UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sulfide stress cracking resistance of API-X100 high strength low alloy steel in H2S environments Almansour, Mansour A. 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2008_spring_almansour_mansour.pdf [ 4.81MB ]
Metadata
JSON: 24-1.0066205.json
JSON-LD: 24-1.0066205-ld.json
RDF/XML (Pretty): 24-1.0066205-rdf.xml
RDF/JSON: 24-1.0066205-rdf.json
Turtle: 24-1.0066205-turtle.txt
N-Triples: 24-1.0066205-rdf-ntriples.txt
Original Record: 24-1.0066205-source.json
Full Text
24-1.0066205-fulltext.txt
Citation
24-1.0066205.ris

Full Text

SULFIDE STRESS CRACKING RESISTANCE OF API-X100 HIGH STRENGTH LOW ALLOY STEEL IN H2S ENVIRONMENTS By MANSOUR A. ALMANSOUR B.Sc., King Fand University of Petroleum and Minerals, 2000  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIRMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE In THE FACULTY OF GRADUATE STUDIES (Materials Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA November 2007 © Mansour A. Almansour, 2007  ABSTRACT:  Sulfide Stress Cracking (SSC) resistance of the newly developed API-X100 High Strength Low Alloy (HSLA) steel was investigated in the NACE TM0177 "A" solution. The NACE TM0177 "A" solution is a hydrogen sulfide (H2S) saturated solution containing 5.0 wt.% sodium chloride (NaC1) and 0.5 wt.% acetic acid (CH 3 COOH). The aim of this thesis was to study the effect of microstructure, non-metallic inclusions and alloying elements of the X100 on H 2 S corrosion and SSC susceptibility. The study was conducted by means of electrochemical polarization techniques and constant load (proof ring) testing. Microstructural analysis and electrochemical polarization results for X100 were compared with those for X80, an older generation HSLA steel. Uniaxial constant load SSC testing was conducted using X100 samples and the results were compared with those reported for older generation HSLA steels.  Addition of H2S to the NACE TM0177 "A" solution increased the corrosion rate of X100 from 51.6 to 96.7 mpy. The effect of H2S on the corrosion rate was similar for X80. The corrosion rate for X80 increased from 45.2 to 80.2 mpy when H 2 S was added to the test solution. Addition of H2S enhanced the anodic kinetics by forming a catalyst (FeHS ads, 1 -  on the metal surface and as a result, shifted the anodic polarization curve to more current densities. Moreover, the cathodic half cell potential increased due to the decrease in pH, from 2.9 to 2.7, which shifted the cathodic polarization curve to more current densities.  i  The increase in both the anodic and cathodic currents, after H2S addition, caused the rise in the corrosion current density.  In H2S saturated NACE TM-0177 "A" solution, the X100 steel corrosion rate was higher than the X80 steel by 20%. Longer phase boundaries and larger nonmetallic inclusions in the X100 microstructure generated more areas with dissimilar corrosion potentials and therefore, a stronger driving force for corrosion. Higher density of second phase regions and larger nonmetallic inclusions acted as an increased cathode area on the X100 surface which increased the cathodic current density and consequently, increased the corrosion current density.  Proof ring tests on the X100 gave a threshold stress value,  C5th,  of 46% YS, 343.1 MPa  (49.7 ksi). The main failure was caused by SSC cracking. SSC nucleated at corrosion pits on the metal surface and microcracks in the metal body and propagated perpendicular to the applied stress. Hydrogen Induced Cracking (HIC) was observed in the X100. HIC cracks nucleated at banded martensite-ferrite interfaces and propagated along the rolling direction parallel to the applied tensile stress through the softer ferrite phase.  When compared to older HSLA grades, the X100 tested in this study had a high SSC susceptibility and therefore, is not be recommended for H2S service applications. The high X100 SSC susceptibility was caused by the material high corrosion rates in H 2 S media which formed corrosion pits that acted as crack initiation sites on the metal surface and provided more hydrogen that migrated into the steel. In addition, the X100  ii  inhomogeneous microstructure provided a high density of hydrogen traps in front of the main crack tip which promoted SSC microcrack formation inside the metal. Microcracks in the metal body connected with the main crack tip that originated from corrosion pits which assisted SSC propagation.  iii  TABLE OF CONTENTS: ABSTRACT ^  i  TABLE OF CONTENTS ^  iv  LIST OF TABLES ^  vi  LIST OF FIGURES ^  vii  ACKNOWLEDGMENT ^  x  1 INTRODUCTION^  1  2 LITERATURE REVIEW ^  6  2.1 HSLA steels ^ 2.1.1 Definition of HSLA steels ^ 2.1.2 Processing of HSLA steels ^  6 6 7  2.2 Electrochemical background ^ 2.2.1 Corrosion in H2S environments ^ 2.2.2 Electrochemical measurements ^ 2.2.2.1 Pourbaix diagrams ^ 2.2.2.2 Polarization diagrams ^  12 12 17 17 19  2.3 Sulfide Stress Cracking (SSC) ^ 2.3.1 SSC definition ^ 2.3.2 SSC process ^ 2.3.3 Stress intensity at the SSC crack tip ^ 2.3.4 Testing for SSC susceptibility ^ 2.3.5 Factors influencing SSC in HSLA steels ^ 2.3.5.1 Effect of material strength ^ 2.3.5.2 Effect of microstructure ^ 2.3.5.3 Effect of material homogeneity ^  21 21 25 26 31 33 33 34 38  3 EXPERIMENTAL PROCEDURES ^  40  3.1 Tested materials ^  40  3.2 Microstructure analysis ^  41  3.3 Electrochemical testing ^  42  iv  3.3.1 Specimen preparation ^ 3.3.2 Test solution ^ 3.3.3 Cell setup ^ 3.3.4 Electrochemical techniques ^ 3.3.4.1 Open circuit potential measurements ^ 3.3.4.2 Potentiodynamic polarization tests ^  42 43 43 45 45 45  3.4 SSC constant load tests ^ 3.4.1 Proof ring devices ^ 3.4.2 Test solution ^ 3.4.3 Test procedure ^ 3.4.4 Failure detection ^ 3.4.5 Failure analysis ^  46 46 50 50 51 51  4 RESULTS AND DISCUSSION ^  52  4.1 Metallurgical examination ^ 4.1.1 Microstructure and strength ^ 4.1.2 Nonmetallic inclusions ^  52 52 57  4.2 Electrochemical behaviour ^ 4.2.1 Corrosion potential ^ 4.2.2 Anodic polarization behaviour ^ 4.2.2.1 X100 ^ 4.2.2.2 X80 ^ 4.2.3 Surface characterization ^ 4.2.4 Mechanistic insights ^ 4.2.4.1 Effect of H2S on corrosion rates ^ 4.2.4.2 Comparison between X100 and X80 ^  62 62 64 64 66 68 73 74 78  4.3 Proof ring testing^ 83 4.3.1 Threshold stress determination for API-X100 ^ 83 4.3.2 Crack characterization ^ 85 4.3.3 Mechanistic insights ^ 89 4.3.4 Comparison between X100 and other HSLA steels ^ 95 4.4 Implications of results on X100 susceptibility ^  97  5 SUMMARY AND CONCLUSION ^  99  6 REFERENCES ^  104  v  LIST OF TABLES Table 2.1 Time to failure results (hours) for as received and quenched-tempered X60 ^  39  Table 3.1 Chemical analysis with carbon equivalence (C.E.) for the X80 and X100 steels ^  40  Table 3.2 X100 proof ring specimens and applied stresses ^ 49 Table 4.1 Average yield and tensile strength values (engineering and true) for X100 and X80 ^  56  Table 4.2 Average hardness values and standard deviation in Rockwell C for X80 and X100 ^  56  Table 4.3 Summary of electrochemical test results conducted on X100 and X80 in H2S-free and H2S-saturated NACE TM0177 "A" solution ^ 73 Table 4.4 Proof ring test results tested in NACE TM-0177 "A" (23 °C) ^ 83  vi  LIST OF FIGURES Fig. 1.1 History of the development of HSLA linepipe steels ^ 2 Fig. 2.1 Temperature-time profile for rolling of a microalloyed steel plate ^ 8 Fig. 2.2 Different approaches to produce API-X100 steel ^ 10 Fig. 2.3 Hydrogen migration into the steel ^  13  Fig. 2.4 Effect of added oxidizer. An increase in pH (pH line 2 < PH line 1) will result in an increase in E con. and i coff ^  15  Fig. 2.5 E-pH diagram showing the thermodynamic stability regions for surface sulfides ^  18  Fig. 2.6 Pourbaix diagram (E-pH) for the iron-water system ^ 19 Fig. 2.7 Schematic polarization diagram for iron dissolution (anode) and hydrogen evolution (cathode) ^  20  Fig. 2.8 Polarization test set up ^  21  Fig. 2.9 Hydrogen accumulation and crack initiation at SSC crack tip ^  22  Fig. 2.10 SSC and HIC cracks ^  24  Fig. 2.11 Stresses acting in front of a crack (mode I) ^  27  Fig. 2.12 Crack tip stress distribution ^  28  Fig. 2.13 Center crack with uniform loading ^  29  Fig. 2.14 Diagram of uniaxial tensile testing (constant load) ^ 32 Fig. 2.15 Elongation (in NACE TM-0177 "A" solution) / Elongation (in air) for API-X70 with different microstructures ^  35  Fig 3.1 Specimens used for microstructural analysis (left) and polarization analysis (right) ^  42  vii  Fig. 3.2 Electrochemical polarization experimental setup ^ 44 Fig. 3.3 Proof ring testing device ^  47  Fig. 3.4 X100 tensile specimen ^  48  Fig. 4.1 X80 optical imaging showing granular bainite (dark) and ferrite (light) microstructure ^  53  Fig. 4.2 X100 optical imaging showing martensite (dark), bainite and ferrite (light) microstructure ^  53  Fig. 4.3 Engineering stress-stain curve for in air for (a) X100 and (b) X80 ^ 55 Fig. 4.4 Spherical non-metallic inclusions in (a) X80 and (b) X100 (SEM) ^  57  Fig. 4.5 EDX spectra for (a) X80 and (b) X100 ^  58  Fig 4.6 Stringers found in the X100 steel, arrow indicating rolling direction (SEM) ^  59  Fig. 4.7 Initiated cracks at the interface between an inclusion and the steel matrix in X100 specimen, arrow indicating rolling direction (SEM) ^ 61 Fig. 4.8 E ocp for X100 versus SCE in (H2S-saturated and H2S-free) NACE TM0177 "A" solution ^  63  Fig. 4.9 E ocp for X80 versus SCE in (H2S-saturated and H2S-free) NACE TM0177 "A" solution ^  64  Fig. 4.10 X100 polarization in H2S-saturated NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH) ^  65  Fig. 4.11 X100 polarization in H2S-free NACE TM-0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH) ^  66  Fig. 4.12 X80 polarization in H2S-saturated NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH) ^  67  Fig. 4.13 X80 polarization in H2S-free NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH) ^  68  Fig. 4.14 Corrosion product EDX for X100 in (a) H2S-free, (b) H2S-saturated NACE TM-0177 "A" solution and (c) Polished uncorroded X100 specimen ^ 70  viii  Fig. 4.15 Corrosion product EDX for X100 in (a) H2S-free and (b) H2S-saturated NACE TM-0177 "A" solution ^ 71 Fig. 4.16 Polarization diagrams of (a) X100 and (b) X80 in NACE TM0177 "A" solution (H2S-free and H2S-saturated) ^  74  Fig. 4.17 Evans diagrams for (a) X100 and (b) X80 in NACE TM0177 "A" solution (H2S-free and H2S-saturated) ^  75  Fig. 4.18 Anodic overpotential (q) vs. current density for X100 and X80 in H2 S-free and H2S-saturated NACE TM-0177 "A" solution ^ 77 Fig. 4.19 Polarization diagram of X100 and X80 in H2S-saturated NACE TM0177 "A" solution ^  79  Fig. 4.20 Evans diagrams for X100 and X80 in NACE TM0177 "A" solution (H2S-saturated) ^  79  Fig. 4.21 Pitting in (a) X80 and (b) X100 in H2S saturated NACE solution (SEM) ^  82  Fig. 4.22 Proof ring time to failure graph for the X100 in H2S-saturated NACE solution ^  84  Fig. 4.23 Fracture face longitudinal section for 80%YS specimen ^ 86 Fig. 4.24 Longitudinal section of fracture face (80%YS specimen) showing transgranular crack propagation (light) ^  87  Fig. 4.25 Longitudinal section of the fracture face of an X100 specimen loaded at 65%YS showing a HIC crack connected to the SSC crack (light) ^ 88 Fig. 4.26 Effect of applied stress and rolling direction on HIC propagation ^ 88 Fig. 4.27 Longitudinal section of X100 specimen loaded at 65% YS showing SSC crack and surface inclusions (dark regions) (light) ^ 89 Fig. 4.28 Longitudinal section of an X100 specimen loaded at 65% YS showing main SSC crack (right of image) and small microcracks (left of image) (light) ^  90  Fig. 4.29 SSC brittle regions connected by a ductile section found in the fracture face of X100 specimen loaded at 80% YS (SEM) ^ 91 Fig. 4.30 Optical imaging of a transverse section of an X100 specimen loaded at 50% YS showing HIC cracks ^  ix  92  ACKNOWLEDGMENT I would like to thank my supervisor, Dr. Akram Alfantazi, for his guidance throughout the course of this project. I would also like to express my deep appreciation to Dr. Joseph Kish for his valued assistance. This report would have not been achievable without his contributions.  My thanks extend to CANMET labs represented by Dr. Mimoun Elboujdaini for supplying the materials required to undertake this project and for allowing me to use their H2S lab for the electrochemical polarization experiments. In addition, I would like to express my appreciation to Bodycote labs represented by Heath Walker for conducting the proof ring experiments at such a short notice and prompt delivery of the test data. I also would like to thank the examining committee members, Dr. Steve Cockcroft, Dr. Chad Sinclair and Dr. GOran Femlund for simulating valuable discussions and comments that benefited the quality of this report. My acknowledgment extends to my sponsoring company, Saudi Aramco, for the post graduate degree scholarship it granted me.  Finally, I would like to express my deepest gratitude to my wife, Jumana, and family. Their patience, support and encouragement were the most important aid throughout the period of my study.  x  Chapter 1  1 Introduction:  High Strength Low Alloy (HSLA) steels are extensively used in oil and gas transportation pipelines. HSLA steels have a low price-to-yield strength ratio, pipelines with minimum yield strength (a y) of 70 ksi (483 MPa) are readily available in the oil and gas industry. In addition to material strength, HSLA steels provide good weldability because of their low carbon contents. The approach to generate HSLA steels involves a combination of lower carbon content and fine grain size by microalloying along with thermomechanical rolling or accelerated cooling (Ref. 1). API grade HSLA steels are designated by their minimum yield strength value. For example, API-X100 steel is designed to have a yield strength of 100 ksi (690 MPa).  The developmental history of HSLA steels is shown schematically in Figure 1.1. In the seventies, hot rolling and normalizing was replaced by thermomechanical rolling. The latter process enabled materials up to X70 (a y = 70 ksi) to be produced from steels that are microalloyed with niobium and vanadium and have a reduced carbon content. An improved processing method, consisting of thermomechanical rolling plus subsequent accelerated cooling, emerged in the eighties. By this method, it has become possible to produce higher strength materials like X80 (a y = 80 ksi). The X80 has a further reduced carbon content and thereby excellent field weldability. Additions of Mo, Cu and Ni enable the strength to be raised to that of grade X100 (a y = 100 ksi), when the steel is  1  processed to plate by thermomechanical rolling plus modified accelerated cooling (Ref. 2).  TM and intensive accelerated cooling  2005 2000 1995 -  Year  1990  -  1985  -  TM treatment  1980  Hot rolled and normalized  1975 1970 1965  TM treatment and accelerated cooling  -  (  1960 X52^X60^X70^X80^X100  API grade Fig. 1.1 History of the development of HSLA linepipe steels.  