
metals; the metal with the lower electrode potential acts as the anode and corrodes
preferentially, while the metal with the higher electrode potential acts as the cathode and is
protected from corrosion. The degree of corrosion is most intense at the contacting point and
decreases further away from the point of contact. Galvanic corrosion in amine treating plants
has been observed in parts where both stainless steel and carbon steel were utilised. Examples
of locations in which galvanic corrosion occurred are in the stainless steel valve coupled with
P a g e | 64
a carbon steel disk (DuPart et al., 1993) and in the stainless steel regenerator trays in contact
with a carbon steel regenerator wall. The carbon steel (anode) preferentially corrodes while the
stainless steel (cathode) remains protected.
As the metal owes its corrosion resistance to the formation and resistance of protective scales,
the removal of such scales in localised areas can lead to accelerated attack. Erosion corrosion
is the deterioration of the metal caused by high amine solution velocities, solution turbulence
and impingement of gas and amine on metal surfaces. Rapid motion of the solution and any
abrupt changes in the direction of flow could either erode the surface of the metal directly or
remove any protective film protecting the piping and equipment from corrosion and finally
corrode the underlying metal. Similarly to the case of pitting corrosion, solid impurities such as
scale, dust and amine degradation products aggravate erosion. In parts of the plant in which
there is a considerable loss of pressure, ‘cavitation’ erosion can occur. Low pressure (below the
saturation vapour pressure of the flowing amines) causes vapour or gas bubbles to be nucleated
from the flowing solution which then bombard the metal surface. Areas that are at risk of
erosion-corrosion are piping with bend fittings, and amine pumps, while the areas damaged by
the cavitation are the absorber bottom, the flash drum, the rich amine pump, the heat exchanger
tubes, the regenerator and the reboiler (Dingman et al., 1966, Smith and Younger, 1972; Heisler
and Weiss, 1975; Gerus 1981; Dupart et al., 1993)
Four types of carbon steel cracking mechanisms have been identified in amine scrubbing plants.
The first three cracking mechanisms are associated with the entry of hydrogen atoms into the
carbon steel lattice. These mechanisms are known as the sulfide stress cracking (SSC),
hydrogen induced cracking (HIC) and the stress-orientated hydrogen induced cracking
(SOHIC) and they all are associated with the production of hydrogen from aqueous–H2S
solutions. However, in such environments, CO2 is considered to be a contributing factor rather
than a principal factor and as the removal of H2S is not of particular concern in this work, it
P a g e | 65
will not be discussed further. The fourth mode of cracking, stress corrosion cracking (SCC),
occurs as a result of a combined influence of tensile stress and a corrosive environment.
Examples of such tensile stress include residual stress from welding and fabrication or
externally applied stress. The most common type of stress-corrosion cracking thought to occur
on an amine plant is the Alkaline Stress-Corrosion Cracking (ASCC) (Kohl & Nielsen, 1997).
ASCC occurs due to a film rupture mechanism, and areas that are prone to high residual stress
slip, causing the protective passive film to break, exposing bare underlying metal, which then
undergoes corrosion and forms cracks. The passive film may re-form, but the residual stress
causes the film to rupture repeatedly leading to cracking. An explosion and a fire caused by the
rupture of an amine LPG treater in 1984, resulted in the death of 17 people (McHenry et al,
1986) and was found to have been most likely caused by stress corrosion cracking. This
unfortunate disaster and reports of cracking within other amine systems, led to a detailed
investigation of carbon steel cracking within amine units. A survey of 294 units conducted by
the National Association of Corrosion Engineers (NACE) in 1992 showed that 70% of the SCC
cases were found in MEA plants, 24 % due to the use of DEA, while the residual 6 % were
from plants utilising the other amine solvents (Dupart et al., 1993). Cracks were found to occur
in every type of equipment, even at low temperatures, but no correlations were established
between cracking and operating conditions, such as amine concentration, CO2 loading, use of
corrosion inhibitors or types of carbon steel etc.

P a g e | 66
Table 3-2: Summary of Amine Scrubbing Plant Experience on Corrosion
Plant Solvent Corrosion Type Susceptible area Reported cause Source

• Erosion • Rich lean heat exchanger • High solution velocity
Sour gas treating plant in MEA (Dingman, 1966)
• Letdown valve • Changes in direction of fluid flow
Texas • Piping downstream of • Contamination with solids such as iron
letdown valve oxide
• Erosion • Rich lean heat exchanger • Contaminants in the amine solution

24 sour gas treating plants in DEA (Smith, 1972)
• SCC • Stripper • High velocity flowrates
Western Canada • Reboiler • Chloride ions evolved gasket fabrication
• Rich solution piping material between heat exchanger plates
• Stainless steel heat exchanger
plate
• Uniform • Stripper, weld seam, joints • Cavitation due to vapour flash

Natural gas treating plants at MEA (Heisler and Weiss
• Pitting and tray
Aderlas, Austria • Erosion 1975)
• Pitting • Absorber bottom • Acid gas flashing

Sour gas treating plants DIPA (Schmeal et al.,
1978)
• N/A • Stripper • N/A

Ram river gas plant located at DEA (Hall and Barron
• Rich lean heat exchanger
Canadian Rocky Mountains • Hot side of the cooler 1981)
Fronthills
P a g e | 67
Table 3-3: Summary of Amine Scrubbing Plant Experience on Corrosion (Continued)
Plant Solvent Corrosion Type Susceptible area Reported cause Source

• Pitting
Natural gas treating plants N/A N/A • N/A (Gerus 1981)
• Erosion
• Pitting • Absorber bottom • Aqueous CO2 flashing

MDEA (Dupart et al., 1993)
• Erosion • Cavitation
• High velocity impingement
• Erosion • Wall of absorber bottom
MDEA Turbulent flow (Dupart et al., 1993)
• Uniform • Absorber bottom

MEA Oxygen contaminants (Dupart et al., 1993)
• Galvanic
• Pitting • Stripper at carbon steel; tray • Aqueous CO2 flashing

MDEA (Dupart et al., 1993)
• Erosion and 304 stainless steel valve • High CO2 loading
• Galvanic opening • Insufficient stripping steam
• Shell side of the heat
exchanger
• 304 stainless steel steel valve
connected to carbon steel disk
P a g e | 68
3.4. Corrosion Mechanisms in Amine Plants
As depicted in Table 3-2, most of the equipment and piping in alkanolamine treating plants is
typically constructed of carbon steel. (Kohl & Nielsen, 1997) suggested that it might be possible
to build the entire alkanolamine plant completely from carbon steel by maintaining a low amine
stripper temperature, minimizing amine solution concentration and CO2 loading. However, it is
common practice to construct the parts of the plant that are more susceptible to corrosion with
stainless steel or more corrosion–resistant alloys, to allow for operation under more aggressive
conditions. The parts of the plant schematically shown in Figure 3-3 marked in the bold line
show the parts that are typically constructed from stainless steel as opposed to carbon steel. The
red and blue dotted lines represent the units succeptible to wet acid gas corrosion and amine
solution corrosion respectively.
Figure 3-3: Flow diagram of a typical amine plant showing principal areas of corrosion and
suggested materials of construction (Kohl & Nielsen, 1997).
P a g e | 69
(Kohl & Nielsen, 1997) suggested that corrosion within amine scrubbing plants can be
classified into two broad categories:
• Wet acid gas corrosion of the carbon steel is the reaction of CO2 with iron in an aqueous
environment, in which little or no amine is present.
• Amine solution corrosion is the corrosion of carbon steel in the presence of aqueous
amine.
These two mechanisms will be discussed in detail below.
3.4.1. Wet Acid Gas Corrosion
Wet acid gas corrosion is most likely to occur in the parts of the plant in contact with aqueous
phases with a high concentration of acid gases such CO2, e.g., at the top of the regenerator and
at the bottom of the absorber (Veawab and Aroonwilas, 2002). Due to the severe corrosion
issues associated with oil and gas production and transportation facilities, the corrosion
mechanisms of carbon steel in CO2–H2O systems under varying conditions, such as pressure,
temperature and pH, has been reported extensively in recent literature. Some of the most
important findings feature the works of (Bockris et al., 1961, de Waard & Milliam, 1975; de
Waard et al, 1991; Dugstad et al, 1991; Drazic et al and Nesic et al, 1996). Thus, most of the
publications on corrosion within the context of amine scrubbing technology have drawn
conclusions from the vast research conducted on carbon steel corrosion in other CO2-H2O
systems, rather than strictly in CO2 loaded amine systems.
In spite of the immense amount of research on CO2 corrosion phenomena, including
experimental research and the development of predictive models, the corrosion process is yet
to be fully understood and there are still some ambiguities about the corrosion mechanisms
and the effect of operating parameters. The mechanisms that have been proposed for wet acid
P a g e | 70
gas corrosion in amine systems differ depending on the operating conditions such as type of
amine being used and absence or presence of oxygen.
The corrosion of carbon steel in CO2 - H2O environments is considered to be a heterogeneous
process with a number of chemical, electrochemical and transport processes occurring
simultaneously and the overall corrosion rate is controlled by the slowest of these steps. When
the transport processes occur at a faster rate than the interfacial electrochemical reactions, the
corrosion rate is governed by the electrode kinetics and is said to be under ‘activation’ or
‘charge transfer’ control. On the other hand, when the electrochemical reactions proceed at a
higher rate, the corrosion rate is controlled by mass transport processes and the overall process
is under transport control.
3.4.1.1. Chemical reactions
Whilst dry carbon dioxide gas itself is not corrosive, on dissolving in water it forms a weak
carbonic acid. This aqueous phase in contact with the carbon steel equipment enables
electrochemical reactions to occur, resulting in corrosion of the steel. Different mechanisms
have been proposed for the CO2 induced corrosion process, but they all involve either carbonic
acid (H2CO3) or hydrogen carbonate ions (HCO3-).
The concentration of the aqueous CO2 at a given CO2 partial pressure and temperature can be
calculated from Henry’s law:
!"#(%) ↔ !"#(()) (3. 1)
!"# = , ∙ ./01 (3. 2)
here, [CO2] is the concentration of CO2 in the aqueous phase, H is the Henry’s constant and
PCO2 is the partial pressure of CO2 in the gas phase.

P a g e | 71
When CO2 is the only acid gas present, it is hydrated in water to form carbonic acid according
to the following reaction:
!"#(%) + ,# "(()) ↔ ,# !"3 (()) (3. 3)
The carbonic acid then partially dissociates to form hydrogen carbonate ion (HCO3-) which
then dissociates to carbonate ion via the following two steps:
,# !"3 (()) ↔ ,!"3 4 (()) + ,5 (()) (3. 4)
,!"3 4 (()) ↔ !"3 #4 (()) + ,5 (()) (3. 5)
3.4.1.2. Electrochemical Reactions at the Carbon Steel Surface
The corrosion of carbon steel occurs by the oxidation of iron via reaction (3. 6):
67 → 67 #5 (()) + 27 4 (3. 6)
The electrons released by the anodic reaction must then be consumed by a cathodic reaction
of equal rate. Earlier studies conducted by researchers in the field have shown that the
corrosion rate of carbon steel in aqueous CO2 environments, in the absence of oxygen and at
ambient temperature, is controlled primarily by the kinetics of one or several reactions
resulting in the evolution of hydrogen. (Schwenk, 1974) concluded that H+ ions are reduced
via reaction (3. 7).
2,5 + 27 4 ↔ ,# (3. 7)
The hydrogen evolution reaction is considered to be the dominant reduction reaction, the rate
of which is limited by the mass transfer of H+ ions to the iron surface from the bulk solution.
However, it has been established that the mechanism in which H+ ions are supplied to the
electrode surface in CO2-containing solutions differs from that occurring in strong acid
P a g e | 72
solutions, in which diffusion of H+ ions from the bulk solution is the principal rate controlling
reaction. There have been several additional electrochemical reactions proposed and several
inconsistencies in the open literature about the species reacting at the electrode to form
hydrogen. While there are still several different mechanistic interpretations of CO2 corrosion,
there appears to be a general agreement on the corrosion rates observed under similar
conditions (Eriksrud and Sontvedt, 1983; Ikeda et al., 1984; Schmitt, 1983; Dunlop et al., 1984;
Murata et al., 1985). The corrosion of steel was found to obey the relationship developed by
(de Waard and Milliams, 1975).
log = = 0.67 (log ./01 ) + ! (3. 8)
For this relationship, different validity limits were established by each researcher: (Schmitt,
1983) PCO2 < 0.2 MPa; (Eriksrud and Sontvedt, 1983) PCO2 < 0.1 MPa, 20oC; (Dunlop et al.,
1984) 82oC, velocity below critical (VC = 7.6/√p) ; (Ikeda et al., 1984) PCO2 = 0.2 MPa, 60oC;
(Murata et al., 1985) PCO2 = 0.7 MPa, 60oC.
(Schwenk, 1974; Schmitt and Rothman, 1977; Fisher and Siedlarek, 1978) each observed a
limiting cathodic current density, in solutions fully saturated with CO2, at pH 4, that was
several orders of magnitude higher than could be expected in solutions of equal pH in the
absence of CO2. At the low temperature conditions under which the measurements were
conducted (≤ 20oC), varying the solution flow rate was found to have little to no effect on the
corrosion rate, so excluding the diffusion of H+ ions as the rate controlling step. As a result, it
was concluded that it was not strictly the hydrogen evolution reaction in the sense of a Volmer
or Bockris mechanism (Section 3. 11) which acts as the rate controlling reaction, instead it was
a reaction which precedes the charge transfer reaction. It is possible that the presence of
carbonic acid and hydrogen carbonate ions formed from the hydration of CO2 provides a
reservoir of H+ at a particular pH, thereby increasing the corrosion rate by serving as a source
P a g e | 73
of H+ via reactions (3. 4) and (3. 5), which are subsequently reduced according to the hydrogen
evolution reaction equation (3. 7) (Schwenk, 1974). Thus for a solution of pH > 4, CO2 being
present would also lead to a much higher corrosion rate than would be found in a strong acid
of equal pH.
(Schwenk, 1974) proposed that the limiting current is determined by the diffusion of CO2 to
the surface of the steel. However, this appears to be improbable as the concentration of CO2 in
CO2-saturated solutions is approximately 27 mol m-3 and the diffusion coefficient at 298 K is
about 10-9 m s-1 (Schmitt, 1983; Schwenk, 1974). This results in a diffusion limiting current
density of 52 A m-2, which is in the order a factor higher than the limiting current density
observed experimentally in quiescent solutions (Eriksrud and Sontvedt, 1983). As the observed
current densities were lower than the predicted limiting currents due to diffusion of CO2, a
purely diffusion controlled limiting by the diffusion of CO2 is improbable.
The corrosion rate at 20oC was found to be independent of solution flow rate, suggesting that
the rate-determining step was either a chemical or electrochemical reaction. If a chemical
reaction was the rate-determining step, then the corrosion rate would be determined by a
limiting current that was reaction controlled and independent of solution flow. However,
(Eriksrud and Sontvedt, 1983) discovered that the corrosion rate was independent of the
limiting current density, which increased with solution flow, so the corrosion rate was not
controlled completely by a chemical reaction. As a chemical rate determining reaction was
excluded, (Eriksrud and Sontvedt, 1983) and (Schmitt and Rothman, 1977) suggested that it
was more likely that a flow independent electrochemical reaction acts as the rate determining
step. As there was a small effect of solution flow rate on the limiting current at low flow rates,
they assumed that diffusion was partly controlling and there was a change from partial
diffusion control at low flow rates to charge-transfer control at higher flow rates. As the
P a g e | 74
diffusion rate increases with flow rate, on attaining a certain flow rate, the charge transfer
reaction becomes the slow step, making the corrosion rate independent of flow rate.
According to studies by (Eriksrud and Sontvedt, 1983), there also appeared to be a change in
mechanism with change in temperature from 20oC to 60oC. Whilst at low temperatures (≤
20oC), the corrosion rate was flow independent, at higher temperatures (≥ 60oC) it appeared to
be mass transport limited. This suggested that the corrosion process was controlled by the
kinetics of a reaction that increased with temperature. As the rate of the charge transfer
controlled reaction increases significantly more with temperature than diffusion rates, so at
20oC the rate determining step was the charge transfer reaction but at elevated temperatures,
diffusion became the slowest step and therefore rate-determining.
According to studies by (Gray et al., 1990; Stern, 1955), the slow hydration of CO2 to form
carbonic acid was considered to be the rate determining step, due to its slow chemical reaction
kinetics. Based on an assumption of the hydration of CO2 as the rate determining step, with
the rate coefficient for the forward-going reaction at room temperature given as 3×10-2 s-1
(Fisher and Siedlarek, 1978), (Eriksrud and Sontvedt, 1983) calculated a limiting current
density of 10 mA m-2, a value that is closer to the observed value for the limiting current density
at 20oC in stagnant solutions. However, this theory can also be discounted, as the limiting
current was found to be dependent of the hydrodynamic conditions of the test solution, i.e. the
rotation rate of the electrode, as mentioned previously. This led (Schmitt and Rothman, 1977)
to suggest that the observed limiting current results from both limiting transport rates of H+
ions and a rate limited chemical reaction. The interpretation of this ‘rate limited reaction’ is
also a source of discrepancies in the literature. Based on observed and calculated limiting
current densities, (Eriksrud and Sontvedt, 1983) suggested that the hydration of CO2 to form
carbonic acid is a homogeneous reaction, while (Schmitt and Rothman, 1977) proposed that
P a g e | 75
the overall cathodic reaction was controlled by the heterogeneously catalysed hydration of
chemisorbed CO2.
Three different reaction mechanisms were proposed by (Schmitt, 1983):
1) Hydrogen evolution by the reduction of H+ ions from the bulk solution after the
dissociation of H2CO3 and diffusion to the metal surface. The hydrogen ions resulting
from the dissociation of carbonic acid are transported from the bulk solution to the
metal surface. Adsorbed molecular hydrogen yields from chemical recombination
according to the Tafel reaction or electrochemical recombination according to the
Heyrovsky reaction (Section 3. 11) and desorbs from the surface to allow subsequent
adsorption and reduction of hydrogen ions. In this reaction sequence, there is no
significant difference between the hydrogen evolution reaction in CO2 solutions and
diluted acid solutions.
!"#(BCD) + ,# " ↔ ,# !"3(BCD) (3. 9)
,# !"3(BCD) ↔ ,5 (BCD) + ,!"3 4 BCD

(3. 10)
,5 (BCD) ↔ ,5 ((EB) (3. 11)
,5 ((EB) + 7 4 ↔ ,((EB) (3. 12)
2) CO2 molecules adsorb on the metal surface and subsequent hydration yields
chemisorbed carbonic acid. These species then dissociate while still adsorbed to give
hydrogen ions, which are subsequently reduced to yield adsorbed hydrogen atoms. It
is important to note a significant difference of reaction sequence (2) from sequence (1)
is that the hydration of CO2 occurs as a heterogeneous reaction on the metal surface.
P a g e | 76
!"#(BCD) ↔ !"#((EB) (3. 13)
!"#((EB) + ,# " ↔ ,# !"3 ((EB) (3. 14)
,# !"3 (EB
↔ ,5 ((EB) + ,!"3 4 (EB
(3. 15)
3) Hydrogen evolution from the direct reduction of adsorbed carbonic acid:
,# !"3 (EB
+ 7 4 ↔ ,((EB) + ,!"3 4 (EB
(3. 16)
The direct reduction of carbonic acid to form hydrogen carbonate ions and hydrogen (reaction
scheme 3) was also proposed by (de Waard and Milliams, 1975) as an explanation for higher
rates of corrosion in CO2 containing solutions. The charge transfer reaction (3. 16) was also
assumed to be the rate determining step. From their experimental data (de Waard and Milliams,
1975) concluded that hydrogen evolution in CO2 solutions proceeds in a ‘catalytic’ manner, in
which adsorbed hydrogen carbonate from reaction (3. 15) is neutralized by a hydrogen ion to
re-form adsorbed H2CO3 from which more hydrogen can be directly or indirectly produced.
,!"3 4 (EB
+ ,5 ((EB) ↔ ,# !"3 ((EB) (3. 17)
(Schmitt and Rothman, 1977) also concluded that the adsorption of CO2 on the metal surface
must occur prior to hydration. This restriction seems unlikely as radiotracer studies conducted
by (Wieckowski et al., 1983) indicated that CO2 absorption is not required for the corrosion
reactions to proceed.
(Wieckowski et al., 1983) suggested an alternative scheme in which H2CO3 and HCO3-
reactions occur on the metal surface, by a catalytic mechanism in which the reactant is
regenerated by water.
P a g e | 77
,# !"3 + 7 4 ↔ ,((EB) + ,!"3 4 (3. 18)
,!"3 4 + ",4 ↔ !"3 #4 + ,# " (3. 19)
Or
,!"3 4 + 7 4 ↔ !"3 #4 + ,((EB) (3. 20)
!"3 #4 + ,# " ↔ ,!"3 4 + ",4 (3. 21)
Whilst (Schmitt, 1983) favoured a heterogeneously catalysed hydration of chemisorbed CO2
as the rate controlling step, on the basis that the limiting cathodic current densities were
strongly affected by the type of steel, (Eriksrud and Sontvedt, 1983) argued otherwise. They
were able to support their claims of a homogeneous hydration of CO2 on the basis that the
observed cathodic limiting current in CO2 solutions was the same as that found by (Schwenk,
1974) and others for platinum and steel. They remarked that a chemical reaction in solution
would be independent of the type of metal, while a heterogeneous adsorption reaction will
often depend on the metal type, since some sort of bonding is involved. They also noted this
as the reason for the influence of metal type on the hydrogen evolution rate, due to the
difference in the binding energy of adsorbed H atoms.
(Berry, 1982) gave an alternative explanation for the higher rates of corrosion observed for
CO2 solutions than those in strong acidic solutions with the same bulk pH values. The depletion
of hydrogen ions from the solution at the iron surface caused the solution adjacent to the
corroding steel surface to become less acidic, which in turn decreased the rate of corrosion.
(Barry, 1982) suggested that this decrease in local H+ ion concentration is counterbalanced by
the evolution of gases which increases turbulence of the liquid, thereby decreasing the
alkalinity by providing fresh acid solution to the metal surface. It was also proposed that in a
weakly ionized acid such as carbonic acid, there is a large amount of dissolved but un-ionized
P a g e | 78
CO2 in the solution, which can then go on to react via reaction (3. 3) to provide hydrogen ions,
replacing those consumed by the corrosion reaction. On the other hand, only relatively low
concentrations of a strongly ionized acid are required to produce the same pH as the weakly
ionized solution and the depletion of hydrogen ions can reduce the corrosion rate.
(Ogundele and White, 1986) suggested that due to the large hydrogen carbonate concentrations
the direct reduction of the hydrogen carbonate ion may also be significant. However, their
experiments covered only a narrow pH range under quiescent solution conditions.
2,!"3 + 27 4 ↔ ,# F + !"3 #4 (GH) (3. 22)
(Gray et al., 1990) also proposed the same electrochemical reduction of hydrogen carbonate
ions with different electrochemical reactions being dominant in different pH regions. They
proposed that in CO2 containing solutions at low pH (pH < 4), that hydrogen evolution from
H+ reduction is dominant and the hydration of CO2 is not thermodynamically favourable and
a very small concentration of carbon dioxide is present as carbonic acid. At an intermediate
pH range (4 < pH < 6), they proposed the direct reduction of carbonic acid in addition to
hydrogen ions. At a higher pH range (6 < pH < 10) in which hydrogen carbonate ions are
present in significantly higher concentrations, they expected HCO3- to be directly involved in
the reduction reaction. However, it is still debatable which one or more of the numerous
proposed mechanisms is appropriate, so the debate continues.
3.4.2. Amine Solution Carbon Steel Corrosion
(Lang and Mason, 1958; MacNab and Treseder, 1971 and Kohl and Riesenfeld, 1985) each
proposed that while alkanolamines contribute to corrosion by reacting with CO2, they are not
corrosive themselves. Amines are not intrinsically corrosive due to their high pH, they become
so only when they absorb CO2, leading to the second type of corrosion encountered: amine
P a g e | 79
solution corrosion. This corrosiveness is due partly to CO2-loaded amines having a significantly
lower pH. This type of corrosion is most prominent in the hot bottom section of the stripper and
the rich solution side of the heat exchanger. (Riesenfeld and Blohm, 1951) firstly noted that
substantial amine corrosion is usually accompanied by the evolution of acid gases from the rich
amine solution. Based on their observations, (Riesenfeld and Blohm, 1951) identified the
presence of CO2 as the principal cause of carbon steel corrosion in amine solutions.
CO2 is evolved from the rich amine solution via the following reactions:
IJ,3 5 + +,!"3 4 ↔ IJ,# + !"# + ,# " (3. 23)
IJ,3 5 + I# J!"# 4 ↔ !"# + IJ,# (3. 24)
(Kosseim et al., 1984) suggested an explanation for amine acid gas corrosion that provides an
alternative source for the proton required for the corrosion reaction. They suggested that due to
the high concentration of protonated amine (RNH3+) resulting from the CO2 absorption
reaction, it is likely that RNH3+ can provide protons for the corrosion reactions. Thus, they
proposed the following corrosion mechanism involving the protonated amine via reaction (3.
25); however, it has not been featured in any other work.
67 + IJ,3 5 ↔ 67 #5 + ,# + IJ,# (3. 25)
The reaction scheme implies that the corrosion rate should increase in proportion to the
concentration of protonated amine(s).
3.5. Factors Affecting CO2 Corrosion in Amine Plants
Certain trends have been identified concerning the corrosive nature of amines used in CO2
absorption. The following factors have been shown to influence amine solution corrosion (Kohl
& Nielsen, 1997):

P a g e | 80
• High operating temperature
• High rich and lean amine loading
• Type of amine
• Amine solution concentration
• Amine solution contaminants such as degradation products and heat-stable salts
• Surface films
3.5.1. Temperature
The amine solution temperature has been identified as one of the most important factors
influencing steel corrosion rates. Increasing the solution temperature enhances corrosion
reaction kinetics with rates doubling with temperature increases of 10 - 20oC (Kittel et al.,
2010). However, measuring the effect of temperature within the amine treating plant is made
complex by the temperature difference between the absorber which operates at 40oC and the
reboiler at 120oC. The difference in temperature has a considerable effect on the chemistry of
the amine, in particular the CO2 loading. Hence, a great difference in corrosion rates can be
found throughout the amine scrubbing plants. Based on plant experience, the most susceptible
areas to corrosion have been observed as the sections operated at elevated temperatures, such
as the heat exchanger and the stripper. (Veawab et al., 1999) made potentiodynamic
measurements on carbon steel in three different amines, MEA, DEA and AMP under varying
temperature conditions (25-80oC) and found that increasing the solution temperature
accelerated the corrosion rate. (Kladkaew et al., 2009) also carried out similar tests on carbon
and stainless steel samples in 5 M MEA solutions, obtaining similar results.
(de Waard and Milliams, 1975) observed that corrosion rates in CO2 containing solutions
increased with temperature up to 60oC, reaching a maximum between 60oC to 70oC and then
decreased with additional increases in temperature. (De Waard and Lotz, 1993), Nesic and
P a g e | 81
Lunde, 1993) and (Vuppu and Jepson, 1994) also observed similar effects of temperature on
corrosion rates. Figure 3-4 depicts the effect of temperature and CO2 partial pressure on the
corrosion rate of carbon steel in aqueous CO2 systems as observed by (De Waard and Lotz,
1993). It was observed that the corrosion rate at any particular CO2 partial pressure initially
increases with temperature until it reaches a maximum then starts to decrease with further
increases in temperature. Whilst the initial increase in corrosion rates was attributed to increases
in mass transfer and charge transfer rates, (De Waard and Lotz, 1993) suggested that the
observed decrease at elevated temperatures (> 70oC) was due to the formation of a protective
film of corrosion products such as siderite (FeCO3) or magnetite (Fe3O4) on the steel surface
which act as barriers to the oxidising species in the solution and restricts access to the metal
surface. At temperatures above 80oC, the iron carbonate solubility decreases, making the
formation of FeCO3 scales more likely (Ikeda et al., 1984) . (De Waard and Lotz, 1993) also
proposed a method for the prediction of carbon steel corrosion rates in aqueous CO2 solutions,
by converting the data they collected to the form of a nomograph, as shown in Figure 3-5. The
decrease in corrosion rates at elevated temperatures was also taken into account in the
nomograph by a scale factor from 0.1 to 1, which is used as a multiplier to decrease the
corrosion rate calculated from the chart in the temperature region beyond the temperature
predicted to correspond to the maximum corrosion rate.

P a g e | 82
Figure 3-4: Effects of temperature and CO2 partial pressures on carbon steel corrosion rates (de
Waard and Lotz, 1993)
Figure 3-5: Nomograph for corrosion rates of carbon steel in aqueous CO2 environments (de
Waard and Lotz, 1993).
P a g e | 83
3.5.2. CO2 Loading
CO2 loading has been identified as one of the other primary factors affecting steel corrosion
rates in amine scrubbing plants (Pearce and DuPart, 1987). Increasing the CO2 loading increases
the amine solution’s aggressiveness, so accelerating corrosion rates (Polderman et al., 1955;
Fochtman et al, 1963). This was supported by evidence that rich amine solutions are more
corrosive than lean amine solutions (Kohl & Nielsen, 1997). Thus, there have been a few
recommended guidelines for the CO2 loading that should keep the corrosion rates in amine
scrubbing plants to acceptable levels (Kohl & Nielsen, 1997). The maximum CO2 loading in
the rich amine solutions should be limited to 0.25 - 0.40 mol CO2 (mol amine)-1 for MEA, 0.33
- 1.00 mol CO2 (mol amine)-1 for DEA, 0.45 - 0.50 mol CO2 (mol amine)-1 for MDEA, 0.25 -
0.45 mol CO2 (mol amine)-1 for DGA and 0.50 – 0.85 mol CO2 (mol amine)-1 for DIPA.
The increase in corrosion rates can be attributed partly to the decrease in the amine solution pH
with increasing CO2 loading. The pH is considered to impact both the rates of electrochemical
reactions at the steel surface and the precipitation of corrosion products.
The total carbonate species in the aqueous solution can be represented by:
!K LM = !"# () + IJ,!"" 4 () + ,# !"3 ()

+ ,!"3 4 ()
+ [!"3 #4 ()
] (3. 26)
At 25oC, in an aqueous CO2 environment within a low pH range (2 < pH < 4), the corrosion of
steel is considered to be limited by the diffusion of H+ to the surface (Gray et al., 1989).
2,5 + 27 4 ↔ ,# (3. 27)
At the intermediate pH range, the direct reduction of carbonic acid is thought to occur at the
steel surface (de Waard and Milliams, 1975). Reaction (3. 28) is considered to be limited by the
hydration of carbon dioxide to form carbonic acid.

P a g e | 84
2,# !"3 (()) + 27 4 ↔ 2,!"3 4 (()) + ,# (F) (3. 28)
It has also been suggested that the direct reduction of hydrogen carbonate ions and water can
become important at a higher pH range (6 < pH < 10).
2,!"3 (()) + 27 4 ↔ !"3 #4 (()) + ,#(%) (3. 29)
2,# " + 27 4 → ,#(%) + 2",4 (()) (3. 30)
Where reactions (3. 29) and (3. 30) are thought to be charge transfer controlled. The
concentration of HCO3- increases with pH, and at pH 6 and 0.1 MPa CO2, the concentration of
HCO3- becomes more than 300 times the H2CO3 concentration. Although the concentration of
HCO3- is significantly higher ≥ 6, the low dissociation constant makes hydrogen carbonate
(pKa2 ca. 10) a poor proton donor, so it has also been deliberated that the contributions from
these reactions might be insignificant. It is also important to note that it is difficult to distinguish
experimentally between any of the mechanisms for hydrogen evolution.
The solubility of iron carbonate is also strongly dependent on the pH and CO2 partial pressure,
and decreases with pH for the pH range (4 < pH < 7). The solubility of FeCO3 decreases by a
factor of about 5 with a pH increase of 4-5, and a factor of 100 per unit pH increase at pH > 5
(Dugstad, 1992). A lower solubility results in a higher FeCO3 supersaturation on the steel
surface, so accelerating the precipitation and deposition processes (Dugstad, 1992). The
probability of FeCO3 formation on the steel surface is dramatically increased at pHs > 5, and
many reports confirm this by indicating significantly lower corrosion rates in the pH range 5.5-
6.
Based on experimental studies, (Veawab et al., 1999) found that in MEA solutions, increasing
the CO2 loading from unloaded solutions to 0.40 mol CO2 (mol amine)-1 resulted in increases
P a g e | 85
in corrosion rates from negligible values to 0.6 mm a-1. They attributed this increase in corrosion
rates with CO2 loading to the increasing concentrations of RNH3+ and HCO3- that dissociate via
reaction (3. 31) and (3. 32) to produce more hydrogen ions.
IJ,3 5 ⟷ IJ,# + ,5 (3. 31)
,!"3 4 ↔ !"3 #4 + ,5 (3. 32)
(Veawab et al., 1999) also proposed that corrosion rates increased due to a corrosion mechanism
involving the direct reaction of Fe with hydrogen carbonate ions via reaction (3. 33). Whilst
this reaction scheme is indeed plausible, the formation of passive FeCO3 due to increasing CO2
loading as they had proposed, would lead to an overall decrease not increase in corrosion rates
as they had observed.
67 + 2,!"3 4 ⟷ 67!"3 + !"3 #4 + ,# (3. 33)
(Sun et al, 2011) also observed an increase in corrosion rates with CO2 loading and suggested
that this was because hydrogen carbonate ions act as the primary oxidising agent in aqueous
amine-CO2 solutions and are reduced via reaction (3. 33).
3.5.3. Amine Type
The type of amine also has an important effect on corrosion rates in amine scrubbing plants.
Pure unloaded amine solutions are not corrosive and are even likely to offer corrosion inhibition
to the carbon steel surface (Lang and Mason, 1958; Kohl and Riesenfeld, 1985; DuPart et al.,
1993). Several studies have also shown that there is no difference in corrosion rates between
the different types of amines in the absence of CO2 (MacNab and Treseder, 1971; Dow
Chemical Company, 1962; Lang and Mason, 1958; Froning and Jones, 1958; Blanc, Grall,
Demarais, 1982). However, amine solutions containing even the smallest concentrations of
dissolved CO2 become corrosive. In loaded solutions, different amines have been found to
P a g e | 86
exhibit varying degrees of corrosiveness to the carbon steel equipment, with MEA having the
worst reputation for corrosion issues. Primary amines have been found to be the most corrosive,
followed by secondary amines, with tertiary amines being the least corrosive (Kohl and
Riesenfeld, 1985). Although DEA has been found to be less corrosive than MEA, similar results
have been obtained with its use, in terms of increasing corrosion rates with CO2 loading and
temperature. Measurements conducted with MDEA indicated that it was favourable from a
corrosion standpoint, as it has significantly lower corrosion rates than MEA and DEA. The
reason for this is not yet fully understood.
It is considered that the difference in the corrosion behaviours may be partly due to the nature
of the degradation products formed in the different amines. MEA and DEA solutions form
degradation products on absorption of CO2, while MDEA solutions do not. There are still
several questions arising from this, such as if MEA, DEA and MDEA have similar corrosion
rates in unloaded solutions and with basic degradation products in the absence of CO2, why is
MEA more corrosive than DEA and DEA more corrosive than MDEA?
(Teevens, 1990) suggested that reactivity plays a strong role and the difference in the corrosion
rates in the amines might be related to the fact that MEA is a stronger Lewis base than DEA
and MDEA is the weakest (Dupart et al., 1993). The formation of basic degradation products
can be attributed to the reaction of CO2 with primary amines and secondary amines to form
intermediate amides and subsequently their carbamates (RNHCOO-), which then undergo
dehydration to yield various degradation products (Section 3.5.7.). As MEA is a stronger Lewis
base than DEA, it is more likely to form carbamate. The ratio of carbamate to hydrogen
carbonate is 3:1 in MEA, 1:1 in DEA, while MDEA, being a tertiary amine does not form
carbamates and only forms hydrogen carbonate ions (Veldman, 2000). Thus, the relationship
between carbamate and hydrogen carbonate ion concentrations is considered to be a major
factor that determines corrosion (Veldman, 2000) and the presence of carbamates as
P a g e | 87
intermediary salts may somehow influence the corrosivity of MEA and DEA solutions, whilst
the inability of tertiary amines to form carbamates is an important clue for its less corrosive
nature. It is also possible than in tertiary amines, the only intermediary product, hydrogen
carbonate ions react with iron to form a uniform layer of FeCO3 which provides a passive barrier
against additional corrosion, even with high loadings of CO2. However, with primary and
secondary amines that selectively form carbamates over hydrogen carbonates, the layer of
FeCO3 will be less uniform.
(Veawab et al., 1999) also conducted studies to compare corrosion rates in various amines and
found that at full CO2 saturation, MEA was the most corrosive followed by AMP, then DEA
and MDEA. They suggested that the influence of amine type on corrosion behaviour is due to
differences in the amount of CO2 absorbed by the amine solution at full saturation. They
supported this hypothesis by calculating the CO2 loading in the amines and they also determined
that MEA had the highest CO2 loading followed by AMP, then DEA and MDEA. (Veawab et
al., 1999) subsequently attempted to eliminate the influence of CO2 loading by comparing the
corrosion rates in the different amines all with a CO2 loading of 0.20 mol CO2 (mol amine)-1.
They also observed that amine type had an effect on corrosion rates, even though the influence
of CO2 loading was eliminated and proposed that it might be due to the difference in HCO3- ion
concentrations in the solutions. However, this hypothesis is unlikely to be valid, as MDEA has
a higher CO2 loading capacity than primary and secondary amines. According to (Jou et al.,
1995), the absorption of CO2 in MDEA can reach 1.0 mol CO2 (mol amine)-1, whilst (Veawab
et al., 1999) calculated a CO2 loading of 0.243 mol CO2 (mol amine)-1. The hypothesis of
(Veawab et al., 1999) also fails to explain why higher rates of corrosion are not experienced in
MDEA solutions, which form only HCO3- ions, than in primary and secondary amines, which
also form carbamates as intermediary products.

P a g e | 88
3.5.4. Amine Concentration
The concentration of the amine solution has also been found to influence steel corrosion rates.
Generally, on amine scrubbing plants, it would be desirable to employ high amine
concentrations for energy saving purposes. However, from an industrial standpoint, high amine
concentrations have been found to increase the corrosion rates. In order to control and limit the
corrosion rates to acceptable and manageable levels, most amine scrubbing plants are operated
within the recommended guidelines for maximum amine concentrations. Typical amine
concentrations are 10 - 20 wt% for MEA, 20 – 40 wt% for DEA, 50-55 wt% for MDEA, 50-65
wt% for DGA and 20 - 40 wt% for DIPA (Nielsen et al., 1995). However, the experimental
results on the effect of amine concentration have ranged from significant (DuPart et al., 1993;
Veawab et al., 1999), to moderate, to no effect at all (Guo and Tomoe, 1999; Vazquez,et al.,
2000).
3.5.5. Effect of Oxygen
The effects of oxygen has been the source of debate by researchers. In experimental studies
conducted by (Veawab et al., 1999), corrosion rates were found to increase with the presence
of O2, as expected for such an oxidising agent in amine solutions, oxidizing steel surfaces to
form oxide films as in the reactions:
267 + 2,# " + "# → 267 ", # (3. 34)
Q
267(",)# + ,# " + "# → 267(",)3 (3. 35)
#
(Kladkaew et al, 2009) also observed increases in corrosion rates and proposed an alternative
reaction scheme (3. 36), in which higher corrosion rates resulted from higher reduction rates
of dissolved oxygen by:

P a g e | 89
"# + 2,# " + 47 4 ↔ 4",4 (3. 36)
Then, iron oxidation products, Fe(OH)2 and Fe(OH)3 are formed via reaction (3. 34) and (3.
35) respectively.
On the contrary, (Kittel et al., 2010) observed no significant impact of oxygen on the corrosion
rates of carbon steel in CO2-loaded MEA solutions.
3.5.6. Effect of Surface Films
Another important parameter that determines the corrosion rate is the ability of the carbon steel
to form iron carbonate by reaction with any of the forms of dissolved CO2. Under the right
operating conditions of temperature and pH, FeCO3, an insoluble scale can precipitate on the
steel and act as a protective barrier. As the corrosion reaction advances, once the Fe2+ and
CO32- ions exceed the solubility limit, they react to form solid FeCO3 films according to
reaction (3. 37). Based on experimental evidence, it seems clear that the rates of corrosion can
be decreased significantly by precipitation of FeCO3 that is adherent to steel surfaces (Dugstad,
1998; Nyborg, 1998; Nyborg, and Dugstad, 1998).
67 #5 + !"3 #4 ⟷ 67!"3(B) (3. 37)
The iron carbonate decreases corrosion rates by presenting a diffusion barrier for the species
involved in the corrosion process, by blocking a portion of the steel and preventing the
underlying steel from undergoing any further dissolution. (Kohl & Nielsen, 1997;Kladkaew et
al., 2009a) suggested an alternative scheme for ferrous carbonate formation in amine scrubbing
plants which occurs from the reaction of the ferrous ions produced from dissolution of iron
and the hydrogen carbonate ions in solution.

P a g e | 90
67 + 2,!"3 4 → 67!"3 + !"3 #4 + ,# (3. 38)
Whilst (DuPart et al., 1993) and (Rooney and Bacon, 2000) were able to identify the presence
of some insoluble corrosion products such as iron carbonate and iron hydroxide in several parts
of the plant, they were also able to identify soluble corrosion products, such as FeII. However,
there are no studies in the open literature which, quantify these corrosion products. (Burnstein
and Davies, 1980) proposed an alternate reaction scheme which, could account for the
formation of such products as outlined in reactions (3. 39) to (3. 42):
67 + 2,# " → 67(",)# + 2,5 + 27 4 (3. 39)
67 + ,!"3 4 → 67!"3 + ,5 + 27 4 (3. 40)
67(",)# + ,!"3 4 → 67!"3 + ,# " + ",4 (3. 41)
67!"3 + ,!"3 4 → 67(!"3 )# #4 + ,5 (3. 42)
The reduction of undissociated carbonic acid has also been proposed via the following schemes
(Kladkaew et al., 2009):
67 + ,# !"3 → 67!"3 + ,# (3. 43)
67 + 2,# !"3 → 67(,!"3 )# + ,# (3. 44)
(Riesenfeld and Blohm, 1951) also suggested that the dissolved CO2 in the amine solutions
could react directly with carbon steel to form iron carbonate (FeCO3) via the following reaction
(Kladkaew et al., 2009):
67 + !"#(()) + 2,# " ↔ 67!"3(B) + ,# (3. 45)
Another common surface product found in CO2 corrosion of mild and low alloy steels is iron
carbide (Fe3C), which is part of the steel microstructure, so its quantity increases with the
P a g e | 91
carbon content of the steel (Avner, 1997). Fe3C is a metallic conductor with a low hydrogen
overpotential and a much lower overpotential than iron for other cathodic reactions, so it is
more difficult to dissolve and is often left on the steel surface as a consequence of the corrosion
reaction. Its presence can enhance the corrosion rate by up to a reported factor of 3-10 (Crolet
et al., 1996) by establishing a galvanic effect and acting as a cathodic site for the hydrogen
evolution reaction (Crolet et al., 1996; Videm et al., 1996). However, Fe3C does not always
have a detrimental and corrosive effect. Under the right conditions such as high pH,
temperature, it has been reported by (Dugstad, 1998) and (Palacios and Shadley, 1991) that
Fe3C may enhance FeCO3 precipitation by preventing the diffusion of ferrous ions from the
surface. If the iron carbide layer blends into the iron carbonate film in a well distributed
manner, it can also act as an anchor to the protective FeCO3 layer. Both the FeCO3 and Fe3C
films will be discussed in more detail in section 3.6.
3.5.7. Amine Degradation Products
Whilst the main reaction between amines and CO2 is reversible, irreversible side reactions
could also occur, resulting in degradation products from which the amines cannot be easily
recovered. There are several mechanisms through which the amines can degrade and the
degradation products depend on a number of factors such as the degradation mechanism and
the type of amine used for the CO2 capture process.
• Oxidative degradation
• Thermal degradation
• Heat stable salts
• Degradation caused by carbonyl sulfide (COS) and carbon disulfide (CS2)
• Degradation by carbon monoxide (CO)

P a g e | 92
• Sulfur and polysulfide degradation
As the solvents must be replaced after degradation due to the decrease in CO2 absorption
capacity, there are concerns about amine loss, the adverse environmental effects of disposal
and the inevitable operating costs of dealing with the degradation. Whilst these are issues that
can be dealt with relatively easily, the corrosion of the carbon steel equipment is another major
concern closely related to amine degradation that is of a more complex nature. According to
(Gregory and Scharmann, 1937) degradation products enhance the corrosion issue in MEA
plants. On the other hand, the link between corrosion and degradation for other amines remains
inconclusive. (Blanc et al, 1982) attempted to study degradation-induced corrosion in DEA
solutions and concluded that DEA degradation products had little to no effect on carbon steel
corrosion, whilst others (Moore, 1960 and Chakma and Meisen, 1986) argued otherwise and
established that there was a correlation between the two operational issues.
Although the presence of degradation products and contaminants, particularly heat stable salts,
has been reported to increase corrosion rates (DuPart et al., 1993), there is still very limited
relevant information in the open literature. Heat stable salts are the products of alkanolamines
and acids stronger than CO2 and H2S, such as carboxylic acids. There are various sources of
these acids such as makeup water and feed gas streams. They can also be generated within the
plant by chemical reactions with contaminants such as O2, CO, HCN and SO2.
In general, the acids react with the amine by the proton transfer reaction (Veldman, 2000):
IJ,# + ,S ↔ IJ,3 5 S4 (3. 46)
where HA is the acid form of the heat stable salt anion and A- is the heat stable salt anion.
They are called heat stable salts (HSS) since they cannot be regenerated and separated from the
amine through typical regeneration conditions; instead they remain and accumulate in the
P a g e | 93
absorbent. The accumulation of these contaminants within the system has been found not only
to decrease the CO2 absorption capacity of the amines but also to increase corrosion rates
significantly (Tanthapanichakoon et al., 2006). HSSs typically induce corrosion in the system
by decreasing the pH, increasing the conductivity of the amine solution and may also act as
chelating agents, dissolving the protective film on the metal surface (Rooney et al., 1996).
(McCullough and Nielsen, 1996) also suggested the possibility that with some of the weaker
HSS such as formic acid, as the operating temperatures increases during the amine regeneration,
the acids also become less bound to the amine and are temporarily un-neutralized to release the
free acid that could then react directly with the steel. It has been found that areas involved with
high temperature processes such as the reboiler, reboiler piping, lower trays of the regenerator
and the hot side of the heat exchanger, are the most affected by this type of corrosion, especially
pitting.
Whilst it is has been generally accepted that the presence of heat stable salts and amine
degradation products exacerbate corrosion rates, there is no consensus on the corrosion
mechanism. (Tanthapanichakoon and Veawab, 2006) investigated the effect of fourteen heat
stable amine salt compounds on the corrosion of both carbon steel and stainless steel, using
MEA and CO2. The corrosion rates increased in the presence of oxalate, formate, malonate,
glycolate, succinate, and acetate, sulfate and hydrogen chloride; oxalate was the most corrosive
of these heat stable salts followed by formate and malonate. These results were in agreement
with those of (Nainar and Veawab, 2009). The other HSSs investigated such as sulfite, sodium
chloride, sodium thiosulfate, ammonium thiosulfate, sodium thiocyanate, and ammonium
thiocyanate, either decreased corrosion rates or acted as inhibitors.

P a g e | 94
3.6. Review of Corrosion Product Films
Based on experience in the oil and gas industry, uniform protective scales at the bottom of
deep gas wells have been found to result in less severe corrosion rates (Burke and Hausler,
1985). Chemical analysis of these protective scales revealed that they consisted predominantly
of iron carbonate (Burke and Hausler, 1985). Experimental studies showed consistently that
even a thin adherent layer of FeCO3 deposited on the surface of corroding iron or steel can
result in significant decreases in corrosion rates (Dunlop et al., 1984; Ikeda et al., 1984; Videm
and Dugstad, 1989).
The nature of this protective scale and the stability of the film determines the rate of corrosion,
as well as the type of corrosion i.e. general corrosion or mesa attack. The structure, porosity
and morphology, as well as its adherence to the underlying steel surface are much more
important than the thickness in diminishing corrosion rates. The composition of the corrosion
layer can also have a substantial effect on its ability to protect the steel surface (Crolet, 1994).
In some cases, corrosion product layers can be extremely protective, whilst in others not very
(Crolet et al., 1995) and in some cases have been found to actually enhance corrosion rates
(Crolet et al., 1994).
Corrosion product films in CO2-water systems can be divided into four main categories, each
of which will be discussed below:
• Iron carbonate
• Iron carbide
• Iron carbonate and Iron carbide
• Iron carbonate and iron oxide

P a g e | 95
3.6.1. Iron Carbide Films
When steel corrodes in an aqueous CO2 system and ferrous ions are produced via the anodic
dissolution of iron, the FeII ions dissolve leaving uncorroded iron carbide (Fe3C) accumulating
on the surface. Fe3C is a metallic conductor with a low hydrogen overpotential, which forms
an electrically conductive, porous, sponge-like layer on which the cathodic reaction can occur.
It is characterised by a non adherent, black surface film with a thickness typically in the order
of 100 µm (Crolet et al., 1994). The flow rate of the aqueous solution affects the amount of
Fe3C formed; at high solution flow rates (turbulent flow conditions, Reynolds number ≥
4000), the surface film will principally consist of iron carbide and possibly some other
unidentified constituents of alloying constituents. Decreasing flow rates results in an increase
in the amount of Fe3C accumulating on the steel surface ,as well as an enhancement in the rate
of formation of FeCO3 in the product layer. According to (Dugstad and Videm, 1990), it is
possible that the diffusion of FeII ions away from the steel surface is slowed down by the Fe3C
film, allowing the solubility of FeCO3 to be exceeded locally.
The presence of Fe3C can increase the corrosion rate by a factor of 3-10 through galvanic
effects (Crolet et al., 1996). As Fe3C has a significantly lower overpotential for the cathodic
reaction than iron, it can establish a galvanic effect with the uncorroded steel surface thereby
enhancing the corrosion of the metal by accelerating the cathodic reaction. At low FeII ion
concentration of << 1 ppm, pure Fe3C films will be formed, causing an increase in the corrosion
potential and the corrosion rate. As the cathodic reactions occur at the Fe3C film and the anodic
reaction is the dissolution of iron as ferrous ions, there is a distance between the site of
occurrence of the two reactions. The cathodic reaction typically consists of the reduction of
protons resulting in the region around the Fe3C film becoming less acidic or more alkaline. On
the other hand, the region around the anodic site (the steel surface) becomes more acidic
P a g e | 96
(Crolet et al., 1994; Crolet et al., 1995; Dugstad and Videm, 1990). It has been suggested that
this is due to a number of hydrolysis reactions such as equation (3. 47).
67 #5 (()) + ,# " → 67(",)5 (()) + ,5 (()) (3. 47)
The iron carbide layer will have the effect of slowing down the diffusion of the ferrous ions to
the bulk solution, causing their accumulation in the porous film structure. This in turn leads to
supersaturation of the FeII locally and on exceeding the solubility, precipitation of FeCO3.
However, it is possible that the steel surface may be too acidic for FeCO3 precipitation, in spite
of high local FeII ion concentrations (Crolet et al., 1994; Crolet et al., 1995; Dugstad and
Videm, 1990). The FeCO3 will precipitate in the more alkaline cathodic regions. Thus, the iron
carbonate layer will precipitate in or on top of the iron carbide layer, forming a corrosion layer
with poor adherence to the metal surface. Such surface films result in inadequate corrosion
protection resulting in high corrosion rates. The inception of high corrosion rates leads to an
increase in the difference in local pH between the cathodic and anodic sites that in turn
aggravates the formation of the non-protective film.
Under certain conditions, it is possible for iron carbide and iron carbonate to mix and form an
adherent surface film. Fe3C acts as the framework for the film and anchors the FeCO3 layer,
providing a greater resistant to the shear forces caused by turbulent flow conditions. The
formation of such films can significantly decrease the effect of localised forms of corrosion,
such as erosion corrosion.
In summary, it is apparent that the accumulation of Fe3C on the steel surface can have either a
favourable or detrimental effect on corrosion of steel in CO2-H2O systems, depending on the
structure of the film and the conditions under which it is formed. If the iron carbide structure
slows down the diffusion of the FeII ions from the steel surface, it promotes FeCO3
precipitation. If it blends into the film, it can enhance the strength of the film formed.
P a g e | 97
3.7. Iron carbonate Film
FeCO3 also known as siderite or chalybite was discovered by (Baylis, 1926) to play an
significant role in the corrosion process. The precipitated FeCO3 film forms a layer that is
impermeable and retards the corrosion process by leading to diffusion control. In CO2
containing aqueous systems, the corrosion rate is highly dependent on the precipitation rate of
FeCO3 and the morphology of the film, both of which are determined by the FeCO3
supersaturation and the temperature of the system. The morphology and composition of the
film govern the nature of the corrosion process from the worst case scenario to low corrosion
with protective films or mesa type corrosion.
The supersaturation of FeCO3 is the key driving force for precipitation. The precipitation rate
is usually slow; a high level of FeCO3 supersaturation is required to obtain sufficient deposits
on the steel surface. The likelihood of FeCO3 depositing on the surface depends on the kinetics
of precipitation and the way in which it is anchored to the underlying steel surface.
In principle, film precipitation occurs in two steps: nucleation and particle growth so the
morphology of the film is most likely subject to the relative kinetics of these two processes
(Dugtsad et al., 1993; Davis, 1993). It is assumed that the rates of these processes are dependent
on the relative supersaturation (RS) where S is the supersaturation, Q is the concentration of the
solute at any instant, Qeq is the equilibrium solubility and Ksp, the solubility product for FeCO3.
(_4_à )
I7UGVW=7 XYZ7[\GVY[GVW]^ = IX = (3. 48)
_à
The nucleation rate is considered to increase exponentially with supersaturation i.e.:
e
(_4_à )
IGV7 ]b ^YcU7GVW]^ = dQ (3. 49)
_à
Whereas the rate of particle growth depends approximately linearly on the supersaturation:
P a g e | 98
(_4_à )
IGV7 ]b ZG[VWcU7 F[]fVℎ = d# (3. 50)
_à
The value of K2 is typically significantly larger than K1, so growth is predominant at low
relative supersaturation, while at higher supersaturation, the exponential dependence of
nucleation rate causes this process to occur to the near exclusion of particle growth. During the
nucleation process, a colloidal dispersion forms close to the steel surface or inside the film in
which settling is prevented as a result of Brownian motion and electrical repulsion between the
colloid particles. In order to form a stable film, coagulation must occur. Coagulation of these
colloidal dispersions can be accelerated by an increase in temperature, stirring and increasing
ionic strength and decreasing the surface charge. A high supersaturation is required for the
precipitation of FeCO3 (Dugstad, 1992), so it is important to control the degree of
supersaturation on the steel surface to achieve the desired film precipitation rate and properties.
Figure 3-6: Calculated growth rate of iron carbonate as a function of temperature and
supersaturation (Jonson and Tomson, 1991).
P a g e | 99
Based on their studies on the precipitation of FeCO3, (Jonson and Tomson, 1991) calculated
FeCO3 growth rates as shown in Figure 3-6; they then proposed equation (3. 51) to describe
the kinetics of FeCO3 precipitation in pure CO2 corrosion of steel:
i
.I = dh kBl XX m.n − L #
(3. 51)
j
In this expression, the precipitation rate (PR) in kmol m-3 s-1 is a function of the iron carbonate
supersaturation SS, the solubility product Ksp, the temperature (via the rate constant kr which
obeys Arrhenius law) and the surface area to volume ratio (A/V). The supersaturation ratio is
given by equation (3. 52), in which [Fe2+] and [CO32-] are the time-dependent species
concentrations:
pq 1r [/0s 1r ]
XX = (3. 52)
tuv
From Figure 3-6, it is clear that the FeCO3 growth rate is very low at low temperatures, but
significantly increases with increasing temperature. (van Hunnik et al., 1996) extended the
work of (Jonson and Tomson, 1991) on the precipitation kinetics of iron carbonate and found
that equation (3. 51), which describes the effect of temperature on the precipitation kinetics of
FeCO3, did not provide a good fit over a wide range of supersaturation levels. This was due to
(Jonson and Tomson, 1991) having fitted the equation only to experimental data at low levels
of supersaturation i.e. SS < 2; (van Hunnik et al., 1996) considered that the precipitation rate
based equation (3. 51) would be overestimated, especially at larger values of SS.
The supersaturation varies from typically above one for high temperature systems to values as
high as 1000 at low temperatures. Hence, the experimental work conducted by (van Hunnik et
al., 1996) investigated the effect of supersaturation over a wide range, resulting in a nominally
more accurate expression referred to as the ‘KSLA’ equation:

P a g e | 100
i
.I = dh kBl XX − L 1 − XX 4Q (3. 53)
j
In deriving this expression, (van Hunnik et al., 1996) assumed that the precipitation rate for
FeCO3 was controlled by the kinetics of the particle growth, as opposed to nucleation. In
support of this assumption, they were able to model their experimental results sufficiently using
equations describing precipitation growth only, regardless of the steel microstructure involved.
The rate constants kr for all the experiments were obtained from the measured precipitation
rates via the precipitation equations and appeared to obey Arrhenius’s law for temperature
dependence.
4{
dh = Sexp( ) (3. 54)
|K
The values for A and B measured by (Jonson and Tomson, 1991; van Hunnik et al., 1996) are
listed in Table 3-4.
Table 3-4: growth kinetics of FeCO3 , 2σ = 95% confidence level
Johnson and Tomson equation KSLA equation

Average 2σ Average 2σ
A 56.3 4 52.4 3
B 127.3 12 119.8 9
The equation proposed by (van Hunnik et al., 1996) has been shown to give a better overall fit
for the precipitation rate, regardless of the level of supersaturation, whilst the (Jonson and
Tomson, 1991) equation gives a more precise prediction at lower levels of supersaturation.
(van Hunnik et al., 1996) observed that the scale formed was protective only if the precipitation
rate was of the order of the corrosion rate. If the precipitation rate was much less than the rate
of iron dissolution, then the steel substrate would be removed at a higher rate than a dense
FeCO3 film can form.

P a g e | 101
The increase in precipitation rate with temperature has been observed to affect FeCO3 film
properties. At temperatures higher than 60oC, when the precipitation rate is high whilst the
supersaturation is low result in dense, protective, crystalline films. At lower temperatures (<
40oC), the precipitation rate is much lower and so the relative supersaturation can become very
high when dissolved FeII and CO32- accumulate; hence, corrosion films are formed with low
crystallinities. If the crystal growth rate is less than the rate of dissolution, the difference
between the two reactions will result in an iron carbonate film that is porous and non-adherent.
Thus, films formed at temperatures less than 40oC are generally porous, not very adherent, non-
crystalline in nature and considerably less protective than those formed at higher temperatures.
For example, at 40oC, a supersaturation ratio of 50 would be required to obtain a precipitation
rate which matches a corrosion rate of 3 mm a-1 (van Hunnik et al., 1996). It is likely that such
a high supersaturation occurs close to the steel surface when a corrosion film layer restricts
diffusion rates of reactants and products. The maximum supersaturation will then be a balance
between the rate of dissolution, nucleation, precipitation and transport (Dugstad, 1998).
This general precipitation mechanism is supported by many of the observations found in
literature. Between 80oC-100oC, the precipitation is fast and the supersaturation low, hence
dense protective crystalline carbonate layers have been observed, as shown in Figure 3-7.
P a g e | 102
Iron Carbonate
Figure 3-7: FeCO3 embedded in a Fe3C Film; 80oC, 10-100 supersaturation ratio (Dugstad,
1998).
At temperatures between 20oC and 60oC, a Fe3C film is formed that possibly may be filled with
FeCO3, depending on the supersaturation. If the supersaturation ratio is between 10 and 100,
then the Fe3C film will be partially filled with FeCO3 as shown in Figure 3-7.
Iron Carbonate
Iron Carbide
Steel
Figure 3-8: FeCO3 embedded in a Fe3C film; 60oC, 10-100 supersaturation ratio (Dugstad,
1998).
P a g e | 103
The film formed will have a porous structure and the FeCO3 is unlikely to form on the actual
surface of the steel. The FeCO3 will probably precipitate within the outer carbide layer,
gradually precipitating inwards towards the steel surface as time progresses. At supersaturation
ratios between 1-3, the iron carbonate is unlikely to precipitate within the porous iron carbide
structure, leaving an unfilled iron carbide film as shown in Figure 3-9.
Iron Carbide
Steel
Figure 3-9: FeCO3 embedded in a Fe3C film; 60oC, 1-3 supersaturation ratio (Dugstad, 1998).
Iron carbonate precipitation has not been observed within the iron carbide structure at
temperatures slightly below 40oC, even when the solution is highly supersaturated with FeCO3.
However, increasing the temperature significantly increases the growth rate, so much more
FeCO3 has been observed in corrosion films formed at 40oC (Dugstad et al., 1993). It is also
important to note that at higher temperatures (ca. 120oC), the corrosion films formed are
frequently less protective than those formed between 80oC and 100oC.
At temperatures > 150oC, the iron dissolution is prevented by the formation of a compact and
adhesive film layer, which consists primarily of FeCO3. According to (Ikeda et al., 1984), the
steel corrosion is controlled by a high precipitation rate and slow dissolution of the protective
FeCO3 film. They conducted X-ray diffraction analysis of the steel surface and detected only
P a g e | 104
FeCO3 and not Fe3O4, even though the latter phase would have been predicted as the more
thermodynamically stable phase at high pHs (≥ 9).
(van Hunnik et al., 1996) also presented a concept of scaling tendency, which is defined as the
ratio between the precipitation rate and the corrosion rate. They suggested that when the scaling
tendency reached a critical value called ‘the critical scaling tendency’, an iron carbonate film
will start to grow and start to decrease the corrosion rate. For all the experiments conducted by
(van Hunnik et al., 1996), the critical scaling tendency was a value higher than 0.4. They also
proposed that for a scaling tendency of > 0.5, then the protective FeCO3 film would be self-
repairing. The concept of the scaling tendency is useful, but it is subject to several drawbacks,
such as the lack of consideration of the film adherence or the steel microstructure, as it assumes
a smooth surface.
The temperature also has been found to influence the adherence of the corrosion film.
According to (Videm and Dugstad, 1989), the FeCO3 film formed at 40oC can be wiped off
easily with a cloth, whilst at 60oC the films are more adherent and at 90oC the films were so
tough that they could not be scraped off even with a plastic knife
Table 3-5: Standard State Formation Properties of FeCO3 at 298.15 K and 0.1 MPa.
ΔfH0/ kJ mol-1 ΔfG0/ kJ S0/ J mol-1 Cp0/ J mol-1 Reference

mol-1 K-1 K-1
92.9 83.35 (Anderson, 1934)

-747.60 -673.75 92.9 82.09 (Kelly and Anderson, 1935)
-744.8 -673.88 92.9 (Latimer, 1952)
-748.9 (Wilcox and Bromley, 1963)
96.1 (Kostryukov and Kalinkina,
1964)
105.0 (Robie, 1965)
-743.965 -673.749 105.0 (Robie and Wladbaum,
1968)
P a g e | 105
-679.44 (Langmuir, 1969)

-740.57 -666.72 92.9 82.13 (Wagman et al., 1969)
-671.53 (Singer and Stumm, 1970)
-775.71 -697.05 74.1 (French, 1971)
-740.57 -768.26 92.9 82.08 (Barin et al., 1973)
-749.656 -679.440 105.0 82.09 (Helgeson et al., 1978)
-736.985 -666.698 105.0 (Robie et al., 1978)
-666.02 (Reiterer et al., 1980)
-740.57 -666.67 92.9 82.13 (NIST, 1982)
-753.22 -680.03 95.47 82.44 (Robie et al., 1984)
-749 (Stubina and Toguri, 1989)
-761.18 -688.04 95.50 82.27 (Holland, 1990)
-750.6 (Chai and Navrotsky, 1994)
-755.9 -682.8 95.47 82.27 (Robie and Hemingway,
-752.0 -678.9 95.47 1995)
(Preis and Gamsjager, 2002)
The solution pH also affects the corrosion rate strongly by affecting the solubility of FeCO3,
which decreases at high solution pHs. Therefore, increasing the pH has the effect of increasing
both the supersaturation and precipitation rate. FeCO3 formation is unlikely to occur at a pH
value of less than between 5-6, although the exact value will depend on the other conditions
such as temperature, dissolved FeII activity etc.
Several researchers have published and reviewed standard FeCO3 thermodynamic properties
such as entropy, enthalpy of formation, solubility constant and Gibbs energy of formation, for
corrosion modelling and the computation of potential-pH diagrams (Ghosh et al., 1996; Dresel
et al., 1989; Hem, 1967). These thermodynamic properties also have been used in kinetic
models to predict the formation of FeCO3. (Fosobol et al, 2010) presented a comprehensive
review of the thermodynamic properties of FeCO3 by comparing and discussing the most
reliable values, which are summarised in Table 3-5.

