Materials On Black (Powder

Formation of Black Powder Components by Dewing and Hygroscopic Corrosion
Processes
A thesis presented to
the faculty of
the Russ College of Engineering and Technology of Ohio University
In partial fulfillment
of the requirement for the degree
Master of Science
Martin L. Colahan
April 2017
© 2017 Martin L. Colahan. All Rights Reserved

2
This thesis titled
Formation of Black Powder Components by Dewing and Hygroscopic Corrosion
Processes
by
MARTIN L. COLAHAN
has been approved for
the Department of Chemical and Biomolecular Engineering
and the Russ College of Engineering and Technology by
David Young
Research Associate Professor of Chemical and Biomolecular Engineering
Dennis Irwin
Dean, Russ College of Engineering and Technology

3
ABSTRACT
COLAHAN, MARTIN L., M.S., April 2017, Chemical Engineering
Formation of Black Powder Components by Dewing and Hygroscopic Corrosion Processes
Director of Thesis: David Young

The presence of black powder in natural gas pipelines can lead to equipment
erosion, valve failure, instrumentation malfunction, and increased pressure drop.
However, despite its impact on downstream and midstream operations, black powder
production is poorly understood. In the present work, black powder formation as a result
of corrosion was investigated by simulating sales gas conditions in a glass cell. Steel
specimens were systematically exposed to a range of CO2, H2S, and O2 partial pressures at
differing water condensation rates. The potential for hygroscopic material assisting black
powder formation was also investigated. Friable corrosion products found in dewing
conditions consisted of siderite, mackinawite, and hematite. Buckle-driven delamination
was identified as a leading cause of black powder production from FeS. The presence of
hygroscopic NaCl crystals facilitated corrosion at relative humidities as low as 33%, but
flakey corrosion products were only found if deliquescence occurred.

4
DEDICATION
In memory of my beloved aunt:

Jamell Faith Cazer
1973 – 2014
5
ACKNOWLEDGEMENTS
I would like express my deepest gratitude to my family and friends for without their
support none of this work would have been possible.
I would also like to thank my thesis committee members Dr. Marc Singer, Dr.
Rebecca Barlag, Dr. Kevin Crist, and Dr. Srdjan Nesic for their guidance. A special thanks
go to the technical staff at the Institute for Corrosion and Multiphase Technology (ICMT)
namely Mr. Cody Shafer, Mr. Al Schubert, Mr. Alexis Barxias, and Mr. Phil Bullington
for their technical expertise and assistance with my experimental work. I would like to
thank The Petroleum Institute in Abu Dhabi, United Arab Emirates for their financial
support. A special thanks goes to Dr. Bruce Brown of the ICMT, and Dr. Ricardo Nogueira
and Dr. Yves Gunaltun of The Petroleum Institute for their support and insightful
discussions.
Last but certainly not least, I would like to thank my advisor, Dr. David Young.
One could not ask for a better mentor. His steady guidance has been the foundation to all
the work herein and has helped me remember why I fell in love with science and
engineering as a child. Thank you so very much.

6
TABLE OF CONTENTS
Page
Abstract ............................................................................................................................... 3
Dedication ........................................................................................................................... 4
Acknowledgements............................................................................................................. 5
List of Tables ...................................................................................................................... 9
List of Figures ................................................................................................................... 10
Chapter 1: Introduction................................................................................................. 16
Chapter 2: Background and Literature Review ............................................................ 20
2.1 Sales Gas ............................................................................................................ 20
2.2 Corrosion of Steel in Oil and Gas Environments ............................................... 22
2.2.1 Water Chemistry ......................................................................................... 23
2.2.2 Formation of Corrosion Product Layers ..................................................... 25
2.2.3 Effect of O2 on CO2 and H2S Corrosion ..................................................... 27
2.2.4 Crystal Structures of Corrosion Product Species........................................ 28
2.3 Black Powder ..................................................................................................... 29
2.4 Water Availability.............................................................................................. 31
2.4.1 Top-of-the-Line Corrosion.......................................................................... 33
2.4.2 Hygroscopic Uptake of Water by Salts....................................................... 35
2.4.3 Hygroscopic Corrosion ............................................................................... 36
2.5 Gaps in Understanding Black Powder Formation.............................................. 40
Chapter 3: Research Objectives and Hypotheses ......................................................... 41

7
3.1 Objective ............................................................................................................ 41
3.2 Hypotheses ......................................................................................................... 41
3.3 Motivation .......................................................................................................... 42
Chapter 4: Methodology and Experimental Design ..................................................... 43
4.1 Overview ............................................................................................................ 43
4.2 Test Conditions .................................................................................................. 43
4.3 Black Powder Formation in Dewing Conditions ............................................... 46
4.3.1 Dewing Corrosion Equipment .................................................................... 46
4.3.2 Dewing Corrosion Experimental Procedures.............................................. 51
4.4 Black Powder Formation in Hygroscopic Conditions........................................ 52
4.4.1 Hygroscopic Corrosion Equipment and Procedures ................................... 52
4.5 Analytical Methods ............................................................................................ 55
4.6 Safety.................................................................................................................. 56
Chapter 5: Results and Discussion ............................................................................... 58
5.1 Dewing Corrosion .............................................................................................. 58
5.1.1 Effect of H2S Partial Pressure on Dewing Corrosion ................................. 58
5.1.2 Effect of Time on Sour Dewing Corrosion................................................. 66
5.1.3 FeS Layer Buckling .................................................................................... 70
5.1.4 Effect of O2 on Sweet Dewing Corrosion................................................... 72
5.1.5 Effect of Cyclic Water Condensation ......................................................... 77
5.1.6 Black Powder Formation in Dewing Conditions ........................................ 82
5.2 Hygroscopic Corrosion ...................................................................................... 84

8
5.2.1 Salt Layer Generation ................................................................................. 84
5.2.2 Hygroscopic Corrosion in Sweet Environments......................................... 86
5.2.3 Hygroscopic Corrosion in Sour Environments ........................................... 93
5.2.4 Corrosion Below the Deliquescence Relative Humidity .......................... 101
5.2.5 Black Powder Formation through Hygroscopic Corrosion Processes ...... 103
5.3 Experiment Repeatability................................................................................. 104
Chapter 6: Conclusions and Future Work .................................................................. 106
6.1 Summary of Results ......................................................................................... 106
6.2 Hypotheses Revisited ....................................................................................... 106
6.3 Comments on Black Powder Mitigation .......................................................... 108
6.4 Future Work ..................................................................................................... 108
References....................................................................................................................... 110
Appendix A: Peltier Temperature Controller Design ..................................................... 115
Appendix B: Calculations ............................................................................................... 119
Appendix C: Additional Microscopy.............................................................................. 123

9
LIST OF TABLES
Page
Table 1: Typical sales gas specifications pertinent to black powder formation [7].......... 21
Table 2: Typical H2S, CO2, O2, and H2O sales gas specifications and measured levels
within the Saudi Aramco sales gas network [8]................................................................ 21
Table 3: Common phases found in corrosion products formed in CO2, H2S, and O2
corrosion ........................................................................................................................... 29
Table 4: Sales gas composition from which experimental parameters are derived .......... 44
Table 5: Composition of API 5L X65 steel used for testing (wt.%)................................. 45
Table 6: Test matrix for studying the effect of H2S partial pressure on dewing corrosion.
........................................................................................................................................... 59
Table 7: Test matrix for analyzing the effect of time on sour dewing systems ................ 66
Table 8: Test matrix for the investigation of the effect of O2 on sweet dewing corrosion72
Table 9: Test matrix for the investigation of the effect of cyclic water condensation...... 77
Table 10: Hygroscopic corrosion test matrix.................................................................... 84

10
LIST OF FIGURES
Page
Figure 1: Turbine blade eroded due to exposure to black powder. Reproduced with
permission from NACE International, Houston, TX. All rights reserved. Smart, Paper
11089 presented at CORROSION/2011, Houston, TX. © NACE International 2011 [6].
........................................................................................................................................... 18
Figure 2: (a) Measured dew point temperatures and (b) corresponding moisture contents
of sales gas measured during July and February at sampling locations in the Saudi
Aramco sales gas network [8]........................................................................................... 32
Figure 3: Ambient temperatures in the Eastern Province of Saudi Arabia. Points below
the red dotted line indicate times when dew may form in the pipeline. Reproduced with
permission from NACE International, Houston, TX. All rights reserved. A.M. Sherik,
S.R. Zaidi, E.V. Tuzan, J.P. Perez, Paper 08415 presented at CORROSION/2008, New
Orleans, LA. © NACE International 2008 [8].................................................................. 33
Figure 4: Flaky FeS corrosion product produced in marginally sour TLC conditions. (Test
Conditions: Duration: 7 days, pH2S: 0.15 H2S, pCO2: 0.93 bar, Water Condensation Rate:
0.25±0.04 ml/m2/s, Tgas: 40°C, Tsteel: 28°C Tsteel). Reproduced with permission from
NACE International, Houston, TX. All rights reserved. N. Yaakob, F. Farelas, M. Singer,
S. Nesic, D. Young, Paper 7695 presented at CORROSION/2016, Vancouver, BC. ©
NACE International 2016 [37].......................................................................................... 35
Figure 5: (a) Corrosion rates and (b) optical images taken after extraction of specimens
with NaCl deposits exposed to CO2/H2S gas mixtures with the indicated relative
humidity. Reproduced with permission from NACE International, Houston, TX. All
rights reserved. W. Litke, J. Bojes, P. Blais, J. Lerbscher, W. Wamburi, Paper 2339
presented at CORROSION/2013, Orlando, FL. © NACE International 2013. [42] ........ 38
with MgCl2 deposits exposed to CO2/H2S gas mixtures with the indicated relative
humidity Reproduced with permission from NACE International, Houston, TX. All rights
reserved. W. Litke, J. Bojes, P. Blais, J. Lerbscher, W. Wamburi, Paper 2339 presented at
CORROSION/2013, Orlando, FL. © NACE International 2013. [42] ............................ 39
Figure 7: Microstructure of the X65 steel specimens. Ferrite and cementite are found in
the dark and white locations, respectively. ....................................................................... 45
Figure 8: Dewing corrosion glass cell. Image courtesy of Cody Shafer. ......................... 48
11
Figure 9: Specimen holding and control stack including (A) specimen holder, (B) Peltier,
(C) Water cooled heat sink, and (D) Anchor bar and bolts. Image courtesy of Cody
Shafer. ............................................................................................................................... 49
Figure 10: XRD specimen holder. Image courtesy of Cody Shafer. ................................ 49
Figure 11: Hygroscopic corrosion glass cell. Image courtesy of Cody Shafer................. 53
Figure 12: Effect of H2S partial pressure and water condensation rate on corrosion
product morphology.......................................................................................................... 61
Figure 13: XRD data showing the effect of H2S partial pressure and WCR on corrosion
product composition.......................................................................................................... 62
Figure 14: Effect of H2S partial pressure and water condensation rate on corrosion rate.63
Figure 15: Effect of H2S partial pressure and water condensation rate (L = low, H = high)
on the measured and theoretical corrosion product masses. ............................................. 64
Figure 16: Effect of H2S partial pressure on maximum possible black powder production
rate in a 100 km, 42 inch ID pipeline................................................................................ 65
Figure 17: Effect of Time on Corrosion Product Morphology (0.3 mbar H2S, 0.96 bar
CO2, 0 mbar O2, 0.015 ml/m2/s WCR, 25 Tsteel, 30°C Tgas).............................................. 67
Figure 18: Effect of time on corrosion product composition............................................ 68
Figure 19: Effect of time on corrosion rate at 0.3 mbar pH2S and low WCR. ................. 69
Figure 20: Effect of time on (a) corrosion product mass and (b) maximum possible black
powder production rate in a 100 km long, 42 in. ID pipeline at 0.3 mbar pH2S and low
WCR. ................................................................................................................................ 70
Figure 21: Effect of O2 on (a) corrosion rate, (b) measured and theoretical corrosion
product mass, and (c) maximum possible black powder production rate in a 100 km, 42
in. ID pipeline. .................................................................................................................. 73
Figure 22: Surface analysis of corrosion products formed after a 3 day exposure to a 10
mbar O2 – 0.92 bar CO2 atmosphere at the low WCR. ..................................................... 75
Figure 23: (a) XRD and (b) Raman spectroscopy compositional analysis of corrosion
product formed after a 3-day exposure to a 10 mbar O2 – 0.92 bar CO2 atmosphere at the
low WCR. The red hematite spectrum is added as a reference [50]................................. 76
Figure 24: Effect of cyclic water condensation on corrosion product morphology. ........ 79
Figure 25: Effect of cyclic water condensation on the corrosion product composition. .. 80
12
Figure 26: Effect of temperature cycling on (a) corrosion rate, (b) measure and theoretical
corrosion product mass, and (c) maximum possible black powder production rate in a 100
km, 42 in. ID pipeline. ...................................................................................................... 81
Figure 27: SEM analysis of the NaCl layer formed before specimen insertion into test
environment. The specimen was coated with palladium to minimize electron beam
charging............................................................................................................................. 85
Figure 28: Slight localized corrosion on the outside boundary of a salt crystal formed by
drying NaCl....................................................................................................................... 86
Figure 29: (a) Surface microscopy after a 3 day exposures to CO2 atmosphere at 75% RH.
The region outlined in red is enlarged in (b)..................................................................... 87
Figure 30: EDS of corrosion product after a 3 day exposure to a CO2 environment at 75%
RH. NQ: Not Quantified. .................................................................................................. 88
Figure 31: XRD corrosion product composition analysis of specimens recovered after a 3
day exposure to CO2 at 75% RH. Peak labels are S: siderite (FeCO3), H: halite (NaCl),
and Fe: ferrite (Fe). ........................................................................................................... 88
Figure 32: Surface profilometry of specimen recovered after a 3 day exposure to CO2 at
75% RH. Scale bar units are in nm. .................................................................................. 90
Figure 33: Surface microscopy and XRD compositional analysis after a 3 day exposure to
CO2 at 58 and 33% RH. .................................................................................................... 91
Figure 34: Effect of relative humidity on (a) general corrosion rate, (b) measured and
theoretical corrosion product mass, and (c) maximum possible black powder production
rate in a 100 km, 42 in. ID pipeline under hygroscopic CO2 conditions. ......................... 92
Figure 35: Effect of H2S partial pressure on corrosion product morphology in