The application of HSLA steels in oil and gas transportation pipelines has been increasing in the past few years. Requirements for higher strength HSLA steels have risen as a result of the increased global energy demand. Transportation pipelines with higher mechanical limits have been utilized to withstand the increasing line pressures. API-X100 is a newly developed HSLA steel that was designed to fill the gap for these new requirements. One of the important aspects to assure good performance of X100 pipeline steels is to minimize cracking susceptibility when exposed to sour environments. The major cracking phenomenon in sour pipelines is Sulfide Stress Cracking (SSC). There is a need to investigate the SSC resistance of X100, considering the higher strength and  2  hardness values it has relative to older generation HSLA steels. Investigating SSC susceptibility will open doors for future improvements on the X100 in order to generate a more cracking and corrosion resistant material in H 2 S environments.  SSC is an embrittlement phenomenon in which crack failures can occur in stresses well below the yield strength of the material. SSC is a joint process of two factors (Ref 3):  •  H2 S corrosion causing metal dissolution, pit formation and production of hydrogen atoms.  •  Migration of hydrogen atoms into the steel causing metal embrittlement.  A number of experimental procedures have been developed to assess susceptibility to SSC. Results from these tests can be used to compare SSC susceptibility of different metals or in purchase requisitions. One of the most widely used SSC assessment methods is the uniaxial constant load test. Constant load tests generate a threshold stress value (Gth), which is the maximum stress at which no failure occurs during the test duration of 720 hours (1 month) (Ref. 4).  Current work on the X100 has been concentrating on generating the optimum microstructure, welding processes, alloying elements and thermomechanical processing methods for the material (Ref 5). So far, little attention has been made to investigate the newly developed X100 in SSC environments. Some SSC work on the X100 was done back in the early eighties by Borik et al. (Ref 6). These tests were conducted to  3  investigate the possibility to produce a future X100 grade steel (at that time). It should be noted that the X100 used in these studies does not resemble current X100 steels. X100 that were tested, but not implemented, in the eighties were developed by hot forging and hot rolling followed by quenching the tempering. No thermomechanical controlled processing or micro alloying by Ti, V and Nb were used. This material was highly susceptible in sour media and generated very low threshold stress values (a th), 25% of the material yield strength. The existing X100 material is produced by controlled rolling and accelerated cooling processes coupled with controlled addition of microalloying elements such as Ti, V and Nb.  Therefore, SSC studies on the current generation X100 are needed to evaluate its susceptibility in sour environments. This includes evaluating the material's corrosion rates in H2S environments. H2S corrosion will provide hydrogen atoms that would migrate into the steel and cause embrittlement. In addition, pits generated from the corrosion process will act as initiation sites for SSC on the metal surface.  The aim of this study is to examine the SSC resistance of the current generation APIX100. This includes:  • Examining the X100 microstructure and nonmetallic inclusions to observe their effect on SSC susceptibility.  4  •  Studying the material's corrosion response in sour media by developing electrochemical polarization diagrams in H2S-saturated NACE TM-0177 "A" solutions.  •  Proof ring testing to determine the threshold stress value (o th). X100 specimens were tested at varying fractions of the material's yield strength to generate  •  CYth•  Studying the generated cracks and their nucleation/propagation processes.  An older generation HSLA steel, X80, was used for comparison in microstructural and electrochemical examinations. For proof ring testing, experiments were conducted on the X100 and results were compared with earlier work in literature.  5  Chapter 2  2 LITERATURE REVIEW  2.1 High Strength Low Allow (HSLA) Steels:  2.1.1 Definition of HSLA steels:  In recent years, microalloying of plain carbon steels with small amounts (max 0.1 wt. %) of strong carbide and nitride forming elements such as Nb, Ti and V has achieved a great improvement in their mechanical properties. The addition of small amounts of Nb, Ti and V in combination with controlled rolling and accelerated cooling has allowed production of low carbon (0.05 to 0.1 wt.%) plain carbon steels with high yield stresses (up to 100 ksi) and good toughness. These steels are known as High Strength Low Allow (HSLA) steels. The improvement in strength is a result of (Ref. 7):  1- Refinement of the ferrite grain size by the formation of a fine sub-grain structure.  2- Strain-induced precipitation of the carbides and nitrides of the strong carbide/nitride forming elements such as Nb, Ti and V.  6  Many advantages arise from the production of HSLA steels. One advantage is the price to yield strength ratio. HSLA steel is priced from the base price of carbon steels and not from the base price of alloy steels. This gives the opportunity to use thinner pipelines with HSLA steels impacting material, installation and inspection costs. In addition, the low carbon content provides low carbon equivalence values and thus, coupled with their relative thinner wall thickness, welding of HSLA steels is much easier and less expensive. For the reasons mentioned above, big pipeline projects have been depending on HSLA steels as the material of choice (Ref. 8, 9).  2.1.2 Processing of HSLA steels:  A controlled rolling and accelerated cooling process is applied to produce HSLA steels. This process will provide toughness and weldability, as well as, high strength. A Thermomechanical Controlled Process (TMCP) schedule is an elaborate sequence of stages interrupted by cooling periods and cooling after finish rolling either in air or water. Initially, slabs are reheated to temperatures in the range of 1100 to 1250 °C (2010 to 2280 °F) and, as a result, significant amount of the carbonitrides are dissolved as shown in Fig. 2.1 for a temperature-time profile for rolling of a microalloyed steel plate. As temperature decreases during controlled rolling, the carbonitrides become insoluble and precipitate out in austenite (Ref. 10).  Soaking  Roughing  r-r ..,./.. 1R  Finishing Air cooling  Time (min.) -10.-  Fig. 2.1 Temperature-time profile for rolling of a microalloyed steel plate. Adapted from.  The rolling operation generally involves two stages, high temperature rolling (or roughing) and low temperature series of deformation steps (or finishing). During roughing operations, repeated recrystallization refines the coarse reheated austenite grains. The carbonitrides reduce the size of the recrystallized grains by "pinning" the grain boundaries. When rolling starts or continues below TR, recrystallization is no longer possible, and the austenite grains becomes flat in an operation known as "pancaking". The produced ferrite has a very fine grain structure because of the large number of nucleation sites available in the pancaked austenite grains. The fine grain ferrite is a main feature of good toughness properties and high yield strengths. In addition, microalloying elements that didn't precipitate during controlled rolling precipitate in the ferrite and provide additional strength to the HSLA steel (Ref. 11).  8  The small additions of alloying elements in HSLA steels processing have a significant effect on the resulting mechanical and electrochemical properties. In HSLA steel processing, C is kept at low levels to improve weldability. High content of Ni and Mo are provided to increase hardenability. Microalloying with V and Nb is used to induce strengthening via stable carbide and carbonitride precipitations (Ref. 12). Cu additions are used to reduce hydrogen diffusion in the steel by forming a protective CuS layer on the steel surface in the corrosive media of mildly sour environments (between pH 4.0 and 7.0) and decreasing the corrosion rate at open circuit potential (Ref. 13).  Further increase in strength and toughness from older generation HSLA steels, like X70, can be attained by changing the microstructure of the steel from ferrite-pearlite to ferritebainite. This has lead to the development of X80 steel which, compared to X70, has a further reduced grain size and increased dislocation density.  After controlled rolling, accelerated cooling (15 to 20 °C/sec) with a cooling stop temperature of about 500 °C followed by air cooling generates a microstructure of ferrite with a certain amount of bainite. Transformation into bainite is promoted by increasing hardenability by additions of B or Ni and Mo to the steel. It should be noted however that these alloy additions increase the carbon equivalent value, which may affect the steel weldability (Ref 14, 15). The term carbon equivalent (C.E.) estimates the cracking susceptibility of steel after welding and determines whether the steel needs any heat treatment to avoid cracking. Carbon equivalent includes the hardeability effect of the  9  alloying elements by expressing the chemical composition of the steel as a sum of weighted alloying contents (concentration in wt.%) according to the equation below:  C.E. = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15^(2.1)  Additions of Mo, Cu and Ni enable the strength to be raised to that of grade X100 when the steel is processed to plate by controlled rolling and accelerated cooling. These alloying elements increase hardenability and strength of the steel and suppress the formation of lower strength phases (Ref. 16). Three main approaches have been developed with respect to the selection of chemical composition and cooling conditions (Ref. 17). Description of these three approaches is shown in Fig 2.2.  Approach A: High carbon, medieum C.E. and soft AcC. Approach B: Low carbon, low C.E. and strong AcC. Approach C: Medium carbon, medium C.E. and medium AcC.  A^  Cooling rate C °/Sec  A  Carbon equivalent  Fig. 2.2 Different approaches (A, B and C) to produce API-X100 steel.  10  The first approach ("A" in Fig. 2.2) involved a relatively high carbon (about 0.08%) and medium carbon equivalent (about 0.49). The processing was a reheating temperature 1140-1220 °C, finish rolling temperature 680-780 °C, and a cooling rate of 20 °C/sec. This procedure had some difficulties when it came to fulfilling the requirements for toughness to ensure crack arrest. In addition, this approach was detrimental in terms of field weldability.  The second approach ("B" in Fig. 2.2) involved a relatively low carbon (around 0.05%) and low carbon equivalent (about 0.43). The processing had high cooling rate of about 60 °C/sec. The result was an uncontrolled fraction of martensite in the microstructure, which has a detrimental effect on toughness properties.  The third approach ("C" in Fig. 2.2) was the most preferred process. Approach "C" involved a medium carbon content (about 0.06%) and medium carbon equivalent (0.48). This process had a cooling rate range of 30-50 °C/sec. This approach has shown good toughness and satisfactory field weldability (Ref. 18). Approach "C" is not strictly followed and manufacturers tend to modify it in order to reach the desired X100 properties.  11  2.2 Electrochemical background:  2.2.1 Corrosion in H2S environments:  Most of the H 2 S in downstream oil transportation pipelines is produced by Sulfate Reducing Bacteria (SRB). SRB enters the oil well with the make-up water. Make-up water is usually injected to oil well to maintain the well pressure. SRB reduce sulfate to sulfide and then converts sulfide to H 2 S by oxidizing molecular hydrogen or hydrogen released as a result of cathodic reaction during corrosion (Ref 19)  Iron corrodes in H 2 S environments according to the following reaction scheme:  Fe + H 2 S -› FeS + 2H °^(2.2) 2H ° H2  ^  (2.3)  Sulfur acts as a poison to the second reaction and thus, promotes the migration of H ° into the metal lattice (Fig. 2.3). The anodic and cathodic reactions are as follows:  Fe 4 Fe 2+ + 2e - (Anode)  ^  2H + + 2e - 4 2H ° (Cathode)  ^  (2.4) (2.5)  Different H2S concentrations produce different corrosion rates. For example, Radkevych et al. (Ref 20) reported that in a natural gas well containing an H2S concentration of 0.25  12  mol% (steel grade was not identified), a corrosion rate of 20.0 mpy (mils per year) was generated, where 1 mil = 0.001 in. Another well with higher H2S content, 16 mol %, had a more aggressive corrosion rate of 34.6-60.0 mpy. Radkevych reported that corrosion rates in H 2 S environments can reach up to 236-314 mpy (Ref. 20). Theoretically, as corrosion rates become higher, more hydrogen is available to migrate into the metal. Koh et al. (Ref 21) measured the corrosion rates for different HSLA steels with yield strengths that ranged from 431 to 539 MPa (X52 to X80). The steels were tested in the NACE TM-0177 "A" solution at 25 °C. The tested HSLA steels exhibited close corrosion rates that varied from 92.1-110.6 mpy.  Metal  FeS  as  H. II H  IT-Trap  H tI  Fig. 2.3 Hydrogen migration into the steel.  H2S can influence corrosion in three different ways: •  Increase in cathodic half cell electrode potential  •  Formation of surface catalyst  •  Films with low protective abilities and "sulfide-oxide" galvanic couple formation  13  Increase in cathodic half cell electrode potential: H2S is an acid (H2S saturated water will generate a pH of about 4.0). This means that H 2 S introduction into aqueous solutions can depress pH values leading to an increase in corrosion potentials (E con.) and corrosion current densities (icon). This effect is closely related to the solution pH and the concentration of added H2S. Increase in cathodic half cell electrode potential can be explained by the result of decreasing pH on the Nernst equation (Ref 22). For a general reaction such as:  aA + mH + + ne" = bB + dH2O^  (2.6)  The Nernst equation will be as follows:  e = e° + (2.3RT/nF) log ([Al a .[H+] m / [B] b [H 2 O] d )^(2.7)  Where e° is the standard half cell potential, R is the gas constant (8.314 J moi l 1(1 ), T is temperature (K), F is Faraday's constant (96,500 coulomb per equivalent). The quantity 2.3RT/nF is equal to 0.059 V at 25 °C and the activity of water is always defined as unity in aqueous solutions.  