P a g e | 106
3.7.1. Mixed Iron Carbonate and Iron Carbide Film
The most common film found on steel surfaces in CO2 containing environments is a structure
consisting of both iron carbonate and iron carbide, as the corrosion of the iron component
leaves a Fe3C layer on the steel surface. If FeCO3 precipitation occurs, it mixes within the Fe3C
layer to form a mixed film. The structure and degree of protectiveness of the mixed corrosion
film is dependent on the location of the iron carbonate precipitation within the film. If the
FeCO3 precipitation occurs inside the carbide layer, then a protective and adherent film will
form that can often endure the stress of high solution flow.
The structure and nature of the mixed film will also depend on the structure of the porous Fe3C
layer, which is a function of the steel microstructure. Pre-treatments such as heat treatment can
influence the formation and structure of the iron carbide layer (Palacios and Shadley, 1991).
Experimental studies conducted by (Dugstad et al., 1993) on the effect of heat treatments on
the corrosion of carbon steel showed that both corrosion rate and the likelihood of protective
film formation decreased with increasing tempering temperature, so supporting the hypothesis
that the Fe3C structure has a significant effect on protective film formation.
The carbon content of the steel also plays a significant role in the formation and protectiveness
of the mixed films. Mesa attacks are more prevalent in carbon steels with a low carbon content
(< 0.05 %) than in carbon steels with a carbon content of 0.15-0.20 %. Steels with a higher
carbon content have a carbide structure that offers a framework for the precipitation and
accumulation of a dense and adherent mixed (iron carbonate and iron carbide) film, offering
protection to the underlying steel. Conversely, carbon steels with a low carbon content have a
much more porous and open carbide structure, which does not provide an adequate framework
for protective FeCO3 precipitation. The shape and distribution of the carbide phase has an
impact on the adherence of corrosion films, with spherical Fe3C particles resulting in little
P a g e | 107
adherence to the carbon steel surface (Dugstad, 1991). The structure of the mixed film has a
substantial effect on the corrosion behaviour of the steel; the effects of different types of
corrosion product orientation have been summarised by (Crolet et al., 1996) as shown
schematically in Figure 3-10.
Fe3C+FeCO3
Fe3C
Fe3C
Steel
Steel
Non- Protective
Fe3C
Fe3C + FeCO3 Fe3C + FeCO3
Steel Steel
Protective
Figure 3-10: The composition of protective and non-protective iron carbide and iron carbonate
films (Crolet et al., 1996).
3.7.2. Iron Carbonate and Iron Oxide Film
The potential-pH diagram for Fe-CO2-H2O systems is shown in Figure 3-11. Based on
thermodynamic predictions (Figure 3-11) and data from the literature (Ikeda et al., 1984), there
is a comparatively wide region in which oxides and hydroxides such as magnetite (Fe3O4) are
also expected to dominate. However, there is no direct evidence of the presence of magnetite
in corrosion films produced in CO2 containing environments at 100-150oC. The corrosion films
P a g e | 108
formed at 120oC are reportedly less protective than films formed within the temperature region
of 80-120oC. Based on studies conducted by (Dugstad et al., 1993; Ikeda et al., 1984) at 120oC,
despite expectations of Fe3O4 within the corrosion film formed at this high temperature, it was
not observed. At 150oC and pH 4.8, the corrosion product formed was found to consist of only
iron carbonate, whilst at 250oC magnetite was detected as the predominant product (70% Fe3O4
and 30% FeCO3) (Shanon, 1978), so providing evidence for magnetite formation only at
sufficiently high temperatures. However, it is also important to note that the electrode potential
in which these films were formed was not reported.
2.2
H3FeO4+ H FeO
2.0
1.8
2 4
HFeO4-
1.6 FeO42-
1.4
1.2
1.0 Fe3+ Fe3+
0.8
0.6
O2
0.4 Fe2O3 (c) H2O
0.2
Fe2+ Fe(OH)3 (ppt)
Fe(OH)4-
0.0 H+
FeCO3 Fe(OH)4-
-0.2 H2
-0.4 FeCO3 Fe(OH)2
-0.6 Fe3O4
-0.8 Fe(OH)2 Fe(OH)42-
-1.0 Fe
-1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Figure 3-11: Potential-pH diagram for Fe-CO2-H2O System, showing the lines for dissolved
iron activity of 10-4 and (0.1 MPa CO2, 298 K) (Kelsall 2013).
P a g e | 109
In summary, the most important component in corrosion films on steel is FeCO3. A dense and
adherent film of FeCO3 can provide protection to the underlying steel, by decreasing corrosion
rates to < 0.1 mm a-1. However, a mixed FeCO3 and Fe3C film offers the most protection against
localised forms of corrosion, such as erosion or mesa attack. Based on reports from the open
literature, the ideal operating conditions for the formation of a protective film are pH > 5.5, and
an elevated temperature 60 oC < T < 150 oC and a supersaturation (SS ≥ 1) of FeII. The different
types of corrosion films formed in aqueous CO2 environments over a temperature range of
25oC- 150oC are summarised in Table 3-6.
Table 3-6: Types of corrosion films formed from 30-150oC in steel-CO2-H2O systems
Type Category Manifestation Temperature Effect on corrosion

Iron • Porous, (T < 150oC) Increase corrosion rate
Carbide/ sponge like
Cementite appearance
(Fe3C) • Black
Ferrous Porous • Grey black or (60oC < T < 150oC) Not Protective
Carbonate/ golden- brown
Siderite
(FeCO3)
Dense Protective but non-adherent
films can lead to localised
corrosion at high solution flow
Mixed film Fe3C on (60oC < T < 150oC) Not Protective

(Iron top, FeCO3
Carbide + at the
Ferrous bottom
Carbonate)
Mixture Protective
throughout
the film
FeCO3 on Protective and provides

top and resistance to localised
Fe3C at the corrosion as a result of
bottom additional anchoring by Fe3C
film framework
P a g e | 110
3.8. Corrosion Inhibitors for Amine Absorption Plants
Inhibitors are frequently employed to decrease corrosion rates in closed systems such as amine
gas treating plants. It has been found that a corrosion inhibitor is unnecessary in systems
operating with relatively low CO2 loading, low amine degradation or heat stable salt formation,
low amine concentrations and amine solution velocities of ca. 1 m s-1. However, in systems
which do not meet these low corrosion rate requirements, it is necessary to inhibit corrosion.
The corrosion inhibitors function via different methods:
i) Adsorbed as a thin film on the metal surface;
ii) Promoting the formation of a thick layer of corrosion products;
iii) Altering the conditions of the corrosive environment by removing or reacting with
corrosion-promoting species.
There are different ways in which corrosion inhibitors can be classified: by their chemical
nature and termed as organic or inorganic, or they could be divided into oxidising passivators
and adsorption inhibitors, based on their inhibiting function.
There has been extensive research into developing corrosion inhibitors and a few have been
commercialised and patented for amine scrubbing plants. Organic inhibitors include cyclic
amines and salicylic acid, while inorganic inhibitors are vanadium(V), copper(II), antimony(V),
cobalt(II) and sulfur-containing compounds. Inorganic inhibitors are generally preferred to
organic corrosion inhibitors due to the superior inhibiting properties.
The most extensively and successfully used inorganic corrosion inhibitors contain
vanadium(V), in particular sodium metavanadate (NaVO3). These class of corrosion inhibitors
work by raising the steel’s potential to the passive region, so that surfaces are covered by passive
films, thereby decreasing corrosion rates. Such passive films that adhere to steel surfaces,
P a g e | 111
forming protective barriers against corrosion, have been postulated to be Fe2O3. The barrier is
often only a few molecular layers thick and has been found to be effective in decreasing
corrosion rates in general corrosion, but inadequate in preventing localised forms of corrosion.
Film forming corrosion inhibitors have had varying reports about their success and commercial
viability. While the use of inorganic inhibitors have been reported to be effective in the vapour
phase regions of the stripper, they have not worked as well in other areas, especially in the
piping and reboiler, in which corrosion occurs from heat stable salts or CO2. In recent years,
environmental concerns have been expressed about the use of heavy metals, because of their
toxicity.
(Nainar and Veawab, 2009) investigated the effectiveness of sodium metavanadate (NaVO3)
and copper carbonate (CuCO3) as corrosion inhibitors; the latter being a potential substitute for
the more toxic NaVO3. Although CuCO3 was able to constrain corrosion rates of carbon steel
below the acceptable level of 0.254 mm y-1, NaVO3 proved to be a superior inhibitor. Both
inhibitors were more effective in the presence of O2 than in its absence, because O2 is required
to maintain active CuII or VV in the inhibited solutions, to prevent the formation of other
oxidation states of vanadate and copper. Increasing the concentration of NaVO3 also increases
the inhibition, more protective films are produced, so decreasing corrosion rates. The presence
of heat stable salts decreased the inhibitive capabilities of both NaVO3 and CuCO3.
Oxygen scavengers or reducing agents are also used to reduce dissolved oxygen, thereby
inhibiting corrosion. They are particularly effective if they are able to enhance the formation
and stability of the passive film on the metal surface. In the case of iron, magnetite (Fe2O3-FeO/
Fe3O4) and hematite (Fe2O3) are the most stable films, believed to consists of a layer of Fe2O3-
FeO in the inner layer and Fe2O3 in the outer layer. It has been proposed that these passive
layers are formed through the following reaction schemes (Veldman, 2000):
P a g e | 112
67 + ,# " → 67", + ,5 + 7 4 (3. 55)
(FeOH)(EB → (FeO)(EB + H 5 + e4 (3. 56)
(FeO)(EB + ,# " → (FeOO)(EB + 2H 5 + 2e4 (3. 57)
The adsorbed species then combine to form the more passive oxides:
(FeO)(ÅBChÅqE + (FeOO)(ÅBChÅqE ↔ 67# "3 (3. 58)
2 (FeO)(ÅBChÅqE + (FeOO)(ÅBChÅqE ↔ 67# "3 − 67" (3. 59)
Thus, the removal of oxygen at the metal surface would shift the overall equilibrium to favour
the formation of the passive oxides. The selection of an ideal oxygen scavenger must be based
on its ability to remain reactive at all amine unit operating temperatures and work well in both
liquid and vapour phases, while not being unfavourable to the operation. In applications where
passive oxide formation does not occur easily, scavengers complicate the process further and
accelerate corrosion. While oxygen scavengers have been found to be effective, they are not
very useful for preventing pitting or stress cracking and corrosion due to heat stable salts.
(Veawab et al., 1997) investigated the corrosion rates of carbon steel in AMP solutions for
comparison with MEA. While they found that the AMP systems were less corrosive to carbon
steel than MEA, they concluded that the use of inhibitors was still necessary. For this purpose,
they employed sodium metavanadate (NaVO3) and an oxygen scavenger, sodium sulfite
(Na2SO3). The corrosion products formed were observed to adhere to the metal surface and act
as a protective barrier against other corroding agents. They observed that the corrosion products
in the AMP system were more tenacious than the loose products formed in the MEA system.
NaVO3 proved to be the more superior of the inhibitors with more than 94% protection, as
compared to the 90% achieved by Na2SO3 (Veawab et al., 1997). They proposed that this was
P a g e | 113
due to Na2SO3 acting only to remove the dissolved oxygen and does not have that same effect
on the corroding agents such as hydrogen carbonate and water which may have played a vital
role in the corrosion process. (Veawab et al., 1997) also observed that high solution flow rates
resulted in breaking up of the protective film formed by the inhibitors or preventing the film
from being formed completely. The potential-pH diagram for vanadium-water systems is
shown in Figure 3-12. Whilst NaVO3 is soluble in neutral and alkaline solutions, it’s inhibitive
action is attributed to the competitive adsorption of monovanadate anions on the metal surface
and reduction to form insoluble lower oxides (salts) with dissolved metal ions, which act as a
passive layer and consequently decreases the rate of iron dissolution.
VO.H2O23+ VO2.H2O23+
1.6
1.4 VO2+ H2V10O284- V10O286-
1.2
Electrode Potential (SHE) / V
V2O 5.H 2O
1 HV10O285-
0.8
V4O124- HV2O73-
0.6 VO2+
0.4 V6O 13
V2O 4
0.2 V4O92-
0 V3+ V3O 5
-0.2
-0.4
-0.6 V2O 3
-0.8 V2+
-1 V
VO
-1.2
0 1 2 3 4 5 6 7 8 9 10 11 12
pH
Figure 3-12: Potential-PH diagram for the V-H2O system at 298K. Activity of dissolved
species = 0.01 (Kelsall et al., 1992).
P a g e | 114
3.9. Equilibrium Models for Aqueous Amine CO2 Systems
A quantitative understanding of the chemical and phase equilibria within amine-CO2-H2O
systems is vital for the prediction of corrosion within an amine scrubbing plant.
Thermodynamic modelling of the aqueous amine systems enables the speciation and redox
behaviour to be predicted as functions of solution composition and operating parameters such
as CO2 loading and temperature. Whilst experimental data are important in terms of offering
direct knowledge on the solubility of CO2 in amine systems, at ‘extreme conditions’ such as
lean CO2 loading, high operating temperature or very high amine concentrations, it is
sometimes difficult to obtain accurate data because of analytical constraints. Thermodynamic
models have been used to complement experimental data, and the models usually require
experimental data to feed in the parameter regression. Thermodynamic models can then be
used to predict the solution chemistry for an entire range of operating conditions beyond the
experimental measurements.
3.9.1. Solution Chemistry
The equilibrium between the amines, CO2 and H2O is assumed to be governed by the following
chemical reactions:
- Dissolution of carbon dioxide
!"#(%) ↔ !"# (3. 60)
- Formation of carbamate
2IJ,# + ,# " + !"# ↔ IJ,!""4 + IJ,3 5 + ,# " (3. 61)
- Dissociation of protonated amine ion
- IJ,3 5 ↔ IJ,# + ,5 (3. 62)

P a g e | 115
- Hydrolysis of carbamate
IJ,!""4 + ,# " ↔ IJ,# + ,!"3 4 (3. 63)
- Hydrolytic ionization of CO2(aq)
!"# (GH) + ,# " ↔ ,5 + ,!"3 4 (3. 64)
- ÇGV7[ ÉW\\]cWGVW]^
,# " ↔ ,5 + ",4 (3. 65)
- Dissociation of hydrogen carbonate ion
,!"3 4 ↔ !"3 #4 + ,5 (3. 66)
As remarked by Austgen et al, other side reactions can also take place, in addition to the ones
outlined above. However, they also noted that these side reactions are negligible on a
laboratory scale and can be ignored in vapour-liquid equilibrium (VLE) studies.
3.9.2. Thermodynamic Framework
The thermodynamic framework for the VLE of amine-carbon dioxide-water systems should
cover both the chemical and phase equilibria. The chemical equilibria is described by the
reactions (3. 60) - (3. 66) and can be summarised with the mass action law in equation (3. 67).
Ü
k Ñ = GÖ á = (àÖ âÖ )Üá (3. 67)
where K(T) is the temperature-dependent equilibrium constant, ai is the activity of species i, vi
is the reaction stoichiometry of species i, mi is the molality of species i and ϒI is the activity
coefficient of species i.
Henry’s law is usually employed to relate the CO2 partial pressure, PCO2, to the dissolved gas
concentration in the amine solution, [CO2], via equation (3. 68).

P a g e | 116
./01 = ,/01 [!"# ] (3. 68)
where HCO2 is the Henry’s law constant
The equilibrium equations (3. 69) to (3. 73) corresponding to reactions (3.60) – (3.66) are:
|äã1 [ã r ]
kQ = (3. 69)
[|äãs r ]
|äã1 [ã/0s å ]
k# = (3. 70)
[|äã/00 å ]
[ã r ][ã/0s å ]
k3 = (3. 71)
[/01 ]
kç = ,5 [",4 ] (3. 72)
ã r [/0s 1å ]
kn = (3. 73)
[ã/0s å ]
Apart from the equilibria equations, the system is also subject to mass and charge balance
constraints:
Amine material balance:
éèS = IJ,# + IJ,3 5 + [IJ,!""4 ] (3. 74)
Carbon material balance:
éèS ê/01 = !"# + ,!"3 4 + !"3 #4 + [IJ,!""4 ] (3. 75)
Charge balance:
IJ,3 5 + ,5 = ",4 + ,!"3 4 + 2 !"3 #4 + [IJ,!""4 ] (3. 76)
Where [MEA] is the initial amine concentration in kmol m-3, [CO2] is concentration of
dissolved CO2 kmol m-3 and αCO2 is the CO2 loading (mol CO2 (mol amine)-1).
P a g e | 117
There have been several attempts by various researchers to develop predictive models, built
upon a large body of VLE data (Atwood et al., 1957; Danckwerts and McNeil, 1967; Edwards
et al., 1978; Deshmukh and Mather, 1981; Austgen et al., 1989; Arcis et al., 2009) Thus, there
are a number of approaches that can be employed to correlate experimental VLE data. Initially,
curve-fitting attempts were adopted (Mason and Dodge, 1936) which then progressed to more
rigorous thermodynamic models. The major equilibrium models for aqueous CO2-amine
systems are i) The Kent-Eisenberg model ii) The model of (Edwards et al., 1978) iii)
Deshmukh- Mather method (Deshmukh and Mather, 1981) iv) Electrolyte- NRTL (e-NRTL)
model (Chen et al., 1986) v) SAFT-VR model (Galindo., 1998) and other types of model such
as the electrolyte –UNIQUAC model used by (Kaewsichan et al, 2001). The Kent-Eisenberg
model was used in this thesis and is described below.
3.9.2.1. Kent-Eisenberg Model
(Kent and Eisenberg, 1976) proposed the first well-accepted VLE model for aqueous amine
absorption processes. They developed a simple correlation method, which included all the non-
ideality in the liquid phase into definite assigned equilibrium constants. Whilst it was not
developed initially to be a predictive model, it is frequently used especially in industrial
settings.
(Kent and Eisenberg, 1976) made the assumption that provided the MEA-CO2-H2O system
behaves ideally, the Henry’s law constant and equilibrium constants should only be
temperature dependent. The activity coefficients of all the species were set to unity. Initially,
they fitted available published data to determine empirical expressions for the Henry’s law
constant and the equilibrium constants K3 – K5. Then they fitted published equilibrium partial
pressure data for H2S- MEA and CO2-MEA systems to obtain an expression for K1, which
represents the amine protonation and to obtain a similar expression for K2, which represents the
P a g e | 118
carbamate formation. Although only MEA or DEA solutions used for the selective absorption
of mixed CO2 and H2S gases were studied in the original paper by (Kent and Eisenberg, 1976),
this model can easily be extended to other amines, even mixed amine systems. The
temperature-dependent equilibrium constants Ki (i = 1-5) and Henry’s constant HCO2 are
represented by the following empirical expression:
lnkÖ ][,/01 = GÖ + íÖ /Ñ + cÖ /Ñ # + ÉÖ /Ñ 3 + 7/Ñ ç (3. 77)
Where ai- di are constants for T in oR. The values of these constants obtained by (Kent and
Eisenberg, 1976) are listed in Table 3-7.
Table 3-7: Equilibrium constants used in the Kent-Eisenberg model (Kent and Eisenberg,
1976).
Equilibrium Constant / mol L- a b × 10-4 c × 10-8 d × 10-8 e × 10-8

K
1 1 -3.3636 -1.0532 0 0 0
K2 6.69425 -0.556349 0 0 0
K3 241.818 53.6855 -4.8123 1.94 -2.96445
K4 39.5554 -17.7822 1.843 -0.8541 -1.4292
K5 -294.74 65.5893 -5.9667 2.4249 -3.7192
HCO2 22.2819 -2.48951 0.223996 -0.090918 0.12601
The Kent-Eisenberg model has proven to be extremely popular amongst industry experts, due
to due to its relative simplicity and reasonable accuracy. Several researchers in academia have
also used this model to correlate their experimental data. (Kritpiphat and Tontiwachwuthikul,
1996) employed this model for their solubility data of CO2 in AMP, whilst (Park et al., 2002)
fitted their data for CO2 solubility in MEA, DEA, AMP, MEA + AMP and DEA + AMP to
obtain empirical expressions for the equilibrium expressions for the amine protonation (K1)
and carbamate formation (K2). The correlations obtained by (Park et al., 2002) were then
P a g e | 119
compared with experimental data obtained by (Tontiwachwuthikul et al., 1991) which,
achieved reasonable results within an average deviation of ≤ 14.9% .
The Kent-Eisenberg model has been found to be accurate within the CO2 loading range of 0.20
– 0.70 mol CO2 (mol amine)-1 (Weiland et al., 1993), but is inaccurate at more extreme CO2
loadings.
3.10. Review of the mechanisms of iron dissolution and hydrogen evolution
This section reviews the literature on the proposed mechanisms for iron dissolution, as well as
the reported exchange current densities and the corresponding Tafel slopes. Studies conducted
on the exchange current densities and Tafel constants for hydrogen evolution on Fe will also
be discussed. Finally, a review on the few experimental studies of Tafel slopes and exchange
current densities in CO2-H2O systems will be presented.
3.10.1. Mechanism of the dissolution of iron
3.10.1.1. Catalytic mechanism
(Drazic, 1989) provides a comprehensive review of the mechanisms for the dissolution of iron.
(Roiter et al, 1939) published the first set of studies on the anodic dissolution of iron, which
were carried out in acidic iron (II) sulfate (FeSO4) solutions. The experiments were conducted
by anodic constant - current (galvanostatic) measurements, from which they obtained a Tafel
slope of ≈ 60 mV dec-1 and an exchange current density (j0) of 10-8 A cm-2. Subsequently
(Kuznetsov and Iofa, 1947) studied the anodic dissolution of iron in 1 M hydrochloric acid
(HCl) and recorded an anodic Tafel slope of βa ≈ 60 mV dec-1. In a following report (Kuznetsov
et al., 1947), anodic dissolution measurements were conducted in alkaline 2 M sodium
hydroxide (NaOH) and they reported an anodic slope βa ≈ 40 mV dec-1. Thus, a reaction
mechanism in which a second-order reaction dependence on OH- ions was proposed. By

P a g e | 120
comparing their results with those conducted in acidic media, they concluded that these two
reactions either occur via different mechanisms or are strongly affected by the solution pH.
The influence of pH on the anodic dissolution of iron was the motivation behind the first
detailed suggestion for the iron dissolution mechanism. (Hoar and Hurlen, 1959) found the
reaction order witth respect to [OH-] pOH- = 1, with βa = RT / 2F in sulfuric acid solutions.
Soon after, (Bonhoeffer and Heusler, 1956) reported pOH- = 1.6 which they took to be 2 and βa
= RT/2F. Based on this, (Heusler, 1957) proposed a mechanism for Fe dissolution in which
they suggested the participation of a species (FeOH)ads as a catalyst. The mechanism suggests
the simultaneous transfer of two electrons in one reaction step, which had not been previously
featured in electrode kinetics. This mechanism will be described in more detail in section
3.10.1.2.
(Bockris, Despic and Drazic, 1961) also conducted studies of Fe dissolution in acid sulfate,
chloride, perchlorate and nitrate solutions and obtained anodic slopes βa = 2RT/3F and
cathodic slopes βc = 2RT/ F and the reaction order in [OH-] was ≈ 1. Based on this and other
observation, (Bockris, Despic and Drazic, 1961) proposed the well known ‘BDD mechanism’,
a consecutive electrode mechanism established upon the existence of an intermediate species
(Fe(OH)+), produced from (FeOH) which participated in the mechanism sequence but no
longer acted as a catalyst. Unlike the mechanism proposed by (Heusler, 1958), only single
electron steps were considered to be involved.
Table 3-7 summarises the results obtained by various authors on the dissolution of iron in
acidic, neutral and alkaline media, respectively. It appears that amongst other factors, the
condition of the metal surface, the type and the content of the steel also play a significant role
(Akiyama et al., 1970; Felloni, Cozzi and Cammarota, 1967; Burnstein and Kearns, 1984).
Based on a comprehensive collection of Fe dissolution kinetic data, there is a clear dependence

P a g e | 121
of the experimental results on the properties of the iron surface, such as the type of impurities,
the dislocation density and internal stresses and crystallinity. (Lorenz & Eichkorn, 1965;
Hilbert et al., 1971) observed that the same sample of iron can produce diverse kinetic data,
depending on both the surface and internal structure of the metal. (Akiyama et al., 1970) found
that annealing of the iron sample could shift the Tafel slope for the anodic dissolution (βa )
from 30 to 40 mV dec-1, while the duration for which the iron surface was in contact with the
electrolyte could decrease the slope from 60 to 40 mV dec-1.
(Drazic and Vorkapic, 1978) proposed a ‘branching mechanism’ to account for the observed
changes in the rate determining step in the BDD reaction mechanism. (Schweickert et al.,
1980) and (Keddam et al., 1981) proposed an alternative but similar branching mechanism.
In terms of experimental measurements for corrosion in the aqueous amine-CO2 solutions, it
is important to determine if the kinetic data for pure iron can be used reliably as an acceptable
surrogate for those on carbon steel, from which the typical amine scrubbing equipment are
fabricated. As for the anodic reaction, the answer is yes; however, for the cathodic reaction,
evidence exists that the presence of Fe3C may affect the mechanism and may also change the
rate-determining step from that on pure iron (Flitt and Bockris, 1982).
The effect of pH on amine-CO2- H2O systems is a fundamental process parameter, so the effect
of pH on the kinetics and mechanisms of Fe dissolution and passivation will be discussed
below.
P a g e | 122
Table 3-8: Compilation of diagnostic criteria for Fe dissolution and deposition reactions
obtained under different experimental conditions (Drazic, 1989).
Acidic Media
Solution T/ pH βa / j0 / A Reactio Reference
o
C mV cm-2 n order
dec-1 in [OH-]
1.25 M FeSO4 60 10-8 (Roiter et al. , 1939)
0.5 M H2SO4 + 1.2 -4.9 35 – 45 (Bockris, Despic and
0.5 M Drazic, 1961)
Fe2(SO4)3
H2SO4 + K2SO4 0.3-2 40 Hilbert et al., 1971
0.5 M H2SO4 40 10-13 (Foroulis, 1966)
0.5 M H2SO4 + 45 0.8 (Aksut, 1983)
0.5 M Na2SO4
0.5 M H2SO4 30 (Akiyama et al., 1970)
Un-annealed
0.5 M H2SO4 38-45
Annealed
H2SO4 + 0 60 0 (Bala, 1984)
Na2SO4 0.75- 1.2 40 1
2.1 30 2
1 M H2SO4 + 2.4 85 0.3 (El Miligy et al., 1973)
Na2SO4 3.3 36 1.4
1 M H2SO4 + 25 0–4 32 (Voight, 1968)
Na2SO4
0.5 M H2SO4 + 1.2 – 4.9 34 1.75 (Bockris and Kita, 1961)
0.5 M
Fe2(SO4)3
0.5 M H2SO4 + 30 1- 3.3 40 (Kelly, 1961)
Na2SO4
0.5 M H2SO4 + 28 37 (Shemenski et al., 1965)
(0.01-0.5) M
Na2SO4
0.5 M H2SO4 40 45 (Hoar and Holliday, 1953)
1.5 M H2SO4 + < 3.5 38 1.0 (Abdul et al., 1972)
K2SO4 > 3.5 40 1.4
0.25 M H2SO4 25 30 (Felloni, 1968)
+ 1M K2SO4
FeCl2 + KCl + 1-4 29 (Hurlen, 1962)
HCl
1 M HCl + KCl 1.5 – 3 30 (Frankenthal, 1971)
0 - 1.5 60
P a g e | 123
2 M NaCl + 50 0–5 42 10-8 (Podesta and Arvia, 1965)

H2SO4
1 M HCl 60 10-11 (Foroulis, 1966)
Neutral Media
Solution T pH βa / j0 / A Reaction Source
mV cm-2 order in
dec-1 [OH-]
Na2SO4 25 4.7- 6 64 (Lorenz and Vilche,

No stirring 1972)
Na2SO4 66
Stirring
0.5 Na2SO4 20 6.2 – 7.9 60 0 (Drazic and Hao, 1982)
0.5 Na3BO3 + 20 6 – 10 60 (Hoar, 1961)

NaOH + 2.5 M
Na2SO4
Alkaline Media
Solution T pH βa / j0 / A Reaction Source

mV cm-2 order in
dec-1 [OH-]
0.05 – 3M NaOH 20 > 3.5 30-40 (Kabanov et al., 1947)
0.1- 3 M NaOH 25 30 (Hurlen, 1963)
0.1 KOH + 1M 25 65 (Asakura, Nobe, 1971)

KCl
0.05 – 5M KOH 25 65-70 (Drazic and Hao, 1982)

P a g e | 124
3.10.1.2. The mechanism of Fe dissolution in Acidic solutions
(Heusler, 1958) suggested a mechanism for iron dissolution based on the results of steady –
state measurements with a recorded anodic and cathode Tafel slope, βa = βc = 30 mV dec-1 and
pOH- =2 and βa = βc = 30 mV dec-1 from transient measurements in perchlorate solutions.
The rate determining step was proposed as:
67 + ",4 + 67", (EB ⟶ 67",5 + 67", (EB + 27 4 (3. 78)
followed by a fast equilibrium reaction
67",5 ⟷ 67 #5 + ",4 (3. 79)
(FeOH)ads is considered as a catalyst, formed via a slow parallel reaction:
67 + ,# " ⟷ 67", (EB + ,5 + 7 4 (3. 80)
The transient Tafel lines from galvanostatic peak potentials were thought to involve the
catalyst (FeOH)ads, such that:
pö
ï(,Öe = 6 67", (EB 67", (EB,ó(ò ",4 7ôZ (3. 81)
|K
The catalyst (FeOH)ads slowly accumulates and the steady state current density is given by:
#pö
ï(,õõ = dõõ ",4 # 7ôZ (3. 82)
|K
giving βa = 29 mV dec-1 and pOH- = 2.
For the cathodic reaction:
67",5 + 67 67", (EB + 27 4 ⟶ 67 + ",4 + 67 67", (EB (3. 83)
Fe(FeOH)ads is the adsorbed catalyst formed from the reaction:

P a g e | 125
67 + 67",5 + 7 4 ⟷ 67 67", (EB (3. 84)
The steady state current density would be represented by:
#pö
Wïó,õõ = dõõ ",4 # [67 #5 ]7ôZ (3. 85)
|K
giving βc = 30 mV dec-1 and pFe2+= 2 and poH- = 2
The foundation of this mechanism was the observation by (Heusler, 1958) of the transient
Tafel slopes βa = 60 mV dec-1, with an anodic steady-slope of βa = 30 mV dec-1 observed by
(Hoar and Hurlen, 1959) and also by (Heusler, 1957) although it was not published. (Heusler,
1958) found the coefficient δ log jcorr / δ pH to be zero while the mechanism predicts -2/5.
Subsequently, (Heusler, 1958) suggested a physical interpretation for the catalysts (FeOH)ads
and Fe(FeOH)ads, in that they represent activities at kink sites. The catalytic mechanism is based
on an assumption that the surface morphology of iron is independent of the electrode potential;
whereas there is evidence from electron microscopic studies that the distance between edge
sites is dependent on the potential (Bockris et al., 1961).
(Bockris et al., 1961) found several inconsistencies with the catalytic mechanism. Firstly,
highlighted that (Heusler, 1958) neglected the fact that the pH at the interface differs from that
in the bulk solution and depends upon potential. They also pointed out that the mechanism
involves a single reaction step with the simultaneous transfer of two electrons, which results
in calculated heats of reaction that are significantly larger than was observed. Thus, Bockris,
Despic and Drazic, proposed the BDD mechanism, which is based on the concept that Fe
dissolution occurs in two consecutive one-electron-transfer steps. The reaction sequence
shown in equations (3. 86) – (3. 88) was suggested as representing the mechanisms under the
conditions resulting in a reaction order of unity with respect to OH- and βa = 2RT/3F (38 mV
dec-1).
P a g e | 126
5
67 + ,# " ↔ 67", (EB + , + 74 (3. 86)
hEB
67", (EB 67",5 + 7 4 (3. 87)
67",5 + ,5 ⟷ 67 #5 + ,# " (3. 88)
The mechanism was established from experimental data measured in sulfate solutions within a
pH range of 1-5 and the effect of other anions was also studied. (Bocrkis et al., 1961) also made
measurements to obtain the cathodic data for Fe deposition, which included the co-deposition
of H and the resulting pH change. They accomplished this by making parallel measurements
of the hydrogen evolution rate, so the total current density jfe could be deconvoluted from jtotal
and the interfacial pH could be calculated. The cathodic slope βc was calculated to be 38 mV
dec-1 (2RT/3F) which corresponded with the experimental results.
The BDD mechanism was used by several researchers to elucidate their experimental results.
However, there was a significant issue, which was frequently raised about the amount and
stability of the hydroxo intermediate in the proposed mechanism when the reaction occurs in
acidic media. (Batrakov et al., 1968) conducted faradaic impedance measurements and
calculated that the surface coverage (θ) of (FeOH)ads in sulfuric acid solutions varies between
0.1 - 4% for the potential range of -0.2 to + 0.12 V (versus standard hydrogen electrode (SHE)).
The BDD mechanism predicts an anodic slope of 40m V dec-1, but as can be seen from Table
3-7, a few other researchers reported βa of 30 mV dec-1 for steady state measurements that
corresponds with the catalytic mechanism by (Hesuler, 1958). Hilbert et al., 1971) attempted
to explain the inconsistency between the results obtained by the different researchers. They
conducted a series of experiments from which they recorded a βa of 40 mV dec-1 and a reaction
order of 1 with respect to OH- for annealed iron wire, βa of 30 mV dec-1 and a reaction order
of 1.70 with respect to OH- when the wire was subjected to a large strain. (Hilbert et al., 1971)
P a g e | 127
proposed that this was due to a change in mechanism from the BBD to the catalytic mechanism,
resulting from the manifestation of a sizeable amount of ‘active sites’ as a result of the strain
on the wire. Hence, it was concluded that the annealed and less strained material undergoes
anodic dissolution via the BDD mechanism, while heavily strained materials undergo the
catalytic mechanism proposed by (Hesuler, 1958).
(Drazic and Vorkapic, 1978) suggested that the adsorption energy of the hydroxo intermediate
(FeOH)ads increases with strain, leading to a rate-determining desorption reaction (reaction (3.
89)) and the anodic Tafel slope βa shifts to RT/2F.
67", (EB → 67", BCDe (3. 89)
(Drazic and Vorkapic, 1978) suggested that the change in kinetic parameters was less likely to
result from a change in mechanism but from a change in the rate determining step in the same
reaction sequence. They then proposed a modified version of the BDD mechanism after taking
into account the possibility of the presence of adsorbed FeOH, FeOH+ and Fe(OH)2 species
even in acidic media and also the possibility that the desorption of these adsorbed species could
be activated and hence the rate determining step. This modified version of the BDD
mechanism, called the branching mechanism, is shown in Figure 3-12. Consequently, for an
activated surface, the anodic slope βa tends to RT/2F and the reaction order with respect to
OH- can shift towards 2, if the rate constants for the two parallel desorption reactions are of
the same order. (Drazic and Vorkapic, 1978) suggested that there are conditions under which
the branching mechanism occurs and the diagnostic criteria are shown in Table 3-9.
P a g e | 128
Table 3-9: Diagnostic Criteria for the ‘Branching Mechanism’ at Low and High Surface
Coverages if Step 2, 3 or 4 is Rate-Determining’ (Drazic and Vorkapic, 1978).
Rate θ << 1 θ >>1

determining
step αa βa / mV dec-1 Reaction αa βa / mV dec-1 Reaction
order in order in
[OH-] [OH-]
2 1.5 40 1 0.5 120 0
3 2 30 1 0 0
4 2 30 2 0 0
(El Miliguy et al., 1975) suggested that there are three different possible states that can occur:
the activated dissolution, a transition range, pre-passive behaviour and a passive state. In the
activated dissolution region, the BDD mechanism is operative, in the transition range, Fe(OH)2
affects the dissolution process by acting as a membrane inhibitor. The branching mechanism
by (Drazic and Vorkapic, 1978) applies to the activated dissolution and pre-passive regions.
3.10.2. The influence of halide ions on dissolution kinetics
Several researchers such as (Hurlen, 1962; Lorenz, 1965; Heusler and Cartledge, 1961;
Darwish et al., 1973) studied the effect of anions on the dissolution of iron. They all reported
that anions, particularly halide ions in concentrations of < 1 M in mixtures of sulfate or
perchlorate, exhibited some sort of inhibitive effect on Fe dissolution. Anodic polarization
curves obtained by (Heusler and Cartledge, 1961; Schwabe and Voigt, 1969) exhibited an “S”
shape with a degree of hysteresis and a strong dependence on the method of surface
preparation, while the polarization curves obtained by (Hurlen, 1962; Lorenz, 1965; Darwish
et al., 1973) had well defined Tafel regions with higher slopes than in halide-free perchlorate
solutions. The reason for this discrepancy was considered to be that the second group of results
arose from electrode surfaces that have been anodically treated previously for some time,
P a g e | 129
although some of the experimental details are not well defined. The adsorption of halide ions
is relatively slow and the results of the Tafel measurements will depend on the rate at which
the experimental studies are conducted, which explains the reason for a spread in experimental
results.
There have been several mechanisms proposed for the effects of halide ions, all of which make
the assumption that these halide ions are specifically adsorbed and compete with other species
for the surface. The reaction mechanism also incorporates a step in which adsorbed halide ions
form a particular kind of adsorbed complex with Fe and OH species that subsequently desorbs
from the surface. This was incorporated to avoid the argument that the adsorption of halide
ions should become stronger at more positive potentials, while also explaining the reason for
the negative reaction order for halide ions and the S-shaped polarization curves.
(McCafferty and Hackermann, 1972) studied the Fe dissolution kinetics over a wide
concentration range of chloride ions (1 – 6 M). The OH- accelerated mechanism proposed by
(Lorenz, 1965) was assumed with a reaction order of -1 with respect to [H+] and adapted to
take into account the competitive adsorption of OH- and Cl- ions. The proposed reaction
sequence is outlined from (3. 90) – (3. 93).
67 + ,# " ⟷ 67 ∙ ,# "(EB ↔ 67ú 4 (EB + ,# " (3. 90)
67 ∙ ,# "(EB + ú 4 ⟷ 67",4 (EB + ,5 + ú 4 (3. 91)
67ú 4 (EB + 67",4 (EB ⟶ 67",5 + 67 + ú 4 + 27 4 (3. 92)
67",5 + ,5 ↔ 67 #5 + ,# " (3. 93)

P a g e | 130
Then at higher H+ concentrations for which the reaction order with respect to [H+] was 2, a
different, H+ accelerated mechanism was proposed and outlined from (3. 94) – (3. 97).
67 + ,# " ⟷ 67 ∙ ,# "(EB (3. 94)
67 ∙ ,# "(EB + ú 4 ⟷ 67ú 4 (EB + ,# " (3. 95)
67ú 4 (EB + ,5 ⟷ 67ú 4 ∙ ,5 ≡ c]àZU7ô (3. 96)
hEB
67ú 4 ∙ ,5 + ,5 67ú 5 + 2,5 + 27 4 (3. 97)
(Chin and Nobe, 1972) proposed an alternative mechanism based on measurements in chloride
ions within a concentration range of (0.01 – 1.9 M) and a pH range of 0.0 – 1.8.
67 + !U 4 + ,# " ⟷ (67!U",)4 (EB + ,5 + 7 4 (3. 98)
hEB
(67!U",)4 (EB 67!U", + 7 4 (3. 99)
67!U", + ,5 ⟶ 67 #5 + !U 4 + +,# " (3. 100)
The adsorption of (FeClOH)-ads was assumed to follow a Langmuir isotherm.
For concentrated chloride solutions and very small acid concentrations, (Chin and Nobe, 1972)
modified the reaction sequence and suggested the mechanism below:
67 ∙ ,# "(EB + !U 4 ⟷ 67!U(EB + ,# " + 7 4 (3. 101)
67!U(EB + ,5 ⟷ 67!U,5 (3. 102)
hEB
67!U,5 + !U 4 67!U# + ,5 + 7 4 (3. 103)
P a g e | 131
For this mechanisms it is assumed that the surface coverage adsorbed species FeCl, θFeCl →1
and that of FeClH+, θFeClH+ →0 and they follow the Langmuir isotherm.
3.10.3. Mechanisms for Dissolution of Iron in Solutions in the Mid-pH range
There have been only a select number of reports on the mechanistic analysis of the anodic
dissolution of iron in neutral solutions, in spite of the fact that most natural corrosion occurs in
solutions with a close to neutral bulk pH. This may be due to several complications such as the
fact that the pH at the iron⏐solution interface is particularly dependent on the current density
and surface coverage with adsorbed oxide or hydroxide films is significantly higher in acid and
varies with time and potential.
(Hoar and Ferrer, 1965) and (Lorenz and Vilche, 1972) conducted some of the earliest studies
within neutral solutions in which (Hoar and Ferrer, 1965) obtained an anodic Tafel slope βa of
60 mV dec-1 in dilute solutions of several anions, whilst (Lorenz and Vilche, 1972) recorded
βa as 66 mV dec-1 and a reaction order of 0.7 with respect to [OH-] in stirred sulfate solutions
within the pH range of 5-6. The active dissolution of iron in both studies was interpreted by the
BDD mechanism and it was also assumed that increases in the local pH led to higher surface
coverage with intermediate species, following a Tempkin adsorption isotherm and resulting in
a Tafel slope of RT/F.
(Drazic and Hao, 1982) carried out a more in-depth study on the effect of pH in neutral sulfate
solutions (pH 6 - 8). They found that the rate of the anodic reaction was independent of pH and
recorded an anodic Tafel slope of 60 mV dec-1 at lower current densities and 120 mV dec-1 at
higher current densities, the latter also determined by (Freiman and Kolotyrkin, 1965). Thus,
(Drazic and Hao, 1982) proposed an extension to the BDD branching mechanism to
accommodate neutral solutions which corresponds to the surface coverage by intermediate

P a g e | 132
species FeOHads θFeOH →1, a reaction order of 0 with respect to OH- ions and βa as 120 mV
dec-1.
3.10.4. Mechanisms of Iron Dissolution in Alkaline Solutions
More extensive research on Fe dissolution at alkaline pH has been reported than in the other
pH ranges, due to its relevance to Fe-Ni batteries. One of their earliest notable mechanisms is
by (Kabanov et al., 1947) in which two anodic Tafel slopes of 30 and 40 mV dec-1 and pOH- of
2 was observed.
67 + ",4 ↔ 67",(EB + 7 4 (3. 104)
67",(EB + ",4 → 67"(EB + ,# " + 7 4 (3. 105)
67"(EB + ",4 ↔ ,67"# 4 (3. 106)
,67"# 4 + ,# " ↔ 67(",)# + ",4 (3. 107)
Whilst the existence of oxide or hydroxide films was accepted generally within this
mechanism, initially there was a limited use of spectroscopic methods to interrogate the
surfaces. However, (Kolotyrkin et al., 1987) employed modulated reflectance spectroscopy to
examine presence of FeOHads species in 0.1 M NaOH and Neugeberger et al employed internal
reflection studies from which they proposed that HFeO2- (adding water gives Fe(OH)3-?) is the
reaction product and forms Fe(OH)2 on hydrolysis.
(Drazic and Hao, 1982) also conducted an extensive study of activated ARMCO iron in
alkaline solutions from which they obtained a reproducible Tafel line with a slope of about 65
mV dec-1 and a reaction order of 1.1 with respect to [OH-]. With the presence of various
concentrations of [HFeO2-] ions, they observed a negative reaction order of -0.45 with respect
to these species. They proposed a mechanism involving:

P a g e | 133
67 + ,# " ↔ 67",(EB + ,5 + 7 4 (3. 108)
67",(EB + ,# " ↔ 67(",)#(EB + ,5 + 7 4 (3. 109)
67(",)#(EB + ",4 ↔ ,67"# 4 + ,# " (3. 110)
,67"# 4 + ,# " ↔ 67(",)# + ",4 (3. 111)
If the second step (3. 109) in the mechanism is assumed to be rate determining with a surface
coverage of FeOH and FeOH2 between 0.2-0.8 and the adsorption follows a Temkin adsorption
isotherm, the predicted anodic Tafel slope βa is 2.3RT/F and the reaction orders with respect
to [OH-] and [HFeO2-] are equal to 1 and -0.5 respectively, in agreement with the experimental
results. The negative reaction order with respect to HFeO2- ions is in agreement with the effect
of these species as inhibitors for the dissolution process.
(Kabanov et al., 1947) also suggested a similar mechanism to that of (Drazic and Hao, 1982).
They obtained an experimental anodic slope βa of 40 mV dec-1 which is also in accordance
with the proposed mechanism if the surface coverage of the intermediate species, FeOH and
FeOH2 is less than 0.2 and the absorption follows a Langmuir-type isotherm.
P a g e | 134
67 + ,# " ↔ 67",(EB + , 5 + 7 4
Acid and neutral media Alkaline media
hEB
2. 67",(EB ûü 67", 5 (EB + 7 4 hEB
2. 67",(EB + ,# " ûü 67(",)#(EB + , 5 + 7 4
3. 67(",)#(EB + ", 4 ↔ ,67"# 4 + ,# "
lhqóÖlÖò(òÖCe
4. ,67"# 4 + ,# " †⎯⎯⎯⎯⎯⎯⎯⎯ü 67(",)# + ", 4
EqB 3. 67", 4 (EB + ,# " ↔ 67(",)#(EB + , 5

3. 67", 5 (EB ûü 67", 5
EqB
4. 67", 5 + , 5 ↔ 67 #5 + ,# " 4. 67(",)#(EB + , 5 ûü 67(",)#
5. 67(",)# + 2,5 ↔ 67 #5 + 2,# " 67", 4 (EB
Figure 3-13: General reaction scheme for iron dissolution (Drazic, 1989)
3.11. The Mechanism of Hydrogen Evolution Reaction (HER) on Iron
A comprehensive understanding of the hydrogen evolution reaction mechanism was only
achieved after 1950 through pioneering work by (Tafel, 1905; Heyrovsky, 1927; Volmer, 1930;
Funmkin, 1967; Conway, 1965; and Kita, 1966). Mechanistically there are three well-known
steps for the Hydrogen evolution reaction commonly referred to as the Volmer (1), Heyrovsky
(2) and Tafel (3) steps.

P a g e | 135
1. Volmer ∶ ,5 + 7 4 ⟶ , (EB (3. 112)
2. Heyrovsky: , (EB + ,5 + 7 ⟶ ,#(%) (3. 113)
3. Tafel: , (EB + , (EB ⟶ ,#(%) (3. 114)
The hydrogen evolution reaction is not a single step reaction; instead it proceeds in two steps
with only one intermediary reaction (Weber and Peria, 1967; Eastman and Nathan, 1975; Park,
1975). There are two widely accepted pathways for the conversion of H+ ions to molecular
hydrogen. They are the discharge, chemical recombination (Volmer-Tafel mechanism) and the
discharge, electrochemical desorption (Volmer-Heyrowsky) mechanism. (Weber and peria,
1967; Eastman and Nathan, 1975; Park, 1975; Papp and Pritchard)
In the first pathway (1 and 3) as proposed by (Volmer, 1930) and (Tafel, 1905), protons are
discharged onto the metal surface sites to form an adsorbed hydrogen atom by a single electron
transfer (reaction (3. 112)) (Weber and Peria, 1967; Eastman and Nathan, 1975; Park, 1975;
Papp and Pritchard). This is followed by a chemical reaction in which the adsorbed hydrogen
atoms (Hads) diffuse to each other across the electrode surface and undergo a chemical
recombination to form H2(g), which accumulate to form bubbles (reaction (3. 114)).
(Volmer, 1930) suggested the first reaction (3. 112) was the rate determining step, so the
surface coverage of adsorbed hydrogen θH will tend to be small because, as the H atoms are
adsorbed on the surface, they immediately react to form molecular H2. On the other hand,
(Tafel, 1905) considered reaction (3. 114) to be rate determining and the reaction (3. 112) to be
in quasi-equilibrium such that θH will be significantly higher than suggested by (Volmer, 1930).
In the second pathway (1 - 2), the first step is the same, but the electrochemical recombination
mechanism (reaction (3. 113)) consists of a second proton discharge reaction on the adsorbed
Hads produced by the first proton discharge. In this reaction mechanism (Volmer- Heyrovsky),
P a g e | 136
the first step, the proton discharge on empty sites was considered by (Kobosew and Nekrassow,
1972) to be the rate determining step. This is followed by a fast proton discharge onto the
adsorbed hydrogen atoms to form H2, so the surface coverage of adsorbed hydrogen θH will
tend to be low. However, it is more likely that the desorption is slow and becomes the rate
determining step and hence θH →1.
Hydrogen permeation rates through Fe have been measured to elucidate the rate determining
step for both the mechanism involving the chemical recombination and the electrochemical
desorption. It is possible to demonstrate that if the discharge of protons is the rate determining
step while the recombination is the fast reaction:
ØDe∞± p
=− (3. 115)
Øö ç|K
but if the electrochemical desorption at θH = 1 is rate determining then:
ØDe∞±
→ 0 (3. 116)
Øö
Based on experimental observations, from the equilibrium potential down to ca. -0.4 (SHE)
discharge of the protons (the first solution) was the rate determining step whilst at more
negative potentials, electrochemical desorption (the second solution) was the rate determining
step. Similar results have been observed in alkaline solutions (Bockris and Koch, 1961;
Conway and Bai, 1984).
3.12. Experimental work in CO2-containing solutions
(de Waard et al., 1975) established a relationship between the corrosion rate of bare steel in
carbonic acid solutions and the CO2 partial pressure, as discussed in Section 3.4.12. Their work
was based on determining the corrosion rates of cylindrical carbon steel samples in a stirred
glass cell at 0.1 – 1 bar CO2 and at different temperatures. They reported Tafel slopes βa = 40
P a g e | 137
mV dec-1 and βc = 120 mV dec-1. These were the only values for the Tafel slope that were
reported by (de Waard et al., 1975) and the effect of pH and temperature were neglected.
(Wieckowski et al., 1983) studied the behaviour of iron electrodes in CO2 saturated sodium
chloride solutions at 25oC, pH 3.25 with the use of cyclic voltammetry. The measurements
were conducted with the use of stationary and rotating disk (99.8 % purity) iron electrodes. The
effects of pH, CO2 partial pressure and temperature were not considered and Tafel slope
measurements were not conducted.
(Eriksrud et al., 1984) conducted a series of experiments in 4% NaCl, various synthetic and
real formation waters and in an oil/salt/water mixture at 20oC and 60oC. The experiments
involved the use of a flow rig to investigate the effects of flow on CO2 corrosion and the CO2
partial pressure was varied from 1 to 2 bars. The recorded anodic current densities were similar
in both quiescent and flowing conditions, whilst a limiting cathodic current was measured,
which increased with increasing flowrate. For measurements conducted at pH 3.88 and 20oC,
they reported an anodic slope, βa = 40 mV dec-1 and βc → infinity. They reported that at low
temperatures (20oC), the corrosion rates appeared to be independent of the solution flow rate
whilst the corrosion rate is highly flow rate-dependent at higher temperatures (60oC), especially
in the absence of protective layers. However, the effect of pH and temperature on the Tafel
constants was not included in their studies.
(Valentini et al.., 1985) studied the behaviour of polycrystalline iron electrodes in stagnant and
stirred K2CO3 and KHCO3 solutions in the 0-75oC temperature range with the use of cyclic
voltammetry. The potentiostatic studies by (Arvia et al.., 1985) showed three oxidation peaks
in stagnant solutions and two in stirred solutions. The initial oxidation peak was attributed to
the formation of a prepassive layer consisting of an oxide of iron, probably Fe(OH)2. The
second anodic stage was ascribed to the electrodissolution of iron through the prepassive layer.
P a g e | 138
As the reaction proceeded, both the thickness of the prepassive layer and the concentrations of
FeII and anions at the electrode⏐ solution interface increase. Once the concentration of these
ions exceeded the solubility limit for FeCO3, the latter was formed on the iron surface.
Therefore, they assigned the potential range of peak II to the formation a complex layer
structure of an inner hydrous Fe(OH)2 layer and an outer FeCO3 layer. (Castro et al., 1986) also
obtained voltammograms over a wide range of scan rates, electrode rotation rates and ionic
strengths. Their results showed that the dissolution and passivation of iron in the aqueous
solutions containing carbonate and hydrogen carbonate ions in the 7.7 – 9.2 pH range strongly
depended on the hydrodynamics and the electrolyte composition. (Castro et al., 1986) also
found two distinct current maxima in the polarisation curves, which they assigned to formation
of formation of a prepassive Fe(OH)2 inner layer and the simultaneous electrodissolution of
iron through the prepassive layer.
(Simard et al., 1997) investigated the voltammetric response of 1024 mild steel electrodes in
the presence of 0.05 -0.5 M NaCO3 and NaHCO3 at pH 8.9 and 25oC. The voltammograms
displayed an oxidation peak followed by a region of passivity. They observed a linear increase
in the peak height with the concentration of NaHCO3, which they proposed to be as a result of
the formation of a Fe2+-HCO3-/CO32- complex most likely in the form of FeHCO3+(aq). Thus
anodic dissolution peak was ascribed to the electrodissolution of iron in the form of soluble
FeHCO3+ and /or FeII with no significant amount of oxides accumulated on the surface.
(Grey et al., 1990) conducted polarization measurements to identify the active dissolution
mechanisms for carbon steel specimens in aqueous CO2 as a function of pH and temperature.
Aqueous CO2-containing solutions at pH 2 - 6 were prepared by dissolving NaHCO3 in 0.15
NaCl and purging with CO2 and solutions at pH 8 - 11 were prepared with NaHCO3 and
Na2CO3 purged with N2. They suggested that there was a change in the cathodic reaction
mechanism with pH which was directly related to changes in the CO2 speciation. Thus,
P a g e | 139
different reactions were assumed at different pH ranges and the recorded Tafel constants are
shown in Table 3-9. (Grey et al., 1990) proposed that at pHs 2-4, the active dissolution of iron
was enhanced by the chemically–controlled reduction of H2CO3. At higher pHs (6-10), the iron
dissolution was enhanced by the charge-transfer controlled reduction of HCO3-. They also
observed a rapid increase in the corrosion rate from 25oC to 90oC, however, little increases
were observed from 90oC to 125oC, which was attributed to changes in the kinetics of the iron
oxidation reaction.
Table 3-9: Electrochemical parameters derived for the redox reactions from modelling the
polarisation curves, pH 2 -10, 25oC (Gray et al., 1990).
Reaction pH Tafel Exchange Reversibl Limiting Current