deliquescing conditions (75% RH, 0.97 bar CO2). ........................................................... 94
Figure 36: Effect of H2S partial pressure on corrosion product composition. Phases are
M: mackinawite, S: siderite, H: halite, and Fe: ferrite. Peaks labeled with an asterisk (*)
correspond to unknown phases. ........................................................................................ 96
Figure 37: Effect of H2S partial pressure on surface morphology at 58 and 33% RH. .... 97
M: mackinawite, H: halite, and Fe: ferrite........................................................................ 99
Figure 39: Effect of H2S partial pressure on (a) general corrosion rate, (b) measured and
rate in a 100 km, 42 in. ID pipeline under hygroscopic conditions. ............................... 100
13
Figure 40: Peltier controller schematic. .......................................................................... 116
Figure 41: Cross-section view of the XRD specimen holder. Image courtesy of Cody
Shafer. ............................................................................................................................. 118
Figure 42: Conditions: pH2S: 0 mbar, pCO2: 0.96 bar, WCR: 0.05 ml/m2/s, Steel
Temperature: 15°C, Gas Temperature: 30°C, Duration: 3 days. .................................... 123
Temperature: 25°C, Gas Temperature: 30°C, Duration: 3 days. Images (a) and (b) were
from the same test as that shown in Chapter 5, whereas images (c)-(f) were from a test
with inconclusive gravimetric measurements. The corrosion product in the inconclusive
test was of a similar morphology to the CO2/O2 test. ..................................................... 124
Figure 44: Conditions: pH2S: 0.1 mbar, pCO2: 0.96 bar, WCR: 0.05 ml/m2/s, Steel
Temperature: 25°C, Gas Temperature: 30°C, Duration: 6 hours. .................................. 126
Temperature: 25°C, Gas Temperature: 30°C, Duration: 1 day....................................... 127
Temperature: 25°C, Gas Temperature: 30°C, Duration: 10 days. .................................. 130
Temperature: 25°C, Gas Temperature: 30°C, Duration: 6 hours. .................................. 133
14
Figure 55: Conditions: pH2S: 0.3 mbar, pCO2: 0.96 bar, WCR: 0.015/0 ml/m2/s, Steel
Temperature: 25/40°C, Gas Temperature: 30°C, Duration: 3 days, Cycle Period: 1 day.
......................................................................................................................................... 134
......................................................................................................................................... 135
Figure 57: Conditions: pH2S: 0 mbar, pCO2: 0.96 bar, RH: 75%, Temperature: 25°C,
Duration: 3 days, Salt Layer Composition: NaCl. (d) and (e): after DI water rinse, (f):
after Clarke...................................................................................................................... 136
Duration: 3 days, Salt Layer Composition: NaCl. Micrograph (a): after extraction, (b):
after DI water rinse, and (c): after Clarke. ...................................................................... 137
after DI water rinse, and (c): after Clarke. ...................................................................... 138
Figure 60: Conditions: pH2S: 0.1 mbar, pCO2: 0.96 bar, RH: 75%, Temperature: 25°C,
Duration: 3 days, Salt Layer Composition: NaCl ........................................................... 139
Duration: 3 days, Salt Layer Composition: NaCl. .......................................................... 139
Duration: 3 days, Salt Layer Composition: NaCl. (c): after DI water rinse, (d): after
Clarke.............................................................................................................................. 143
Duration: 3 days, Salt Layer Composition: NaCl. Micrographs (c) and (d): after DI water
rinse................................................................................................................................. 144
15
rinse................................................................................................................................. 145
16
CHAPTER 1: INTRODUCTION
The global economy is driven by the consumption of useful forms of energy.
Historically, energy demand in the industrialized world has largely been fulfilled through
the consumption of fossil fuels such as coal, oil, and natural gas. Investments into
alternative energy sources (e.g., wind, solar, and nuclear energy) have been made in an
attempt to diversify the global energy portfolio, but limitations including supply
intermittency, energy storage, and large up-front capital costs currently prevent the overall
transformation from a fossil fuel based economy into an alternative energy based economy.
Unless groundbreaking discoveries are made in alternative energy, the consumption
of fossil fuels is likely to remain a primary source of energy. The United States Energy
Information Administration (EIA) [1] projects global energy consumption to increase by
48% between the years 2012 and 2040 to 815 quadrillion Btu, fossil fuels are expected to
supply 78% of this demand. The energy source with the largest increase in consumption is
projected to be natural gas. The abundance and low cost of natural gas makes it an
economically attractive energy source, particularly for the industrial and electrical power
generation sectors. Governmental policies aimed at reducing greenhouse gas emissions
also make natural gas attractive as it generates significantly less CO2 per kWh produced,
as well as combusting more cleanly, than both coal and liquid fossil fuels [2].
Natural gas is a byproduct of the decomposition of plant and animal matter trapped
underground at high pressures and temperatures over millions of years. The resulting
hydrocarbon-rich mixture is extracted from geologic reservoirs, processed to separate the
more valuable components, and remove the majority of undesired components. The
17
processed methane-rich gas, commonly known as sales gas, is transported to the consumer
through a network of pipelines. In 2008, the EIA estimated the network of natural gas
pipelines in the continental U.S. to be approximately 305,000 miles (490,000 kilometers)
long, but the recent boom in natural gas production from unconventional sources (e.g.,
shale or tight gas reservoirs) means that this length will continue to increase [3].
A challenge to the operation of natural gas pipelines is corrosion, which can weaken
the integrity of the pipeline and ultimately may render the costly structure useless unless
mitigated. Internal corrosion in the oil and gas industry is mostly due to water wetting the
steel surface in the presence of acid gases including carbon dioxide (CO2) and hydrogen
sulfide (H2S). Processed natural gas can contain small concentrations of water, carbon
dioxide, hydrogen sulfide, and potentially oxygen that can be corrosive to the steel pipeline
transporting the gas. Products formed as a result of corrosion may spall off the steel surface
and become entrained in the gas becoming what is commonly known in the oil and gas
industry as “black powder” [4].
Black powder is a collective term used to generally describe particles that can be
entrained in the gas. Black powder is a problem for pipeline operators as it can erode
equipment, induce greater pressure drops, and clog instrumentation [4]. The erosive power
of entrained particles can be seen in Figure 1, which shows a power plant turbine blade
after exposure to black powder. Black powder particles vary in size and can range from
100 nm up 1 mm [5]. The composition of black powder can also vary and can include
corrosion products, salt, dirt, and other materials trapped in the pipeline during
construction. The production of black powder is most significant in gas transmissions

18
pipelines. Transmission pipelines are large capacity pipelines that deliver gas from
processing plants to the consumer or distributor. These pipelines are designed to move
large quantities of natural gas over long distances so, therefore, have large diameters and
high operating pressures.
Figure 1: Turbine blade eroded due to exposure to black powder. Reproduced with
permission from NACE International, Houston, TX. All rights reserved. Smart, Paper
11089 presented at CORROSION/2011, Houston, TX. © NACE International 2011 [6].
Although investigations into the impact of black powder have been performed, its
formation processes remain poorly understood. The present study investigates black
powder formation by examining the effect of CO2, H2S, and O2 on the corrosion of steel in
both water condensing conditions and humid conditions where hygroscopic material assists
wetting the steel surface. Research performed to investigate black powder formation is
described in this thesis as follows:

19
Chapter 2 reviews pertinent literature related to black powder formation and
provides a foundation for work described thereafter.
Chapter 3 describes research objectives and proposes hypotheses regarding black
powder formation.
Chapter 4 discusses the methodologies and experimental equipment used to test
black powder formation in simulated sales gas environments.
Chapter 5 presents and discusses results obtained during experimentation.
Chapter 6 summarizes black powder research described in the preceding chapters
and proposes future work to better understand black powder formation and ideas for
the mitigation of black powder in gas transmission pipelines.

20
CHAPTER 2: BACKGROUND AND LITERATURE REVIEW
Black powder in gas pipelines is a costly problem for which effective mitigation
strategies are desired. Previous studies on black powder have tended to focus on its
behavior rather than formation, consequently, additional information is required to develop
superior mitigation strategies. A review of black powder literature was conducted by Khan
and Al-Shehhi [5] that examined black powder composition, formation, entrainment,
transport, and filtration. Due to the wide variety of topics covered, the review serves as an
excellent resource for engineers in the field with a black powder problem, but the review
of formation mechanisms was limited by the lack of available data in the scientific
literature. Black powder formation through corrosion processes is reviewed in more detail
in this chapter.
2.1 Sales Gas
Compositions will vary, but sales gas is composed primarily of methane (CH4).
Larger hydrocarbon molecules such as ethane (C2H6) can be present, but their
concentrations are very low. Low concentrations of water (H2O), nitrogen (N2), carbon
dioxide (CO2), hydrogen sulfide (H2S), and sulfur-containing species such as methyl
mercaptan (CH3SH, added as a malodorant) are often present in sales gas, but a ceiling
content is generally specified to ensure the processed natural gas meets a gross heating
value and minimizes corrosion. Typical sales gas specifications pertinent to black powder
are given in Table 1 below [7].

21
Table 1: Typical sales gas specifications pertinent to black powder formation [7]
Specification Value
Water Content 4-7 lbm H2O/MMscf
CO2 Content 3-4 mol%
N2 Content 4-5 mol%
Maximum O2 Content 0.01 mol%
H2S Content 0.25-1.0 grain/100 scf
Total Sulfur Content 0.5-20 grain/100 scf
Gross Heating Value 950-1200 Btu/scf
Solids and Free Liquid Commercially free
Typical specifications and measured levels of CO2, H2S, O2, and H2O were
published in a black powder study by Sherik, et al., [8] and are shown in Table 2. H2S,
CO2, and O2 levels were typically within limits, but the water content was consistently over
the 0.112 mg/L (7 lbs/MMscf) limit. The higher than specification water contents may
render the pipeline susceptible for liquid water to form and subsequently allow corrosion
to occur.
Table 2: Typical H2S, CO2, O2, and H2O sales gas specifications and measured levels
within the Saudi Aramco sales gas network [8].
Component Typical Sales Gas Measur ed Levels

Specifications
(maximum limits)
H2S (ppm) 16 1-15
O2 (mol%) 0.02 0.01-0.03
CO2 (mol%) 3 <1.62
H2O (mg/L) 0.112 0.12-0.55
22
2.2 Cor r osion of Steel in Oil and Gas Envir onments
Both CO2 and H2S are considered acid gases due to their ability to form hydrogen
ions in water via dissociation processes. The H+ will react with iron at the steel surface in
an overall reaction given by Reaction 1. This electrochemical reaction can be broken into
an oxidation reaction, Reaction 2, and a reduction reaction, Reaction 3.
Fe s 2H+aq → Fe aq
+
H g (1)
+
Fe s → Fe aq 2푒− (2)
2H+aq 2푒− → H g (3)
If CO2 and H2S are included in the bulk gas, additional electrochemical reactions
need consideration. Carbonic acid has been postulated to directly react with the steel by the
reduction reaction given in Reaction 4, but this direct reaction mechanism is under dispute
in favor of a “buffering” mechanism [9]. Conversely, H2S is known to directly react with
the steel in the overall reaction shown in Reaction 5 [10].
2H CO aq 2푒− → H g 2HCO− aq (4)
Fe s H S aq → FeS s H g (5)
The rate at which the electrochemical reactions encountered in corrosion proceed
is dependent upon the charge transfer rate and the concentration of the oxidants at the steel
surface. The oxidant concentration at the surface is determined by both mass transfer and
the bulk concentration, which is governed by water chemistry. The effect of CO2 and H2S
on the water chemistry is explored further in the following sections [10].

23
2.2.1 Water Chemistry
The acidification of water caused by carbon dioxide takes place through several
steps. These steps include the dissolution of CO2 (Reaction 6), the hydration of CO2 to
form carbonic acid (Reaction 8), the first dissociation step of carbonic acid to form
bicarbonate ions (Reaction 10), and the dissociation of bicarbonate ions to form carbonate
ions (Reaction 12). The equilibrium constants are denoted in the reactions and are also
defined as equilibrium expressions [11], [12].
퐶
CO ⇌ CO (6) 퐾 (7)
g aq 푝
퐶
CO H O 푙 ⇌ H CO (8) 퐾hyd (9)
aq aq 퐶
퐶 퐶
H CO ⇌ HCO− aq H+aq (10) 퐾ca (11)
aq 퐶
퐶 퐶
HCO− aq ⇌ CO −aq H+aq (12) 퐾bi (13)
퐶
where:
퐶 = concentration of species i (molar)
푝 = partial pressure of gas i (bar)
퐾 = equilibrium constant of reaction 푖
With the exception of the lack of a hydration step, the acidification of water due to
H2S is similar to that of CO2. H2S does not require hydration to dissociate. To acidify water,
H2S first dissolves (Reaction 14) and then dissociates into bisulfide ions (Reaction 16),
which can dissociate further into sulfide ions (Reaction 18). Each dissociation step forms
hydrogen ions which can be reduced during the corrosion process [13]–[15].
24
퐶
H Sg ⇌ H S aq (14) 퐾 (15)
푝
퐶 퐶
H S aq ⇌ HS−aq H+aq (16) 퐾hs (17)
퐶
퐶 퐶
HS−aq ⇌ S aq
−
H+aq (18) 퐾bs (19)
퐶
Additionally, water can dissociate (Reaction 20) [15].
H O 푙 ⇌ H+aq OH−aq (20) 퐾w a 퐶 퐶 (21)
Several correlations to calculate the equilibrium constants defined above are given
in Equations 22-29 below [11]–[15]:
14.5 ∗
퐾 10− ( . + . × − . + . ) (22)
1.00258
퐾hyd 2.58× 10− (23)
.
퐾ca 387.6× 10− ( . − . 9 × + . × − . × − . + . ) (24)
.
퐾bi 10− ( . − .9 × + . × − . × − . +. ) (25)
퐾 10− . + . 9 − . × − 9× − .9 l og (26)
퐾 10 . 9 + . − . × − . − . ln (27)
퐾bs 10− .9 + . − . × (28)
퐾w a 10− ( 9. − . 9 + . × ) (29)
where:
푇
퐹= Temperature (°F)
푇= Temperature (K)
푝=partial pressure of CO2 (bar)

25
퐼= Ionic strength (molar)
The ionic strength is calculated with Equation 30 where 푐, 푧, and 푁 are the
concentration (molar) of species 푖

, the charge of species 푖
, and the total number of species,
respectively. Equilibrium is reached when electroneutrality is achieved (Equation 31) [10].
1
퐼 ∑ 푐푧 (30)
2
=
∑ 푐푧 0 (31)
=
2.2.2 Formation of Corrosion Product Layers
If saturation is reached, both iron carbonate and iron sulfide can precipitate. The
precipitation reactions for iron carbonate and iron sulfide along with their attributed
solubility product expressions are given in Equations 32-35 [8, 9].
, +
FeCO s ↔ Fe aq CO −aq (32)
퐾sp, 퐶 퐶 (33)
,
FeS s H+aq ↔ +
Fe aq HS−aq (34)
퐶 퐶
퐾sp, (35)
퐶
Considering temperature and ionic strength, the solubility products 퐾sp, and
퐾sp, are calculated with empirical correlations, which are given in Equations 34 and 35,
respectively [16], [17].

. 9
− 9. 9 − . − + . log + . √ − . (36)
퐾sp, 10
26
. 9
− . (37)
퐾sp, 10 퐾 퐾bs
The solubility products indicate whether precipitation is thermodynamically
possible but do not give any indication on how fast precipitation can occur. Precipitation
kinetics is dependent upon solution supersaturation, temperature, and the onset of
nucleation. If the saturation with respect to a precipitate is greater than 1 then precipitation
may occur. The saturation (S) and precipitation (R) rates of iron carbonate and iron sulfide
are given by Equations 38-41 [18], [19].
퐶 퐶
푆 (38)
퐾sp,
. −
. 푆
푅 푒 퐾 푆 1 (39)
푉 sp,
퐶 퐶
푆 (40)
퐾sp, 퐶
−
, 푆
푅 푒 퐾 푆 1 (41)
푉 sp,
where:
푆 = Saturation of iron carbonate
푅퐹 푂= Rate of iron carbonate formation (mol/m2/s)
푆/ 푉= Surface to volume ratio
푅= Gas constant (J/mol/K)
푆 = Saturation of iron sulfide
푅퐹 = Rate of iron sulfide formation (mol/m2/s)

27
While iron carbonate is considered to be solely produced through precipitation, the
FeS polymorph mackinawite has been postulated to form through both precipitation and a
direct reaction with iron as given by Reaction 42 [20], [21].
Fe s H S aq → FeS s H g (42)
Mackinawite, FeS, is typically the first iron sulfide formed in the corrosion of steel
in H2S-containing environments. The formation of mackinawite through a direct reaction
with iron has been justified by Sun and Nesic [20] due to the fast formation kinetics,
minimal effect of the bulk supersaturation on mackinawite formation, and the similarities
between the crystal structures of ferrite and mackinawite [22].
Two layers are typically observed in iron sulfide corrosion product layers: a thin,
inner layer and a porous, outer layer. Sun and Nesic [20] proposed a mechanism whereby
the thin, inner layer is formed by a reaction of H2S to iron. The outer layer is then formed
by the spallation of this thin inner layer due to internal compressive stresses that arise in
the inner mackinawite layer. The mechanism of mackinawite formation was later
reexamined by Zheng, et al., [21] who postulated that the inner layer forms through
chemisorption and the outer layer is formed by precipitation. The thin inner layer can act
as a preferred nucleation site for the growth of FeS.
2.2.3 Effect of O2 on CO2 and H2S Corrosion
Ingress of oxygen, even at low concentrations, can significantly affect corrosion
processes. The presence of oxygen in pipelines is attributed to ingress at low pressure
points or dissolved in additives such as methanol [23]. A survey of 44 North American

28
natural gas transmission companies found that maximum allowable O2 concentrations were
between 0.01 and 0.10 mol% with a typical value of 0.02 mol% [24].
If oxygen is present in acidic environments such as those associated with CO2 and
H2S then the reduction of oxygen (Reaction 43) may occur [25] given the presence of a
species that can give up electrons (e.g., a corroding metal).
O aq 4H+aq 4푒− → 2H O l (43)
The formation of iron oxides and iron oxyhydroxides proceeds through the
following reactions:
+
4Fe aq 4H+aq O aq
+
→ 4Fe aq 2H O l (44)
+
Fe aq 2H O l → Fe OH s 2H+aq (45)
+
Fe aq 3H O l → Fe OH s 3H+aq (46)
Fe OH s → FeO OH 푠 H Ol (47)
+
2FeO OH s Fe aq → Fe O s 2H+aq (48)
where FeO(OH) can be in the α, ß, or γ form depending on the formation conditions [26].
2.2.4 Crystal Structures of Corrosion Product Species
Corrosion products that form in sales gas conditions may include iron carbonate,
iron sulfide, iron oxide, and iron oxyhydroxide. The phases formed as a product of
corrosion will ultimately depend on both thermodynamics and kinetics, which are, in turn,
determined by pipeline conditions including CO2 partial pressure, O2 partial pressure, H2S
partial pressure, and temperature. The names, formulae, unit cell types, and the
characteristic/related crystal structure of inorganic phases commonly found in corrosion
products formed due to the internal corrosion of pipeline steel are given in Table 3.
29
Table 3: Common phases found in corrosion products formed in CO2, H2S, and O2
corrosion
Phase For mula Cr ystal Structur e Sour ce