For the cathodic reaction of hydrogen evolution (at 25 °C):  2H+ + 2e -  (2.8)  H2^  14  "A" and H2O in equation 2.6 are absent and B is replaced by  H2  (PH2 assumed to be 1  atm). Since pH = -log (11 ± ), Equation 2.7 becomes:  eH+/H2 = eH+/H2 ° - (0.059) pH  ^  (2.9)  As a result, as the concentration of the oxidizer, }1 ± , increases (pH decreases), the potential, eH+,H2, increases. More importantly, the corrosion current density, i con-, will also increase as can be seen in Fig. 2.4.  e H+TH 2 Cathode pH decrease  e H+/H 2 as O  Ecorr2 Ecorrl e Fe /Fe 2+  Anode con' 1^icorr2  Current density  Fig. 2.4 Effect of added oxidizer on the electrochemical behaviour. An increase in pH (pH line 2 < pH line 1) will result in an increase in E con- and icorr•  15  Formation of surface catalyst: Radkevych, (Ref 20), suggested that H2S effects the electrode reactions by forming an intermediate compound, FeHS - ad„ which acts as a surface catalyst. For the anodic reaction:  Fe + H 2 S 4 FeHS - ad s + H ±^(2.10) FeHS - ads^Fel-IS+ads ± 2e^ (2.11) FeHS ± ad s + 11+ - Fe 2+ + H2 S^ (2.12)  The complex FeHS - ad s will decompose allowing H 2 S to regenerate. When the complex is formed on the metal surface, the bond of iron with sulfur decreases the metallic bonds in the metal which facilitates ionization. Films with low protective abilities and "sulfide-oxide" galvanic couple formation: For the corrosion process in gas field equipments after long-term service (> 10 years), steel parts will be covered with a thick layer corrosion products. For example, according to Radkevych et al (Ref 20), corrosion products on pipeline metal surfaces after service in H2S containing gas deposits can be as thick as 2 mm. For this reason, it is necessary to consider the influence of surface sulfides on the corrosion process.  Fig. 2.5 shows the thermodynamic stability regions for the surface sulfides, Different types of HSLA chemistry (X60/70/80/100) will produce slightly different E-pH diagrams, but the general observation will be valid for all. It is shown that those surface  16  sulfides are stable in pH > 4.5. It is important that E-pH diagrams will only show the thermodynamic stability of reaction species and not the rate of their formation (Ref 23).  In order to obtain quantitative data on the chemical composition of products of H 2 S corrosion, analytical calculations of possible products of corrosion with various types of sulfides were conducted (Ref 20). Results showed that H 2 S corrosion of iron generated a significant percentage of oxides alongside the sulfide corrosion products. The potential difference between a steel electrode covered with sulfide and another of oxidized steel electrode gave a difference of 0.1-0.4 V, which means that the iron sulfide will act as a cathode relative to oxidized iron. Radkevych (Ref 20) proposes that the product of H 2 S corrosion generates a lot of "oxide-sulfide" galvanic couples. Thus, hydrogen more intensively diffuses into the metal due to the operation of numerous galvanic couples on the steel surface that increase the corrosion rate.  2.2.2 Electrochemical measurements:  2.2.2.1 Pourbaix (Potential-pH) Diagrams: When investigating SSC on pipeline steels, API-X100 in this case, the use of Pourbaix diagrams is quite useful. Through the use of thermodynamic theory (the Nernst equation), Pourbaix diagrams can be constructed. These diagrams show the thermodynamic stability of species as a function of potential and pH. Fig. 2.6 represents a version of the Pourbaix diagram for the iron-water system at ambient temperature. This diagram determines, by  17  means of potential and pH, the tendency of a metal surface to corrode, passivate or be immune against corrosion. When investigating the API-X100 steel, Pourbaix diagrams can be used to determine fundamental information regarding its corrosion potential in H2S environments. The metal's potential can be measured against a reference electrode (e.g. SHE). With a stable reference electrode, the potential measurements can be used to monitor changes in potential that are due to changes that occur at the metal surface or in the environment during the test. It should be noted that Pourbaix diagrams only show the thermodynamic tendency of the metal (X100 in this case) to immunity, corrosion or passivation. Corrosion rates can not be obtained from Pourbaix diagrams. Corrosion rates (icorr) will be determined by conducting polarization diagrams (Ref 23).  0 -0.2 -  FeS (Passive)  2+ Fe -0.4 - (Corrosion)  2 "-0.6 -0.6 in -0.8 -  ^ FeS Fe +112S  FeS  Fe^Fe + HS (Immunity) -1.2 i 0^2^4^6^8^10^12^14 pH Fig. 2.5 E-pH diagram showing the thermodynamic stability regions for surface Sulfides (dissolved ion activity = 10 -6 g-equiv/L).  18  Fe E-pH Diagram Fe+++  11 0  Fe(OH)3 (Passivity) 1.1.1  Fe++ (Corrosion)  1.2  -IFe02Corrosion  Fe^Fe(OH)2 (immunity)^(Passivity)  im 'it -1^1^2^3 4^6^6^7^8^9 10 11 12 13 14 16 16 pH  Fig. 2.6 Pourbaix diagram (E-pH) for the iron-water system (dissolved ion activity = 10 6 g-equiv/L).  2.2.2.2 Polarization Diagrams:  As discussed in the previous section, Pourbaix diagrams show the tendency of the material to corrode but it won't describe how fast the reaction will be, in other words, the corrosion rate or i con . A useful method to determine the corrosion rate is by constructing a polarization diagram for the environment of interest. When the corrosion reaction starts, the cathodic and anodic half-cell reactions will polarize to a point of intersection, which determines the corrosion rate (i con) and the corrosion potential (E con), as shown in Fig. 2.7.  19  JO, H:r1.•12  cri  >  LL!^  Fe2'  ...1  ill i0^ a.  ----a  FefFe 2 ' — — - et \  •,tor 4.7-e  .-• ''' C' e '-'°' 7 Polarization line  Tafel slope  CURRENT DENSITY (A/cm 2 )  Fig. 2.7 Schematic polarization diagram for iron dissolution (anode) and hydrogen evolution (cathode).  The instrumentation set-up for electrochemical polarization experiments is shown in, Fig. 2.8. The potential of the working electrode (WE), X80/X100 specimens, is measured with respect to a reference electrode (REF) coupled to the polarization cell by a salt-bridge probe. The salt-bridge probe is necessary in order to measure the potential near the working electrode surface. A counter, or auxiliary (AUX), electrode is placed in the polarization cell to transmit impressed currents through the solution, either to or from the working electrode surface.  The potentiostat is the electronic instrument that carries out the electrochemical polarization experiment. It comprises a power supply for impressed current on the 20  electrochemical cell and circuits that measure and control the potential to set values. A high-impedance voltmeter measures the potential between the working electrode and the reference electrode without affecting the potential of the working electrode. A lowresistance ammeter measures the current flow between the counter and working electrode without affecting current flow (Ref 24).  Reference electrode Auxiliary electrode  ^  Working electrode  Fig. 2.8 Polarization test setup.  2.3 Sulfide Stress Cracking (SSC):  2.3.1 SSC definition:  SSC is a cracking phenomenon caused by the joint effect of H 2 S corrosion and hydrogen embrittlement. During exposure of pipeline steels to sour environments, H2S corrosion generates on the steel causing metal dissolution and corrosion pit formation. As part of  21  the corrosion process, hydrogen cathodically evolves on the metal surface and migrate into the steel leading to metal embrittlement.  The generated corrosion pits are usually the initiation site for the main SSC crack at the metal surface. The main crack propagates by connecting with brittle microcracks in the metal body causing crack propagation and pipeline failure at applied stresses well below the material's yield strength. In the absence of hydrogen charging, the material can hold stresses up to the material's yield strength before plastic deformation. In the event of hydrogen charging, a threshold stress value is present above which SSC will propagate. The threshold stress (6th) is a fraction of the yield strength value and its value decreases with increasing susceptibility (Ref. 25), Fig. 2.9. Yield strength  Yield strength  Threshold stress  ^ Applied Applied ^ stresses stresses 11==.  P  A str ess  HH  '  ^  H H H  I stress  Without hydrogen ^With hydrogen charging^ charging  Fig. 2.9 Hydrogen accumulation and crack initiation at SSC crack tip.  22  Many factors control SSC susceptibility. One of the main aspects is corrosion behaviour because this process determines the intensity of hydrogen that migrates into the steel. In addition, corrosion produces corrosion pits that can act as the initial site for cracking. Another important factor is mechanical properties such as hardness and yield strength values. HSLA microstructure and homogeneity are also essential to determine SSC susceptibility.  A number of candidate mechanisms for hydrogen embrittlement have been proposed (Ref. 26). One reasonable aspect of this controversy is that there are several viable mechanisms of hydrogen related failure and that the search for a single mechanism to explain all the observations may not be successful. It is generally accepted that hydrogen migration into susceptible steels causes embrittlement regardless of the exact mechanism. These theories are:  •  Decohesion theory, according to which interatomic bonds in the crystal lattice weaken due to the diffusion of hydrogen into triaxial stressed regions.  •  Enhanced localized plasticity by increasing the mobility of dislocations in hydrogen enriched regions.  •  Pressure theory caused by the combination of hydrogen atoms to form molecular hydrogen inside in steel leading to an increase in internal pressures and void formation.  23  Another form of hydrogen damage that might form in HSLA steels, in addition to SSC, is hydrogen induced cracking (HIC). HIC is generated by hydrogen migration into voids or inclusions followed by combination to form hydrogen gas. The pressure caused by the generated hydrogen gas forms blisters inside the metal that propagate parallel to the rolling direction. In the presence of residual or applied stresses, connection of HIC cracks might occur in a stepwise manner (through pipe thickness) (Ref. 26), Fig. 2.10. It should be noted that in Fig. 2.10, the applied stresses are parallel to the rolling direction. As a result, the generated HIC cracks that propagate along the rolling direction are independent of the applied stresses. In the case where the applied stresses are perpendicular to the rolling direction, HIC cracks are affected by the applied stress and propagate as a result. Corrosion pit ^MC crack SSC crack  Microcracks Rolling direction ■••■1111110.  41  /11111^Applied stress ■=11111110.  Fig. 2.10 SSC and HIC cracks.  24  2.3.2 SSC process:  One of the principle factors that determine the hydrogen damage susceptibility of ferrous alloys is a phenomenon referred to as "trapping". Diffusion studies of iron and steels has shown an initial retardation in diffusion rate or lag time for hydrogen diffusion through these alloys before a steady state diffusivity compatible with that expected theoretically is achieved. This lag time is generally considered to be related to the filling of traps by hydrogen. Trapping will facilitate nucleation of SSC microcracks in the metal body, these microcracks can connect to the main crack tip thereby assisting SSC propagation (Ref. 27).  After diffusing into the metal, hydrogen collects at metal-inclusion (mainly sulfides and oxides) voids which are usually present because of the coefficient of thermal expansion (a) differences between the base metal and inclusions. MnS inclusions are known to be of an elongated shape, therefore, providing a relatively large metal-inclusion void that can trap hydrogen and, ultimately, form a microcrack in the metal body (Ref. 28).  SSC is confined to highly stressed regions where hydrogen accumulation occurs preferentially. The combination of internal or applied stresses and a susceptible microstructure determine the critical hydrogen build-up needed for crack initiation and growth. The microstructural part of SSC susceptibility is controlled by the magnitude and dispersion of Ti and Nb carbides, hard low temperature transformation phases such untempered martensite or bainite and alloying elements such as C and Mn, which give  25  rise to segregation bands (Ref 29). Susceptibility is also affected by the presence and shape of oxide (e.g. Al203) and sulfide (e.g. MnS) inclusions. Ume et al. (Ref 30) have shown that nonmetallic inclusions can raise the SSC and HIC cracking susceptibility.  ath  values for X70 steels dropped from 80%YS to 50%YS due to added sulfur content which gave raise to MnS inclusions.  In general, SSC cracking propagates by bursts where the most favourable oriented microcrack will eventually reach the main crack earlier than the rest. Thus, the exhibited rates of crack growth can be directly related to the relative difficulty for microcracks to reach the main crack under a given applied stress, only those cracks that successfully meet the crack tip are able to keep on growing. SSC crack growth in the steel is mostly of a brittle nature, but it contains some ductile crack growth segments. A ductile segment represents the region between the main crack tip and a microcrack. This region cracks in a ductile manner when it can no longer hold the applied stresses and, as a result, leads to joining the microcrack to the main crack tip (Ref 31, 32).  2.3.3 Stress intensity at the SSC crack tip:  In most cases, SSC cracks initiate at surface imperfections or at pits generated from the corrosion reaction. The stress concentrations at the base of these pits can promote SSC crack nucleation.  26  To illustrate the stresses at the crack tip, a body with arbitrary shape with a crack of arbitrary size subject to arbitrary tensile loading is considered, Fig. 2.11. If the material is assumed to be elastic, following Hook's law, the theory of elasticity can be used to calculates the stress field and the stresses G, and Gy can be obtained (Ref 33)  Fig. 2.11 Stresses acting in front of a crack (mode I).  The crack tip stresses can be biaxial or triaxial (if contraction in thickness is constrained). As a result, there will be stresses in at least the x and y directions. Stresses in the material in Fig. 2.11 can be described by: G, = [K / .N1(2 it r)] cos (0/2) [1 - sin (0/2) sin (30/2)] + G, ^(2.13a) ay = [K / A2 it r)] cos (0/2) [1 + sin (0/2) sin (30/2)] ^(2.13b)  27  For r = x and 0 = 0 (plane through the cracked section), eqn. 2.13a and 2.13b become: a x = K / Ai(2 IC X) Gy =  ^  K / Ai(2 7r X)  (2.14a)  ^  (2.14b)  Here, K is the stress intensity factor. According to the above equation, crack tip stresses will be infinite at x = 0 regardless of the value of K. This is a direct consequence of the use of the theory of elasticity, which states a is linearly proportional to strain, E, without any limitation. In reality, a material exhibits plastic deformation which limits the stress as can be seen in Fig. 2.12 (Ref. 34).  