Constant/ Current e µA cm-2
mV/dec-1 Density/ A cm-2 Potential
mV
Fe/ Fe2+ 2.0 50 0.00002 -924 N/A
4.0 55 0.05 -924
6.2 70 0.046 -924
8.1 72 0.024 -924
10.0 124 0.47 -924
H2O/H2 4.0 160 0.04 -478 N/A
6.2 160 0.2 -608
8.1 125 0.16 -720
10.0 78 0.25 -838
H+/H2 2.0 100 3.8 -361 3200
4.2 120 0.65 -478 32
H2CO3/H2 4.0 130 4.0 -478 144
6.2 160 2.5 -608 55
HCO3-/H2 6.2 175 1.0 -608 N/A
8.1 160 4.6 -720
10.0 133 2.0 -838
P a g e | 140
3.13. References
Abdul Azim, A. A. and Sanad, S. H. 1972. Effects of Acid Concentration. C-content and
temperature on the corrosion rate of steel in HCl. Electrochim. Acta. 12, 313-324.
AEP. 2016. Available from:
https://www.aep.com/environment/climatechange/carboncapture.aspx (Accessed 10th
February 2016).
Akiyama, A., Patterson, R. E. and Nobe, K. 1970. Electrochemical Characteristics of Iron.
Effect of Heat Treatment and Iron Purity. Corrosion 26. 51- 57.
Aksut, A. A. 1983. The Corrosion of Iron in Propargylic Alcohol and Sulphuric Acid and
Reduction Mechanism of Propargylic Alcohol. Electrochim. Acta 28. 1177-1184.
Alstom. 2016. Available from:
http://www.alstom.com/404/?404;http://www.alstom.com:80/power/news-and-
events/press-releases/alstom-chilled-ammonia-process-selected-for-leading-cO2-
capture-plant-in-romania/ (Accessed 6th February 2016).
Anderson, C. T. 1934 The Heat Capacities of Magnesium, Zinc, Lead, Manganese and Iron
Carbonates at Low Temperatures . J. Am. Chem. Soc. 56, 849-851.
Asakura, S. and Nobe, K. 1971. Kinetics of anodic processes on iron in alkaline solution.
ibid., 118, 536- 541.
Atwood, K. Arnold, M. R. and Kindrick, R. C. 1957. Equilibria for the System
Ethanolamine-Hydrogen Sulfide- Water. Ind. Eng. Chem. Res. 49, 1439-1444.
Austgen, D. M., Rochelle, G, T., Peng, X. and Chen, C. 1989. Model of Vapour-Liquid
Equilibria for Aqueous Acid Gas- Alkanolamine Systems Using the Electrolyte- NRTL
Equattion. Ind. Eng. Chem. Res. 28, 1060-1073.
Avner, S. H. 1997. Introduction to Physical Metallurgy, 2nd Edition. Tata McGraw-Hill.
Bai, L., Harrington, D. A. and Conway, B. E. 1987. Behaviour of Overpotential- deposited
species in Faradaic reactions-II. Ac Impedance Measurements on H2 Evolution
Kinetics at Activated and Unactivated Pt cathodes. 32, 1713- 1731.
Barin, I., Knacke, O. and Kubaschewski, O. 1977. Thermochemical Properties of Inorganic
Substances. Berlin, Springer-Verlag.
Bala, H. 1984. The Corrosion of Iron in Sulphate Solutions at pH = 0-2. Electrochim. Acta.
29, 119-129.
Batrakov, V. V., Avad, G. H., Mikhaolova. E. I. and Iofa, Z. A. 1968. Differential Capacity
Curves of Metals of the Iron Group and their Interpretations. Elektrokhimiya, 4, 601.
Baylis, J. R. 1926. Factors Other than Dissolved Oxygen Influencing the Corrosion or Iron
Pipes. J. Indust. Eng. Chem. 18, 4, 370-380.
Benamor, A. and Aroua, M. K. 2005. Modelling of CO2 Solubility and Carbamate
Concentration in DEA, MDEA and their Mixture Using the Deshmukh-Mather Model.
Fluid Phase Equilibria. 231, 150-162.
Berry, W. E. 1982. Effect of CO2 Corrosion of Line Pipe. American Gas Association.
Transmission Conference, Chicago, IL.
P a g e | 141
Bishnoi, S., and Rochelle, G. T. 2002. Absorption of Carbon Dioxide in Aqueous Piperaxine/
Methyldiethanolamine. AlChE Journal. 48, 2788-2799.
Blanc, C., Grall, M. and Demarais, G. 1982. The Part Played by Degradation Products in the
Corrosion of Gas Sweetening Plants Using DEA and MDEA. Presented ar the Laurance
Redi Gas Conditioning Conference (LRGCC) Norman.
J., Scharifker, B. R., Chandreskaran, K. 1987. J. Electrochem. Soc. 134, 1957.
Bockris, J. O’M. and Drazic. D. M. 1962. The Kinetics of Deposition and Dissolution of Iron:
Effect of Alloying Impurities. Electrochim. Acta. 7, 293.
Bockris, J. O’M., Drazic. D. M. and Despic, A.R. 1961. The Electrode Kinetics of the
Deposition and Dissolution of Iron. Electrochim. Acta. 4, 325.
Bockris, J. O’M., and Flitt, H. 1983. Atomistics of Fracture. Plenum Press, New York, p593
Bockris, J. O’M. and Kita, H. 1961. Analysis of Galvanostatic Transients and Application to
the Iron Electrode Reaction. J Electrochem. Soc. 108, 676- 685.
Bockris, J. O’M and Koch, D. F. A. 1961. Comparative Rates of The Electrolytic Evolution of
Hydrogen and Deuterium on Iron, Tungsten and Platinum. J. Phys. Chem. 65 1941-1948.
Bockris, J. O’M. and Reddy, A. Modern Electrochemistry. Plenum Press, Section 10.3, 1970.
Bonhoeffer, K. F. and Heusler, K. E. 1956. Z Phys. Chem. (N.F) 390.
Bottoms, R. R. 1930. Separating acid gases. U.S. Patent 1783901.
Brouwer, J. P., Feron, P. H. M., and Asbroek, N. A. M. 2005. Amino-acid salts for CO2
Capture from flue gases. Fourth Annual Conference on Carbon Capture & Sequestation,
Alexandria Vrginia, 5th May, USA.
Browning, G. J. and Weiland, R. H. 1994. Physical Solubility of Carbon Dioxide in Aqueous
Alkanolamines Via Nitrous Oxide Analogy. J. Chem. Eng. Data, 39, 817–822.
Burke, P. A. 1984. Recent Progress in the Understanding of CO2 Corrosion, Advances in CO2
Corrosion, Vol. 1. Paper 3. NACE, Houston, TX.
Burke, P. A., and Hausler, R. H. 1985. Assessment of CO2 corrosion in Cotton Valley
Limeston Trend, Advances in CO2 corrosion. Advances on CO2 Corrosion. Ed. Hausler,
R. H and Godard, H. P. NACE, Houston, TX.
Burnstein, G. T. and Davies, D. H. 1980. The Effects of Hydrogen Carbonate on the Corrosion
and Passivation of Iron. Corrosion 36, 8, 418.
Burnstein, G. T. and Kearns, M. A. 1984. J Electrochem. Soc 131 991.
Castro, E. B., Valentini, C. R., Moina, C. A., Vilche, J. R. and Arvia, A. J. 1986. The Influence
of Ionic Composition and Passivation of Iron Electrodes in Potassium Carbonate-
Bicarbonate Solutions in The 8.4-10.5 pH Range at 25oC. Corrosion Science. 26, 10,
781-793.
Castro, E. B. and Vilche, J. R. 1992. Study of Iron Electrodissolution in Carbonate /
Bicarbonate Solutions by Electrochemical Impedance Spectroscopy. J.Electroanal.
Chem., 323. 231-246.
Chai, L. and Navrotsky. 1994. Enthalpy of Formation of Siderite and its Application in Phase
Equilibrium Calculation. Am Mineral. 79, 921-929.
P a g e | 142
Chen, C. C. 1982. A Local Composition Model for the Excess Gibbs Energy of Aqueous
Electrolyte Systems. I: Single Completely Dissociated Electrolyte Single Solvent
Systems. AIChE. J. 28, 588-596.
Chin, R. J. and Nobe, K. J. 1972. Electrodissolution Kinetics of Iron in Chloride Solutions. 3.
Acidic Solutions. J Electrochem. Soc. 119, 1457- 1461.
Conway, B. E. 1965.Theory and Principles in Electrode Processes, Ronald Press Co., New
York.
Conway, B. E., Bai, L and Tessier, D. F. 1984. Data Collection and Processing of Open-Circuit
Potential-Decay Measurements using a Digitial Oscilloscope: Derivation of the H-
Capacitance Behaviour of H2-evolving, Ni-based Cathodes. J Electroanal. Chem. 161,
39-49.
American Petroleum Institute. 1958. Corrosion of Oil- and Gas-Well Equipment. NACE.
Dallas, TX. 7
Craig, H. L. and MClaughlin, B. D. 1996. Corrosive Amine Characterization. Paper 394,
Corrosion 96. NACE. Houston, TX.
Crolet, J. L. 1994. Which CO2 Corrosion? Hence, Which Prediction? Predicting CO2 Corrosion
in the Oil and Gas Industry. European Federation of Corrosion Publication, No 13.
London, UK. The Institute of Materials.
Crolet, J. L. and Bonis, M. R. 1986. A Tentative Method for Predicting the Corrosivity of
Wells in New CO2 Fields. Materials Performance. 25, 41-49
Crolet, J. L. and Bonis, M. R. 1989. Basics of the Production of the Risks of Co2 Corrosion in
Oil and Gas Well. Paper 466. Corrosion 89.
Crolet, J. L. and Bonis, M. R. 1991. Prediction of The Risks of CO2 Corrosion in Oil and Gas
Wells. SPE Production Engineering, 1991, 6, 4, 449.
Crolet, J. L., Olsen, S., and Wilhelmsen, W. 1994. Influence of a Layer of Undissolved
Cementite on the rate of CO2 corrosion of carbon steel. Paper 4. Corrosion 94. NACE.
Houston, TX.
Crolet, J. L., Olsen, S., and Wilhelmsen, W. 1995. Observation of Multiple Steady-State in the
CO2 Corrosion of Carbon Steel. Paper 188. Corrosion 95. NACE. Houston, TX.
Crolet, J. L., Trebenot, N., and Nesic, S. 1996. Role of Conductive Corrosive Products on the
Protectiveness of corrosion Layers. Paper 2, Corrosion 96. NACE. Houston, TX.
Cullinane, J. T. and Rochelle, G. T. 2003. Properties of concentrated aqueous potassium
carbonate/piperazine for CO2 capture. Proceedings, Second Annual Conference on
Carrbon Sequestration. Alexandria, VA.
Danckwerts, P. V. and McNeil, K. M., 1967a. The absorption of carbon dioxide into aqueous
amine solutions and he effects of catalysis. Transaction of Institution of Chemical
Engineers. 45, T32- T-49.
Danckwerts, P. V. and McNeil, K. M., 1967b.The effects of catalysis on rates of absorption of
CO2 into aqueous amine-potash solutions. Chemical Engineering Science. 22, 925- 930
Dang, H. & Rochelle, G. T. 2003. CO2 Absorption Rate and Solubility in
Monoethanolamine/Piperazine/Water. 38, 337–357.
P a g e | 143
Darwish, N. A., Hilbert, F., Lorenz, W. J. and Roswag, H. 1973. The Influence of Chloride
Ions on the Kinetics of Iron Dissolution. Electrochim. Acta. 18, 421- 425.
Davis, S. H. 1993. Handbook of Crystal Growth. Ed. Hurle, D. T. J. , vol 1b. North Holland.
860.
Dawodu, O. F. & Meisen, A. 1996. Degradation of alkanolamine blends by carbon dioxide.
Can. J. Chem. Eng., 74, 960-966.
DeBerry, D. W., Clark, W. S. and Yost, A. 1979. Corrosion Due to the Use of Carbon Dioxide
for Enhanced Oil Recovery. DOE Final report MC/ 08442, SumX Corporation. Austin,
TX. 19, 342
Department of Trade and Industry. 2005. Optimising CO2 Capture from Pulverised Coal Plants:
Post Combustion Capture with amine solvents. London. DTI.
Deshmukh, R. D. and Mather, A. E. (1981). A Mathematical Model for Equilibrium Solubility
of Hydrogen Sulfide and Carbon Dioxide in Aqueous Alkanolamine Solutions. Chem.
Eng. Sci. 36, 355-362.
de Waard, C., and Milliams, D. E. 1975. Carbonic Acid Corrosion of Steel. Corrosion 31, 131.
de Waard, C., Milliams, D. E., and Lotz, U. 1991. Predictive Model for CO2 Corrosion
Engineering in Wet Natural Gas Pipelines, Paper 577. Corrosion 91. NACE. Cincinnati,
OH.
de Waard, C., and Lotz, U. 1993. Prediction of CO2 Corrosion of Carbon Steel, Paper 69.
Corrosion 93. Houston, TX: NACE.
de Waard, C., Lotz, U., and Dugstad, A. 1995. Influence of Liquid Flow Velocity on CO2
Corrosion: A Semi-Empirical Model. Paper 128, Corrosion 95. NACE. Orlando, FL.
Dingman, J. C., Allen, D. and Moore, T. F. 1966. Minimize Corrosion in MEA Units.
Hydrocarbon Processing 45, 9, 285-290.
Dow Chemical Company. 1962. Gas Conditioning Fact Book. 145-234, Midland MI.
Drazic, D. M. Modern Aspects of Electrochemistry. No 19, Plenum Press, New York, 1989.
Drazic, D. M. and Hao, C. S. 1982. The Anodic Dissolution Process on Active Iron in Alkaline
Solutions. Glasnik Hem. Drusiva Beograd 47, 649.
Drazic, D. M. and Vorkapic. L. Z. 1977. Z. Hem, Glas. Hem. Drus., Beograd 42, 545.
Dresel, P. E. 1989. The Dissolution Kinetics of Siderite and its Effect on Acid Mine Drainage.
PhD thesis, Pennsylvania State University, University Park, PA, USA.
Drogowska, M., Brossard, L. and Menard, H. 1998. 304 Stainless Steel Oxidation in Sulfate
and Sulfate + Bicarbonate Solutions. Journal of Applied Electrochemistry. 28, 491-501.
Dugstad, A. 1992. The Importance of FeCO3 Supersaturation on the CO2 Corrosion of Carbon
Steel. Paper 14. Corrosion 88. NACE. New Oreleans, LA.
Dugstad, A. 1997. Formation of protective corrosion filsm during CO2 corrosion of carbon
steel. Conference of European Fedration of Corrosion. London, UK. Insitutue of
Mateirlas.
Dugstad, A. 1998. Mechanism of Protective Film Formation During CO2 Corrosion of Carbon
Steel. Paper 31. Corrosion 98. NACE. Hoston, TX.
P a g e | 144
Dugstad, A., Hemmer, H. Kooyman, H. Wilhelmensen,W. 1993. “Kjeller Sweet Corrosion III-
Final Report”, IFE report no. IFE/KR/F-93/091, Insitute for Energy Technology. Kjeller,
Norway.
Dugstad, A., Lunde, L., and Videm, K. 1991. Influence of alloying elements upon the CO2
corrosion rate of low alloyed carbon steels. Paper 473, Corrosion 91, NACE. Houston,
TX.
Dugstad, A., Lunde, L., and Videm, K. 1994. Parametric Study of CO2 Corrosion of Carbon
Steel. Paper 14, Corrosion 94. NACE. Houston, TX. 1994.
Dugstad, A., Videm, K. 1990. “Kjeller Sweet Corrosion II- Final Report”, IFE report no.
IFE/KR/F-90/008, Insitute for Energy Technology. Kjeller, Norway.
Dunlop, A., Hassell, H, L., and Goddard, H, P. 1984. Fundamental Considerations in Sweet
Gas Well Corrosions, Advances in CO2 Corrosion 1. Ed. Hausler, R. H and Godard, H.
P. Paper 52, Corrosion 1, , NACE, Houston.
gas treating plants Parts 1&2. Hydrocarbon Processing.
Eastman, D. E. and Nathan, M. I. 1975. Phys. Today. 28, 44.
Edwards, T. J., Maurere, G., Newman, J. and Prausnitz, J. M. 1978. Vapor-Liquid Equilibria
in Multicomponent Aqueous Solutions of Voltaile Weak Electrolytes. AIChE J. 24, 966-
976.
El Miligy, A. A., Geana, D. and Lorenz, W. J. 1975. A Theoretical Treatment of the Kineics of
Iron Dissolution and Passivation. Electrochim. Acta. 20, 273- 281.
El Miligy, A. A., Hilbwert, F. and Lorenz, W. J. 1973. J. Electrochem. Soc. 120, 247-251.
Eichkorn, G., Lorenz, W. J., Albert, L. and Fischer, H. 1968. Electrochim. Acta 13 183.
Eriksrud, E., and Sontvedt, T. 1983. Effect of flow on CO2 Corrosion Rates in Real and
Synethetic Formation Waters. Proceedingss of the Advances in CO2 Corrosion 1,
Symposium on CO2. Corrosion 83. Hosuton, TX.
Felloni, L. 1968. The Effect of pH on the Electrochemical Behaviour of Iron in Hydrochloric
Acid. Corr. Sci. 8, 133- 148.
Felloni, L., Cozzi, A., and Cammarota, G. P. 1967. Optical Microscopy Studies of Anodic
Dissolution of Iron in Sulfuric and Hydrochloric Acid Solutions. Electrochim.
Metallorum. 2, 104.
Fisher, W. and Siedlarek. 1978. Werkstoffer und Korrosion, 28, 822.
Flitt, H. J. and Bockris, J. O’M. 1982. Effect of organic inhibitors on the ingress of H into
metals. Int. J. Hydrogen Energy. 7, 411.
Fluor. Available: http://www.fluor.com/econamine/Pages/default.aspx (Accessed 17th
January 2016).
Fochtman, E. G., Langion, W. M. and Howard, D. R. 1963. Continuous Corrosion
Measurements. Chem. Eng. 70, 140.
Foroulis, Z. A. 1966. Anodic Protection-Theoretical Considerations. J. Electrochem. Soc. 113,
532-536.
P a g e | 145
Frankenthal, R. P. 1971. Passivation of Iron in Borate Buffer Solution. Electrochimica. Acta.

16, 1845-1857.
Frankenthal, R. P. and Milner, P. C.1986. Hydrogen Evolution Kinetics on a High- Carbon
Steel and on Tin in Seawater. Corrosion 42, 51.
French, B. M. 1971. Stability Relations of Siderite (FeCO3) in the System Fe-CO Am. J. Sci.
271, 37-78.
Freiman, L. I. and Kolotykin, Y, M. 1965. Zasch. Met. 1, 161.
Froning, H. R. and Jones, J. H. 1958. Corrosion of Mild Steel in Aqueous Monoethanolamine.
Ind. Eng. Chem. 50, 1737-1738.
Frumkin, A. N. 1961. Advances in Electrochemistry and Electrochemical Engineering. Vol 1,
Interscience, Ne York. 65.
Gabrielsen, J., Michelsen, M. L., Stenby, E. H and Kontogeorgis, G. M. 2005. A Model for
Estimating CO2 Solubility in Aqueous Alkanolamines. Ind. Eng. Chem. Res. 44, 3348-
3354.
Galindo, A., Davies, L. A., Gil-Villegas, S. and Jackson, G. 1998. The Thermodynamic of
Mixtures and the Corresponding Mixing Rules in the SAFT-VR Approach for Potentials
of Variable Range, Molecular Physics, 93, 241-252.
Gerus, B. R. D. 1981. Detection and Mitigation of Weight Loss Corrosion in Sour Gas
Gathering Systems. In H2S Corrosion in Oil and Gas Production – A Compilation of
Classic Papers. NACE, Houston, TX. 888-903.
Ghosh, M. M., O’Connor, J. T. and Engelbrecht, R. S. J. San. Eng. ASCE, 1996. 92, 199-213.
Global CCS Institute. 2011. Global Status of Large-Scale Integrated CCS Projects- December
2011 Update. (Online) Australia. Available from:
http://www.globalccsinstitute.com/publications/global-status-large-scale-integrated-
ccs-projects-december-2011-update (Accessed 15th January 2016)
Global CCS Institute. 2011. The Global Status of CCS: 2011. (Online) Australia. Available
from: http://cdn.globalccsinstitute.com/sites/default/files/publications/22562/global-
status-ccs-2011.pdf (Accessed 15th January 2016)
Global CCS Institute. 2012. CO2 Capture Technologies: Post Combustion Capture (PCC).
(Online) Australia. Available from:
http://www.globalccsinstitute.com/publications/co2-capture-technologies-post-
combustion-capture-pcc (Accessed 17th January 2016)
Global Climate and Energy Project. 2005. An Assessment of Carbon Capture Technology and
Research Opportunities. Stanford University. (Online) Stanford. Available from:
https://gcep.stanford.edu/pdfs/assessments/carbon_capture_assessment.pdf
Goff, G. S. & Rochelle, G. T. 2004. Monoethanolamine degradation: O2 mass transfer effects
under CO2 capture conditions. Industrial & Engineering Chemistry Research, 43, 6400-
6408.
Gray, L. G. S., Anderson. B. G., Danysh, M. J., and Tremaine P. G. 1989. Mechanism of
Carbon Steel Corrosion in Brines Containing Dissolved Carbon Dioxide at pH 4. Paper
464, Corrosion 89. NACE. New Orleans, Louisiana.
P a g e | 146
Gray, L. G. S., Anderson. B. G., Danysh, M. J., and Tremaine P. G. 1990. Effect of pH and
Temperature on the Mechanism of Carbon Steel Corrosion by Aqueous Carbon Dioxide,
Paper 40, Corrosion 90. NACE. Las Vegas.
Gregory, L. B. and Scharmann, W. G. 1937. Carbon Dioxide Scrubbing by Amine Solutions.
Ind. Eng. Chem. Res. 29, 514-519.
Guo, X. P. and Tomoe, Y. 1999. The Effect of Corrosion Product Layers on the Anodic and
Cathodic Reactions of Carbon Steel in CO2- saturated MEA solutions at 100 degrees C.
Corrosion Science 41, 7, 1391-1402.
Hackerman, N. and Schmidt, H, R. 1949. Kinetics of the Corrosion Process in Condensate Gas
Wells. Ind. Eng. Chem. 41, 1712-1716.
Hall, W. D and Barron, J, G. 1981. Solving Gas Treating Problems - A Different Approach.
Proceedings of Gas Conditioning Conference, 31st Annual, University of Oklahoma, C1-
C13.
Hausler, R. H., and Goddard, H. P. 1984. ‘Fundamental aspects of CO2 corrosion’ in Advances
in CO2 Corrosion,1, page 52, NACE. Houston, TX.
Helgeson, D. C., Delany, J. M., Nesbitt, H. W. and Bird, D. K. 1978. Am. J. Sci. A, 278A, 1-
229.
Hem, J. D. 1967. Principles and Application of Water Chemistry. Ed S. D. Faust and J. V.
Hunter. John Wiley and Sons, Inc. New York.
Heisler, L., Weiss, I. H. Operating Experience at Aderklaa with Alkanolamine Gas Treating
Plants for Sour Natural Gas Sweetening. Proceedings of Gas Conditioning Conference,
25th Annual, University of Oklahoma. 1975, H1 – H22.
Hemmer, H., Kooyman, H., and Wilhelmsen, W. 1993. “Kjeller Sweet Corrosion I- Final
Report”, IFE report no. IFE/KR/F-93/091, Insitute for Energy Technology. Kjeller,
Norway.
Heusler, K. E. 1957. Doctoral Thesis, University of Gottingen.
Heusler, K. E. 1958. Elektrochem. 582, 62.
Heusler, K. E. 1964. Chem-Ing-Tech. 3, 589
Heusler, K. E. and Cartledge, G. H. 1961. The Influence of Iodide Iona and Carbon Monoxide
on the Anodic Dissolution of Active Iron. J. Electrochem. Soc. 108, 732- 740.
Heusler, K. E. and Huerra, D. Kinetics and Mechanisms of the Anodic Dissolution of Metallic
Glasses. 1989. Electrochem. Soc. 136, 65.
Heyrovsky, J. 1927. Recl. Trav. Chim. Pays-Ba, 46. 582.
Holland, T. J. B. and Powell, R. 1990. An Enlarged and Updated Internally Consistent
Thermodynamic Dataset eith Uncertainties and Correlations: The System K2O-Na2O-
CaO-MgO-MnO-FeO-Fe2O3-Al2O3-TiO2-SiO2-C-H2-O2. J Metam. Geol. 8, 89-124.
Hikita, H., Asai, S., Katsu,Y. And Ikuno, S. 1979. Absorption of Carbon Dioxide into Aqueous
Monoethanolamine Solutions. AlChE Journal. 25, 793-800.
Hilbert, F., Miyoshi, Y., Eichkorn, G. and Lorenz, W. J. 1971. Correlation between the kinetics
of electrolytic dissolution and deposition of iron. J Electrochem. Soc 118, 1919 - 1927
P a g e | 147
Hoar, T. P. 1961. ‘Corrosion Problems in Chile’ Anti- Corrosion Methods and Materials. 8,
12, 367-370.
Hoar, T. P. and Holliday, R. D. 1953. The Inhibition by Quinolines and Thioureas of the Acid
Dissolution of Mild Steel. 3, 502-513.
Hoar, T. P. and Hurlen, T. 1959. Proceedings of the VIIth Meeting of CITCE, Madrid, 1956,
Butterworths, London. 445.
Hoar, T. P. and Ferrer, L. The Relationships Between Anodic Passivity, Brightening and
Pitting. Corrosion Science 1. 1965. 49
Hobler, H. and Schonewolf, H. 1974. How to Handle Sour Natural Gas. Petro. Int. 14, 1, 10.
Hofmeyer, B. G. Scholten, H. G. Lloyd, W. G. 1965. Contamination and Corrosion in
Monoethalamine Gas Treating Solutions. The Dow Chemical Company.
Hurlen, T. 1962. Kinetics of Metal/ Metal-ion Electrodes: Iron, Copper, Zinc.Electrochim.
Acta, 7, 653.
Hurlen, T. 1963. Anodic Behaviour of Iron in Alkaline Solutions. Electrochim. Acta, 8, 609.
Hurlen, T. 1966. Reaction Models for Dissolution and Deposition of Solid Metals,
Electrochim, Acta. 11, 1205.
Ikeda, A., Ueda, M., and Mukai, S. 1984. CO2 Behaviour of Carbon and Cr Steels. Advances
in CO2 Corrosion 1. Ed. Hausler, R. H and Godard, H. P. NACE. Houston, Tx.
Isaacs 1980. Solubility of mixtures of hydrogen sulfide and carbon dioxide in a
monoethanolamine solution at low partial pressures. Journal of chemical & engineering
data, 25, 118-120.
Johnson, M. L., Tomson, M. B. 1991. Ferrous Carbonate Precipitation Kinetics and its Impact
on CO2 Corrosion, Paper 268, Corrosion 91. NACE. Houston, TX.
Jou, F. Y., Mather, A. E. and Otto, F. D. 1995. The solubility of CO2 in a 30 mass percent
monoethanolamine solution. Can. J. Chem. Eng., 73 (1), 140- 147.
Kabanov, B., Burnstein, R., and Frumkin, A. 1947. Kinetics of Electrode Processes on the Iron
Electrode. Disc. Faraday Soc. 1, 259-269.
Kaewsichan, L., Al-Boferseb, O., Yesavage, V. F. and Selim, M. S. 2001. Prediction of the
Solubility of Acids Gases in Monoethanolamine (MEA) and Methyldiethanolamine
(MDEA) Solutions Using the Electrolyte- UNIQUAC model. Fluid Phase Equilibria,
183-184, 159-171.
Keddam, M., Mattos, O. R. And Takenouti, H. 1981. Reaction Model for Iron Dissolution
Studied by Electrode Impedance. J. Electrochem. Soc. 128 257.
Kent, R. L. and Eisenberg, B. 1976. Better Data for Amine Treating. Hydro. Proc., 55, 87-90.
Kelley, K. K. and Anderson, C. T. 1935. Metal Carbonates: Correlations and Applications of
Thermodynamic Properties. Bulletin 384, in Contributuins to the Data on Theoretical
Metallurgy IV. 1-73; US Department of the Interior, Bureau of Mines.
Kelly, E. J.1961. J Electrochem. Soc., 122, 124.
Kelsall G.H., Francis P.A. and Thompson I. 1993. The redox chemistry of H2S oxidation by
the British Gas Stretford Process. IV. Thermodynamics of V-S-H2O Systems and the
P a g e | 148
Electrochemical Behaviour of V(V)/V(IV) couples in Alkaline Solutions,

J.Appl.Electrochem., 23(5), 417-426.
Kennard, M. L. and Melsen, A. 1985. Mehanisms and Kinetics of Diethanolamine Degradation.
Ind, Eng. Chem. 24, 129-140.
Kita, H. 1966. Period Variation of Exchange Current Density of Hydrogen Electrode Reaction
with Atomic Number and Reaction Mechanism. J Electrochem. Soc., 113, 1095.
Kittel, J., Fleury, E., Vuillemin, B., Gonzalez, S., Ropital, F. and Oltra, R. 2010. Corrosion in
alkanolamine used for acid gas removal: From natural gas processing to CO2 capture.
Materials and Corrosion.
Kobosew, H. And Nekrassow, P. 1972. Electrochim Acta. 17, 369
Kohl, A. & Nielsen, R. 1997. Gas Purification (5th edition). Houston: Gulf Publishing
Company. Houston, TX.
Kohl, A. & Riesenfeld, F. C. 1985. . Gas Purification (4th edition). Houston: Gulf Publishing
Kolotyrkin, Y. M., Lazorenko- Manevich, R, M. and Sokolova, L. A. 1987. Spectroscopic
Studies of Water Adsorption on Iron Group Metals. J.Electroanal. Chem. 228, 301-328.
Kosseim, A. J., McCullough, J. G., and Butwell, K. F. 1984. Corrosion-Inhibited Amine Guard
ST process, Chemical Engineeering Progress, 64-71.
Kladkaew, N., Idem, R., Tontiwachwuthikul, P., Saiwan, C. 2009a. Corrosion Behaviour of
Carbon Steel in the Monoethanolamine- H2O- CO2- O2- SO2 System. Ind. Eng. Chem.
Res. 48, 8913-8919.
Kladkaew, N., Idem, R., Tontiwachwuthikul, P., Saiwan, C. 2009b. Corrosion Behaviour of
Carbon Steel in the Monoethanolamine- H2O- CO2- O2- SO2 System: Products, Reaction
pathways and kinetics. Ind. Eng. Chem. Res. 48, 10169-10179.
Kostryukov, V. N. and Kalinkina, I, N. 1964. Zhurnal Fizicheskoi Khimii. 38, 780-781.
Kritpiphat, W. and Tontiwachwuthikul, P. 1996. New Modified Kent-Eisenberg Model for
Predicting Carbon Dioxide Solubility in Aqueous 2-Amino-2- Methyl-1-Propanol
(AMP) solutions. Chem. Eng. Commun., 144, 77-83.
Kuranov, G., Rumpf, B., Maurer, G. and Smirnova, N. 1997. VLE Modeling for Aqueous
Systems Containing Methyldiethanolamne , Carbon Dioxide and Hydrogen Sulfide.
Fluid Phase Equilibria. 136, 147-162.
Kuznetsov, V. A., Iofa, Z. A. and FizKhim, Zh. 1947. 31, 201.
Kvarekval. J. A Kinetic Model For Calcluating Profiles and Fluxes Across of CO2 related
species across the Nernst Diffusion Layer. Paper 5. NACE 97.
Lang, F. S. and Mason, J. F. 1958. Corrosion in Amine Gas Treating Solutions. Paper 2.
Corrosion 14, NACE. 65-68.
Langmuir, D. 1969. US Geol. Surv. Prof. Pap., 650B, 180-184.
Latimer, W. E. 1952. The Oxidation States of the Elements and their Potentials in Aqueous
Solutions. 2nd edition. Prentice Hall InC. Englewood Cliffs, NJ.
P a g e | 149
Legrand, L., Savoye, S., Chausse, A. and Messina, R. 2000. Study of Oxidation Products
Formed on Iron in Solutions Containing Bicarbonate/ Carbonate. Electrochim. Acta, 46.
111-117.
Lepaumier, H., Picq, D. & Carrette, P.-L. 2009a. New Amines for CO2 Capture. II. Oxidative
Degradation Mechanisms. Industrial & Engineering Chemistry Research, 48, 9068-
9075.
Lepaumier, H., Picq, D. & Carrette, P. L. 2009b. New Amines for CO2 Capture. I. Mechanisms
of Amine Degradation in the Presence of CO2. Industrial & Engineering Chemistry
Research, 48, 9061-9067.
Li, C. X and Furst, W. 2000. Representation of CO2 and H2S Solubility in Aqueous MDEA
Solution Using an Electrolyte Equation of State. Chem. Eng. Sci. 55, 2975-2988.
Liu, Y., Zhang, L. and Watanasiri, S. 1999. Representing Vapor-Liquid Equilibrium for an
Aqueous MEA-CO2 System Using the Electrolyte Nonrandom-Two-Liquid Model. Ind.
Eng. Chem. Res., 38, 2080-2090.
Litschewski, M. J. 1996. More Experiences with Corrosion and Fouling in a Refinery Amine
System. Paper 96391, Corrosion 96, NACE.
Lorenz, W. J. 1965. Corros. Sci. 5, 121
Lorenz, W. J. and Eichkorn, J. 1965. Discussion of ‘The Active Iron Electrode, I. Iron
Dissolution and Hydrogen Evolution Reactions in Acidic Sulfate Solutions’.
Electrochem. Soc. 112 1255.
Lorenz, W. J and Vilche, J. R. 1972. pH Effects on Iron Dissolution in The Active Region.
Corros. Sci. 12, 785.
MacNab, A. J. and Treseder, R. S. 1971. Materials Requirements for a Gas Treating Process.
Materials Performance. 10, 1 21-26.
Mason, J. W. and Dodge, B. F. 1936. Equilibrium Absorption of Carbon Dioxide by Solution
of Ethanolamines. AIChE J., 32, 27-48
McCafferty, E. and Hackerman, N. 1972. Double Layer Capacitance of Iron and Corrosion
Inhibition with Polymethylene Diamines. 119, 146-154.
McCullough, J. G. and Nielsen, R. B. 1996. Contanimation and Purfication of Alkaline Gas
Treating Solutions. Paper 96396, Corrosion 1996. NACE.
McHenry, H. I., Shives, T. R., Read, D. T., McColskey, D, T., Brady, C. H. and Purtscher, P.
T. 1986. Examination of a Pressure Vessel that Ruptured at the Chicago Refinery of the
Union Oil Company on July 23, 1984. National Bureau of Standards Report. 86, 3049.
Mitchell, R. 2008. Mitsubishi Heavy Industries’ carbon capture technology. Carbon Capture
Journal. (Online) 1, 2-5. Available from: https://www.mhi-
global.com/products/pdf/articles_05.pdf ( 5th February 2016).
Murata, T., Sato, E., and Matsuhashi, R. 1985. Factors Controlling Corrosion of Steels in CO2-
Saturated Environments. Advances in CO2 Corrosion. NACE. Houston, TX.
Nainar, M, and Veawab, A. 2009. Corrosion in CO2 Capture Unit Using MEA- Piperazine
blends. In: Gale, J., Herzog, H. and Braitsch, J. (eds) Greenhouse Gas Control
Technologies 9. Amsterdam: Elsevier Science Bv.
P a g e | 150
Nesic, S. and Lunde, L. 1993. CO2 Corrosion of Carbon Steel in Two Phase Flow. Corrosion
93. 49, 640
Nesic, J., Postlethwaite and S. Olsen. 1996. An Electrochemical Model for Prediction of
Corrosion of Mild Steel in Aqueous Carbon Dioxide Solutions. Paper 4. Corrosion 52.
280-294.
Nielsen, R. B., Lewis, K. R., McCullough, J. G. and Hansen, D. A. 1995. Corrosion in Refinery
Amine Systems. Paper 571, Corrosion 95. NACE.
Nyborg, R. 1998. Initiation and Growth of Mesa Corrosion Attack During CO2 Corrosion of
Carbon Steel. Paper 48. Corrosion 98. NACE, Houston, TX.
Nyborg, R., and Dugstad, A. 1998. Mesa Corrosion Attack in Carbon Steel and 0.5%
Chromium Steel, Corrosion 98, Paper 29, NACE. Houston, TX.
Ogundele, G. and White, W. 1986. Some Observations on Corrosion of Carbon Steek in
Aqueous Environments Containing Carbon Dioxide. NACE Corrosion 42, Paper 2.
Toronto, Canada.
Palacios, C. A. and Shadley, J. R. 1991. Characteristics of Corrosion Scales on Steels in CO2-
Saturated NaCl brine. Corrosion 47, 122-127.
Papp, H. and Pritchard, J. 1975.The adsorption of Xe and Co on a Cu (311) single crystal
surface. Surf. Sci. 53, 371.
Park, R. L. 1975. Inner-shell Spectroscopy. Phys. Today. 28, 52.
Park, J. Y. E., Yoon, S. J. & Lee, H. 2003. Effect of Steric Hindrance on Carbon Dioxide
Absorption into New Amine Solutions:? Thermodynamic and Spectroscopic Verification
through Solubility and NMR Analysis. Environmental science, 37, 1670-1675.
Park, S. H., Lee, K. B., Hyun, J. C. and Kim, S. H. 2002. Correlation and Prediction of the
Solubility of Carbon Dioxide in Aqueous Alkanolamine and Mixed Alkanolamine
Solutions. Ind. Eng. Chem. Res., 41, 1658-1665.
Pearce, B., and Dupart, M. S. 1987. Corrosion in Gas Conditioning Plants- An Overview. Paper
39, Corrosion 87. NACE.
Podesta, J. J. and Arvia, A. J. 1965. Kinetics of the Anodic Dissolution of Iron in Concentrated
Ionic Media: Galvanostatic and Potentiostatic Measurements. Electrochim. Acta. 10, 171
-182.
Polderman, L. D., Dillon, C. P. and Steele, A. B. 1955. Degradation of Methanolamine in
Natural Gas Treating Service. Oil Gas Journal 16, 5, 180-183.
Powerspan Clean Energy Technology. 2016. Available from:
nd
http://powerspan.com/technology/eco2-co2-capture/. Accessed 2 February 2016.
Preis, W. and Gamsjager, H. 2002. Phys. Chem. Chem. Phys. 4, 4014-4019.
Reiterer, F. Johannes, W. Ganjager, H. 1981. Mikrochim. Acta. 1, 63-72.
Riesenfeld, F. C. and Blohm, C. L. 1951. Corrosion Resistance of Alloys in Amine Gas
Treating Systems. Petroleum Refiner 30, 10, 107-115.
Robie, R. A. 1965. Heats and Free Energies of Formation of Troilite, Herzenbergite, Magnesite,
and Rhodochrosite Caalculated From Equilibrium Data. US Geol. SurveyProf., paper
525-D, 65-72.
P a g e | 151
Robie, R. A. and Hemingway, B. S. 1995. Thermodynamic Properties of Minerals and Related

Substances at 298.15 K and 1 Bar (105 Pascals) Pressure and at Higher Temperatures.
USGS B 2131, Government Printing Office, Washington, DC, USA.
Robie, R. A., Hemingway, B. S. and Fisher, J. R. 1978. Thermodynmic Properties of Minerals
and Related Substances at 298.15K and 1 bar (105 Pascals) pressure and at higher
tempertature. USGS B 2131, Government Prinitng Office, Washington, DC, USA.
Robie, R. A., Haselton, H, T. and Hemingway, B. S. 1984. Am Mineral., 69, 349-357.
Robie, R. A. and Waldbaum, D. R. Thermodynamic Properties of Minerals and Related
Substances at 298.15oK (25oC ) and one atmosphere (1.013 bars) pressure and at higher
temperatures. USGS B 1259, Government Printing Office, Washington, DC, USA, 1968.
Rochelle, G. T. 2009. Amine Scrubbing for CO2 Capture. Science, 325, 1652-1654.
Roiter, V. A., Yuza, V. A. Poluyan, E. S., FizKhim, Zh. 1939. 13 605.
Rooney, P. C. and Bacon, T. R. 2000. Corrosion in Alkanolamine Plants: Causes and
Minimization. Paper 00494, Corrossion 2000. NACE. Houston, TX.
Rooney, P. C., Bacon, T. R. and DuPart, M. S. 1996. Effect of Heat Stable Salts on Solution
Corrosivity of MDEA-based Alkanolamine Plants Hydrocarbon Processing 75, 3, 95-
99.
Rooney, P. C., Bacon, T. R. and DuPart, M. S. 1997. Effect of Heat Stable Salts on Solution
Corrosivity of MDEA-based Alkanolamine Plants. Part II. Hydrocarbon Processing 76,
4, 65-69.
Rooney, P. C., DuPart, M. S. and Bacon, T. R. 1998. Oxygen’s Role in Alkanolamine
Degradation. Hydrocarbon Processing. 77, 109-113.
Sada, E., Kumuzawa, H. and Butt, M. A. 1976. Gas Absorption with Consecuive Chemical
Reactions: Absorption of Carbon Dioxide into Aqueous Amine Solutions. The Candaian
Journal of Chemical Engineering. 54, 421- 424.
Schmeal, W. R., MacNab, A. J., and Rhodes, P. R. 1978. Corrosion in Amine/ Sour Gas
Treating Contractors. Chem. Eng. Progr. Vol 74, 3, 37-42.
Schmidt, K. A. G., Maham, Y. and Mather, A. E. 2007. Use of The NRTL Equation for
Simultaneous Correlation of Vapour-Liquid Equilibria and Excess Enthalpy Applications
to Aqueous Alkanolamine Systems. Journal of Thermal Analysis and Calorimetry, 89
(1), 61-72.
Schmitt, G. 1983. Fundamental Aspects of CO2 Corrosions. Paper 43, Corrosion 83. Houston,
TX: NACE.
Schmitt, G., and Rothman, B. 1977. Studies on the Corrosion Mechanism of Unalloyed Steel
in Oxygen-fee Carbon Dioxide Solutions. Part I. Kinetics of the liberation of hydrogen,
Werkstoffe Korros. 28, 816.
Schmitt, G., and Rothman, B. 1977. Studies on the Corrosion Mechanism of Unalloyed Steel
in Oxygen-fee Carbon Dioxide Solutions. Part II. Kinetics of iron disssolution.
Werkstoffe Korros. 29, 98.
Schmitt, G., and Rothman, B. 1978. Corrosion of unalloyed and low-alloyed steels in carbonic
acid solutions. Werkstoffe Korros. 29, 237.
Schwabe, K. and Voigt, C. 1969. Electrochim. Acta. 14 853-869.
P a g e | 152
Schwenk, W. 1974. Corrosion of Unalloyed Steel in Oxygen-free Carbonic Acid

Solutions.Werkstoffe und Korros. 25, 643.
Schweickert, H., Lorenz, W. J. and Friedburg, H. 1980. Discussion of Impedance Measurments
of the Anodic Iron Dissolution. J. Electrochem. Soc. 127. 8. 1693-1701
Shao, R. & Stangeland, A. 2009. Amines Used in CO2 Capture – Health and Environmental
Impacts, Bellona. 3-26.
Shemenski, R. M., Beck, F. H., and Fontana, M. G. 1965. Dissolution Kinetics and Polarization
of Iron Whiskers, Corrosion 21. 39-47.
Simard, S., Drogowska, M. and Menard, H. 1996. Electrochemical Behaviour of 1024 Mild
Steel in Slightly Alkaline Bicarbonate Solutions. Journal of Applied Electrochemistry.
27, 317-324.
Simard, S. and Menard, H. 1997. Electrooxidation of 1024 Mild Steel in Slightly Alkaline
Phosphate and Bicarbonate Solutions. 1998. Journal of Applied Electrochemistry. 28,
593-600.
Singer, P. C. and Stumm, P. C. 1970.Solubility of Ferrous Iron in Carbonate Bearing Waters.
J. Am. Water Works Assoc. 62, 198-202.
Smith, R. F. and Younger, A. H. 1972. Operating Experiences of Canadian Diethanolamine
Plants. Proceedings of Gas Condition Conference, 22nd Annual, p 1-17.
Soosaipakasam, I. R. & Veawab, A. 2008. Corrosion and polarization behavior of carbon steel
in MEA-based CO2 capture process. International Journal of Greenhouse Gas Control,
2, 553-562.
Stern, M, 1955. The Electrochemical Behavior, Including Hydrogen Overvoltage, of Iron in
Acid Environments. J. Electrochem, Soc. 102, 609.
Strazisar, B. R., Anderson, R. R. & White, C. M. 2002. Degradation of monoethanolamine used
in CO2 capture from flue gas of a coal-fired electric power generating station. Abstracts
of Papers of the American Chemical Society, 223, 54-56.
Strazisar, B. R., Anderson, R. R. & White, C. M. 2003. Degradation pathways for
monoethanolamine in a CO2 capture facility. Energy & Fuels, 17, 1034-1039.
Stubina, N. M. and Toguri, J. M. 1989. Trans. ISS, 10, 87-90.
Sun, Y., Remias, J. E., Neathery, J. K., and Liu, K. 2011. Electrochemical Study of Corrosion
Behaviour of Carbon Steel A106 and Stainless Steel 304 in Aqueous Monoethanolamine.
Corrosion Engineering, Science and Technology. 46, 724-731.
Tafel, J. 1905. Z. Physik. Chem,. 50, 641.
Tanthapanichakoon, W., Veawab, A. and McGarvey, B. 2006. Electrochemical Investigation
on the Effect of Heat-Stable Salts on Corrosion in CO2 Capture Plants Using Aqueous
Solution of MEA. Industrial & Engineering Chemistry Research, 45, 2586-2593.
Teevens, P. J. 1990. Toward a Better Understanding of the Cracking Behaviour of Carbon Steel
in Alkanolamine Sour Gas Sweetenining Units: Its Detection, Monitoringm and How to
Avoid It. Paper 198, Corrosion 90. NACE.
Tontiwachwuthikul, P., Meisen, A. and Lim, C. J. (1991) Solubility of CO2 in 2- Amino-2-
methyl-1-propanol solutions. J. Chem. Eng. Data, 36(1), 130-133.
P a g e | 153
Ueda, M., and Ikeda, A. 1996. Effect of microstructure and Cr content in steel on CO2
corrosion, Paper 13, Corrosion. NACE.
Valentini, C. R., Moina, C. A., Vilche, J. R. and Arvia, A. J. 1985. The Electrochemical
Behaviour of Iorn in Stagnant and Stirred Potassium Carbonate- Bicarbonate Solutions
in the 0-75oC Temperature Range. Corrosion Science 25, 985-997.
van Hunnik, E. W. J., Pots, B. F. M., and Hendriksen, E. L. J. A. 1996. The formation of
protective corrosion product layers in CO2 corrosion, Paper 6, Corrosion 96, NACE.
Houston, TX.
Vatankhah, G., Drogowska, M., Menard, H. and Brossard, L. 1998. Electrodissolution of Iron
in Sodium Sulfate and Sodium Bicarbonate Solutions at pH 8. Journal of Applied
Electrochemistry. 28, 173-183.
Vazquez, R. C., Rios, G., Trejo, A., Rincon, R. E., Uruchurtu, J., and Malo, J. M. 2000. The
Effect of Diethanolamine Solution Concentration in the Corrosion of Steel. Paper 696
NACE.
Veawab, A. & Aroonwilas, A. 2002. Identification of oxidizing agents in aqueous amine-CO2
systems using a mechanistic corrosion model. Corrosion Science, 44, 967-987.
Veawab, A., Tontiwachwuthikul, P. and Bhole, S. D. 1997. Studies of corrosion and corrosion
control in a CO2-2-amino-2-methyl-1-propanol (AMP) environment. Industrial &
Engineering Chemistry Research, 36, 264-269.
Veawab, A., Tontiwachwuthikul, P. and Chakma, A. 1999. Corrosion Behaviour of Carbon
Steel in the CO2 Absorption Process Using Aqueous Amine Solutions. Industrial &
Veldman, R. 2000. Alkanolamine Solution Corrosion Mechanisms and Inhibition From Heat
Stable Salts and CO2. Corossion 2000- NACE International.
Videm, K., and Dugstad, A. 1987. Effect of Flowrate, pH, Fe2+, Concentration and Steel
Quality on the CO2 corrosion of Carbon Steels. Paper 42, Corrosion 87. Houston, TX:
NACE.
Videm, K., and Dugstad, A. 1989a. Corrosion of Carbon Steel in an Aqueous Carbon Dioxide
Environment Part 1: Film Formation. Materials Performance. 28, 63-67
Videm, K., and Dugstad, A. 1989b. Corrosion of Carbon Steel in an Aqueous Carbon Dioxide
Environment Part 2: Solution effects. . Materials Performance. 28, 46.
Videm, K., and Dugstad, A. 1990. Film Covered Corrosion, Film Breakdown and Pitting
Attack of Carbon Steels in Aqueous CO2 Environemnents. Paper 186. Corrosion 88.
NACE. Houston, TX.
Videm, K., Kvarekvaal, T., Perez, T., and Fitzdimons, G. 1996. Surface Effects on the
Electrochemistry of Iron and Carbon Steel Electrodes in Aqueous CO2 Solutions, Paper,
1, Corrosion 96, NACE. Houston, TX.
Voight, C. 1968. Electrochim. Acta. 13, 2037-2050.
Volmer, T. and Erdey-Gruz. 1930. Phys. Chem. Chem. Abt. A, 150, 203.
Vrachnos, A., Kontogeorgis, G. and Voutsas, E. 2006. Thermodynamic modeling of acidic gas
solubility in aqueous solution of MEA, MDEA and MEA-MDEA blends. Ind. Eng.
Chem. Res., 45, 5148-5154.
P a g e | 154
Vuppu, A. K. and Jepson, W. P. 1994. Study of Sweet Corrosion in Horizontal Multiphase,

Carbon Steel Pipelines. Offshore technology Conference. 7494, 265-276.
Wagman, D. D., Evans, W. H., Parker, V. B., Schumm, R. H., Halow, I., Bailey, S. M.,
Schumm, R. H. 1969. Selected Values of Chemical Thermodynamic Properties, Tables
for Elements 35 through 53 in the Standard Order of Arrangement. National Bureau of
Standards Technical Note 270-4, US Government Printing Office, Washington, DC,
USA.
Weber, R. E. and Peria, W. T.1967. Use of LEED Aparatus for the Detection and Identification
of Surface Contaminants. J Appl. Phy. 38, 4355.
Weiland, R. H., Chakravarty, T. and Mather, A. E. (1993) Solubility of carbon dioxide and
hydrogen sulfide in aqueous alkanolamines. Ind. Eng. Chem. Res., 32(7), 1419- 1430.
Wieckowski, A., Ghali, E., Szklarczyk, M., and Sobkowski. J. 1983. The Behaviour of Iron
Electrode in CO2 Saturated Neutral Electrolyte-1. Electrochim Acta. 28, 1619.
Wilcox, D. E. and Bromley, L. A. J. Indust. Eng. Chem. 1963, 55, 32-9.
Yeh, J. T. & Pennline, H. W. 2001. Study of CO2 Absorption and Desorption in a Packed
Column. Energy, 15, 274-278.
P a g e | 155
4. Chapter 4
Experimental Methods and Materials

Kinetic studies were conducted to identify the effect of temperature, CO2 loading, pH, amine
type, amine concentration and oxygen concentration on iron corrosion in aqueous amine
solutions. Based on the collected data, corrosion mechanisms and rates, in principle, may be
determined for every set of experimental conditions. The experiments were conducted with the
expectation that a trend in corrosion rates would be identified, which would enable future
prediction as functions of these process conditions. In order to be able to deduce trends, it was
deemed important to obtain a large body of data covering a wide range of conditions typical of
amine scrubbing processes. Due to the substantial amount of experimental data required, the
experimental techniques were selected based on their ability to acquire useful data in the most
effective and efficient way. Hence, techniques such as weight loss measurement were
excluded.
This chapter details the experimental methods and materials used to obtain micro-kinetic data
for the following purposes:

P a g e | 156
i. To obtain the oxidation and reduction kinetics of iron in aqueous MEA-H2O-CO2
systems as functions of temperature (25-80oC), CO2 loading (0-0.6 mol CO2 (mol
amine)-1), pH (8.10-12.55), MEA concentration (5-60 wt%) and oxygen concentration.
ii. To determine and assign the oxidation and reduction reaction mechanisms of iron in
aqueous MEA-H2O- CO2 systems.
iii. Determination of corrosion rates via the Tafel extrapolation method as functions of the
experimental variables listed above.
iv. As inputs for the development of a mechanistic model for the uniform corrosion of iron
in CO2-loaded MEA systems
The mechanism for iron corrosion was characterized further with the use of a quartz crystal
microbalance, the mass detection device discriminating between the dissolution reaction
occurring in CO2-loaded MEA solutions from the predominantly passivation reactions
occurring in unloaded solutions.
This chapter also details the mechanical design, configuration and operation of an
electrochemical reactor developed for the measurement of dissolved iron concentrations and
electrode currents as functions of time, under typical CO2 absorption conditions. During the
constant potential measurements, samples were taken from the anolyte compartment at fixed
time intervals to be analysed subsequently by Inductively Coupled Plasma Spectroscopy
Optical Emission Spectroscopy (ICP-OES). The dissolved FeII concentrations determined by
ICP-OES enabled partial current densities leading to iron dissolution to be deconvoluted from
measured current densities.

P a g e | 157
4.1. Materials
This section details the main chemicals and materials used within the course of this project, the
purity and the suppliers from which they were obtained.
4.1.1. Chemicals
The CO2 absorption solvents tested within this project include the benchmark amine
monoethanolamine (MEA), methyldiethanolamine (MDEA), 2-amino-2-methyl-1-propanol
(AMP) and aminoethylpiperazine (AEP). A ferrous sulfate (FeSO4) bath was used for iron
plating in the Electrochemical Quartz Crystal Microbalance measurements described in
Section 4.4.4 below. In the absence of CO2, pH adjustments to the amine solution were made
with 1 M sulphuric acid (H2SO4). A solution of 1 M Na2SO4 was used as the supporting
electrolyte for all the electrochemical measurements. The reagent grade and sources of all the
chemicals used are given in Table 4-1.
Table 4-1: Common reagents
Reagent Reagent Grade Supplier

Monoethanolamine AnalaR, ≥ 99.0 % Sigma Aldrich
Methyldiethanolamine AnalaR, ≥ 99.0 % Sigma Aldrich
2-Amino-2-methyl-1-propanol AnalaR, ≥ 99.0 % Sigma Aldrich
Aminoethylpiperazine AnalaR, ≥ 99.0 % Sigma Aldrich
Iron(II) Sulfate Heptahydrate AnalaR, ≥ 99.0 % Sigma Aldrich
Sulfuric Acid AnalaR, ≥ 99.0 % Sigma Aldrich
Sodium Sulfate AnalaR, ≥ 99.0 % Fisher Scientific
4.1.2. Standard reagents
The standard reagents listed in Table 4-2 were used for calibration of the Inductively Coupled
Plasma Spectroscopy Optical Emission Spectrometer (ICP-OES) prior to commencing
analyses of dissolved metal concentrations.

P a g e | 158
Table 4-2: Standard reagents
Reagent Reagent Grade Supplier

Iron Standard solution for ICP 1000 mg dm-3 Fe in Sigma Aldrich
Phosphorus, Potassium, Silica, Aluminum, 1000 mg dm-3 Fe in SPEX
2% HNO3
Antimony, Arsenic, Barium, Beryllium, Boron, 5% HNO3 / Tartaric CertiPrep
Calcium, Chromium, Cobalt, Copper, Iron, Acid / HF
Lead, Lithium, Magnesium, Manganese,
Molybdenum, Nickel, Selenium, Silver,
Sodium, Strontium, Thallium, Tin, Vandium,
Zinc Standard for ICP
4.1.3. Electrode Materials
Where required, the electrodes were cut into the desired size to suit the experimental procedure.
Although carbon steel is the most commonly used material for pipe work and equipment in
CO2 absorption plants, pure iron was used as the working electrode for most of the experiments
reported in this thesis to elucidate its electrochemical behaviour in aqueous amine-CO2
solutions before considering the additional complexity of alloying elements. Iron cylindrical
discs of diameter 6.4 mm (0.322 cm2) and a 3.125 cm2 platinum flag was used as the working
and counter electrodes for the rotating disc electrode experiments detailed in Section 4.4.2.1.
The working electrode for the Electrochemical Quartz Crystal microbalance experiments
described in Section 4.4.4 was made partially in-house by electroplating iron on Au-TiO2
coated Quartz Crystals supplied by Metrohm Autolab B.V., Utrecht (NL). For the flow reactor
experiments in Section 4.5.1, three additional types of working electrode materials were tested
in addition to pure Fe, namely carbon steel 1020 (Carbon Steel S355) , Carbon Steel 1038
P a g e | 159
(Carbon Steel EN8) and stainless steel AISI 304. The working electrodes used each had a
surface area of 400 cm2 and a thickness of 10 mm, whilst the platinised titanium electrode of
equal surface area had a thickness of 1 mm. The electrode materials, their chemical
composition and suppliers are listed in Table 4-3.
Table 4-3: Electrode Materials
Materials Material Grade Supplier

Iron Slug ≥ 99.95 Alfa Aesar
Platinum foil 99.85% Goodfellow
Au coated, AT-Cut Quartz Crystals Not specified Metrohm
Iron Sheet 99.8% Goodfellow
Carbon Steel S355 Iron ≥ 97.6% West Yorkshire Steel
Carbon ≤ 0.20%
Manganese ≤ 1.60%
Silicon ≤ 0.55%
Phosphorous ≤ 0.025%
Sulfur ≤ 0.025%
Carbon Steel EN8 Iron ≥ 98.18% West Yorkshire Steel
Carbon: 0.35 - 0.45%
Manganese: 0.60 - 1.00%
Silicon: 0.05 - 0.35%
Phosphorous: 0.06%
Sulfur: 0.06%
Stainless Steel Iron ≥ 66.35% West Yorkshire Steel
Carbon ≤ 0.08%
Chromium ≤ 18 - 20%
Manganese ≤ 2%
Silicon ≤ 1.00%
Phosphorous ≤ 0.045%
Sulfur ≤ 0.030%
Nickel ≤ 8 - 10.5%
Platinised Titanium Sheet Not specified Magneto Special
Anodes B.V.
P a g e | 160
4.1.4. Cation Exchange Membrane
A Nafion membrane (Ion Power Inc., New Castle, DE, (USA)) was used in the electrochemical
flow reactor described in Section 4.5. The chemically stable Nafion membrane is a
perfluorosulfonic acid based polymer, which allows only the passage of cations due to the
negative charge of the acid groups grafted onto the polymer backbone. The membranes were
cut into the required size and conditioned by soaking in hot water at 80oC for 1 hour, to ensure
maximum swelling. After the conditioning process, the membrane when not in use was soaked
in solution to maintain the efficiency and performance.
Table 4-4: Cation exchange membrane
Membrane Membrane Type Supplier

Nafion 424 Perfluorinated acid based Ion Power
4.1.5. Gases
The gas supply set included 99.8% CO2, compressed air (21.5% ± 0.5% O2 and 78.5% Ni) and
99.998% N2 (The BOC Group plc., London, UK).
4.2. Electrode and Solution Preparation
4.2.1. Electrodes
Prior to conducting the experiments, the exposed surface of the embedded iron disc in the
rotating disc electrode described in section 4.4.2.1 was polished mechanically until it attained
a mirror finish with the use of an Alpha 2-speed grinder-polisher polishing machine (Buehler
Ltd., Lake Bluff, IL, (USA)). The Fe disk surface was polished on 600 grit CarbiMet® paper
disks with 300 nm followed by 50 nm particle diameter aluminum oxide powder (Sigma
Aldrich Co., Suffolk (UK)), wetted in semi-pure water. The electrode surface was then cleaned
P a g e | 161
by thoroughly rinsing in ultra pure water. The water was purified (PURELAB Ultra, ELGA
LabWater Ltd., UK) by incorporating a series of water purification technologies namely
reverse osmosis followed by ion exchange and adsorption on a bed of activated carbon
particles.
Following the process of polishing and cleaning, the electrode was immersed in an ultrasonic
bath (Thermo Scientific) for 5 minutes to remove any adhering aluminum oxide particles
before commencing the electrochemical procedure.
The pure Fe, carbon steel 1020 (Carbon Steel S355), Carbon Steel 1038 (Carbon Steel EN8)
and stainless steel AISI 304 electrodes used for the flow reactor in Section 4.5.2 were prepared
by wet grinding on 600 grit CarbiMet® paper and degreasing with acetone as suggested by
(Bockris et al., 2000).
4.2.2. Aqueous Amine Solutions
The aqueous amine solutions were classified into three different categories:
i. Unloaded amine solutions containing no CO2.
ii. CO2-loaded amine solutions containing desired amounts of CO2.
iii. Fully saturated amine solutions containing the maximum amount of dissolved CO2.
The unloaded amine solutions were prepared by simply diluting the pure amine solution with
de-ionized water to the desired mass concentration. A 30 wt% solution of MEA was selected
as the benchmark for the experiments as it represents the standard operating conditions for the
CO2 absorption process. For a loaded solution, the desired CO2 loadings (αCO2) were achieved
by varying the relative CO2 / air or CO2 / N2 gas flow rates with the use of Platon NGX Glass
Variable Area Flowmeters (Roxspur Measurement & Control Ltd, Sheffield, UK). The flow
rates were set according to the CO2 solubility data in 30 wt% MEA obtained by (Tong, Trusler
et al. 2012) and (Jou, Mather et al. 1995), included in Appendix I. As an example, according
P a g e | 162
to Jou et al, a CO2 partial pressure 8.96 Pa would be required to achieve a CO2 loading of 0.203
mol CO2 (mol MEA)-1 at 40oC. Fully saturated amine solutions were obtained by maintaining
a constant CO2 flow until the pH of the solution remained unchanged for a period of 30 minutes.
4.3. Instrumentation for Electrochemical Measurements
The electrochemical measurements reported in this thesis were carried out principally with
Metrohm PGSTAT 30 Autolab potentistat/galvanostat controlled using the Autolab NOVA
software (Metrohm Autolab B.V., Utrecht (NL)). The chemicals were weighed for solution
preparation with the use of a Mettler-Toledo AG245 balance supplied by Mettler-Toledo Ltd.,
Leicester, (UK). Following the weighing procedure, the solutions were prepared by diluting to
the desired concentration with the use of de-ionised water. A solution of 1 M Na2SO4 was used
as the base electrolyte for all measurements. Silver / silver chloride (AgCl|Ag) electrodes
supplied by Cole-Parmer Instrument Co., London (UK)) and Metrohm were used as reference
electrodes for all experiments. All potentials reported in this work were measured against the
AgCl|Ag electrode. Throughout the experiments, the pH of the amine solutions was monitored
with a pH meter (Hanna Instruments Ltd., (UK)). To ensure accuracy in measurements, the pH
meters were calibrated regularly by a three-point calibration with pH 4.0, 7.0 and 10.0 buffer
solutions.
4.4. Experimental Methods for Micro-kinetic studies on Iron
4.4.1. Cyclic Voltammetry
Cyclic voltammetry is a commonly-used electrochemical technique with which useful
information about the system can be obtained quickly and with a high level of precision. The
technique entails measurement of the current response under potentiodynamic conditions

P a g e | 163
imposed by a potentiostat whereby the potential of the working electrode is cycled linearly
with time at a scan rate (v / mVs-1) expressed by equation (4. 1).
! " = ! 0 ± &" (4. 1)
Whilst in linear voltammetry, the potential is swept in only one direction between two specified
potential limits, in cyclic voltammetry it can be swept backwards and forwards between the
two limits over an appropriate number of cycles, depending on the system being studied. The
negative-going potential sweep induces the reduction of the species that may have been
oxidized previously in the positive-going sweep. The scan limits were set based on
thermodynamic predictions with the use of potential-pH diagrams for Fe-H2O and Fe-H2O-
CO2 systems. The potential cycling across the boundaries between different Fe species / phases
results in cyclic voltammograms (CV) of current against potential which provides insight into
the nature of the oxidation and reaction occurring at the electrode surface, the potentials at
which they occur and the amount of charge and monolayers involved in such processes, and
whether adsorbed products accumulated on successive cycling.
4.4.2. Kinetic Studies Using Rotating Disc Electrodes
Levich (Albery and Hitchman, 1971; White et al., 1984) solved the hydrodynamics of
horizontal, downward-facing rotating discs rotating about a vertical axis in an ‘infinite’ liquid
(i.e. no wall effects) behaves like a pump with axial flow to the centre of the disc and radial
flow across it, resulting in mass transport rates to its surface that are independent of radial
position. The requirement for zero liquid velocity at the disc/solution interface, gives rise to a
hydrodynamic boundary layer and corresponding diffusion layer, with a thickness δN remains
stationary relative to the electrode. As electro-active species must be transported to the
electrode | solution interface from the bulk solution, the maximum attainable current is limited
by the rate of mass transport. As discussed in Chapter 2, if the electrode overpotential is

P a g e | 164
increased, surface concentrations of reacting species B decrease progressively to zero, at which
a transport limited current iL is reached:
*+ ,-[/]1234
'() (4. 2)
56
Where DB is the diffusion coefficient of species B, and δN is the Nernst diffusion layer
thickness.
Side view
Side view Flow
FlowPattern
Pattern
Electrode
Electrode DiscDisc
rotation
rotation
End view
End view
Disc
Disc Electrode
Electrode
Insulating
InsulatingMaterial
Material
Stainless steel
Stainless Steel Conductor
Figure 4-1: Schematic of the side and end view of the RDE and the flow pattern induced by
rotation of an RDE system
A rotating disc electrode is constructed by embedding a small cylindrical rod of the desired
electrode material centrally into an insulating sheath such as Teflon. The top of the sheath is
attached to a rotator whilst the bottom in horizontal alignment with the disc surface is immersed
into the electrochemical cell that contains the solution of interest. As the rotating disk electrode
starts to rotate about the vertical axis, the rotation spins the solution out from the surface of the
P a g e | 165
cylinder in a radial direction, effectively drawing fresh material from the bulk towards the
surface of the disk. The rotation creates a solution pattern as shown in Figure 4-1.
For the purposes of simplification, the experiments should be designed for laminar flow
conditions only. During laminar flow conditions, the solution flow is characterized by smooth
movement in layers along streamlines as described in Figure 4-2 unlike the random and chaotic
motion in turbulent flow. Prediction of the nature of solution flow can be readily obtained with
the dimensionless parameter, the Reynolds number, defined for the RDE system by:
:;<
78 = (4. 3)
=
Where rc is the radius of the cylinder, f is the angular rotation speed (Hz) and υ is the solution
viscosity. If Re does not exceed about 1 × 105, the solution flow induced by rotation is laminar
(Fisher, 1996) and has the pattern shown schematically in Figure 4-1.
>( = 1.555BC DEFG/I & JK/L MN K/G (4. 4)
Equation (4. 4) is the Levich equation predicting the transport limiting current (IL) generated
with a rotating disc electrode, and is a function of the rotation frequency, f, the kinematic
viscosity of the solution, υ, the diffusion coefficient of the electrolyte species, D, and the
electron stoichiometry, υe of the electrode reaction. Based on the Levich equation, for a mass
transport controlled reaction, a plot of IL against the square root of the rotation rate should result
in a straight line passing through the origin.
In order to understand the increase in limiting current density with rotation rate, it is important
to recall that the velocity of the solution as it approaches the electrode surface increases with
rotation speed leading to improved convective mass transport of the electro-active species and
a simultaneous decrease in the diffusion layer thickness, as defined by equation (4. 5):
P a g e | 166
K
OP = FK/I & K/L N JK/G (4. 5)
K.QQR
i.e. increasing rotation rates (f) decreases δN, so enabling higher diffusion limited currents.
In general, there are several advantages associated with the use of rotating disc electrodes such
as the ease of fabrication and surface preparation. As discussed, the flow conditions are well
defined and a laminar flow regime can be attained over an extensive range of rotation rates for
small electrodes. In addition, the mass transport characteristics are established and can be easily
varied by altering the rotation rate. However, there are also a few disadvantages associated
with the use of rotating disc electrodes. The accumulation of bubbles at the electrode surface
resulting from the hydrogen evolution reaction is principally relevant in acidic solutions.
Bubbles are particularly adherent to certain insulating materials such as PTFE and produces a
significant and detectable amount of noise in the measured currents, thus indicating disruption
due to hydrogen bubble.
4.4.2.1. Experimental procedure
The rotating disc electrode was fabricated in-house by embedding a cylindrical, irremovable
Fe disc of 99.95% purity and 6.4 mm diameter in a cylindrical Teflon sheath. The rotating disc
was connected to a modulated speed electrode rotator, Model E6 (Pine Instrument Company,
Raleigh, NC, (USA)). The connection between the rotating disc electrode and the rotator was
established via a conducting stainless steel cylindrical sheath.
The experiments were conducted in a standard three compartment glass cell fabricated by
Cambridge Glassblowing Ltd (UK). A schematic of electrochemical cell and RDE system is
shown in Figure 4-2.