Ir on Car bonate
Siderite FeCO3 Trigonal; Calcite [27]
Ir on Sulfide
Mackinawite FeS Tetragonal; 2D Layer; anti-Lead Oxide [28]
Greigite Fe3S4 Cubic; Spinel [29]
Pyrite FeS2 Cubic; Rock Salt [29]
Fe1-xS Monoclinic, Hexagonal; Nickel
Pyrrhotite [29]
(x = 0 to 0.17) Arsenide
Troilite FeS Hexagonal; Nickel Arsenide [29]
Cubic FeS FeS Cubic; Zinc Blende [29]
Ir on Oxide
Hematite α-Fe2O3 Hexagonal; Corundum [26]
Magnetite Fe3O4 Cubic; Spinel [26]
Wüstite FeO Cubic; Rock Salt [26]
Maghemite γ-Fe2O3 Cubic; Spinel (Fe – deficient) [26]
Ir on Oxyhydr oxide
Goethite α-FeO(OH) Orthorhombic; Diaspore [26]
Akagenite ß-FeO(OH) Monoclinic; Nickel Arsenide [26]
Lepidocrocite γ-FeO(OH) Orthorhombic; Boehmite 2D layer [26]
2.3 Black Powder
Sherik, et al., [8] collected black powder samples from Saudi Aramco sour gas and
sales gas pipelines for analysis by X-ray diffraction (XRD) and X-ray fluorescence (XRF).
The sour gas pipeline black powder samples were predominantly composed of FeS and
30
FeS2 with minor amounts of Fe3O4, FeO(OH), and FeCO3. XRD of black powder samples
from the sales gas pipelines detected Fe3O4, α-FeO(OH), and γ-FeO(OH) with small
amounts of FeCO3. However, sulfur was detected in all collected samples by XRF, which
suggests the black powder may contain an amorphous sulfur-containing species. Elemental
sulfur was detected with XRD in some samples, but it is unclear if it was detected in the
sour gas black powder or the sales gas black powder.
Yamada, et al., [30] examined the composition of black powder produced in a
Japanese sales gas network. The black powder was determined to be predominantly
magnetite, but small amounts of hematite, goethite, lepidocrocite, and siderite were also
found. Iron sulfide and elemental sulfur was detected in trace amounts in pipelines that
historically transported H2S-containing sales gas. Both mill scale and the corrosion of
pipeline steel due to oxygen and water vapor ingress is attributed to be a major source of
black powder due to the high quantity of magnetite collected. Ingress of O2 and H2O during
pipeline construction and maintenance is suggested to be the primary cause of corrosion.
The identification of iron oxide and iron oxyhydroxides as the major components
of black powder in sales gas pipelines containing O2 and liquid water were used by Sherik,
et al., and Yamada, et al., to justify the corrosion of steel due to oxygen as the primary
source of black powder. However, the presence of lepidocrocite and elemental sulfur
suggests the formation of iron oxide and iron oxyhydroxide from an iron sulfide precursor
likely occurred. The formation of lepidocrocite in sales gas conditions is largely attributed
to the oxidation of mackinawite, which is shown in Reaction 49 [31], [32]. Magnetite can
then form through the reaction of lepidocrocite and ferrous ions (Reaction 50) [33].
31
8FeS s 4H O l 6O g → 8훾-FeO OH s S s (49)
+
2훾-FeO OH s Fe aq → Fe O s
2H+aq (50)
The oxidation of iron sulfide is not the only pathway to elemental sulfur. Elemental
sulfur may be obtained through the oxidation of H2S as shown in Reaction 51.
8H S aq 4O aq → 8H O l S s (51)
The oxidation of iron carbonate may also occur and is reported to follow Reaction
52.
6FeCO s O g → 2Fe O s 6CO g (52)
2.4 Water Availability
Measured dew points of water in the sales gas network were reported by Sherik, et
al., [8] and are given in Figure 2. The sales gas moisture content consistently exceeded the
maximum moisture level of 7 lbs/MMscf (0.112 mg/L) set by sales gas specifications. The
dew point temperature of water was generally lower than 10°C, but several dew points
exceeded 15°C, which may lead to water condensation depending on the pipeline steel
temperature.
32
(a) (b)
Figure 2: (a) Measured dew point temperatures and (b) corresponding moisture contents
of sales gas measured during July and February at sampling locations in the Saudi
Aramco sales gas network [8].
Sherik compared the measured water dew points to meteorological data to examine
the potential for water condensation on the steel pipeline. As indicated by Figure 3, winter
ambient temperatures in Saudi Arabia were often below the water dew point temperatures
measured; therefore, water condensation was likely. If the ambient temperature serves as
an adequate proxy for the steel temperature, then water condensation is likely to be cyclical
where water condensation can be expected to occur during cool time periods, such as
overnight, and not occur during warm time periods such as during the day. Cyclical water
wetting may affect the corrosion product layer morphology due to fast precipitation during
drying. The effect of temperature fluctuations on the corrosion product morphology is

33
unknown, but corrosion in dewing conditions has been studied in top-of-the-line corrosion
studies.
Figure 3: Ambient temperatures in the Eastern Province of Saudi Arabia. Points below
the red dotted line indicate times when dew may form in the pipeline. Reproduced with
permission from NACE International, Houston, TX. All rights reserved. A.M. Sherik,
S.R. Zaidi, E.V. Tuzan, J.P. Perez, Paper 08415 presented at CORROSION/2008, New
Orleans, LA. © NACE International 2008 [8].
2.4.1 Top-of-the-Line Corrosion
Prior to processing, natural gas contains condensable species including water,
hydrocarbons, and organic acids. If the gas is saturated with water and the steel temperature
is cooler than the gas temperature, then the water can condense on the steel and cause
corrosion. Corrosion in these conditions can be severe near the top of the pipeline where
34
nonvolatile corrosion inhibitors cannot reach, therefore, this type of corrosion has been
called top-of-the-line corrosion (TLC) [34].
Water condensation plays a critical role in TLC. A high water condensation rate
(WCR) would limit the ability of a corrosion product to precipitate due to dilution and thus
leave the steel unprotected. Conversely, a low WCR would allow for higher ferrous ion
concentrations and faster precipitation rates, potentially generating a protective corrosion
product layer [35]. A model for determining the WCR based on heat transfer was developed
by Zhang, et al. [36].
Studies have been performed that investigated the formation of corrosion products
in TLC conditions since corrosion products can offer some protection against corrosion.
Yaakob, et al., examined the effect of iron sulfide on localized corrosion in marginally sour
conditions similar to sales gas conditions without O2. Steel specimens mounted in the lid
of a glass cell with the polished side facing down were exposed to water-saturated CO2/H2S
mixtures with H2S partial pressures of 0, 0.015, 0.03, 0.08, and 0.15 mbar. To force water
condensation on the specimens, their surfaces were cooled to below the 40 or 60°C gas
temperature. Severe localized corrosion was found at 0.015 and 0.03 mbar H2S and was
attributed to failures within the FeS corrosion product layer. The corrosion product layers
obtained at 0.08 and 0.15 mbar H2S were more protective as no localized attack was
observed, but the FeS layers were flaky and appeared to readily detach from the steel
surface. Figure 4 shows the corrosion product morphology from the 40°C Tgas, 0.15 mbar
H2S partial pressure experiment. Corrosion product spallation appears to have occurred
during either test exposure or the isopropanol rinse after specimen extraction, as evidenced
35
by the visibility of the inner FeS layer seen in the right half of Figure 4 [37]. The role of
H2S on the development of a corrosion product layer susceptible to spallation is still
unknown.
Figure 4: Flaky FeS corrosion product produced in marginally sour TLC conditions. (Test
Conditions: Duration: 7 days, pH2S: 0.15 H2S, pCO2: 0.93 bar, Water Condensation Rate:
0.25±0.04 ml/m2/s, Tgas: 40°C, Tsteel: 25°C). Reproduced with permission from NACE
International, Houston, TX. All rights reserved. N. Yaakob, F. Farelas, M. Singer, S.
Nesic, D. Young, Paper 7695 presented at CORROSION/2016, Vancouver, BC. ©
NACE International 2016 [37]
2.4.2 Hygroscopic Uptake of Water by Salts
During the commissioning phase, a pipeline undergoes hydrotesting by flooding
with water and subsequent pressurization to check for leaks. Due to volume requirements
and water costs, a pipeline operator may elect to use seawater, which can leave a
36
hygroscopic salt residue on the pipe wall after the seawater is drained. If not properly
rinsed, the salt residue can accumulate water from a humid sales gas stream.
The amount of moisture extracted from the surrounding atmosphere by a
hygroscopic material is dependent primarily on its affinity toward water, relative humidity
(RH), and temperature. Soluble materials with a high affinity toward water can dissolve if
the availability of water is sufficiently high through a process called deliquescence. The
RH whereby a material deliquesces is known as the deliquescence relative humidity
(DRH). Efflorescence is the reverse process of deliquescence, where an electrolyte
crystallizes due to drying. The DRH and the efflorescence relative humidity (ERH) are not
necessarily the same as hysteresis between the DRH and the ERH exists due to
crystallization kinetics [38]. Hygroscopic uptake of water by chloride and perchlorate salts
in Martian soil has been hypothesized to be the reason for recurring slope lineae, dark
streaks on Martian slopes only present during warmer seasons [39]. The potential for
corrosion due to hygroscopic salts is examined below.
2.4.3 Hygroscopic Corrosion
Hygroscopic salts and their potential to lead to corrosion have been examined in
atmospheric corrosion studies where the deposition of salt aerosols on metals can lead to
corrosion in oxic conditions. The corrosive effect of hygroscopic NaCl particles on mild
steel in atmospheric environments was investigated by Schindelholz, et al. [40]. NaCl was
deposited onto mild steel specimens and subjected to air with relative humidities ranging
from 1.5% to 90% RH for up to 300 days, demonstrating that water adsorbed on the NaCl
particles and the steel can initiate corrosion at relative humidities as low as 33%. The
37
initiation of corrosion was attributed to water adsorbed to the NaCl crystals and the steel
surface with capillary condensation, which accumulates water at the salt/steel interface.
The effect of CO2 and H2S on hygroscopic corrosion is even less understood.
Kolts [41] investigated the corrosion of sales gas pipelines due to salt deposits left behind
by a seawater hydrotest. Steel samples with and without dried synthetic seawater deposits
were exposed to RH-controlled CO2. Kolts found salts can induce corrosion in the
conditions tested but oxygen ingress is suspected to occur which will skew results.
Litke, et al., [42] investigated the corrosion of steel under hygroscopic salts in
CO2/H2S environments with humidity-controlled autoclaves. In their study, they placed
salt dried under vacuum onto UNS G10180 steel coupons and then inserted the specimens
into autoclaves with CO2 and H2S. The humidity of the autoclaves was controlled with
methanol/water or ethylene glycol/water mixtures placed in the bottom of the autoclave
before loading the samples. Figure 5 and Figure 6 show optical images of the exposed
surface after extraction from a 14 day exposure to a gas with 15 psig H2S and 57 psig CO2.
General and localized corrosion rates for the NaCl and MgCl2 deposit tests were measured
and are shown in Figure 5 and Figure 6, respectively. Both corrosion and corrosion product
formation were found to occur in the conditions tested, however, reservations exist due to
the measurement of similar general corrosion rates in the NaCl experiments at 86% and
0.7% RH and the appearance of corrosion product at 0.7% RH in the NaCl experiments.
The surface of the NaCl/0.7% RH specimen is expected to look more like the surface of
the MgCl2/0.7% RH specimen with the white crystalline salt still on the surface. Corrosion
underneath the NaCl crystals at 0.7% RH should be negligible due to the minute amount
38
of water in the vapor phase. The unexpected behavior is likely due to the ethylene glycol
added for RH control.
(a)
(b)
with NaCl deposits exposed to CO2/H2S gas mixtures with the indicated relative
humidity. Reproduced with permission from NACE International, Houston, TX. All
rights reserved. W. Litke, J. Bojes, P. Blais, J. Lerbscher, W. Wamburi, Paper 2339
presented at CORROSION/2013, Orlando, FL. © NACE International 2013. [42]

39
(a)
(b)
with MgCl2 deposits exposed to CO2/H2S gas mixtures with the indicated relative
humidity Reproduced with permission from NACE International, Houston, TX. All rights
reserved. W. Litke, J. Bojes, P. Blais, J. Lerbscher, W. Wamburi, Paper 2339 presented at
CORROSION/2013, Orlando, FL. © NACE International 2013. [42]
Both Kolts and Litke, et al., found corrosion to occur due to salt deposits in humid
conditions related to gas production, but it is still unknown if corrosion induced by
hygroscopic salts can contribute to black powder production as neither Kolts nor Litke
examined the morphology or compositions of the corrosion products formed.

40
2.5 Gaps in Under standing Black Powder For mation
The literature on black powder formation is missing a vital step: spallation. A
corrosion product can grow in a pipeline, but if that corrosion product cannot spall then it
cannot become entrained in the gas and thus form black powder. However, to understand
how corrosion product spallation leads to black powder, the processes that lead to corrosion
product formation in sales gas conditions must be better understood.
Water availability has been identified as the primary unknown for corrosion in sales
gas conditions since the concentrations of CO2, H2S, and sometimes O2 are known to be
adequate for signification corrosion to occur. If the water content of the sales gas is high,
then water may condense onto the steel surface. However, if the water content is below the
thermodynamic dew point temperature, then a hygroscopic material like salt must be
present for corrosion to occur. Cyclical temperature fluctuations were also identified as
potentially changing the corrosion product morphology. Spalled corrosion products were
obtained in dewing conditions with H2S, but the effect of acid gas concentration and water
condensation rates on the production of friable corrosion products is still unknown. While
some information on corrosion product formation in dewing conditions is available, little
is known about the potential for hygroscopic corrosion in CO2 and H2S conditions to
produce friable corrosion products.

41
CHAPTER 3: RESEARCH OBJ ECTIVES AND HYPOTHESES
3.1 Objective
The primary objective of this work is to examine the formation of corrosion
products, associated with black powder when spalled, developed through dewing and
hygroscopic processes. The overall goal is to gain a better understanding of the role of
water availability and H2S concentration on the corrosion of sales gas pipelines.
This study is broken into two parts. Firstly, corrosion under dewing conditions is
examined where water is forcibly condensed by cooling steel specimens in simulated sales
gas conditions. The objective is to examine the corrosion products produced in a worst-
case scenario; the effect of cyclic water condensation on corrosion product formation is
also examined. Secondly, hygroscopic corrosion is examined where the objective is to
determine if salts on the surface of the steel may lead to black powder production in
conditions where water would not be expected to condense.
3.2 Hypotheses
Based on the literature review, the following hypotheses were formulated:
1. Pipeline conditions define what thermodynamically stable corrosion products will
form on the steel surface although kinetically favored compounds will also be
encountered.
2. The formation of corrosion products follows a cyclical pattern that correlates with
the wet/dry cycle caused by external temperature fluctuations.
3. The cyclic behavior of water condensation and evaporation in sales gas pipelines
prevents corrosion from reaching steady-state resulting in the formation of a

42
kinetically favored, and potentially friable, corrosion product susceptible to
spallation and entrainment in the bulk gas.
4. Corrosion products will form on the steel surface in non-water saturated sales gas
pipelines if a hygroscopic material is present on the steel surface.
3.3 Motivation
The Abu Dhabi based natural gas company Gasco is reporting black powder
formation in their sales gas pipelines. Filtration and pigging are currently employed to
minimize the impact to downstream operations, but are costly. A joint research partnership
between the Abu Dhabi Petroleum Institute and the Ohio University Institute for Corrosion
and Multiphase Technology has been tasked with examining the role of corrosion products
on black powder formation in the Gasco pipelines as well as to examine potential methods
of black powder mitigation.