Plastic k4 zone  ^Mow  r  Fig. 2.12 Crack tip stress distribution.  Since the stresses everywhere in an elastic body are proportional to the applied load, the crack tip stress is proportional to the applied stress a:  ay a [a / \I(2 a x)]^  (2.15)  28  For a very large (infinite) panel subjected to tensile loading with a nominal stress of a, Fig. 2.13, the crack tip stresses also depends on the crack length, a. The stresses are higher when "a" is longer so the crack size "a" must appear in the nominator in equation 2.15:  cry = (C a \la) / \1(2 a x)  (2.16)  K = C a Ala  (2.17)  -  Where "C" is a dimensionless factor.  Fig. 2.13 Center crack with uniform loading.  In the practical use of the equations above, all "C" terms are divided by \in and Ana) is -  substituted for Va to compensate. The function C/Jir is then renamed 13", the geometry .  factor:  29  Gy = (13 6  \Ina) / \/(2 7C X) = K / Ai(2 7E X)^(2.18) -  K =13 a pia^  (2.19)  It must be noted that equations 2.18 and 2.19 represent the crack tip stresses and stress intensity for all crack problems. The equations above are derived from the general solution for an arbitrary crack in an arbitrary body with arbitrary mode I loading (Ref. 35). For any crack in any particular problem, only the function f3 or its functional value needs to be derived. 13 is a function of the specimen geometry. The function 13 has already been calculated for many configurations (Ref 36).  Fracture occurs when the stresses at the crack tip becomes too high for the material to withstand. As the stress intensity factor, K, determines the entire crack tip stress field, fracture occurs when K becomes too high for the material. This critical value of K is known as the threshold stress intensity factor Kissc• "I" here refers to mode I (tensile loading) fracture and "SSC" refers to Sulfide Stress Cracking. The unit for the threshold stress intensity factor is ksi Niin or MPa gym. This follows from the definition of K which .  is the product between stress (ksi) and square root of crack length.  30  2.3.4 Testing for SSC susceptibility:  Standardized testing of metals for resistance to cracking failure under the combined action of stress and corrosion in H25 environments is very important to asses the SSC susceptibility of any material of interest and to assure safe operation. In addition, these tests facilitate conformity in testing so that data from different sources can be compared on a common basis. Standardized tests also aid the evaluation and selection of all type of metals and alloys regardless of their form or application for service in H25 environments. Most material standards include SSC testing requirements according to test methods that are described in NACE TM-0177 "Laboratory testing of materials for resistance to specific forms of environmental cracking in H2 S environments" (Ref 4). One of the most popular methods in the NACE standard is the uniaxial tensile test (constant load test).  The uniaxial tensile test evaluates SSC cracking susceptibility in H2S environments under an applied stress. Since the SSC phenomena produces cracks in materials bearing applied stresses below their yield strengths, test specimens are loaded to stress levels below the material's yield strength value. SSC susceptibility is determined by time-to-failure. A specimens loaded to a particular stress level give either a pass or fail test result after a specified test period.  When testing multiple specimens at varying stress levels, an apparent threshold stress for SSC can be obtained, a t h. The threshold stress represents the maximum stress above which SSC occurs on a smooth surface. The NACE TM-0177 standard indicates that a th  31  is the stress value below which no SSC failure is detected after the 720 hours (1 month) test period (Ref 4). This test is popular because the stress pattern is simple and uniform. In addition, the magnitude of the applied stress can be accurately determined. Figure 2.14 shows a schematic arrangement of test equipment for the uniaxial tensile test.  Force  Fig. 2.14 Diagram of uniaxial tensile testing (constant load).  Proof rings are one of the most economical and effective uniaxial stress methods used to determine SSC susceptibility of metals in H2 S service (Ref. 37). Proof rings are specifically designed to meet the NACE TM-0177 standard, and are less expensive to purchase, maintain, and operate than dead-weight testers. Proof rings are fabricated from precision-machined alloy steel and are available in many load ranges. Tension on the  32  Proof ring is quickly and easily adjusted using a standard wrench on the tension-adjusting screw and lock nut. A thrust bearing distributes the load and prevents seizure. Each individually calibrated proof ring is accompanied by a conversion chart that can accurately determine specimen load.  2.3.5 Factors influencing SSC susceptibility in HSLA steels:  Various metallurgical variables can be controlled to maximize HSLA resistance to SSC. The strength level (commonly measured by hardness) is a widely used criteria to predict performance in SSC media. NACE recommends that carbon and low alloy steels used in H2 S environments should have a hardness value of 22 HRC or less (Ref. 38). This hardness value has been challenged by many because a material susceptibility to SSC should not be determined from its hardness value alone since microstructure is very important and essentially controls the material's SSC properties (Ref 39).  2.3.5.1 Effect of material strength:  In general, SSC susceptibility increases with increasing strength. This can be seen very clearly in materials with the same microstructure. Hardie et al. (Ref 40) tested three API HSLA pipeline steels (X50, X60 and X70) from the point of view of their susceptibility to SSC. The microstructure of the three steels consisted of ferrite-pearlite bands. In this study, the degree of embrittlement tended to increase with strength level. The relationship between yield strength and SSC should not be taken when comparing different  33  microstructures because, even though YS plays a major role in SSC susceptibility, other microstructural features like banding, non-metallic inclusion and carbide distribution play an important role in SSC susceptibility.  2.3.5.2 Effect of microstructure:  In the work of Koh et al. (Ref. 41), different microalloyed HSLA steels with different microstructures were examined in an H 2 S environment. The four microstructures were coarse ferrite-pearlite, fine ferrite-pearlite, acicular ferrite and ferrite-bainite. SSC was evaluated by Constant Elongation Rate Test (CERT) in NACE TM-0177 "A" solution.  Elongation to failure in the test solution was measured and compared with those measured in air, Fig. 2.15. In general, all failures in test solution occurred much sooner than in air. Acicular ferrite was the least sensitive to cracking while ferrite-bainite was the most sensitive.  34  ^  F. oC  0.3 -  0.25  C  O  0.2  C O  10.15 00  ^  ^0.1  CO 0C  0.05 0^ M1^M2^M3^M4 Material  Fig. 2.15 Elongation (in NACE TM-0177 "A" solution) / Elongation (in air) for API-X70 with different microstructures (M1 = ferrite+coarse pearlite, M2 = ferrite + fine pearlite, M3 = acicular ferrite, M4 = ferrite + bainite).  In all microstructures, cracks nucleated at Al-Ca oxide inclusions. In addition to nonmetallic inclusions, microcracks nucleated at grain boundaries and at the needle shaped second phases in coarse ferrite-pearlite and ferrite-bainite steels respectively. These additional crack nucleation sites seemed to contribute to the microscopic advance of cracks. In addition, high SSC susceptibility was observed in coarse ferrite-pearlite and ferrite-bainite steels. Crack nucleation at grain boundaries in coarse ferrite-pearlite was due to the grain boundary network of cementite and it was concluded that segregation of cementite at grain boundaries must be avoided to increase SSC resistance. In ferritebainite, needle shaped second phase was sensitive to crack nucleation. These needle shaped second phases seemed to be related to the formation of interlath martensite and 35  retained austenite constituents (M/A). It was concluded that retardation of M/A constituents formation as a second phase is desirable to increase SSC resistance. On the other hand, fine-ferrite-pearlite and acicular ferrite steels didn't form the grain boundary cementite or the interlath M/A constituents, so that their resistance to SSC was higher. It has been shown, (Ref. 42), that upper bainite was more susceptible to SSC and showed lower threshold stress intensity factor (Kissc) values when compared to acicular ferrite and tempered martensite. The negative effect of bainite was related to the large matrixcarbide interface where crack could nucleate, thus, providing large hydrogen trapping sites.  Zhao et al. (Ref. 43) also investigated the role of microstructure on SSC. In a bent beam test conducted on three HSLA microstructures (acicular ferrite, ultrafine ferrite and ferrite-pearlite), acicular ferrite possessed the best SSC resistance by having the highest critical stress. Ultrafine ferrite was in second position while ferrite-pearlite steel was the most susceptible to SSC as it exhibited more cracks at even less applied stresses than the ultrafine ferrite microstructure. Zhao noted that a ferrite-banded pearlite microstructure is susceptible to SSC cracking. Zhao also noted that carbides and MnS inclusions precipitate along the banded pearlite. Hydrogen easily diffuses to the interfaces between these precipitates and banded pearlite reaching the critical hydrogen concentration. Furthermore, SSC propagation occurs normally through the susceptible cracking path provided by the segregated pearlite bands, because most of the microcracks formed in crack initiation can grow along the layered pearlite, which accordingly promotes the main cracks to propagate. On the other hand, there is no susceptible banded pearlite  36  microstructure in ultrafine ferrite or acicular ferrite so the harmful effects on SSC resistance caused by coarsening carbides and MnS inclusions can almost be neglected. Even if microcracks initiate due to HE, they can be retarded to further propagate by numerous grain boundaries because of the fine grain size.  The benefit of acicular ferrite comes from its good dispersion of carbonitrides and high dislocation density. Nb and V carbonitrides in acicular ferrite delay crack initiation through increasing the number and distribution of hydrogen trapping sites, instead of localized hydrogen accumulation (Ref. 44). High-angle boundaries in acicular ferrite act as obstacles to cleavage propagation, forcing the cleavage crack to change the microscopic plane of propagation in order to accommodate the new local crystallography (Ref 45, 46).  In addition to acicular ferrite, tempered martensite has shown good SSC resistance when compared to a lath martensite microstructure. Quenched and tempered martensite is a good candidate in H2S environments because of its dispersed carbon rich regions and carbides (dispersed hydrogen traps). Quenching is detrimental to SSC because it generates a susceptible lath martensite microstructure. In the as quenched condition, it is expected that carbon segregation at grain and interlath boundaries combines with internally stressed martensite to give a more susceptible steel condition. Multiple cracks can then appear along martensite interlath boundaries. Albarran (Ref 47) showed that lath martensite cracks at lower stress intensity factors than tempered martensite and with higher crack propagation rates.  37  2.3.5.3 Effect of material homogeneity:  In addition to microstructure, the distribution of different phases in the steel can contribute to the material's susceptibility in H2S environments. A banded microstructure provides high hydrogen trapping site density in front of the main crack tip yielding lower threshold stresses, cr th. In a banded microstructure, reaching the critical hydrogen concentration, for microcrack formation, will be easier because more hydrogen accumulates in these traps. A homogeneous microstructure is beneficial to steels because it provides dispersed hydrogen traps in which, reaching the critical hydrogen concentration for a single trap site is more difficult. Albarran and Flores (Ref. 48) observed that an API X52 and X80 with a banded ferrite-pearlite microstructure developed longer SSC cracks that followed the banded ferrite. Significant microcracks ahead of the crack tips were observed. On the other hand, X65 samples with homogenous acicular ferrite microstructure had the lowest crack propagation rate.  Carneiro et al. (Ref. 49) tested the effect of homogeneity in H 2 S environments for API X60 steels. X60 specimens were given different thermal treatments to produce different microstructures. The as received material consisted of a banded ferrite-pearlite microstructure while quenching and tempering produced a homogenous ferrite-pearlite microstructure. The homogenous microstructure generated the highest a th value (70% of YS) as seen in Table 2.1. In addition, banded microstructures were more susceptible to HIC cracking. HIC cracks initiated at hard phase constituents and propagated in a  38  direction parallel to the stress direction. Corrosion pits and HIC cracks acted as SSC nucleation sites.  X60  100  Applied stress (% of yield strength) 85 70  60  As received  7  114  365  720  Quenched and tempered  22  28  720  ---  Table 2.1 Time to failure results (hours) for as received and quenched-tempered X60. Adapted from (Ref. 49).  39  Chapter 3 3 EXPERIMENTAL PROCEDURES:  The aim of these experiments is to study the behaviour of the current generation X100 pipeline steel in SSC media. An older generation of HSLA steels, X80, was used for comparison in microstructural and electrochemical examinations. For proof ring testing, experiments were conducted for the X100 and results were compared with earlier work on older generation HSLA steels reported in literature.  3.1 Tested materials  X100 and X80 pipeline materials were used in this study. The two materials were supplied by CANMET Materials Laboratories. In order to determine the chemical composition (wt.%) of the specimens to be used in the experiments, X100 and X80 samples were sent to International Plasma Labs (IPL) for chemical analysis. The specimens were analyzed by inductive coupled plasma (ICP) and LECO carbon analysis. Results for both materials are shown in Table 3.1. Composition (Wt. %)  C.E.  C  Mn  Mo  Ni  S  Al  Cu  Ti  Nb  V  X80  0.09  1.42  0.21  0.002  <0.01  0.04  0.19  0.01  0.07  0.022  0.39  X100  0.10  1.67  0.21  0.09  <0.01  0.02  0.24  0.01  0.043  0.003  0.47  Table 3.1 Chemical analysis with carbon equivalence (C.E.) for the X80 and X100 steels.  