P a g e | 167
A
A
B A-A –Water
Waterbath
bath
B G B-B -Reference
Referenceelectrode
C-electrode
Rotating disc electrode
D
D D-C –Counter
Rotatingelectrode
disc
E-electrode
Glass Luggin probe
F-D -Glass frit electrode
Counter
G- Aqueous solution
C E- Glass Luggin
C
probe
F – Glass frit
E
E G – Aqueous solution
G F
F
G
Figure 4-2: Rotating disc electrode system.
The electrochemical cell was immersed in a temperature controlled water bath fitted with a
o
temperature controller to maintain constant temperature within ± 0.1 C throughout
experiments. A silver-silver chloride electrode and a platinum flag, custom made by Cambridge
Glassblowing Ltd, were used as reference and counter electrodes, respectively. For
experiments conducted under anoxic conditions, the cell was bubbled with nitrogen gas for 20
minutes before each experiment to minimise dissolved oxygen concentrations. Before the start
of each experiment, trapped bubbles within the Luggin probe were removed by tilting the cell
until it appeared to be free of bubbles. Prior to conducting the experiments, the exposed surface
of the embedded iron disk was prepared according to the procedure described in Section 4.2.
All kinetic measurements included were conducted at 6 different rotation rates i.e. 0 rpm, 60
rpm, 240 rpm, 540 rpm, 960 rpm and 1500 rpm, altering mass transport rates for studying
diffusion-limited phenomena. The electrochemical measurements were conducted with the use
of the computer controlled Metrohm PGSTAT 30 Autolab potentistat/galvanostats described
in section 4.3.
P a g e | 168
During the course of experiments, special precautionary measures were taken such as daily
maintenance on the solution level in the AgCl|Ag reference electrodes and pH meter and
adequate storage of the probes in their respective storage fluids to prevent crystallization.
Whilst cyclic voltammetry was the primary electrochemical technique of choice for
characterizing the behaviour of iron in aqueous amine solutions with the use of rotating disk
system, electrochemical parameters such as the scan rate, rotation rate and positive potential
scan limit were used to probe the kinetics and mechanisms of iron in aqueous MEA as functions
temperature (25-80oC), CO2 loading (0-0.6 mol CO2 (mol amine)-1), pH (8.10-12.55), MEA
concentration (5-60 wt%) and oxygen concentration. Electrode potential-pH and activity-pH
diagrams of iron-water–CO2 systems were used to assist with reaction assignments. Analysis
of the resulting kinetic data via the Tafel slope extrapolation method described in section 4.4.3
enabled corrosion rates to be predicted as functions of the experimental variables.
4.4.2.2. Repeatability and Reproducibility
At the start of all the electrochemical measurements reported in this thesis, after the mechanical
polishing and cleaning procedure, the iron electrode was also subjected to a potential pre-
treatment at -1.4 V for 120 s to reduce any oxides that may have formed spontaneously in air.
The polishing method and electrochemical cathodic reduction method chosen for the iron
electrode resulted in very good repeatability and reproducibility of the experimental results
obtained.
Due to the extensive number of experiments, each experiment was also repeated at least twice
especially if the experimental data resulted in an outlier from a trend.
4.4.3. Tafel Slopes Extrapolation Method for Corrosion Rate Determination
The Tafel line extrapolation method, which involves plotting ln(current density) against
potential is a method used for estimating the corrosion rate from experimental current-potential
P a g e | 169
data. At the corrosion potential (Ecorr), the anodic and cathodic rates are exactly equal as such,
the measured current density is zero. All the electrons released by the oxidation of the metal
during the dissolution reaction are consumed by the reduction reaction. At significant
overpotentials, in the potential region where the rate of a single partial corrosion reaction is
predominant, the log current-potential plot for reactions under pure activation control result in
straight lines called Tafel lines. A relation analogous to the Butler-Volmer equation describes
the experimental current density potential curve.
YZ , JY< ,
S = STU;; exp ] − exp ] (4. 6)
[\ [\
Thus when E >> Ecorr, the Tafel equation is given by:
YZ ,
S_ = STU;; exp ] (4. 7)
[\
Thus when E << Ecorr, the Tafel equation is given by:
JY< ,
−ST = STU;; exp ] (4. 8)
[\
According to the Tafel equation (4. 6), the value of the anodic and cathodic current intersection
is the corrosion current density jcorr. Suggesting that jcorr can be determined by extrapolating
the linear segments of the anodic and cathodic Tafel region back to their intersection, at the
corrosion potential (Ecorr). The Tafel equation holds true for overpotentials within the range of
± 250 mV from the Ecorr. At larger ‘corrosion overpotentials’ (over 250 mV) from Ecorr, the
measured current is a reflection of the kinetics for either the anodic or cathodic corrosion
reactions. Figure 4-3 shows an example of a Tafel plot with extrapolation of the anodic and
cathodic curves back to the corrosion potential Ecorr.

P a g e | 170
By assuming uniform metal dissolution across the surface of the electrode, Faraday’s law can
be used to calculate the corrosion rate from the current density with equations (4. 9) and (4.
10).
jFe ⋅ M Fe
rFe / m s-1 = (4. 9)
ν e F ⋅ ρ Fe
jFe ⋅ M Fe
rFe / m a -1 = 3.154 × 107 (s a -1 ) (4. 10)
ν e F ⋅ ρ Fe
Where rFe is the corrosion rate in mm a-1, jFe is the corrosion current density in mA cm-2, MFe
is the molar mass of iron and ρ is its density in g cm-3.
lnln|j||j|
Cathodic branch
Cathodic branch Anodic
Anodicbranch
branch
lnlnjcj= (αaFη)
c = (α
/ RT
cFη)/ RT lnlnjaja==(α(αaFη)/
a
Fη) RT
/ RT
jcorrjcorr
← EE(negative)
(negative) E (positive) →
E (positive)
EEcorr
corr
Figure 4-3: Tafel plot for the determination of corrosion current density and hence rate.
Whilst the Tafel extrapolation method has received general acceptance for the graphical
determination of corrosion rates, the method can lead to various values (Badea et al., 2010),
since inappropriate selection of the Tafel slopes can change corrosion rates by a factor of up to
P a g e | 171
10 (Badea et al., 2010). Thus, care was taken when applying the Tafel extrapolation, at least
one of the branches if not both of the polarization curves exhibited Tafel behaviour, i.e.
linearity on the semi-logrithmic scale over several decades of current density. In addition, the
extrapolation was set to begin within 50-100 mV from the corrosion potential.
4.4.4. Electrochemical Quartz Crystal Microbalance
The Electrochemical Quartz Crystal Microbalance (EQCM) was developed to monitor
frequency and hence mass changes in situ and effectively in real time; hence, it has found wide
acceptance for the study of interfacial phenomena. The EQCM technique provides the
capability to measure nanogram mass changes to the electrode, which can then be correlated
to the amount of charge passed during electrochemical measurements. The EQCM consists of
a thin piezoelectric quartz crystal sandwiched between two metal electrodes as schematically
shown in Figure 4-4 (a). As the crystal is a piezoelectric material, it undergoes a slight
vibrational deformation in the presence of an electric field. The application of an alternating
electric field across the crystal through the two metal electrodes causes mechanical oscillation
of the crystal at its resonant frequency, fo, which is sensitive to changes in the mass on the
surface of the crystal. One side of the crystal can then be used as the working electrode from
which current transients and changes to the mass of the electrode may be measured as functions
of potential (Buttry & Ward 1992) which makes this a suitable method for the verification of
the dissolution and (or) passivation of iron.
For most Quartz Crystal Microbalance (QCM) applications, the AT-cut quartz crystals are the
most commonly used. The quartz crystals are cut so that the thickness shear mode (TSM) of
vibration dominates and is most responsive to changes in mass. The AT-cut means that the
crystals, which are typically a few tenths of a mm in thickness are manufactured by slicing
through a quart rod at a 35o15´ angle from the crystalline axis. The AT-cut is also particularly
P a g e | 172
suitable as it provides stable oscillation with reproducible resonant frequencies, which are
somewhat insensitive to temperature changes near room temperature. The vibrational motion
of the quartz crystal produces an acoustic wave, which extends through the thickness of the
crystal as shown schematically in Figure 4-4 (b).
Deposit
AT-Cut Quartz
Crystal
Electrode
Quartz Crystal
Electrode
Electrode
Figure 4-4: Acoustic waves propagating through a quartz crystal, electrodes and deposit
The thickness of the crystal, Wq, is half of the wavelength of the acoustic wave, λ:
b
à = (4. 11)
G
The resonant frequency, fo, can be calculated equation (4. 12) and is expressed as function of
velocity of shear wave, vtr and crystal thickness, Wq and the wavelength of the standing wave,
λ.
cde cde
NU = = (4. 12)
b Gfg
The velocity (vtr) of the wave is the speed of sound through a material expressed in equation
(4. 13), where µq and ρq are the shear modulus and density of the quartz crystal:
P a g e | 173
ide
&h; = (4. 13)
jde
Thus,
ide K
NU = (4. 14)
jde Gfg
In the thickness shear mode, both the underlying crystal and any mass adhering to the surface
will oscillate at the same lateral displacement and frequency. Hence, an assumption is made
for small changes (≤ 2% of the quartz crystal mass), the addition of mass from external deposits
can be treated as a mass change of the quartz crystal. If the crystal and the mass bound to the
surface oscillate with no energy loss, then the process is said to be elastic. Thus, the expression
for the fo in equation (4. 15) takes into account the increase in the thickness of the crystal.
∆: ∆f G:l ∆f
= = (4. 15)
:l fg ig /jg
Thus,
G:l m ∆f
∆N = − (4. 16)
ig /jg
The negative sign results from the fact that the acoustic wavelength increases with increasing
thickness of the quartz, resulting in a decrease in the resonant frequency of the quartz crystal.
The change in thickness can be related to the change in mass by taking into account the density
of the deposited mass, ρq, and the area of the electrode, A.
∆n
∆` = (4. 17)
-jg
Substituting equation (4. 17) into equation (4. 16) results in:
P a g e | 174
G:l m ∆n
∆N = − (4. 18)
- ig jg
Thus,
-∆: ig jg
∆o = (4. 19)
G:l m
Equation (4. 18), results in the well-established linear Sauerbray equation, which relates with
nanogram sensitivity, the Δm to the shift in the resonant frequency of the quartz crystal.
pq
∆N = −∆o (4. 20)
-
Where Cf is the sensitivity constant and is expressed in units of ng cm-2 Hz-1.
G:l m
r: = (4. 21)
ig jg
The sensitivity constant of the crystal is proportional to the square of fo, so higher sensitivity
can be achieved by selecting crystals with higher resonant frequencies. The resonant
frequencies of crystals used within EQCM techniques are typically in the range of 5-20 MHz
(Marx 2003). The typical values used for the electrode area of an AT-quartz, A is 0.196 cm2,
density, ρq is 2.648g cm-3 and the sheer modulus, µq is 2.947 × 1011 g cm-1 s-2 (Buttry and Ward
1992).
Whilst some QCM applications have been designed to measure Δm by incorporating a mass
amplification feature into the signal to be transduced by the crystal, it is preferable that the
measured quantity during experiments is the frequency change and not the mass change.
Although a change in mass is the most obvious factor affecting the resonant frequency, it may
be subject to other obscuring effects, which may result in misinterpretation of the results. For
example, the frequency to mass relationship based on the Sauerbray equation (4. 20) is valid
only for mass changes of less than 2% of the original crystal mass, as the mass deposited on
P a g e | 175
the crystal moves synchronously with the resonance of the quartz crystal. Excessive mass
loading may result in the outer region of the deposit exhibiting a phase delay with respect to
the resonator, resulting in changes to the sensitivity of the crystal and thus inaccurate and
incomprehensible frequency changes.
4.4.4.1. Experimental Procedure
The EQCM apparatus (Metrohm Autolab B.V., Utrecht (NL)) consisted of an EQCM cell
fabricated from polypropylene, an EQCM oscillator, a spring-mounted Au/TiO2 coated quartz
crystal, an Au wire counter electrode and a AgCl|Ag reference electrode. The quartz crystal
possessed a resonant frequency, fo, of 6 MHz, the frequency resolution was 0.07 Hz, with a 1
Hz accuracy of frequency measurements. The sensitivity coefficient of Cf was specified as
0.0815 Hz cm2 ng-1 and a maximum measurable frequency change was set as 80 kHz. The
working electrode AEQCM was 0.35 cm2 and the maximum volume of the EQCM cell was
specified as 3 cm3.
The EQCM cell was connected to a computer controlled Metrohm PGSTAT 30 Autolab
potentiostat/ galvanostat in conjunction with a Metrohm - Autolab EQCM module. During the
electrochemical measurements, one side of the AT-cut quartz crystal was exposed to the
aqueous solution whilst the opposing side remained dry, in air.
The first step was to verify the stability of the frequency response with time and study the noise
levels obtained on the Au electrode, by measuring the current and frequency response to the
potential scan in the potential region 0 V ≤ E ≤ +0.3 V during which no significant frequency
change was expected or detected.
As there are no commercially available Fe electrodes compatible with the EQCM cell supplied
by Metrohm, the Fe electrodes were produced by electroplating the gold electrode in a 300 g
dm-3 ferrous sulphate solution at pH 3.2. This was achieved in two steps:
P a g e | 176
1) Firstly by conducting a potential scan in the potential region -1.4 V ≤ E ≤ 0.4 V to
ascertain the values of the electrode potentials at which Fe electrodeposition occurred
in the ferrous sulphate solution.
2) The Au electrode was electroplated with Fe by applying a potential of -1.0 V for 200 s,
during which the frequency response was expected to decrease with time. The total
mass of Fe deposited on the electrode surface was evaluated from the frequency
response. The total charge passed calculated by integrating the area under the current-
time graph, was used to evaluate the change in Fe mass that would have been detected
with a charge yield of unity. The ratio between the detected mass and the expected
mass (based on a charge yield of unity) was used to evaluate the experimental charge
yield for iron electrodeposition.
On completion of the deposition process, the FeSO4 solution in the EQCM cell was emptied
and swiftly replaced with the aqueous amine solution used as the electrolyte solution. Cyclic
voltammetry experiments were conducted on the Fe electroplated electrodes to determine the
effects of CO2 loading, pH and MEA concentration on the kinetics and mechanisms of iron. In
conjunction with predictions from potential-pH diagrams, the change in the mass on the
electrode was evaluated with the Sauerbray equation, allowing deconvolution of the Fe partial
oxidation current densities leading to dissolution of iron and (or) the adsorption of products on
the electrode surface.
Unfortunately, due to the high cost of the EQCM apparatus, incompatibility of the aqueous
ethanolamine solution with polypropylene (the fabrication material for the EQCM cell) and the
consumable nature of the crystals, this technique could be used only to obtain the kinetics of
the iron dissolution and passivation reactions for a limited number of experiments.
Furthermore, the design of the EQCM apparatus was suitable only for measurements under
P a g e | 177
quiescent conditions and could not be modified to investigate the effects of other operating
parameters such as temperature.
4.4.4.2. Special precautionary measures
A Faraday cage was used to enclose the EQCM equipment in order to minimize the electrical
noise from surrounding equipment. Each experiment was repeated at least three times to ensure
reproducibility. Prior to the start of voltammetric measurements in amine solutions, the
potential of the Fe electrode was held at -1.0 V for 50 s to reduce any oxides that may have
formed spontaneously during the transfer of electrolyte solutions from FeSO4 to MEA. Due to
the small size of the EQCM cell, maintaining a constant flow of CO2 to the solution throughout
the experiment was not a procedure that could be conducted feasibly, so was not carried out.
However, the pH of the test solution at the start and ending of the experiment were measured
and found unchanged, so it can be assumed that there was no change in the CO2 loading of the
solution due to the short timescale of the experiment. Before the start of every experiment,
about 15 minutes was given to allow the EQCM driving frequency to stabilize until the
frequency shift per minute was less than 1 Hz. In order to reduce measurement errors, the
crystals were kept clean and free of any extraneous materials by observing precautionary
procedures such as using gloves and tweezers for handling, storing the crystals in the original
packaging and the use of high purity water for cleaning.
4.5. Electrochemical Flow Reactor for Time-dependent Analysis of Iron
Corrosion in Flow Reactor Mode
4.5.1. Reactor design, configuration and materials
In order to measure dissolved iron concentrations in the aqueous amine solution as functions
of the operating parameters such as electrode potential, solution composition and flow rate,
P a g e | 178
temperature and CO2 loading, a parallel plate flow reactor was designed and fabricated in-
house principally from PTFE (PAR Group Ltd., Lancashire, UK). PTFE was selected to
withstand the highly chemically aggressive and corrosive nature of MEA at pH 8.01 and the
other amine solutions tested within this project.
Support plate
Platinised Titanium counter electrode

Catholyte outlet
NAFION permeable membrane
Anolyte outlet Reference electrode compartment
Fe working electrode
Support Plate
Catholyte inlet
Anolyte inlet
Figure 4-5: Schematic diagram of membrane-divided electrochemical flow reactor.
Figure 4-5 shows a schematic diagram of the electrochemical flow reactor, consisting of a
rectangular PTFE reactor body made of two separate compartments, each holding a rectangular
plate electrode held in place by a set of G-clamps and sealed with custom made PTFE
encapsulated Viton o-rings (PAR Group Ltd., Lancashire, UK). Unless otherwise specified, the
anode was an iron plate with a purity of 99.8% (Goodfellow Cambridge Ltd., Huntingdon, UK
), 400 mm in length, 100 mm in width with a thickness of 10 mm with a geometric surface area
of 0.04 m2 on each side. The cathode consisted of a 2.5 µm platinised titanium sheet (Magneto
Chemie, Holland, The Netherlands) also with the same dimensions.

P a g e | 179
+ -
Fe anode Platinised titanium cathode
Reference electrode
Luggin electrode
Cation-permeable
membrane
Anolyte Outlet Catholyte Outlet
F Flowmeter F
Anolyte Catholyte
Reservoir Reservoir
Figure 4-6: Schematic diagram of the experimental set-up of the electrochemical flow reactor
Once the reactor was assembled, the electrode area for both the anode and cathode exposed to
the test solution was 0.01 m2 and the inter-electrode gap was 50 mm. Both the anolyte and
catholyte were both operated in batch recycle mode with separate reservoirs. The volume of
the anode and cathode compartments of the reactor were identical at 0.25 dm3 and the total
volume of the anolyte and catholyte solutions was 4 dm3. As shown schematically in Figure 4-
6, 6.35 mm diameter tubular PTFE (Cole-Parmer Instrument Co., London, UK) were used to
connect the reactor, reservoirs and a Masterflex Precision peristaltic pump (Cole-Parmer
Instrument Co., London (UK)), to circulate the amine solutions from their reservoirs through
the bottom of the anode and cathode chamber, out from the top and back into the reservoirs.
P a g e | 180
At the inlet (bottom) of the anolyte and catholyte compartments, flow distributors shown
schematically in Figure 4-5 were utilised to enable an even distribution of the aqueous solution
flow rate. The solution flow rate was set and varied by the pumps but still measured by a Titan
810-VOP-O-PTM micro-turbine flow meter (Titan Enterprises Ltd., Dorset, UK) integrated
into the flow circuit and the data was logged by the data acquisition device NI DAQPad-6015
and the LabView software supplied by National Instruments. An AgCl|Ag reference electrode
was placed in the reference electrode compartment with a hole of 1 mm drilled in the body of
the anolyte compartment and emerging close to the surface of the iron electrode, to ensure
liquid connection between the reference electrode and the electrolyte solution in the reactor.
4.5.2. Experiment Procedure
4.5.2.1. Electrochemical Technique - Chronoamperometry
The flow reactor shown schematically in Figure 4-7 was designed to characterize the kinetics
of iron in aqueous amine solutions as functions of electrode potential, amine concentration,
amine type, temperature, flow rate and CO2 loading. The electrode potential in the flow reactor
was controlled with respect to the AgCl|Ag reference electrode by the computer-controlled
Metrohm PGSTAT 30 Autolab potentiostat/galvanostat (connected to a 10A current booster)
and the current was recorded as a function of time. The series of constant potential
measurements were conducted in the flow reactor, whilst recording the current with time, so
that the charge passed could be determined by integration. Inductively Coupled Plasma
Spectroscopy Optical Emission Spectroscopy (ICP-OES; Perkin Elmer Optima DV 2000,
Perkin Elmer Waltham, MA, USA) was used to determine the concentrations of dissolved FeII
in 1 cm3 samples taken from the reactor outlet enabling discrimination between partial current
densities leading to dissolution from those resulting in adsorbed products.

P a g e | 181
In addition to pure Fe, measurements were also conducted on Carbon Steel 355 (carbon steel
1020), Carbon steel EN8 (carbon steel 1038) and stainless steel AISI 304 plates with the same
dimensions as the Fe anode.
Throughout the entire duration of the experiment, the pH of anolyte and catholyte was
monitored to ensure constant CO2 loading. During the experiments, solution samples were
taken from the top of the anolyte compartment and the samples were taken at 10 minute
intervals for all the experiments conducted, with the exception of the longer term measurements
during which samples were taken once day over a period of 28 days. Immediately after
obtaining the samples, they were diluted in 1 M H2SO4 at a ratio of 1:10 before the dissolved
metal concentration was analysed by ICP-OES.
(a) (b) (c) (d)
Figure 4-7: (a) The assembled electrochemical flow reactor in action (b) The anolyte
compartment (c) The catholyte compartment (d) A Fe plate (anode)
4.5.2.2. Charge Yield Determination
The charge yield Φe is the ratio of the theoretical charge that would be required for the amount
of dissolved iron obtained (QFeII) to the actual total charge passed (Q) through the reactor
during the time t.

P a g e | 182
tuvww
ΦC = (4. 22)
t
x,C ww for the amount of FeII dissolved in the amine solution was calculated from the
stoichiometry of the half-cell reaction based on Faraday’s law:
nuvww ѵv , {uvww ѵv ,
x,C ww = = (4. 23)
ѵuvww z ѵuvww
where mFeII is the mass and nFeII is the number of mols of the FeII formed during the experiment,
M is the Fe molar mass, F is the Faraday constant, νe is the stoichiometric number (2) of
electrons for the reaction producing FeII.
The total amount of charged passed through the reactor was calculated from the current passed
during the time (t) via equation (4. 24).
x = > |" (4. 24)
4.5.2.3. Analytical Technique - Inductively Coupled Plasma Spectroscopy
Optical Emission Spectroscopy
The ICP-OES used for this thesis was supplied from (Perkin Elmer Optima DV 2000, Perkin
Elmer Waltham, MA, USA). The ICP-OES technique involves the application of a high
frequency and highly energetic magnetic field to a stream of argon to excite the gas and create
a plasma. A peristaltic pump is used to pump a liquid sample into the system, which was
nebulised and sprayed into the plasma flame leading to the excitation of the sample. This
resulted in an emission spectrum showing peaks at characteristic wavelengths indicative of the
solution composition. The concentration of the dissolved metals in the solution samples at these
wavelengths was determined by the calibration of emission intensity.
As discussed in 4.5.2.1, the aliquots obtained from the flow reactor outlet were diluted by a
factor of 10 in 1 M H2SO4 to prevent the dissolved Fe2+ ions from oxidising in air and forming
P a g e | 183
Fe(OH)3 / Fe2O3 particles forming that otherwise may have blocked the nebuliser. Prior to
commencing the sample measurement, a four-point calibration was carried out using diluted
ICP standard solutions (Table 4-2) which represented concentrations below and above the
expected dissolved metal ion concentrations in the samples. For repeatability purposes, each
sample was measured three times and the arithmetic mean was calculated.
4.5.2.4. Reactor Characterization
The (mean) mass transport rate coefficient km of the flow-through reactor was characterized
using the transport controlled reduction of a known concentration of hexacyanoferrate(III) ions,
a model fast reaction, as a function of solution flow rate and at a constant potential of -0.8 V
(AgCl|Ag):
3− 4−
Fe (CN )6 + e− ⎯⎯
→ Fe (CN )6 (4. 25)
For this purpose, 2 dm3 of solution containing 50 mol m-3 K3Fe(CN)6, 100 mol m-3 K4Fe(CN)6
and a supporting electrolyte of Na2CO3 at pH 11.3 was circulated through the reactor at various
solution linear velocities between 0.1 to 0.8 mm s-1. The charge passed due to the reduction of
hexacyanoferrate(III) to (II) was measured and subsequently used to evaluate km via the
equation (4. 26):
x = "}n EM&C D (4. 26)

P a g e | 184
4.6. References
Albery, W. J. and Hitchamn, M. L. 1971. Rotating Ring-Disc Electrodes. Clarendon Press,

Oxford.
Badea, G, E., Caraban, A., Sebesan, M., Dzitac, S., Cret, P., Setel, A. (2010). “Polarisation
measurements used for corrosion rates determination.” Journal of Sustainable Energy,
1: 121-127.
Bockris, J., Reddy, A. & Gamboa-Aldeco, M. (2000). Modern Electrochemisty, Vol.2A:
Fundamentals of electrodics, 2nd edn, Kluwer Academic / Plenum Publishers, New
York.
Buttry, D. & Ward, M. (1992), “Measurement of interfacial processes at electrode surfaces
with the electrochemical quartz crystal microbalance”, Chem. Rev. 92, 1355.
Christensen, P. & Hamnett, A. (1994). Techniques and Mechanism in Electrochemistry,
Blackie Academic and Professional, Glasgow, UK.
Fisher, A.C. 1996, Electrode Dynamics, Oxford University Press, New York.
Jou, F.-Y., et al. (1995). "Solubility of CO2 in a 30 mass percent monoethanolamine solution."
Canadian Journal of Chemical Engineering 73(1): 140-147.
Marx, K. A. (2003). "Quartz Crystal Microbalance: A Useful Tool for Studying Thin Polymer
Films and Complex Biomolecular Systems at the Solution−Surface Interface."
Biomacromolecules 4(5): 1099-1120.
Tong, D., et al. (2012). "Solubility of carbon dioxide in aqueous solution of monoethanolamine
or 2-amino-2-methyl-1-propanol: Experimental measurements and modelling."
International Journal of Greenhouse Gas Control 6: 37-47.
White, R. E., Bockris, J. O'M., Conway, B. E. and Yeager, E. 1984. Experimental Methods in
Electrochemistry, Comprehensive Treatise of Electrochemistry, Vol. 8. Plenum, New
York.
P a g e | 185
Chapter 5
Electrochemical Study of Iron

Electrodes in MEA-H2O-CO2 Systems
The corrosion and passivation of iron and carbon steel in aqueous CO2-containing
environments has been researched for decades, resulting in numerous publications reporting a
variety of corrosion rates and various mechanisms, as discussed in the Literature Review,
Chapter 3. Despite the large amount of data reported from extensive investigations, the
mechanistic interpretation of the processes involved in iron corrosion remains poorly
understood.
Due to the use of alkanolamines for the post-combustion capture of dilute CO2 from flue gas
streams being heralded as one of the most promising CO2 capture technologies currently being
developed, there have been several recent attempts at establishing the corrosion mechanism of
carbon steel in alkanolamine-H2O-CO2 systems, notably by (Veawab et al., 1999), (Kittel,
2010) and (Kladkaew et al., 2009a). These experimentally-based investigations aimed at
establishing the impact of crucial operating parameters, such as amine concentration,
temperature, CO2 loading, and oxygen concentration. While there have been some systematic
P a g e | 186
trends governing the corrosion behaviour of these chemical solvents that have been identified,
such as the detrimental effect of increasing CO2 loading and temperature, the results of these
studies have also resulted in conflicting and contradictory information, particularly about the
effects of amine and oxygen concentrations. In addition, a consensus has yet to be reached on
the actual mechanism of corrosion in these amine solutions. The models proposed depended
on the amine type and absence/ presence of oxygen. The previous studies on corrosion in
amines have been based solely on electrode polarization measurements of carbon steel samples
in quiescent solutions; the influence of the hydrodynamic conditions has been completely
neglected.
In order to assign the corrosion mechanism(s), as well as firmly establish and quantify the
effects of the critical operating conditions, the electrochemical behaviour of rotating iron disc
electrodes in 5-60wt% MEA-H2O-CO2 aqueous solutions, in the pH range of 8.01 to 12.55 and
from 18oC to 80oC, was investigated using cyclic and linear scan voltammetry, as described in
Chapter 4. In addition, some measurements were made of iron electrodeposited on gold
electrodes of an electrochemical quartz crystal microbalance (EQCM).
This section reports the experimental results obtained from a vast number of experiments over
a wide range of amine concentrations, temperatures, CO2 loadings, O2 presence, as well as
electrode rotation and potential scan rates and positive potential limits. The kinetic analysis of
the resulting data was also conducted and reported in terms of corrosion rates as functions of
the process parameters studied. The details of the electrochemical results such as the open
circuit potential (OCP), corrosion potential (Ecorr) and corrosion current (jcorr) for each of the
experiments conducted are listed in Appendix (V).
This Chapter is dedicated entirely to establishing the corrosion behaviour of iron in (aqueous)
MEA solutions, as it is the most extensively used alkanolamine and acts as the benchmark
solvent for the CO2 capture process. Comparisons with the corrosion behaviour in other
P a g e | 187
commercially-available solvents for the process will be discussed in Chapter 7. The majority
of voltammetric experiments for which results are reported below were conducted in an
aqueous solution of 30 wt% MEA at 25oC and full CO2 saturation (pH 8.01); this particular
MEA concentration was selected as it corresponds to the industry standard operating condition,
above which corrosion rates are believed to increase dramatically.

P a g e | 188
5. Electrochemical Behaviour of Iron in MEA-H2O-CO2 Systems
5.1. Fe in MEA+ H2O (absence of CO2)
To understand the corrosion behaviour of iron in aqueous MEA solutions, firstly an aqueous
solution of MEA was studied in the absence of CO2. A stationary Fe rotating disc electrode
immersed in 30wt% MEA solution of pH 12.55 attained an open circuit potential (OCP) of
+0.43 V. The subsequent cyclic voltammogram for a potential scan rate (dE/dt) of 10 mV s-1
is shown in Figure 5-1. On a positive-going potential scan from the negative potential limit of
-1.4 V, current densities initially decreased exponentially, then at potentials > ca. -1.03 V
(indicated by the red dashed line in Figure 5-1), net reduction reactions (resulting in negative
current densities by convention) changed to net oxidation reactions (positive current densities).
This region of oxidation reaction also marked the start of passivation extending over 1.6 V
(from -1.03 V to +0.6 V at the end of which is demarcated by a blue dashed line in Figure 5-
1). The potential-pH diagram for Fe-H2O systems (Figure 5-5) was used to assign reactions to
the main features of Figure 5-1 by approximating equilibrium potentials with potentials of
(partial) current onsets, a discussion of which follows.
At pH 12.55, the equilibrium potential of water reduction E"# /"# % is -0.92 V, so reduction
current densities at more negative potentials (region A) were assigned to the evolution of
hydrogen by the direct reduction of water via reaction (5. 1).
2() * + 2, - → () +2*(- (5. 1)
For electrode potentials more negative than ca. -1.03 V, net reduction occurred, after which
there was an onset of net oxidation (region B) due to the oxidation of elemental Fe by reaction
(5. 2). For a very small electrode potential window, it is possible that this was the only reaction
operating.
P a g e | 189
/, + 2() * → /,(*()) + 2(2 + 2, - (5. 2)
80
EFe(OH)2/Fe(OH)3 = -0.70 V EH2O/O2 = +0.29 V
60
EFe3O4/ Fe2O3 = -0.75 V
Current Density / A m-2
40
EFe(OH)2/ Fe3O4 = -1.01 V
20 EFe/Fe(OH)2 = -1.03 V
-20
-40
EH2/H2O = -0.94 V
-60
A B C
-80
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Electrode Potential (AgCl|Ag) / V
Figure 5-1: Cyclic voltammogram of a stationary iron RDE in 30 wt% MEA, at 10 mV s-1
(T = 25oC, pH 12.55).
However, starting at potentials less negative than -1.01 V, a secondary reaction (5. 3) also
occurred, predicted according to Figure 5-5, to result in the formation of magnetite (Fe3O4).
3/,(*()) → /,4 *5 + 2() * + 2(2 + 2, - (5. 3)
As the positive-going potential scan continued, the formation of hematite (Fe2O3) and ferric
hydroxide (Fe(OH)3) also became thermodynamically favourable reactions. Under these
circumstances, there was a very slow increase in current density, but because of the slow
potential scan rate, there was no distinctive current density peak due to the formation of the
thermodynamically stable III phase hematite (Fe2O3) and metastable ferric hydroxide
(Fe(OH)3). On increasing the electrode potential to values positive of the equilibrium potential
for the formation of Fe(OH)3, ca. -0.70 V, passivation of the iron electrode continued with
P a g e | 190
anodic current densities remaining essentially constant at 3.2 A m-2. At electrode potentials >
ca. +0.60 V, current densities started to increase exponentially (region C) due to evolution of
oxygen from water, for which the equilibrium potential, 67# 8/8# is +0.29 V at pH 12.22. After
reversing the potential scan from the positive potential limit of +1.0 V, no reduction current
density peak was evident at slow potential scan rates.
2/,4 *5 + () * → 3/,) *4 + 2(2 + 2, - (5. 4)
/,(*()) + () * → /,(*()4 + (2 + , - (5. 5)
Oxidation and reduction charge densities passed within the potential range of -1.0V and +0.6V
in the voltammogram were respectively 418 C m-2 for oxidation (Qox), corresponding to ca.
38.5 monolayers of Fe assuming a perfectly planar Fe | solution interface, and 114 C m-2 for
reduction (Qred). The Qox / Qred ratio of ca. 4 reflected the irreversibility of the reactions
involved. In particular, this highlights the charge transfer controlled nature of this redox
process, i.e. it reveals the extent to which the species oxidised on the positive-going scan were
reduced in the subsequent negative-going scan. The initial investigation demonstrated the result
for a stationary electrode at a single temperature. The roles of rotation speed and effect of
temperature were studied subsequently.
Passivation current densities appeared to be independent of the rotation rate, implying that the
passivation process is not controlled by diffusion but by the underlying charge transfer process.
The dependence of current densities on rotation rate was evident only in the hydrogen evolution
region at highly negative potentials (E ≤ -1.0 V) and the oxygen evolution region at highly
positive potentials (E ≥ +0.6 V). In both cases, current densities increased with rotation speed
(Figure 5-2 - 5-4), which also shifted the onset of oxidation to more positive potentials, as
evident in Figure 5-2 (25 oC) in which oxidation current densities occurred at > -0.32 V for
1500 rpm as opposed to -1.03 V at 0 rpm.

P a g e | 191
At low temperatures (25 oC) for a rotating disc electrode in 30 wt% MEA solution, there was
no evidence of a current density peak during the positive-going potential scan for all rotation
rates used in this work (0-1500 rpm). This demonstrated that aqueous MEA solutions are not
inherently corrosive in the absence of CO2. However, for solution temperatures of 40 oC and
60 oC, there were suggestions of active–passive behaviour with a slightly visible current density
peak at ca. -0.10 V for higher rotation rates (>240 rpm). Increasing temperature had no notable
effect on the magnitudes of passive current densities.
50
25oC
40
30
20
10
0
-10
0rpm
-20
60rpm
-30 240rpm
540rpm
-40 960rpm
A B 1500rpm C
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-2: Influence of electrode rotation rate on voltammogram of iron RDE in 30 wt% MEA,
at 10 mV s-1 (T = 25 °C, pH 12.55)
P a g e | 192
50
40oC
40
30
20
10
0
-10 0 rpm
-20 60rpm
240rpm
-30 540rpm
960rpm
-40 1500rpm
A B C
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-3: Influence of rotation rate on the voltammogram obtained on an iron RDE in 30 wt%
MEA, at 10 mV s-1 (T = 40 °C, pH 12.27)
50
60oC
40
30
20
10
0
-10
0rpm
-20 60rpm
240rpm
-30 540rpm
960rpm
-40 1500rpm
A B C
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-4: Influence of electrode rotation rate on voltammograms of iron RDE in 30 wt%
MEA, at 10 mV s-1 (T = 60oC, pH 12.10)
P a g e | 193
2.2
H3FeO4+
H2FeO4
2.0 HFeO4-
1.8
1.6 FeO42-
1.4
1.2
1.0
Fe3+ Fe3+
0.8
O2
0.6
Fe2O3 (c)
0.4
H 2O
0.2 Fe(OH)3 (ppt)
Fe(OH)4-
0.0
H+ Fe2+
-0.2 H2 Fe(OH)4-
-0.4
Fe(OH)2
-0.6 Fe3O4
-0.8 Fe(OH)2 Fe(OH) 2-
-1.0
Fe 4
-1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Figure 5-5: Stable (black) and meta-stable (red) potential-pH diagrams for Fe-H2O system for
dissolved iron activity of 10-6, 298 K and 0.1 MPa (Kelsall 2011); red lines correspond to meta-
stable behaviour of iron hydroxides rather than the more stable oxides.
P a g e | 194
5.2. Fe in MEA + H2O + CO2
Iron in an aqueous solution of MEA at pH 12.55 showed predominantly passive behaviour,
whereas addition of CO2 decreased the pH to 8.01 at which active iron (II) dissolution and
hence spontaneous corrosion is thermodynamically favourable, as predicted by Figure 5-5. This
change in pH can be explained by considering the main reactions of CO2 absorption in
alkanolamines. MEA (RNH2) reacts with water to form protonated amine and hydroxide ions
evidenced by the high pH measured in the absence of CO2. CO2 is absorbed by MEA and reacts
to form carbamate ions (RNHCOO-), protonated amine (RNH3+) and hydrogen carbonate ions
(HCO3-). The RNH3+ and HCO3- ions in turn may dissociate and produce hydrogen ions (H+),
resulting in the observed pH decrease in aqueous MEA solutions on addition of CO2.
Reaction of MEA with water:
9:() + () * ↔ 9:(4 2 + *(- (5. 6)
Reaction of MEA with water and CO2:
29:() + () * + <*) ↔ 9:(<**- + 9:(4 2 + () * (5. 7)
9:(<**- + () * ↔ 9:() + (<*4 - (5. 8)
- Dissociation of protonated amine ions
9:(4 2 ↔ 9:() + (2 (5. 9)
- Dissociation of hydrogen carbonate ions

P a g e | 195
(<*4 - ↔ <*4 )- + (2 (5. 10)
1.0
0.9
RNH2
RNH3+
0.8
0.7
Mole Fraction
0.6
HCO3-
0.5
0.4
0.3
CO2
0.2
RNHCOO-
0.1
CO32-
0.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
CO2 loading / mol CO2 (mol amine)-1
Figure 5-6: Liquid-phase composition of 30wt% MEA solution loaded with CO2 at 25°C.
Compositions were calculated with the Kent-Eisenberg Model (Kent and Eisenberg, 1976).
The cyclic voltammogram recorded for a stationary iron electrode in 30 wt% MEA at 25oC
where the solution is fully saturated with CO2 and has a pH 8.01 is reported in Figure 5-7 and
Figure 5-8. The iron electrode attained an OCP of -0.71 V after being cleaned cathodically at -
1.4 V for 120 s to reduce any oxides that may have formed spontaneously in air. The electrode
potential was scanned under the same constraints as for those solutions not containing CO2 to
provide a direct comparison (-1.4 V to +1.0 V, dE/dt = 10 mV s-1). The potential-pH diagram
for Fe-H2O-CO2 systems (Figure 5-9) was used to assign reactions to the main features of
Figure 5-8.
P a g e | 196
20
(II)
10
(I)
(III)
0
(V)
(IV)
-10
A B C
-20
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Figure 5-7: Cyclic voltammogram of stationary iron RDE in 30 wt% MEA + H2O + CO2;
10 mV s-1 (T = 25 °C, pH 8.01).
50
EFe3O4/ Fe2O3= -0.48V
40
EFe2+/Fe3O4 = -0.66V
30
EFe(OH)2/Fe(OH)3= -0.43V
EFe/Fe(OH)2 = -0.76V
20
EFe/Fe2+ = -0.79V
10
EFe/FeCO3 = -0.88V
0
-10
-20 EH2/H2+= -0.67V
-30
-40
A B C
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Figure 5-8: Cyclic voltammogram of a stationary iron RDE in 30 wt% MEA + H2O + CO2, at
10 mV s-1 (T = 25 °C, pH 8.01) with reactions assigned.
P a g e | 197
Large cathodic current densities resulted from potentials between the negative potential limit
and -0.75 V (AgCl|Ag) (region A), due to hydrogen evolution, firstly from the reduction of
water (E"# /"# % = −0.68 V) (5. 15) and secondly from protons resulting from the reaction of
CO2 in the aqueous MEA solution. The dissociation of protonated amine (RNH3+) (5. 9) and
HCO3- formed from the reaction of CO2 in aqueous MEA (5. 10) and CO2 in water (5. 13) serve
as an additional source of H+ which are consequently reduced via equation (5.14).
Hydration of CO2:
<*) (D) ↔ <*) (EF) (5. 11)
Hydrolytic Ionisation of CO2 (aq):
<*) (EF) + () * ↔ (2 + (<*4 - (5. 12)
(<*4 - ↔ (2 + <*4 )- (5. 13)
Hydrogen Evolution reduction reactions:
2(2 + 2, - → () (5. 14)
2H) O + 2e- → H) + 2OH - (5. 15)
With increasing potentials, oxidation current densities resulted at potentials ≥ -0.78 V (region
B); predictions from the potential-pH diagram (Figure 5-8) for Fe-H2O-CO2 systems at pH
8.01, suggested that this is due to the active dissolution of iron to FeIIaq species (6JK#L / =
−0.79 V) via reaction (5. 16).

P a g e | 198
Dissolution of Fe:
/, ↔ /, )2 + 2, - (5. 16)
The cyclic voltammogram was characterised by three oxidation peaks at ca. -0.72 V (peak I)
and a maximum current density of 16.2 A m-2 at ca. -0.49 V (peak II), followed by another
broad peak at ca. -0.31 V. Similar behaviour has been reported (Castro et al., 1986) for iron in
stagnant and stirred KHCO3 / K2CO3 at pH 8.9, whereas others (Clerbois and Massart, 1962),
(Thomas et al., 1970), (Simard et al., 1998) reported only two distinct current maxima.
The j-E data in Figure 5-7 exhibited an active-passive transition, indicated by the blue dashed
line; The active region was followed by sharp decrease in anodic current to 3.2 A m-2 at
potentials > -0.3 V and a passive region extending over 1.1 V. Passivation became evident at
potentials at potentials more positive of those corresponding to peak (III) at which oxidation to
a Fe(III) phase was thermodynamically favourable. In the passive region, current densities
remained relatively constant until the scan was reversed at the positive potential limit. From
the Fe-H2O-CO2 potential-pH diagram in Figure5-9, passivation can be ascribed to formation
of an adsorbed film of Fe2O3 (6OPQ 8R /OP# 8Q = −0.48 V) and Fe(OH)3 (6OP 87 # /OP(87)Q =
−0.43 V). This potential-independent passive region ensued even on the negative-going
potential scan until low cathodic current densities occurred at ca. -0.52 V < E < -0.02 V.
Unexpectedly, on the negative-going potential scan to potentials < -0.52 V, the oxidation
process was suddenly reactivated, leading to net oxidation current densities, causing an anodic
peak (IV) at ca. -0.59 V. After this abrupt change, anodic current densities decreased and
changed to a net reduction process (region A, indicated by the red dashed line). There was a
reduction peak (V) at ca -0.80 V before hydrogen evolution current densities were detected at
the most negative potentials.

P a g e | 199
Successive potential cycling resulted in very little change in both anodic and cathodic current
densities and the voltammograms were invariant with cycle number, suggesting that at this
chosen scan rate of 10 mV s-1, the adsorbed species from the anodic scan were completely
reduced at the set negative potential limit of -1.4 V.
The total oxidation charge density was significantly higher than that corresponding to the
reduction reactions (Qox/Qred =12, in the potential range of -0.78 V to 1.00 V) resulting most
probably due to active dissolution of Fe to FeIIaq within the active region.
At the end of each experiment, it was evident to the naked eye that the electrode surface had
been etched strongly, with evidence of pitting and a general change in surface roughness.
2.2
2.0 H3FeO4+ H2FeO4
HFeO4-
1.8
1.6 FeO42-
1.4
1.2
1.0
0.8
Fe3+ Fe3+
0.6 O2
0.4
Fe2O3 (c)
H2O
0.2
0.0
H+ Fe2+ Fe(OH)3 (ppt)
Fe(OH)4-
-0.2 H2 FeCO3 Fe(OH)4-
-0.4 FeCO3
-0.6 Fe2(OH)2CO3 Fe(OH)2
Fe3O4
-0.8 Fe Fe(OH)2 Fe(OH)42-
-1.0
-1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Figure 5-9: Potential-pH diagram for Fe-CO2-H2O System, showing the lines for dissolved iron
activity of 10-4 and (0.1 MPa CO2, 298 K).
P a g e | 200
5.3. Effect of electrochemical parameters
On establishing the typical behaviour of a stationary Fe electrode in CO2-loaded MEA
solutions, the effects are described below of potential scan rate, rotation rate and positive
potential limit to elucidate the mechanisms of the various electrode processes of iron in MEA-
H2O-CO2 systems.
5.3.1. Effect of electrode rotation rate
As discussed previously in the Experimental Methods section of Chapter 4, a rotating disc
electrodes (RDE) system offers the benefit compared to the use of a stationary electrode of an
easily changeable and predictable mass transport regime and hence mass transport limited
current densities (jL), for which Levich published an analytical solution for steady-state laminar
flow conditions:
TU = 1.554XK /<Y)/4 X -Z/[ \ Z/) (5. 17)
Hence, for mass transfer controlled reactions, current densities are proportional to the square
root of the rotation rate (f).
Figure 5-10 shows the effects of rotation rate on the cyclic voltammetric behaviour of a Fe
RDE in 30wt% MEA saturated in CO2 at pH 8.01. Within the active dissolution region (region
B), both the peak current densities and charges associated with peaks II and IV increased
significantly with increasing rotation rate, while there was no effect on peak I or on passive
current densities at potentials > ca. -0.3 V (AgCl|Ag) and peak (III) disappeared completely.
P a g e | 201
80 (II)
60
40
(I)
20
(III)
0
(V) (IV) 0 rpm
60rpm
-20 240rpm
540rpm
A B C 960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-10: Effect of rotation rate on cyclic voltammograms of an iron RDE in 30wt% MEA
+ H2O + CO2, at 10 mV s-1 (T = 25oC, pH 8.01).
Figure 5-10 shows the effects of rotation rate on the peak current densities (jP). Current
densities of Peak (II) and Peak (IV) increased linearly with f1/2. At constant potential scan rate,
temperature and pH, jpII vs f1/2 was linear for the whole range of experimental conditions used.
This increase in current densities with electrode rotation rate suggested that the diffusion of
FeII from the electrode into the solution controlled the magnitudes of Peak (II) and Peak (IV),
confirming that anodic dissolution was mass transport controlled. The absence of an obvious
reduction current density peak due to:
→ 2 Fe 2+ + 3H 2O
Fe2O3 + 6 H + + 2e − ⎯⎯ (5. 18)
→ Fe 2+ + 3H 2O
Fe(OH )3 + 3H + + e − ⎯⎯ (5. 19)
In Figure 5-10 suggested that it was masked by a larger oxidation current density:
P a g e | 202
→ Fe2+ + 2e −
Fe ⎯⎯ (5. 20)
leading to a measured net positive current density in the potential range
E FeCO3 Fe > E > E Fe2O3 FeCO3

, as removal of (hydr-)oxide by reactions (5. 18) and (5. 19)
reactivated reaction (5. 20), giving rise to peak IV. While there was a linear relationship
between the jpII and f1/2, the lines did not intercept the origin, as current densities due to the
process corresponding to peak II were still finite in quiescent solutions (jpII ≠ 0 for f1/2 = 0).
This suggested that the diffusion-limited dissolution process occurred in parallel with a
kinetically-limited oxidation process. In the whole range of solution composition studied, the
value of jpII increased linearly with f1/2, but the gradients of the lines and their intercepts to the
origin of coordinate increased with temperature, reflecting the temperature dependence of the
FeII diffusion coefficient:
k BT
DFe2 + ,T = (5. 21)
6π rηT
80 (a) 20 (b)
18
Peak Current Density (jp) / A m-2
Peak Current Density (jp)) / A m-2
70 16
60 Peak (II)
14
12 Peak (V)
50 10
40 8
6
30 4
20 2
0
Peak (I) Peak (IV)
10 -2
-4
0
-6
0 1 2 3 4 5 6
0 1 2 3 4 5 6
(Rotation rate / s-1)1/2 (Rotation rate / s )
-1 1/2
Figure 5-11: Effect of rotation rate on jP for Fe RDE in 30wt% MEA + H2O + CO2, at 10 mV s-
1
(T = 25 °C, pH 8.01) in the a) positive-going and b) negative-going potential scan.
P a g e | 203
5.3.2. Effect of positive potential limit
To further establish the oxidation mechanisms for iron in aqueous MEA solutions, the
participation and stability of adsorbed or passive film-forming species was studied, by varying
the electrode potential perturbation applied to the iron electrode. Figure 5-12 shows the current
density response to varying the positive potential limit Es,a on the subsequent negative-going
potential excursion for a Fe electrode rotated at 1500 rpm (dE/dt = 10 mV s-1) in 30 wt% MEA
+ H2O + CO2. As the positive potential limit Es,a was decreased, the apparent magnitude of the
re-activation current peak (IV) increased on the subsequent negative-going scan. When Es,a was
set to a potential (-0.4 V) just beyond the potential of peak II, an anodic current peak was
detected with no prior cathodic current. At more positive limit potentials (-0.2 V < Es,a < 1.0 V),
there was a broad reduction current (III) before the oxidation current peak (IV). The charge
density of the anodic current peak IV decreased while the charge of the cathodic current peak
III increased with more positive values of Es,a.
The magnitudes of peaks I and V did not change on varying Es,a, suggesting that they were
conjugated in nature and can be assigned to the oxidation and subsequent reduction of an inner
surface layer. Peak I was assigned to the oxidation of iron to form the first layers of a pre-
passive insoluble phase such as Fe(OH)2 as predicted by the potential-pH diagram of the Fe-
H2O-CO2 system (Figure 5-9):

P a g e | 204
→ Fe2+ + 2e −
Fe ⎯⎯ (5. 22)
Fe + HCO3− ⎯⎯
→ FeCO3 + H + + 2e − (5. 23)
E FeCO3 Fe (SHE) / V = −0.413 − 0.0296 pH − 0.0296log ( HCO3− ) (5. 24)
3Fe2O3 + 2 H + + 2e − ←⎯
⎯⎯
⎯→ 2 Fe3O4 + H 2O (5. 25)
Fe3O4 + 5H + + 3HCO3− + 2e − ←⎯
⎯⎯
⎯→ 3FeCO3 + 4 H 2O (5. 26)
And or: Fe2O3 + 4 H + + 2 HCO3− + 2e − ←⎯

⎯⎯
⎯→ 2 FeCO3 + 3H 2O (5. 27)
EFe2O3 FeCO3 (SHE) / V = 0.659 − 0.118 pH + 0.0592log ( HCO3− ) (5. 28)
Where VFe O = 30.274 cm3 mol-1 ; VFe O = 44.524 cm3 mol-1 ; VFeCO = 29.378 cm3 mol-1
2 3 3 4 3
(Haynes and Lide, 2011).
(Castro et al., 1991) reported a similar peak (I), attributed to the formation of a monolayer of
Fe(OH)2, in voltammograms of Fe in hydrogen carbonate / carbonate solutions at pH 8.9; this
inner passive layer of Fe(OH)2 formation was believed to occur simultaneously with the active
dissolution of Fe to FeII, while the potential region of peak (II) and above could be associated
with the growth of the outer passive layer.

P a g e | 205
80
(II)
60
40
(a)
20
(I) (V) (b) (c) (d) (e)
0
(IV) (III)
-20
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Figure 5-12: Cyclic voltammograms for Fe | 30wt% MEA + H2O + CO2, at 10 mV s-1
(T = 25 °C, pH 8.01) with positive potential limits: (a) -0.40 V, (b) -0.20 V, (C) +0.20 V (d)
+0.60 V (e) +1.0 V.
In general, the electrochemical response during the negative-going potential scan to changes in
Es,a could be attributed to changes in the stabilisation of the oxidation products formed during
the prior positive-going potential scan. The increase in the magnitude of the reactivation peak
(IV) with decreasing positive potential limits Es,a was due to the decreasing amount of
passivating film formed and so less charge was required subsequently for its reductive removal
by reaction (5. 18) and (5. 19), thereby reactivating the oxidation reaction (5. 20) anodic peak
(IV) was shown clearly to result from cathodic reduction of the passive film with the anodic
dissolution of the Fe to soluble FeII(aq).

P a g e | 206
5.3.3. Effect of potential scan rate
The potentiodynamic behaviour of Fe in 30 wt% MEA at pH 8.01 also varied with the potential
scan rate, but the anodic part of the voltammograms generally exhibited the same form and the
onset of oxidation occurred at the same potential for all scan rates (Figure 5-13).
At constant rotation rate, the magnitude of peak I depended linearly on the scan rate and the
potentials corresponding to the peaks also shifted with changes in scan rate which is indicative
of a process that is not diffusion limited but rather controlled by the charge-transfer process
(Figure 5-14). The negative-going potential scan at 100 mV s-1 exhibited cathodic current
densities independent of rotation rate with a current density peak at -0.39 V, the associated
charge of which decreased at 50 mV s-1. Current densities detected during the negative-going
potential scan due to re-activation of the oxidation process occurred only at lower scan rates
(1-50 mV s-1), with oxidation charge densities increasing and reduction charge densities
decreasing with decreasing scan rate. The current density of peak (II) appeared to be
independent of potential scan rate between 10 – 100 mV s-1, but increased with scan rate at
lower scan rates. The current densities of peak II did not increase linearly with either the
potential scan rate or its square root, as it was likely to result from diffusion controlled
dissolution occurring simultaneously with passivation of the iron surface by (hydr-)oxides. The
potential corresponding to the current peak II became more positive with increasing scan rate,
due to the latter process; hence for near steady-state measurements, a potential scan rate of
10 mV s-1 was maintained for all other experiments. The magnitude of peak II increased linearly
with ν for the range of potential scan rates used (Figure 5-15 and Figure 5-16) and the variation
in peak II with scan rate at each rotation rate used is also reported in Appendix (II). The total
anodic charge density (Qox) also increased with decreasing scan rate and is tabulated below:
P a g e | 207
Table 5-1: Effects of rotation and potential scan rates on anodic charge / C m-2 passed with Fe
RDE in 30 wt% MEA + H2O + CO2 (T = 25°C, pH 8.01).
Potential scan rate / mV s-1 100 50 10 5 1

Rotation rate / s-1
25 2.00 3.23 11.53 15.35 62.35
16 1.82 2.85 11.12 13.79 54.31
9 2.00 2.78 10.22 13.05 44.14
4 1.92 2.59 8.70 10.76 29.91
1 1.67 2.56 7.86 8.77 27.07
0 1.98 2.46 7.44 8.47 20.12
P a g e | 208
80 80
60 (a) (b)
60
40 40

20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
-120 -120
-140 -140
100mV/s 10mV/s
-160 -160
50mV/s 5mV/s
-180 -180
10mV/s 1mV/s
-200 -200
-220 -220
-1.4 -1 -0.6 -0.2 0.2 0.6 1 -1.4 -1 -0.6 -0.2 0.2 0.6 1
Electrode Potential (AgCl|Ag) / V Electrode Potential (AgCl|Ag) / V
Figure 5-13: Effect of potential scan rate on voltammograms for a Fe RDE in 30wt% MEA +
H2O + CO2 (T = 25oC, pH 8.01, f = 1500 rpm,) (a) 100 ≤ v ≤10 mV s-1 (b) 10 ≤ v ≤1 mV s-1.
35
Peak Current Density (jpI) / A m-2
30
25
20
15
10
0
0 10 20 30 40 50 60 70 80 90 100
Potential scan rate / mV s-1
Figure 5-14: Effect of potential scan rate on current density for peak I of stationary Fe electrode
in 30wt% MEA + H2O + CO2 (T = 25 °C, pH 8.01).
P a g e | 209
Peak Current Density (jpII) / A m-2 80
70
60
50
40
30
20 100mv/s
50mV/s
10 10mV/s
0
0 1 2 3 4 5 6
(Rotation rate / s-1)1/2
Figure 5-15: Effects of rotation and potential scan rates on current density of peak(II), jPII, on
Fe in 30wt% MEA + H2O + CO2, at scan rates 100 ≤ v ≤10 mV s-1
80
Peak Current Density (jpII) / A m-2
70
60
50
40
30
10mV/s
20 5mV/s
10 1mV/s
0
0 1 2 3 4 5 6
(Rotation rate / s-1)1/2
Figure 5-16: Effects of rotation and potential scan rates on peak (II) current density jPII of Fe
in 30wt% MEA + H2O + CO2, at various scan rates 10 ≤ v ≤ 1 mV s-1
P a g e | 210
5.4. Effect of Process Parameters
Results are reported below for the effects of CO2 loading, amine concentration, solution
temperature and oxygen presence on the electrochemical behaviour of iron in MEA-H2O- CO2
system using 30wt% MEA at full CO2 saturation 25oC as the reference point.
5.4.1. Effect of temperature
A variation in solution temperature has a significant effect on the electrochemical behaviour of
iron in aqueous MEA-CO2 systems. Figure 5-17 shows the effect of temperature in the range
25oC ≤ T ≤ 80oC on cyclic voltammograms for an iron electrode rotated at 1500 rpm in 30 wt%
MEA fully saturated with CO2 at a pH 8.01, as previously over the same potential range -1.4 V≤
E ≤ +1.0 V; similar results were obtained for other rotation rates (see Appendix II).
200
(II)
150
100
50
(I) (III)
0
(IV)
-50
80C 70C 60C
50C 40C 25C
-100
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-17: Effect of aqueous solution temperature on cyclic voltammograms for iron RDE in
30 wt% MEA + H2O + CO2 at pH ~8.01; 10 mV s-1, f = 1500 rpm.
The following observations were made from the experimental results:

P a g e | 211
(i) The onset of oxidation occurred at essentially the same potential (-0.75 ± 0.1 V) for
the temperature range studied.
(ii) Increasing temperatures resulted in peak (I), peak (II) and peak (IV) being displaced
towards less positive potentials. Peak (I) also became less pronounced with
increasing temperature and eventually disappeared at 80oC. This displacement of
the peak potentials could be explained by the change in the standard potential of the
AgCl/Ag reference electrode with temperature. The standard potential (the potential
at which the chloride activity is 1 mol kg-1) of the AgCl/Ag electrode is a function
of temperature. The correlation for the standard potential of the AgCl/Ag electrode
as a function of temperature, for 0 ≤ T ≤ 95 °C is given by (Bard, 1985) :
6 % ] = 0.23695 − 4.8564 ×10-5 _ − 3.4205 ×10-[ _ ) − 5.869×10-` _ 4
(5. 29)
This correlation was used to produce the data in the table below:
Table 5-2: Effect of temperature on the peak (II) electrode potential after taking into account
the shift in electrode potential of AgCl/Ag with respect to standard hydrogen electrode (SHE)
Temperature, T /°C AgCl/Ag vs. SHE /V Ep(II) vs. Ag/AgCl /V Ep(II) vs. SHE at T / V
o
C
25 0.2224 -0.453 -0.675
40 0.2120 -0.463 -0.675
50 0.2045 -0.471 -0.676
60 0.1965 -0.481 -0.678
70 0.1881 -0.490 -0.678
80 0.1791 -0.499 -0.678
(iii) As expected, current densities increased with temperature in the active region, peak
(1), peak (III) and the anodic reactivation peak (IV), whereas passive current
densities were independent of the temperature. At elevated temperatures, hydrogen

P a g e | 212
evolution current densities were increased, while oxygen evolution current densities
were detected only at T ≥ 50 C and increased with temperature.