43
CHAPTER 4: METHODOLOGY AND EXPERIMENTAL DESIGN
4.1 Over view
Corrosion of steel in sales gas conditions was examined by exposing steel
specimens to CO2, CO2/H2S, or CO2/O2 mixtures in dewing or hygroscopic conditions.
Under dewing conditions, water condensation was forced by cooling the steel specimens
to a temperature lower than the gas temperature in a water-saturated environment. In
hygroscopic conditions, NaCl was deposited onto steel specimens that were then placed in
a relative humidity-controlled environment that allows deliquescence of the deposited
NaCl or maintains a solid salt layer on the steel. Upon specimen extraction, corrosion
product morphology was examined with scanning electron microscopy, and corrosion
product composition was obtained with energy dispersive X-ray spectroscopy, Raman
spectroscopy, and X-ray diffraction. Localized corrosion was characterized by surface
profilometry.
4.2 Test Conditions
Experimental conditions were designed to simulate a GASCO sales gas
environment known to produce black powder. Pipeline conditions used to determine
experimental parameters are shown in Table 4. Oxygen and water contents were not
measured. The 40 inch (1 m) GASCO pipeline transported the gas at a flow rate of 550
MMscfd at an operating pressure of 40-42 bar and an operating temperature of 33-50°C.
From the specified gas concentration and the highest operating pressure, the partial
pressures of CO2 and H2S were calculated to be approximately 0.55 bar and 0.3 mbar,
respectively.
44
The experimental gas composition was primarily based on the partial pressure of
H2S measured in the pipeline. The experimental H2S partial pressures were 0, 0.1, 0.3, and
1.0 mbar to examine corrosion in simulated GASCO conditions, and the effect of increased
and deceased H2S contents. The H2S was diluted with CO2 for a CO2 partial pressure of
0.96 to 0.98 bar. Corrosion at 0.1 – 1 mbar H2S is known to be H2S dominant, so dilution
over the GASCO conditions should not affect results. Tests with O2 were conducted at 10
mbar O2, the O2 partial pressure expected if the O2 content was approximately 0.02 mol%.
Table 4: Sales gas composition from which experimental parameters are derived
Component Units of Measur ement Content

Methane mol % 93.60
Ethane mol % 03.25
Propane mol % 00.58
i-Butane mol % 00.10
n-Butane mol % 00.16
i-Pentane mol % 00.03
n-Pentane mol % 00.02
Hexane+ (C6 and higher) mol % 00.01
Carbon Dioxide mol % 01.31
Nitrogen mol % 00.84
Hydrogen Sulfide ppm 7
Carbonyl Sulfide ppm 3
Mercaptans ppm 5
The steel used for all experiments is API 5L X65 mild steel, a grade of steel
typically used for gas transmission pipelines. The steel composition is listed in Table 5.
The steel is a low carbon steel with a ferritic-pearlitic microstructure. The microstructure
is shown in Figure 7, which was determined after etching with a 2% Nital solution (HNO3
45
in CH3CH2OH). Cementite (Fe3C) and ferrite are found in the white and dark regions,
respectively.
Table 5: Composition of API 5L X65 steel used for testing (wt.%)
C Al As Co Cu Mn Mo Nb Ni P
0.05 0.033 0.015 0.012 0.14 1.21 0.16 0.03 0.38 0.004
S Sb Si Sn Ti V Zn Fe
<0.001 0.035 0.25 0.012 0.01 0.04 0.004 Balance
Figure 7: Microstructure of the X65 steel specimens. Ferrite and cementite are found in
the dark and white locations, respectively.

46
4.3 Black Powder For mation in Dewing Conditions
Dewing conditions were examined by exposing specimens to water-saturated CO2,
CO2/H2S or CO2/O2 mixtures. Water condensation was stimulated onto the steel specimens
by cooling their surfaces to 15°C or 25°C for high or low water condensation rates (WCR),
respectively. The water-saturated gas was generated by bubbling CO2/H2S mixtures
through deionized (DI) water heated at a temperature to maintain the gas temperature at
30°C. A 30°C gas temperature was selected by rounding down the 33°C GASCO sales gas
operating temperature; the temperature most likely to condense water. The apparatus and
procedures for testing in dewing conditions are based largely on those used by Yaakob, et
al., [37] to test marginally sour TLC conditions. Modifications, described below, were
made to better control the steel temperature.
4.3.1 Dewing Corrosion Equipment
A glass cell setup, as shown in Figure 8, was used for testing black powder
formation in water-condensing conditions. Gas, metered by a rotameter at the desired CO2,
H2S, and O2 concentrations, was bubbled through 1 liter of heated deionized (DI) water
located in the bottom of the glass cell to heat and saturate the gas with water. The
temperature of the water, which is controlled by the hotplate, is set such that the gas above
the solution is held at the desired temperature of 30°C.
Two specimens were included in each test: a 1.25-inch-diameter × 0.5-inch-thick
(31.7 mm × 12.7 mm) cylindrical specimen for surface and corrosion rate analysis, and a
0.5×0.5×0.08 inch³ (12.7×12.7×2 mm³) specimen for XRD compositional analyses. The
cylindrical specimens were inserted directly into PEEK specimen holders designed to
47
provide a gas-tight seal to the lid while exposing a specimen surface for temperature
control. The smaller XRD specimens were suspended from the lid with a specially designed
holder, shown in Figure 10, where the steel specimens were held by a magnet within an
aluminum body, which aids heat transfer to the Peltier thermoelectric cooler. The steel
specimens were coated with Xylan on surfaces in contact with the holder to prevent
galvanic effects. Additional design parameters can be found in Appendix A.

48
Figure 8: Dewing corrosion glass cell. Image courtesy of Cody Shafer.

49
C
B
A
Figure 9: Specimen holding and control stack including (A) specimen holder, (B) Peltier,
(C) Water cooled heat sink, and (D) Anchor bar and bolts. Image courtesy of Cody
Shafer.
Figure 10: XRD specimen holder. Image courtesy of Cody Shafer.
Specimen temperature was controlled using a specially designed Peltier control
system. The temperature is controlled by measuring the temperature with a thermistor
mounted to the side of the cylindrical specimens or within the aluminum body of the XRD
50
specimen holder, computing the output with a PID (proportional-integral-derivative)
feedback control loop, and supplying the computed power to a thermoelectric cooler in
contact with the specimen. The thermoelectric cooler, more commonly known as a Peltier
after the French physicist Jean Charles Athanase Peltier, acts as a heat pump when
electrical power is applied. By controlling the amount of power supplied to the Peltier, the
amount of heat transferred can be controlled. Power to the Peltiers is supplied by an
external AC to DC power supply controlled with pulse width modulation. During
operation, a significant amount of heat is generated, so a water-cooled heatsink is placed
over the Peltier to prevent it from overheating. Temperature control was primarily done
with a TC-48-20 Peltier controller by TE Technology®, but very similar results were
obtained using a controller developed by the author of this thesis based on an Arduino®
microcontroller.
To examine the effect of cyclic temperature fluctuations, the temperature of the
steel specimen must be periodically changed. To control the temperature changes with a
Peltier, the polarity of the Peltier must be switched. Changing the polarity of the Peltier
changes the direction heat is pumped and, therefore, switches the hot and cold sides of the
Peltier. The TC-48-20 Peltier controller cannot change the polarity internally, so a double-
pole-double-throw mechanical relay was connected between the controller and the relay.
MOSFETs in an H-bridge configuration were used to control both the power and the
polarity in the Arduino controller. Additional information regarding the controllers can be
found in Appendix A.
51
4.3.2 Dewing Corrosion Experimental Procedures
The glass cell shown in Figure 8 was cleaned then 1 liter of deionized water
introduced. The glass cell was then placed on the hot plate and the DI water heated to 35°C
to maintain the gas temperature at setpoint. The glass cell was purged with CO2 for at least
3 hours by sparging the CO2 through the DI water.
Prior to insertion, the specimens were coated with Xylan®, a fluoropolymer, to
prevent the specimens from corroding on undesired surfaces. Following routine
procedures, the specimens were then ground with silicon carbide abrasive papers to a 600
grit finish with water as the coolant. The water was rinsed from the surface with
isopropanol immediately after grinding to prevent unwanted corrosion. The specimens
were cleaned in isopropanol in an ultrasonic bath, dried with cool air, and then the initial
mass was recorded. Specimens were stored in nitrogen until insertion, which was ordinarily
less than an hour after polishing.
The specimens were inserted into the specially designed holders outlined above and
mounted to the glass cell lid. The Peltiers and heatsinks were placed over the specimens,
and the resultant stack fastened to the lid. The thermistors and Peltiers were connected to
the controllers, and specimen temperature control was started. The glass cell was allowed
to purge for an additional 30 minutes with CO2 before H2S was then added into the system.
The H2S concentration was set by mixing a CO2 and a CO2/H2S stream metered by a
rotameter upstream of the glass cell. The H2S concentration was measured with
colorimetric gas detector tubes (GASTEC 4HM) and adjusted as necessary. Gas was
continuously purged through the glass cell throughout the experiments. Effluent gas
52
containing H2S was passed through a sodium hydroxide solution and finally an activated
carbon scrubber. Test conditions were monitored and adjusted as needed to maintain
experimental parameters.
After the allotted exposure time, the specimens were removed and immediately
rinsed with isopropanol to remove water and prevent unwanted oxidation. The rinse was
performed by targeting the top of the specimen with a stream of isopropanol from a squirt
bottle and allowing the isopropanol to fall over the rest of the specimen. The wall shear
stress of the isopropanol rinse was calculated to be about 1 Pa, roughly the same wall shear
stress in the GASCO pipeline. Whether or not corrosion product was found in the waste
isopropanol was recorded. The specimens were dried under nitrogen, optical images were
taken, and the specimen with corrosion product was weighed. The specimens were then
stored in a desiccator under nitrogen awaiting further analysis.
4.4 Black Powder For mation in Hygr oscopic Conditions
The effect of salts on the corrosion of steel in unsaturated conditions was examined
by depositing salts onto steel specimens and exposing the specimens to relative humidity-
controlled CO2, CO2/H2S and CO2/O2 mixtures. The three relative humidities tested were
75%, 58%, and 33% controlled through the use of saturated aqueous solutions of NaCl,
NaBr, and MgCl2, respectively.
4.4.1 Hygroscopic Corrosion Equipment and Procedures
Hygroscopic corrosion tests were performed in a glass cell with the specimens
located on a platform above the saturated salt solutions in the bottom of the glass cell, as
shown in Figure 11. The apparatus consisted of a glass cell, TeflonTM lid, lid clamp, gas
53
inlet, gas outlet, specimen platform, and thermocouple. A large hole was made in the lid
for fast specimen insertion and extraction. The large hole was sealed with a rubber stopper,
which was clamped down, if necessary, to ensure a gas-tight seal.
Figure 11: Hygroscopic corrosion glass cell. Image courtesy of Cody Shafer.
Prior to testing, the glass cell, lid and stage are thoroughly cleaned with DI water
and isopropanol then dried. In the bottom of the glass cell, 100 ml of DI water was mixed
with salt at 125% of the solubility limit in water at 25°C. The stage is then placed into the
glass cell and the lid is fastened. The glass cell is purged with N2 for a minimum of 4 hours
to remove oxygen.
54
X65 steel specimens, the same geometry as those used in dewing conditions, were
coated with Xylan on all surfaces where corrosion is undesired. The steel surface was
ground with carbide abrasive paper and then polished to a mirror finish by polishing with
a 0.25 µm diamond suspension. The specimens were then thoroughly rinsed with DI water
and ultrasonically cleaned in isopropanol. Residual isopropanol was evaporated with cool
air, and the specimen mass was recorded. The specimens were stored in N2 until salt was
deposited onto the surface or the clean specimen was inserted into the glass cell.
Salt layers were generated by drying a NaCl/DI water solution placed on the steel
specimens with N2. A 3.5 wt.% NaCl solution was prepared prior to experimentation and
used for all experiments; this is a similar salinity to seawater. The salt solution was purged
with N2, and 250 µl of solution was deposited onto the steel surface with a pipette. The
specimen was immediately placed in a N2 environment, and the water was then spread over
the entire surface by tilting the specimens. The salt solution was dried by passing dry N2
over the wetted steel surface. Once the salt crystallized, the specimen mass with salt was
measured, photographed, and then placed in the N2-purged glass cell on the platform salt-
side up.
The glass cell was allowed to purge with N2 for another 30 minutes before
replacement with the test gas, which was continuously sparged into the system throughout
each experiment. The dry test gas entered the glass cell through a tube with the outlet
approximately 1 cm above the saturated salt solution. The gas was not bubbled through the
salt solution to prevent aerosol formation, which can contaminate the steel surface; this
phenomenon was observed in preliminary experiments with saturated MgCl2. A small gas
55
flow rate was maintained throughout the experiment to prevent a decrease in the gas RH.
In sweet conditions, a hygrometer (GE Panametrics Moisture Monitor Series 35) was
placed downstream of the glass cell to verify the relative humidity, but the hygrometer
could not be used in H2S testing due to chemical incompatibility. Specimen extraction was
performed in the same manner as outlined in the dewing corrosion procedures.
4.5 Analytical Methods
Scanning electron microscopy (SEM) and optical microscopy were performed to
examine the corrosion product morphology. SEM was performed with a JEOL JSM-
6390LV. SEM sampling locations were selected to be representative of the specimen
surface at large. Hygroscopic corrosion specimens were sputter-coated with palladium to
minimize charging in the SEM. Cross-sectional SEM was performed after first mounting
the specimen in epoxy to stabilize the corrosion product and cutting the specimen normal
to the steel surface with a diamond-tipped saw blade. The cross-section was polished to a
mirror finish through the use of consecutively finer grit silicon carbide abrasive paper and
diamond suspensions. The cross-sectioned surface was sputter-coated with palladium to
minimize charging.
Corrosion product/phase composition was determined locally with energy
dispersive X-ray spectroscopy (EDS), Raman spectroscopy, and, more generally, by X-ray
diffraction (XRD). EDS was performed with an EDAX Genesis attachment on the SEM.
Raman spectroscopy was conducted with a Bruker Senterra with a 785 nm wavelength
laser. XRD was performed on a Rigaku Ultima IV with a CuKα source over 10-70° 2θ at
a scan rate of 1° per minute. The corrosion product was removed in accordance with ASTM
56
G1-03 [43], and the corrosion rate and corrosion product mass was calculated by
comparing the after removal mass to the initial mass and the after extraction mass,
respectively. A detailed description of calculation methodologies can be found in Appendix
B. Localized corrosion was examined by optical surface profilometry using an Alicona
InfiniteFocus microscope.
4.6 Safety
The primary risk during the work conducted for this thesis is exposure to hydrogen
sulfide through inhalation. H2S atmospheres above 100 ppm are considered immediately
dangerous to life and health. To protect against H2S inhalation, experiments with H2S
concentrations above 0.1 mbar (100 ppm) H2S were conducted in a segregated laboratory
at the Ohio University Institute for Corrosion and Multiphase Technology, which has a
safety system designed for H2S experimentation.
The safety system contained multiple levels of defense. The first level is, of course,
maintaining a well-sealed glass cell. If the gas inlet flow rate did not match the gas outlet
flow rate, then H2S was not added into the system. The second level was the vent hood in
which the glass cell was placed. Any leakage of H2S from the glass cell would be largely
contained by the vent hood and sent to an activated carbon scrubber. Then, if necessary,
the effluent gas is passed through a combustion system and released. The third level of
defense was personal protective equipment (PPE). PPE worn included both an H2S badge
sensor and a self-contained breather apparatus (SCBA). The H2S badge was worn during
any entry into the H2S room, regardless of whether or not experiments were being
conducted. The badges have two audible and tactile alarms with the low alarm at 10 ppm
57
and a high alarm at 15 ppm. The badges were checked daily with a 25 ppm H2S/air mixture
to ensure they functioned properly. The SCBA was worn during any entry in the H2S room
while an H2S experiment was conducted. Air was supplied to the SCBA wearer from
cylinders located outside the H2S room through hoses to prevent H2S inhalation in the event
of a release. Lastly, in case all previous safety layers failed, the buddy system was
employed where trained personnel were stationed outside the H2S room in visual contact
with the person inside and ready to enter the laboratory at a moment’s notice in the event
of incapacitation. H2S experiments below the 100 ppm H2S threshold were conducted in a
vent hood outside the H2S laboratory. A badge was still worn when working in the vent
hood in case of an accidental gas release.
The other primary safety concern is hydrochloric acid. Concentrated hydrochloric
acid is a primary component of the solution, commonly known as Clarke solution, used to
remove corrosion product. Any work with hydrochloric acid was done in a vent hood to
minimize exposure to HCl fumes. Nitrile gloves were also worn to prevent skin contact.
58
CHAPTER 5: RESULTS AND DISCUSSION
5.1 Dewing Cor r osion
If pipeline conditions are favorable, water may condense on the steel wall and allow
CO2, H2S, and O2 to corrode the steel and develop corrosion products, which may spall and
develop black powder. Experiments were performed to investigate the formation of
corrosion products in dewing sales gas conditions. The results of the experiments are
reported and discussed below.
5.1.1 Effect of H2S Partial Pressure on Dewing Corrosion
To examine the effect of H2S on the corrosion of pipeline steel in water condensing
conditions, experiments were conducted at varying H2S partial pressures. H2S partial
pressures tested were 0, 0.1, 0.3, and 1 mbar with calculated water condensation rates
(WCR) of 0.015 and 0.05 ml/m2/s. Experimental conditions are summarized in the test
matrix shown in Table 6.