40  3.2 Microstructure analysis:  After mounting, the specimens underwent the following steps before examining their microstructures under the optical microscope:  •  Wet-ground in a series of progressively finer silicon carbide papers up to 1200 grit finish.  •  Polishing in 6 um diamond suspension.  •  Cleaning with ethyl alcohol swabbing followed by drying in warm air.  •  Polishing in 11,tm suspension.  •  Repeat cleaning step with ethyl alcohol swabbing followed by drying in warm air.  •  Etching in 2% nital (2 ml nitric acid + 98 ml ethyl alcohol).  •  Cleaning in water followed by alcohol swabbing then air drying.  In order to examine the as-polished surface for non-metallic inclusions, the steps above were repeated except for the etching step. Specimen etching was replaced by a final polishing step using 0.05 i-tM colloidal silica followed by alcohol swabbing then air drying. All surfaces were examined under light microscopy (LM). Scanning electron microscopy (SEM) equipped with an energy dispersive X-ray spectrometer (EDX) was used to investigate as polished specimens and non-metallic inclusions.  41  3.3 Electrochemical testing:  3.3.1 Specimen preparation:  Several flat disc specimens were machined from the X80 and X100 steels with nominal dimensions of 16 mm diameter and 3 mm thickness as seen in Fig 3.1. The working electrode surface was prepared as follows: •  Wet ground to a 1200-grit silicon carbide (SiC) finish.  •  Degreased in acetone  •  Rinsed with distilled water  •  Dried with alcohol under a stream of warm air  Mounted specimen  Polarization specimen  1.6 c  83  cm  Fig 3.1 Specimens used for microstructural analysis (left) and polarization analysis (right).  42  3.3.2 Test solution:  NACE TM0177 test solution "A" (Ref. 4) was used in all electrochemical tests in this study. Test solution "A" consisted of an acidified H2S saturated aqueous environment containing 5.0 wt% NaC1 and 0.5 wt% CH3COOH. The solution pH after saturation was expected to range between 2.6 and 2.8 and increase to below 4.0 as the test progresses. pH values were measured with a combination pH electrode (glass membrane with Ag/AgC1 reference element) calibrated with pH value of 7.0 standard buffer solution. H 2 S gas, NaCl and CH3COOH were reagent-grade chemicals. For each test, 600 ml of test solution were used. A fresh solution was used in every new test. To study the electrochemical behaviour in the absence of H2S, the same NACE TM-0177 "A" solution composition was used but without the addition of H2S.  3.3.3 Cell setup:  In this study, all electrochemical tests were conducted at 23 °C and atmospheric pressure. Tests were made using a standard glass cell containing the working electrode (specimen) and a graphite counter electrode. Potentials were measured with reference to a saturated calomel reference electrode (SCE) interfaced to the test solution via a salt bridge that terminated about 2 mm from the specimen, Fig. 3.2.  43  Reference electrode  Working electrode  Fig. 3.2 Electrochemical polarization experimental setup.  A SolatronTM potentiostat system (Model 1286) was utilized to perform and analyze the potentiodynamic polarization curves. The system was coupled with a software package having a combination of CorrWareTM (for hardware control and data acquisition) and CorrViewTM (for data comparison and analysis).  After pouring the solution and sealing the cell, the cell was deaerated by argon for 1 hour to eliminate any oxygen interference with the electrochemical reaction. After purging, H2S was bubbled in to the cell at a flow rate of 50 cc/min for 30 minutes before starting the test. H2S bubbling was eliminated when testing the X100 in the H 2 S free NACE TM0177 "A" solution.  44  3.3.4 Electrochemical techniques:  3.3.4.1 Open circuit potential measurements:  After preparing and sealing the electrochemical cell, the test specimen was immersed in the test solution for 33 minutes in order to measure the open circuit potential (E 0q ). Eocp measurement was made between the working electrode (specimen) and the reference electrode. No current was passed in this test to the working electrode. The aim of this test was to find the potential at which the anodic and cathodic reaction currents at the working electrode/solution interface were balanced. E ocp measurements are needed prior to the electrochemical polarization tests to insure stability.  3.3.4.2 Potentiodynamic polarization tests:  After reaching a stable open circuit potential (E ocp ), the electrode potential was swept potentiodynamically at a scan rate of 0.66 mV/sec from an initial potential of -0.15 V to 0.5 V versus open circuit potential unless otherwise indicated. All potentials in this test were measured with respect to the saturated calomel electrode (E = 0.241 V vs. standard hydrogen electrode SHE). The corroded surface was examined under scanning electron microscopy (SEM) equipped with an energy dispersive X-ray spectrometer (EDX).  The corrosion rates, in mils (0.001 inch) per year (mpy), for the tested materials were calculated by the following equation:  45  Corrosion rate (mpy) = (0.129 x icon- x EW) / (D)  ^  (3.1)  Where:  icon = corrosion current density in pA/cm 2 .  EW = the equivalent weight, indicating the mass of metal in grams, which is oxidized. For carbon steels, EW is approximately 28 grams (assuming Fe oxidation state of +2). D = density in g/cm 3 . Density for carbon steel is 7.84 g/cm 3 .  3.4 SSC constant load testing:  3.4.1 Proof ring devices:  Time to failure data, by SSC constant load testing, was conducted at Bodycote labs while crack imaging, crack analysis and threshold stress determination were conducted at UBC. In this study, all proof ring tests were conducted at room temperature. A calibrated proof ring was used to test X100 specimens at constant load (CL). Fig. 3.3 shows a description of the different components of the proof ring. The used proof rings were specifically designed to meet the NACE standards (Ref 4, 38). Standard tensile specimen dimensions were used as seen in Fig. 3.4. Each individually calibrated proof ring was made using precision-mechanics alloy steel and was accompanied by a calibration curve showing the load versus deflection. The test specimens were loaded under uniaxial tension. The  46  tensile load on the proof ring was quickly and easily adjusted using a standard wrench on the tension-adjusting screw and lock nut. A thrust bearing distributed the load and prevented seizure. The environmental test chamber was secured by 0-ring seals that prevented any leakage during testing.  Dial  Environment ( harnber  Fig. 3.3 Proof ring testing device.  47  Fig. 3.4 X100 tensile specimen.  The amount of deflection needed to apply the desired load with the proof ring was determined by using the calibration curves of each proof ring. The magnitude of the applied stress was based on the ambient-temperature yield strength (YS) of the tested X100 specimens. Yield strength values for the X100 were generated by pulling X100 tensile specimens in air using a 50 kN InstronTM universal tensile machine. Tension testing was conducted in accordance with ASTM standard E8-04, standard test method for tension of metallic materials (Ref 50).  Loads for stressing the specimens were determined from the equation:  (3.2)  P=SxA^ Where:  48  P = Load S = applied stress A = actual cross section area of the gauge section (31.67 mm 2 )  The specimens were loaded at stress values equivalent to different percentages of the material's YS value and the corresponding time-to-failure (TTF) was recorded. Table 3.2 lists the different stresses applied on the X100 specimens. Determination of SCC susceptibility using this technique was based on the TTF for a maximum test duration of 720 hours (30 days). An automatic timer attached to the test specimen recorded the TTF. The cracking susceptibility was expressed in terms of a threshold stress (a th) for SCC below which cracking did not occur during the maximum test duration.  # specimens  Applied stress (%YS)  3 3 3 3 3 Total # of specimens : 15  80% 65% 50% 40% 30% Test termination: Specimen failure or after 720 hrs. (1 month), which ever occurs first.  Table 3.2 X100 proof ring specimens and applied stresses.  49  3.4.2 Test solution:  NACE TM0177 test solution "A" (Ref. 4) was used in all electrochemical tests in this study. This is the same test solution used in the electrochemical measurements. A fresh solution was used in each test. In addition, initial and final pH values were measured for each test.  3.4.3 Testing procedure:  The test procedure was conducted according to the following sequence: •  The minimum gauge diameter of the test specimen was measured and the test specimen load was calculated accordingly for the desired stress level.  •  The tensile test specimen was cleaned and placed in the test vessel. The test vessel was sealed to prevent air leaks into the vessel during the test.  •  After loading and purging the test vessel, the test vessel was immediately filled with deaerated test solution. The solution was poured to a level where solution-gas interface didn't contact the gauge section of the test specimen. Deaerated solution was prepared in a sealed vessel that was purged with inert gas.  •  The solution, once in the test vessel, was purged with inert gas for at least 20 min to insure that the test solution is oxygen free before introducing H 2 S.  •  The test solution was then saturated with H 2 S at a rate of 150 mL/min. for 30 minutes. A continuous flow of H2S through the test vessel and outlet trap was maintained for the duration of the test at a low flow rate (a few bubbles per minute). This maintained  50  the H 2 S concentration and a slight positive pressure to prevent air from entering the test vessel through small leaks. • The test was terminated either after specimen failure (double ended fracture) or 720 hours (1 month), which ever occurred first.  3.4.4 Failure detection:  Following exposure, the surface of the gauge section of the non-failed test specimens were cleaned and inspected for evidence of cracking. The specimens were cleaned by an inhibited acid consisting of a stock solution of HC1 + 3.5 g/1 of methenamine diluted with an equal amount of distilled water. The failure criterion was either complete separation of the test specimen or visual observation of cracks on the gauge section of the test specimen at 10X magnification after completing the 720 hour test duration (1 month).  3.4.5 Failure analysis:  In order to analyze the generated cracks, specimens were cut in both the longitudinal and transverse directions to observe the generated cracks. Fracture faces were examined under the SEM microscope to observe brittle and ductile regions. Transverse and longitudinal cut specimens were ground to 1200 grit finish, polished with 6 and 1 micron diamond suspension then etched in 5% nital. The specimen and generated cracks were then examined under the light microscope.  51  Chapter 4  4 RESULTS AND DISCUSSION  4.1 Metallurgical examination:  4.1.1 Microstructure and strength:  Optical microscopy images for the X80 and X100 microstructures are shown in Fig. 4.1 and Fig. 4.2 respectively. The X80 microstructure was composed mainly of ferrite and regions of granular bainite, which is composed of low carbon bainite and martensite. On the other hand, the X100 steel showed a more complex microstructure consisting of martensite, bainite and ferrite. Bainite was observed to grow from the prior austenite grain boundaries that were elongated due to rolling.  The chemistry variation between the two steels is one of the factors affecting the differences in microstructure between the X80 and X100 steels. It can be seen from the chemical analysis of the two materials that the X100 possessed higher Cu, Mn, Mo and Ni contents. These alloying elements tend to increase the metal hardenability and shift the continuous cooling transformation (CCT) diagram line to the right. Shifting of the CCT line in this manner increases the formation of low temperature transformation microstructures like bainite and martensite under the same cooling rate (Ref. 51).  52  Fig. 4.1 X80 light microscopy imaging showing granular bainite (dark) and ferrite (light) microstructure.  Fig. 4.2 X100 light microscopy imaging showing martensite (dark), bainite and ferrite (light) microstructure.  53  In that essence, the X80 produced the ferrite-granular bainite microstructure because the X80 CCT line promoted the formation of higher temperature transformation temperatures due to the material's less alloying elements, compared to X100, and limited the possibility of forming larger amounts of bainite and martensite to very high cooling rates.  The resulting differences in microstructure had its effect on the material's strength values. The yield strength for the X100 was determined by pulling X100 tensile specimens in air as shown in Fig. 4.3a. The engineering yield strength for the X100 was 743 MPa (107.8 ksi), Table 4.1. X80 tensile specimens were also pulled in air to verify the material's yield strength value as shown in Fig. 4.3b and Table. 4.1. The X80 stressstrain curve generated an engineering yield strength value of 579 MPa (84.0 ksi). The generated yield strength values are in agreement with HSLA specifications. It should be noted that specimen elongation during the test reduces the cross sectional area. As a result, the true stress applied on the gauge section is higher than the engineering stress determined from the stress-strain curve, which is based on the original cross section area. Assuming that no volume change occurs during the test, the true stress  (Gtrue),  engineering  stress (a erig) and engineering strain (E eng ) are related by the following equation:  true = Geng ( 1  +  ceng)  (4.1)  The yield strength value for X80 shows the effect of HSLA microstructure on strength. For API grade HSLA steels, increase in strength from older generation HSLA steels, like X70, is attained by changing the microstructure from ferrite-pearlite to ferrite-bainite.  54  1000 ^ 900 —  200 100  0  ^  5^10^15  ^  20^25^30  Strain (%)  (a) Engineering stress-strain curve for X100 1000 900 800 700 O-  M  w  600 500  in 400 300 200 100 0 0  5  10  15  20  25  30  Strain %  (b) Engineering stress strain curve for X80 Fig. 4.3 Engineering stress-stain curve for in air for (a) X100 and (b) X80.  55  Specimen Type  Yield MPa (Ksi) (engineering)  Yield MPa (Ksi) (true)  UTS MPa (Ksi) (engineering)  UTS MPa (Ksi) (true)  743 (107.8)  746 (108.2)  850 (123.3)  893 (129.5)  579 (84.0)  581 (84.3)  675 (97.9)  745 (108.1)  X100 6.35 mm 0 X80 6.35 mm 0  Table 4.1 Average yield and tensile strength values (engineering and true) for X100 and X80.  