°
(iv) In the 25-80oC range studied, there was a linear increase in the magnitude of peak
II with f1/2 and gradients of the lines and intercepts increased with temperature.
(v) The number and distribution of peaks in the voltammograms remained the same in
the 25-80 oC range studied, and the relationship between the current density for peak
II and T-1 was found to be linear, as shown in the Arrhenius plot for various rotation
rates in Figure 5-19. The apparent activation energy, Ea related to each peak was
calculated from the gradients of the Arrhenius plots and listed in Table 5-2.
180
160
Peak Current Density / A m-2
140
120
100 18C
25C
80
40C
60 60C
40 80C
20
0
0 1 2 3 4 5
Rotation Frequency 0.5/ Hz0.5

Figure 5-18: Effect of rotation rate and temperature (18 oC ≤ T ≤ 80 oC) on jPII for Fe RDE in
30 wt% MEA + H2O + CO2, at 10 mV s-1 (pH 8.01).
P a g e | 213
180
1500rpm
Peak Current Density (jpII) / A m-2
160 960rpm
540rpm
140
240rpm
120
100
80
60
40
20
0
0.0028 0.0029 0.003 0.0031 0.0032 0.0033 0.0034 0.0035
Temperature-1/ K-1
Figure 5-19: Arrhenius plot for jpII 30 wt% MEA + H2O + CO2, at 10 mV s-1 (pH 8.01) for
varying aqueous solution temperatures (18 oC ≤ T ≤ 80 oC) and rotation speeds (60 rpm ≤ f ≤
1500 rpm).
Table 5-3: Activation energy associated with jpII for Fe RDE in 30 wt% MEA + H2O + CO2,
at 10 mV s-1 (pH 8.01) for varying aqueous solution temperatures (18 oC ≤ T ≤ 80 oC) and
rotation speeds (60 rpm ≤ f ≤ 1500 rpm)
Rotation rate / rpm Activation Energy (Peak II) / kJ mol-1

1500 20.4
960 18.4
540 17.9
240 11.6
60 0.9
Eact < 30 kJ mol-1, implies process was controlled by diffusional process(es) rather than
chemical reaction control, i.e. breaking of chemical bonds, which normally requires
> 60 kJ mol-1.
P a g e | 214
5.4.2. Effects of CO2 loading and pH
From the voltammetric behaviour of iron in 30 wt% MEA, it was inferred that the decrease in
pH due to absorption of CO2 facilitated oxidative dissolution, whereas alkanolamines alone are
not intrinsically corrosive. In order to further validate this behaviour, an Electrochemical
Quartz Crystal Microbalance (EQCM) was used to measure the resulting mass loss and
discriminate between passivation and dissolution processes with a resolution on the nanogram
(ng cm-2) scale. It has been employed effectively in the study of FeCO3 passive film formation
by (Fajardo et al.) and the growth and dissolution of passive Fe-Cr alloys by (Hamm et al.,
2002) & (Schmutz and Landolt, 1999)
As there are no commercially available Fe quartz crystal electrodes compatible with the EQCM
cell utilized for this research, iron electrodes were produced by electroplating the gold electrode
at a potential of -1.0 V in a 300 g dm-3 ferrous sulfate solution at pH 3.2 and 25 °C. Figure 5-
20 shows the polarisation behaviour and the accompanying mass change of the gold quartz
crystal in that solution.
On the positive-going potential scan starting from the negative potential limit (-1.4 V),
reduction current densities were detected initially with a corresponding increase in mass due to
the reduction of the FeIIaq species to Fe, which was deposited on the surface of the gold crystal
via reaction (5. 30):
/, )2 + 2, - ↔ /, (5. 30)
On scanning the potential to values positive of the equilibrium potential for iron dissolution to
FeIIaq (6OP #L /OP = −0.79 V), oxidation current densities were recorded accompanied by an
expected mass loss, until the electrode reached the set positive potential limit. On the
subsequent negative-going potential excursion, current densities and the resulting mass change
P a g e | 215
remained essentially constant until the potential reached the equilibrium potential 6OP #L /OP ,
after which reduction of the dissolved FeIIaq species to Fe occurred with an increase in mass.
600 6000
Fe2+(aq) + 2e- ← Fe
400 5000
4000
∆m / mg m-2
200
3000
0
2000
EFe/Fe2+ = -0.79 V
-200
1000
-400 0
Fe2+(aq) + 2e- →Fe
-600 -1000
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4
Figure 5-20: Cyclic voltammogram and corresponding massogram for Au quartz crystal in
300 g dm-3 aqueous FeSO4 at pH 3.2; at 10 mV s-1, T = 25 C). °
From the voltammogram, it was deduced that the reduction of FeII from a solution of FeSO4
would occur at potentials negative of the equilibrium potential. Hence, a potential of -1.0 V
was selected for iron deposition; the resulting current density-time and Δm-time data are shown
in Figure 5-21. Δm increased linearly with time and a total mass of 5838 mg m-2 was deposited
in 200 s. The total mass increase detected by the EQCM corresponded to 89.3% of the total
charge passed (2.26 ×104 C m-2) in the current-time data.

P a g e | 216
-70 6000
5000
-80
Fe2+ + 2e- → Fe
4000
∆m / mg m-2
-90
3000
-100
2000
-110
1000
-120 0
0 50 100 150 200
Time / s
Figure 5-21: Potentiostatic deposition of Fe on Au quartz crystal in 300 g dm-3 FeSO4 at pH

3.2, - 1.0 V (AgCl|Ag) and T = 25 °C.
After the deposition process was completed, the FeSO4 solution in the quartz crystal cell was
emptied and swiftly replaced with unloaded 30wt% MEA at pH 12.55. Figure5-22 shows the
resulting cyclic voltammogram of the electroplated Fe crystal swept at a rate 100 mV s-1 from
-1.0 V to a positive limit of +0.1 V and negative limit of -1.2 V. Figure 5-23 shows the mass
change corresponding to the voltammogram in Figure 5-22. Prior to the voltammetric
procedure, the electrode was reduced at -1.0 V for 60 s to reduce any oxides that may have
formed spontaneously on the surface. At the starting potential of -1.0 V, the voltammogram
showed oxidation current densities due to the oxidation of Fe to Fe(OH)2 (6OP(87)# /OP =
−1.03 V) and the subsequent oxidation to Fe3O4 (6OPQ 8R /OP(87)# = −1.01 V). The first cycle
produced 3 anodic peaks, while on the second cycle, the first and third peak became less evident
and only the second peak was prominent. This broadening effect of the peaks, as well as
decreasing current densities, was due to the accumulation of passivating species on the surface
which were not completely reduced on the subsequent negative-going potential scan. In
P a g e | 217
accordance with the results from the cyclic voltammogram, on the positive-going cycle, there
was a net increase in mass due to the adsorption of oxides/hydroxides on the surface, and a loss
in mass on the negative-going cycle as these oxides/hydroxides were reduced back to elemental
Fe (Figure 5-23). On closer inspection of the voltammogram and the resulting Δm, it can be
seen that between the starting potential of -1.0 V and -0.6 V (peak II), increasing current density
resulted in an increase in mass and at -0.6 V≤ E ≤+0.1 V, both current densities and Δm
remained essentially constant due to passivation of the electrode surface. On the subsequent
negative-going potential scan, passivation current densities and Δm were initially constant until
reaching a potential of ca. -0.56 V, below which current densities and Δm started to decrease
due to reduction of the species that had been oxidised during the previous positive-going scan.
In general, there was a net accumulation of mass from cycle to cycle.
10
EFe(OH)2/Fe3O4 = -0.70 V
8
6
4
2
EFe3O4/Fe2O3 = -0.75V
0
-2 EFe(OH)2/Fe3O4 = -1.01V
-4
-6
-8 EH2/H2O = -0.94 V
-10
EFe/Fe(OH)2 =-1.03V
-12
-14
-16
-18
-20
-1.2 -1.1 -1.0 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1
Figure 5-22: Cyclic voltammogram obtained on the deposited Fe quartz crystal in 30 wt% MEA
+ H2O, at 100 mV s-1 (T = 25 °C, pH 12.55) with equilibrium reaction potentials assigned.
P a g e | 218
30
25
20
∆m / mg m-2
15
10
0
-1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1
Fe quartz crystal in 30 wt% MEA + H2O, at 100 mV s-1 (T = 25 °C, pH 12.55) with reactions
assigned.
The voltammetric / Δm-E behaviour was completely different in CO2 loaded 30 wt% MEA
solutions at pH 8.01 (Figure 5-24 and Figure 5-25). The voltammogram (dE/dt = 100 mV s-1)
exhibited 3 oxidation peaks at ca. -0.67 V, -0.51 V and -0.28 V and two reduction peaks at ca.
-0.44 V and -0.91 V, as in the voltammogram recorded with the RDE under the same conditions
(f = 0) and scan rate. However, unlike the voltammogram recorded in the absence of CO2, there
was a net loss in mass during the positive-going potential scan corresponding to the oxidative
dissolution of Fe as aqueous FeII. From the potential at which oxidation current densities
resulted from ca. -0.88 V to -0.38 V just after the appearance of peak II, there was a mass loss
of 9.28 mg m-2 which corresponded to about 52 % of the total anodic charge passed (60.84 C m-
2
). In the passive region extending from -0.4 V to the positive potential limit of +0.5 V, both
current densities and Δm remained essentially constant, even on the negative-going scan until
reaching a potential of -0.4 V. At more negative potentials, reduction current densities resulted
and Δm continued to decrease due to a reduction of the species formed back to elemental iron.
Suddenly, at ca. -0.85V there is an increase in Δm even with increasing reduction current. This
P a g e | 219
was due to the deposition of Fe from dissolved FeIIaq in the solution (6OP #L /OP = −0.79 V).
From cycle to cycle, there was a slight increase in current densities and a significant mass loss.
40
EFe3O4/Fe2O3 = -0.48V
30
EFe2+/Fe304 = -0.66V EFe(OH)2/Fe(OH)3= -0.43 V
20
(II) (III)
EFe/Fe2+ = -0.79V
10 EFe/FeCO3 = -0.88V (I)
-10
(V) (IV)
-20
EH2/H2O = -0.92 V
-30
-40
-50
-1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

H2O + CO2 at pH 8.01, 100 mV s-1, T = 25 °C; equilibrium reaction potentials assigned.
5
0
-5
-10
∆m / mg m-2
-15
-20
-25
-30
-35
-40
-45
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6

Fe quartz crystal in 30 wt% MEA + H2O + CO2, at 100 mV s-1 (T=25 °C, pH 8.01).
P a g e | 220
The effects of CO2 loading / pH was studied further with the RDE by varying the CO2-air
mixture to achieve the desired CO2 loading. The reported pH values were recorded at 25 C. As °
mentioned previously, in the absence of CO2 at pH 12.55, the oxidation part of voltammogram
displayed only a passive region; the same was found for low CO2 loading at pH 11.00 (Figure
5-27).
The characteristic active-passive behaviour of Fe electrodes in MEA solutions loaded with CO2
became evident only at pH 10.01 (Figure 5-28), at which a barely visible and broad current
density peak at ca. -0.67 V appeared due to the active oxidative dissolution of Fe to FeII. The
current density and charge resulting from the active current peak continued to increase with
CO2 loading (decreasing pH), with the highest value for the active current peak being recorded
at the lowest pH, when the solution was fully saturated with CO2 at pH 8.01 (Figure 5-34). The
same was found for the re-activation current density peak on the negative-going scan which
became evident only at pH ≤ 8.43 and increased with CO2 loading. It is also worth noting that
increasing the CO2 loading and reducing the pH led to a positive potential shift in the onset of
oxidation, the active current density peak and re-activation peak.
40
EFe(OH)2/Fe(OH)3 = -0.70V
pH 12.55
30
EFe304/Fe203 = -0.75V EH2O/O2 =+0.29V
20
EFe(OH)2/Fe304 = -1.01V
10 EFe/Fe(OH)2 = -1.03V
-10
0 rpm
60rpm
-20 240rpm
EH2O/H2 = -0.94 V
540rpm
-30 960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
P a g e | 221
Figure 5-26: Effect of rotation rate on cyclic voltammograms of iron RDE in 30 wt% MEA at
pH 12.55 and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned.
40
EFe(OH)2/Fe(OH)3 = -0.60V pH 11.00
30 EFe(OH)2/Fe(OH)3 = -0.66V
20 EFe(OH)2/Fe304 = -0.92V EH20/O2= +0.38V
10
-10
0rpm
EH2/H+ = -0.85V 60rpm
-20 240rpm
540rpm
-30 960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned.
40
EFe304/ Fe203 = -0.60V pH 10.01
30
EFe(OH)2/Fe304 = -0.86V
EFe(OH)2/Fe(OH)3 = -0.54V
20 EFe/Fe(OH)2= -0.88V
EH2O/O2 = +0.44V
10 EFe/FeCO3 = -0.91V
-10
0rpm
-20 EH2/H2O = -0.79V 60rpm
240rpm
-30 540rpm
960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
and T = 25 C; 10 mV s-1, with equilibrium potentials of reactions assigned.
P a g e | 222
40
pH 9.54
30
EFe304/Fe203= -0.57V
20
EFe(OH)2/Fe304 = -0.83V EFe(OH)2/Fe(OH)3 = -0.52V

EFe/Fe(OH)2 = -0.85V EH2/O2 = +0.47V
10
EFe/FeCO3 = - 0.89V
0
-10 0rpm
EH2/H+ = -0.76V 60rpm
-20 240rpm
540rpm
-30 960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
40
pH 9.06
30 EFe304/Fe203 =- 0.54V
EFe(OH)2/Fe304 = -0.80V EFe(OH)2/Fe(OH)3 = -0.49V
20 EFe/Fe(OH)2 = -0.82V
EH20/O2 = +0.50 V
10 EFe/FeCO3 = -0.88V
-10
0rpm
EH2/H2O = -0.73 V 60rpm
-20 240rpm
540rpm
-30 960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
P a g e | 223
40
EFe304/Fe203 = -0.52 V pH 8.78
30
EFe(OH)2/Fe(OH)3 = -0.47 V
20 EH2O/O2 = +0.50 V
EFe(OH)2/ Fe304 = -0.79V
EFe/Fe(OH)2 = -0.81 V
10
EFe/FeCO3= -0.87 V
0
-10
EH2/H2O = -0.72 V 0 rpm
-20 60rpm
240rpm
-30 540rpm
960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
and T = 25 °C; 10mVs-1, with equilibrium potentials of reactions assigned.
60
EFe304/Fe203 = -0.50 V pH 8.43
40
EFe2/Fe304= -0.76V
20 EFe/Fe(OH)2 = -0.79 V
EH20/O2 = +0.53 V
EFe/Fe2+ = -0.79 V
EFe/FeCO3 = -0.86 V
0
0rpm
60rpm
EH2/H2O = -0.70 V
-20 240rpm
540rpm
960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
P a g e | 224
60
pH 8.17
40 EFe2+/ Fe3O4 = -0.70 V
EFe/Fe(OH)2= -0.77 V
EFe/Fe2+ = -0.79 V
20 EH2O/O2 = 0.55 V
EFe/FeCO3 = -0.85 V
0
0rpm
60rpm
EH2/H+ = -0.68 V 240rpm
-20 540rpm
960rpm
1500rpm
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-33: Cyclic Voltammogram obtained on a stationary iron RDE in 30 wt% MEA+ CO2
at pH 8.17 and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned.
80 EFe3O4/ Fe2O3= -0.48 V
60
EFe2+/Fe3O4 = -0.66 V EFe(OH)2/Fe(OH)3= -0.43 V

40 EFe/Fe(OH)2 = -0.76 V
EFe/Fe2+ = -0.79 V
20 EH2O/O2 = +0.56 V
EFe/FeCO3 = -0.88 V
-20 EH2/H2+= -0.67 V
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-34: Cyclic Voltammogram obtained on a stationary iron RDE in 30wt% MEA+ CO2
at pH 8.01 and T = 25 °C; 10 mV s-1, with equilibrium potentials of reactions assigned.
P a g e | 225
80
60
Increasing CO2 loading
40
20
0 8.78
8.64
-20 8.43
8.17
8.01
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-35: Effect of pH on voltammograms of a stationary iron RDE in 30wt% MEA+ CO2
at T = 25 °C; 10 mV s-1, f = 1500 rpm.
0.0025
Peak Current Density jpII / A m-2
pH 10.01
0.002
pH 9.51
pH 9.02
0.0015
pH 8.78
pH 8.43
0.001
pH 8.17
pH 8.01
0.0005
0
0 1 2 3 4 5
(Rotation rate / s-1)0.5
Figure 5-36: Effects of rotation rate and pH on peak current density jPII of Fe in 30 wt% MEA
+ H2O + CO2; potential scan rate: 10 mV s-1.
P a g e | 226
Peak Potential EpII (AgCl|Ag) / V -0.4
-0.45
-0.5
-0.55
-0.6
-0.65
8 8.5 9 9.5 10 10.5
pH
Figure 5-37: Effect of aqueous solution pH on peak (II) electrode potential EPII.
Whilst it was established that MEA solutions became corrosive only on absorption of CO2, it
was yet to be determined if this is due primarily to the shift in pH of the system or as a result
of the aqueous CIV species (HCO3-, CO32-) and amine species (RNH3+) resulting from the
chemical reactions of MEA, CO2 and H2O (Reactions 5. 6 – 5. 15). For this purpose, the effect
of pH on the voltammetric behaviour of iron in MEA was also studied in the absence of CO2
using both the RDE (Figure 5-38) and the EQCM (Figure 5-39 & Figure 5-40). The pH of the
unloaded 30 wt% MEA solution was adjusted to the desired pH of 8.01 by addition of aliquots
of 1.0 M H2SO4 solution at 25 C. The MEA solution with H2SO4 resulted in a radically different
°
voltammogram from that of the solutions loaded with CO2 at the same pH. The onset of
oxidation occurred at the same potential for both solutions and current densities continued to
increase due to the dissolution of Fe to FeIIaq (6OP #L /OP = −0.79 V) until it reached a plateau
at ca. -0.56 V, possibly limited by a dissolution-precipitation process due to the formation of
(hydr-)oxides. As predicted by the potential-pH diagram for the Fe-H2O system (Figure 5-5),
this was due to the formation of Fe(OH)2, which was also observed during the experiments as
P a g e | 227
a green rust on the surface of the electrode and in the solutions. Unlike the previous
voltammogram in CO2, in MEA + H2SO4 solutions, after scanning past the potential range for
the active dissolution of Fe to FeII, current densities started to increase and there was a second
active peak ca. +0.30 V at high rotation rates (540-1500 rpm). At lower rotation rates (0-
240 rpm), the second peak broadened. This was also evidenced in the Δm response of the
EQCM (Figure 5-40) where it can be seen that there was a continued mass loss during the
positive-going potential scan even at potentials up to +0.15 V, unlike the case of MEA-CO2
systems in which mass loss ceased at potentials ≥ -0.4 V. In total, the mass lost during the
anodic scan in MEA + H2SO4 solution was about 16.5 times that of the MEA + CO2 systems.
Further positive-going potential scan resulted in a short passive region from potentials greater
ca. +0.4 V for the RDE and +0.15 V in the case of the EQCM. It is important to note that the
current densities and charge of the peak associated with Fe→FeII were much higher in the
presence of CO2. The contrary was the case for the rest of the positive-going potential scan, for
which higher current densities were recorded in sulfuric acid due to active dissolution as
opposed to passivation in the CO2 system. The negative-going potential scan was also
characterised by a reduction current peak at ca. -0.41 V and not a re-activation peak as evident
in solutions loaded with CO2.

P a g e | 228
150 250
1500 rpm 960 rpm
200

100
150
50
100
0 50
0
-50
MEA + CO2 pH 8.01 -50 MEA + CO2 pH 8.01
MEA + H2SO4 pH 8.01 MEA + H2SO4 pH 8.01
-100 -100
-1.4 -1 -0.6 -0.2 0.2 0.6 1 -1.4 -1 -0.6 -0.2 0.2 0.6 1
200 300
540 rpm 240 rpm
Current Density / A m-2 200

100
100
0
0
-100 MEA + CO2 pH 8.01
MEA + CO2 pH 8.01
-100 -200
-1.4 -1 -0.6 -0.2 0.2 0.6 1 -1.4 -1 -0.6 -0.2 0.2 0.6 1
300 300
60 rpm 0 rpm
200 200
100 100
0 0
-100 -100
MEA + CO2 pH 8.01 MEA + CO2 pH 8.01
-200 -200
-1.4 -1 -0.6 -0.2 0.2 0.6 1 -1.4 -1 -0.6 -0.2 0.2 0.6 1
Figure 5-38: Effects of CO2: – Comparisons between cyclic voltammogram for iron RDE in
30 wt% MEA+ CO2 and 30 wt% MEA+ 1 M H2SO4; 10 mV s-1 (T = 25 °C, f = 0- 1500 rpm).
P a g e | 229
200
EFe304/Fe2O3 = -0.48 V
150
EFe2+/Fe304 = -0.66 V
100
EFe/Fe(OH)2 = -0.76 V
50
EFe/Fe2+ = -0.79 V
0
-50
-100
-150
-200
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
H2SO4 at pH 8.01 and T = 25oC; 100 mV s-1, with equilibrium potentials of reactions assigned.
10
0
-10
-20
-30
-40
∆m / mg m-2
-50
-60
-70
-80
-90
-100
-110
-120
-130
-1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Figure 5-40: Mass change corresponding to voltammograms (Figure 5-39) of electrodeposited
Fe quartz crystal in 30 wt% MEA + H2O + CO2 at pH 8.01 and T = 25 C; 100 mV s-1. °
P a g e | 230
5.4.3. Effect of MEA concentration
The MEA concentration in the aqueous solution also appeared to have important effects on the
electrochemical behaviour of iron. Figure 5-41 & Figure 5-42 are voltammograms for an iron
RDE rotated at 1500 rpm in 5-30 wt% and 30-60 wt% MEA consecutively (25 C) that had been
°
fully saturated with CO2. As shown in Appendix (III), similar results were obtained for other
rotation rates.
The anodic and cathodic sections of the voltammograms generally exhibited the same form,
with the same number of peaks occurring at essentially the same potentials (Ep) as for 30 wt%
MEA and the onset of oxidation occurred at the same potential for each of the MEA
concentrations. However, MEA concentrations had a significant effect on anodic current
densities (Region B) associated with the Fe-FeIIaq dissolution process. Contrary to expectations
reported in the literature (Kittel, 2010, Veawab et al., 1999, Kladkaew et al., 2009a) of anodic
current densities increasing with MEA concentration, there were two distinct effects:
(i) Firstly, at lower concentrations (5-30 wt %), the anodic current densities increased
with increasing MEA concentration. The peak potential also shifted slightly to
lower potentials at higher MEA concentrations for 25-60 C, but at 80oC, the
°
maximum current density occurred in 40 wt% solution and was verified with the
EQCM. The voltammograms (Figure 5-54) obtained from the EQCM were similar
to those obtained from the RDE at 0 rpm. With increasing MEA concentration, the
Δm-potential plot (Figure 5-45 and Figure 5-46) showed a net loss of mass during
the positive-going potential scan.
(ii) At higher concentrations (30-60 wt% for 25-60oC, and 40-60 wt% for 80oC), anodic
current densities decreased and peaked at higher potentials with increasing amine
concentrations.
P a g e | 231
The linear relationship between the magnitude of the peak current density (jPII) and the square
root of RDE rotation rate (f1/2) was maintained for the entire range of experiments, regardless
of the amine concentration and temperature.
While the anodic current densities ascribed to the dissolution process (region B) were affected
considerably, the passivation process (region C) appeared to be completely independent of
MEA concentration.
80
60
MEA %
40
20
0
5wt%
10wt%
-20
A B C 20wt%
30wt%
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 25 °C; 10 mV s-1.
P a g e | 232
80
60
40
MEA%
20
30wt%
-20 40wt%
B C 50wt%
A 60wt%
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
RDE rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 25 °C; 10 mV s-1.
P a g e | 233
140 200
25C 10 wt% MEA 20 wt% MEA
120 40C
60C 150
100 80C
jp(II) / A m-2
jp(II) / A m-2
80
100
60
40 50
20
0 0
0 1 2 3 4 5 0 1 2 3 4 5
(Rotation rate / s-1)1/2 (Rotation rate / s-1)1/2
200 250
30 wt% MEA 40 wt% MEA
200
150
jp(II) / A m-2
jp(II) / A m-2
150
100
100
50 50
0 0
0 1 2 3 4 5 0 1 2 3 4 5
200 200
50 wt% MEA 60 wt% MEA
150 150
jp(II) / A m-2
jp(II) / A m-2
100 100
50 50
0 0
0 1 2 3 4 5 0 1 2 3 4 5
Figure 5-43: Effects on jPII of rotation rate, temperature (25-80 °C) and aqueous MEA
concentration (10-60 wt%) for Fe RDE in MEA + H2O + CO2 at pH ̴ 8.01; 10 mV s-1.
P a g e | 234
5wt% MEA 10wt% MEA

20 20 20 20
0 0 0 0

-20 -20 -20 -20
∆m / mg m-2
∆m / mg m-2
-40 -40 -40 -40
-60 -60 -60 -60
-80 -80 -80 -80
-100 -100 -100 -100
-120 -120 -120 -120
-140 -140 -140 -140

-1.4 -1 -0.6 -0.2 0.2 0.6 -1.4 -1 -0.6 -0.2 0.2 0.6
20wt% MEA 30wt% MEA

20 20 20 20
0 0 0 0
-20
-20 -20 -20
∆m / mg m-2
-40
∆m / mg m-2
-40 -40 -40

-60
-60 -60 -60
-80
-80 -80 -80
-100
-100 -100 -100
-120
-120 -140 -120 -120
-140 -160 -140 -140

-1.4 -1 -0.6 -0.2 0.2 0.6 1 -1.4 -1 -0.6 -0.2 0.2 0.6 1
Figure 5-44: Cyclic voltammograms and corresponding mass change of electrodeposited Fe

quartz crystal in (5-30 wt%) MEA + H2O + CO2 at pH ̴ 8.01 and T = 25 C; 100 mV s-1.
P a g e | 235
30
20
10
-10 5wt%
10wt%
-20 20wt%
30wt%
-30
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-45: Cyclic voltammogram of electrodeposited Fe on quartz crystal in (5-30) wt%

MEA + H2O at pH = 8.01 and T = 25 °C; 100 mV s-1.
20
Area specific mass change, ∆m / mg m-2
-20
5wt%
10wt%
-40
20wt%
30wt%
-60
-80
-100
-120
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 5-46: Mass changes corresponding to voltammograms (Figure 5-45) of electrodeposited

Fe quartz crystal in (5-30 wt%) MEA + H2O + CO2 at pH 8.01 and T = 25 °C; 100 mV s-1.
P a g e | 236
To clarify the effect of oxygen, cyclic voltammetry experiments for iron in 30 wt% MEA
solutions fully loaded with CO2 (25-80 °C) were repeated in the absence of O2 by purging
solutions with N2 prior to loading with CO2. During experiments, the cell was covered with
cling film and a nitrogen blanket was maintained to minimise oxygen ingress.
Cyclic voltammograms (Figure 5-47) obtained in the absence of oxygen generally exhibited
the same form and the onset of oxidation occurred at the same potential as those obtained in
the presence of oxygen; similar results were obtained for other rotation rates. However, the
magnitudes of current densities of peaks (I) & (II) in the active region were lower in the absence
of oxygen, the difference being more distinct at elevated temperatures (60-80oC). In the passive
region, the magnitudes of potential-independent current densities were also marginally lower
in the absence of dissolved oxygen.

P a g e | 237
150
100 40 °C
25 °C 100
50
50

0
0
-50
-50
-100
-100
-150 -150
-200 25C O2 1500rpm -200

25C N2 1500rpm O2 40C 1500rpm
N2 40C 1500rpm
-250 -250
-1.4 -1 -0.6 -0.2 0.2 0.6 1 -1.4 -1 -0.6 -0.2 0.2 0.6 1
200 200
60 °C 80 °C
150 150
100 100
50 50
0 0
-50 -50
-100 -100
-150 -150
-200 O2 60C 1500rpm -200

O2 80C 1500rpm
N2 60C 1500rpm N2 80C 960rpm
-250
-250
-1.4 -1 -0.6 -0.2 0.2 0.6 1
-1.4 -1 -0.6 -0.2 0.2 0.6 1
Figure 5-47: Effect of oxygen concentration and temperature on cyclic voltammograms of iron
RDE rotating at f = 1500 rpm in 30 wt% MEA+ CO2 at pH (7.24-8.01) ; 10 mV s-1.
P a g e | 238
5.5. Interpretation and Discussion
Cyclic voltammograms obtained over a wide range of potential scan rates, RDE rotation rates
positive potential limits, CO2 loadings, MEA concentrations and temperatures showed that the
active to passive transition of iron was strongly dependent on both the local hydrodynamic
conditions and the composition of the MEA solutions.
Iron in aqueous MEA solutions in the absence of CO2 exhibited passive behaviour, implying
minimal corrosion of the underlying metal. However, CO2 absorption in aqueous MEA reaction
resulted in the formation of protonated amine (RNH3+) and hydrogen carbonate ions (HCO3-)
that dissociate producing hydrogen ions (H+). The resulting decrease in pH from 12.55 in
unloaded solutions to 8.1 in fully saturated solutions, as well as the increased equilibrium
concentrations of oxidising species (H+ and HCO3-) shifted the system from a passive to an
active state, in which corrosion was thermodynamically and kinetically favourable.
Anodic dissolution region
Voltammograms of iron in MEA solutions at full saturation of CO2 exhibited at least two
anodic current peaks within the potential range of the active-passive transition, suggesting that
the dissolution and passivation of metal involves a minimum of two stages.
The first anodic stage occurred in the potential range of peak I, at which only FeII species are
likely as oxidation products, according to predictions from the potential-pH diagram of iron in
H2O-CO2 (Figure 5-8). The charge and magnitude of current density peak (I) in
voltammograms of iron in MEA solutions were essentially independent of RDE rotation rate
and pH, but increased linearly with potential scan rate, as expected for a surface transformation.
However, the current density of peak (I) was temperature dependent and the corresponding
charge increased with temperature. These results agree with those reported by (Castro et al.,
P a g e | 239
1986, Castro et al., 1991, Valentini et al., 1985) for iron in hydrogen carbonate/carbonate
solutions at pH 8-9.
Based on these observations, as well as thermodynamic predictions, the first stage of oxidation
of Fe (Peak I) can be assigned to a surface transformation reaction: the formation of a pre-
passive layer of Fe(OH)2 by the overall reaction:
Fe + 2H) O ↔ Fe(OH)) + 2H 2 + 2e- (5. 31)
As the iron surface was covered by the Fe(OH2) layer, the current decreased once the potential
of peak I was exceeded.
The second oxidation stage, linked to the higher anodic peak (II) occurred in the potential range
of -0.48 V to -0.52 V, at which only FeII species can be formed according on
thermodynamically predictions. There was a linear relationship between the current density at
peak II (jpII) with rotation rate (f1/2) and jp ≠ 0 at f1/2 = 0, suggesting that diffusion of ionic
species at the iron surface plays a role in limiting the rate of the oxidation process. The resulting
mass loss from EQCM results (Figure 5-24) clearly showed that reaction 5. 19 occurred
simultaneously with the oxidative dissolution of Fe via reaction (5. 32):
Fe ↔ Fe)2 + 2e- (5. 32)
Therefore, peak II can be attributed to the dissolution of iron through the pre-passive layer most
likely in the form of soluble FeIIaq species. As the reaction proceeded, both the thickness of the
Fe(OH)2 pre-passive layer and the concentration of FeII at the electrode-solution interface
increased. The dependence of jpII on rotation rate suggested that the latter reaction was the
major contributor to current densities in this potential range. The same was also found for the
anodic re-activation reaction that occurred during the negative-going potential sweep (peak
IV).
P a g e | 240
At the outer layer, dissolved FeII can be precipitated as FeCO3 if the concentration of the former
at the iron | solution interface exceeds the saturation value of the latter. On close inspection of
the Δm data from the EQCM study of iron under stationary conditions, (Figure 5- 24) there was
a mass gain in the region of peak 3 due to the precipitation reaction (5. 33).
/, )2 + (<*4 - + *(- → /,<*4 + () * (5. 33)
The formation of FeCO3 was predicted by the potential-pH diagram in Figure 5-8. FeCO3
precipitation was induced only by high local concentration of both the ferrous ions and
carbonate / hydrogen carbonate ions. As the rotation rate was increased, the local
concentrations of the soluble species near the electrode surface were decreased, deferring or
even preventing precipitation, so explaining the disappearance of peak III (Figure 5-10) on
rotating the RDE. The FeII activity-pH diagram for FeCO3 and Fe(OH)2 precipitation in H2O-
CO2 systems is shown in Figure 5-48.
1E+0
1E-1 FeCO3
1E-2
Dissolved FeII Activity
1E-3 Fe(OH)2 (ppt)

1E-4
1E-5
1E-6
1E-7
1E-8
1E-9
1E-10
1E-11
1E-12
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
pH
Figure 5-48: Activity-pH diagram of Fe-H2O-CO2 at 0.1 MPa CO2; 298 K (Kelsall, 2015).
In addition to FeCO3, another likely corrosion product of iron in CO2-containing systems is
iron hydroxycarbonate Fe2(OH)2CO3, for which the International Mineralogy Association
(IMA) approved the name ‘chukanovite’. The formation of Fe2(OH)2CO3 according to

P a g e | 241
reactions (5. 34) or (5. 35) was also predicted by the potential-pH diagram in (Figure 5-8) for
Fe-H2O-CO2 systems, from pH = 7-8 to pH 11.5-12, depending on the CO2 partial pressure.
However, the precipitation of chukanovite occurs preferentially to siderite at higher FeII
activities.
→ Fe2+ + FeCO3 + 2 H 2O
Fe2 (OH )2 CO3 + 2 H + ⎯⎯ (5. 34)
←⎯⎯
Fe2+ + FeCO3 + 2OH − ⎯⎯

→ Fe OH )2 CO3
←⎯⎯ 2( (5. 35)
At equilibrium: log ( Fe2+ ) = 10.70 − 2 pH (5. 36)
so higher FeII activities / concentrations are predicted to favour chukanovite rather than siderite.
The reactions that occurred within the active region were affected significantly by temperature;
the charge density of peak II more than doubled at 80oC compared to at 25oC.
In order to establish the effect that CO2 had on the corrosion behaviour of iron in MEA systems,
it was important to distinguish between the effects of pH and the effect oxidising species
resulting from the absorption of CO2 (HCO3-, CO32-, RNH3+). Hence, the behaviour was studied
of iron in CO2-free solutions of MEA that had been adjusted by the addition of H2SO4 to the
same pH as that of fully CO2 loaded solutions. The results clearly showed that in the absence
of CO2, even at the same pH, the electrochemical behaviour was radically different. As
observed in MEA-CO2-H2O systems, there was also an anodic current peak in the potential
range associated with dissolution of Fe to FeIIaq, but current densities were significantly less in
the absence of CO2; the differences were more evident at higher rotation rates (Figure 5-38).
The presence of CO2 increased the Fe oxidation rate predominantly by increasing the rate of
the hydrogen evolution reaction (HER):

P a g e | 242
2H 2 + 2e- → H) (5. 37)
The rate of the dissolution reaction was limited by the mass transport of H+ ions from the bulk
solution to the surface of the iron electrode. In solutions of pH ˃ 4, this limiting flux of H+ ions
is usually quite small. In the case of MEA-H2O-CO2, the flux of H+ ions was increased by the
presence of RNH3+ and HCO3- resulting from the absorption of CO2 in MEA (5. 7 – 5. 8) and
from the reaction of H2O and CO2 (5. 12). This explains the increased dissolution current
densities in the presence of CO2, as compared to the same MEA solution pH-adjusted with
sulfuric acid to the same pH.
The presence of HCO3- and RNH3+ in the solution can increase Fe dissolution rates through
dissociation via reactions (5. 9) – (5. 10) which serve as additional sources of H+ ions and were
consequently reduced via reaction (5. 37). Thus, it was unsurprising when the dissolution
current increased with increasing CO2 loading and decreasing pH (Figure 5-26 and Figure 5-
35).
The dependence of the iron dissolution kinetics on ionic composition was also studied using
cyclic voltammetry in sodium hydrogen carbonate-carbonate solutions at pH 8.1 (Appendix
IV). The voltammograms obtained were similar to those of iron in MEA+CO2+H2O systems.
There was also a large anodic peak (II) at the same potential which corresponded to the
dissolution of Fe to FeIIaq. In addition to the observed dependence of peak (II) on rotation rate,
its magnitude and associated charge increased linearly with HCO3- ion concentration. As this
set of results was obtained over a range of HCO3- / CO32- concentration ratio at a practically
constant pH, it can be concluded that the change in the kinetics of the dissolution reaction was
related principally to a process under HCO3- ion transport control.

P a g e | 243
It has also been suggested that the charge and current density of the peak increasing with
hydrogen carbonate concentration could be due to the removal of the pre-passive Fe(OH)2 layer
according to the following dissolution reaction (Castro et al., 1986):
Fe OH ) + HCO4 - → CO4 )- + OH - + Fe)2 + H) O (5. 38)
The presence of carbonate and hydrogen carbonate ions also contributed, by precipitating the
Fe2+ resulting from the dissolution of the underlying metal and enhancing the transport of ions
through the complex passive layer consisting of an Fe(OH)2 inner layer and an outer layer of
FeCO3.
In summary, it can be said that reactions (5. 18), (5. 19) and (5. 38) coupled with reaction (5.
32) result in the corrosion of the metal, as evident from Figure 5-24 and Figure 5-25, showing
EQCM data of mass loss in the potential range of peak II. Within this region, there was
competition between the dissolution of iron and the protective passive layer formation.
Passivation region
In the MEA solution, Fe dissolution was inhibited by the formation of a passive film,
characterized by the presence of peak III, which was evident only when the RDE was stationary
and disappeared when it was rotated. The magnitudes of the current densities detected within
the passive region were proportional to the scan rate and were practically independent of the
rotation rate, solution pH and the CO2 loading. This implied that the rate-determining step in
the passive region of peak III was indeed a surface transformation reaction. This implied that
the passive layer was an oxide or hydroxide film and HCO3-/ CO32- played no part in the
passivation process.
The potential-pH diagram for the relevant pH region predicts that passivation of iron / mild
steel at sufficiently anodic potentials, can be assigned to the formation of a Fe2O3 / Fe3O4 film.
At low potentials, the presence of Fe3O4 is thermodynamically favourable (at potentials ≥ -

P a g e | 244
0.66 V, pH 8.1) while Fe2O3 is more stable at higher potentials (≥ -0.48 V). The formation of
(metastable) FeIII hydroxide (Fe(OH)3) was also possible at potentials > -0.43 V. The total
charge density involved in passive film formation corresponded to ca. 5 monolayers.
On negative-going potential sweeps to potentials below the passive region, the j-E response
was dependent on the anodic potential limit; a reactivation of the anodic process (peak IV)
occurred at about the same potential range in which the dissolution reaction occurred on the
positive-going potential scan. This sudden and unexpected anodic current was probably due to
reductive dissolution of the passive film, exposing the underlying metal, which then oxidised
producing net current densities that decayed with decreasing potential. That decay could also
have resulted from FeII surface concentrations inhibiting further dissolution because of e.g.
FeCO3 formation. In summary, the passive film becomes increasingly stable as the positive
potential limit was increased (Figure 5-12).
5.6. Kinetic analysis
From the potentiodynamic study of iron in MEA solutions, it was apparent that operating
parameters such as temperature, CO2 loading, amine concentration and oxygen content
significantly altered the electrochemical response. The changes in the electrochemical
behaviour could be directly related to the impact that the operating conditions had on the
chemistry and the ionic composition of the aqueous solution. A Kent-Eisenberg type model,
explained in detail in Chapter 6, was employed to estimate the concentrations of these chemical
species in the bulk solution at each operating condition. Each ionic species has a different effect
on the corrosion of iron, with a strong dependence on the operating conditions especially the
amine concentration, pH (CO2 loading) and temperature. Results reported in this section aimed
to relate the Fe oxidation kinetics to the operating conditions. Details of the experimental
results are summarised in Appendix (V). The results were analysed and corrosion rate
P a g e | 245
measurements were conducted using Tafel plots, as described previously in the experimental
methods Chapter 4.
5.6.1. Effect of CO2 loading
CO2 loading is an important parameter affecting iron / steel oxidation rates in alkanolamines;
most of those commonly used in CO2 absorption plants are not intrinsically corrosive, but
become so on absorption of CO2, as found for MEA. Figure 5-49 shows the effect of CO2
loading on the voltammetric behaviour of iron in MEA+ H2O systems, demonstrating that
higher CO2 loading resulted in higher cathodic and anodic current densities, implying higher
corrosion rates. In short, increases in CO2 loading led to a significant rise in the corrosion rates.
As shown in (Figure 5-50), the corrosion rate of an iron electrode rotated at 1500 rpm (dE/dt =
10 mV s-1) in 30 wt% MEA + H2O (25 oC) in the absence of CO2 had a comparably negligible
corrosion rate of 0.278 mm y-1 which, increased to 6.189 mm y-1 at full CO2 saturation under
the same conditions. The impact of CO2 on the corrosion of iron in the MEA-H2O-CO2 system
can be explained by the absorption reaction of CO2 in MEA via reaction schemes (5. 7) – (5.
10) and the reaction of CO2 in water (5. 11) – (5. 13). As the CO2 loading increases, there are
higher amounts of hydrogen carbonate ion (HCO3-) and protonated amine (RNH3+) which then
subsequently dissociate to form the hydrogen ion H+. Enhancement in the hydrogen evolution
reaction (5. 14) with increasing concentrations of HCO3- and RNH3+ results in expediting the
kinetics of the anodic reactions, which in the active region primarily consists of the dissolution
of iron.
P a g e | 246
6 6
5 5
4
Ln (Current Density) / A m-2
Ln (Current Density) / A m-2

3
3
2
2
1
1
0
0 pH 8.01
pH 8.07 -1
-1 pH 8.43 -2
pH 8.52 pH 8.01
-2 pH 8.64 -3 pH 10.01
pH 8.78 pH 12.55
-3 -4
-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Figure 5-49: Effect of pH on polarisation behaviour of an iron RDE in 30 wt% MEA+ CO2, at
10 mV s-1; T = 25 oC, f = 1500 rpm.
The influence of CO2 loading was evaluated further by calculating the equilibrium
concentrations of the ionic species using the Kent Eisenberg model. It can be seen from (Figure
5-6) that increases in CO2 loading yields higher amounts of hydrogen carbonate ion and
protonated amine and thus hydrogen ions (Figure 5-51). This increase in H+ ions at higher CO2
loadings is also evidenced experimentally as the pH decreases with CO2 loading. Unloaded 30
wt% MEA+ H2O (25oC) solutions have a pH of 12.55 but the pH gradually lowers with
increasing CO2 loading, reaching a minimum of pH 8.01 at full saturation. This change in pH
shifts the MEA + H2O+ CO2 aqueous system from the ‘passive’ region in which corrosion is
thermodynamically unfavourable to the ‘active region’ where corrosion is predicted.

P a g e | 247
Predicted corrosion rate / mm y-1 7.0

6.0
5.0
4.0
3.0
2.0
1.0
0.0
8 9 10 11 12 13
pH
Figure 5-50: Effect of pH on oxidation / corrosion rates of an iron RDE in 30 wt% MEA+ CO2,
at 10 mV s-1 (T = 25 oC, f = 1500 rpm).
4
H+ concentration x 108 / M
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 5-51: Effect of CO2 loading on hydrogen ion concentration based on the Kent-Eisenberg
model for 30 wt% MEA solution and 25 oC.
5.6.2. Temperature effect
The experimental results showed, as expected, that over the range 25 and 80 oC, temperature
had a very significant role on the oxidation rate of iron in MEA-H2O-CO2 systems. On a typical
CO2 absorption plant, the absorber operates at ca. 40 oC and the heat exchanger at 80 oC.
P a g e | 248
The voltammograms (Figure 5-52 & Figure 5-55) for various concentrations of MEA showed
that increasing the temperature had a substantial effect on both the anodic and cathodic reaction
rates with increases in current densities at higher temperatures. For example, the corrosion rate
of an iron electrode rotated at 1500rpm (dE/dt = 10 mVs-1) in 30 wt% MEA + H2O + CO2 at
full saturation of CO2 increased from 6.189 mm y-1 at 25oC to 17.675 mm y-1 at 80oC. Similar
results were obtained at all rotation rates and MEA concentrations studied within this work
(Appendix V).
The increase in corrosion rate with temperature can be explained by the role of temperature in
the reaction kinetics as temperature generally expedites the rate of any reaction (Levenspiel,
1999). Elevated temperatures increase the rate of all processes involved in the corrosion
processes including the mass transport of the species, chemical reactions in the bulk solution
and the electrochemical reactions at the iron surface. In this particular case, reactions (5. 7) and
(5. 8) are accelerated with increasing temperature which in turn increases the amounts of
protonated amine (RNH3+) and hydrogen carbonate ion (HCO3-) in solution. The effect of
temperature on the equilibrium H+ concentration in solution, calculated by the Kent-Eisenberg
model is also shown in Figure 5-56. The increased amounts of RNH3+ and HCO3- ions result
in a shift in order to maintain the equilibrium of reactions (5. 9) and (5. 10), therefore producing
more H+ ions in the system. This increase in hydrogen ion concentration was also evidenced
experimentally by the reduction of solution pH with increasing temperature at full CO2
saturation (Table 5-4).

P a g e | 249
6
5
Ln ( Current Density / A m-2 )
4
3
2
1
18C
0 25C
-1 40C
60C
-2
80C
-3
-1.4 -1.3 -1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0
Figure 5-52: Effect of solution temperature on polarisation curves of iron RDE in 30 wt% MEA
+ H2O + CO2 at pH 7.24 - 8.01; 10 mV s-1, f = 1500 rpm.
20 20
5wt% 30wt%
18 10wt% 18 40wt%
16 20wt% 50wt%
30wt% 16
Corrosion rate / mm y-1
60wt%
14 14
12 12
10
10
8
8
6
6
4
4
2
2
0
15 25 35 45 55 65 75 85 0
15 25 35 45 55 65 75 85
Temperature / oC
Temperature / oC
Figure 5-53: Effect of solution temperature on the corrosion rate obtained on iron RDE in 5-60
wt% MEA + H2O + CO2 at pH 7.24 - 8.01; 10 mV s-1, f = 1500 rpm.
P a g e | 250
[H+] ion concentration x 108 / M 8
0
20 30 40 50 60 70 80
Temperature / oC
Figure 5-54: Effect of temperature on hydrogen ion concentration [H+] based on the Kent-
Eisenberg model for 30 wt% MEA solution and 25oC.
Table 5-4: Effect of solution temperature on the pH for 30 wt% MEA at full CO2 saturation.
Temperature/ oC pH
25 8.01
40 7.83
60 7.49
80 7.24
P a g e | 251
6 6
Ln (Current Density) / A m-2)

Ln (Current Density / A m-2 )
5 5
4 4
3 3
2
2
1
1
0
0 25C -1 25C
-1 40C -2 40C
-2 5 wt% MEA 60C -3 60C
80C 10 wt% MEA 80C
-3 -4
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
6 6

5 5
4 4
3 3
2 2
1 1
0 25C 0
40C 25C
-1 -1 40C
-2 60C 60C
20 wt% MEA -2 40 wt% MEA
80C 80C
-3 -3
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
6 6
5 5
4 4
3
3
2
2
1
1
25C 0 25C
0 40C -1 40C
-1 60C 60C
-2
50 wt% MEA 80C 60 wt% MEA 80C
-2 -3
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Figure 5-55: Effect of solution temperature on polarisation curves of iron RDE in 5-60wt%
MEA + H2O + CO2 at pH 7.24-8.01; 10 mV s-1, f = 1500 rpm.
P a g e | 252
5.6.1. Effect of MEA Concentration
The effects of amine concentration on iron / steel corrosion rates have been the subject of a few
report (Kohl and Riesenfeld, 1985), (Kohl and Nielsen, 1997), (Veawab et al., 1999),
(Kladkaew et al., 2009a) and (Kittel, 2010). While (Kladkaew et al., 2009b) & (Kittel, 2010)
used MEA alone, (Veawab et al., 1999) & (Nainar and Veawab, 2009) also conducted studies
on other commercially-available amines, such as AMP, DEA and MDEA. Based on the
previously conducted studies on MEA, 20-30 wt% has been considered to be the maximum
concentration range, above which the corrosion rate was thought to increase drastically. In
order to verify this, seven experiments were conducted in 5-60 wt% MEA, fully saturated in
CO2 at varying temperatures and rotation rates (25-80 oC, 0-1500 rpm). The resulting
polarization curves are presented in Figure 5-56 – Figure 5-59. For all the experiments
conducted, the open circuit potential measured at the start of each experiment was very close
(± 0.03 V) to the observed corrosion potential obtained from the polarization curves. Based on
the Tafel plots, it was evident that MEA concentration had a significant effect on anodic current
densities. According to Figure 5-60, MEA concentrations affected corrosion rates in two
distinctive ways:
i) At lower concentrations, corrosion rates increased with increasing amine concentration,
for concentrations < 30 wt%, for all temperatures and mass transport rates studied. This
initial increase in corrosion rate with amine concentration was also evident from the
polarisation curves (Figure 5-56 and Figure 5-59). As the MEA concentration was
increased from 5 wt% to 30 wt%, anodic current densities increased in the active region
of the polarisation curve. The increase in corrosion rate with MEA concentration can be
explained by taking into account the vapour-liquid equilibrium of the MEA-H2O-CO2
system. Based on simple stoichiometry of reactions (5. 7) – (5. 10) and the equilibrium
concentrations of the chemical species in solution calculated by the Kent-Eisenberg

P a g e | 253
Model, there was a corresponding increase in the total amount of CO2 absorbed with
increasing amine concentration, i.e. increased CO2 loading with MEA concentration.
This increased CO2 loading directly affected the predicted concentrations of protonated
amine (RNH3+) and hydrogen carbonate ion (HCO3-), and H+ ions (Figure 5-61).
ii) As shown by Figure 5-60, corrosion rates increased dramatically between 20-30 wt% and
reached a peak at 30 wt%, above which it decreased gradually with increasing MEA
concentration. This type of behaviour was also reported by (Veawab et al., 1999) with
AMP at concentrations > 3 kmol m-3 (ca. 18.2 wt%), but not for MEA systems. Veawab
et al studied the corrosion behaviour in MEA only at concentrations between 1-5 kmol m-
3
(ca. 6-30 wt%) and also reported increasing corrosion rates with concentration, as in
Figure 5-60 for concentrations ≤ 30 wt%.
This result is in complete contradiction with recommendations from the literature Kohl and
Nielsen, 1997 that 20 to 30 % should be the maximum MEA concentration, due to concerns
about corrosion. Veawab et al attributed the decreasing corrosion rates in AMP solutions at
concentrations ≥ 18.2 wt% to the decreasing amount of water available. The decrease in
corrosion rates at the maximum corrosion rates could result from two separate causes. Firstly,
H2O acts as an additional oxidising agent in the system via (5. 39):
2() * + 2, - ↔ () + 2*(- (5. 39)
Secondly, this retardation in corrosion rates at high amine concentrations could also be due to
mass transport limitations as the solutions become viscous with increasing concentration. This
will be discussed in further detail in Chapter 6 & 7.
Depending on the amine concentration, some differences were evident in the passive region at
electrode potentials ˃ -0.40V. At potentials ca. > -0.40 V, measured current densities from
secondary oxidation reactions resulted in accumulation of a passivating layer. At lower

P a g e | 254
temperatures (25-40 °C), it was difficult to identify the effect of MEA concentration on passive
layers, while at higher temperatures (60-80 °C), current densities in the passive region followed
the same trend as in the active region; i.e. increasing passivating current densities with
increasing concentration for ≤ 30 wt% MEA and decreasing passivating current densities with
increasing concentration for ≥ 30 wt% MEA. Kittel et al reported the effect of MEA
concentration (20, 30 & 40 wt% at a temperature of 80oC) on steel corrosion rates and on
passive layer formation but were unable to do so due to low reproducibility of their
experimental results.
P a g e | 255
6 6
5 5

4
4
3
3
2
2
1
1
0
-1 5wt% 0 30wt%
40wt%
10wt% -1
-2 50wt%
20wt%
-3 -2 60wt%
30wt%
-4 -3
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Figure 5-56: Effect of MEA concentration on the polarisation curves obtained on iron RDE in
MEA + H2O + CO2 at pH 8.01 and T = 25oC; 10 mVs-1, f = 1500 rpm.
6 6
5 5
4
Ln (Current Density) / A m-2 )
4
3
3
2
2
1
1
0 30wt%
0 30wt
-1 40wt%
%
50wt%
-1 5wt% -2 60wt%
-2 -3
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
+ CO2 at pH 7.83 and T = 40 oC; 10 mV s-1, f = 1500 rpm.
P a g e | 256
6 6
5 5

4 4
3 3
2 2
1 1
0 0
-1 -1 30wt%
5wt%
10wt% 40wt%
-2 20wt% -2 50wt%
30wt% 60wt%
-3 -3
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
+ CO2, at 10 mV s-1 ; pH 7.49, f = 1500 rpm, T = 60oC.
6 6
5
5
Ln (Current Density) / A m-2 )
4
4
3
3 2
1
2 5wt%
10wt% 0 30wt%
20wt% 40wt%
1 30wt% -1 50wt%
60wt%
0 -2
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
+ CO2 at pH 7.24 and T = 80 °C; 10 mV s-1, f = 1500 rpm.
P a g e | 257
20 25 oC 20
1500 rpm 25 oC 960 rpm
40 oC

40 oC
15 60 oC
80 oC 15 60 oC
80 oC
10 10
5 5
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
[MEA] / wt% [MEA] / wt%
16 16
25 oC 540 rpm 25 oC 240 rpm
14 14
40 oC 40 oC
12 60 oC 12 60 oC
10 80 oC 10 80 oC
8 8
6 6
4 4
2 2
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
15 15
25 oC 60 rpm 25 oC 0 rpm
40 oC 40 oC
60 oC 60 oC
10 80 oC 10 80 oC
5 5
0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Figure 5-60: Effect of MEA concentration on the corrosion rate obtained on iron RDE in MEA
+ H2O + CO2 at pH 7.24 – 8.01 and T = 25-80oC; 10 mV s-1, f = 0-1500 rpm.
P a g e | 258
[HCO3-] ion concentration / M 6
0
10 20 30 40 50 60
MEA Concentration / wt%
5
[RNH3+] ion concentration / M
0
10 20 30 40 50 60
2.5
[H+] ion concentration x 108 / M
1.5
0.5
0
10 20 30 40 50 60
Figure 5-61: Effect of MEA concentration on the species concentrations of the aqueous bulk
solution at chosen CO2 loading of 0.6 CO2 loading / mol CO2 (mol amine)-1, T = 25 oC,
calculated by the Kent-Eisenberg Model (Kent and Eisenberg, 1976).
P a g e | 259
5.6.2. Effect of Mass Transport
Surprisingly, the effects of mass transport conditions on the corrosion of Fe in aqueous
alkanolamine-CO2 systems, does not appear to have been reported. However, the polarisation
curves in Figure 5-62 show that increasing RDE rotation rates / mass transport rates affected
both cathodic and anodic current densities and hence accelerated corrosion rates significantly.
For example, the corrosion rate of a stationary Fe electrode in 30wt% MEA + H2O + CO2 at
full saturation of CO2 and 25 °C was 4.345 mm y-1 and increased to 6.189 mm y-1 when the
RDE rotated at 1500 rpm. Similar results were obtained at all rotation rates, temperatures and
MEA concentrations studied in this project.
The increase in corrosion rates can be attributed partly to the increased mass transport of the
oxidising species, resulting from chemical reactions in the bulk solution, to the surface of the
iron. This is due to the rate of the dissolution reaction being limited by mass transport rates of
H+ and HCO3- ions from the bulk solution to the surface of the iron electrode.
The introduction of rotation also increased corrosion rates by preventing or decreasing the rate
of FeCO3 precipitation, which is predicted thermodynamically to occur when the local
concentrations of Fe2+ and CO32- ions exceed the solubility product of FeCO3. As the rotation
rate increases, this local concentrations of the soluble species near the electrode surface
decreases, ultimately preventing the formation of the protective FeCO3 film.

P a g e | 260
6
1500rpm
5 960rpm
540rpm
4 240rpm
60rpm
3 0rpm
2
1
0
-1
-2
-1.4 -1.3 -1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0
Figure 5-62: Effect of rotation rate on polarisation curves of iron RDE in 30 wt% MEA + H2O
+ CO2, at 10 mV s-1; pH 8.01, f = 0-1500 rpm, T = 25 oC.
6
2
30 wt%
1 20wt %
10 wt%
5 wt%
0
0 1 2 3 4 5
(Rotation rate / s-1)0.5
Figure 5-63: Effect of rotation rate on corrosion rates of iron RDE in 5-30 wt% MEA + H2O +
CO2 at pH 8.01; 10 mV s-1, f = 0-1500 rpm, T = 25 oC.
P a g e | 261
The effect of oxygen on corrosion rates of iron / steel in alkanolamine solvents used for CO2
capture has been much debated. Initially and surprisingly, it was assumed that oxygen did not
participate in the corrosion process (Kohl and Nielsen, 1997). However, several publications
(e.g. Veawab et al., 1999) & (Kittel, 2010) reported that, as expected, the presence of oxygen
increased corrosion rates in amine solutions.
Figure 5-64 shows the effects of dissolved oxygen on the polarisation curves as functions of
temperature. It can be seen that at 25 °C, there was a significant increase in both anodic and
cathodic current densities in the presence of oxygen. However, at higher temperatures ≥ 40 C,

°
there was no obvious effect of oxygen on cathodic current densities, while anodic current
densities were slightly higher with aqueous MEA solutions in which oxygen was present. In
the passive region, current densities were generally higher when oxygen was present.
From the analysis of corrosion rates, it was evident that the presence of oxygen induced higher
corrosion rates. For example, the corrosion rate of iron in an anoxic 30 wt% MEA system at
25 oC had a corrosion rate of 5.761 mm y-1, which increased to 6.189 mm y-1 in the presence
of oxygen. Veawab et al attributed the corrosion rate increase they measured at higher
concentrations of O2 to the formation of Fe(OH)2 and Fe(OH)3 via reactions (5. 40) and (5. 41):
2/, + 2() * + *) → 2/,(*()) (5. 40)
2/,(*()) + 2() * + 1/2*) → 2/,(*()4 (5. 41)
Whereas additional oxidant would be expected to increase cathodic currents, apparent anodic
(net) currents would be expected to decrease, rather than increase, as measured. One possible
reason could be that the soluble H2O2 intermediate of oxygen reduction ultimately to water,
P a g e | 262
reacts homogeneously with Fe2+ species, precipitating ‘Fe(OH)3’ in a less compact form than
Fe2O3, so enabling slightly higher net oxidation current densities:
2() * + *) + 2, - → () *) + 2*(- (5. 42)
/, → /, )2 + 2, - (5. 43)
2/, )2 + () *) + 4*(- → 2/,(*()4 (5. 44)
() *) + 2, - → 2*(- (5. 45)
The electrons for reduction of the hydrogen peroxide intermediate by the reaction (5. 45) could
also be provided by oxidation reactions of the alkanolamine, with its own intermediates being
oxidised and so contributing to the net increase in iron oxidation current densities:
-P c -7 L
9 − <() :() 9 − <() : 2 () 9 − <( = :() (5. 46)
-P c -7 L
9 − <( = :() 9 − <( = : 2 () + () * 9<( = * + :(4 (5. 47)
9<( = * + () * → 9<**- + 3(2 + 2, - (5. 48)

P a g e | 263
Although MEA itself was not oxidised on iron for the range of potentials used, partial reduction
of oxygen to peroxide that diffused away from the electrode, reacting homogeneously with
MEA and producing intermediates, such as from reactions (5. 46 – 5. 48) above, would have
caused the diffusion of those intermediates to the electrode. Their further oxidation, could have
resulted in the measured net increase in current densities in the presence of oxygen.
While this may be the case for the passive region, the increase in corrosion rates at active
dissolution potentials was most likely due to oxygen acting as an additional oxidant, by reaction
(5. 49):
2() * + *) + 4, - → 4*(- (5. 49)

P a g e | 264
6 6
25C N2 1500rpm 40C N2 1500rpm
25C O2 1500rpm 40C O2 1500rpm
5 5
ln (Current density / A m-2 )

4 4
3 3
2 2
1 1
0 0
-1 -1
-2 -2
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
6 6
60C N2 1500rpm
60C O2 1500rpm
5
5
4
4
3
3 2
1
2
0
1
-1 80C N2 1500rpm
80C O2 1500rpm
0 -2
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0
Figure 5-64: Effect of oxygen on the polarisation curves obtained on iron RDE in 30wt% MEA
+ H2O + CO2 at pH 7.24-8.01 and T = 25-80oC; 10 mV s-1, f = 1500 rpm.
P a g e | 265
5.7. Summary & Conclusions
The electrochemical behaviour of Fe was investigated using a rotating Fe disc electrode (RDE)
in aqueous MEA-CO2 solutions using voltammetry, as functions of potential sweep rate
(dE/dt), rotation rate (f), positive potential limit (Ea), solution temperature, CO2 loading, MEA
concentration and oxygen content. In the absence of CO2, aqueous MEA had a pH of 12.55, so
Fe oxidation resulted in passivation, probably due to Fe2O3 / Fe(OH)3 formation. However,
when CO2 was absorbed, the pH decreased to ca. 8, causing iron to dissolve at potentials in the
range ca. -0.8 to -0.5 V (AgCl|Ag), above which it passivated, as predicted thermodynamically
from potential-pH diagrams. Cyclic voltammograms exhibited an anodic peak on negative-
going potential sweeps from potentials corresponding to the passivation; this probably resulted
in net currents due to reductive dissolution of the iron (III) (hydro-)oxide, followed by oxidative
dissolution of the Fe substrate.
Kinetic analysis of the experimental results showed that temperature and CO2 loading were the
most significant parameters affecting the corrosion of iron or steel in MEA-CO2-H2O. For
example, corrosion rates of iron in fully CO2-saturated MEA solutions increased from 6.19 mm
y-1 at 25oC to 17.68 mm y-1 at 80oC. Whilst, the corrosion rate in unloaded MEA solutions at
25oC had a comparably negligible corrosion rate of 0.28 mm y-1. Corrosion rates increased with
increasing CO2 loading, due to the accompanying decrease in pH and increased concentration
of species such as RNH3+, HCO3- and H+. As an additional oxidant, dissolved oxygen enhanced
Fe corrosion rate.
At amine concentrations < 30 wt%, corrosion rates increased with increasing concentration,
while at concentrations > 30 wt%, corrosion rate decreased gradually, implying more complex
behaviour than for Fe in CO2-H2O alone. As solution viscosities (µ) increased at MEA
concentrations > 30 wt%, mass transport rates of HCO3- and H+ ions and O2 to the Fe surface
P a g e | 266
and FeII away from it, would have decreased, as predicted by the Stokes-Einstein equation:
Di = kBT (6π rη ) . Such high concentrations would also have affected dissolved reactant
activities, including that of H2O which also acts as an oxidant for Fe.
As expected for a system exhibiting dissolution-precipitation-passivation behaviour, current
densities due to oxidative dissolution increased with increasing mass transport rate coefficients,
which decreased FeII concentrations at the Fe|solution interface, minimising or even preventing
precipitation of FeCO3.
To discriminate between partial current densities resulting in FeII dissolution from those
leading to adsorbed products, a flow reactor was used in batch recycle mode with a stirred tank
reservoir, from which solution samples were taken for subsequent analysis by ICP-OES to
determine dissolved FeII concentrations in the amine solution after constant potential
electrolyses as functions of the process parameter. This enabled the charge yield for dissolution
to be determined by relating the amount of charge passed to the amount of iron dissolved,
enabling more accurate estimates of corrosion rates.