59
Table 6: Test matrix for studying the effect of H2S partial pressure on dewing corrosion.
Par ameter Value

Test Mater ial API 5L X65
Total Pr essur e (bar ) 1
H 2S Par tial Pr essur e (mbar ) 0, 0.1, 0.3, 1
O 2 Par tial Pr essur e (mbar ) 0
CO 2 Par tial Pr essur e (bar ) 0.96
N2 + Ar Par tial Pr essur e (bar ) 0
Test Dur ation (days) 3
Water Condensation Rate (ml/m 2/s) 0.015, 0.05
Steel Temper atur e (oC) 25, 15
Gas Temper atur e (oC) 30
After extraction from the glass cell, the specimens were analyzed using SEM.
Micrographs of the corrosion product on the steel surface after the extraction of specimens
are shown in Figure 12. In CO2 conditions, minimal corrosion product other than residual
iron carbide was present. Some iron carbonate was present at the low WCR, whereas iron
carbide was only present at the high WCR. At 0.1 mbar H2S, a friable corrosion product
was present at both WCRs, but buckling was present in regions presumed to be the edges
of water droplets at the low WCR. At higher H2S partial pressures, FeS spallation was
much more pronounced. A higher degree of corrosion product layer buckling was observed
at 0.3 mbar H2S than at 0.1 mbar, but the buckles were confined to specific regions.
Buckling was also present at the high WCR at 0.3 mbar H2S. At 1.0 mbar H2S, severe
buckling was observed over the specimen surface at the low WCR. At the high WCR,
60
buckling still occurred, but was not nearly as severe. Buckle-driven spallation is examined
further in a later section.
Diffraction patterns of specimens recovered from the H2S experiments examining
the effect of pH2S are shown in Figure 13. At the low WCR, XRD did not pick up any
corrosion product on the specimens analyzed from the CO2 experiment, but mackinawite
(FeS) was detected in tests with H2S. Increasing the H2S partial pressure saw an increase
in mackinawite peak intensity, but the relative peak intensities did not match the
mackinawite XRD reference data [28] indicating preferred orientation of the mackinawite
layer on the steel substrate. This has been postulated to be due to a growth mechanism
favoring parallel [100] and [001] iron and mackinawite planes, respectively, at the steel
surface. At the high WCR, mackinawite was detected at 0.3 and 1.0 mbar H2S, but not at
0.1 mbar.
61
0 mbar – Low WCR 0 mbar – High WCR
0.1 mbar – Low WCR 0.1 mbar – High WCR
Figure 12: Effect of H2S partial pressure and water condensation rate on corrosion
product morphology.
62
Figure 13: XRD data showing the effect of H2S partial pressure and WCR on corrosion
product composition
63
The effect of H2S concentration on corrosion rate is shown in Figure 14. For both
WCRs the addition of 0.1 and 0.3 mbar H2S generally decreased the corrosion rate due to
the corrosion product offering more protection. However, increasing the H2S partial
pressure to 1.0 mbar saw an uptick in the corrosion rate, particularly for the low WCR. The
dramatic increase in corrosion rate at 1.0 mbar H2S and low WCR is presumed to be due
to the much higher degree of spallation observed at 1.0 mbar H2S than at 0.3 and 0.1 mbar
H2S, since a spalling layer cannot offer as much protection to the steel as a more stable
layer. The spallation of a corrosion product would expose the steel surface to further attack.
The smaller uptick in corrosion rate observed in the 1.0 mbar and high WCR experiment
relative to its low WCR counterpart is likely due to the lesser degree of corrosion product
spallation.
Figure 14: Effect of H2S partial pressure and water condensation rate on corrosion rate.
64
Figure 15: Effect of H2S partial pressure and water condensation rate (L = low, H = high)
on the measured and theoretical corrosion product masses.
The measured corrosion product mass and theoretical corrosion product mass are
shown in Figure 15. The measured corrosion product mass is the mass of corrosion product
removed with the inhibited hydrochloric acid (Clarke solution) in the procedure described
by ASTM G1. The theoretical corrosion product mass is determined by calculating the
maximum amount of corrosion product possible derived from iron lost due to corrosion.
The theoretical corrosion product mass calculation assumes all iron lost due to corrosion,
not in the form of Fe3C, is converted to FeCO3 under sweet conditions and FeS under sour
conditions. Knowledge about the theoretical corrosion product mass is useful as a
comparison against the measured corrosion product mass as it gives some insight into the
amount of iron that was lost in removed condensed water or spalled corrosion product. As
65
evidenced by the large difference in the measured and theoretical corrosion product mass,
most of the iron available for corrosion product formation on the steel surface was lost.
The difference in the theoretical and measured corrosion product masses generally
decreased with increasing H2S concentration, but this is expected since corrosion product
layers are easier to produce in sour systems than sweet systems at the temperatures tested.
The maximum black powder production rate shown in Figure 16 is calculated by
extrapolating the theoretical corrosion product mass averaged by the experimental duration
to a hypothetical 100 km length of a 42 inch (106.7 cm) ID pipeline. The maximum black
powder production rates varied from 900 to 2000 kg/day for a 100 km, 42-inch ID pipeline.
These numbers are high, but they represent a worst-case scenario where the entire
hypothetical pipeline is experiencing water condensation.
Figure 16: Effect of H2S partial pressure on maximum possible black powder production
rate in a 100 km, 42 inch ID pipeline.

66
5.1.2 Effect of Time on Sour Dewing Corrosion
The effect of time was examined by exposing steel specimens to 0.3 mbar H2S and
0.96 bar CO2 in an O2-free environment for up to 10 days at the low WCR. Experimental
parameters are shown in the test matrix shown in Table 7.
Table 7: Test matrix for analyzing the effect of time on sour dewing systems
Par ameter Value

H 2S Par tial Pr essur e (mbar ) 0.3

Test Dur ation (days) 0.25, 1, 3, 7, 10
Water Condensation Rate (ml/m 2/s) 0.015
Steel Temper atur e (oC) 25
The effect of time on the corrosion product layer morphology is shown in Figure
17. After 6 hours, the only visible corrosion product was confined to the structures shown.
Flakes were already beginning to form, but no buckling was observed. FeS layer buckling
was evident after 1 day where the buckled corrosion product was evident only surrounding
the flakey structures seen after 6 hours. The severity of buckling intensified with time
developing large buckled regions after 7 days. Similar sized features were not observed
after 10 days as they likely had been removed from the surface due to spallation.
67
0.25 Days 1 Day
3 Days 7 Days
10 Days
Figure 17: Effect of Time on Corrosion Product Morphology (0.3 mbar H2S, 0.96 bar
CO2, 0 mbar O2, 0.015 ml/m2/s WCR, 25 Tsteel, 30°C Tgas)
The effect of time on the corrosion product composition is shown in Figure 18. No
change in phase identity associated with corrosion product was observed, however,
68
absolute peak intensities generally increased as the corrosion product developed. No XRD
measurements were performed on the specimen retrieved after 6 hours.
Figure 18: Effect of time on corrosion product composition
Time dependency relating to corrosion rate is shown in Figure 19. The corrosion
rate decreased with time from 0.28 mm/yr to a steady corrosion rate of 0.10 mm/yr after
corrosion product, which can offer some protection, had time to develop on the steel
surface. The error in the measurement decreased significantly over time since mass
measurement uncertainties had less impact on the error-propagated result as time increased.
69
Figure 19: Effect of time on corrosion rate at 0.3 mbar pH2S and low WCR.
As shown in Figure 20, the measured corrosion product mass on the steel surface
is approaching a steady value near 10 g/m2 while the theoretical corrosion product mass
becomes approximately linear with time. This combined with the changes in corrosion
product morphology over time indicate the corrosion product formation rate and spallation
rate are approaching equilibrium and may ultimately lead to a black powder production
rate of ca.1000 kg/day for a 100 km, 42 in. ID pipeline. Note, the errors in the measured
and theoretical corrosion product mass measurements were relatively constant since they
are based soley on gravimetric measurements and did not consider time, whereas the
maximum possible black powder production rate is based on the corrosion rate, which had
time to negate some of the effects of mass measurement error propagation.

70
(a) (b)
Figure 20: Effect of time on (a) corrosion product mass and (b) maximum possible black
powder production rate in a 100 km long, 42 in. ID pipeline at 0.3 mbar pH2S and low
WCR.
5.1.3 FeS Layer Buckling
The increased degree of buckling with increasing H2S partial pressure and time in
the FeS layers is related to the accumulation of compressive stress accompanying FeS layer
growth. Spontaneous buckling and layer delamination occurs to release stored energy
within the intrinsically stressed FeS layer.
Stress accumulation may be caused through a variety of pathways. In the Sun-Nesic
model [20] of iron sulfide layer formation, stress accumulation, which leads to cracking
and spallation, is postulated to be due to volume differences between the ferrite and the
mackinawite. The ratio of the scale volume to the substrate volume, more commonly
known as the Pilling-Bedworth ratio (PBR) [44], for FeS relative to ferrite is about 2.6.
Since the PBR is greater than 1, compressive stress within the layer is expected to
71
accumulate and lead to layer failure. However, the PBR is not especially useful for layers
grown through outward cation diffusion, which is the case for most sulfides. Iron cations
may diffuse faster than the much larger sulfide anions. Layers predominantly grown via
outward cation diffusion can grow outward from the scale surface and, therefore, are not
volume limited as required to apply the PBR [45].
Sun and Nesic suggest epitaxial stresses, which arise due to unit cell size
differences may lead to stress accumulation [20]. However, the adherence of the sulfide to
the metal is expected to be fairly weak due to void formation as the steel substrate corrodes,
which may limit epitaxial stresses [38]. A perhaps more plausible method of stress
accumulation may be due to crystal growth at cracks, pores, and grain boundaries. Crystal
growth due to the addition of ferrous and sulfide ions to an already formed crystal may lead
to crystals pushing against one another therefore compressively stressing the layer. Crystal
growth at grain boundaries and cracks has been used to explain the buckling of chromium
oxide scales and oxide scales on Fe-Cr-Al alloys [45]–[48].
In all H2S experiments, black particles were present in collected fluid after the
isopropanol rinse of extracted specimens. The wall shear stress imposed by the isopropanol
was estimated to be approximately 1 Pa based on the flow of a falling laminar film [49].
This shear stress is similar to the wall shear stress in the GASCO pipeline. It is unlikely
this small shear stress would cause damage to a corrosion product layer experiencing little
to no intrinsic stress, so it is proposed that black powder formation due to H2S is
predominantly caused by intrinsic stresses within the layer generated by layer growth rather
72
than the extrinsic stresses caused by fluid flow. Further study is needed to better quantify
the accumulation of stress as the layer grows to better predict spallation.
5.1.4 Effect of O2 on Sweet Dewing Corrosion
The effect of oxygen was examined by exposing steel specimens at the low WCR
of 0.015 ml/m2/s to a CO2/air mixture with a 10 mbar O2 partial pressure. Test conditions
are shown in Table 8.
Table 8: Test matrix for the investigation of the effect of O2 on sweet dewing corrosion
Par ameter Value

H 2S Par tial Pr essur e (mbar ) 0

O 2 Par tial Pr essur e (mbar ) 0, 10
CO 2 Par tial Pr essur e (bar ) 0.96, 0.92
N2 + Ar Par tial Pr essur e (bar ) 0, 0.04
Water Condensation Rate (ml/m 2/s) 0.015
Steel Temper atur e (oC) 25
73
(a)
(b) (c)
Figure 21: Effect of O2 on (a) corrosion rate, (b) measured and theoretical corrosion
product mass, and (c) maximum possible black powder production rate in a 100 km, 42
in. ID pipeline.
The addition of oxygen to a CO2 environment saw, as expected, a dramatic increase
in corrosion rate and corrosion product mass, as shown in Figure 21(a) and Figure 21(b).
The measured corrosion product mass was much closer to the theoretical corrosion product
74
mass than the CO2-only experiment indicating the water chemistry was more favorable for
precipitation.
The dramatic increase in the corrosion rate and corrosion product mass resulted in
development of a bilayered corrosion product, as shown in the SEM cross-section in Figure
22(a). The inner layer shown in Figure 22(b) was well adhered to the steel, but the outer
layer, Figure 22(c), was more loosely bound. Red flakes of corrosion product were
observed in the collected isopropanol after rinsing, which suggests black powder
components would readily detach.
The origin of the outer layer is unclear. The outer layer is presumed to have grown
on the gas/liquid interface since a flat surface is present on the outer layer, whereas the
inner layer is more rough. It is postulated that the precipitation of iron carbonate in the bulk
occurred away from the steel surface whereupon it collected at the gas/water interface due
to gravity.
75
(a)
(b) (c)
Figure 22: Surface analysis of corrosion products formed after a 3 day exposure to a 10
mbar O2 – 0.92 bar CO2 atmosphere at the low WCR.
Seemingly contradictory results were obtained from corrosion product composition
analysis with XRD and Raman spectroscopy as shown in Figure 23. Siderite (FeCO3) was
detected with XRD, but hematite (Fe2O3) was detected by Raman spectroscopy. It is
postulated that an amorphous or nanocrystalline hematite layer was present on the
corrosion product surface, which cannot be detected by XRD but is detected by Raman
spectroscopy. Indeed, the high XRD baseline is also indicative of the presence of an
76
amorphous corrosion product. As Raman is more suface specific, the iron carbonate
beneath the iron oxide was undetected.
Figure 23: (a) XRD and (b) Raman spectroscopy compositional analysis of corrosion
product formed after a 3-day exposure to a 10 mbar O2 – 0.92 bar CO2 atmosphere at the
low WCR. The red hematite spectrum is added as a reference [50].
Under pure CO2 environments, corrosion products were minimal and consisted
primarily of iron carbide. However, the predominance of iron carbonate over iron
oxyhydroxides and iron oxides even in the presence of dissolved O2 suggests black powder
77
formation may still be possible under pure CO2 conditions if conditions are favorable for
precipitation to occur into a morphology, such as the bilayered morphology seen above,
capable of producing detached to marginally attached particles of corrosion products.
5.1.5 Effect of Cyclic Water Condensation
Pipeline conditions are not static. Temperatures are expected to change throughout
the year and even throughout the day due to meteorological conditions. To examine the
effect of temperature fluctuations on the corrosion product, the temperature of the
specimens was changed every 12 hours between 25°C and 40°C. These temperature
changes are supposed to mimic temperature changes that occur due to the day/night cycle.
Experimental conditions are summarized in the test matrix in Table 9.
Table 9: Test matrix for the investigation of the effect of cyclic water condensation
Par ameter Value
H 2S Par tial Pr essur e (mbar ) 0.3
Test Dur ation (days) 3, 7
Water Condensation Rate (ml/m 2/s) 0.015, 0
Steel Temper atur e (oC) 25, 40
Temper atur e Cycle Per iod (days) 1 (change every 12 hours)
78
SEM surface analysis was performed after each experiment to examine the effect
of temperature cycling on the corrosion product morphology. The surfaces of specimens
that experienced cyclic water condensation are compared to their constant WCR
counterparts in Figure 24. Overall, FeS spalling was more apparent in the constant WCR
experiments than the cycling experiments. The less sheet-like characteristics of the cycled
corrosion product layers, as compared to the layers formed under constant water
condensation rates, is likely a result of high FeS precipitation rates as water was quickly
removed from the steel surface. It is postulated that precipitated corrosion product would
have settled on layers similar to those observed from the constant WCR experiments,
thereby reinforcing the structural integrity of the layer making spalling less pronounced.
79
Constant WCR – 3 Days Cycled WCR – 3 Days
Constant WCR – 7 Days Cycled WCR – 7 Days
Figure 24: Effect of cyclic water condensation on corrosion product morphology.
The effect of cyclic water condensation on the corrosion product composition
determined by X-ray diffraction is shown in Figure 25. Mackinawite was the primary
corrosion product detected, but siderite was also detected in the corrosion product
developed after 7 days under cycling conditions. Peak broadening is apparent in the
temperature cycled experiments. It is postulated that the fast precipitation rates caused by
the drying process results in growth of small mackinawite crystals or induces lattice strain
80
as the ions coalesce in thermodynamically unfavorable locations thus inducing peak
broadening.
Figure 25: Effect of cyclic water condensation on the corrosion product composition.
81
(a)
(b) (c)
Figure 26: Effect of temperature cycling on (a) corrosion rate, (b) measure and theoretical
corrosion product mass, and (c) maximum possible black powder production rate in a 100
km, 42 in. ID pipeline.