Table 4.2 shows the hardness values obtained for the two materials using the Rockwell C scale. The higher hardness values for the X100 are attributed to the low temperature transformation microstructures in the steel. Lower hardness values for the X80 steel are a result of its ferrite-granular bainite microstructure which was expected to generate lower hardness values than the X100. It should be noted that the X100 hardness values exceeded the maximum hardness value of HRC 22 as recommended by NACE for pipeline materials to be used in H 2 S environments (Ref. 38). Specimen  Hardness (HRC) X80  X100  1 2 3 4 5 6 7 8 9 10 Average Standard deviation  18.3 18.8 17.8 17.9 18.2 17.7 19.3 18.2 18.6 19.1 18.4 0.54  23.3 25.2 23.4 23.5 25.2 23.8 23.8 24.3 25.3 24.4 24.2 0.78  Table 4.2 Average hardness values and standard deviation in Rockwell C for X80 and X100.  56  4.1.2 Nonmetallic inclusions  As can be seen in Fig. 4.4a and 4.5a for the X80 steel, the non metallic inclusions were spherical Al-Ca-O-S inclusions. The spherical inclusions were randomly dispersed in the metal and showed no signs of clustering. Inclusion diameter ranged between 2.0-4.5 pm. The average spherical inclusion diameter was 3.5 gm. No elongated MnS inclusions were observed in the steel.  n  (a)  :2 I^L, ^•  (b) Fig. 4.4 Spherical non-metallic inclusions in (a) X80 and (b) X100 (SEM).  57  Most of the examined inclusions in the X100 steel, Fig 4.4b and 4.5b, showed a similar spherical shape to those observed in the X80 samples. Although the X100 and X80 samples had a similar inclusion shape, the size of these non-metallic inclusions was different. Spherical inclusion diameter in the X100 ranged between 6.0-9.5 gm. The average inclusions diameter for X100 was 8.0 gm. Therefore, the average inclusion diameter in the X100 was higher than the X80 average inclusion diameter, 3.5 lam.  a  kee  2  (a) Counts  Fe  N.,  2  (b) Fig. 4.5 EDX spectra for (a) X80 and (b) X100.  58  In addition to spherical inclusions, the X100 steel showed some stringers which were not observed in the X80 steel, Fig 4.6. Those stringers are most likely a result of rolling when large oxide inclusions are broken into a line of smaller particles. The broken small particles connecting the other larger pieces of the stringers are evidence that these inclusions are hard and brittle. Stringers formed by softer inclusions would have generated a smooth "line like" connection between the larger pieces of the inclusion. No MnS inclusions were found in the X100 specimens.  nnnn 2nkV  acwffl  Fig 4.6 Stringers found in the X100 steel, arrow indicating rolling direction (SEM).  EDX spectra for the X80 and X100 spherical inclusions showed that the inclusions consisted of aluminium, calcium, oxygen and smaller amounts of sulfur. This inclusion chemistry is typical for Al killed and Ca treated carbon steels and mainly consists of 59  calcium aluminate (CaO-Al 2 0 3 ) and calcium sulfide (CaS) (Ref 52). As for the stringer inclusions found in the X100 steel, the bright regions (larger parts) of the stringer consisted of MgO-Al 2 0 3 the while darker regions (smaller parts) consisted of CaO-Al203. Nuspl et al.(Ref. 52) showed that brittle stringers found in carbon steel pipeline materials mainly consisted of MgO, Al 2 0 3 and CaO. In addition, since both steels had a low sulfur content, most of the sulfur is bound in calcium sulphides (Ref 53). This explains why no elongated MnS inclusions were found in the steel as most of the sulfur reacted with calcium to form CaS instead of MnS.  Stress localization at the inclusion/matrix interface is a preferential site for crack initiation and hydrogen trapping. This arises from:  (1) Differential thermal contraction of inclusions and matrix during cooling (2) Concentration of remote applied stresses.  Gaps and cracks initiate at the interface between the inclusion and the steel matrix during hot deformation such as rolling. An example of this was observed in some of the inclusions in the X100 material, Fig. 4.7. The effect of inclusions on the crack initiation depends on chemistry, size, density, location, and morphology of the inclusion. Large and brittle Al203 inclusions are more detrimental to fracture toughness than soft MnS inclusions. However, it should be noted that elongated sulfide inclusions, such as MnS, are very effective in promoting hydrogen embrittlement because of its planner shape which is easier to trap hydrogen atoms (Ref 54).  60  Fig. 4.7 Initiated cracks at the interface between an inclusion and the steel matrix in X100 specimen, arrow indicating rolling direction (SEM).  The large inclusions found in the X100 steel are expected to reduce cracking resistance in sour environments much more than the small inclusions. Dub et al. (Ref 55) stated that CaO-Al203 inclusions have a lower thermal expansion coefficient than iron (matrix). As a result, gaps and cracks formed at large inclusions after cooling or during rolling can act as hydrogen accumulation sites. Finally, large surface inclusions can provide a larger inclusion/matrix interface that can preferentially corrode and form corrosion pits on the surface (Ref 55).  61  4.2 Electrochemical behaviour:  This section demonstrates the results generated from the electrochemical polarization testing that was conducted on the X100 and X80 HSLA steels. The aim of these tests is to study the effect of X100 and X80 microstructures on corrosion rates. The NACE TM0177 "A" solution was used in these tests. To examine the effect of H2S, testing in H2S free NACE solutions were conducted and compared with results generated in saturated H2S environments.  4.2.1 Corrosion potential:  Fig. 4.8 shows the open circuit potential (E ocp ) measurement for the API X100 steel over 1800 seconds (30 min.) of immersion in the (H2S-saturaded) NACE TM0177 "A" solution. Potential readings were slightly fluctuating in the first 200 seconds. The readings then started to decrease until a steady state open circuit potential, E  ocp ,  reading  of -0.61 V was reached. E ocp for the X100 steel in the (H2S-free) NACE TM0177 "A" solution yielded a similar trend with an open circuit potential value of -0.63 V (average value).  62  Fig. 4.8 E ocp for X100 versus SCE in (H 2 S-saturated and H2S-free) NACE TM0177 "A" solution (5.0% NaCI + 0.5% CH3COOH), T = 23 °C.  The X80 E ow measurements in the (H2S-saturaded) NACE TM0177 "A" solution showed a similar decreasing potential trend before reaching a steady state value, Fig. 4.9. With the X80 steel, potential values kept decreasing before reaching a steady state value of 0.63 V. E ocp for the X80 steel in the (H2S-free) NACE TM0177 "A" solution was -0.65 V.  63  Fig. 4.9 E ocp for X100 versus SCE in H2S-saturated NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH), T = 23 °C.  4.2.2 Anodic polarization behaviour:  4.2.2.1 X100:  The polarization diagram for the X100 steel in H2S-saturated NACE TM0177 "A" solution is shown in Fig. 4.10. The anodic curve showed linear Tafel behaviour until about -0.5 V. At that potential, dark corrosion products formed on the surface, the corroded surface became darker with increased applied potentials and showed a black color appearance. After -0.5 V, current density continued to increase with increasing potential but at a lower rate. Some of the produced corrosion product started to spall off the specimen and mix with the test solution. The corrosion current density (i co ,) was found to be 210 RA/cm 2 by extrapolating the linear Tafel slopes. The initial pH was 2.7.  64  The pH at the end of the test was 2.9. The increase in bulk pH was due to the consumption of hydrogen ions to form hydrogen gas (atomic or molecular) as part of the cathodic reaction:  2H + + 2e"  H2  or 2H  ^  (4.2)  X100 (H2 S -saturated) 0D -0.1-02 -0 3 -  5 -0 4 w -0 .6 -0 -0 .7 -0 2  -  ^ ..t 1.0E08^1.0E07^1.0E06^1.0E06^1.0 504^10E03  -0 D ^i  1.0E02^1.0E01^1.0E00  (Ak m2)  Fig. 4.10 X100 polarization (3 measurements) in H 2 S-saturated NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH), T = 23 °C.  To study any influence of H 2 S on the polarization behaviour of the X100 steel, a round of polarization tests were conducted using H 2 S-free NACE TM0177 "A" solution. The test solution had an initial pH of 2.9. The pH value increased during the test to a 3.0 (caused by hydrogen ion consumption). Fig. 4.11 shows the resulting polarization behaviour  65  without H2S. With no H2S, the cell had a lower open circuit potential, -0.63V, than the value obtained in the H 2 S saturated test. By extrapolating the linear Tafel slopes, the corrosion current density (i con ) was found to be 112 liA/cm 2 . At potentials above -0.30 V, the color of the metal surface became darker and turned black at the end of the test.  X100 (H2S-free) 0 -0.1 -02 -03  5 -0.4 LU  -0 -0.7 -02 -OD ^• 1.0E-08 1.0E-07 1.0E-06 1,0E-05 1.0E-04 1.0E-03 1.0E02 1.0E-01 1.0E+00 i (A fo m2)  Fig. 4.11 X100 polarization (3 measurements) in H2S-free NACE TM-0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH), T = 23 °C.  4.2.2.2 X80:  Even though the X80 exhibited lower open circuit potentials than the X100, the general structure of the polarization diagrams for the X80 steel in saturated H 2 S solutions were similar to the X100, Fig. 4.12. The anodic polarization curve started off linearly until 0.51V where, similar the X100, current density continued to increase with increasing  66  potential but at a lower rate. The corrosion current density (i cc,n) was 174 iA/cm 2 . Like the X100, corrosion product accumulated on the surface until it reached a point where some of it started to spall off the specimen surface and mix with the solution. The initial pH was 2.7 and increased to 2.9 at the end of the test due to consumption of hydrogen ions (cathodic reaction).  X80 (H2 S-saturated) D -0.1 -0 2 -0 3 5 -04 w -0 .5 -016 -0.7 -0 2 -0 9 111608^10E07^1.0E013^1.0E06^1.0 504^1.0E03^1.0E02^1.0E01^1.0E4-00 i (A fc m2)  Fig. 4.12 X80 polarization (3 measurements) in H 2 S-saturated NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH 3 COOH), T = 23 °C.  Like the X100, polarization tests in H2S-free NACE TM-0177 "A" solutions were conducted to observe the influence of H2S on the polarization behaviour of the X80 steel. Polarization of X80 in H 2 S-free NACE TM-0177 "A" solution had an open circuit potential of -0.65 V. Initial and final pH were 2.9 and 3.0 respectively. Fig. 4.13 shows the resulting polarization behaviour of X80 without H2S. The corrosion current density  67  ^  (icorr)  was found to be 98 ttA/cm 2 . The color of the specimen surface became dark at  potentials above -0.33 V and became black at the end of the test.  X80 (H 2 S-free) ^0  ^  -02  7 7  -03  —  -0.1  ▪  -0.4 —  ;7; - 0 b  -  -0e6 — -0.7  -  -013 -0.9 ^ 1.0E-09  1.0E-07 1 DE-06 16E-05 1.0E-04 1.0E-03 1.0E02 1.0E-01 1.0E+00  i (A k m2)  Fig. 4.13 X80 polarization (3 measurements) in H2S-free NACE TM0177 "A" solution (5.0% NaC1 + 0.5% CH3COOH), 23 °C.  4.2.3 Surface characterization:  After the polarization tests, X100 and X80 corroded specimens from the two environments (with and without H 2 S) were examined using an energy dispersive X-ray spectrometer (EDX). Fig 4.14 and 4.15 show the different surface chemistries found on the specimens.  68  Counts  0  ^ ^ ^ ^ ^ keV 4 8 6 2  (a) H2 S-free  0  ^ ^ ^ ^ ^ 8 keV 4 6 2  (b) H2 S-saturated  69  2  ^  4  ^ ^ ^  8  8  keV  (c) Polished uncorroded X100 specimen Fig. 4.14 Corrosion product EDX for X100 in (a) H2S-free, (b) H2S-saturated NACE TM-0177 "A" solution and (c) Polished uncorroded X100 specimen was used as a reference.  Counts 4000  Fe  2  ^  4  ^  (a) H2S-free  70  a  ^  keV  Coures  3000  2000 -  1000  2  ^  8  ^  keV  (b) H2S-saturated Fig. 4.15 Corrosion product EDX for X80 in (a) H2S-free and (b) H 2 S-saturated NACE TM-0177 "A" solution.  It can be seen that in H2S-free NACE TM-0177 "A" solution, the EDX spectra was mainly composed of iron and oxygen. These products on the metal surface did not promote passivity, as can be seen in the (H2S-free) polarization diagrams for the two materials (Fig. 4.11 and 4.13). In the (H2S-free) polarization diagrams, the measured anodic current densities kept on increasing with potential and the resulting anodic polarization curve didn't show any sign of passivity. The presence of these products on the metal was most likely caused by precipitation of Fe(OH)2 on the surface. Fe(OH)2 was likely produced by the reaction between Fe 2+ and H 2 O:  2H + + 2e" -3 H2^(corrosion cathodic reaction)^(4.3) Fe^Fe 2+ + 2e  ^  (corrosion anodic reaction)^(4.4)  Fe 2+ + 2H 2 0 Fe(OH) 2 + 2H +^(4.5)  71  In H 2 S-saturated NACE TM0177 "A" solutions, the EDX spectra showed significant amounts of sulfur. After the consumption of H + , produced by CH3CHOOH, has reached a certain degree (increase in pH), FeS was likely formed by the reactions below:  Cathodic reaction:  H2 S + H2O 4 H + + HS - + H 2 O HS - + H 2 O 4 H+ + S 2- + H2 O 2H + + 2C 4 H2 or 2H  ^  ^  ^  (4.6) (4.7) (4.8)  Anodic reaction:  Fe 4 Fe 2+ + 2C  ^  (4.9)  Overall:  Fe + H2 S 4 FeS + 2H  ^  (4.10)  In the EDX spectra, the Mo and S peaks are very close and some times a Mo peak might be labelled by EDX as an S peak and vice versa. In the H2S-free EDX spectra, this peak was most representing Mo because the concentration of Mo in both steels (0.21 wt.%) is much higher than S (<0.01 wt.%). In addition, the H2S-free environment is not expected to promote any sulfide film formation. In H2S-saturated environment, the peak was likely  72  representative of S because the presence of H 2 S generates a sulfide corrosion product, which increases the S peak in the EDX spectra.  4.2.4 Mechanistic insights:  Table 4.3 summarizes the results obtained from the electrochemical tests on both the X100 and X80. It can be seen from Table 4.3 that the starting pH values were lower in H2 S environments. Both E con. and icon increased when H2S was added to the test solution. It can also be seen that anodic Tafel slopes, r3 a , were close in value. Likewise, values for the cathodic Tafel slopes,  P c , were also similar. Addition of H2S to the NACE TM0177  "A" solution increased the corrosion rate of X100 from 51.6 to 96.7 mpy. The effect of H2 S on the corrosion rate was similar for X80. The corrosion rate for X80 increased from 45.2 to 80.2 mpy. pH Initial/final  Emil'  X80 (H 2 S-free solution)  (V)  (pA/cm2)  fla (V/dec.)  Pc (V/dec.)  Corrosion rate (mpy)  2.9 / 3.0  -0.65  98  0.065  -0.121  45.2  X80 (H 2 S-saturated solution)  2.7 / 2.9  -0.63  174  0.063  -0.112  80.2  X100 (H 2 S-free solution)  2.9 / 3.0  -0.63  112  0.066  -0.123  51.6  X100 (H 2 S-saturated solution)  2.7 / 2.9  -0.61  210  0.065  -0.111  96.7  iCOrr  Table 4.3 Summary of electrochemical test results conducted on API-X100 and X80 in H2S-free and H2S-saturated NACE TM0177 "A" solution.  