P a g e | 267
5.8. References
BARD, A.J., PARSONS R., JORDAN J. (Eds.), 1985. Standard Potentials in Aqueous
Solution, Marcel Dekker, New York p.848?.A.J.Bard, R.Parsons and J.Jordan (eds.),
'Standard Potentials in Aqueous Solution', Marcel Dekker, New York, 1985.
BIRCH, A. S. 2002. Ph.D. Thesis, Modelling Carbon Dioxide Induced Corrosion of Oil
Pipelines. Imperial College of Science, Technology and Medicine
CASTRO, E. B., VALENTINI, C. R., MOINA, C. A., VILCHE, J. R. & ARVIA, A. J. 1986.
The influence of ionic composition on the electrodissolution and passivation of iron
electrodes in potassium carbonate-bicarbonate solutions in the 8.4–10.5 pH range at
25°C. Corrosion Science, 26, 781-793.
CASTRO, E. B., VILCHE, J. R. & ARVIA, A. J. 1991. Iron dissolution and passivation in
K2CO3-KHCO3 solutions. rotating ring disc electrode and XPS studies. Corrosion
Science, 32, 37-50.
CLERBOIS, F. & MASSART, J. 1962. Comportement anodique des metaux en milieu aqueux-
I. Utilisation des courbes potentiostatiques a la determination des mecanismes de
corrosion. Corrosion Science, 2, 1-20.
FAJARDO, V., BROWN, B., YOUNG, D., NE & SRDJAN Study of the Solubility of Iron
Carbonate in the Presence of Acetic Acid Using an EQCM. NACE International.
HAMM, D., OGLE, K., OLSSON, C. O. A., WEBER, S. & LANDOLT, D. 2002. Passivation
of Fe–Cr alloys studied with ICP-AES and EQCM. Corrosion Science, 44, 1443-1456.
HAYNES, W. M. & LIDE, D. R. 2011. Handbook of chemistry and physics: a ready-reference
book of chemical and physical data. CRC Press, 92nd. Edition, Boca Raton, Florida.
J. KITTEL, E. F., B. VUILLEMIN, S. GONZALEZ, F. ROPITAL AND R. OLTRA 2010.
Corrosion in alkanolamine used for acid gas removal: From natural gas processing to
CO2 capture. Materials and Corrosion, 63, 223-230.
KENT, R. L. & EISENBERG, B. 1976. Better Data for Amine Treating. Hydrocarbon
Processing, 55, 87.
KLADKAEW, N., IDEM, R., TONTIWACHWUTHIKUL, P. & SAIWAN, C. 2009a.
Corrosion Behavior of Carbon Steel in the Monoethanolamine−H2O−CO2−O2−SO2
System. Industrial & Engineering Chemistry Research, 48, 8913-8919.
KLADKAEW, N., IDEM, R., TONTIWACHWUTHIKUL, P. & SAIWAN, C. 2009b.
Corrosion Behavior of Carbon Steel in the Monoethanolamine−H2O−CO2−O2−SO2
System: Products, Reaction Pathways, and Kinetics. Industrial & Engineering
Chemistry Research, 48, 10169-10179.
KOHL, A. & NIELSEN, R. 1997. Gas Purification (5th edition). Houston: Gulf Publishing
LEVENSPIEL, O. 1999. Chemical Reaction Engineering, 3rd Edition, Wiley & Sons, Inc., New
York.
NAINAR, M. & VEAWAB, A. 2009. Corrosion in CO2 Capture Process Using Blended
Monoethanolamine and Piperazine. Industrial & Engineering Chemistry Research, 48,
9299-9306.
P a g e | 268
SCHMUTZ, P. & LANDOLT, D. 1999. Electrochemical quartz crystal microbalance study of

the transient response of passive Fe 25Cr alloy. Electrochimica Acta, 45, 899-911.
SIMARD, S., MENARD, H. & BROSSARD, L. 1998. Localized corrosion of 1024 mild steel
in slightly alkaline bicarbonate solution with Cl− ions. Journal of Applied
Electrochemistry, 28, 151-160.
THOMAS, J. G. N., NURSE, T. J. & WALKER, R. 1970. Anodic Passivation of Iron in
Carbonate Solutions. British Corrosion Journal, 5, 87-92.
VALENTINI, C. R., MOINA, C. A., VILCHE, J. R. & ARVIA, A. J. 1985. The
electrochemical behaviour of iron in stagnant and stirred potassium carbonate-
bicarbonate solutions in the 0–75°C temperature range. Corrosion Science, 25, 985-
993.
VEAWAB, A., TONTIWACHWUTHIKUL, P. & CHAKMA, A. 1999. Corrosion Behavior
of Carbon Steel in the CO2 Absorption Process Using Aqueous Amine Solutions.
P a g e | 269
6. Chapter 6
Mechanistic Model for the Prediction of

Corrosion of Iron in MEA +H2O+ CO2
solutions
As discussed in the literature review (Chapter 3), the corrosion of Fe in MEA-CO2-H2O systems
has been studied by a few researchers, but the underlying mechanism is not very well
understood. This has led to electrochemical reactions being assigned without appropriate
verification methods being carried out. The arbitrary and sometimes uninformed assignment of
chemical species as principal oxidising agents can be detrimental to the implementation of
corrosion control techniques. Therefore, it was also important to identify the oxidising agents
which participate in the corrosion reaction, the influence that operating parameters have on the
concentration of such species and the resulting impact on the corrosion rate.
A mechanistic model for the uniform corrosion of Fe in aqueous MEA-CO2 solutions was
developed, describing the electrochemical reactions at the metal surface and the chemical
reactions occurring in the bulk solution, aiming to predict corrosion rates as well as the
concentrations of the solution species at various operating conditions and hence corrosion rates.
P a g e | 270
6.1. Integrated Model for Fe corrosion in MEA+ H2O+CO2
During corrosion of iron in aqueous MEA loaded with CO2, chemical, electrochemical and
transport processes occur simultaneously at the Fe | solution interface.
Chemical Reactions in the Bulk Solution
During the absorption reaction of CO2 in MEA, the following chemical reactions occur:
- Dissolution of carbon dioxide
!"#(%) ↔ !"# (6. 1)
2)*+# + +# " + !"# ↔ )*+!""- + )*+. / + +# " (6. 2)
- Dissociation of protonated amine ion
- )*+. / ↔ )*+# + +/ (6. 3)
)*+!""- + +# " ↔ )*+# + +!". - (6. 4)
- Hydrolytic ionization of CO2(aq)
!"# (12) + +# " ↔ +/ + +!". - (6. 5)
- 31456 7899:;8148:<
+# " ↔ +/ + "+- (6. 6)
- Dissociation of bicarbonate ion
+!". - ↔ !". #- + +/ (6. 7)

P a g e | 271
When the chemical reactions are very fast compared to the other processes which are occurring
simultaneously, the chemical equilibrium of the system is maintained. However, when the
chemical reactions progress relatively slowly, the faster processes such as the mass transport or
electrochemical reactions can lead to a local offset in equilibrium. The rate of the chemical
reactions occurring can also change the rate of the electrochemical reactions occurring on the
surface of the iron. An example of this is when there is a high local concentration of species
such as Fe2+ and CO32- ions and the solubility limit is exceeded, leading to the precipitation of
FeCO3. The model took into account the homogeneous chemical reactions (6. 1)- (6. 7)
occurring in the solution, but it excluded heterogeneous chemical reactions such as the
precipitation of FeCO3. The model also made an assumption that all the dissolved CO2 within
the system are present as Carbon (IV) species.
Electrochemical Reactions at the Iron Surface
The model assumed that corrosion rates were spatially uniform across the iron surface, which
oxidised, driven primarily by the hydrogen evolution reactions:
2H / + 2e- → H# (6. 8)
2H# O + 2e- → H# + 2OH - (6. 9)
There are three well known steps for the hydrogen evolution reaction commonly known as the
Volmer (1), Heyrovsky (2) and Tafel (3) steps.
1. Volmer ∶ +/ + 5 - ⟶ + JKL (6. 10)
2. Heyrovsky: + JKL + +/ + e- ⟶ +#(%) (6. 11)
3. Tafel: + JKL + + JKL ⟶ +#(%) (6. 12)

P a g e | 272
As discussed in the Literature Review there are two widely accepted pathways for the hydrogen
evolution process. They are the discharge, chemical recombination (Volmer-Tafel mechanism)
and the discharge, electrochemical desorption (Volmer-Heyrowsky) mechanism. (Weber and
Peria, 1967; Park, 1975). Whilst the hydrogen evolution reaction proceeds in two steps with
one intermediary reaction (Weber and Peria, 1967; Eastman and Nathan, 1975; Park, 1975), the
mechanistic model makes an assumption that it occurs via a single step two-electron process
(reaction 6.8).
The concentration of H+ ions was increased by the dissociation reactions of RNH3+ (6. 3) and
HCO3- (6. 7) resulting from the reactions of CO2 in MEA and H2O.
It was also clear from the cyclic voltammetric results in Chapter 5 that the reduction of oxygen,
if present, also played a significant role in the corrosion of iron in MEA-H2O-CO2 systems, so
it also was included in the model:
2H# O + O# + 4e- → 4OH - (6. 13)
The results obtained from the combined use of the RDE and EQCM (Chapter 5) and electrode
potential-pH diagrams indicated that the dissolution of iron occurs primarily within the
potential range where only FeII is predicted thermodynamically. Therefore, it was assumed in
the model that the electrochemical dissolution of iron in MEA- CO2-H2O systems according to
equation (6. 14) is the dominant anodic reaction:
Fe ↔ Fe#/ + 25 - (6. 14)
The model was focused on the anodic dissolution region of the cyclic voltammetry experiments
and did not take into account the passive region resulting from adsorption of products such as
iron(III) (hydr-)oxide at higher potentials.

P a g e | 273
6.1.1. Model scheme
The model scheme, shown schematically in Figure 6-1 below, illustrates how both chemical
and electrochemical reactions occur simultaneously at the iron metal | amine solution interface.
It includes five chemical reactions typical of the MEA-H2O-CO2 absorption process (Table 6-
1): i) dissociation of the protonated amine (RNH3+), ii) hydrolysis of carbamate (RNHCOO-),
iii) hydrolysis of CO2, iv) dissociation of water, v) dissociation of bicarbonate ions. 11 chemical
species were considered in the solution, namely MEA (RNH2), carbamate (RNHCOO-),
protonated amine (RNH3+), un-dissociated water (H2O), dissolved CO2 (CO2(aq)), bicarbonate
ion (HCO3-), carbonate ions (CO32-), hydrogen ions (H+), hydroxyl ions (OH-), dissolved
oxygen (O2) and ferrous ions (Fe2+).
At the iron ǀ solution interface, electrons produced by the oxidation of iron to ferrous ions by
reaction (6. 14) were consumed at the corrosion potential at an exactly equal and opposite rate
by reduction of: hydrogen ions (H+), water and dissolved oxygen.

P a g e | 274
Aqueous MEA-CO2 solution
Chemical reactions occurring in the bulk solution

)*+. / ↔ )*+# + + /
)*+!""- + +# " ↔ )*+# + +!". -
!"# (12) + +# " ↔ + / + +!". -
+# " ↔ + / + "+-
+!". - ↔ !". #- + + /
H2 2OH- H2 2OH-
Fe2+
H+ 2H2O 1/2O2 H2O
2e-
Electrochemical reactions at Fe surface
Iron
Fe
Figure 6-1: Schematic representation of the corrosion process in aqueous MEA-CO2 solutions,
including the chemical reactions occurring at the iron surface
P a g e | 275
Table 6-1: Chemical and electrochemical reactions represented in the mechanistic corrosion
model
Type of Reaction Description Reaction
reaction
1. Dissociation of protonated
Chemical )*+. / ↔ )*+# + +/
Reactions
amine.
2. Hydrolysis of carbamate. )*+!""- + +# " ↔ )*+# + +!". -
3. Hydrolysis of carbon dioxide. !"# (12) + +# " ↔ +/ + +!". -
4. Water dissociation +# " ↔ +/ + "+-
5. Dissociation of hydrogen +!". - ↔ !". #- + +/
carbonate ions
Electrochemical 1. Dissolution of iron X5 ↔ X5 #/ + 25 -
Reactions
2. Reduction of hydrogen ions 2+/ + 25 - → +#
3. Reduction of water 2H# O + 2e- → H# + 2OH -
4. Reduction of oxygen 2+# " + "# + 45 - → 4"+-

P a g e | 276
6.1.2. Model Structure and Development
The model was structured with three main components: inputs, the thermodynamic and
electrochemical models followed by the outputs, as demonstrated in Figure 6-2. The inputs
included information on the operating conditions of the aqueous MEA-CO2 solution such as
temperature, CO2 loading and pH, MEA concentration, mass transport conditions (electrode
rotation rate) and oxygen content.
Firstly, concentrations of species at the iron | solution interface were established using
thermodynamic vapour-liquid equilibrium (VLE) principles of a Kent-Eisenberg based model,
assuming ideal behaviour, so implying that the activity coefficients of all the chemical species
were unity. The thermodynamic model integrated VLE data from literature, mass balances for
MEA and carbon (IV) species, as well as the principle of electro-neutrality for charged species
in solution.
Once the concentrations for each of the species had been determined, based on the parameters
from the ‘input’, the electrochemical model for corrosion of iron in the aqueous environment
could then be established, incorporating the oxidation of iron coupled to reduction of the
oxidising agents at the iron surface. Standard equilibrium potentials (E0O/R) for all
electrochemical reactions considered were calculated from Gibbs energies of formation (Allen
Bard, 1985). Concentrations of species involved in electrode reactions enabled equilibrium
potentials (EO/R) to be calculated using Nernst equations, so overpotentials (η) could be defined
for each reaction as a function of applied potential (E). Where available for relevant
compositional conditions, values for exchange current densities (j0) and anodic (αa) and
cathodic (αc) transfer coefficients were taken from the literature. The mass transport
contributions to the corrosion behaviour were established by using the Levich equation for a
rotating disc electrode or the experimentally-derived laminar flow mass transport correlation
P a g e | 277
for the flow-through reactor. The model output included the speciation of the amine solution
and current density – electrode potential data defined by Butler-Volmer equations (6. 15) for
each of the electrochemical reactions (i), so enabling corrosion rates to be predicted.
⎡ ⎧α F ⎫ ⎧ −α F ⎫⎤
ji = j0,i ⎢exp ⎨ a ,i ηi ⎬ − exp ⎨ c ,i ηi ⎬⎥ (6. 15)
⎣ ⎩ RT ⎭ ⎩ RT ⎭⎦
P a g e | 278
Inputs
• Solution temperature
• MEA Concentration
• CO2 loading in solution
Thermodynamic Model of MEA- CO2-H2O
• Based on the Kent-Eisenberg model

• Speciation in solution
• Mass balance for Carbon
• Mass balance for MEA
• Charge balance (Electroneutrality of charged species)
Electrochemical Model for Fe in MEA- CO2-H2O
• Nernst Equation
• Kinetics of Fe oxidation and reduction of oxidising
agents (Butler-Volmer)
• Levich Equation
• Mixed potential theory
Outputs
• Speciation in the bulk solution

• Polarisation curves
• Corrosion rate
Figure 6-2: Flow diagram showing the model structure

P a g e | 279
6.1.3. Thermodynamic Modelling of Aqueous MEA-CO2 system
Thermodynamic modelling of the MEA-CO2-H2O system enabled the speciation and redox
behaviour to be predicted as functions of process parameters. There are a few experimental
techniques that could be employed to measure the vapour-liquid (VLE) data for the absorption
of CO2 in amine systems (Tong, 2012), however they are tedious and time-consuming. In
addition, experimental techniques do not provide accurate data under certain conditions, such
as low CO2 loading, high system temperatures and very high amine concentrations, due to
analytical constraints (Chakma and Meisen, 1990). Therefore, a few mathematical models have
been developed to obtain accurate information on the solubility of CO2 in aqueous MEA
solutions over the full range of operating conditions required for the design of the equipment
used in alkanolamine plants.
Different approaches to thermodynamic modelling have been proposed by researchers to
correlate experimental VLE data, the most popular of which include: i) Kent-Eisenberg model
(Kent and Eisenberg, 1976), ii) model of (Edwards et al., 1978), iii) Deshmukh-Mather model
(Deshmukh and Mather, 1981) iv) Electrolyte-NRTL (e-NRTL) model (Austgen et al., 1989),
v) SAFT-VR model (MacDowell, 2010).
The Kent and Eisenberg model, the first VLE model for alkanolamine absorption of CO2, was
adopted because of its computational simplicity. It is a solubility model, which follows a
regression-based approach, based on equilibria of the ionic species in the liquid phase. The
equilibrium constants were fitted to available CO2 partial pressure and vapour-liquid
equilibrium data for the absorption of CO2 in alkanolamines. The model could then be used to
predict the solubility of CO2 in amine solutions even at operating conditions outside of the range
of existing data.
P a g e | 280
Table 6-2: Equilibrium expressions for chemical reactions
Reaction Description Equilibrium Constants
1. Dissociation of protonated amine )*+# [++]

YZ =
[)*+. / ]
)*+# [+!". - ]
2. Hydrolysis of Carbamate Y# =
[)*+!""- ]
3. Hydrolysis of carbon dioxide +/ [+!". - ]

Y. =
[!"# ]
4. Water dissociation
Y^ = +/ ["+- ]
5. Dissociation of bicarbonate ion +/ [!". #- ]

Y_ =
[+!". - ]
6. Amine mass balance àb = )*+# + )*+. / + [)*+!""- ]
7. CO2 mass balance àb cdef = !"# + +!". - + !". #-
+ [)*+!""- ]
8. Charge Balance )*+. / + +/
= "+- + +!". -
+ 2 !". #- + [)*+!""- ]
The equilibrium between the solution species is governed by the reactions (1-5) in Table 6-1
and the corresponding equilibrium constants (Ki) are expressed in Table 6-2. (Austgen et al.,
1989) noted that other side reactions such as degradation reactions may also take place
P a g e | 281
simultaneously; however, rates of these side reactions are negligible on a laboratory scale so
they were ignored within this model.
Kent and Eisenberg assumed that the MEA-H2O-CO2 system behaves ideally (activity
coefficient of unity) and the equilibrium constants (Ki) are dependent only on the temperature
of the system. The equilibrium constants used within this model were represented by the
empirical equation:
g<Yh = 1h i -Z + jh g:ki + ;h i + 7h (6. 16)
Where ai - di are constants and the temperature unit is Kelvin. Values of the constants were
taken from literature and are listed in Table 6-3. Mass balances of the MEA species (reaction
7) and the carbon (IV) species (reaction 8) were included, as well as a charge balance (Equation
9) in order to maintain the system’s electroneutrality.
Table 6-3: Temperature dependent equilibrium constants for chemical reactions 1-5
Parameter ai bi ci di Source
K1 -17.3 0 0.05764 -38.846 (Hamborg and Versteeg, 2009)
K2 -1545.3 0 0 2.151 (Haji-Sulaiman and Aroua, 1995)
K3 -12092.1 -36.7816 0 235.482 (Edwards et al., 1978)
K4 -13445.9 -22.4773 0 140.932 (Edwards et al., 1978)
K5 -12431.7 -35.4819 0 220.067 (Edwards et al., 1978)
6.1.4. Electrochemical model for Iron in MEA-CO2-H2O
As discussed previously, the electrochemical reactions occurring at the iron surface are the
dissolution of iron and the reduction of the oxidising agents such as H+, H2O and O2. The
oxidising agents were identified from the results of the cyclic voltammetry experiments in
chapter 5 which clearly indicated that the rate of the electrochemical reactions at the iron
P a g e | 282
surface depend on the applied electrode potential, the concentrations of the species at the
interface, as well as the operating conditions such as temperature and pH.
The mechanistic model was tested by simulating anodic and cathodic polarisation curves based
on the pre-calculated concentrations of the species involved in the corrosion reactions. This was
achieved by varying the electrode potential from the equilibrium potentials (EO/R) and hence
the overpotentials (η) for the reactions considered, and predicting the resulting current densities
(j) from the Butler-Volmer equations for mixed charge transfer-mass transfer control by
equation (6. 17):
rr st wx st
mn opq - opq y
uv uv
l = zn x st z wx st
(6. 17)
Z/ opq r - n opq y
z{,u uv z{,} uv
Where
~ = a − ae/Å (6. 18)
The Butler-Volmer equation (6. 17) was expressed for each of the electrochemical reactions
occurring at the iron surface; e.g. for the anodic iron dissolution reaction:
mn,sÖfÑ opq Ür fÑ ásÖfÑ -opq Üy fÑ ásÖfÑ

lÇÉ fÑ = sÖ
zn
sÖ
(6. 19)
Z- sÖfÑ opq -Üy fÑ ásÖfÑ
z{,} sÖ
sÖfÑ
and for the cathodic reactions (6. 8), (6. 9) & (6. 13) respectively:
mn,â opq Ürâ áâf - opq -Üyâ áâf

làf = zn
f f
nâ
f
z (6. 20)
âf f opq -Ü
Z/ opq Ürâ áâf - yâ á â f
z{,u f z{,} f
âf Ñ â
mn,} opq Ür} á}f - opq -Üy} á}f

lef = f
zn
f f
(6. 21)
}f
Z- opq -Üy} á}f
z{,} f
}f
P a g e | 283
Where the limiting diffusion current density (jL) for a RDE system was defined by the Levich
equation as:
lä = ±1.554çÉ Xbé#/. è -Z/ê ë Z/# ! (6. 22)
The diffusion coefficients for the solution species in listed in Table 6-4 were obtained from the
literature. The kinematic viscosities of the MEA solutions were derived from the dynamic
viscosity and density measurements reported by (Amundsen et al., 2009) which have been listed
in Table 6-6 to Table 6-9. In the absence of available data for the diffusion coefficient, density
and dynamic viscosity at elevated temperatures, the equations displayed in Table 6-5 were used
to obtain the approximate values.
Table 6-4: Diffusion Coefficients for species in water at 25oC
Species Diffusion coefficient / m2 s-1 Source

Fe2+ 0.72×10-9 (Newman and Thomas-Alyea, 2004)
H+ 9.31×10-9 (Newman and Thomas-Alyea, 2004)
O2 1.92×10-9 (Han and Bartels, 1996)
CO2 (aq) 1.94×10-9 (Tamimi et al., 1994)
H2CO3 2.00×10-9 (Kvarekval, 1997)
HCO3- 1.11×10-9 (Newman and Thomas-Alyea, 2004)
CO32- 0.92×10-9 (Kvarekval, 1997)
Table 6-5: Liquid properties as a function of temperature (Haynes and Lide, 2011)
Parameter Expression
Dynamic Viscosity/ kg m-1 s-1 Z..#ïï #ñ..Z_-ó -ò.òòZò_.(#ñô.Z_-ó)f
í = 0.001002 ∙ 10 ó-Zêô.Z_
Density/ kg m-3 ö = (753.596 + 1.87748i − 0.003564i # )
Diffusion coefficient / m2 s-1 i í#ñô.Z_

éó = é#ñô.Z_
298.15 í
P a g e | 284
Table 6-6: Density ρ/ kg m3 for MEA +H2O from (T= 25- 80oC) as a function of MEA mass
fraction w (Amundsen et al., 2009).
T/ oC w
20 wt% 30 wt% 40 wt% 50 wt% 70 wt% 90 wt% 100 wt%
25 1005.3 1010.6 1015.8 1020.8 1026.2 1020.0 1012.3
40 999.1 1003.4 1007.7 1011.7 1015.5 1008.4 1000.3
50 994.3 998.1 1001.8 1005.3 1008.2 1000.6 992.3
70 983.0 985.8 988.9 991.5 993.0 984.6 976.0
80 976,6 979.4 981.9 984.2 985.0 976.4 967.8
Table 6-7: Density ρ / kg m3 for 30 wt% MEA +H2O from (T= 25- 80oC) as a function of
temperature and CO2 loading (α =0.1 - 0.5) (Amundsen et al., 2009).
T/ oC α
0.1 0.2 0.3 0.4 0.5
25 1028.0 1048.0 1070.0 1095.7 1121.1
40 1021.0 1041.0 1062.9 1088.5 1114.0
50 1016.0 1035.5 1058.0 1083.0 1108.0
70 1004.0 1024.0 1046.4 1071.9
80 997.0 1017.6 1040.2 1066.0
Table 6-8: Dynamic viscosity μ / mPa s-1 for MEA +H2O from (T= 25- 80oC) as a function of
MEA mass fraction w (Amundsen et al., 2009).
T / oC w
20 wt% 30 wt% 40 wt% 50 wt% 70 wt% 90 wt% 100 wt%
25 1.70 2.48 3.58 5.51 12.46 19.40 17.90
40 1.18 1.67 2.28 3.39 6.69 10.20 9.61
50 0.95 1.33 1.75 2.54 4.94 7.06 6.72
70 0.67 0.92 1.14 1.57 2.79 3.81 3.69
80 0.58 0.77 0.95 1.28 2.18 2.93 2.85
P a g e | 285
Table 6-9: Dynamic viscosity μ / mPa s-1 for 30 wt% MEA +H2O from (T= 25- 80oC) as a
function of temperature and CO2 loading (α =0.1 - 0.5) (Amundsen et al., 2009).
T / oC α
0.1 0.2 0.3 0.4 0.5
25 2.6 2.9 3.1 3.5 3.9
40 1.7 2.0 2.0 2.4 2.7
50 1.4 1.6 1.6 1.9 2.1
70 0.9 1.1 1.1 1.3 1.5
80 0.8 0.9 0.9 1.1 1.3
The electrochemical parameters EO/R, j0 and α are contingent upon the electrochemical reaction
and are dependent on the temperature and the concentration of the species at the surface of the
metal.
EO/R was calculated from the Nernst equations in equation (6. 23) to (6. 25), assuming that
activities could be approximated by concentrations:
Åó
aÇÉ fÑ /ÇÉ = a ò ÇÉ fÑ /ÇÉÇÉ + ln [X5 #/ ] (6. 23)
#Ç
Åó † âf
aà Ñ /àf = a ò à Ñ /àf − ln (6. 24)
#Ç àÑ f
Åó
aef /àf e = aeòf /àf e + ln °ef ∗ [+ / ]^ (6. 25)
^Ç
Where E0 is the standard electrode potential for each of the reactions under standard conditions
(298.15 K, 0.1 MPa) calculated from standard Gibbs energies for formation.
-∆•°
a° = (6. 26)
¶Ö Ç
At elevated temperature conditions, the Gibbs energy change (ΔGT0) was calculated by
determining the heat capacity at constant pressure from the equation:

P a g e | 286
d(Zò© )
!ß = b + ® 10-. i + + é 10-ê i # (6. 27)
óf
The heat capacity values were obtained from the HSC software database and have been listed
in the Appendix VI. In addition to the following equations for entropy (S), enthalpy (H) and
Gibbs energy of formation (ΔfG):
ó d´
™ = ™#ñô.Z_ + #ñô.Z_ ó
7i (6. 28)
ó d´
+ = ∆+#ñô.Z_ + #ñô.Z_ ó
7i (6. 29)
¨≠ = ¨+ − i¨™ (6. 30)
The anodic and cathodic transfer coefficients are defined by equations (6. 31) and (6. 32)
(Guidelli et al., 2014).
αa = ( RT F )( d ln ja dE ) (6. 31)
αc = − ( RT F )( d ln jc dE ) (6. 32)
According to (Bockris et al., 1961) α, the transfer coefficient is 0.5 for cathodic reactions and 2
for anodic reactions. The values expressions for j0 were obtained from (Nordsveen et al., 2003)
and are listed in Table 6-10.
At the (open circuit) corrosion potential, the rate of production of electrons by oxidation must
be equal and opposite to their rate of consumption by reduction:
lJ = ld (6. 33)
As iron dissolution is the main oxidation reaction in the active region:
lÇÉ = lÆ (6. 34)

P a g e | 287
As H+, H2O and O2 all participate as oxidising agents the total reduction rate will be:
ld = là Ñ + làf e + lef (6. 35)
The values of jcorr and corrosion potential (Ecorr) were determined from the intersection point of
the simulated anodic and cathodic polarization curves.

P a g e | 288
Table 6-10: Kinetic parameters of the electrochemical reactions
Reaction j0 / A m-2 !",$%& -./,012 -.3456,$%& Tref / oC k0 / s-1 -Ea / ΔH /
/ ) *+, kJ kJ
-1
mol mol-1
Fe ?@ABC6
D
?@ /
+8.F
∆L 1 1 0.53 10-9.1 1 50 40
78,9: ;/ = 78,9: ;/=:> exp − −
oxidation ?
@ABC 6 ,=:> ?@ /,=:> M O O=:>
H+ ?@ /
8.F
∆L 1 1 0.05 - - 20 - - 30
78,@ / = 78,@ /,=:> exp − −
reduction ?@ /,=:> M O O=:>
H 2O 78,@;B = 3×10+F A m+D - - - - - -

reduction
O2 −VW - - - - 1.69 2.00

78,B; = U8 ?B; exp
reduction MO ×103 ×104
HCO3- ?@ /
8.F
∆L 1 1 0.15 - 0.5 50 - - 50
7X,@ / = 7X,@ /,=:> exp − −
reduction ?@ /,=:> M O O=:>
P a g e | 289
6.2. Model Verification by Comparison with Experiments
The validity of the corrosion model was tested by comparing the model predictions with the
experimental results. The voltammetric data simulated for Fe in 30 wt% MEA+H2O solution
by the electrochemical model showed good agreement with the experimental data, specifically
within cathodic potentials and at active dissolution potentials. However, the model did not take
into account the precipitation of FeCO3 / Fe(OH)2 or passivation by adsorbed species at higher
potentials. Details of the model results are summarised in Appendix VII.
6
ln ( Current density / A m-2 )
4
Ecorr
3
jcorr
2
0 Experiment
Model
-1
-1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4
30 wt% MEA+ CO2, at pH=8.1 and T = 25 oC; 10 mV s-1, f = 1500 rpm. Black dotted lines
represent the Tafel extrapolation of the anodic and cathodic linear sections to find the
intersection at Ecorr and icorr.
The model-predicted voltammetric data indicated slightly higher corrosion rates, within
+25.2% of the corrosion rates calculated from experimental data. However, due to the
complexity of the process(es) and the precipitation of FeCO3 and Fe(OH)2 which had been
neglected within the model, this result was acceptable.

P a g e | 290
As observed experimentally, the mass transport conditions had a considerable effect on the
anodic and cathodic current densities which increased with increasing mass transport rate
(Figure 6-4). Thus the corrosion rates (Figure 6-5) also increased with mass transport rates due
to the increased transport of the oxidising species resulting from the reactions in the bulk
solution to the surface of the iron.
5 5
540 rpm 5
960 rpm
4 4 240 rpm
ln (Current Density / A m-2 )

4
3
3
3
2
2 2
1
0 1 1
-1 0 0
-2 -1 -1
-3
-1 -0.8 -0.6 -0.4 -2 -2
-1 -0.8 -0.6 -0.4 -1 -0.8 -0.6 -0.4
Electrode Potential Electrode Potential Electrode Potential
(AgCl|Ag) / V (AgCl|Ag) / V (AgCl|Ag) / V
Experiment Experiment Experiment
Model Model Model
30 wt% MEA+ CO2, at pH=8.1 and T = 25 oC; 10 mV s-1, f = 960 rpm, 540 rpm and 240 rpm.
9
Experiment
8
Model
7
6
5
4
3
2
1
0
1500 960 540 240
Rotation rate / rpm

Figure 6-5: Comparisons of experimental and model-predicted corrosion rates as a function of
rotation rate for iron RDE in 30 wt% MEA+ CO2, at pH=8.1 and T = 25 oC; 10 mV s-1.
P a g e | 291
6.3. Effect of CO2 loading
As described in Chapter 5, the chemical composition of the MEA solution changes depending
on the CO2 loading; the speciation of 30 wt% MEA with CO2 loading at 25oC is shown in
Figure 6-6 and listed in Table 6-11.
1.0
0.9
RNH2
0.8 RNH3+
0.7
Mole Fraction
0.6
HCO3-
0.5
0.4
0.3 CO2 (aq)
0.2 RNHCOO-
0.1
CO32-
0.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 6-6: Predicted effect of CO2 loading on liquid phase compositions of 30 wt% aqueous
MEA solution at 25 °C, as predicted by the Kent-Eisenberg Model (Kent and Eisenberg, 1976).
Figure 6-8 shows the model-predicted effect of CO2 loading on corrosion rates, which
increased mainly because of increased rates of hydrogen evolution. As shown in Table 6-11,
increased CO2 loading was predicted to cause higher equilibrium concentrations of H+, RNH3+
and HCO3-; the last two species were produced by reactions of CO2 and MEA-H2O and acted
as proton sources by the dissociation reactions (6. 3) and (6. 7), leading to higher rates of
hydrogen evolution.
P a g e | 292
6 4 3
α= 0.4
α =0.6 α =0.2
5 3

2
n (Current Density / A m-2 )

4 2
3 1
1
2 0
0
1 -1
-1
0 -2
-1 -2
-3
-1.2 -1 -0.8 -0.6 -0.4 -1 -0.9 -0.8 -0.7 -0.6 -0.5
-1 -0.9 -0.8 -0.7 -0.6 -0.5
Experiment Experiment
Experiment
Model Model Model
8
Corrosion rate / mm y--1
7 Experiment
Model
6
0
0.6 0.4 0.2
Figure 6-8: Comparisons of experimental and model-predicted corrosion rates for iron RDE in
P a g e | 293
(Waard and Milliams, 1975), (Nordsveen et al., 2003) suggested that the presence of hydrogen
bicarbonate can also increase the hydrogen evolution rate by undergoing direct reduction via
reaction:
2"#$% & + 2( & → "* + 2#$% *& (6. 36)
6
5
Ecorr
4
3
2 jcorr
1
Experiment
0
Model
-1
-1.3 -1.2 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2
Figure 6-9: Comparison of model-predicted and experimental current density – electrode

potential data for iron RDE in 30 wt% MEA+ CO2 at pH=8.1 and T = 25 oC; 10 mV s-1,
f = 1500 rpm. Direct reduction of HCO3- included.
However, hydrogen evolution from weak acids in general involves fast chemical pre-equilibria,
prior to the reduction of the resulting protons. Nonetheless, the direct reduction of HCO3- was
implemented in the model by including the cathodic current density and reversible electrode
potential for HCO3- reduction.
45, :;< => ? 1 & :;< &=@ ? 1

7890 1 7890 1 7890 7890 1 7890
,-./0 1 = C5 C5 (6. 37)
7890 1 7890 1
AB :;< => ? 1 & FGH &=@ ? 1
CD,E 7890 1 7890 CD,9 7890 1 7890
"2 7890 1
L NO S7K ∗[./0 K1 ]K
I-./0 1 /-K = I-./ 1
0 /-K
− ln (6. 38)
*P -./0 1 K
P a g e | 294
This resulted in current density-electrode potential data that agreed less well with experimental
data than model predictions excluding HCO3- reduction and a slightly higher corrosion rate of
19.06 mm y-1 in comparison to 7.75 mm y-1 in the absence of HCO3- reduction.
6.4. Effect of Temperature
As expected, increasing the temperature increased model-predicted corrosion rates, as shown
in Figure 6-11. The effect of temperature on the equilibrium H+ ion concentration as a function
of CO2 loading, calculated by the Kent-Eisenberg Model is also shown in Figure 6-12.
However, it is important to note that the model showed a closer agreement to the experimental
results at low temperatures (25 - 40 oC) than at higher temperatures. This discrepancy at higher
temperatures is probably as a result of the shortcomings of the Kent-Eisenberg model to predict
the solution species concentrations at higher temperatures.
6 6 6
40 oC 60 oC 80 oC
5 5 5
4 4
4
3 3
3
2 2
2
1 1
1
0 0
-1 0
-1
-1.2 -1 -0.8 -0.6 -0.4 -1.1 -1 -0.9 -0.8 -0.7 -0.6 -0.5
-1.2 -1 -0.8 -0.6
Experiment Experiment Experiment
Model Model Model
Figure 6-10: Comparison between predicted and measured voltammetric sweep for iron RDE
in 30 wt% MEA+ CO2, at pH=8.1 10 mV s-1, f = 1500 rpm as a function of temperature.
P a g e | 295
25
Experiment
Model
Corrosion rate / mm year-1
20
15
10
0
25 40 60 80
Temperature / °C
Figure 6-11: Comparison of experimental and model-predicted Fe corrosion rates in
30 wt% MEA solution at different temperatures.
45
25C
H+] ion concentration x 108 / M
40 60C
35 80C
100C
30
25
20
15
10
5
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 6-12: Effect of temperature and CO2 loading on hydrogen ion concentration [H+]
predicted by Kent-Eisenberg model for 30 wt% aqueous MEA solution.
P a g e | 296
Table 6-11: Predicted species concentration / kmol m-3 for 30 wt% MEA-H2O-CO2 systems at 25oC
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
RNH2 4.32E+00 3.70E+00 3.16E+00 2.64E+00 2.21E+00 1.76E+00 1.44E+00 1.19E+00 1.04E+00 9.21E-01
+
RNH3 6.08E-01 1.17E+00 1.66E+00 2.15E+00 2.58E+00 3.03E+00 3.37E+00 3.63E+00 3.80E+00 3.93E+00
-
RNHCOO 2.72E-01 6.34E-01 9.49E-01 1.20E+00 1.31E+00 1.10E+00 7.92E-01 4.75E-01 2.62E-01 5.07E-02
-
HCO3 1.47E-01 2.53E-01 4.07E-01 6.07E-01 8.63E-01 1.51E+00 2.16E+00 2.76E+00 3.13E+00 3.40E+00
2-
CO3 7.78E-02 1.33E-01 1.37E-01 1.56E-01 1.90E-01 1.99E-01 2.01E-01 1.95E-01 1.98E-01 1.85E-01
CO2 2.64E-03 -1.97E-02 6.33E-03 3.16E-02 1.38E-01 1.86E-01 3.48E-01 5.72E-01 9.11E-01 1.26E+00
+
H 2.00E-12 9.77E-12 7.45E-10 1.71E-09 4.86E-09 1.06E-08 1.74E-08 2.57E-08 3.27E-08 3.80E-08
-
OH 5.01E-03 1.02E-03 1.34E-05 5.83E-06 2.06E-06 9.42E-07 5.75E-07 3.89E-07 3.05E-07 2.63E-07
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
RNH2 4.24E+00 3.68E+00 3.13E+00 2.60E+00 2.13E+00 1.73E+00 1.40E+00 1.16E+00 9.67E-01 7.69E-01
+
RNH3 6.77E-01 1.16E+00 1.65E+00 2.14E+00 2.61E+00 3.02E+00 3.37E+00 3.62E+00 3.84E+00 4.06E+00
-
-
HCO3 1.17E-01 2.13E-01 3.68E-01 5.72E-01 8.80E-01 1.47E+00 2.14E+00 2.70E+00 3.36E+00 3.55E+00
2-
CO3 1.02E-01 1.22E-01 1.33E-01 1.47E-01 1.70E-01 1.86E-01 1.91E-01 1.95E-01 1.89E-01 1.68E-01
CO2 -1.64E-03 2.57E-03 9.14E-03 2.91E-02 8.91E-02 1.92E-01 3.36E-01 5.86E-01 8.59E-01 1.11E+00
+
H 3.98E-10 1.15E-09 2.05E-09 4.65E-09 1.13E-08 2.04E-08 3.16E-08 5.01E-08 1.00E-07 2.05E-07
-
OH 2.51E-05 8.73E-06 4.89E-06 2.15E-06 8.86E-07 4.91E-07 3.16E-07 2.00E-07 1.00E-07 4.87E-08
P a g e | 297
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
RNH2 4.22E+00 3.66E+00 3.09E+00 2.55E+00 2.07E+00 1.67E+00 1.35E+00 1.11E+00 8.97E-01 8.51E-01
+
RNH3 6.74E-01 1.14E+00 1.63E+00 2.13E+00 2.60E+00 3.01E+00 3.35E+00 3.62E+00 3.86E+00 3.91E+00
-
RNHCOO 3.04E-01 7.07E-01 1.05E+00 1.32E+00 1.43E+00 1.22E+00 8.99E-01 5.73E-01 3.44E-01 2.36E-01
-
HCO3 9.83E-02 1.82E-01 3.27E-01 5.11E-01 8.21E-01 4.04E-01 2.07E+00 2.68E+00 3.18E+00 3.34E+00
2-
CO3 9.90E-02 1.10E-01 1.14E-01 1.38E-01 1.62E-01 1.83E-01 1.86E-01 1.82E-01 1.67E-01 1.65E-01
CO2 1.73E-03 1.82E-03 7.33E-03 3.09E-02 8.68E-02 1.93E-01 3.47E-01 5.69E-01 8.11E-01 1.26E+00
+
H 1.00E-09 1.40E-09 3.50E-09 1.07E-08 1.86E-08 2.77E-08 4.76E-08 7.87E-08 1.58E-07 3.16E-07
-
OH 1.00E-05 7.17E-06 2.86E-06 9.34E-07 5.37E-07 3.62E-07 2.10E-07 1.27E-07 6.31E-08 3.16E-08
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
RNH2 4.20E+00 3.56E+00 3.02E+00 2.48E+00 2.01E+00 1.61E+00 1.31E+00 1.07E+00 8.93E-01 8.49E-01
+
RNH3 6.72E-01 1.19E+00 1.64E+00 2.13E+00 2.59E+00 3.00E+00 3.33E+00 3.59E+00 3.80E+00 3.86E+00
-
-
HCO3 8.05E-02 1.65E-01 2.85E-01 4.60E-01 7.60E-01 1.34E+00 1.99E+00 2.59E+00 3.05E+00 3.22E+00
2-
CO3 9.22E-02 7.70E-02 9.50E-02 1.27E-01 1.53E-01 1.73E-01 1.82E-01 1.79E-01 1.70E-01 1.72E-01
CO2 1.98E-04 3.18E-03 6.41E-03 2.16E-02 8.45E-02 1.96E-01 3.68E-01 5.93E-01 8.71E-01 1.32E+00
+
H 2.51E-09 3.98E-09 6.24E-09 7.04E-09 2.23E-08 4.15E-08 7.38E-08 1.47E-07 2.50E-07 4.30E-07
-
OH 3.98E-06 2.51E-06 1.60E-06 1.42E-06 4.48E-07 2.41E-07 1.35E-07 6.82E-08 4.00E-08 2.33E-08
P a g e | 298
6.5. Effect of MEA Concentration
According to the model predictions, corrosion rates decreased with increasing amine
concentration. Due to limitations on the available data for the dynamic viscosity and density
measurements of MEA solutions loaded with CO2, the model was tested for (20-50) wt% MEA
concentration and compared to the experimental results in Figure 6-13. In spite of the increase
in the equilibrium concentration of H+ ions predicted at higher MEA concentrations, there was
a decrease in the corrosion rate which can be attributed to the change in viscosity with MEA
concentration (Figure 6-14). This model-predicted decrease in corrosion rates with increasing
MEA concentration provides direct evidence for the experimental observation at higher MEA
concentrations resulting from the change in viscosity of the solution. Increased viscosity of the
solution resulted in the dissolution reaction being limited by the decreased mass transport of
the oxidising species H+, H2O and O2 to the iron electrode.
12
Fe Corrosion Rate / mm year-1
Experiment
10
Model
0
20 30 40 50
Figure 6-13: Comparison of experimental and model-predicted corrosion rates at different

MEA concentrations, T = 25oC, CO2 loading = 0.5 mol CO2 (mol amine)-1.
P a g e | 299
20
Dyanmic Viscosity / mPa s-1
18
16
14
12
10
8
6
4
2
0
0 10 20 30 40 50 60 70 80 90 100
Figure 6-14: Dynamic viscosity of 30 wt% MEA at 25 °C and CO2 loading (α =0.5), as a
function of mass fraction (Amundsen et al., 2009).
P a g e | 300
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
RNH2 2.99E+00 2.96E+00 2.34E+00 1.88E+00 1.54E+00 1.30E+00 1.08E+00 8.75E-01 7.41E-01 6.63E-01
RNH3+ 2.41E-01 1.84E-01 8.24E-01 1.29E+00 1.63E+00 1.87E+00 2.10E+00 2.31E+00 2.45E+00 2.54E+00
-
-
HCO3 1.13E-01 4.82E-01 2.50E-01 2.38E-01 3.97E-01 8.65E-01 1.41E+00 1.95E+00 2.25E+00 2.40E+00
2-
CO3 3.26E-03 4.63E-03 9.56E-03 1.97E-02 2.35E-02 5.00E-03 2.40E-03 1.98E-02 1.36E-02 3.42E-02
CO2 2.60E-03 3.00E-03 5.40E-03 8.50E-03 1.90E-02 8.84E-02 1.89E-01 2.95E-01 5.08E-01 7.71E-01
+
H 3.16E-10 3.46E-10 6.31E-10 1.26E-09 2.20E-09 8.70E-09 1.63E-08 4.57E-08 1.26E-07 2.51E-07
-
OH 3.16E-05 2.89E-05 1.58E-05 7.94E-06 4.54E-06 1.15E-06 6.13E-07 2.19E-07 7.94E-08 3.98E-08
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
RNH2 5.97E+00 5.86E+00 4.48E+00 3.66E+00 2.92E+00 2.43E+00 1.94E+00 1.59E+00 1.32E+00 1.12E+00
+
RNH3 4.40E-01 2.32E-01 1.69E+00 2.50E+00 3.24E+00 3.74E+00 4.25E+00 4.63E+00 4.94E+00 5.17E+00
-
-
HCO3 2.86E-01 1.12E+00 9.67E-01 1.21E+00 1.65E+00 2.39E+00 3.23E+00 3.94E+00 4.49E+00 4.85E+00
2-
CO3 1.31E-03 3.78E-03 7.56E-03 9.24E-03 1.88E-02 2.95E-02 3.82E-02 3.98E-02 3.42E-02 3.08E-02
CO2 3.40E-03 9.90E-03 9.40E-03 4.69E-02 7.39E-02 2.14E-01 3.62E-01 6.35E-01 9.80E-01 1.40E+00
+
H 1.00E-10 6.31E-10 2.36E-09 9.27E-09 1.47E-08 2.83E-08 5.18E-08 7.94E-08 1.11E-07 1.44E-07
P a g e | 301
OH- 1.00E-04 1.58E-05 4.23E-06 1.08E-06 6.78E-07 3.53E-07 1.93E-07 1.26E-07 9.01E-08 6.94E-08
o
Table 6-17: Predicted species concentration / M for 50 wt% MEA-H2O-CO2 systems at 25 C
Species α
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
RNH2 7.08E+00 6.25E+00 5.31E+00 4.48E+00 3.73E+00 3.15E+00 2.64E+00 2.39E+00 1.92E+00 1.95E+00
+
RNH3 9.19E-01 1.53E+00 2.37E+00 3.10E+00 3.86E+00 4.44E+00 4.99E+00 5.26E+00 5.79E+00 5.75E+00
-
RNHCOO 3.93E-01 9.16E-01 1.27E+00 1.61E+00 1.70E+00 1.49E+00 1.17E+00 8.43E-01 5.84E-01 4.90E-01
-
HCO3 3.65E-01 8.84E-01 1.20E+00 1.85E+00 2.26E+00 3.02E+00 3.86E+00 4.43E+00 5.15E+00 5.22E+00
2-
CO3 8.44E-04 1.01E-03 5.11E-03 1.43E-02 1.58E-02 3.00E-04 2.30E-03 4.03E-02 9.60E-03 4.96E-02
CO2 1.31E-02 2.11E-02 3.22E-02 4.13E-02 1.46E-01 3.97E-01 7.01E-01 1.24E+00 1.63E+00 2.43E+00
+
H 3.16E-10 3.98E-10 1.74E-09 6.40E-09 1.85E-08 4.16E-08 6.63E-08 1.05E-07 1.32E-07 1.68E-07
-
OH 3.16E-04 2.51E-05 5.75E-06 1.56E-06 5.41E-07 2.40E-07 1.51E-07 9.56E-08 7.59E-08 5.94E-08
P a g e | 302
P a g e | 303
6.6. Discussion and Conclusions
A mechanistic model for the prediction of corrosion of iron in MEA-CO2-H2O was developed
by taking into account the following cathodic reactions: hydrogen evolution, direct water
reduction, oxygen reduction and hydrogen carbonate ion reduction. Iron dissolution was the
only anodic reaction considered within the model. The data obtained from the model was used
for:
• Prediction of corrosion rates;
• Elucidating the governing corrosion mechanism and identify the primary oxidising
species;
• Determination of the electrochemical kinetics as a function of operating parameters:
CO2 loading, temperature, electrode rotation rate and MEA concentration.
The mechanistic model incorporated mathematical equations for the chemical and
electrochemical reaction rates and mass transport processes governing the rates of iron
corrosion in MEA-CO2 systems. The equilibrium concentrations of the aqueous species in
solution calculated using the Kent-Eisenberg method showed good agreement with
experimental data for the pH and CO2 loading. Concentrations of MEA, the protonated amine
(RNH3+) and hydrogen bicarbonate ions (HCO3-) and thus hydrogen ions (H+) were predicted
to increase with CO2 loading and temperature.
Based on the results of the speciation model, a simplified electrochemical model was developed
for the ‘uniform’ corrosion of iron in aqueous MEA-CO2 solutions. Model-predicted steady
state current density - electrode potential data were compared with corresponding experimental
data, as functions of MEA concentration, temperature, CO2 loading and electrode rotation rate.
P a g e | 304
Values of exchange current densities and transfer coefficients for iron oxidation and reduction
of protons, water and oxygen were obtained from the literature.
Model predictions were in modest to fairly good agreement with experimental data, with the
model generally predicting higher corrosion rates than those determined experimentally,
especially at higher temperatures and low MEA concentrations. The discrepancy was attributed
to the model omitting the effects of FeCO3 precipitation, which would have partially passivated
the dissolution of Fe to FeII. In addition, the diffusion coefficients used with the model were
obtained at low temperatures and a theoretical relationship was used to extend their validity to
higher temperatures.
In addition to predicting corrosion rates, the mechanistic model was useful in elucidating the
effects of CO2 and of MEA concentrations on the iron corrosion rates and mechanism.
Contrary to suggestions by (Waard and Milliams, 1975) and (Veawab et al., 1999) that the
direct reduction of HCO3- participates in the corrosion reaction and acts as the primary oxidant
in MEA systems, the model predicted that the effect of HCO3- ions in solution was to provide
an additional source of H+ ions through dissociation.
Model predictions showed fair agreement with experimental results, the increased viscosities
at MEA concentrations ≥ 30 wt% decreasing corrosion rates. According to experimental data
from (Amundsen et al., 2009), the viscosity of 50 wt% aqueous MEA solution at 25 °C loaded
with 0.5 mol CO2 (mol amine)-1 was more than three times that of 20 wt% MEA under the
same conditions. As the solution became more viscous, oxidant diffusion rates decreased,
thereby decreasing rates of the dissolution and overall corrosion process.
Comparisons between model predictions and the experimental data revealed the strengths of
the model such as the ability to clarify the impact of operating parameters on the predicted
P a g e | 305
corrosion rates. However, it was also evident from the comparisons that the model was also
weaker in determining the corrosion rates at high solution temperatures and low MEA
concentrations, which is primarily attributed to the limitations of the Kent-Eiseneberg model
to provide reliable thermodynamic data for extreme conditions.

P a g e | 306
6.7. References
BARD, A. J.,, PARSONS, R. and JORDAN, J. 1985. Standard Potentials in Aqueous Solution.
848? A.J.Bard, R.Parsons and J.Jordan (eds.), 'Standard Potentials in Aqueous
Solution', Marcel Dekker, New York, 1985.
AMUNDSEN, T. G., ØI, L. E. & EIMER, D. A. 2009. Density and Viscosity of
Monoethanolamine + Water + Carbon Dioxide from (25 to 80) °C. Journal of Chemical
& Engineering Data, 54, 3096-3100.
AUSTGEN, D. M., ROCHELLE, G. T., PENG, X. & CHEN, C. 1989. Ind. Eng. Chem. Res.,
28, 1073.
BOCKRIS, J. O'M., DRAZIC, D. & DESPIC, A. R. 1961. The electrode kinetics of the
deposition and dissolution of iron. Electrochimica Acta, 4, 325-361.
CHAKMA, A. & MEISEN, A. 1990. Improved Kent-Eisenberg model for predicting CO2
solubilities in aqueous diethanolamine (DEA) solutions. Gas Separation &
Purification, 4, 37-40.
DESHMUKH, R. D. & MATHER, A. E. 1981. A mathematical model for equilibrium
solubility of hydrogen sulfide and carbon dioxide in aqueous alkanolamine solutions.
Chemical Engineering Science, 36, 355-362.
EDWARDS, T. J., MAURER, G., NEWMAN, J. & PRAUSNITZ, J. M. 1978. Vapor-liquid
equilibria in multicomponent aqueous solutions of volatile weak electrolytes. Aiche
Journal, 24, 966-976.
GUIDELLI, R., COMPTON RICHARD, G., FELIU JUAN, M., GILEADI, E., LIPKOWSKI,
J., SCHMICKLER, W. & TRASATTI, S. 2014. Defining the transfer coefficient in
electrochemistry: An assessment (IUPAC Technical Report). Pure and Applied
Chemistry 86, 245-256.
HAJI-SULAIMAN, M. Z. & AROUA, M. K. 1995. EQUILIBRIUM OF CO2 IN AQUEOUS
DIETHANOLAMINE(DEA) AND AMINO METHYL PROPANOL (AMP)
SOLUTIONS. Chemical Engineering Communications, 140, 157-171.
HAMBORG, E. S. & VERSTEEG, G. F. 2009. Dissociation Constants and Thermodynamic
Properties of Amines and Alkanolamines from (293 to 353) K. Journal of Chemical &
Engineering Data, 54, 1318-1328.
HAN, P. and BARTELS, D. M. 1996. Temperature Dependence of Oxygen Diffusion in H2O
and D2O. J. Phys. Chem. 100 (13), 5597-5602.
HAYNES, W. M. & LIDE, D. R. 2011. Handbook of Chemistry and Physics. CRC Press, 92nd.
Edition, Boca Raton, Florida.
KENT, R. L. & EISENBERG, B. 1976. Better Data for Amine Treating. Hydrocarbon
Processing, 55, 87.
KVAREKVAL, J. 1997. A Kinetic Model For Calculating Concentration Profiles And Fluxes
Of CO2-Related Species Across The Nernst Diffusion Layer. Corrosion/97, Paper No.
5. NACE International. Houston, Texas.
P a g e | 307
MACDOWELL, N. 2010. The integration of advanced molecular thermodynamics and

process modelling for the design of amine-based CO2 capture processes. PhD Thesis,
Imperial College London.
NEWMAN, J. & THOMAS-ALYEA, K. E. 2004. Electrochemical Systems, United States of
America, John Wiley & Sons, Inc.
NORDSVEEN, M., NEŠIĆ, S., NYBORG, R. & STANGELAND, A. 2003. A Mechanistic
Model for Carbon Dioxide Corrosion of Mild Steel in the Presence of Protective Iron
Carbonate Films—Part 1: Theory and Verification. Corrosion, 59, 443-456.
PARK, R. L. 1975. Inner-shell Spectroscopy. Phys. Today. 28, 52.
TAMIMI, A., RINKER, E.B. and SANDALL, O.C. 1994. Diffusion coefficients for
hydrogen sulfide, carbon dioxide, and nitrous oxide in water over the temperature
range 293- 368 K, J. Chem. Eng. Data 39, 330–332.
TONG, D. 2012. Development of Advanced Amine Systems with Accurate Vapour-liquid

Equilibrium Measurement PhD PhD Thesis, Imperial College London.
VEAWAB, A., TONTIWACHWUTHIKUL, P. & CHAKMA, A. 1999. Corrosion Behavior
of Carbon Steel in the CO2 Absorption Process Using Aqueous Amine Solutions.
WAARD, C. D. & MILLIAMS, D. E. 1975. Carbonic Acid Corrosion of Steel. Corrosion, 31,
177-181.
WEBER, R. E. and PERIA, W. T.1967. Use of LEED Aparatus for the Detection and
Identification of Surface Contaminants. J Appl. Phy. 38, 4355.
P a g e | 308
7. Chapter 7
Analysis of Time-dependent Corrosion

of Iron in MEA- H2O- CO2 Systems
This section details the results of 9 sets of experiments carried out in a flow reactor, which
was designed and fabricated to enable measurement of dissolved iron concentrations under
conditions typical of amine-based CO2 absorption. This aimed to enable discrimination
between current densities leading to dissolution from those resulting in adsorbed products,
such as iron (hydr-)oxide or carbonate. This was done to investigate the components of the
corrosion processes systematically and elucidate their kinetics, mechanisms and rate
determining steps(s), by combining electrochemical techniques with subsequent chemical
analysis of solutions by inductively coupled plasma optical emission spectrophotometry
(ICP-OES).
The mechanisms of iron and carbon steel corrosion in aqueous-CO2 systems has been
deliberated and debated for several decades, with particular attention being given to CO2
P a g e | 309
absorbing amine solutions in recent years. Extensive research of CO2-induced steel corrosion
has been carried out, primarily because of its great importance in the oil and natural gas
industry; unfortunately, conflicting corrosion rates as functions of process parameters and
various mechanisms have been reported. In order to reconcile the seemingly contradictory
reports, it is important to note that the term ‘corrosion’ in CO2 systems is as a result of
several reactions as outlined in Chapter 5 & 6. Several CO2 dependent chemical,
electrochemical and mass transport processes occur simultaneously in the aqueous solution at
the iron ǀ solution interface. Each of these reactions responds differently to changes in CO2
loading, temperature, solution flow rate and other operating parameters, which may result in
changes to the concentrations in solution species. Whilst insoluble corrosion products such as
FeCO3 and Fe(OH)2 have been identified in various locations on amine plants (DuPart and
Rooney, 1993), there are also soluble products such as Fe2+ resulting from the corrosion
reactions. So far, very few publications have addressed the experimental quantification of
these soluble products.
The ICP-OES was used to determine dissolved (FeII) concentrations in the amine solution
after constant potential electrolyses (chronoamperometry), enabling the charge yield for
dissolution to be determined by relating the amount of charge passed to the amount of iron
dissolved. The combined use of conventional electrochemical methods, as well as mass loss
determination by chemical analysis provided a more accurate method for dissolution rate
measurement than the commonly used Tafel extrapolation method.
The first task was to study the effect of the electrode potential on the corrosion and
passivation behaviour. A matrix of experiments was designed to establish the effects on the
corrosion rate of operating parameters: solution temperature, CO2 loading, amine
concentration, oxygen content and solution flow rates. The effects of these experimental
P a g e | 310
variables had been studied previously with the RDE system (Chapter 5), which was not able
to determine partial current densities due to FeII dissolution.
Once the baseline of the corrosion behaviour of iron in the industry standard solvent MEA
had been firmly established, a series of experiments were conducted to compare that to the
behaviour in other promising and commercially available solvents: (MDEA), (AEP) and
(AMP). According to (Kohl and Nielsen, 1997), most of the equipment and piping in CO2
absorption plants are fabricated from carbon steel, so it was also imperative to establish the
corrosion rate of carbon steel under similar operating conditions. Most of the previous studies
on steel corrosion in aqueous MEA have focused on carbon steel with 0.20 % carbon (Nainar
and Veawab, 2009; Veawab et al., 1999), which together with carbon steel with 0.35 %
carbon and stainless steel, was included in this study.
The majority of the experiments were based on ca. 1 hour electrochemical measurements on
iron / steel specimens with the surfaces ‘cleaned’ beforehand, representative of new
processing equipment and piping at the start of plant operation. In order to simulate typical
CO2 absorption plant conditions, it was also deemed important to carry out long term
measurements, over a period of 28 days. These series of experiments, conducted at open
circuit potential, aimed to establish the longer term corrosion behaviour and to gain
information on the effects of degradation products accumulating in the aqueous solution.
7.1. Electrochemical Behaviour of Iron in Large Scale Electrode
Figure 7-1 shows the cyclic voltammogram (dE/dt = 10 mV s-1) recorded for the iron
electrode, as 4 dm3 of aqueous solution containing 30 wt% MEA fully saturated with CO2
(25oC, pH 8.03) was circulated through the reactor at a linear velocity of 0.5 mm s-1. The iron
P a g e | 311
electrode was ‘cleaned’ cathodically by applying a potential of -1.00 V (AgCl|Ag) for 10
minutes to reduce any (hydr-)oxides formed spontaneously in air;
EAgCl Ag (SHE) / V = 0.2197 − 0.0592log (Cl − ) . After this procedure, the open circuit potential was
measured as -0.75 V. The electrode potential was scanned from a negative potential limit of -
1.00 V to a positive limit of 0.00 V, a potential window in which the anodic behaviour of iron
could be studied without the evolution of oxygen (EO2/H2O = +0.55 V). The potential-pH
diagram for a Fe-H2O-CO2 system (Figure 5-9) was used to assign reactions to the main
features of Figure 7-1.
25
(II)
20
EFe2+/Fe3O4 = -0.66 V
15 EFe/Fe(OH)2 = -0.76 V
EFe/Fe2+ = -0.79 V (IV) (III)

10
EFe/FeCO3 = -0.86 V (I) EFe3O4/ Fe2O3= -0.48 V
5
0
EH2/H+= -0.67 V EFe(OH)2/Fe(OH)3= -0.43 V
-5
-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0
Figure 7-1: Cyclic voltammograms of an iron electrode in 30 wt% MEA + H2O + CO2 at
10 mV s-1 (T = 25oC, pH 8.03) with a linear velocity of 0.5 mm s-1.
The cyclic voltammogram exhibited the same shape and form as those obtained from the
RDE system under similar operating conditions (25 °C, pH 8.01). On the positive-going
potential sweep from the negative potential limit of -1.00 V, cathodic current densities
resulted from hydrogen evolution from the reduction of H2O and H+ ions. At potentials > -
P a g e | 312
0.78 V, anodic current densities were measured due to the dissolution of iron (EFe2+/Fe = -
0.79 V) occurring simultaneously with other oxidation reactions, such as the precipitation of
FeCO3 (EFeCO3/Fe = -0.86 V, based on [HCO3-] of 1.96 M obtained from the Kent-Eisenberg
model described in Chapter 6). The anodic domain was characterised by three features: at ca -
0.72 V (I), a current density peak of 23.7 A m-2 at ca. -0.51 V (peak II), and (III) at ca. -
0.28 V. The current density peak (II) was attributed to the dissolution of iron by reaction (7.
1) and FeCO3 formation when its solubility product was exceeded sufficiently; at more
positive potentials, this was followed by a passive region due to the formation of (hydr-
)oxides on the iron surface.
!" ↔ !" $% + 2" ( (7. 1)
On the negative-going scan from the positive potential limit of 0 V, there was a reduction
peak (IV) at ca. -0.42 V followed by an additional peak (V) at ca. -0.54 V, as previously
detected with the RDE system. As described in Chapter 5, this peak resulted from the
simultaneous reduction of (hydr-)oxide such as Fe2O3 and Fe(OH)3 that reactivated the
oxidative dissolution reaction of Fe to FeII, giving rise to a net positive current density.
7.2. Effect of Electrode Potential
The anodic domain of iron in aqueous MEA-CO2 systems was studied by
chronoamperometry, recording current responses over a 1 hour period to a range of
potentials. To complement the chronoamperometric data, known aliquots of the aqueous
amine solution were extracted from the reactor outlet at 10 minute intervals for subsequent
analysis by the ICP-MS to measure dissolved FeII ion concentrations. This enabled the
P a g e | 313
potential range for active iron dissolution to be defined within which subsequent experiments
could be conducted.
14 22
-0.55V 20 -0.55V
12
-0.60V

18 -0.50V
10 -0.65V
16 -0.45V
-0.70V
8
-0.75V 14
-0.40V
6 -0.80V 12
10 -0.20V
4
8
2 6
0 4
2
-2
0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
Figure 7-2: Effect of potential on chronoamperograms obtained on iron electrode at -

0.80 V ≤ E ≤ -0.20 V in contact with 30 wt% MEA + H2O + CO2 (T = 25oC, pH 8.03) with a
linear velocity of 0.5 mm s-1.
Figure 7-2 shows the chronoamperometric (j-t) data for Fe in 30 wt% MEA fully saturated
with CO2 (25 C, pH 8.03), for potentials between -0.80 V and -0.20 V. Potentials of -0.80 V
°
and -0.75 V produced low cathodic (negative) current densities, resulting in relatively small
concentrations of dissolved FeII, as shown in Figure 7-4. The cathodic current measured was
due to the applied potential at -0.80 V being more negative than the equilibrium potential for
iron dissolution (EFe2+/Fe = -0.79 V at dissolved FeII activity of 10-4 and pH 8.03) and while
that of -0.75 V was less negative than EFe2+/Fe, the applied overpotential was not significant
enough to induce a net anodic current. However, the total dissolved FeII ions in solution
increased significantly as the potential became less negative. The total amount of dissolved
iron detected in the solution at the end of the measurement was 130.4 mg m-2 at -0.80 V and
P a g e | 314
315.6 mg m-2 -0.75 V. As EH2O H2 > EFe2+ Fe , net current densities occurred in the potential
window EH2O H2 > E > EFe2+ Fe .
16000
Charge Density / C m-2

16000
14000
14000
12000
12000
10000
10000
8000
6000 8000
4000 6000
2000 4000
0 2000
-2000 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time/ s Time / s
-0.55V -0.60V -0.65V -0.55V -0.50V -0.45V
-0.70V -0.75V -0.80V -0.40V -0.20V
Figure 7-3: Effect of potential on the charge density-time data for iron electrode at -0.80 V ≤
E ≤ -0.20 V in 30 wt% MEA + H2O + CO2 (T = 25oC, pH 8.03) with a linear velocity of
0.5 mm s-1.
At potentials less negative than -0.70 V, net anodic (positive) current densities were
measured due to the dissolution of iron as iron(II)aq, the subsequent precipitation of FeCO3,
Fe2(OH)2CO3, Fe(OH)2 and (or) passivation of oxides on the electrode surface. The current
densities initially exhibited a maximum value, followed by a sharp decrease within 100 s,
after which it continued to decline at a slower rate. This decrease in current density with time
was indicative of a passivation or precipitation process on the surface of the iron electrode.
The net charged passed was anodic (calculated by integrating current-time data) and
increased from -0.70 V to -0.55 V, then started to decrease from -0.55 V to -0.20 V (Figure 7-
3).
The total concentration of FeII ions measured also increased within the potential range of -
0.70 V to -0.55 V and then decreased gradually between -0.55 V and -0.20 V (Figure 7-4).
P a g e | 315
The charge yield of the dissolution reaction was calculated by relating the charge that would
have been passed theoretically for the dissolution of elemental Fe to the measured FeII ions in
solution by the ICP-OES, to the total charge passed (recorded in the chronoamperogram).
4000 4000
FeII dissolved/ mg m-2

3000 3000
2000 2000
1000 1000
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time/ s Time/ s
-0.55V -0.60V -0.65V -0.55V -0.50V -0.45V
-0.70V -0.75V -0.80V -0.40V -0.20V
Figure 7-4: Effect of potential on FeII concentration-time data for iron electrode at -
0.80 V ≤ E ≤ -0.20 V in contact with 30 wt% MEA + H2O + CO2 (T = 25 C, pH 8.03) with a °
linear velocity of 0.5 mm s-1.
The dissolution charge yield increased from -0.70 V to -0.55 V and then decreased at
potentials less negative than -0.55 V. The observed decrease in dissolution rate and anodic
current densities at potentials less negative than -0.55 V coincided with the occurrence of the
anodic peak (II), which was attributed to the active-passive transition of iron due to its
passivation by oxides and hydroxides. Figure 7-5 shows images of the iron electrode after
constant potential electrolyses and demonstrates the increasingly passive behaviour at
potentials less negative than -0.55 V. This behaviour validated the results from the cyclic
voltammograms shown in Figure 7-1. Thus, the potential of -0.60 V was selected as a suitable
potential for the subsequent constant electrode potential electrolyses in the active region, in
which iron dissolution dominated the resulting anodic current. This chosen potential also falls
within the small potential range (-0.88 V ≤ E ≤ -0.52 V) in which the precipitation of
P a g e | 316
protective FeCO3, Fe2(OH)2CO3 and Fe(OH)2 ((7. 2) – (7. 4) is predicted thermodynamically
according to the potential-pH diagram for Fe-CO2-H2O systems, an example of which is
shown in Figure 5-9.
!" $% (+,) + ./0 $( (+,) ↔ !"./0 (1) (7. 2)
Fe2+ + FeCO3 + 2OH − ⎯⎯

→ Fe OH )2 CO3
←⎯⎯ 2( (7. 3)
!" $% (+,) + 2/2( (+,) ↔ !"(/2)$ (+,) (7. 4)
-0.70V
-0.60V
-0.50V
-0.40V
Figure 7-5: Images of iron electrode in 30 wt% MEA + H2O + CO2 at pH 8.03 and T = 25oC,)
with a solution velocity of 0.5 mm s-1 (Re = 1.69) after constant electrolyses at -0.70 V, -
0.60 V, -0.50 V & -0.40 V.
P a g e | 317
7.3. Effect of Solution Velocity
7.3.1. Reactor Characterisation with Mass Transport Correlations
As outlined below, the hydrodynamic behaviour of the reactor was characterised using a set
of mass transport equations, which relate the hydrodynamic conditions to the geometry of the
reactor and the characteristics of the amine solution.
The equivalent hydrodynamic diameter, de for the reactor compartment of electrode-
membrane distance d and electrode width w was obtained with the following equation (7. 5):
6(78911 147:;9<+= +84+) $>?