82
The effect of temperature cycling on corrosion rate, corrosion product mass, and
maximum possible black powder production rate is illustrated in Figure 26. As would be
expected for the case of a steel surface water wetted for about half the experimental
duration, the corrosion rates of the cycled specimens were less than those from the constant
WCR conditions. However, while the corrosion rate from the 3 day-cycled experiment was
about half of its constant WCR counterpart, the corrosion rate from the 7 day – cycled
experiment was 36 ± 4 % of the 7 day – constant WCR corrosion rate suggesting that the
solid corrosion product produced during the drying phase is more protective than the
buckled layers produced under constant water condensation rates.
5.1.6 Black Powder Formation in Dewing Conditions
Corrosion products susceptible to black powder formation were produced in
conditions favorable to corrosion product precipitation and iron sulfide film forming
conditions. The precipitated corrosion product formed in oxic conditions consisted of two
layers. The outer layer is the more likely layer to detach and form black powder as the inner
layer is attached to the steel with no evidence of inner layer delamination. The outer layer
appeared only loosely held to the steel surface with portions projected away from the
surface, which would make the layer more susceptible to removal by a flowing gas stream.
The bilayered morphology is postulated to only occur at the top of the pipeline (i.e., 12
o’clock position) where iron carbonate, which nucleated away from the steel surface, can
collect at the gas/liquid interface. Away from the top of the line, gravity will likely draw
precipitated products back to the steel surface to merge with the inner layer.
83
Spalling corrosion product films were ubiquitous in H2S-containing conditions
under constant water condensation rates. These thin films were likely a results of the direct
reaction of H2S to the steel, which forms a framework for iron sulfide precipitation as per
Zheng’s model of sour corrosion [21]. The spallation of these thin films occurs as intrinsic
stresses, which accumulate during layer growth, then relax during the buckling process. A
gas pipeline experiencing water condensation would likely have to contend with this
method of black powder formation over the entire pipe wall as there is no bilayered
morphology. It appears the only requirement to form these types of layers is a water-wetted
surface and the presence of H2S.
Corrosion products formed under dewing conditions include siderite, mackinawite,
and hematite. Of the three, mackinawite is metastable and its transformation to more
thermodynamically favored corrosion products is likely hindered at the temperatures in
which the experiments were conducted.
Spallation was not as pronounced under temperature cycling condition where it
appears corrosion product formed through precipitation served to structurally reinforce the
iron sulfide layer. No major change in corrosion product composition was obtained due to
temperature cycling.
To mitigate the formation of black powder due to dew formation within the
pipeline, proper gas dehydration is absolutely necessary. Dehydrating the gas to below
100% RH, however, may not be completely effective. The potential for corrosion below
100% RH is investigated in the following section.

84
5.2 Hygr oscopic Cor r osion
The presence of hygroscopic material such as a salt is hypothesized to sustain
corrosion in water unsaturated conditions, which can lead to black powder formation. The
possibility of hygroscopic salts sustaining corrosion at a relative humidity (RH) below
100% was investigated by depositing NaCl on polished steel specimens and exposing the
specimens to differing H2S partial pressures and relative humidities. The test matrix that
guided hygroscopic corrosion experimentation is shown in Table 10, below.
Table 10: Hygroscopic corrosion test matrix
Par ameter Value

H 2S Par tial Pr essur e (mbar ) 0, 0.1, 0.3, 1.0

Relative Humidity (% ) 33, 58, 75
Satur ated Salt Solution MgCl2, NaBr, NaCl
Temper atur e (oC) 25
Salt Layer NaCl
5.2.1 Salt Layer Generation
Surface microscopy of the salt layer, generated by drying a 3.5 wt% NaCl/deionized
water solution onto a steel surface polished to a mirror finish, is shown in Figure 27. The
85
solution was purged with N2 prior to deposition, and drying was performed by passing dry
N2 over the steel specimen and allowing the water to evaporate, slowly.
Figure 27: SEM analysis of the NaCl layer formed before specimen insertion into test
environment. The specimen was coated with palladium to minimize electron beam
charging.
As seen in Figure 27, the salt was predominantly grouped into roughly cubic
structures approximately 100 µm in width. A slight salt residue is apparent on the steel
surface between the large salt crystals.
The salt layer generation method produced some pitting on the outside of the border
of crystals, as shown in Figure 28. The micrographs shown in Figure 28 were taken after
the rinsing the salt layer shown in Figure 27 with N2-purged DI water. The methodology
employed during salt generation, sputter coating, SEM, and salt removal of the salt layer
86
in the micrographs is the same as during experimentation, therefore the attack morphology
would propagate to the experiments, whose results are shown in subsequent sections.
Attack beyond that shown in Figure 28 is a result of exposure to the atmosphere dictated
by the hygroscopic corrosion experiments. No change in steel mass was measured as a
result of salt layer deposition and removal.
(a) (b)
Figure 28: Slight localized corrosion on the outside boundary of a salt crystal formed by
drying NaCl.
5.2.2 Hygroscopic Corrosion in Sweet Environments
Surface microscopy of specimens exposed to CO2 atmospheres under deliquescing
conditions (75% RH) is shown in Figure 29. The occurrence of deliquescence is evident at
75% due to the small size of the cubic salt crystals relative to those seen in Figure 27.
Distinct regions were found under deliquescing conditions, which are the darker and the
lighter regions seen in Figure 29(a) and better detailed in the enlarged micrograph in Figure
29(b). More corrosion product is found in the darker shaded regions, whereas coverage was
87
sparser in the circular region where islands of corrosion product were obtained amidst a
material structurally consistent with iron carbide. EDS analysis of a similar region shown
in Figure 30 indicates the corrosion product is compositionally consistent with iron
carbonate, which was also more generally detected by XRD, as shown in Figure 31.
(a) (b)
Figure 29: (a) Surface microscopy after a 3 day exposures to CO2 atmosphere at 75% RH.
The region outlined in red is enlarged in (b).

88
Figure 30: EDS of corrosion product after a 3 day exposure to a CO2 environment at 75%
RH. NQ: Not Quantified.
Figure 31: XRD corrosion product composition analysis of specimens recovered after a 3
day exposure to CO2 at 75% RH. Peak labels are S: siderite (FeCO3), H: halite (NaCl),
and Fe: ferrite (Fe).

89
The reason for the non-uniform corrosion product is unclear but is attributed to the
presence of a large salt crystal that once covered the circular area. The circular shape
compared to the typical cubic shape is due to the rounding of the NaCl crystal edges during
deliquescence, a phenomenon witnessed by Wise, et al., [51] with environmental
transmission electron microscopy.
Significant corrosion took place beneath the salt crystal as evident by the
approximately 1 µm difference between regions beneath salt crystals and the surrounding
regions in the surface profilometry shown in Figure 32. Deeper localized corrosion was
also observed typically beneath the smaller crystals than the larger crystals.
90
Figure 32: Surface profilometry of specimen recovered after a 3 day exposure to CO2 at
75% RH. Scale bar units are in nm.
Corrosion product was not directly observable by SEM after a 3-day exposure to
CO2 at 58% and 33% RH, as shown in Figure 33. The large NaCl crystals were still present
at 58% and 33% RH indicating deliquescence did not occur, however, a ring surrounding
the salt crystal is apparent at 58% RH. Also, no corrosion product was detected by XRD
beneath the DRH of NaCl in CO2 conditions.

91
0 mbar H2S – 58% RH 0 mbar H2S – 33% RH
Figure 33: Surface microscopy and XRD compositional analysis after a 3 day exposure to
CO2 at 58 and 33% RH.
The effect of relative humidity on the general corrosion rate due to hygroscopic
corrosion in CO2 atmospheres is shown in Figure 34(a). As expected, the corrosion rate
increased with the availability of water in the atmosphere; however, the corrosion rates
were all very small. The corrosion rate at 33% RH cannot be distinguished from no
corrosion due to its small value and gravimetric measurement uncertainty. The small
corrosion rates are partially owed to the nonuniform coverage of the salt layer making
92
hygroscopic corrosion an inherently localized system. Despite these reservations, a general
corrosion rate is still useful for comparison purposes, so the value is still provided.
Figure 34: Effect of relative humidity on (a) general corrosion rate, (b) measured and
rate in a 100 km, 42 in. ID pipeline under hygroscopic CO2 conditions.

93
The effect of relative humidity on the corrosion product mass and black powder
production rate is shown in Figure 34(b) and Figure 34(c), respectively. Below the DRH,
the measured corrosion product mass was higher than the theoretical corrosion product
mass. This is believed to be due to the presence of NaCl in the corrosion product that could
not be rinsed off before the corrosion product was removed.
5.2.3 Hygroscopic Corrosion in Sour Environments
The morphology of corrosion products obtained in deliquescing conditions is
shown in Figure 35. Flakey corrosion products potentially susceptible to spallation were
only obtained at 0.3 and 1.0 mbar H2S where the corrosion product appeared to create a
shell around the location once occupied by a salt crystal. Corrosion product formation atop
of the salt crystal is postulated to be due to ferrous ion diffusion through the water layer
surrounding the deliquescing salt crystal. FeS precipitation then occurred with sulfides
already present. The shell of corrosion product would then be all that remains once the salt
crystal fully deliquesces. No evidence of a friable corrosion product was found at 0.1 mbar
H2S.
94
0 mbar H2S 0.1 mbar H2S
0.3 mbar H2S 1.0 mbar H2S
Figure 35: Effect of H2S partial pressure on corrosion product morphology in
deliquescing conditions (75% RH, 0.97 bar CO2).
The effect of H2S partial pressure on the composition of corrosion products formed
under deliquescing conditions is shown in Figure 36. With the increase of H2S partial
pressure to 0.1 mbar, mackinawite was the dominant phase, however the [001] peak at
17.6° 2θ was quite broad. At 0.3 mbar and 1.0 mbar H2S, both siderite and mackinawite
were present with an additional unknown phase(s). The baseline of the diffraction patterns
95
collected at 0.3 and 1.0 mbar H2S were high indicating the presence of amorphous or
nanocrystalline material.
96
M: mackinawite, S: siderite, H: halite, and Fe: ferrite. Peaks labeled with an asterisk (*)
correspond to unknown phases.

97
0 mbar H2S – 58% RH 0 mbar H2S – 33% RH
0.1 mbar H2S – 58% RH 0.1 mbar H2S – 33% RH
Figure 37: Effect of H2S partial pressure on surface morphology at 58 and 33% RH.
98
The effect of H2S partial pressure at relative humidities below the DRH on surface
morphology is examined in Figure 37. Little to no corrosion is observable in the
micrographs, however XRD detected the presence of mackinawite at 0.3 and 1.0 mbar H2S
at 58% RH as shown in Figure 38. As was the case in their 75% RH analogous experiments,
the mackinawite peaks were very broad. No corrosion product was detected by XRD at
33% RH regardless of H2S partial pressure.

99
M: mackinawite, H: halite, and Fe: ferrite.

100
Figure 39: Effect of H2S partial pressure on (a) general corrosion rate, (b) measured and
rate in a 100 km, 42 in. ID pipeline under hygroscopic conditions.

101
The effect of H2S partial pressure on the general corrosion rate, corrosion product
mass, and maximum possible black powder production rate is shown in Figure 39. At 0.3
and 1.0 mbar H2S, the corrosion rates were close to those seen in dewing conditions,
however, the corrosion rates saw a large drop as the RH fell below the deliquescence
relative humidity (DRH) of NaCl (75% RH). Corrosion was measurable at 0.3 mbar H2S
and 33% RH, but the corrosion rates at 0, 0.1 and 1.0 mbar and 33% RH measured could
not be distinguished from a corrosion rate of zero due to measurement uncertainties. The
measured corrosion product mass was consistently higher than the theoretical mass, but
this is likely due to the presence of NaCl in the measured corrosion product that was not
quantified.
5.2.4 Corrosion Below the Deliquescence Relative Humidity
Based on the results above, it is apparent that corrosion occurred below the DRH,
but corrosion products were most readily observed if deliquescence occurred.
Schindelholz, et al., [40] examined the effect of relative humidity on the corrosion of steel
beneath NaCl particles in oxic environments. Corrosion below the DRH was attributed by
Schindelholz, et al., to adsorbed water on both the salt and the steel, deliquescence-
efflorescence hysteresis, and changing chemistries due to corrosion.
The initiation of corrosion due to the hygroscopic salt particles was attributed to
the presence of adsorbed water on both the salt and the steel surface and capillary
condensation at the gas/salt/metal interface. Both the salt and the steel surface can adsorb
water. NaCl has been shown to produce adsorbed water layers with properties similar to
bulk water at a relative humidity of 40-50% [52]. Similarly, water adsorption on iron was
102
found to lead to a potentially corrosive water layer at about 50% RH [53]. The occurrence
of corrosion at 33% RH in the oxic environments tested was attributed to the interaction of
the two water layers on the salt and the steel through capillary condensation at the confined
salt/steel interface [40].
Once initiated, the sustainment of corrosion is expected to be due to deliquescence-
efflorescence hysteresis and the changing electrolyte chemistries as corrosion occurs.
Deliquescence-efflorescence hysteresis is a depression in the efflorescence relative
humidity relative to the deliquescence relative humidity and is typically caused by
crystallization kinetics. In short, the driving force for crystallization at the DRH is
frequently insufficient for nucleation and crystal growth, therefore a lower RH for
efflorescence is necessary. However, the presence of corrosion products and metal surfaces
can limit the effect of deliquescence-efflorescence hysteresis as heterogeneities may serve
as nucleation sites [40], [54].
In the limited water available for corrosion to occur, a small amount of corrosion
can significantly impact the water chemistry. The addition of ferrous ions to water can
depress the effective DRH allowing corrosion to occur at even lower RH values than those
necessary to initiate corrosion. The formation of other deliquescing phases such as
FeCl2·4H2O, which has a DRH of approximately 55% [55], may indeed allow for
additional water accumulation. The presence of FeCl2·4H2O is reported to allow active
corrosion of archaeological iron artifacts down to 20% RH and induce the “weeping” of
the artifacts, whereby a highly acidic solution with high Fe2+ and Cl– concentrations is
formed as a result of exposure to a humid atmosphere [56], [57]. The formation of

103
additional species leading to locally decreased RH values is potentially important for black
powder production through hygroscopic corrosion processes as pipeline conditions are not
static. Corrosion beneath a salt crystal present on the steel surface may initiate while gas
conditions are relatively humid, but the change in chemistry and deliquescence-
efflorescence hysteresis could sustain deliquesced water even if bulk gas RH is decreased.
5.2.5 Black Powder Formation through Hygroscopic Corrosion Processes
For black powder to form via corrosion at RH values below 100%, the species on
the surface must assist in the accumulation of moisture. NaCl deposits on the steel surface
were found to lead to corrosion and corrosion product formation at relative humidities as
low as 33%, however, corrosion products that can potentially lead to black powder were
only found after 3 days if the salt was able to deliquesce.
From an engineering standpoint, the formation of friable corrosion products at
deliquescence makes black powder prediction computationally simpler. Rather than trying
to predict the corrosion due to adsorbed water layers on salt particles of variable size, one
can assume full deliquescence is needed thereby allowing the thermodynamics of salt
deliquescence in the particular system to be the deciding factor in the prediction of black
powder formation. However, this still requires some knowledge of the composition of
materials present on the steel surface. A sea salt residue will have a different DRH than a
pure NaCl residue. Luckily, the thermodynamics of salt deliquescence is relatively well
established due to its importance in atmospheric processes and food sciences.
The experiments discussed above do not necessarily negate the potential for black
powder to form without the presence of a hygroscopic salt. Other materials such as glycols
104
or corrosion products may assist in the wetting of the steel surface and sustaining corrosion.
Water accumulation leading to black powder may be facilitated by a corrosion product
present upon a steel surface. A porous corrosion product may lead to water accumulation
through capillary condensation. A liquid can spontaneously form in a nanoscopic pore at
RH values much less than 100% due to increased intermolecular interactions within the
confined space [38].
Glycols, especially triethylene glycol (TEG), are commonly used to dry natural gas
to specification. Glycol carryover from gas processing plants can lead to corrosion and
possibly black powder production through co-condensation of glycol and water and/or
through the hygroscopic uptake of water by glycol. Deposited TEG/H2O mixtures of >95%
TEG in a 0.4 mbar H2S/CO2 (1 bar total pressure) atmosphere were found to sustain
corrosion rates of 0.6 – 30 µm/year [58]. For reference, a 1 µm/yr general corrosion rate
over a 100 km, 42 in. ID pipeline would yield a maximum possible black powder
production rate of approximately 10 kg/day. Mixtures of TEG with black powder and other
residual components will result in a black sludge, which requires pigging for it to be
removed [59]. If this sludge is present within the pipeline or glycol carryover is suspected,
then a thorough cleaning with methanol or another solvent is recommended before inhibitor
application.
5.3 Exper iment Repeatability
Great care was taken to ensure consistency in laboratory practices so that repeatable
results were obtained, but the results herein are generally from one experiment per set of
conditions. The complexity of each experiment, and particularly the encountered