73  4.2.4.1 Effect of H2S on corrosion rates: Fig. 4.16 shows the polarization diagrams of (a) X100 and (b) X80 with and without H2S (NACE TM0177 "A" solution). Schematic Evans diagrams for each metal (with and without H2S) were also drawn as shown in Fig. 4.17. X100 ( S -saturated ) X100 (H2S free) 0D ^ -0.1 -0 2 -0 3 -0.4 "'  -0,5-00-0.7 -0 2 -0  9  10E08^1.0E07^1.0E06^1.0E06^12E04^1.0E03^1.0E02^1.0E01  ^  1.0E1-00  i (Akm2)  (a) X10 (H2 S -saturated )  ----X00 (H 2 S-free) OD ^ -0.1 -0 2 -02 -0.3 -0.4 w  -0.6 -00 -0.7 -0.8 -0 9 ^ 12E08  1.0E07^1.0E06^1.0E06^1.0E04^1.0E03^1.0E02^1.0E01^1.0E4-00 i (Akm2)  (b) Fig. 4.16 Polarization diagrams of (a) X100 and (b) X80 in NACE TM0177 "A" solution (H2S-free and 1 12S-saturated), T = 23 °C. -  74  u.i  Log i  X 1 00  (a)  (b) Fig. 4.17 Evans diagrams for (a) X100 and (b) X80 in NACE TM0177 "A" solution (H 2 S-free and H2S-saturated). E = potential, Log i = log value of current density.  75  The Evans diagrams demonstrated two effects of H 2 S. When H2S was added, the anodic line was accelerated and produced more anodic current densities (shifted to the right). In addition, the cathodic line produced more current densities by shifting upwards.  The effect of H2S on the anodic reaction was likely caused by H2S chemisorbtion on the metal surface. This catalyzes the anodic discharge reaction as seen in reaction 2.10 to 2.12 and generates more anodic current densities for the same potential by reducing the activation energy required for the anodic reaction.  Fe + H2 S^FeHS - ads  (2.10)  FeHS - ads^4 FeHS ± ad s + 2e"  (2.11)  FeHS + ad s +1-1± 4 Fe 2+ + H 2 S  (2.12)  Fig. 4.18 shows a plot of anodic overpotential (i) vs. current density, where ri = E oc p.  Eapplted  -  It can be seen that, for the same II, the anodic polarization curve shifted to more  current densities in H2S-saturated NACE TM-0177 "A" solution. Cheng et al. (Ref. 56) tested the anodic behaviour of an iron electrode (99.99% purity) in H2S media (acidic environments with pH < 5). The study showed that increasing H 2 S concentration shifted the anodic polarization curve to the right due to the catalyzing effect on the dissolution reaction of iron.  76  X100 ( H2 S -saturated ) X100 (H 2 S- free) 0.8  5 0.5 ? 0.4  -  0.3 -  ‹Y; 0.20.1 0^ 1.0E08  1.0 6.07^1.0E08^1.0E05^1 DE04  1.0E03^1.0E02^1.0E01^1.0B-00  i (Pabm2)  (a) X100 ^X80 ( H2 S -saturated ) XBO (112S - free) 0.6 5 0.5 .  la. 0.4 0.3 0.2 0.1 0^ 1.0E08  1.0E07^1.0E08^1.0E1)6^1.0E04^1.0E03^1.0E02^1.0E01^1.0E00 i (Akm2)  (b) X80 Fig. 4.18 Anodic overpotential (i) vs. current density for (a) X100 and (b) X80 in H2S-free and H2S-saturated NACE TM-0177 "A" solution, T = 23 °C.  In addition to the effect on the anodic reaction, H 2 S also affected the cathodic reaction. The Evans diagrams in Fig. 4.17 show a shift of the cathodic lines to more current densities. This effect is due to an increase in the cathodic half cell potential, eH+/H2, because of the decrease in pH that is observed with H2S saturation. Table 4.3 shows that  77  pH values after H2S addition decreased from 2.9 to 2.7 for the two materials. The Nernst equation for hydrogen evolution is as follows:  eH+/H2 = eH+/H2  ° - (0.059) pH  ^  (2.9)  The standard half cell potential for hydrogen evolution, eH+/H 2 °, equals zero (Ref. 23). Applying the test pH values in the above equation shows that, for X100 and X80, the cathodic half cell potential, eH+/H2 increased from -0.171 to -0.159 V when H 2 S was .  added to the test solution. The increase in potential, about 7.0%, can have a significant effect on the corresponding results from the "Log i" axis in the Evans diagram. This can be seen from the increase in corrosion rate with added H2S. The corrosion rate increased from 45.2 to 80.2 mpy for X80 and from 51.6 to 96.7 mpy for X100, Table 4.3.  4.2.4.2 Comparison between X100 and X80:  Fig. 4.19 shows the polarization diagram of X100 and X80 in H2S saturated NACE TM0177 "A" solution. Schematic Evans diagrams were drawn as shown in Fig. 4.20. Even though the X100 and X80 were tested in the same environment, the X100 corrosion rate was higher by 20% (Table 4.3). It can be seen from the Evans diagram that there was a significant change in the cathodic line which shifted to higher current densities for X100. Unlike the Evans diagrams that were generated with and without H2S, Fig. 4.17, the increase in current densities can not be attributed to an increase in  78  eH+/H2  because in  ^  saturated H2S tests, there was no change in pH so the cathodic half cell potential will be the same for both metals. X100 (H2S -saturated ) X80 (H2S-saturated) On ^ -0.1  -  -0 2  -  -013 -0.4 w  -0,5-0.6 -0.7 -DS -0.9 ^ 1.0508^113507^1.0506  1.0505^1.0504^15503  1.0502^1.0501  1.0E+00  i (A km2)  Fig. 4.19 Polarization diagram of X100 and X80 in H2S-saturated NACE TM0177 "A" solution, T = 23 °C.  Fig. 4.20 Evans diagrams for X100 and X80 in NACE TM0177 "A" solution (H2S-saturated). E = potential, Log i = log value of current density, T = 23 °C.  79  Microstructural differences between the X100 and X80 might have caused the shift in the cathodic line and the resulting increase in corrosion rates. As can be seen in Fig. 4.1 and 4.2 for the microstructure of the two materials, the interfacial area between ferrite and the other phases is higher in the X100. Phase boundaries are areas of inhomogeneous corrosion potentials. As a result, ferrite at phase boundaries have a stronger driving force to corrode than ferrite in the homogeneous matrix because second phases acts as a local cathode to the surrounding ferrite. The shift of the cathodic line in the Evans diagram can then be attributed to an increased cathodic area in the X100. An increase in cathodic areas tends to shift the cathodic line further to the right, rather than upward, yielding higher corrosion potentials and higher corrosion rates.  Reformatskaya et al. (Ref. 57) showed that the corrosion rate of carbon steel was affected by corrosion potential inhomogeneity between ferrite and the other phases in the steel. The study showed that corrosion rate increases with an increase in phase boundaries between ferrite and second phases. These phase boundaries increase with increasing second phase grain sizes or by a higher dispersion of these phases in the steel.  The nonmetallic inclusion diameter difference between the two steels may have also played a role in the increased corrosion rate in the X100. Inclusion diameter in the X100 ranged between 6-9.5 pm while for the X80, inclusion diameter ranged between 2-4.5 p,m. Furthermore, the average inclusion diameter in the X100 steel, 8.0 wn, was higher than those in the X80, 3.5 [n. The effect of surface inclusions is associated with preferential dissolution of the surrounding metal matrix. Corrosion of ferrite at inclusion-  80  matrix interfaces develops more rapidly than ferrite in the matrix. Due to the higher dissolution rates in that area, a narrow corroded area forms around the inclusion and the contact between the inclusion and the matrix breaks forming a local pit at the inclusion site (Ref. 58). Because of the larger inclusion diameter in the X100, the inclusion-matrix interface is larger which leads to a higher corrosion rate than the X80. This is accomplished by providing larger cathodic area (inclusion) relative to any single nucleation site in the surrounding metal. Fig. 4.21 shows the corroded metal surface of X100 and X80 after polarization testing in H2S-saturated NACE TM-0177 "A" solution. For both steels, the surface consisted of randomly distributed pits. However, the corrosion pits in the X100 were bigger than those in the X80. X100 average pit diameter, d ot ,  was 0.36 mm while X80 average pit diameter was 0.27 mm. Therefore, corrosion pits  in X100 were 25% larger than those in X80. In addition, X100 had a corrosion pit density, p o t, of 2.13 pit/mm2 which was 17% more that the pit density in X80, 1.82 pit/mm 2 .  81  (a)  (b) Fig. 4.21 Pitting in (a) X80 and (b) X100 in H2S saturated NACE solution (SEM).  82  4.3 Proof ring testing:  4.3.1 Threshold stress 6 th determination for API-X100:  As seen in Table 4.4, fifteen X100 tensile specimens were grouped into five sets that were stressed at different percentages of the yield strength (80, 65, 50, 40 and 30%) in the NACE TM0177 "A" solution (saturated H2S) until breakage or a maximum test duration of 720 hours (1 month). %YS  Specimen  1 2 3 1 65 2 3 1 50 2 3 1 40 2 3 1 30 2 3 Total number of specimens = 15 80  API-X100 proof ring results Time to failure pH (hours) initial/final 2.7/2.9 26 2.7/2.8 9 2.7/2.8 24 2.7/3.1 45 2.7/3.1 68 51 2.7/3.1 2.7/4.0 --2.7/3.5 300 2.7/3.2 245 2.7/3.8 --2.7/3.8 --2.7/3.7 2.7/3.7 --2.7/3.7 --2.7/3.7 ---  Result fail fail fail fail fail fail pass fail fail pass pass pass pass pass pass  Overall result fail fail  fail pass pass  Table 4.4 Proof ring test results tested in NACE TM-0177 "A" (23 °C). Dashed lines (---) in the time to failure column indicate that the specimen survived the 720 (1 month) test duration with no cracking.  83  ^  The results from Table 4.4 are plotted in Fig. 4.22. It can be seen that the time to failure increased with a decrease in the applied load. In addition, the change of time to failure values between two consecutive applied stresses seemed to widen as the threshold stress value was approached. Since an applied stress of 50%YS generated a "fail" result while 40%YS generated a "pass" result, the threshold stress will be between these two points. Proof ring time to failure  90 ^ 80 70 - 60 50 • 40  0  30  0  cp  co  20 10 0 ^ ^ 0 100 200 300 400 500 600 700 800 900 1000 Time (hours)  Fig. 4.22 Proof ring time to failure graph for the X100 in H 2 S-saturated NACE TM-0177 "A" solution (dark = fail, clear = pass, dashed line = 720 hrs. max. test duration). Dashed line indicates the 720 hours (1 month) test duration).  The LOGEST function in Microsoft Exce1TM was used to generate an equation to fit the proof ring results. The LOGEST function calculated the exponential curve that fitted the data from the proof ring experiments and returned an array of values that described the curve. The generated equation was:  84  Time to failure (hours) = (3.41x10 5 ) x (0.875 %11s )^(4.11)  Equation 4.11 predicts that time to failure, in NACE TM-0177 "A", reaches 720 hours (1 month) when the applied stress is 46% of the yield strength or 343.1 MPa (49.7 ksi). Therefore, the threshold stress value (a th) for the X100 steel quoted here is 46% YS. As per equation 4.11, X100 loaded below 46%YS should always pass the NACE TM-0177 test. Implication of this conclusion is seen in specimens loaded to 40%YS and 30%YS in which all specimens passed the test.  4.3.2 Crack characterization:  Fig. 4.23 shows a longitudinal section for a failed tensile specimen. In all cracked specimens, the main failure mechanism was SSC cracking that propagated perpendicular to the applied stress. The SSC crack was transgranular as seen in Fig. 4.24. All samples that fractured into two parts exhibited the same fracture morphology. SSC initiated at the surface and propagated by brittle fracture perpendicular to the applied stress. The crack kept on propagating until the specimen could not hold the applied load any longer. At that stage, the specimen broke mechanically into two parts by ductile fracture.  85  (a)  (b) Fig. 4.23 Fracture face longitudinal section for 80%YS specimen. SSC = Sulfide Stress Cracking, DF = Ductile Failure.  86  Fig. 4.24 Longitudinal section of fracture face (80%YS specimen) showing transgranular crack propagation (light).  In addition to SSC, Hydrogen Induced Cracking (HIC) sites were observed as well. These HIC cracks were observed to propagate parallel to the rolling direction of the material and in some cases were connected to the main crack that caused failure, as seen in Fig. 4.25. In this study, the applied stresses were parallel to the rolling direction. Since HIC cracks propagate along the rolling direction, HIC cracks in this case were not affected by the applied stresses and its propagation was due to internal hydrogen pressure originating from inside the HIC crack. If the applied stresses were perpendicular to the rolling direction, applied stresses, in this case, would have an effect on HIC propagation, Fig. 4.26.  87  Fig. 4.25 Longitudinal section of the fracture face of an X100 specimen loaded at 65%YS showing a HIC crack connected to the SSC crack (light).  I MC  Applied stress  HIC  --4111■1111111111m----  Rolling direction  Rolling direction  Fig. 4.26 Effect of applied stress and rolling direction on HIC. 88  4.3.3 Mechanistic insights:  Corrosion pits played an important role in SSC nucleation on the metal surface. Fig. 4.27 shows a SSC cracking site at the metal surface. The stress concentration in the pit gave rise to SSC crack nucleation and subsequent propagation.  Fig. 4.27 Longitudinal section of X100 specimen loaded at 65% YS showing SSC crack and surface inclusions (dark regions) (light).  In addition to the main SSC crack at the metal surface, nucleation of microcracks inside the metal was observed. SSC crack nucleation inside the metal produced small brittle cracks in the matrix. Crack propagation likely occurred by bursts in which the most favourable oriented microcrack connected to the main crack before the rest. Fig. 4.28  89  shows the main SSC propagating in the metal. Microcracks can also be seen near the crack tip of the main crack.  Fig. 4.28 Longitudinal section of an X100 specimen loaded at 65% YS showing main SSC crack (right of image) and small microcracks (left of image) (light).  The area connecting the main crack to a microcrack fractured in a ductile fashion. When a microcrack is close enough to the main SSC crack, the stress intensity on the area connecting the two brittle cracks exceeds the material's tensile strength leading to mechanical fracture. Therefore, the crack growth can be related to how easy a microcrack can reach the main crack under a given stress intensity factor (K 1 ) and only those cracks that connect with the main crack are able to grow. As a result, the SSC crack in the X100 steel was mostly of a brittle nature but contained segments of ductile crack growth. Fig. 4.29 shows an SEM image taken from a fractured surface that shows two brittle sections connected by a ductile region.  90  x 50 0 0000 20 k V 1 00.0 m Fig. 4.29 SSC brittle regions connected by a ductile section found in the fracture face of X100 specimen loaded at 80% YS (SEM). B = brittle (SSC), D = ductile.  HIC cracks found in the tested specimens were nucleated at banded martensite-ferrite interfaces and non-metallic inclusions and propagated along the rolling direction parallel to the applied tensile stress through the softer ferrite phase. Fig. 4.30 shows HIC cracks found in specimens stressed at 50% of the yield strength. After hydrogen accumulation, internal pressure in these traps increased due to the reaction between hydrogen atoms to form molecular hydrogen leading to HIC folination.  91  Fig. 4.30 Transverse section of an X100 specimen loaded at 50% YS showing HIC cracks (light).  The existence of HIC cracks can affect the material's cracking susceptibility in sour environments. Residual or applied stresses can increase the stress field surrounding the cracks. This generates localized yielding which can produce a second crack perpendicular to the tensile stress direction.  