34 = = (7. 5)
>4::4? @48;A4:48 (>%?)
The mass transport coefficient km, is a heterogeneous rate constant which characterises the
mass transport conditions of the reactor under defined conditions such as amine composition,
temperature and solution velocity. The mass transport coefficient, km,x at distance ‘x’ from the
inlet of the reactor, was estimated from the appropriate correlation for the (laminar)
hydrodynamic properties of the reactor and is a function of dimensionless quantities that were
used to characterise fluid flow regimes in the flow reactor:
• Reynolds number, defined as the ratio of the inertial forces to viscous forces, where u
is the linear velocity and v is the kinematic viscosity of the solution:
D?E
Re = (7. 6)
F
• Schmidt number: the ratio of the (momentum) kinematic viscosity, v to the diffusivity
coefficient, Di for species i, and is used to characterise fluid flow regimes in which
there are simultaneous momentum and mass diffusion convection processes.

P a g e | 318
F
Sc = (7. 7)
IJ
• Sherwood number represents the ratio of the total mass transfer rate (convective plus
diffusion) to the diffusion transfer rate alone:
MN,P ?E
ShL = (7. 8)
IJ
250
Sherwood Number, Sh
200
150
100
50
883
830
777
0
724
671
618
0.015
565
0.06
512
459
0.105
406
0.15
353
300
0.195
247
0.24
194
Reynolds number, Re
141
0.285
Distance from inlet / m

88
Figure 7-6: Dependences of Sherwood number on Reynold number and distance from inlet of
the flow reactor based on the reactor geometry, solution flowrate 0.01 - 0.1 m s-1 and MEA
solution viscosity (3.22×10-6 m2 s-1).
For a parallel plate electrochemical reactor with a small inter-electrode gap, the
dimensionless group correlation for fully developed laminar flow (Re < 2000) is given by
equation (7. 9) (Pickett, 1979; Walsh, 1993; Rousar et al., 1971; Rousar, 1986):
?E V/0
ShL = 1.23 Re ∙ Sc ∙ (7. 9)
U
The Reynolds and Sherwood numbers were calculated from equations (7. 6) and (7. 9), based
on the reactor geometry and solution velocity of 0.01 - 0.1 m s-1. The kinematic viscosity of
P a g e | 319
30 wt% MEA at 25oC with 0.5 mol CO2 (mol amine)-1 was calculated using published
viscosity and density data (Amundsen et al., 2009). The results are shown in Figure 7-6.
The mass transport coefficients, km (displayed in Figure 7-8 were then determined by
manipulations of the dimensionless group mass transport correlations (7. 9) and equation (7.
8) and averaged over the electrode surface.
7.3.2. Reactor Characterisation by the Reduction of Hexacyanoferrate(III)
For hydrodynamic systems in which the mass transport correlations cannot be predicted
adequately from published correlations, mass transport coefficients can also be determined
experimentally by using a diffusion limited current technique, usually based on the reduction
of hexacyanoferrate(III) ions (Fe(CN)63-) or the oxidation of hexacyanoferrate(II) ions
(Fe(CN)6-4) at a platinum, gold or nickel electrode (Walsh, 1993; Colli et al., 2013; Wragg
and Leontaritis, 1997):
!"(.X)Y 0( + " ( ↔ !"(.X)Y 6( (7. 10)
According to (Walsh, 1993), the reduction of hexacyanoferrate(III) ions gives better results
than the oxidation of hexacyanoferrate(II) ions, as metal oxide formation may affect the
kinetics of the latter, depending on the working electrode material. Due to the alkalinity of
the hexacyanoferrate(III/II) solutions, at the appropriate potentials, hexacyanoferrate(III) can
be reduced with close to 100 % current efficiency. In addition, in the absence of the hydrogen
evolution reaction, the reduction of Fe(CN)63- displays a distinctive mass transport limited
current, as shown in Figure 7-7.

P a g e | 320
1.0
0.8
0.6
Fe(CN)64- + e- → Fe(CN)63-
Current Density/ A m-2
0.4
0.2
0.0
-0.2
Z[(\])^ _( + [( → Z[(\])^ a(
-0.4
-0.6
-0.8
-1.0
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Figure 7-7: Cyclic voltammograms of an iron electrode in contact with 50 mol m-3 of
K3Fe(CN)6 / 100 mol m-3 of K4Fe(CN)6 and 1000 mol m-3 of Na2CO3 at pH 11.3 with a flow
rate of 0.1 mm s-1.
Thus, as a method of comparison to the mass transport correlations described above, the flow
reactor was also characterised by the reduction of hexacyanoferrate(III). 2 dm3 of solution
containing 50 mol m-3 K3Fe(CN)6, 100 mol m-3 K4Fe(CN)6 and a supporting electrolyte of
Na2CO3 at pH 11.3 was circulated through the reactor at various solution linear velocities
between 0.1 to 0.8 mm s-1. Cyclic voltammograms were determined from a negative potential
limit of -0.80 V to a positive limit of 0.80 V and the mass transport limiting reduction current
plateaux were measured as functions of solution velocity (Figure 7-7).
The (local) mass transport coefficient km,x is related the mass transport limited current density
jLIM,x to the reactant concentration change in the bulk solution (Wragg and Leontaritis, 1997),
where c∞ is the bulk concentration of the electro-active species.

P a g e | 321
cdef,P
bA,U = (7. 11)
<g7h
By relating the magnitude of the mass transport limited current to the solution flow rates, the
mass transport coefficient, km for hexacyanoferrate(III) was obtained from (7. 11) and the
corresponding values for FeII were calculated using (7. 12). The mass transport coefficient
values for Fe(CN)6 and FeII obtained by the mass transport correlation and reduction of
hexacyanoferrate (III) are shown as a function of solution flow rates in Figure 7-8. The
diffusion coefficients for the Fe(CN)6-3 ions and FeII ions were obtained from (Books et al.,
2006) and (Reddi and Inyang, 2000) respectively.
IpEij
bA,g4 ij = bA,g4 kl m
no (7. 12)
I no
pE qr m
The mass transport coefficient, km is also given by equation (7. 13), from which the thickness
of the Nernst diffusion layer δN was calculated and displayed in Figure 7-9.
IpEij
bA,g4 ij = (7. 13)
sr
There was a discrepancy between the values of the mass transport coefficient calculated by
the reduction of Fe(CN)63- and the values predicted by mass transfer correlations, due to flow
not being well developed, because of entry and exit flow dominating mass transport within
the short reactor (Pickett and Ong, 1974).
As shown in Figure 7-9, increasing solution velocity decreased the thickness of the Nernst
diffusion layer at the iron metal ǀ solution interface.

P a g e | 322
4.8E-6 6E-6
Mass transportation, km / m s-1
5E-6
(correlations), km / m s-1
Mass transportation
4.6E-6
4E-6
4.4E-6 3E-6
2E-6
4.2E-6 Fe(CN)6 3-
Fe2+ 1E-6
Fe2+ (correlations)
4.0E-6 0E+0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Linear velocity / mm s-1

Figure 7-8: Mass transfer coefficient versus linear velocity for Fe(CN)63- measured during
hexacyanoferrate(III) reduction and the corresponding values calculated for FeII with
comparisons to the calculations based on mass transport correlations.
1.62E-4 3.0E-4
Diffusion layer thickness, δN / m
Diffusion layer thickness

2.5E-4
(correlattion), δN / m
1.60E-4
2.0E-4
1.58E-4
1.5E-4
1.56E-4
1.0E-4
Fe2+
1.54E-4
Fe2+ (correlations) 5.0E-5
1.52E-4 0.0E+0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Linear velocity / mm s-1
Figure 7-9: Effect of solution velocity on the diffusion layer thickness for FeII measured by
the reduction of hexacyanoferrate(III) with comparisons to the calculations based on mass
transport correlations.
P a g e | 323
7.3.3. Effect of solution flow rate on iron corrosion in MEA+ CO2 solutions
Figure 7-10 shows the effect of solution linear velocity on current densities of iron at -
0.6 V (AgCl|Ag) in contact with aqueous 30wt% MEA + CO2, current densities increasing
with increasing flow rates.
9
8 0.8 mm s-1
Current Density/ A m-2
7 0.7 mm s-1
6 0.5 mm s-1
5 0.3 mm s-1
4 0.2 mm s-1
3
2
1
0
0 500 1000 1500 2000 2500 3000 3500 4000
Time / s
Figure 7-10: Effect of solution velocity on chronoamperometric data for iron electrode in 30
wt% MEA + H2O + CO2 at pH 8.03 and T = 25 °C,); -0.60 V (AgCl|Ag).
The bulk concentration of FeII ions increased with solution velocity (Figure 7-13), partly due
to the enhanced mass transport of the oxidising species (H+) to the surface of the iron. The
charge density (Figure 7-12) and the charge yield of the iron dissolution reaction also
increased with solution flow rate, implying that the FeCO3 protective layer formation/
precipitation of Fe(OH)2 was impeded by solution flow. FeCO3 precipitation occurred only if
the local concentrations of FeII and CO32- ions in the aqueous solution exceeded its solubility
product, with sufficient supersaturation to provide a driving force. The temperature
dependent solubility product, Ksp for FeCO3 formation is 1.27 ×10-11 at 25oC (calculated from
thermodynamic data in (Benezeth et al., 1985)). The condition for Fe(OH)2 precipitation is
outlined in equation (7. 18). In order to determine if the ‘local’ conditions at the iron
electrode met the requirements for FeCO3 and Fe(OH)2 nucleation, the FeII ion concentrations
P a g e | 324
at the electrode surface [Fe2+(x=0)] was estimated from the flux, Xg4 ij and the measured FeII
ion concentration in the bulk solution [Fe2+(x=δN)] by taking into account the concentration
profile across the Nernst diffusion layer.
MN
!" → 2" ( + !" $% (Utu) !" $% (Utsr ) (7. 14)
Figure 7-11 shows a schematic representation of the concentration profile of FeII across an
idealised Nernst diffusion layer at an iron electrode.
[Fe2+]x=0
FeCO3
Fe(OH)2
[Fe2+]x=δN
δN
x=0 x = δN
Figure 7-11: Idealised Nernst diffusion layer and concentration profile for FeII ions required
for FeCO3 and Fe(OH)2 precipitation.
A series of equations were employed for the calculation of [Fe2+(x=0)] and are outlined in (7.
15) – (7. 17) below:
The flux of FeII ions from the electrode ǀ solution interface (x=0) into the solution:
P a g e | 325
cpE cv,pE wLx (ypE zpE )

Xg4 ij = = (7. 15)
$g $g
~ g4 ij g4 ij Pv ( g4 ij PÄ
Xg4 ij ≈ −}g4 ij ≈ −}g4 ij r
(7. 16)
~U sr
At steady-state and in the absence of FeII accumulation/precipitation:
cpE sr
!" $% Utu ≈ !" $% Utsr + (7. 17)
$gIpEij
Precipitation of Fe(OH)2 may occur if:
!" $% Utu ≥ 10 V$.É6($@Ñ ÖÜá 3Ö(0 (7. 18)
Precipitation of Fe2(OH)2CO3 may occur if:
!" $% Utu ≥ 10 Vu.àu($@Ñ ÖÜá 3Ö(0 (7. 19)
Precipitation of FeCO3 may occur if:
!" $% Utu ∙ ./0 $( ≥ â1@ (7. 20)
However, all three processes require supersaturation as a driving force for phase formation.
The resulting calculations for the FeII concentration at the surface of the electrode, [Fe2+](x=0)
obtained at each linear solution velocity was summarised in Table 7-1. The equilibrium
concentration of the CO32- ions was calculated by the Kent Eisenberg model described in
Chapter 6. At each of the flow rates studied within this work, the [Fe2+](x=0), [CO32-] and [OH-
] species exceeded the solubility products, Ksp, of [Fe2+](x=0) [CO32-] and/or [Fe2+(x=0)](x=0)
[OH-]2 , i.e. the thermodynamic condition required for Fe(OH)2 and FeCO3 precipitation to
occur.
P a g e | 326
Turbulent solution flow conditions (Re > > 2200 for channel flow) at higher flow rates has
been reported (Revie, 2015) to cause mechanical disruption to the FeCO3 film, decreasing the
barrier of the passive film to the diffusion of oxidising agents to the underlying iron surface.
However, this was unlikely to be the case in this flow system, as the hydrodynamic
conditions remained under laminar flow conditions ( Re ≤ 6.47 ) even at the maximum flow
rate studied (Figure 7-10). However, the increase in dissolution charge yield for FeIIaq
formation with increasing solution velocity was most probably due to a decrease in the
supersaturation level at the metal ǀ solution interface, through the increased mass transport of
FeII away from the iron surface.
14000 14000
Charge Density / A s m-2
Charge Density/ A s m-2
12000 12000
10000
10000
8000
8000
6000
4000 6000
2000 4000
0 2000
0 1000 2000 3000 4000
0
Time / s 0 0.2 0.4 0.6 0.8 1
0.8 mm s-1 0.7 mm s-1
Flowrate/ mm s-1
0.5 mm s-1 0.3 mm s-1
Figure 7-12: Effect of solution velocity on charge density-time data for iron electrode in 30
wt% MEA + H2O + CO2 at pH 8.03 and T = 25 oC, at -0.60 V (AgCl|Ag).
P a g e | 327
Fe dissolved / mg m-2
3000 3.5
Coorosion rate/ mm y-1

2500 3
2000 2.5
1500 2
1000 1.5
1
500
0.5
0
0
0 1000 2000 3000 4000
0 0.2 0.4 0.6 0.8 1
Time/ s
0.8 mm s-1 0.7 mm s-1 Flowrate / mm s-1
0.5 mm s-1 0.3 mm s-1
Figure 7-13: Effect of solution velocity on time dependence of mass of FeII dissolved from
iron electrode in 30 wt% MEA + H2O + CO2 at pH 8.03 and T = 25 oC; -0.60 V (AgCl|Ag).
Table 7-1: Effect of solution velocity on the conditions required for FeCO3 and Fe(OH)2
precipitation by determining the [FeII] at the iron electrode surface, equilibrium [CO32- ] from
the Kent Eisenberg model and the results from iron electrode in 30 wt% MEA + H2O + CO2
(T = 25 oC, pH 8.03) at -0.60 V (AgCl|Ag).
-1 2+ 2+ 2+ 2- 2+ 1 2+ 2-
Flowrate / mm s Time / s [Fe ]x=δN [Fe ]x=0 [Fe ]x=0[CO3] [Fe ]x=0 [Fe ]x=0[CO3]
0.80 600 2.97E-05 4.01E-03 7.98E-04 4.34E-03 8.64E-04
1200 7.52E-05 3.81E-03 7.58E-04 4.38E-03 8.70E-04
1800 1.14E-04 3.69E-03 7.33E-04 4.15E-03 8.25E-04
2400 1.59E-04 3.68E-03 7.31E-04 3.81E-03 7.58E-04
3000 2.06E-04 3.74E-03 7.43E-04 3.20E-03 6.36E-04
3600 2.45E-04 3.64E-03 7.25E-04 2.58E-03 5.12E-04

0.70 600 4.00E-05 3.86E-03 7.68E-04 3.89E-03 7.73E-04
1200 7.52E-05 3.65E-03 7.26E-04 3.71E-03 7.37E-04
1800 1.17E-04 3.53E-03 7.02E-04 3.59E-03 7.14E-04
2400 1.58E-04 3.36E-03 6.68E-04 3.02E-03 6.01E-04
3000 1.89E-04 3.45E-03 6.86E-04 2.64E-03 5.24E-04
3600 2.20E-04 3.46E-03 6.88E-04 2.45E-03 4.87E-04

0.50 600 2.60E-05 3.72E-03 7.40E-04 3.56E-03 7.09E-04
1200 5.93E-05 3.50E-03 6.97E-04 3.36E-03 6.67E-04
1800 9.32E-05 3.37E-03 6.71E-04 3.23E-03 6.43E-04
2400 1.34E-04 3.20E-03 6.36E-04 3.07E-03 6.10E-04

1
Calculations based on km from the reactor characterisations based on mass transport correlations.
P a g e | 328
3000 1.67E-04 3.29E-03 6.55E-04 3.16E-03 6.28E-04

3600 1.91E-04 3.29E-03 6.55E-04 3.16E-03 6.29E-04

0.30 600 2.58E-05 3.59E-03 7.14E-04 3.40E-03 6.75E-04
1200 5.97E-05 3.43E-03 6.81E-04 3.21E-03 6.39E-04
1800 6.55E-05 3.32E-03 6.60E-04 3.11E-03 6.19E-04
2400 1.14E-04 2.80E-03 5.56E-04 2.96E-03 5.90E-04
3000 1.47E-04 2.44E-03 4.86E-04 3.05E-03 6.06E-04
3600 1.63E-04 2.27E-03 4.51E-04 3.06E-03 6.09E-04

0.20 600 2.88E-05 3.27E-03 6.51E-04 3.36E-03 6.68E-04
1200 5.65E-05 3.31E-03 6.58E-04 3.19E-03 6.35E-04
1800 7.53E-05 3.14E-03 6.25E-04 3.10E-03 6.16E-04
2400 1.03E-04 2.89E-03 5.76E-04 3.10E-03 6.16E-04

3000 1.25E-04 2.44E-03 4.85E-04 3.16E-03 6.28E-04
2
3600 1.46E-04 1.97E-03 3.93E-04 3.08E-03 6.14E-04

7.4. Temperature Effect - Full CO2 saturation
Figure 7-14 shows the effect of temperature on current density-time data at a potential of -
0.60 V (AgCl|Ag) in 30 wt% MEA solution at full CO2 saturation. 40 oC represented the
absorber conditions while the highest temperature 80 oC was representative of the rich-lean
heat exchanger. The chronoamperograms maintained a similar characteristic shape, but the
initial maximum current density increased with temperature. The current density decreased
rapidly until it reached a minimum current density (close in value at all temperatures), which
remained constant for the duration of the experiment.
2
Based on the Kent-Eisenberg Model, [CO32-] was 0.20 M for aqueous 30wt% MEA+CO2 , 25oC at pH 8
P a g e | 329
60
80C
50 60C
40C
40
25C
30
20
10
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-14: Effect of temperature on current density – time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 oC, pH 8.03) with a velocity of 0.5
mm s-1.
Both the total charge (Figure 7-15) and dissolved FeII ion concentration (Figure 7-16)
increased with increasing temperature. The change in dissolved FeII with time exhibited a
seemingly linear behaviour at 25 C, but then dissolution rates at higher temperatures (40-
°
80 C) decreased. The calculated dissolution charge yield for FeIIaq species also decreased
°
with temperature; for example, the charge yield decreases from 77.9 % at 40 C to 65.3 % at °
80 C, due to increased rates of adsorbed product formation.

°
The following expression for the temperature dependence of the solubility product of FeCO3
was obtained by (Benezeth et al., 2009) and the calculated values for the temperature range
studied within this work are shown in Table 7-3.
(V
logVu â1@ = ç + é ∙ è/â + 3 ∙ logVu (è/â) (7. 21)
Elevating the temperature increased the rate of all the processes involved in the corrosion of
iron in the aqueous amine solution including mass transport rates, chemical reactions in the
bulk solution and electrochemical reactions at the iron ǀ solution interface. As the kinetics of
each of these processes increased with temperature, depending on the solubility product of
P a g e | 330
ferrous carbonate (FeCO3) film formation, increased temperature can either increase or
decrease the corrosion rate.
40000 40000
Total Charge Density / C m-2

30000
30000
20000
20000
10000
10000
0
0 600 1200 1800 2400 3000 3600 0
20 30 40 50 60 70 80
Time/ s
80C 60C 40C 25C Temperature / 0C
Figure 7-15: Effect of temperature on the charge density-time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25–80 oC, pH 8.03) with a solution velocity
of 0.5 mm s-1.
Increasing the temperature promoted higher concentrations of FeII ions in the aqueous amine
solution, and as the local concentrations exceeded the solubility product, FeCO3, the
protective film precipitated. The solubility product (Ksp) is related to the supersaturation (S)
via reaction (7. 22).
g4 ij Pv ∙ kën io
ê= (7. 22)
íìî
Once the supersaturation value S exceeded unity, the FeCO3 film could have begun to
precipitate on the surface of the iron, probably leading to a decrease in corrosion rates.
However, at low concentrations of Fe2+(x=0) and CO32- that are only slightly above the
solubility limit and at low temperatures, the rate of FeCO3 precipitation is slow. Thus, the
precipitation kinetics were deemed even more important in determining the protectiveness of
the FeCO3 film than just the thermodynamics of the system. (Johnson and Tomson, 1991)
P a g e | 331
discovered that the most important parameters influencing the kinetics of FeCO3 precipitation
are the supersaturation and temperature and they proposed an equation (7. 23) for the
precipitation rate:
úin.v ú
ñ
ïg4kën = ∙ expö6.õ( ùû ∙ â1@ ∙ (ê i − 1)$ (7. 23)
ó
The nucleation and growth of the FeCO3 film are slow processes that are enhanced by
increasing the solution temperature. The nucleation of crystalline films such as FeCO3 has
proven to be challenging to model mathematically and it is also generally accepted that the
rate of precipitation is dependent primarily on the crystal growth rate (supersaturation)
instead of the nucleation rate. In addition to the rate equation (7. 23) proposed by (Johnson
and Tomson, 1991), (van Hunnik and Hendriksen, 1996) also proposed a more frequently
used equation for the crystal growth rate of FeCO3; both equations have been used to evaluate
the effect of temperature on the rate of corrosion based on the Fe2+(x=0) and CO32- obtained
from the experiments.
According to (van Hunnik and Hendriksen, 1996):
úúü.†
ñ
ïg4kën = ∙ expö$.6( ùû ∙ â1@ ∙ (ê − 1)(1 − ê (V ) (7. 24)
ó
The FeCO3 film growth rate obtained from equation (7. 23) and (7. 24) based on the Fe2+(x=0)
data listed in Table 7-2 is plotted as a function of the temperature in Figure 7-17. The results
obtained are in agreement with reports from literature that substantial rates of film formation
are achieved only at higher temperatures (> 60 C) and high supersaturation, S (Nesic and
°
Lee, 2003). The solubility of FeCO3 is also dependent on the temperature and decreases with
increasing temperature (Table 7-3).

P a g e | 332
Although the dissolution rate increased significantly at higher temperatures (i.e. increased
concentration of Fe2+ ions in the bulk solution), the formation rate of FeCO3 also accelerated
which resulted in a decrease in the dissolution rate at elevated temperatures. This decrease in
corrosion rate occurred as a result of the FeCO3 layer acting as a diffusion barrier to the
oxidising species and covering at least part of the iron surface, thereby inhibiting the
dissolution of underlying metal. Figure 7-18 shows images of the iron electrode after the
chronoamperometric experiments; the passive film formation was more evident at higher
temperatures. It is also important to note that the film formed at 60oC was non-adherent to the
metal surface and could easily be wiped off with a cloth, whereas the films formed at 80oC
were so tough they could be removed only by mechanical polishing.
7000
8
6000 7
FeII dissolved / mg m-2
5000 6
4000 5
3000 4
2000 3
1000 2
1
0
0 1000 2000 3000 4000 0
Time / s 20 30 40 50 60 70 80
80C 60C 40C 25C Temperature / 0C
Figure 7-16: Temperature effect on FeII concentration-time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 C, pH 8.03) with a solution velocity of °
0.5 mm s-1.
P a g e | 333
0.30
Rate of FeCO3 precipitation / mol m-3 s-1
Johnson and Tomson

0.25
van Hunnik and Hendriksen
0.20
0.15
0.10
0.05
0.00
20 30 40 50 60 70 80
Temperature / °C
Figure 7-17: Effect of temperature on the FeCO3 precipitation rate at iron electrode in contact
with 30 wt% MEA + H2O + CO2 (T = 25 C, pH 8.03) with a solution velocity of 0.5 mm s-1
°
after constant potential electrolyses at -0.60 V (AgCl|Ag).
25oC
60oC
80oC
Figure 7-18: Images of iron electrode in contact with 30 wt% MEA + H2O + CO2 (T = 25 –
80 C, pH 8.03) with a solution velocity of 0. 5 mm s-1 after constant electrolyses at -0.60 V
°
(AgCl|Ag).
P a g e | 334
Table 7-2: Effect of temperature on the conditions required for FeCO3 precipitation by
determining the [FeII] at the iron electrode surface, equilibrium [CO32- ] from the Kent
Eisenberg model and the results from iron electrode in contact with 30 wt% MEA + H2O +
CO2 (T = 25 °C, pH 8.03) at -0.60 V.
Temperature
o 2+ 2+ 2- 2+ 2- 2+ 1 2+ 2- 1
/ C Time / s [Fe ]x=δN [Fe ]x=0 [CO3] [Fe ]x=0[CO3] [Fe ]x=0 [Fe ]x=0[CO3]
25 600 2.60E-05 3.72E-03 1.99E-01 7.40E-04 3.56E-03 7.09E-04
1200 5.93E-05 3.50E-03 1.99E-01 6.97E-04 3.36E-03 6.67E-04
1800 9.32E-05 3.37E-03 1.99E-01 6.71E-04 3.23E-03 6.43E-04
2400 1.34E-04 3.20E-03 1.99E-01 6.36E-04 3.07E-03 6.10E-04
3000 1.67E-04 3.29E-03 1.99E-01 6.55E-04 3.16E-03 6.28E-04
3600 1.91E-04 3.29E-03 1.99E-01 6.55E-04 3.16E-03 6.29E-04

40 600 8.74E-05 6.67E-03 1.86E-01 1.24E-03 6.39E-03 1.19E-03
1200 1.80E-04 4.60E-03 1.86E-01 8.57E-04 4.41E-03 8.22E-04
1800 2.28E-04 3.20E-03 1.86E-01 5.97E-04 3.07E-03 5.73E-04
2400 2.70E-04 2.25E-03 1.86E-01 4.19E-04 2.17E-03 4.04E-04
3000 3.13E-04 2.02E-03 1.86E-01 3.77E-04 1.95E-03 3.63E-04
3600 3.40E-04 1.80E-03 1.86E-01 3.35E-04 1.73E-03 3.23E-04

60 600 1.37E-04 7.63E-03 1.83E-01 1.39E-03 7.31E-03 1.34E-03
1200 2.68E-04 5.13E-03 1.83E-01 9.38E-04 4.92E-03 9.00E-04
1800 3.54E-04 3.42E-03 1.83E-01 6.26E-04 3.29E-03 6.02E-04
2400 4.07E-04 2.77E-03 1.83E-01 5.07E-04 2.67E-03 4.88E-04
3000 4.65E-04 2.46E-03 1.83E-01 4.50E-04 2.38E-03 4.35E-04
3600 5.23E-04 2.30E-03 1.83E-01 4.21E-04 2.23E-03 4.07E-04

80 600 1.85E-04 7.68E-03 1.73E-01 1.33E-03 7.36E-03 1.27E-03
1200 3.43E-04 5.21E-03 1.73E-01 9.01E-04 5.00E-03 8.65E-04
1800 4.35E-04 3.51E-03 1.73E-01 6.07E-04 3.37E-03 5.84E-04
2400 4.84E-04 2.85E-03 1.73E-01 4.93E-04 2.75E-03 4.76E-04
3000 5.44E-04 2.54E-03 1.73E-01 4.40E-04 2.46E-03 4.25E-04
3600 5.70E-04 2.35E-03 1.73E-01 4.07E-04 2.28E-03 3.94E-04

Table 7-3: Effect of temperature on the solubility product, Ksp (calculated from
thermodynamic data in (Benezeth et al., 2009))
Temperature / oC Ksp
25 1.27E-11
40 8.83E-12

60 4.90E-12

80 2.48E-12

P a g e | 335
7.5. Effect of CO2 loading
The effect of CO2 loading (α) on the corrosion rate of iron in the aqueous 30wt% MEA
solutions was investigated at loadings of 0.4, 0.2, 0.0 and compared to the results at full CO2
loading (α ≈ 0.6). The desired CO2 loadings were achieved according to the CO2 solubility
data in 30 wt% MEA obtained by (Tong et al., 2012) and (Jou et al., 1995), included in
Appendix (I). As an example, according to Jou et al., a CO2 partial pressure 8.96 Pa would be
required to achieve a CO2 loading of 0.203 mol CO2 (mol MEA)-1 at 40oC.
There was a significant difference between the current-time behaviour of the iron maintained
potentiostatically at -0.60 V in loaded and unloaded amine solutions (Figure 7-19). In the
absence of CO2 (α = 0.0, pH 12.31), the current-time transients resulted in a net cathodic
charge which increased in value with increasing temperature. Although MEA in itself is non-
corrosive in the absence of CO2 and iron dissolution was not predicted thermodynamically at
pH 12.3, at lower temperatures (25-40 °C), there was a relatively small amount of dissolved
FeII ions detected in the amine solution that increased linearly with time (Figure 7-21). At
elevated temperatures (60-80oC), [FeII] decreased with time, i.e. there was a net loss of
dissolved FeII, presumably due to its precipitation / adsorption. In unloaded MEA solutions,
total dissolved FeII concentrations decreased with increasing temperature.
In CO2 loaded MEA solutions (α = 0.4, pH 9.90 and α = 0.2, pH 10.70), the current-time
transients maintained the characteristic anodic behaviour as seen previously at full CO2
saturation (α = 0.6, pH 8.03) (Figure 7-22 & Figure 7-25). The net anodic charge and total
dissolved FeII ion concentration increased with CO2 loading and linearly with temperature
(Figure 7-23 - Figure 7-27). For example, the total dissolved FeII concentration at 80oC at the
end of the chronoamperometric measurement was 1.534 g m-2 at α = 0.2, 2.815 g m-2 at
P a g e | 336
α = 0.4, 6.364 g m-2 at α = 0.8. The transformation in the corrosion behaviour of iron in MEA
solutions on absorption of CO2 was due to the changes in the ionic composition of the amine
solution resulting from the reaction of the amine and CO2. As described previously in
Chapter 5 and Chapter 6, as the CO2 loading is increased, so do equilibrium concentrations
of hydrogen carbonate ion (HCO3-) and protonated amine (RNH3+), which subsequently
dissociate to form hydrogen ions (H+). The reduction of H+ ions serves as the main reduction
reaction, so increasing its rate by increasing its effective concentration, thereby increasing
dissolution rates.
At lower temperatures, it was difficult to distinguish the effect of CO2 loading on the
dissolution charge yield; however, at higher temperatures (60-80 oC), both the total charge
and the dissolution charge yield increased significantly with CO2 loading. It is clear from the
results (Figure 7-28) that the dissolution reaction was enhanced by CO2 absorption; for
example, the total charge for the chronoamperogram of iron in 30 wt% MEA at full saturation
(80 oC) was 33 691 C sm-2 (dissolution charge yield was 65.3 %), while that of α = 0.2 was
21 020 C m-2 (dissolution charge yield is 25.3 %).
In addition to iron dissolution, the solubility of FeCO3 is also strongly dependent on the CO2
loading and pH of the system. According to (Nordsveen et al., 2003), FeCO3 precipitation can
occur only at a pH which exceeds a critical value, which depends on the temperature and FeII
concentration. At a pH above this critical pH, once the saturation of FeII and CO32- is
exceeded, precipitation can occur. According to thermodynamic predictions from the
potential-pH diagram for Fe-CO2-H2O (Figure 5-9), this critical pH is ca. 6.4 and for P(CO2)
= 0.1 MPa. The rate of FeCO3 precipitation also increased with CO2 loading (Figure 7-33)
due to the increased concentration of CO32- ions in the bulk solution and FeII ions at the
surface of the iron from the increased corrosion rate.

P a g e | 337
0.4
25C
0.3
40C
0.2
60C
0.1
80C
0
-0.1
-0.2
-0.3
-0.4
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-19: Chronoamperometric data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt%
MEA + H2O (T = 25-80 °C, pH 12.31, α = 0) with a solution velocity of 0.5 mm s-1.
0
Total Charge Density / C m-2 0
-200 -200
-400 -400
-600
-600
-800
-800
-1000
-1200 -1000
0 600 1200 1800 2400 3000 3600
-1200
Time / s 20 30 40 50 60 70 80
25C 40C 60C 80C Temperature / °C
Figure 7-20: Charge density-time data for iron electrode at -0.60 V (AgCl|Ag) in contact with
30 wt% MEA + H2O (T = 25-80 °C, pH 12.31, α = 0) with a solution velocity of 0.5 mm s-1.
600
0.6
500
0.5
Corrosion rate/ mm y-1
400
Fe dissolved/ mg m-2
300 0.4
200 0.3
100 0.2
0 0.1
-100 0
-200
-0.1
-300
0 600 1200 1800 2400 3000 3600 -0.2
-0.3
Time/ s 20 30 40 50 60 70 80
25C 40C 60C 80C Temperature/ °C
Figure 7-21: Corrosion-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA +
H2O (T = 25-80 °C, pH 12.31, α = 0) with a solution velocity of 0.5 mm s-1.
P a g e | 338
9
80C
8
60C
7
40C
6
5
4
3
2
1
0
-1
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-22: Chronoamperometric data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt%
MEA + H2O (T = 25-80 °C, pH 10.70, α = 0.2) with a solution velocity of 0.5 mm s-1.
25000 25000

20000 20000
15000 15000
10000 10000
5000
5000
0
0 600 1200 1800 2400 3000 3600 0
-5000 20 30 40 50 60 70 80
Time / s -5000
Figure 7-23: Charge density-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt%
1600 1.8
1400 1.6
1200 1.4
1000 1.2
800 1
0.8
600
0.6
400
0.4
200
0.2
0 0
0 600 1200 1800 2400 3000 3600 20 30 40 50 60 70 80
Time/ s
MEA + H2O (T = 25-80 °C, pH 10.70, α = 0.2) with a solution velocity of 0. 5 mm s-1.
P a g e | 339
16
14 80C
12 60C
10 40C
25C
8
6
4
2
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-25: Chronoamperogram for iron electrode at -0.60 V (AgCl|Ag) in 30 wt% MEA +
H2O (T = 25-80 °C, pH 9.90, α = 0) with a solution velocity of 0.5 mm s-1.
30000
30000 Total Charge Density / C m-2

25000 25000
20000 20000
15000
15000
10000
10000
5000
5000
0
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
Figure 7-26: Charge density-time data for iron electrode at -0.60 V (AgCl|Ag) in 30 wt%
MEA + H2O (T = 25-80°C, pH 9.90, α = 0.4) with a solution velocity of 0.5 mm s-1.
3000 3.5
2500 3
2000 2.5
1500 2
1000 1.5
1
500
0.5
0
0 600 1200 1800 2400 3000 3600 0
20 30 40 50 60 70 80
Time / s
P a g e | 340
α=0.60 α=0.40 α=0.60 α=0.40 (b)

α=0.20 α=0.00
α=0.20 α=0.00
2200
4000
25oC 40oC
2000
3500
1800

1600 3000
1400
2500
1200
2000
1000
800 1500
600 1000
400
500
200
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
6000 7000
60oC
80oC
5000 6000
5000
4000
4000
3000
3000
2000
2000
1000
1000
0 0
-1000 -1000
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
α=0.60 α=0.40 α=0.60 α=0.40
α=0.20 α=0.00 α=0.20 α=0.00
Figure 7-28: Effect of CO2 loading on the FeII concentration-time data for iron electrode at -
0.60 V (AgCl|Ag) in contact with 30 wt% MEA + H2O (pH 8.03-12.3, α = 0-0.6) with a
solution velocity of 0.5 mm s-1 for a) 25 oC b) 40 oC c) 60 oC d) 80 oC.
P a g e | 341
12
17
11 25oC
α=0.60 16 40oC α=0.60
10 15 α=0.40
α=0.40 14
9 α=0.20 13
α=0.20

α=0.00 α=0.00
8 12
11
7
10
6 9
5 8
7
4 6
3 5
4
2 3
1 2
1
0
0
-1 -1
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
35
60oC 55
30 α=0.60 50 80oC α=0.60
α=0.40 45 α=0.40
25 α=0.20 α=0.20
40
α=0.00 α=0.00
20 35
30
15 25
20
10
15
5 10
5
0
0
-5 -5
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
Figure 7-29: Effect of CO2 loading on current density - time data for iron electrode at -0.60
mm s-1 for a) 25 oC b) 40 oC c) 60 oC d) 80 oC.
P a g e | 342
12000 20000
o
25 C 40o C
10000
15000

8000
10000
6000
4000 5000
2000 0
0
-5000
-2000 0 1000 2000 3000 4000
0 1000 2000 3000 4000
Time / s
Time / s
α=0.60 α=0.40 α=0.60 α=0.40
α=0.20 α=0.00 α=0.20 α=0.00
30000 35000
o 80oC
60 C 30000
25000
25000
20000
20000
15000
15000
10000
10000
5000 5000
0 0
-5000
-5000
0 1000 2000 3000 4000
0 1000 2000 3000 4000
Time / s
Time / s
α=0.00 α=0.20 α=0.60 α=0.40
α=0.40 α=0.60 α=0.20 α=0.00
Figure 7-30: Effect of CO2 loading on charge density-time data for iron electrode at -0.60
mm s-1 for a) 25 oC b) 40 oC c) 60 oC d) 80 oC
P a g e | 343
7000
80C 60C 40C 25C
6000
5000
4000
3000
2000
1000
0
-1000
-0.1 0.1 1E-15
0.2 0.3 0.4 0.5 0.6
CO2 loading / mol CO2 (mol MEA)-1
Figure 7-31: Effect of CO2 loading and temperature on total dissolved [Fe2+] from iron
electrode in 30 wt% MEA + H2O (T = 25-80 oC, α = 0-0.6) at -0.60 V (AgCl|Ag) with a
solution velocity of 0.5 mm s-1.
35000
80C 60C 40C 25C
30000
25000
20000
15000
10000
5000
0
-5000
-0.1 1E-15 0.1 0.2 0.3 0.4 0.5 0.6
Figure 7-32: Effect of CO2 loading and temperature on total charge density for iron electrode
at -0.60 V (AgCl|Ag) in 30 wt% MEA + H2O (T= 25-80 oC, α = 0-0.6) with a solution
velocity of 0.5 mm s-1.
P a g e | 344
0.30
Rate of FeCO3 precipitation / mol m-3 s-1
Johnson and Tomson

0.25
van Hunnik and Hendriksen
0.20
0.15
0.10
0.05
0.00
0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6
Figure 7-33 Effect of CO2 loading on the FeCO3 precipitation rate at the iron electrode in 30
wt% MEA + H2O + CO2 (T = 80 °C, pH 8.03) with a solution velocity of 0.5 mm s-1 after
constant potential electrolyses at -0.60 V (AgCl|Ag).
7.6. Effect of Oxygen Content
Figure 7-34 shows the effect of oxygen on current density - time data for iron at -0.6 V in
30 wt% MEA at full CO2 saturation (pH = 8.03, α = 0.60, T = 25-80 °C). An anoxic system
was created to exclude the influence of oxygen by bubbling high purity nitrogen, prior to CO2
being introduced to the system and maintaining a flow of nitrogen instead of air throughout
the reactor system. Although the absence of oxygen had a considerable effect on the
corrosion of iron in aqueous solutions, the chronoamperograms exhibited similar forms in
both atmospheres, suggesting that the mechanism for iron dissolution and precipitation was
the same in the presence and absence of oxygen. The data displayed a maximum initial value
which decreased rapidly due to the passivation of the metal surface; however, the minimum
current density was significantly greater in the presence of oxygen.

P a g e | 345
24
Air
20
Nitrogen
16
12
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-34: Effect of oxygen on current density – time data for iron electrode at -0.60
V (AgCl|Ag) in 30 wt% MEA + H2O (T = 40oC, pH 8.03, α = 0.6) with a solution velocity of
0.5 mm s-1.
Thus, the possibility of Fenton’s reaction was considered. The oxidation of organic substrates
by FeII ions and H2O2 (resulting from the reduction of oxygen via reaction (7. 23)) was first
described by (Fenton, 1876) after observing the oxidation of tartaric acid by hydrogen
peroxide in the presence of Ferrous ions. Fenton’s mechanism is initiated by the formation of
hydroxyl radical via the reaction (7. 26).
/$ + 22% + 2" ( → 2$ /$ (7. 25)
Fenton’s reaction:
Fe2+ +H 2O2 → Fe3+ + HO • + OH − (7. 26)
Fe3+ +H 2O2 → Fe 2+ + HOO • + H + (7. 27)
Initiating radicals hydroxyl (HO•) and hydroperoxyl (HOO•) could then react with MEA to
form oxidative degradation products such as ammonia or organic acids via a hydrogen
abstraction mechanism as shown in Figure 7-35. (Petryaev et al., 1984) proposed that at pHs
≥ 6, MEA forms a five-member ring by forming hydrogen bonds between HN---O or OH---
P a g e | 346
N. Free radicals such as hydroxyl and hydroperoxyl then abstract a hydrogen atom from the
nitrogen, α-carbon or the °-carbon (Lawal and Idem, 2006). The newly formed amine radical
can then transfer the radical internally through the ring structure, which subsequently results
in cleavage of the N- - -C bond (Chi and Rochelle, 2002). The main degradation products
formed are an amine, an aldehyde or aldehyde radical. The aldehyde radical, if formed, would
then abstract hydrogen from another molecule of MEA, forming an MEA radical and
aldehyde.
Figure 7-35: Hydrogen abstraction mechanism for the oxidative degradation of MEA
(Petryaev et al., 1984)
A similar effect of oxygen was observed for the CO2 loadings of 0.20 and 0.40. The increase
in corrosion rate in the presence of oxygen is attributed to the increased concentration of
oxidising agents in the aqueous amine solution. As described in Chapter 5, the presence of
oxygen enhances the rates of the O2 reduction reaction, which in turn increases the iron
corrosion rate at open circuit.

P a g e | 347
60000
80C O2
50000
80C N2
40000 60C O2
60C N2
30000
40C O2
20000 40C N2
25C O2
10000
25C N2
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-36: Effect of CO2 loading on charge-time data for iron electrode at -
0.60 V (AgCl|Ag) in 30 wt% MEA + H2O (T= 40 oC, pH 8.03, α = 0.6) with a solution
velocity of 0.5 mm s-1
7000
80C O2
6000
80C N2
5000 60C O2
60C N2
4000
40C O2
3000 40C N2
25C O2
2000
25C N2
1000
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-37: Effect of CO2 loading on FeII concentration - time data for iron electrode at -
0.60 V in 30 wt% MEA + H2O (T = 40 oC, pH 8.03, α = 0.6) with a solution velocity of 0.5
mm s-1
P a g e | 348
7.7. Effect of MEA concentration
The effect of MEA concentration on the oxidation kinetics of iron and precipitation of FeCO3
was consistent with that of voltammetric behaviour of the Fe RDE (Chapter 5). The
chronoamperometric experiments were carried out in 5 – 50 wt% MEA at full CO2 saturation
(25 – 80 oC); Figure 7-38 - to Figure 7-55 summarise the results.
The current-time transients (Figure 7-38 and Figure 7-41) for low MEA concentrations (5 –
10 wt%) exhibited current densities that rose initially to a maximum, followed by sharp
decreases to current densities values close to zero. At 80 oC, the chronoamperometric data
also displayed a sharp peak at 789 s (5 wt%) and 600 s (10 wt%), signifying dissolution of
the underlying iron surface which was being covered progressively with a FeCO3 passive
layer. The formation of FeCO3 was also evidenced in FeII concentration-time curves: at low
temperatures (25-40 oC), FeII concentrations increased linearly with time, but at higher
temperatures (60-80 oC), FeII dissolution rates reached a maximum, then started to decline
sharply. At high temperatures, this loss of FeII in the aqueous amine solution could be
attributed to the precipitation of FeCO3 on the surface, which acted as a local sink for FeII and
CO32- ions.
At higher MEA concentrations (≥ 20 wt%), the chronoamperometric data were similar to
those for 30 wt%; the current density decreased with time to a certain value, at which it
remained approximately constant, and the dissolved [FeII] concentration in solution increased
with time and temperature.
At full CO2 loading and temperatures of 25-80 °C, FeII concentrations increased from 5 –
40 wt% and then decreased at concentrations ˃40 wt%. It is also important to note that there
was a significantly larger increase in FeII concentrations between 20 wt% and 30 wt%. This
P a g e | 349
trend was similar to that from the voltammetric results with the rotating disc electrode;
however, the FeII ion concentration was found to be a maximum at [MEA] of 40 wt% as
opposed to 30 wt% with the RDE system.
The charge density-time data had the same form as the FeII concentration-time plots, in which
the total charge increased at lower concentrations (≤40 wt%), then decreased at higher MEA
concentrations. However, the charge yield due to dissolution decreased with increasing
[MEA]; for example, at 25 °C the charge yield due to dissolution was 31.25% at 5 wt% MEA
and 20.8% at 50 wt%.

P a g e | 350
12
10 80C 60C 40C 25C

8
6
4
2
0
-2
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-38: Effect of temperature on current density –time data for iron electrode at -
6000 6000
5000
5000
4000
4000
3000
3000
2000
2000
1000
1000
0
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
Figure 7-39: Effect of temperature on charge density-time data for iron electrode at -
1200 0.8
800 0.6
400
0.4
0
0.2
-400
0
-800
-0.2
-1200
0 600 1200 1800 2400 3000 3600 -0.4
20 30 40 50 60 70 80
Time / s
P a g e | 351
18
Current Density / A m2
16 80C 60C 40C 25C

14
12
10
8
6
4
2
0
-2
0 600 1200 1800 2400 3000 3600
Time/ s
Figure 7-41: Effect of temperature on current density - time for iron electrode at -
7000
7000
6000
6000
5000 5000
4000 4000
3000 3000
2000
2000
1000
1000
0
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
0.6 V (AgCl|Ag) in 10wt% MEA + H2O + CO2 (T = 25-80oC, pH 8.03) with a solution
1400 1.2
1200
1
1000
0.8
800
600 0.6
400 0.4
200 0.2
0
0 600 1200 1800 2400 3000 3600 0
20 30 40 50 60 70 80
Time / s Temperature / °C
80C 60C 40C 25C
P a g e | 352
60
Current Density / A m2
50 80C 60C 40C 25C
40
30
20
10
0
0 600 1200 1800 2400 3000 3600
Time / s
30000 30000

25000 25000
20000 20000
15000
15000
10000
10000
5000
0 5000
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
3000 3.5
2500 3
2000 2.5
1500 2
1000 1.5
500 1
0 0.5
0 600 1200 1800 2400 3000 3600
0
Time/ s 20 30 40 50 60 70 80
P a g e | 353
50
80C
40 60C
40C
30 25C
20
10
0
0 600 1200 1800 2400 3000 3600
Time / s
0.6 V (AgCl|Ag) in 40 wt% MEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution
80000 80000
60000 60000
40000
40000
20000
20000
0
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
Figure 7-48: Effect of temperature on charge density-time data for iron electrode -
0.6 V (AgCl|Ag) in 40 wt% MEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution
9000 10
8000
7000 8
6000
5000 6
4000
4
3000
2000
2
1000
0 0
0 600 1200 1800 2400 3000 3600 20 30 40 50 60 70 80
Time / s
Figure 7-49: Temperature effect on the FeII concentration-time data for iron electrode at -
0.6 V (AgCl|Ag) in 40 wt% MEA + H2O + CO2 (T =25-80 C, pH 8.03) with a solution °

P a g e | 354
35
30 80C 60C 40C 25C

25
20
15
10
5
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-50: Effect of temperature on current density - time data for iron electrode at -
0.6 V (AgCl|Ag) in 50 wt% MEA + H2O + CO2 (T = 25-80 C, pH 8.03) with a solution °
70000
70000
60000 60000
50000 50000
40000 40000
30000
30000
20000
20000
10000
0 10000
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
80C 60C 40C 25C Temperature/ 0C
3500 4.0
3000 3.5
2500 3.0
2000 2.5
1500 2.0
1000 1.5
500 1.0
0 0.5
0 600 1200 1800 2400 3000 3600 0.0
Time / s 20 30 40 50 60 70 80
P a g e | 355
2500 4000
o
25 C 40 oC
3500
2000

3000
2500
1500
2000
1000 1500
1000
500
500
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
50wt% 40wt% 30wt% 50wt% 40wt% 30wt%
20wt% 10wt% 5wt% 20wt% 10wt% 5wt%
7000 9000
50 oC 80 oC
8000
6000
7000
5000
6000
4000 5000
4000
3000
3000
2000
2000
1000 1000
0 0
-1000
-1000 0 600 1200 1800 2400 3000 3600
0 600 1200 1800 2400 3000 3600
Time / s Time / s
50wt% 40wt% 30wt% 50wt% 40wt% 30wt%
20wt% 10wt% 5wt% 20wt% 10wt% 5wt%
Figure 7-53: Effect of [MEA] on FeII concentration-time data for iron electrode at -
0.6 V (AgCl|Ag) in 5-50wt% MEA + H2O + CO2 (pH 8.03) with a solution velocity of 0.5
mm s-1 .
P a g e | 356
24
20 40 oC
25 oC 22
18
20
16 18

14 16
12 14
12
10
10
8
8
6 6
4 4
2 2
0 0
-2
-2
0 600 1200 1800 2400 3000 3600
0 600 1200 1800 2400 3000 3600
Time / s
Time / s
50wt% 40wt% 30wt% 50wt% 40wt% 30wt%
20wt% 10wt% 5wt% 20wt% 10wt% 5wt%
45 55
80 oC
60 oC 50
40
45
35
40
30 35
25 30
20 25
15 20
15
10
10
5
5
0 0
-5 -5
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s
Time / s
50wt% 40wt% 30wt%
50wt% 40wt% 30wt%
20wt% 10wt% 5wt% 20wt% 10wt% 5wt%
Figure 7-54: Effect of [MEA] on current density - time data for iron electrode at -
0.6 V (AgCl|Ag) in 5 – 50 wt% MEA + H2O + CO2 (pH 8.03) with a solution velocity of 0.5
mm s-1.
P a g e | 357
20000 35000
25oC 40oC
18000
30000

16000
14000 25000
12000
20000
10000
8000 15000
6000
10000
4000
5000
2000
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
50wt% 40wt% 30wt% 50wt% 40wt% 30wt%
20wt% 10wt% 5wt% 20wt% 10wt% 5wt%
60000 80000
60oC 80oC
70000
50000
60000
40000
50000
30000 40000
20000 30000
20000
10000
10000
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
50wt% 40wt% 30wt%
40wt% 30wt% 20wt%
20wt% 10wt% 5wt% 10wt% 5wt% 50wt%
Figure 7-55: Effect of [MEA] on the charge density-time data for iron electrode at -0.6 V
(AgCl|Ag) in 5 - 50wt% MEA + H2O + CO2 (pH 8.03) with a solution velocity of 0.5 mm s-1.
P a g e | 358
7.8. Effect of Metal Type
Although carbon steel is the most commonly used material for the piping and equipment in
CO2 absorption plants (DuPart and Rooney, 1993), pure iron was used for most experiments
for which results are reported in this thesis, to elucidate its electrochemical behaviour in
aqueous amine-CO2 solutions, before introducing the additional complexity of alloying
elements. However, most of the published experimental results have been focused on the
behaviour of carbon steel 1020 (Carbon Steel S355), so it was included in this work for
comparison purposes. Carbon steel 1038 (Carbon Steel EN8) is an additional type of steel
that presently is being utilised in amine-based absorption processes (Sun et al., 2011);
however, its corrosion behaviour in amine solvents has yet to be reported. Comparisons were
also made to stainless steel AISI 304, one of the most commonly utilised stainless steels, used
in the most corrosive locations in CO2 absorption plants.
The constant potential measurements of carbon steel S355 and EN8 in aqueous 30 wt% MEA
solutions fully loaded with CO2 resulted in chronoamperograms similar to that of iron (Figure
7-57 & Figure 7-60). The total anodic charges (Figure 7-58 & Figure 7-61 ) and FeII
concentration-time data (Figure 7-59 & Figure 7-62) also increased with temperature.
Overall, the corrosion rates in S355 and EN8 were slightly lower than that of pure Fe; for
example, at 80 oC, the total dissolved FeII concentration at the end of the procedure was 6.364
g m-2 in iron, 5.865 g m-2 in S355 and 5.544 g m-2 for EN8.
Figure 7-63 - Figure 7-65 demonstrate the effect of temperature on stainless steel 304 in
30 wt% MEA at full CO2 saturation; as expected, FeII concentrations were significantly lower
than for iron and carbon steel. In contrast to the corrosion behaviour in iron and carbon steel,
the total charge and FeII dissolution rate decreased with increasing temperature. As predicted
by the potential-pH diagram in Figure 7-56 for the Fe-Cr-Ni-H2O system, the results
P a g e | 359
suggested that the ca. 18 % Cr in the bulk AISI 304 / UNS S30400 stainless steel (0.07 % C,
0.75 % Si, 2.00 % Mn, 0.045 % P, 0.030 % S, 17.5/19.5 % Cr, 8.0/10.5 % Ni, 0.10 % N)
resulted in a passive layer whose composition was dominated by chromium. As a more active
metal than iron, the surface concentration is enhanced to > 50 %; the results in Figure 7-63
suggest the passive layer hindered the iron dissolution process, increasingly so with
increasing temperature. The concentrations of chromium and manganese in solution were
also monitored by the ICP, but were considered insignificant. Hence, stainless steel offers a
potential solution to corrosion in alkanolamine-based CO2 capture processes, but at
significantly greater capital cost (ca. $2 000 t-1) compared to carbon steels (ca. $200 t-1; LME,
2016).
1.8
1.6 CrO42- Ni(OH)3
1.4 HCrO4-
1.2
1 CrO2
0.8
Fe3+ Fe O NiFe2O4
2 3
0.6 Fe 2+
0.4
0.2
Cr2O3 NiO CrO43-
Ni2+
0 Cr(OH)2+
-0.2 Cr3+ Fe(OH)2
-0.4 FeCr2O4
Fe2+
-0.6 Fe3O4
-0.8 Cr2+
-1
-1.2
-1.4 FeCrNi Alloy
-1.6
0 2 4 6 8 10 12 14
pH
Figure 7-56: Potential-pH diagram for Fe-Cr-Ni-H2O system at 298 K, 0.1 MPa and
dissolved activities = 10-4.
P a g e | 360
30
25 80C 60C 40C 25C

20
15
10
5
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-57: Effect of temperature on current density – time data for carbon steel S355
50000
50000
40000 40000
30000 30000
20000
20000
10000
10000
0
0 600 1200 1800 2400 3000 3600
0
Time/ s 20 30 40 50 60 70 80
Figure 7-58: Effect of temperature on charge density – time data for carbon steel S355
electrode at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 oC, pH 8.03) with a
6000 7
5000 6
5
4000
4
3000
3
2000
2
1000
1
0
0 600 1200 1800 2400 3000 3600 0
20 30 40 50 60 70 80
Time / s
Temperature / 0C
80C 60C 40C 25C
Figure 7-59: Temperature effect on the dissolved FeII concentration-time data for carbon steel
S355 electrode at -0.6 V (AgCl|Ag) in 30wt% MEA + H2O + CO2 (T = 25-80 C, pH 8.03) °
with a solution velocity of 0.5 mm s-1.