105
difficulties with H2S environments, limited the number of experiments that could be
conducted for corrosion rate and corrosion product mass measurements. However, the
features observed in the above microscopy were consistent with those seen in experiments
with inconclusive gravimetric measurements though otherwise fine. SEM images from
those tests and additional locations in the specimens shown above are shown in Appendix
C.
106
CHAPTER 6: CONCLUSIONS AND FUTURE WORK
6.1 Summar y of Results
Under dewing conditions, the two corrosion product morphologies were identified
which can lead to black powder production. The first is formed due to corrosion product
precipitation into a dual-layered morphology. The second is a result of the formation of
mackinawite films that buckle and spall due to compressive stress generation due to layer
growth.
Potential was found for black powder to form through hygroscopic corrosion if
deliquescence occurs. Without deliquescence the corrosion product was minimal due to a
decreased opportunity for corrosion to occur.
6.2 Hypotheses Revisited
The hypotheses formulated at the start of this study are revisited here in light of the
results shown above.
Pipeline conditions define what thermodynamically stable corrosion products will
form on the steel surface although kinetically favored compounds will also be
encountered.
The composition of the corrosion product did indeed change with the differing
conditions tested. Mackinawite was the predominant phase detected under sour conditions,
iron carbonate was found under sweet conditions, and a mix of iron carbonate and hematite
was found under oxic conditions. Also, a mixture of thermodynamically-favored species
and kinetically-favored species was also obtained. Mackinawite is a kinetically favored
phase as compared to the more thermodynamically favored pyrite. The preference for
107
mackinawite is likely due to the temperatures being too low for meaningful phase
transformation during the timeframe of an experiment.
With regards to black powder formation, the corrosion product composition seemed
to have some effect on the propensity of the layer to spall. Buckling was readily observed
in sour conditions, and loose layers were found in the oxic environments. However,
corrosion products which form through a precipitation-only mechanism did not seem to be
so readily removed compared to mackinawite, which forms through a combination of a
direct reaction with the steel and precipitation. Precipitation directly onto the steel surface
appeared to have developed a much more adherent corrosion product layer than the
mackinawite formed under H2S conditions.
The formation of corrosion products follows a cyclical pattern that correlates with
the wet/dry cycle caused by external temperature fluctuations.
The cyclic behavior of water condensation and evaporation in sales gas pipelines
prevents corrosion from reaching steady-state resulting in the formation of a
kinetically favored, and potentially friable, corrosion product susceptible to
spallation and entrainment in the bulk gas.
Corrosion and subsequent corrosion product formation did indeed seem to have
correlated with the wet/dry cycle based on the 50% reduction in corrosion rate after 3 days
in a temperature cycled condition compared to the 3 days, constant WCR experiment, but
the cyclic water condensation seemed to have an opposite effect than hypothesized on the
development of friable corrosion product layers. Less buckling was observed in the cycled
108
conditions than the constant temperature due to corrosion product precipitation upon
dehydration reinforcing preexisting corrosion product layers.
Corrosion products will form on the steel surface in non-water saturated sales gas
pipelines if a hygroscopic material is present on the steel surface.
The development of corrosion products did indeed occur under non-water saturated
conditions due to the presence of hygroscopic salts on the steel surface. Corrosion products
were more prominent, however, when the deliquescence occurred.
6.3 Comments on Black Powder Mitigation
Black powder mitigation efforts should emphasize proper gas dehydration as the
primary means of black powder mitigation. However, periodic upsets may allow water to
condense despite best efforts in dehydration. As a means of insurance against black powder
formation if liquid water was to condense, the use of corrosion inhibitors, particularity
volatile corrosion inhibitors (VCIs), should minimize the effects of corrosion and therefore
limit the production of resulting black powder. Finding an appropriate VCI to mitigate
corrosion should become a priority for future black powder research.
6.4 Futur e Wor k
Suggested research topics to extend this work are as follows:
Characterization of the mechanical properties of corrosion products formed under
dewing conditions.
Determining the role of iron sulfide growth mechanisms (i.e., epitaxial growth,
precipitation) on stress generation and layer failure.
Black powder mitigation through volatile corrosion inhibitors

109
Model development for the prediction of black powder formation.

110
REFERENCES
[1] “International Energy Outlook 2016,” U.S. Energy Information Administration,

Washington, D.C., 2016.
[2] U.S. Energy Information Administration, How much carbon dioxide is produced per
kilowatthour when generating electricity with fossil fuels? [Online]. Available:
http://www.eia.gov/tools/faqs/faq.cfm?id=74&t=11. [Accessed: 13-Nov-2016].
[3] Energy Information Administration, About U.S. Natural Gas Pipelines -

Transporting Natural Gas [Online]. Available:
https://www.eia.gov/pub/oil_gas/natural_gas/analysis_publications/ngpipeline/inde
x.html. [Accessed: 02-Jun-2016].
[4] R. Baldwin, “Black Powder in the Gas Industry - Sources, Characteristics and
Treatment,” Mechanical and Fluids Engineering Division Southwest Research
Institute, San Antonio, TA 97-4, 1997.
[5] T. S. Khan and M. S. Al-Shehhi, “Review of black powder in gas pipelines – An

industrial perspective,” J. Nat. Gas Sci. Eng., vol. 25, pp. 66–76, Jul. 2015.
[6] J. Smart, “Black Powder Movement in Gas Pipelines,” in CORROSION/2011,

Houston, TX: NACE International, 2011, Paper No. 11089.
[7] S. Mokhatab, W. A. Poe, and G. Zatzman, Handbook of natural gas transmission

and processing, Second edition. Amsterdam: Gulf Professional Publishing, 2012.
[8] A. M. Sherik, S. R. Zaidi, E. V. Tuzan, and J. P. Perez, “Black powder in gas

transmission systems,” in CORROSION/2008, Houston, TX, 2008, Paper No.
08415.
[9] T. Tran, B. Brown, and S. Nesic, “Corrosion of Mild Steel in an Aqueous CO2
Environment–Basic Electrochemical Mechanisms Revisited,” in
CORROSION/2015, Houston, TX, 2015, Paper No. 5671.
[10] M. Nordsveen, S. Nesic, R. Nyborg, and A. Stangeland, “A Mechanistic Model for

Carbon Dioxide Corrosion of Mild Steel in the Presence of Protective Iron
Carbonate Films—Part 1: Theory and Verification,” Corrosion, vol. 59, no. 5, pp.
443–456, 2003.
[11] J. Oddo and M. Tomson, “Simplified Calculation of CaCO3 Saturation at High

temperatures and Pressures in Brine Solutions,” J. Pet. Technol., vol. 34, no. 7, pp.
1583–1590, 1982.
111
[12] D. A. Palmer and R. Van Eldik, “The chemistry of metal carbonate and carbon
dioxide complexes,” Chem. Rev., vol. 83, no. 6, pp. 651–731, 1983.
[13] O. M. Suleimenov and R. E. Krupp, “Solubility of Hydrogen Sulfide in Pure Water

and in NaCl Solutions, from 20 to 320°C and at Saturation Pressures,” Geochim.
Cosmochim. Acta, vol. 58, no. 11, pp. 2433–2444, 1994.
[14] O. M. Suleimenov and T. . Seward, “A Spectrophotometric study of hydrogen

sulphide ionisation in aqueous solutions to 350°C,” Geochim. Cosmochim. Acta, vol.
61, no. 24, pp. 5187–5198, 1997.
[15] Y. Kharaka, W. Gunter, P. Aggarwal, E. Perkins, and J. Dedraal, “SOLMINEQ.88:

A Computer Program for Geochemical Modeling of Water-Rock Interactions,” U.S.
Geological Survey, Alberta Research Council, Menlo Park, CA, 88–4227, 1988.
[16] W. Sun, S. Nešić, and R. C. Woollam, “The effect of temperature and ionic strength
on iron carbonate (FeCO3) solubility limit,” Corros. Sci., vol. 51, no. 6, pp. 1273–
1276, Jun. 2009.
[17] L. G. Benning, R. T. Wilkin, and H. . Barnes, “Reaction pathways in the Fe–S

system below 100°C,” Chem. Geol., vol. 167, no. 1–2, pp. 25–51, Jun. 2000.
[18] W. Sun and S. Nešić, “Kinetics of Corrosion Layer Formation: Part 1—Iron
Carbonate Layers in Carbon Dioxide Corrosion,” CORROSION, vol. 64, no. 4, pp.
334–346, Apr. 2008.
[19] S. Nesic, Y. Zheng, B. Brown, and J. Ning, “Advancement in Predictive Modeling

of Mild Steel Corrosion in CO2 and H2S Containing Environments,” in
[20] W. Sun and S. Nešic, “A mechanistic model of uniform hydrogen sulfide/carbon

dioxide corrosion of mild steel,” Corrosion, vol. 65, no. 5, pp. 291–307, 2009.
[21] Y. Zheng, J. Ning, B. Brown, D. Young, and S. Nesic, “Mechanistic Study of the
Effect of Iron Sulfide Layers on Hydrogen Sulfide Corrosion of Carbon Steel,” in
[22] D. W. Shoesmith, P. Taylor, M. G. Bailey, and D. G. Owen, “The Formation of

Ferrous Monosulfide Polymorphs during the Corrosion of Iron by Aqueous
Hydrogen Sulfide at 21°C,” J. Electrochem. Soc., vol. 127, no. 5, pp. 1007–1015,
May 1980.
[23] J. Boivin and S. Oliphant, “Sulfur Corrosion Due to Oxygen Ingress,” in

112
[24] A. Sherik, “Black Powder in Gas Transmission Pipelines,” in Oil and Gas Pipelines,
R. W. Revie, Ed. John Wiley & Sons, Inc., 2015, pp. 423–436.
[25] M. G. Fontana, Corrosion Engineering, 3rd ed. New York, NY: McGraw-Hill
Education, 1986.
[26] R. M. Cornell and U. Schwertmann, The iron oxides: structure, properties,

reactions, occurrences, and uses, 2nd ed. Weinheim: Wiley-VCH, 2003.
[27] B. Lavina et al., “Structure of siderite FeCO3 to 56 GPa and hysteresis of its spin-
pairing transition,” Phys. Rev. B, vol. 82, no. 6, Aug. 2010.
[28] A. R. Lennie, S. A. T. Redfern, P. F. Schofield, and D. J. Vaughan, “Synthesis and

Rietveld crystal structure refinement of mackinawite, tetragonal FeS,” Mineral.
Mag., vol. 59, no. 397, pp. 677–684, 1995.
[29] D. Rickard and G. W. Luther, “Chemistry of Iron Sulfides,” Chem. Rev., vol. 107,
no. 2, pp. 514–562, Feb. 2007.
[30] J. Yamada, H. Kaneta, and K. Nakayama, “Analysis of Black Powder in Natural Gas
Pipelines,” in CORROSION/2011, Houston, TX, 2011, Paper No. 11088.
[31] J.-A. Bourdoiseau, M. Jeannin, R. Sabot, C. Rémazeilles, and P. Refait,

“Characterisation of mackinawite by Raman spectroscopy: Effects of crystallisation,
drying and oxidation,” Corros. Sci., vol. 50, no. 11, pp. 3247–3255, Nov. 2008.
[32] B. Craig, “Corrosion Product Analysis - A Road Map to Corrosion in Oil and Gas
Production,” Mater. Perform., vol. 41, no. 8, pp. 56–58, 2002.
[33] Y. Tamaura, K. Ito, and T. Katsura, “Transformation of γ-FeO(OH) to Fe3O4 by

adsorption of iron(II) ion on γ-FeO(OH),” J. Chem. Soc. Dalton Trans., no. 2, pp.
189–194, Jan. 1983.
[34] N. Yaakob, “Top of the Line Corrosion in CO2/H2S Environments,” Ph.D.

dissertation, Dept. Chem. Eng., Ohio Univ., Athens, OH, 2015.
[35] S. Olsen and A. Dugstad, “Corrosion Under Dewing Conditions,” in

CORROSION/1991, Cincinnati, OH, 1991, Paper No. 472.
[36] Z. Zhang, D. Hinkson, M. Singer, H. Wang, and S. Nešić, “A mechanistic model of

top-of-the-line corrosion,” Corrosion, vol. 63, no. 11, pp. 1051–1062, 2007.
113
[37] N. Yaakob, F. Farelas, M. Singer, S. Nesic, and D. Young, “Localized Top of the
Line Corrosion in Marginally Sour Environments,” in CORROSION/2016,
Houston, TX, 2016, Paper No. 7695.
[38] E. Schindelholz and R. G. Kelly, “Wetting phenomena and time of wetness in

atmospheric corrosion: a review,” Corros. Rev., vol. 30, no. 5–6, Jan. 2012.
[39] L. Ojha et al., “Spectral evidence for hydrated salts in recurring slope lineae on
Mars,” Nat. Geosci., vol. 8, no. 11, pp. 829–832, Sep. 2015.
[40] E. Schindelholz, B. E. Risteen, and R. G. Kelly, “Effect of Relative Humidity on

Corrosion of Steel under Sea Salt Aerosol Proxies I. NaCl,” J. Electrochem. Soc.,
vol. 161, no. 10, pp. C450–C459, 2014.
[41] J. Kolts, “Design for Internal Corrosion Resistance of Sales Gas Pipelines,”
presented at EUROCORR, Nice, France, 2004.
[42] W. Litke, J. Bojes, P. Blais, J. Lerbscher, and W. Wamburi, “Aggressive Corrosion

Associated with Salt Deposits in Low Water Content Sour Gas Pipelines,” in
[43] “Standard Practice for Preparing, Cleaning and Evaluating Corrosion Test
Specimens,” ASTM Standard G1-03, 2011.
[44] R. E. Bedworth and N. B. Pilling, “The Oxidation of Metals at High Temperatures,”

J. Inst. Met., vol. 29, no. 3, pp. 529–582, 1923.
[45] N. Birks, G. H. Meier, and F. S. Pettit, Introduction to the High Temperature

Oxidation of Metals, Second Edition, 2nd ed. Cambridge University Press, 2006.
[46] K. P. Lillerud and P. Kofstad, “On High Temperature Oxidation of Chromium I .

Oxidation of Annealed, Thermally Etched Chromium at 800°–1100°C,” J.
Electrochem. Soc., vol. 127, no. 11, pp. 2397–2410, Nov. 1980.
[47] P. Kofstad and K. P. Lillerud, “On High Temperature Oxidation of Chromium II .

Properties of and the Oxidation Mechanism of Chromium,” J. Electrochem. Soc.,
vol. 127, no. 11, pp. 2410–2419, Nov. 1980.
[48] F. A. Golightly, F. H. Stott, and G. C. Wood, “The Relationship Between Oxide

Grain Morphology and Growth Mechanisms for Fe-Cr-Al and Fe-Cr-Al-Y Alloys,”
J. Electrochem. Soc., vol. 126, no. 6, pp. 1035–1042, Jun. 1979.
[49] R. B. Bird, W. E. Stewart, and E. N. Lightfoot, Transport phenomena, Rev. 2. ed.