P a g e | 361
20
80C 60C 40C 25C

15
10
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-60: Effect of temperature on current density – time data for carbon steel EN8
40000 40000
1 h Charge Density / C m-2

35000 35000
30000 30000
25000 25000
20000
20000
15000
10000 15000
5000 10000
0 5000
0 600 1200 1800 2400 3000 3600 0
Time/ s 20 30 40 50 60 70 80
Figure 7-61: Effect of temperature on charge density – time data for carbon steel EN8
solution velocity of 0. 5 mm s-1.
6000 7
5000 6
4000 5
3000 4
2000 3
1000 2
0 1
0 600 1200 1800 2400 3000 3600 0
20 30 40 50 60 70 80
Time/ s
Figure 7-62: Temperature effect on FeII concentration-time data for carbon steel EN8
electrode at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a
P a g e | 362
6
5 25C 40C 60C 80C

4
3
2
1
0
-1
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-63: Effect of temperature on current density - time data for stainless steel electrode
1000
Charge Density/ C m-2
1000 1 h Charge Density / C m-2

800 800
600 600
400
400
200
200
0
0 1000 2000 3000 4000 0
Time/ s 20 30 40 50 60 70 80
Figure 7-64: Effect of temperature on charge density – time data for stainless steel electrode
at -0.6 V (AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25-80 C, pH 8.03) with a solution °
600 0.6
500 0.5
400 0.4
300
0.3
200
0.2
100
0 0.1
0 600 1200 1800 2400 3000 3600
0
Time/ s 20 30 40 50 60 70 80
Figure 7-65: Temperature effect on dissolved FeII concentration - time data for stainless steel
P a g e | 363
2500 4000
25oC 40oC
3500
2000

3000
1500 2500
2000
1000
1500
500 1000
500
0
0 600 1200 1800 2400 3000 3600 0
0 600 1200 1800 2400 3000 3600
Time / s Time / s
Fe S335 Fe S335
EN8 Stainless Steel EN8 Stainless Steel
6000 7000
60oC 80oC
5000 6000
5000
4000
4000
3000
3000
2000
2000
1000 1000
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
Fe S355 Fe S335
EN8 Stainless Steel EN8 Stainless Steel
Figure 7-66: Effect of steel composition on the corrosion-time curve obtained on iron
electrode at -0.6 V (AgCl|Ag) in contact with 30 wt% MEA + H2O + CO2 (pH 8.03) with a
P a g e | 364
7.9. Effect of Amine Type
As discussed in the literature review (Chapter 3), there are several other amines and blends
that are being developed currently and considered to replace MEA as the benchmark for the
CO2 absorption process. Chronoamperometric measurements were made with the iron
electrode at -0.60 V in solutions of 30 wt% methyldiethanolamine (MDEA), 30 wt% 2-
amino-2-methyl-1-propanol (AMP) & 30 wt% aminoethylpiperazine (AEP) at full CO2
saturation (25-60 C). A total amine concentration of 30 wt% and full CO2 saturation provided
°
directly comparable conditions to the previous studies on MEA, from which the influence of
amine type could be considered with a negligible influence of operating parameters. The
results for the MDEA system are displayed in (Figure 7-67 - Figure 7-69), AEP in (Figure 7-
70 - Figure 7-72) and AMP in (Figure 7-73 - Figure 7-75).
The relative dissolution rates of iron in the amine solutions depended on both the amine type
and temperature (Figure 7-77). In comparison to the other amines, MEA was the most
‘aggressive’ towards iron at elevated temperatures (60-80 oC), while the highest FeII
concentrations were determined in MDEA at lower temperatures (25-40 oC).
Sterically-hindered amines such as AMP and cyclic amines AEP have gained substantial
interest recently due to superior CO2 absorption capacity, as well as low degradation rates.
Significantly lower corrosion rates were observed in both AEP and AMP solutions, with
AMP solutions yielding the lowest corrosion rates at all temperatures studied.
Whilst the charge density resulting from iron in MEA and AEP solutions were very close in
magnitude, the charge yield due to dissolution in AEP was significantly lower than that of
MEA. For example, the total charge in MEA solutions was 3.4×104 C m-2 with a dissolution
yield of 65.3 %, while the total charge in AEP was 3.6×104 C m-2 and the yield was 29.8 %,
P a g e | 365
resulting in lower corrosion rates in in AEP. The lowest charge current densities and
dissolution rates were determined in aqueous AEP solutions.
Although AEP has not been featured in any previous corrosion studies based on the CO2
capture process, (Veawab et al., 1999) conducted potentiodynamic tests on carbon steel in
MEA, MDEA and AMP. The results by (Veawab et al., 1999) indicated that the highest
corrosion rates were in MEA, followed by AMP and then MDEA had the lowest corrosion
rate (MEA > AMP > MDEA). Whilst MEA solutions exhibited the highest dissolution rates
at higher temperatures (60-80 °C), this was followed by MDEA, then AEP and AMP
(MEA > MDEA > AEP > AMP). At lower temperatures (25-40 °C) the corrosion rates in the
amine solutions are in the order MDEA > MEA > AEP > AMP.
The decrease in corrosion rates in MDEA at elevated temperatures, in comparison to MEA is
consistent with the results of (Veawab et al., 199), (Tomoe et al., 1996) and (DuPart and
Rooney, 1993). It was suggested that MDEA, a tertiary amine, is less corrosive due to its
inability to form carbamates (RNHCOO-), which would subsequently undergo hydrolysis to
form HCO3-. (Veawab et al., 1999) also proposed that the influence of the amine type on the
dissolution rates was as a result of the total amount of CO2 absorbed. They measured a total
CO2 loading of 0.243 mol CO2 (mol MDEA)-1, which was significantly lower than the
0.565 mol mol-1 loading in MEA solutions under CO2 saturation. However, it is possible to
achieve significantly higher CO2 loadings in MDEA solutions. Although tertiary amines have
a low reactivity to CO2 due to the inability to form carbamates, there is also preferential
formation of HCO3- via the direct hydrolysis of CO2. In fact, one of the main advantages
associated with the use of tertiary amines is their relatively superior capacity to absorb CO2
up to a loading of 1 mol CO2 (mol tertiary amine)-1. Hence, it was expected that the corrosion
rate would increase in the MDEA systems as observed at 25 °C and 40 °C.

P a g e | 366
AMP solutions also have a higher absorption capacity due to the lower stability of the
carbamates formed, which readily hydrolyses to form HCO3- ions. Thus, the reason for
significantly lower corrosion rates observed in AMP systems is unclear, as it was also
expected that higher equilibrium concentrations of HCO3- ions would result in higher
dissolution rates.
Whilst it was evident from the results that the corrosion behaviour of iron in amine solutions
can be attributed partly to the CO2 loading capabilities, it was also clear that the amines also
have additional and amine-specific interactions with iron.

P a g e | 367
25
80C
20
60C
15 40C
10 25C
5
0
-5
0 600 1200 1800 2400 3000 3600
Time / s
(AgCl|Ag) in 30 wt% MDEA + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity
of 0.5 mm s-1.
25000 25000
20000 20000
15000
15000
10000
10000
5000
5000
0
0 600 1200 1800 2400 3000 3600 0
Time / s 20 30 40 50 60 70 80
Figure 7-68: Effect of temperature on charge density - time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% MDEA + H2O + CO2 (T = 25-80 °C, pH 8.03) with a solution velocity
of 0.5 mm s-1.
6000
8
5000
6
4000
3000 4
2000
2
1000
0 0
0 600 1200 1800 2400 3000 3600 20 30 40 50 60 70 80
Time / s Temperature/ oC
80C 60C 40C 25C
(AgCl|Ag) in 30wt% MDEA + H2O + CO2 (T = 25oC, pH 8.03) with a solution velocity of
0.5 mm s-1.
P a g e | 368
30
80C
25
60C
20
40C
15 25C
10
5
0
0 600 1200 1800 2400 3000 3600
Time / s
Figure 7-70: Effect of temperature on current density-time data of iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% AEP + H2O + CO2 (T = 25 - 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1.
40000 40000
30000
30000
20000
20000
10000
10000
0
0 600 1200 1800 2400 3000 3600 0
Time/ s 20 30 40 50 60 70 80
(AgCl|Ag) in 30 wt% AEP + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1.
3500 4.0
Corrosion rate/ mm y--1
3000
2500 3.0
2000
1500 2.0
1000
500 1.0
0
0 600 1200 1800 2400 3000 3600 0.0
Time / s 20 30 40 50 60 70 80
80C 60C 40C 25C Temperature / oC
(AgCl|Ag) in 30 wt% AEP + H2O + CO2 (T = 25 oC, pH 8.03) with a solution velocity of 0.5
mm s-1.
P a g e | 369
30
80C
25
60C
20 40C
15 25C
10
5
0
-5
0 600 1200 1800 2400 3000 3600
Time / s
(AgCl|Ag) in 30wt% AMP + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity of
0.5 mm s-1.
3000 3000
Charge Density/ C m-2

2500 2500
2000 2000
1500
1500
1000
1000
500
0 500
0 1000 2000 3000 4000 0
Time / s 20 30 40 50 60 70 80
(AgCl|Ag) in 30 wt% AMP + H2O + CO2 (T = 25 – 80 oC, pH 8.03) with a solution velocity
of 0.5 mm s-1.
1400 1.4
1200
Corrosion rate/ mm y--1
1.2
1000
800 1.0
600 0.8
400 0.6
200 0.4
0
0.2
-200
0 600 1200 1800 2400 3000 3600 0.0
Time / s 20 30 40 50 60 70 80
(AgCl|Ag) in 30 wt% AMP + H2O + CO2 (T = 25 oC, pH 8.03) with a solution velocity of
0.5 mm s-1.
P a g e | 370
14 18
25oC MEA 40oC
13 MDEA MEA
16
12 AEP MDEA
11 AMP 14
AEP
AMP

10
12
9
8 10
7
6 8
5 6
4
3 4
2 2
1
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
35 50
60oC 80oC
MEA MEA
MDEA 45 MDEA
30
AEP AEP
AMP 40
AMP
25 35
30
20
25
15
20
10 15
10
5
5
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time/ s
Figure 7-76: Effect of amine type on current density – time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% [Amine] (T = 25-80 oC, pH 8.03) with a solution velocity of 0.5 mm s-1.
P a g e | 371
3000 25oC 4500

40oC
4000
2500
3500

2000 3000
2500
1500
2000
1000 1500
1000
500
500
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
MEA MDEA MEA MDEA
AEP AMP AEP AMP
6000 7000
60oC 80oC
5000 6000
5000
4000
4000
3000
3000
2000
2000
1000
1000
0
0
0 600 1200 1800 2400 3000 3600
0 600 1200 1800 2400 3000 3600
Time / s
Time / s
MEA MDEA MEA MDEA
AEP AMP AEP AMP
Figure 7-77: Effect of amine type on FeII concentration-time data for iron electrode at -0.6 V
(AgCl|Ag) in 30 wt% [Amine] (T = 25 - 80oC, pH 8.03) with a solution velocity of 0.5 mm s-
1
.
P a g e | 372
25000 20000
25oC 40oC

20000
15000
15000
10000
10000
5000
5000
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time / s
MEA MDEA AEP AMP MEA MDEA AEP AMP
35000 40000
60oC 80oC
30000 35000
30000
25000
25000
20000
20000
15000
15000
10000
10000
5000
5000
0 0
0 600 1200 1800 2400 3000 3600 0 600 1200 1800 2400 3000 3600
Time / s Time/ s
MEA MDEA AEP AMP MEA MDEA AEP AMP
Figure 7-78: : Effect of amine type on charge density-time data for iron electrode at -0.6 V
(AgCl|Ag) in contact with 30 wt% [Amine] (T =25 - 80oC, pH 8.03) with a solution velocity
of 0.5 mm s-1.
P a g e | 373
7.10. Longer Term Corrosion Measurements
The corrosion behaviour of both iron and carbon steel S355 in aqueous 30 wt% MEA
solution at full CO2 saturation (α = 0.60, T = 25oC) was determined at open circuit
potential over a 28 day period and Fe2+ concentrations were measured using ICP-OES. In
order to maintain a constant loading, CO2 was purged continuously in the amine solution
while the pH of the solution was monitored throughout the duration of the experiment.
Figure 7-79 shows the open circuit potential (OCP) measured and the resulting corrosion-
time curves over the 28 day period are shown in Figure 7-80.
-0.3
Open Cell Potential (AgCl|Ag) / V
Carbon Steel S355

-0.4
Iron
-0.5
-0.6
-0.7
-0.8
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Time / Days
Figure 7-79: Time dependence of open circuit potentials of iron and steel electrodes at -0.6 V
(AgCl|Ag) in 30 wt% MEA + H2O + CO2 (T = 25 oC, pH 8.03) with a solution velocity of 0.5
mm s-1.
The iron electrode maintained an OCP of -0.76 ± 0.07 V (AgCl|Ag) for the first 22 days,
which then increased slowly from a value of -0.75 V to -0.68 V by the end of the experiment.
On the other hand, the steel sample displayed an OCP that remained relatively constant at -
P a g e | 374
0.73 ± 0.03 V for the first 21 days then at day 22 increased in value suddenly to -0.39 V,
followed by a slow increase to -0.31 V at the end the end of the experiment.
140
120
100
80
60
40
Carbon Steel S355
20 Iron
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Time / Days
Figure 7-80: Time-dependence of iron(II) concentrations dissolved from iron and steel S355
electrodes at -0.6 V (AgCl|Ag) in contact with 30 wt% MEA + H2O + CO2 (T = 25 oC, pH
8.03) with a solution velocity of 0.5 mm s-1.
Initially, the measured FeII concentration was almost identical for steel and iron samples.
However, as time progressed, the dissolution rates on both electrodes diverged, with the rate
of dissolution being generally higher for the steel sample than for iron. However, while the
rate of iron corrosion appeared to increase with time, that of steel appeared to decrease after
22 days, coinciding with the sudden change in the steel’s OCP and can be attributed to the
formation of a protective passive layer. A thick black corrosion product which covered the
entire surface of the steel exposed to the amine solution was visible even to the naked eye,
whereas the passive layer formed on the iron sample was less visible (Figure 7-82).
Although the characterization of the passive film was not within the scope of this work,
according to the potential-pH diagram for Fe-CO2-H2O systems, at the initial OCP of -0.73±
0.03V, the carbon steel sample as for iron, was predicted thermodynamically to correspond to
P a g e | 375
active iron dissolution and FeCO3 precipitation. Lastly, it is also important to note that as
time progressed, the initially clear solution amine solution with a slight yellow tint became
increasingly darkened in colour with time (Figure 7-81). It was unclear if this was due to an
accumulation of contaminants resulting from the oxidative and CO2–induced degradation of
MEA, or from the accumulation of corrosion products in solution.
Figure 7-81: Images of the flow reactor system for the longer term studies at open circuit
potential for a) 7 days b) 14 days c) 28 days
P a g e | 376
Iron
Steel
Figure 7-82: Images of the iron and carbon steel S355 electrodes after immersion in 30 wt%
MEA + H2O + CO2 (T = 25oC, pH 8.03) with a solution velocity of 0.5 mm s-1 at open circuit
potential for 28 days.
P a g e | 377
7.11. References
Amundsen, T. G., Øi, L. E. & Eimer, D. A. 2009. Density and Viscosity of

Monoethanolamine + Water + Carbon Dioxide from (25 to 80) °C. Journal of
Chemical & Engineering Data, 54, 3096-3100.
Azoulay, I. Rémazeilles, C. Refait, Ph. 2012. Determination of standard Gibbs free energy of
formation of chukanovite and Pourbaix diagrams of iron in carbonated media,
Corrosion Science 58. 229–236.
Bard, A.J., Parsons, R. and Jordan, J. 1985. Standard potentials in aqueous solution. Marcel
Dekker, New York.
Books, L.S., Cherie Harris & Kasem, K. K. 2006. Electrochemical Behaviour of Hexacyano
Iron (III/II) in Frozen Aqueous Electrolytes. American Journal of Undergraduate
Research 5.
Chi, S. & Rochelle, G. T. 2002. Oxidative Degradation of Monoethanolamine. Industrial &
Colli, A. N., Toelzer, R., Bergmann, M. E. H. & Bisang, J. M. 2013. Mass-transfer studies in
an electrochemical reactor with a small interelectrode gap. Electrochimica Acta, 100,
78-84.
Dupart, M. S. & Rooney, P. C. 1993. Corrosion in Alkanolamine Plants: Causes and
Minimization. Hydrocarbon Processing.
Fenton, H. J. H. 1876. On a New Reaction of Tartaric Acid. Chem News, 33, 190-190.
Johnson, M. L. & Tomson, M. B. 1991. Ferrous carbonate precipitation kinetics and its
impact on CO2 corrosion NACE International Houston, Texas, 91.
Jou, F.-Y., Mather, A. E. & Otto, F. D. 1995. Solubility of CO2 in a 30 mass percent
monoethanolamine solution. Canadian Journal of Chemical Engineering, 73, 140-
147.
Kohl, A. L. & Nielsen, R. B. 1997. Gas Purification (5th edition). Houston: Gulf Publishing
Lawal, A. O. & Idem, R. O. 2006. Kinetics of the Oxidative Degradation of CO2 Loaded and
Concentrated Aqueous MEA-MDEA Blends during CO2 Absorption from Flue Gas
Streams. Industrial & Engineering Chemistry Research, 45, 2601-2607.
LME. 2016. Retrieved from https://www.lme.com/. ( Accessed 21st January 2016).
Nainar, M. & Veawab, A. 2009. Corrosion in CO2 Capture Process Using Blended
Monoethanolamine and Piperazine. Industrial & Engineering Chemistry Research,
48, 9299-9306.
Nesic, S. & Lee, K.-L. J. 2003. A Mechanistic Model for Carbon Dioxide Corrosion of Mild
Steel in the Presence of Protective Iron Carbonate Films—Part 3: Film Growth
Model. Corrosion, 59, 616-628.
Nordsveen, M., Nesic, S., Nyborg, R. & Stangeland, A. 2003. A Mechanistic Model for
Carbon Dioxide Corrosion of Mild Steel in the Presence of Protective Iron Carbonate
Films—Part 1: Theory and Verification. Corrosion, 59, 443-456.
P a g e | 378
Petryaev, E. P., Pavlov, A. V. and Shadyro. 1984. Homolytic Deamination of Amino

Alcohols. Zh Org Khim, 20, 29-34.
Pickett, D. J. 1979. Electrochemical Reactor Design. Elsevier, Amsterdam, The Netherlands.
Pickett, D. J. and Ong, K. L. 1974. The Influence of Hydrodynamic and Mass Transfer
Entrance Effects on the Operation of a Parallel Plate Electrolytic Cell. Electrochimica
Acta, 19, 875-882.
Reddi, L. M. & Inyang, H. I. 2000. Geoenvironmental Engineering: Principles and
Applications. CRC Press.
Revie, R. W. 2015. Oil and Gas Pipelines: Integrity and Safety Handbook, Wiley.
Rousar, I., Hostomsky, J., Cezner, V. & Štverak, B. 1971. Limiting Local Current Densities
for Electrodes Located on the Walls of a Rectangular Channel with Laminar Flow;
Asymptotic Solution and Experimental Verification. Journal of The Electrochemical
Society, 118, 881-884.
Rousar, M., AND Kimla 1986. Electrochemical Engineering, New York, Elsevier
Sun, Y., Remias, J. E., Neathery, J. K. & Liu, K. 2011. Electrochemical study of corrosion
behaviour of carbon steel A106 and stainless steel 304 in aqueous monoethanolamine.
Corrosion Engineering, Science and Technology, 46, 724-731.
Tomoe, Y., Shimizu, M. & Kaneta, H. 1996. Active Dissolution and Natural Passivation of
Carbon Steel in Carbon Dioxide-Loaded Alkanolamine Solutions. NACE
International.
Tong, D., Trusler], J. P. M., Maitland], G. C., Gibbins, J. & Fennell, P. S. 2012. Solubility of
carbon dioxide in aqueous solution of monoethanolamine or 2-amino-2-methyl-1-
propanol: Experimental measurements and modelling. International Journal of
Greenhouse Gas Control, 6, 37-47.
Van Hunnik, E. W. J. & Hendriksen, E. L. J. A. 1996. The formation of protective FeCO3
corrosion product layers. NACE International Houston, Texas, 96.
Veawab, A., Tontiwachwuthikul, P. & Chakma, A. 1999. Corrosion Behavior of Carbon
Steel in the CO2 Absorption Process Using Aqueous Amine Solutions. Industrial &
Walsh 1993. A First Course in Electrochemical Engineering.
Wragg, A. A. & Leontaritis, A. A. 1997. Local mass transfer and current distribution in
baffled and unbaffled parallel plate electrochemical reactors. Chemical Engineering
Journal, 66, 1-10.
P a g e | 379
8. Chapter 8
Conclusions and Future Work

The principal results of the project are summarised in the three objectives:
1. Quantification of the electrochemical kinetics and mechanisms of iron oxidation and hence
corrosion rates, in aqueous MEA-CO2-H2O systems, as a function of experimental
variables.
2. Mechanistic modelling for iron oxidation kinetics in aqueous MEA-H2O-CO2.
3. Description of the dependence of iron(II) dissolution rates on experimental variables, using
a specially-designed electrochemical flow reactor with subsequent analysis of dissolved
iron(II) concentrations, enabling deconvolution of total oxidation current densities in the
small scale electrochemical experiments, into rates of reactions leading to adsorbed and
dissolved products.
P a g e | 380
8.1. Summary of this project
As iron is the main component of carbon steel materials used for fabrication of most equipment
in amine scrubbing plants, this project aimed to elucidate experimentally, supported by
thermodynamic and electrochemical modelling, the electrochemical behaviour of iron in
aqueous MEA-CO2-H2O solutions.
8.2. Rotating Disc Electrode and Electrochemical Quartz Crystal Results
One of the main objectives of this project was to quantify the electrochemical behaviour of iron
in aqueous MEA-CO2-H2O systems and hence its corrosion rates, as functions of experimental
variables. Hence, the kinetics of iron oxidation in aqueous MEA solutions were determined by
cyclic voltammetry with a RDE and an EQCM as functions of electrochemical parameters.
Electrode potential-pH and activity-pH diagrams for Fe-H2O systems and Fe-water–CO2
systems were used to assist with reaction assignments. The results obtained showed that the
active to passive transition of iron in MEA was strongly dependent on both the hydrodynamic
conditions of the system and on the composition of MEA solutions. In unloaded MEA solutions
at pH ca. 12, iron exhibited passive electrochemical behaviour implying a negligible corrosion
rate. However, on loading the MEA solution with CO2, protonated amine (RNH3+) and
hydrogen carbonate ions (HCO3-) are formed, which subsequently dissociate to produce
hydrogen ions (H+). This resulted in a decrease in the bulk pH from ca. 12 in unloaded solutions
to ca. 8 in fully CO2-saturated solutions, switching the system from a passive to an active state,
in which active dissolution was thermodynamically favourable.
Based on the voltammograms of iron in MEA solutions fully loaded with CO2, it was
established that the active dissolution and passivation of iron involves a minimum of two stages
due to the appearance of at least two anodic current peaks. The first anodic stage occurred in a
potential range in which only FeII oxidation products were predicted thermodynamically. This
P a g e | 381
first anodic stage was also assigned to a surface transformation as the charge and magnitude of
the current peak within this stage was independent of the RDE rotation rate and pH whilst it
increased linearly with potential scan rate and temperature. As a result of these observations,
and in accordance with the thermodynamic predictions, the first stage of oxidation was
assigned to formation of a pre-passive layer of Fe(OH)2 by
Fe + 2H& O ↔ Fe(OH)& + 2H , + 2e- (8. 1)
The second oxidation stage also occurred within a potential range at which only FeII species
are predicted; current densities (jpII) increased linearly with rotation rate (f1/2), suggesting that
diffusion at the electrode surface limited iron oxidation rates. The results from the EQCM
clearly indicated a mass loss associated with this second stage, so it was attributed to the iron
dissolution as the predominant process, through this passive layer via the reaction:
Fe ↔ Fe&, + 2e- (8. 2)
The anodic reaction rates within the active region were also strongly dependent on the solution
temperature, as the charge density within the region of iron dissolution doubled at 80oC in
comparison to 25oC.
The effect of pH was distinguished from that of CO2 by studying the behaviour of iron in CO2-
free solutions of MEA which had been adjusted to the same pH as that of fully CO2 saturated
solutions with the addition of H2SO4. In the absence of CO2, even at the same pH, the
electrochemical behaviour was fundamentally different, resulting in significantly lower current
densities in the potential range associated with the dissolution of Fe to FeII(aq).
The rate of iron dissolution was controlled primarily by the kinetics of one or several reactions
which result in the evolution of hydrogen. The hydrogen evolution reaction is considered to
the predominant cathodic reaction, the rate of which is limited by the diffusion of H+ ions to
the iron electrode surface from the bulk solution.

P a g e | 382
2H , + 2e- → H& (8. 3)
However, it has been established that the mechanism by which H+ ions are supplied to the
electrode surface in aqueous CO2 solutions differs from that in strong acid solutions, in which
the diffusion of H+ ions is the principal rate controlling reaction. In MEA-CO2-H2O systems,
the flux of H+ ions was increased by the presence of the reaction products, RNH3+ and HCO3,
resulting in high dissolution current densities in the presence of CO2 as compared to those in
MEA solutions with pHs adjusted by H2SO4. The presence of the protonated amine and
hydrogen carbonate ions provides a reservoir of H+ ions, thereby increasing the corrosion rate
by serving as an additional source of H+ ions via dissociation. Measurements in NaHCO3 /
Na2CO3 solutions at pH ca. 8 showed a linear dependence of jpII on HCO3- concentrations, also
leading to the conclusion that the kinetics of the iron dissolution is related to a process under
HCO3- ion transport control.
In MEA-CO2-H2O solutions, the Fe dissolution process was inhibited by the formation of a
passive film. At the outer layer, precipitation of FeCO3 is thermodynamically feasible if the
concentration of dissolved FeII(aq) ions at the iron | solution interface exceeds the solubility
limit. The appearance of a third oxidation stage in quiescent solutions accompanied with a
relative mass gain within this region implied the precipitation of FeCO3 via:
/0 &, (12) + 345 &- (12) ↔ /0345 (6) (8. 4)
or,
/0 &, + 7345 - + 47- → /0345 + 7& 4 (8. 5)
The precipitation of FeCO3 on the iron surface was induced only at high local concentrations
of both FeII(aq) and CO32- / HCO3- ions. Thus, the disappearance of the third oxidation stage
P a g e | 383
with rotation of the RDE was attributed to dispersion of FeII(aq), decreasing the local
supersaturation and so preventing FeCO3 precipitation.
Based on the potentiodynamic study of iron in aqueous MEA solutions, it was clear that
operating parameters such as temperature, CO2 loading amine concentration and oxygen
content significantly altered the electrochemical response. The changes in the electrochemical
behaviour can be attributed to the impact on the solution chemistry and ionic composition. A
Kent-Eisenberg type model was used to estimate the concentrations of the solutions species in
the bulk solution as a function of the operating parameters. The electrochemical measurements
were analysed using the Tafel extrapolation method; analysis of the resulting kinetic data
enabled corrosion rates to be predicted as functions of the experimental variables.
Kinetic analysis of the experimental results showed that temperature and CO2 loading were the
most significant parameters affecting corrosion rates of iron / steel in MEA-CO2-H2O. For
example, corrosion rates of iron in fully CO2-saturated MEA solutions increased from
6.2 mm a-1 at 25oC to 17.7 mm a-1 at 80oC, compared with 0.3 mm a-1 in unloaded EA solutions
at 25oC. Increasing the CO2 loading increased the iron oxidation rate, principally by shifting
the solution pH and increasing the rate of the hydrogen evolution reaction. At amine
concentrations less than 30 wt%, the corrosion rates increased with increasing MEA
concentration, while at concentrations more than 30 wt%, corrosion rates decreased implying
a more complex behaviour that had not been observed previously in Fe-MEA-H2O studies in
which it had been concluded that corrosion rates increased with amine concentration.
As expected, addition of (dissolved) oxygen increased iron corrosion rates, e.g. from 5.8 mm
a-1 to 6.2 mm a-1 in 30 wt% MEA system at 25oC, due to its acting as an additional oxidant.
As expected for a system exhibiting dissolution-precipitation-passivation behaviour, current
densities due to iron dissolution increased with increasing mass transport rate coefficients,
P a g e | 384
which also decreased FeII concentrations at the iron | solution interface, minimising or even
preventing FeCO3 precipitation.
8.3. Mechanistic Modelling
A mechanistic model was developed for the uniform corrosion of iron in MEA-CO2-H2O based
on the proposed corrosion mechanisms from the voltammetric results. The model was
constructed by taking into account the CO2 absorption equilibrium reactions and the
electrochemical reactions at the iron | solution interface, with the aim to predict concentrations
of the solution species and corrosion rates as a function of operating conditions.
Equilibrium concentrations of the amine species (RNH2, RNH3+, RNHCOO-), carbon(IV)
species (HCO3-, CO32-) and hydrogen ions (H+) were calculated with the Kent-Eisenberg type
model, the results of which were in accordance with the experimental data for the pH and CO2
loading. The electrochemical reactions incorporated in the model were the anodic dissolution
of iron and the cathodic reduction of H+, direct water reduction and the reduction of oxygen.
The corrosion model was developed by simulating polarization curves based on the species
concentrations and the transport limited current densities of the iron RDE defined by the Levich
equation. Values of exchange current densities and transfer coefficients for iron oxidation and
reduction of protons, water and oxygen were obtained from the literature.
The thermodynamic speciation model predictions agreed well with the experimental data based
on the solution pH and CO2 loading and the concentrations of RNH3+, HCO3- and consequently
H+ were predicted to increase with CO2 loading and temperature. The electrochemical model
predictions were in modest to fairly good agreement with experimental data, with the model
generally predicting higher corrosion rates than experimental data. This discrepancy was
attributed to the model neglecting the effect of FeCO3 precipitation, which would have partially
passivated the iron surface and impeded the active dissolution process.
P a g e | 385
In addition to predicting corrosion rates, the mechanistic model was also used to deconvolute
the effects of CO2 loading and MEA concentrations on the iron corrosion rates and
mechanisms. Contrary to suggestions in literature that the direct reduction of HCO3-
participates in the corrosion process and acts as the primary oxidant in CO2 loaded amine
systems, the direct reduction of HCO3- did not have to be considered to account for the higher
corrosion rates, which would otherwise have been significantly higher than the experimental
results. The model predicted that the primary role of HCO3- in the corrosion process was to
provide an additional reservoir of H+ ions via dissociation.
The mechanistic model also clarified the decrease in corrosion rates observed at higher MEA
concentrations as being due to increased viscosities, thereby decreasing rates of oxidant
diffusion, FeII dissolution and corrosion.
8.4. Electrochemical Flow reactor
The final objective for this project was to investigate the components of the corrosion process
and elucidate their kinetics, mechanisms and rate determining steps. Hence, it was important
to combine electrochemical techniques with chemical analysis of sample solutions by
inductively coupled plasma optical emission spectrophotometry (ICP-OES) to discriminate
between total oxidation current densities from those leading to dissolution, rather than adsorbed
products such as Fe(OH)2 or FeCO3. Thus, an electrochemical flow reactor was designed and
built in-house to enable the measurement of dissolved iron concentrations under typical amine-
based CO2 absorption conditions. ICP-OES was used to measure dissolved FeII(aq)
concentrations in the amine solutions after constant potential measurements, enabling the
charge yield for the dissolution to be determined by relating the amount of charge passed to the
amount of dissolved FeII (aq) measured.

P a g e | 386
The mass transport behaviour of the reactor was characterised as a function of solution flow
rate by the transport controlled reduction of hexacyanoferrate(III) ions at a platinised titanium
electrode. This enabled predictions of the conditions required for FeCO3 and Fe(OH)2
precipitation, as well as the rate of FeCO3 precipitation from the current fluxes and measured
dissolved FeII(aq) concentrations in solution.
Based on measurements conducted to study the effect of the electrode potential, the potential
of -0.60 V (AgCl|Ag) was chosen as a suitable potential for the subsequent constant electrode
potential measurements within the active region, in which iron dissolution dominated the
anodic current measured. In the potential range -0.88 V ≤ E ≤ -0.52 V, precipitation of
protective FeCO3 and Fe(OH)2 was predicted thermodynamically.
In a complementary set of results to those obtained using the RDE, increasing solution flow
rates increased corrosion rates, partly by enhanced mass transport of the oxidising species to
the surface of the iron and prevention or decreased of FeCO3 supersaturation.
Increasing the temperature and CO2 loading enhanced iron dissolution rates, whilst the rate of
formation of FeCO3 was also increased. The supersaturation of FeCO3 was exceeded at lower
temperatures (≤ 60oC) and high CO2 loading. However, the precipitation rate was significantly
lower than the corrosion rate, resulting in the formation of non-adherent FeCO3 films, in
agreement with reports from the literature. At 80oC and high CO2 loadings, a more adherent
film was formed, resulting in a significant decrease in corrosion rates, in accordance with the
conditions of formation for a protective FeCO3 film reported in the literature.
As for the voltammetric measurements, corrosion rates increased with MEA concentrations
from 5 – 40 wt%, then corrosion rates decreased at concentrations > 40 wt%. At low MEA
concentrations, increasing equilibrium concentrations of H+ ions with increasing MEA
concentrations, led to increased corrosion rates. However, increased viscosity of the solutions
P a g e | 387
at MEA concentrations > 40 wt%, resulted in dissolution rates being limited by the decreased
mass transport rates of the oxidising species H+, H2O and O2 to the iron electrode. It is
important to note that concentrations of MEA solutions used in amine scrubbing plants have
been limited to ca. 30 wt %, specifically to limit corrosion rates. Hence, this finding that
corrosion rates are lowered at higher MEA concentrations due to increasing viscosity may be
of significant economic impact to the CO2-capture industry, as higher CO2 absorption
capacities are attainable at higher amine concentrations, thereby decreasing sizes and hence
costs of absorber-desorber equipment.
The type of amine also affected corrosion rates. Significantly lower corrosion rates were
measured in the other commercially-available solvents tested: methyldiethanolamine (MDEA),
2-amino-2-methyl-1propanol (AMP) and aminoethylpiperazine (AEP). Iron corrosion rates in
the amine solutions fully saturated with CO2 decreased in the order:
MEA > MDEA > AEP > AMP at higher temperatures (60-80 °C) and
MDEA > MEA > AEP > AMP at lower temperatures (25-40 C). Thus, sterically hindered
°
amines such as AMP and cyclic amines AEP may provide a suitable alternative to MEA, due
to their high CO2 absorption capacity and resistance to degradation. However, all these benefits
associated with the use of alternative amines come at a considerably higher cost. Carbon steel
exhibited similar dissolution rates to those of iron, whereas those for stainless steel were
significantly lower. Hence, stainless steel offers a potential solution to corrosion in the amine
scrubbing process, but at significantly higher costs.
8.5. Future Work Proposals
Being the most extensively used amine for CO2 capture, MEA acts as the benchmark for the
amine scrubbing process, so was the primary focus of most of the studies within this thesis. In
more recent years, several other promising solvents such as methyl diethanolamine (MDEA),
P a g e | 388
2-amino-2-methyl-propaol (AMP) and piperazine (PZ) have been poised possibly to replace
MEA as the primary solvent of choice, based on their efficiencies and superior CO2 adsorption
capacities. However, the electrochemical kinetics of iron and steel in these alternative amines
has yet to be studied significantly.
Due to the limitations of the electrochemical apparatus, the experimental studies in this thesis
were restricted to a maximum of 80oC, whilst the operating temperature of an absorption plant’s
stripper is typically 120oC. Based on the results from the higher temperature measurements
(60-80oC) in the electrochemical flow reactor, it is possible that the corrosion rates may
decreased significantly under amine regenerating conditions, due to the precipitation of FeCO3.
Therefore, it may also be important to investigate the kinetics at higher solution temperatures
(80oC-120oC).
Whilst the soluble corrosion products resulting from the corrosion reaction have been
quantified within this thesis, efforts must also be devoted to identifying the insoluble corrosion
products such as FeCO3 and Fe(OH)2 as functions of the operating conditions. In a
complementary set of studies to the electrochemical techniques used within this thesis, it is also
recommended that analytical techniques such as in-situ (Fourier Transform Infrared) FT-IR
spectroscopy be employed to characterize the corrosion products formed on the surface of the
electrode during electrochemical oxidation as function of experimental variables. This can be
used to define further the favourable conditions under which protective FeCO3 layers may be
formed, which is key to decrasing corrosion rates in CO2 - containing systems.
Based on the comprehensive results on the kinetics of iron in MEA-CO2 systems, the effect of
MEA concentrations on predicted iron corrosion rates raises questions on the effectiveness of
the MEA concentrations in amine scrubbing plants being limited to 30wt%, specifically to limit
corrosion issues. However, it is also important to note that concentrated amine solutions are
also more prone to degradation. According to some suggestions from the literature, degradation
P a g e | 389
products and contaminants, particularly heat stable salts, enhance corrosion rates in MEA
plants either by decreasing solution pHs or by forming complexes with iron. Ultimately, future
research should aim at addressing both the degradation and corrosion processes in amine
scrubbing plants, leading to the development of more robust, less prone to degradation and less
corrosive amines for the CO2 absorption process.
The mechanistic corrosion model needs to be developed further into a macro-kinetic model
describing the electrochemical flow reactor, in order to predict the time-dependent iron
corrosion in amine-CO2 systems. The model should also be extended to include the use of other
amines and to determine the conditions under which FeCO3 could precipitate.
Finally, based on the information obtained, strategies to mitigate or minimise steel corrosion
and amine decomposition should be devised.

P a g e | 390
9. Chapter 9
Appendices
P a g e | 391
I. CO2 Solubility Data
25oC from (Jou et al., 1995)
PCO2 (kPa) αCO2
0.00213 0.211
0.00927 0.323
0.06 0.439
0.392 0.494
2.8 0.54
11.8 0.583
55.1 0.648
97.2 0.694
297 0.809
596 0.844
996 0.923
2996 1.039
6402 1.102
6425 1.109
40oC from (Jou et al., 1995) and (Tong et al., 2012)
Reference: Jou et al., 1995 Reference: Tong et al., 2012
PCO2 (kPa) αCO2 PCO2 (kPa) αCO2
0.00147 0.0888 3.95 0.530
0.00896 0.203 19.12 0.585
0.0677 0.365 71.50 0.632
0.604 0.461 101.00 0.639
2.57 0.513 159.40 0.668
8.09 0.557 161.52 0.670
36.1 0.609 211.92 0.687
103 0.646 297.02 0.706
293 0.709 408.17 0.748
593 0.794
993 0.844
2992 0.965
5986 1.049
9969 1.097
14945 1.132
19914 1.18
P a g e | 392
PCO2 (kPa) αCO2
0.00428 0.0564
0.0193 0.119
0.0579 0.206
0.528 0.389
2.01 0.438
11 0.504
34.1 0.565
82 0.575
282 0.642
582 0.701
982 0.749
2977 0.88
5968 0.942
9959 1.027
14931 1.062
19893 1.109
PCO2 (kPa) αCO2

0.00296 0.0174
0.00484 0.0236
0.0465 0.0781
0.0992 0.118
0.278 0.187
2.67 0.348
16 0.46
56 0.517
235 0.576
958 0.68
2951 0.797
5943 0.875
9925 0.953
14902 0.998
P a g e | 393
II. Cyclic Voltammetry Data

Table 9-5: Effects on jPII of rotation rate, temperature (25-80 °C) and aqueous MEA
concentration (10-60 wt%) for Fe RDE in MEA + H2O + CO2 at pH ̴ 8.01; 10 mV s-1.
Temperature/ oC 25 40 60 80 25 40 60 80
[MEA ]/ wt% 5 10
RPM
1500 48.01 61.00 54.80 59.36 59.76 79.34 95.68 136.85
960 38.30 53.62 44.48 49.14 53.31 66.70 76.39 116.49
540 26.24 41.55 33.07 37.33 42.49 49.76 58.62 101.08
240 14.11 29.02 20.54 28.49 31.57 31.61 41.79 88.77
60 6.32 18.21 9.49 23.28 14.81 19.68 28.03 63.43
0 2.41 17.56 13.41 11.74 18.05 18.87 49.47
[MEA ]/ wt% 20 30
RPM
1500 68.21 84.10 142.55 185.57 72.02 119.53 154.46 172.64
960 59.40 72.15 126.20 168.89 63.19 101.58 132.05 155.43
540 48.79 58.67 103.36 148.89 52.87 85.25 110.66 143.73
240 34.35 50.45 83.21 123.00 40.17 57.04 74.05 109.26
60 26.93 30.53 65.39 85.13 30.39 38.30 54.29 83.93
0 22.68 26.76 65.39 77.56 32.80 33.89 54.29
[MEA ]/ wt% 40 50
RPM
1500 55.47 113.83 188.11 216.02 41.71 81.40 138.83 201.85
960 40.49 92.16 154.24 209.12 37.75 62.28 114.12 162.01
540 32.04 70.18 111.80 201.64 32.56 51.83 91.49 115.14
240 19.37 50.22 86.33 164.59 26.42 41.69 70.08 88.53
60 12.05 26.48 56.91 112.35 22.34 28.64 49.25 75.54
0 12.05 21.64 27.48 42.75 100.17
[MEA ]/ wt% 60
RPM
1500 37.39 81.40 122.95 175.75
960 34.30 72.35 104.37 156.23
540 30.60 57.20 80.67 118.64
240 24.26 41.12 66.97 89.45
60 16.18 27.10 44.43 69.63
0 33.68 42.33
P a g e | 394
III. Effect of MEA Concentration
80
60
40
20
0
5wt%
10wt%
-20
20wt%
30wt%
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

80
60
MEA%
40
20
0
30wt%
40wt%
-20
50wt%
60wt%
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

P a g e | 395
140
120
100
80
60
40
20
0 5wt%
10wt%
-20 20wt%
30wt%
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 40 °C; 10 mV s-1
140 30wt%
120 40wt%
50wt%
100
MEA% 60wt%
80
60
40
20
-20
-40
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
RDE rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 40 °C; 10 mV s-1
P a g e | 396
250 5wt%
10wt%
20wt%
200
30wt%
150
100
50
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
250 30wt%
40wt%
50wt%
200 60wt%
150
100
50
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
P a g e | 397
250 5wt%
10wt%
20wt%
200
30wt%
40wt%
150
100
50
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
rotating at f = 1500 rpm in MEA+ CO2 at pH ̴ 8.01 and T = 80 °C; 10 mV s-1
250
40wt%
50wt%
200 60wt%
150
100
50
-50
-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
P a g e | 398
IV. The influence of ionic composition on the electrochemical behaviour of iron

in sodium carbonate-hydrogen carbonate solutions at pH 8.01
40
30 (II)
20
(III)
10
(I)
0
0.148M NaHCO3/0.001M Na2CO3
0.148 M NaHCO3/ 0.001 M Na2CO3
-10 0.295MNaHCO2/0.002M Na2CO3
0.295 M NaHCO3/ 0.002 M Na2CO3
(IV) 0.500M NaHCO3/ 0.003M Na2CO3
0.500 M NaHCO3/ 0.003 M Na2CO3
-20 0.750M NaHCO3/ 0.005M Na2CO3
0.750 M NaHCO3/ 0.005 M Na2CO3
1.00M NaHCO3/ 0.007M Na2CO3
1.000 M NaHCO3/ 0.007 M Na2CO3
-30
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Figure 9-9: Effect of NaHCO3 + Na2CO3 concentration on cyclic voltammograms of an iron

40
Peak Current Density (jpII) / A
35
30
25
m-2
20
15
10
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
NaHCO3 concentration/ M
Figure 9-10: Effect of NaHCO3 concentration on current density for peak II of stationary Fe
electrode in NaHCO3 + Na2CO3 solutions (T = 25 °C, pH 8.01).
P a g e | 399
V. Kinetic Data obtained from Polarization curves

rotating at f = 1500 rpm in 10 wt% MEA+ CO2 at pH ̴ 8.01; 10 mV s-1
Temperature/ Rotation ba/ mV bc/ mV Ecorr / Corrosion rate/
o
C / RPM dec-1 dec-1 mV jcorr/ µA cm-2 mm y-1
25 1500 128.39 144.77 -740.30 374.33 4.385
25 960 117.13 136.84 -725.57 347.69 4.072
25 540 107.35 139.49 -730.84 344.44 4.034
25 240 65.84 130.35 -747.18 296.95 3.478
25 60 215.41 123.99 -755.47 283.09 3.316
25 0 245.03 118.81 -777.85 240.80 2.820
40 1500 94.12 148.17 -746.31 481.38 5.639
40 960 111.46 129.92 -752.49 424.39 4.971
40 540 102.94 129.64 -758.02 370.92 4.345
40 240 110.16 127.81 -761.12 348.95 4.087
40 60 156.88 159.85 -762.24 332.15 3.890
40 0 128.93 140.65 -772.37 317.03 3.713
60 1500 84.70 125.21 -772.85 532.85 6.241
60 960 66.64 127.36 -776.90 511.77 5.995
60 540 80.93 111.68 -777.04 427.51 5.007
60 240 78.97 109.01 -778.10 403.25 4.723
60 60 79.73 99.05 -778.77 358.28 4.197
60 0 95.02 100.83 -778.35 331.54 3.883
80 1500 57.51 123.13 -789.93 578.84 6.780
80 960 51.35 136.53 -779.85 564.85 6.616
80 540 57.34 123.91 -786.58 547.49 6.413
80 240 60.07 113.73 -789.93 501.81 5.878
80 60 63.84 103.28 -789.97 468.58 5.489
80 0 78.53 98.41 -788.28 441.75 5.174
Temperature/ Rotation/ ba/ mV bc/ mV Ecorr/ Corrosion rate/
o
C rpm dec-1 dec-1 mV jcorr/ µA cm-2 mm y-1
25 1500 122.77 110.83 -750.79 480.35 5.626
25 960 137.35 134.47 -749.00 437.29 5.122
25 540 137.25 143.46 -749.62 419.73 4.916
25 240 108.59 195.49 -743.11 387.53 4.539
25 60 145.02 187.07 -749.76 356.84 4.180
25 0 149.44 202.31 -749.41 321.32 3.764
40 1500 98.06 149.885 745.19 544.17 6.374
40 960 131.29 127.88 -750.07 514.11 6.022
40 540 122.95 134.89 -754.25 481.41 5.639
40 240 134.35 121.80 -752.29 467.98 5.482
P a g e | 400
40 60 124.64 133.53 -760.65 396.75 4.647

40 0
60 1500 91.85 121.83 -774.41 663.52 7.772
60 960 92.07 125.04 -774.44 672.44 7.877
60 540 83.33 124.89 -770.01 595.97 6.981
60 240 75.28 126.30 -774.96 569.34 6.669
60 60 74.60 116.32 -775.51 536.18 6.280
60 0 66.99 135.98 -775.33 517.06 6.057
80 1500 57.89 119.62 -799.15 781.03 9.149
80 960 59.98 125.26 -805.54 749.26 8.777
80 540 68.81 110.58 -802.51 715.47 8.381
80 240 65.13 88.57 -806.6 623.78 7.307
80 60 67.82 71.30 -803.86 582.92 6.828
80 0 73.44 102.71 -803.88 547.27 6.410
Rotation/ ba/ mV bc/ mV Corrosion rate/

Temperature/ oC rpm dec-1 dec-1 Ecorr mV jcorr/ µA cm-2 mm y-1
18 1500 127.49 146.54 -776.17 423.63 4.96
18 960 136.72 153.09 -753.13 392.16 4.59
18 540 141.15 162.70 -755.96 342.45 4.01
18 240 135.64 157.88 -765.53 280.45 3.29
18 60 157.54 139.09 -804.06 183.05 2.14
18 0
25 1500 137.44 142.42 -768.02 528.32 6.19
25 960 120.36 147.16 -747.14 486.72 5.70
25 540 132.36 142.31 -751.05 463.49 5.43
25 240 157.00 146.39 -756.02 460.32 5.39
25 60 160.20 140.67 -760.57 389.29 4.56
25 0 156.27 130.63 -760.26 370.98 4.35
40 1500 119.05 168.12 -766.50 832.17 9.75
40 960 117.78 154.17 -768.26 701.76 8.22
40 540 108.39 128.29 -767.91 594.08 6.96
40 240 123.99 146.60 -771.99 567.75 6.65
40 60 140.51 119.38 -762.74 540.01 6.33
40 0 125.53 121.58 -764.60 451.90 5.29
60 1500 97.27 143.02 -779.84 1154.40 13.52
60 960 91.49 151.40 -784.03 1103.00 12.92
60 540 98.18 139.67 -784.03 1078.80 12.64
60 240 99.17 138.35 -784.38 1006.30 11.79
60 60 74.30 150.92 -784.62 935.30 10.96
P a g e | 401
60 0 90.47 136.34 -780.25 805.32 9.43

80 1500 97.74 170.50 790.43 1508.80 17.67
80 960 79.93 115.78 -808.53 1375.10 16.11
80 540 88.29 167.98 -789.17 1293.80 15.16
80 240 74.92 155.75 -787.83 1215.20 14.24
80 60 85.94 166.32 -786.48 1153.70 13.51
80 0 109.73 163.06 -786.21 1008.80 11.82
Temperature/ Rotation/ ba/ mV bc/ mV Ecorr/ jcorr/ µA cm- Corrosion rate/
o
C rpm dec-1 dec-1 mV 2
mm y-1
18 1500 127.49 146.54 -776.17 423.63 4.96
18 960 136.72 153.09 -753.13 392.16 4.59
18 540 141.15 162.70 -755.96 342.45 4.01
18 240 135.64 157.88 -765.53 280.45 3.29
18 60 157.54 139.09 -804.06 183.05 2.14
18 0
25 1500 137.44 142.42 -768.02 528.32 6.19
25 960 120.36 147.16 -747.14 486.72 5.70
25 540 132.36 142.31 -751.05 463.49 5.43
25 240 157.00 146.39 -756.02 460.32 5.39
25 60 160.20 140.67 -760.57 389.29 4.56
25 0 156.27 130.63 -760.26 370.98 4.35
40 1500 119.05 168.12 -766.50 832.17 9.75
40 960 117.78 154.17 -768.26 701.76 8.22
40 540 108.39 128.29 -767.91 594.08 6.96
40 240 123.99 146.60 -771.99 567.75 6.65
40 60 140.51 119.38 -762.74 540.01 6.33
40 0 125.53 121.58 -764.60 451.90 5.29
60 1500 97.27 143.02 -779.84 1154.40 13.52
60 960 91.49 151.40 -784.03 1103.00 12.92
60 540 98.18 139.67 -784.03 1078.80 12.64
60 240 99.17 138.35 -784.38 1006.30 11.79
60 60 74.30 150.92 -784.62 935.30 10.96
60 0 90.47 136.34 -780.25 805.32 9.43
80 1500 97.74 170.50 790.43 1508.80 17.67
80 960 79.93 115.78 -808.53 1375.10 16.11
80 540 88.29 167.98 -789.17 1293.80 15.16
80 240 74.92 155.75 -787.83 1215.20 14.24
80 60 85.94 166.32 -786.48 1153.70 13.51
80 0 109.73 163.06 -786.21 1008.80 11.82
P a g e | 402
RDE rotating at f = 1500 rpm in 50 wt% MEA+ CO2 at pH ̴ 8.01; 10 mV s-1
Temperature/ Rotation/ ba/ mV bc/ mV Ecorr/ jcorr/ µA cm- Corrosion rate/
o
C rpm dec-1 dec-1 mV 2
mm y-1
25 1500 108.005 130.575 -751.65 288.56 3.380
25 960 117.46 105.415 -763.48 282.87 3.313
25 540 133.07 140.57 -757.53 278.03 3.256
25 240 118.7 126.03 -753.67 273.6 3.205
25 60 130.38 126.84 -754.09 260.92 3.056
25 0 152.78 125.04 -764.11 249.96 2.928
40 1500 125.12 89.455 -761.21 397.96 4.661
40 960 92.55 114.915 -763.88 379.82 4.449
40 540 98.915 112.865 -766.19 357.67 4.189
40 240 115.835 142.98 -774.96 340.01 3.982
40 60 101.905 113.51 -768.34 323.43 3.788
40 0 115.61 146.5 -775.44 311.87 3.653
60 1500 89.62 131.37 -780.25 861.48 10.091
60 960 92.635 134.595 776.54 822.35 9.633
60 540 82.98 133.44 -783.43 785.34 9.199
60 240 89.05 141.21 -782.93 703.5 8.241
60 60 92.48 155.49 -787.05 655.45 7.678
60 0 87.355 162.635 -787.57 595.97 6.981
80 1500 93.29 104.93 -793.61 1115.9 13.072
80 960 92.645 89.765 -802.07 963.67 11.288
80 540 78.925 94.725 -807.13 830.99 9.734
80 240 66.015 127.225 -810.48 779.16 9.127
80 60 127.16 168.63 -811.58 741.8 8.689
80 0 79.725 104.115 -802.64 689.52 8.077
RDE rotating at f = 1500 rpm in 60 wt% MEA+ CO2 at pH ̴ 8.01; 10 mV s-1
Temperature/ Rotation ba/ mV bc/ mV Ecorr/ Corrosion rate/ mm
o
C / rpm dec-1 dec-1 mV jcorr/ µA cm-2 y-1
25 1500 119.35 106.655 756.54 279.99 3.279
25 960 110.475 119.585 -756.09 270.35 3.166
25 540 119.5 99.415 -756.61 243.58 2.853
25 240 138.63 111.87 -759.94 230.27 2.697
25 60 149.585 137.06 -769.99 223.48 2.617
25 0 175.59 125.525 -787.49 204.79 2.398
P a g e | 403
40 1500 98.315 88.85 -759.57 373.95 4.380

40 960 101.68 90 -756.67 363.72 4.260
40 540 101.46 95.675 -762.29 359.41 4.210
40 240 112.315 88.975 -766.8 333.74 3.909
40 60 130.465 95.815 -772.98 320.79 3.757
40 0 126.51 132.36 -777.7 303.17 3.551
60 1500 72.275 80.895 -786.06 773.85 9.065
60 960 72 120.375 -784.54 737.99 8.645
60 540 83.16 119.41 -787.24 675.81 7.916
60 240 80.47 152.79 -788.83 634.54 7.433
60 60 75.645 156.885 -800.03 612.27 7.172
60 0 68.47 110.59 -784.89 550.62 6.450
80 1500 60.02 148.485 -794.91 965.61 11.311
80 960 69.86 116.43 -796.68 846.28 9.913
80 540 71.7 116.405 -798.19 800.55 9.377
80 240 65.375 122.965 -796.38 712.93 8.351
80 60 75.505 161.98 -802.69 639 7.485
80 0 78.28 145.235 -802.69 578.05 6.771
VI. Thermodynamic Data

Table 9-12: Thermodynamic properties of species in Fe-CO2-H2O systems at 298K (Bard et
al., 1985).
Species DfG0 / KJ mol-1 DfH0 / KJ mol-1 S0 / J mol-1 K-1

Fe 0.00 0.00 27.29
Fe2+ -91.50 -92.40 -106.00
FeOH+ -277.40 -324.70 -29.00
Fe(OH)2 (aq) -441.00 -29.10
Fe(OH)2 (ppt) -492.02 -574.00 87.90
Fe(OH)3- -605.10 29.90
Fe(OH)42- -769.90
HFeO2- -376.40 63.00
FeO22- -301.00 -98.00
FeO -251.45 -272.04 60.75
Fe0.942O -245.14 -266.27 57.49
Fe3O4 -1015.36 -1118.40 146.40
Fe2O3 -743.61 -825.50 87.40
FeOOH
Fe2O3.H2O -984.30 -111.80 118.80
Fe3+ -16.70 -50.20 -280.30
FeOH2+ -229.41 -290.80 -142.00
Fe(OH)2+ -438.10
Fe2(OH)24+ -467.27 -612.10 -356.00
Fe(OH)3 (aq) -660.00 75.40
Fe(OH)3 (ppt) -7055350.00 -833.00 105.00
P a g e | 404
Fe(OH)4- -842.20 24.50

FeO42- -322.38
HFeO4- -364.05
H2FeO4 -384.03
H3FeO4+ -393.16
FeCO3 -666.72
Fe2(OH)2CO3 -1171.5
H 2O -237.18 -285.83 69.91
OH- -157.29 -229.99 -10.75
H+ 0.00 0.00 0.00
O2 16.30 -11.70 110.90
O2 0.00 0.00 205.03
H2 0.00 0.00 130.68
CO32- -527.90
HCO3- -587.06
H2CO3 -607.63
CO2 -386.20 -412.90 121.00
CO2 -394.38 -393.51 213.64
CO -137.27 -110.52 197.91
CH4 -50.79 -74.85 186.20
HCHO -129.70
HCHO -110.00 -116.00 218.60
HCOOH -335.70 -362.60 251.00
HCOOH -346.00 -409.00 128.90
HCOO- -351.00 -410.00 91.60
(COOH)2 -697.90 -826.80 120.00
(COOH)2 -697.00 -818.26
- -690.86 -818.80
H(COO)2- 2
(COO)2 -674.04 -818.80 45.60
CH3OH -166.30 -238.64 127.00
CH3OH -161.90 -201.20 238.00
CH3OH -174.50 -246.00 132.30
CH3CH2OH -174.80 -277.63 161.00
CH3CH2OH -168.60 -235.30 282.00
CH3CH2OH -177.00 -288.10
C6H12O6 -917.22
Specific heat capacity values at constant pressure were calculated from the following expression:
!(10. )
!" = $ + & 10*+ , + + 1 10*2 , 0
,0
The coefficients of Cp were taken from the HSC software database.

P a g e | 405
Table 9-13: Specific Heat Capacity Coefficients(HSC Software Database, 2016)
Species A B C D
H 2O 0.023 7.976 5.507 7.229
CO3 2- -0.064 -106.242 5.548 97.445
HCO3 - -0.035
H2CO3
CO2 0.243
CO2 (g) 0.042 15.100 -8.891 -2.908

H2(g) 16.920 61.459 0.59 -79.559
Fe 0.015 28.079 1.548
Fe2+
Fe3+
Fe(OH)2
FeOH+
Fe2(OH)24+
Fe(OH)2(s) 0.074 81.534 1.833 -31.895
Fe(OH)3(s) 0.086 123.240 -15.121 -42.263
FeCO3(s) 0.049 112.098
FeO 20.188 -21.411 -3.306 0.167
Fe3O4 0.475 -873.665 -120.520 801.73
Fe2O3 0.098 77.822 -14.853

Labiyi F 2016 PHD Thesis

Uploaded by

Copyright:

Available Formats

Labiyi F 2016 PHD Thesis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Labiyi F 2016 PHD Thesis

Uploaded by

Copyright:

Available Formats

Electrochemical Kinetics and Mechanisms of

Iron Oxidation in CO2-containing Aqueous

Department of Chemical Engineering and Chemical Technology

A thesis submitted for the degree of Doctor of Philosophy of Imperial College

Department of Chemical Engineering at Imperial College London. The information used or

acknowledged in the list of references.

Imperial College of Science, Technology and Medicine

Department of Chemical Engineering

Electrochemical Kinetics and Mechanisms of Iron Oxidation in CO2-containing Aqueous

2016 Folasade Labiyi

necessitating the development of low-carbon technologies. Despite its energy requirements,

The passive electrochemical behaviour of Fe in MEA at pH ca. 12 switched to active dissolution

defined by the Levich equation.

platinised titanium cathode and a cation-permeable Nafion membrane. Inductively coupled

plasma optical emission spectrophotometry (ICP-OES) was used to determine FeII

concentrations in the amine solution after potentiostatic electrolyses as a function of solution

from measured current densities.

reductions of H+, H2O and O2.

commercially available solvents tested namely methyldiethanolamine (MDEA), 2-amino-2-

methyl-1propanol (AMP) and aminoethylpiperazine (AEP). Carbon steel exhibited similar

behaviour and their reported efficiencies in the absorption process.

opportunity and granting me the capability to finish.

and moral support from the following people.

obstacles that I faced throughout my PhD.

(EPSRC) for sponsoring my research.

issues as well as fun social activities that provided a much-needed distraction.

and tolerant throughout.

Declaration of Originality .......................................................................................................... I

List of Figures ........................................................................................................................ XII

List of Tables ..................................................................................................................... XXIX

Nomenclature .................................................................................................................. XXXIII

1. Chapter 1 Introduction ................................................................................................... 1

1.1. Motivation for carbon capture and sequestration........................................................ 1

1.2. Operational difficulties in CO2 absorption process .................................................... 6

1.3. Corrosion in Amine Scrubbing Plants ........................................................................ 7

1.4. Existing Corrosion Control ......................................................................................... 8

1.5. Limitation to Available Corrosion Data ..................................................................... 9

1.6. Project Aims.............................................................................................................. 10

1.7. Project Scope and Objectives.................................................................................... 10

1.7.1. Project Tasks ......................................................................................................... 11

1.8. Thesis outline: ........................................................................................................... 12

1.9. References ................................................................................................................. 14

2. Chapter 2 Principles of Electrochemical Systems for Corrosion Studies ................. 16

2.1. Thermodynamics of Electrochemical reactions ........................................................ 18

2.1.1. Nernst Equation .................................................................................................... 19

2.1.2. Pourbaix Diagrams................................................................................................ 20

2.2. Kinetics of Electrochemical Reactions ..................................................................... 26

2.2.1. Rate of reaction: Faraday’s Laws of Electrolysis ................................................. 26

2.2.2. Rate limiting Mechanisms .................................................................................... 28

2.2.3. Charge Transfer Controlled Reactions.................................................................. 28

2.2.4. Mixed Potential Theory ........................................................................................ 33

2.3. Mass transport ........................................................................................................... 36

2.3.1. Transport Processes in Electrochemical Systems ................................................. 36

2.3.2. Electrochemical Reaction Coupled to Mass Transport ......................................... 43

2.4. References ................................................................................................................. 46

3. Chapter 3 Literature Review ........................................................................................ 47

3.1. The CO2 Absorption Process ................................................................................... 48

3.2. Solvents for CO2 Absorption ................................................................................... 53

3.2.1. CO2 Absorption Reactions ................................................................................... 53