New York: Wiley, 2007.
114
[50] RRUFF, “Hematite X050102.” [Online]. Available: https://rruff-

2.geo.arizona.edu/m/rruff_id/desc/X050102. [Accessed: 26-Nov-2016].
[51] M. E. Wise, S. T. Martin, L. M. Russell, and P. R. Buseck, “Water Uptake by NaCl

Particles Prior to Deliquescence and the Phase Rule,” Aerosol Sci. Technol., vol. 42,
no. 4, pp. 281–294, Mar. 2008.
[52] M. C. Foster and G. E. Ewing, “Adsorption of water on the NaCl(001) surface. II.
An infrared study at ambient temperatures,” J. Chem. Phys., vol. 112, no. 15, pp.
6817–6826, 2000.
[53] S. Lee and R. W. Staehle, “Adsorption of Water on Copper, Nickel, and Iron,”
CORROSION, vol. 53, no. 1, pp. 33–42, Jan. 1997.
[54] E. Schindelholz, L. Tsui, and R. G. Kelly, “Hygroscopic Particle Behavior Studied

by Interdigitated Array Microelectrode Impedance Sensors,” J. Phys. Chem. A, vol.
118, no. 1, pp. 167–177, Jan. 2014.
[55] G. M. Richardson and R. S. Malthus, “Salts for static control of humidity at

relatively low levels,” J. Appl. Chem., vol. 5, no. 10, pp. 557–567, Oct. 1955.
[56] S. Turgoose, “Post-Excavation Changes in Iron Antiquities,” Stud. Conserv., vol.

27, no. 3, pp. 97–101, 1982.
[57] L. S. Selwyn, P. I. Sirois, and V. Argyropoulos, “The corrosion of excavated

archaeological iron with details on weeping and akaganéite,” Stud. Conserv., vol.
44, no. 4, pp. 217–232, Jan. 1999.
[58] A. Dugstad and H. Sirnes, “Experimental Study of Black Powder in field TEG
solutions,” in CORROSION/2011, Houston, TX, 2011, Paper No. 11083.
[59] A. M. Sherik, A. Rasheed, and A. Jabran, “Effect Of Sales Gas Pipelines Black
Powder Sludge on the Corrosion of Carbon Steel,” in CORROSION/2011, Houston,
TX, 2011, Paper No. 11087.
[60] E. O. Lebigot, Uncertainties: a Python package for calculations with uncertainties.

Available: https://pythonhosted.org/uncertainties/
115
APPENDIX A: PELTIER TEMPERATURE CONTROLLER DESIGN
Specimen Temperature Controller Design
Dewing corrosion specimen temperature was performed with a specially designed
control system. The system was built around a TC-48-20 PID (Preportional, Integral,
Derivative) Peltier controller by TE Technology®. Key features of the controller include
heating and cooling control modes, 2 thermistor inputs, a pulse-width modulated output, 2
temperature alarms, and a computer interface through serial RS-232 communications. The
two alarms included in the controller monitor their respective thermistor input and activate
if the temperature exceeds a user-defined value. If an alarm is triggered, then a transistor
activates and allows electrical current to pass to ground completing a circuit. The alarms
were designed for controlling fans or other loads. For instance, if a high temperature is
measured on the Peltier heatsink by the secondary thermistor, then a fan on the heat sink
will be activated allowing for better heat transfer.
A general schematic of the Peltier controller used for experimentation is shown in
Figure 40. Specimen temperature is measured with the primary thermistor (10 kΩ at 25°C)
held in contact with the specimen. The thermistor bulb is mounted through a hole in the
PEEK specimen holder and held in place with hot glue. The hot glue acts as both support
for the thermistor and as thermal insulation isolating the sensor from external influences.
The PID controller computes the pulse-width modulation (PWM) duty cycle required to
maintain the temperature setpoint which modulates the 12 VDC power supply voltage to
supply the required powered to the Peltier.

116
Figure 40: Peltier controller schematic.
A limitation of the TC-48-20 Peltier controller is that the positive and negative
Peltier terminals are fixed preventing the polarity change required for temperature cycling
experiments. To overcome this limitation, a double-pole-double-throw (DPDT) relay was
installed between the controller and the Peltier in the configuration shown in Figure 40,
above. Break-before-make contacts are required in the relay to prevent short circuits. A
diode was placed in a reverse bias orientation to minimize the effects of voltage spikes due
to a collapsing magnetic field in the relay’s inductor. The relay is activated by triggering
the secondary alarm which allows current to pass through the relay inductor thereby
activating the relay. Since measurements are required to activate the alarm, the secondary
117
thermistor was replaced with a 10kΩ resistor to measure a constant 25°C secondary
temperature. By changing the secondary alarm temperature to above or below 25°C, then
the alarm can be deactivated or activated, respectively.
Since the Peltier controller was to be used in an H2S laboratory, the controller was
housed in a sealed steel box which isolates the controller and potential sparks within the
box from a potentially flammable environment. Panel mount connectors were installed in
the box door to provide a connection from the controller to the Peltier, primary thermistor,
and the computer kept outside the H2S laboratory. The serial RS-232 connection and
software included with the controller allowed for specimen temperature monitoring and
temperature cycling without the need to enter the H2S laboratory.
A secondary Peltier controller was developed based upon an Arduino Uno
microcontroller. As the microcontroller cannot supply the electrical power to the Peltier, a
MOSFET H-bridge motor driver (Pololu High-Power Motor Driver 18-25) was installed
between the Peltier and the power supply for power modulation and Peltier polarity control.
The Arduino controller acts in a similar manner as the TC-48-20 controller in that a PID
algorithm computes the duty cycle for the PWM output. The AC output is passed to the
MOSFET gates in the motor driver along with a binary signal which determines the polarity
of the Peltier. Controller behavior was programmed in C/C++.
XRD Specimen Holder Design
Specimen size limitations in the X-ray diffractometer prevented the use of the large
cylindrical specimens the dewing corrosion apparatus was designed around, so a new
holder was developed to hold smaller specimens within a body close to the same size as
118
the cylindrical specimens. A cross-sectional view of the apparatus is shown in Figure 41.
The specimen holder holds a 0.5×0.5×0.08 inch³ (12.7×12.7×2 mm³) specimen within a
square cutout in the center of an aluminum body with a 1.25-inch-diameter × 0.625-inch-
thick (31.7 mm × 15.9 mm) cylindrical specimen. Aluminum was chosen for its thermal
conductivity and corrosion resistance. The specimen was coated with Xylan to prevent
galvanic effects between the aluminum and the steel. A magnet within a stainless steel part
embedded within the aluminum holds the specimen to the holder. Threads in the aluminum
body and the stainless steel part allow for specimen height adjustment. Lubrication was
applied to the threads to prevent galling. A hole slightly offset from center was drilled to
place a thermistor near the steel specimen for temperature control. The aluminum body
was fitted into a PEEK sleeve designed to provide a seal with the stainless steel glass cell
lid. O-rings were embedded within the body and PEEK sleeve to prevent gas leakage.
Figure 41: Cross-section view of the XRD specimen holder. Image courtesy of Cody
Shafer.
119
APPENDIX B: CALCULATIONS
Three primary specimen mass measurements were taken for each specimen: (1)
initial mass, 푚 (g); (2) post-extraction mass, 푚 (g); and (3) mass after corrosion product
removal, 푚푓 (g). The procedures for calculating the general corrosion rate, measured
specimen mass, and theoretical specimen mass from those measurements are detailed in
the following sections.
Corrosion Rate
The corrosion rate is calculated with Equation 53:
푚 푚푓
CR 365 (53)
푡
휌st l 퐴s
where:
CR = Corrosion rate (mm/yr)
푡 = Experiment duration (days)
휌st l = Steel density (kg/m3)
퐴s = Wetted steel surface area (m2)
Measured Corrosion Product Mass
The measured corrosion product mass is the mass of corrosion product removed
per unit area. The measured corrosion product mass, in g/m2, is calculated with Equation
54:
푚 푚푓
푚m as (54)
퐴푠
120
Theoretical Corrosion Product Mass
The theoretical corrosion product mass is less straightforward to calculate than the
measured corrosion product mass. The theoretical corrosion product mass is the mass of
corrosion product that would be produced based on the mass of steel lost due to corrosion.
The X65 steel includes alloying elements (i.e., C, Mn, Cu, Ni, etc.), so the measured
corrosion product mass should account for these species. The steel components are grouped
into three categories: iron, carbon, and other (which includes all other alloying elements).
The mass lost due to corrosion in these three categories is calculated by multiplying the
total mass lost due to corrosion by the mass fraction (푥) of that category:
∆푚 푚 푚푓 (55)
푚 푥 ∆푚 (56)
푚 푥∆푚 (57)
푚ot h r 푥ot h r ∆푚 (58)
Carbon within the steel is assumed to all be present as iron carbide (Fe3C).
Realistically, some carbon will be interstitially intercalated through the ferrite, however
this mass is assumed to be small and therefore negligible. The mass of iron carbide is
therefore determined by Equation 59. The mass of iron available for iron carbonate, iron
sulfide, or iron oxide production is thereby the difference in the total mass of Fe and the
mass of Fe within the iron carbide (Equation 60). 푀 is the molar mass of species 푖
.
푚푀
푚 (59)
푀
푚 푀
푚 , P 푚 3 (60)
푀
121
The dominant corrosion product is ultimately dependent upon the conditions in
which the corrosion product is developed. The corrosion product with identity 푗is assumed
to be Fe2O3 in oxic environments, FeS in sour conditions, and FeCO3 in sweet conditions.
The mass of the iron corrosion product with 푛moles of Fe per mole 푗produced is calculated
as follows:
푚 , P푀
푚 (61)
푛푀
The total theoretical mass of corrosion product per unit area is calculated by
summing the individual corrosion product components and dividing by the specimen area
as in Equation 62.
푚 푚 푚ot h r
푚t h or (62)
퐴s
Maximum Possible Black Powder Production Rate
The theoretical mass of corrosion product is extrapolated over a straight, 42 in.
(1.07 m) ID, 100 km pipeline with area 퐴pip for the experimental duration to determine
the maximum possible black powder production rate, 푚̇ BP.
푚t h or 퐴pip
푚̇ BP 10− (63)
푡
Error Analysis
Corrosion rate and corrosion product mass measurements are highly dependent upon
gravimetric measurements. Mass measurements were all conducted on a Mettler Toledo
AB204-S, which has a readability and repeatability (standard deviation) of 0.1 mg. The
uncertainty in the mass measurement was propagated throughout the corrosion rate and
corrosion product mass measurements with the assistance of the Python package
122
Uncertainties [60]. The Uncertainties package uses linear error propagation theory to
calculate the standard deviation of a mathematical expression with inputs that include a
probability distribution function. The calculated results were reported using a 95%
confidence interval.
123
APPENDIX C: ADDITIONAL MICROSCOPY
Many more SEM images were taken than those shown in Chapter 5. The images
selected for viewing in the main body of the text were selected based on how representative
the micrograph is of the specimen at large and their content in relation to understanding
black powder formation. Addition SEM surface microscopy is shown in this appendix for
context. The micrographs are grouped by experiment condition and, unless specified
otherwise, were taken after the specimen was extracted.
Dewing Cor r osion
0 mbar H2S – High WCR
Temperature: 15°C, Gas Temperature: 30°C, Duration: 3 days.

124
0 mbar H2S – Low WCR
(a) (b)
(c) (d)
(e) (f)
Temperature: 25°C, Gas Temperature: 30°C, Duration: 3 days. Images (a) and (b) were
from the same test as that shown in Chapter 5, whereas images (c)-(f) were from a test
125
with inconclusive gravimetric measurements. The corrosion product in the inconclusive
test was of a similar morphology to the CO2/O2 test.
0.1 mbar H2S – High WCR
0.1 mbar H2S – Low WCR

126
0.3 mbar H2S – Low WCR – 6 hours
Temperature: 25°C, Gas Temperature: 30°C, Duration: 6 hours.

127
0.3 mbar H2S – Low WCR – 1 day
Temperature: 25°C, Gas Temperature: 30°C, Duration: 1 day.

128
0.3 mbar H2S – Low WCR – 3 days

129

130

131

132
1.0 mbar H2S – Low WCR

133
10 mbar O2 – Low WCR – 3 days
Temperature: 25°C, Gas Temperature: 30°C, Duration: 6 hours.

134
0.3 mbar H2S – Temperature Cycling – 3 days
135
0.3 mbar H2S – Temperature Cycling– 7 days
136
Hygr oscopic Cor r osion
0 mbar H2S – 75% RH – 3 days
(a) (b)
(c) (d)
(e) (f)
Duration: 3 days, Salt Layer Composition: NaCl. (d) and (e): after DI water rinse, (f):
after Clarke.
137
(a) (b)
(c)
after DI water rinse, and (c): after Clarke.

138
after DI water rinse, and (c): after Clarke.

139
0.1 mbar H2S – 75% RH – 3 days
Duration: 3 days, Salt Layer Composition: NaCl
Duration: 3 days, Salt Layer Composition: NaCl.

140

141

142

143
(a) (b)
(c) (d)
Duration: 3 days, Salt Layer Composition: NaCl. (c): after DI water rinse, (d): after
Clarke.
144
(a) (b)
(c) (d)
rinse.
145
(a) (b)
(c) (d)
rinse.
Thesis and Dissertation Services

Materials On Black (Powder

Uploaded by

Copyright:

Available Formats

Materials On Black (Powder

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

Materials On Black (Powder

Uploaded by

Copyright:

Available Formats

Formation of Black Powder Components by Dewing and Hygroscopic Corrosion

the Russ College of Engineering and Technology of Ohio University

of the requirement for the degree

© 2017 Martin L. Colahan. All Rights Reserved

This thesis titled

Formation of Black Powder Components by Dewing and Hygroscopic Corrosion

has been approved for

the Department of Chemical and Biomolecular Engineering

and the Russ College of Engineering and Technology by

Research Associate Professor of Chemical and Biomolecular Engineering

Dean, Russ College of Engineering and Technology

COLAHAN, MARTIN L., M.S., April 2017, Chemical Engineering

Formation of Black Powder Components by Dewing and Hygroscopic Corrosion Processes

Director of Thesis: David Young

erosion, valve failure, instrumentation malfunction, and increased pressure drop.

conditions consisted of siderite, mackinawite, and hematite. Buckle-driven delamination

flakey corrosion products were only found if deliquescence occurred.

In memory of my beloved aunt:

support none of this work would have been possible.

engineering as a child. Thank you so very much.

List of Tables ...................................................................................................................... 9

List of Figures ................................................................................................................... 10

Chapter 2: Background and Literature Review ............................................................ 20

2.1 Sales Gas ............................................................................................................ 20

2.2 Corrosion of Steel in Oil and Gas Environments ............................................... 22

2.2.1 Water Chemistry ......................................................................................... 23

2.2.2 Formation of Corrosion Product Layers ..................................................... 25

2.2.3 Effect of O2 on CO2 and H2S Corrosion ..................................................... 27

2.2.4 Crystal Structures of Corrosion Product Species........................................ 28

2.3 Black Powder ..................................................................................................... 29

2.4 Water Availability.............................................................................................. 31

2.4.1 Top-of-the-Line Corrosion.......................................................................... 33

2.4.2 Hygroscopic Uptake of Water by Salts....................................................... 35

2.4.3 Hygroscopic Corrosion ............................................................................... 36

2.5 Gaps in Understanding Black Powder Formation.............................................. 40

Chapter 3: Research Objectives and Hypotheses ......................................................... 41

3.1 Objective ............................................................................................................ 41

3.2 Hypotheses ......................................................................................................... 41

3.3 Motivation .......................................................................................................... 42

Chapter 4: Methodology and Experimental Design ..................................................... 43

4.1 Overview ............................................................................................................ 43

4.2 Test Conditions .................................................................................................. 43

4.3 Black Powder Formation in Dewing Conditions ............................................... 46

4.3.1 Dewing Corrosion Equipment .................................................................... 46

4.3.2 Dewing Corrosion Experimental Procedures.............................................. 51

4.4 Black Powder Formation in Hygroscopic Conditions........................................ 52

4.4.1 Hygroscopic Corrosion Equipment and Procedures ................................... 52

4.5 Analytical Methods ............................................................................................ 55

Chapter 5: Results and Discussion ............................................................................... 58

5.1 Dewing Corrosion .............................................................................................. 58

5.1.1 Effect of H2S Partial Pressure on Dewing Corrosion ................................. 58

5.1.2 Effect of Time on Sour Dewing Corrosion................................................. 66

5.1.3 FeS Layer Buckling .................................................................................... 70

5.1.4 Effect of O2 on Sweet Dewing Corrosion................................................... 72

5.1.5 Effect of Cyclic Water Condensation ......................................................... 77

5.1.6 Black Powder Formation in Dewing Conditions ........................................ 82

5.2 Hygroscopic Corrosion ...................................................................................... 84

5.2.1 Salt Layer Generation ................................................................................. 84

5.2.2 Hygroscopic Corrosion in Sweet Environments......................................... 86