Hydrogen can diffuse rather quickly into the steel matrix. Koh et al. (Ref. 41) measured the diffusivity of hydrogen, at 25 °C, in several HSLA steels, X52 to X80, with varying microstructures, ferrite-pearlite, acicular ferrite and ferrite-bainite, using the NACE TM0177 "A" solution (the same solution used in the polarization and proof ring tests). The study showed that the lowest hydrogen diffusivity in the tested HSLA steels was 3.5  92  x 10 -6 cm2/sec. Diffusivity of hydrogen is closely related to the trapping sites provided by various defects such as vacancies, inclusions and precipitation in the steel.  In order to demonstrate the rate of hydrogen diffusion into the test specimen, a GurneyLurie chart for a cylindrical shape was used (Ref. 59). A chart for a long cylinder was used so the effect of the cylinder ends can be neglected. The Gurney-Lurie charts represent solutions for unsteady state mass transfer problems for several simple shapes by relating the relative time (XD) to the unaccomplished concentration change (Y). XD and Y are dimensionless ratios described as follows:  Y = unaccomplished concentration change = (CH s — CH) / (CHs CH0)  ^  (4.12)  XD = relative time = (D • t) / x1 2^(4.13)  Where: CHs concentration of hydrogen at the surface. CH0 = concentration of hydrogen at the specimen center at t = O. CH = unknown hydrogen concentration at desired t and x. D = hydrogen diffusivity. x i = specimen radius. t = time.  When t = 0, CH 0 = O. As a result, Equation 4.12 will be rearranged as follows:  93  (4.14)  CH = (1 —^CHs^  In the proof ring tests conducted in this study, x 1 = 0.318 cm. In addition, the earliest time to failure was 9 hours, Table 4.4, so t = 32400 sec. (9 hrs.). It is assumed that the test solution is well mixed. Therefore, for a hydrogen diffusivity value of 3.5 x 10 -6 cm2 /sec:  XD = (D t) / x 1 2 = (3.5 x 10 -6 32400) / (0.318 2 ) = 1.12^(4.15)  At the specimen center, the calculated XD will yield Y = 0.003. Therefore, after 9 hours:  CH = (1 — 0.003) CHs CH = 0.997 CHs^CH = 99.7% CHs^(4.16)  It is shown that after 9 hours, the hydrogen concentration at the specimen center was 99.7% of the hydrogen concentration at the surface. To further illustrate the fast diffusion of hydrogen through the tested specimens, the time that hydrogen will need to reach the center was examined by determining the value of XD at Y = 1.0 (Y value below which CH at the center will be above 0). At Y = 1.0, XD = 0.09. Applying Equation 4.13:  XD = 0.09 = (3.5 x 10 -6 t) / (0.318 2 )^t = 2592 seconds = 43.2 minutes^(4.17)  It is seen that hydrogen will reach the specimen center after 43.2 minutes. It should be noted that the specimen radius was very small (3.18 mm). This explains why little time was required for hydrogen to reach the specimen center line.  94  4.3.4 Comparison between X100 and other HSLA steels:  The X100 microstructure, mechanical and electrochemical properties played an important role in its susceptibility to cracking in sour environments. As discussed in section 4.3.1, the X100 exhibited a threshold stress,  ath,  of 46% of yield strength. Tests were conducted  in the early eighties by Borik et al., (Ref. 6) to investigate the possibility to produce a future X100 grade steel (at that time) as a candidate material for deep sea pipelines. The X100 in this study was developed by hot forging and hot rolling followed by quenching in agitated water and tempering at 600 °C. This material doesn't resemble current X100 steels because no controlled rolling and microalloying by Ti, V and Nb were implemented. Tests in the NACE TM0177 "A" solution revealed that the material was highly susceptible in sour media and generated very low threshold stress values (cr th ), 25% YS. Improving ath could only be attained by increasing the tempering temperature to 650 °C but this has lead to a decrease in strength that dropped from the X100 level to the X90 level.  Even though the generated at h value in this study is considered low, it can be seen that controlled rolling and microalloying used on the tested X100 have generated a higher  ath  relative to X100 generated by older methods of production (hot rolling with no microalloying additions). With hot rolling, the final product is attained by accelerated cooling (agitated water) and subsequent tempering of martensite. This process can be used to generate lower HSLA grades. However, when a high grade HSLA, such as X100, is needed, higher alloying additions and faster cooling rates are needed to generate the  95  targeted mechanical properties. The generated microstructures contain a lot of hard phases (e.g. martensite) that can act as hydrogen traps and increase SSC susceptibility. On the other hand, controlled rolling and microalloying with carbonitride forming elements such as V, Nb and Ti will provide an additional source of strengthening through grain refinement and precipitation hardening. As a result, mechanical targets can be met with less aggressive cooling rates and without the need for heat treatment (Ref. 2).  Many HSLA materials were tested in the NACE TM0177 "A" solution for their threshold stress values. Pontremoli et al. (Ref. 60) evaluated a bainitic-ferrite X80 steel in the NACE TM0177 "A" solution. The X80 generated a 6 th value of 65%YS. Carneiro et al. (Ref 49) showed that a banded ferrite-pearlite API-X60 had a 6 t h value of 60% YS. Quenching and tempering treatment for the same material raised 6 th to 70% YS by eliminating the banded structure and promoting a more homogeneous one. Hardness values can not be taken as the sole criteria to determine material susceptibility to SSC as other factors such as microstructure can play an important role in SSC susceptibility. Pontremoli et al. (Ref 60) showed that 6 th for an acicular ferrite X70 (78% YS) was higher than an X65 (69% YS) steel with a ferrite-pearlite microstructure. In addition, the degree of inclusion density can also be a factor in SSC susceptibility. Ume et al. (Ref. 30) showed that 6 t h values for X70 steels dropped from 80%YS to 50%YS due to added sulfur content which gave raise to MnS inclusions.  The X100 microstructure was one of the main causes for the low 6 t h value. A banded structure provides high hydrogen trapping site density in front of the main crack tip and is  96  more susceptible to SSC than homogenous microstructures (Ref. 49). Homogeneous microstructures present dispersed and relatively smaller sites for hydrogen trapping hence, reaching the critical hydrogen concentration for SSC microcrack formation is more difficult. In addition, bands in inhomogeneous microstructures act as nucleation sites for HIC cracks (Ref. 48). HIC can connect with propagating SSC cracks which lead to failure in sour environments at lower applied stresses. In addition to the susceptible microstructure, the higher corrosion rates experienced in the X100 steel is evidence that more hydrogen was available from the corrosion reaction. Hydrogen can migrate into the steel and cause embrittlement after it accumulates in hydrogen traps.  4.4 Implications of results on X100 susceptibility:  The results generated in this study and comparison with older generation HSLA steels revealed that X100 is susceptible in H2S environments. The X100 high susceptibility is demonstrated by its high yield strength value, susceptible microstructure, high corrosion rates and low threshold stress value in the NACE TM0177 "A" environment. Investigating the reasons behind the low SSC susceptibility should open doors for future improvements on the X100 in order to generate a more resistant material in H2S environments.  Manufacturing an X100 grade steel with acceptable resistance to SSC might be possible if the areas that lead to its high susceptibility in this study are improved. This can be achieved by generating methods to produce a more homogeneous microstructure that will  97  minimize local hydrogen build-up that might lead to SSC nucleation. In addition, minimizing the phase boundaries in the steel will depress corrosion rates which will minimize pit formation and reduce the intensity of hydrogen propagation into the steel. Another way to improve SSC resistance is to better control nonmetallic inclusions by reducing inclusion sizes and eliminating any elongated inclusions. A material with less nonmetallic inclusions will have less nucleation sites for corrosion pits on the surface and less hydrogen traps in the metal body.  98  Chapter 5  5 SUMMARY AND CONCLUSION:  In this study, API-X100 HSLA steel was tested for its SSC susceptibility in H 2 S environments. The application of HSLA steels in oil and gas transportation pipelines has increased in the past few years because of an increase in global energy demand. The X100 is a newly developed HSLA steel that was designed to meet the requirement for pipelines with higher mechanical limits to withstand the increasing line pressures. To assure safe and reliable performance, it is necessary to investigate the X100 SSC cracking susceptibility in H2S environments. The need to investigate SSC susceptibility is raised from the material's higher strength and hardness values that might make it more susceptible to cracking than older generation HSLA steels. Investigating SSC susceptibility will open doors for future improvements on the X100 in order to generate a more resistant material in H 2 S environments.  Electrochemical and proof ring tests were conducted on the X100 steel using the NACE TM0177 "A" test solution. X80, an older generation HSLA steel, was used for comparison in the microstructural and electrochemical examinations. Uniaxial (constant load) SSC testing was conducted on the X100 to determine the material threshold stress value (a t h) in H2S environment. Test results and comparison with older generation HSLA steels showed that the X100 displayed low resistance to SSC.  99  The steel had a high corrosion rate in the NACE TM0177 "A" test solution. The effect of corrosion on SSC susceptibility was achieved by generating corrosion pits that will act as initiation sites for SSC cracks on the metal surface. In addition, higher corrosion rates will provide more hydrogen as a result of the cathodic reaction on the metal surface. These hydrogen atoms will migrate into the steel and cause hydrogen embrittlement.  Addition of H 2 S to the NACE TM-0177 "A" solution increased the corrosion rate of X100 from 51.6 to 96.7 mpy. The effect of H 2 S was similar on the X80. The corrosion rate for the X80 when H2S was added to the test solution increased from 45.2 to 80.2 mpy. Addition of H 2 S accelerated the anodic reaction by forming a surface catalyst (FeHS ad s ) on the metal surface which shifted the anodic line polarization curve to more -  current densities. In addition, decreased pH values after H 2 S addition increased the cathodic half cell potential and shifted the cathodic polarization curve to more current densities. The increase cathodic and anodic current densities led to the increase in the corrosion current density, i cor,. In saturated H 2 S NACE TM0177 "A" solutions, the X100 corrosion rate was higher than the X80 steel by 20%. The effect of microstructure was the likely cause of this result. The X100 microstructure exhibited higher interfacial area between phases and larger nonmetallic inclusions. This generated more areas with inhomogeneous corrosion potentials and a stronger driving force for corrosion. Increased areas of carbon rich regions and larger non metallic inclusion diameters acted as an increased cathodic area on the X100 surface. This shifted the cathodic line in the Evans diagram to the right which increased i con- in the X100 when compared to X80.  100  The X100 generated a low threshold stress value, a th, of 46% YS, 343.2 MPa (49.7 ksi). The main failure mechanism in the proof ring specimens was SSC cracking. SSC nucleated at corrosion pits in the metal surface. Nucleation of SSC microcracks inside the metal was observed. SSC crack nucleation inside the metal produced small brittle cracks in the matrix. SSC crack propagation occurred by bursts in which the most favourable oriented microcrack connected to the main crack.  In addition to SSC, HIC was observed as well. HIC cracks were observed to propagate parallel to the rolling direction of the material and in some cases were connected to the main crack that caused failure. HIC cracks nucleated at banded martensite-ferrite interfaces and propagated along the rolling direction parallel to the applied tensile stress through the softer ferrite phase.  When compared to older HSLA grades, the X100 tested in this study had a high SSC susceptibility. SSC susceptibility was caused by the X100 high corrosion rate in the H 2 Ssaturated TM0177 "A" test solution. The increased corrosion rate formed more corrosion pits that acted as crack initiation sites on the metal surface and provided more hydrogen that migrated into the steel. The X100 inhomogeneous microstructure provided a high density of hydrogen traps in front of the main crack tip which promoted SSC microcrack formation inside the metal. Microcracks in the metal body connected with the main crack tip that originated from corrosion pits which assisted SSC propagation.  101  The X100 microstructure played a role in SSC susceptibility. Inhomogeneous microstructure with bands of carbon rich phases acted as SSC and HIC initiation sites inside the metal. Crack nucleation formed more easily in the susceptible microstructures and linkage of microcracks inside the metal with the main crack tip was less difficult.  Manufacturing an X100 grade steel with acceptable susceptibility to SSC might be possible if the areas that lead to its high susceptibility in this study are improved. Methods to produce a homogeneous X100 microstructure need to be generated in order to minimize hydrogen build-up, in front of the main crack tip, that might lead to SSC microcrack nucleation in the metal body. In addition, microstructures with minimized phase boundaries are needed to reduce corrosion rates which will minimize pit formation and reduce hydrogen propagation into the steel. Finally, better control of nonmetallic inclusions by reducing inclusion sizes and eliminating any elongated inclusions are needed in order to produce a material with less nucleation sites for corrosion pits on the surface and less hydrogen traps in the metal body.  For future academic research in the area of SSC and H 2 S corrosion in HSLA steels, the following is recommended: • Effect of possible microstructures for new generation HSLA steels on SSC susceptibility. Small differences in microstructure will produce different response in SSC media. Factors such as non-metallic inclusions, banding, present phases should be investigated.  102  •  Effect of alloying elements on H2S corrosion and SSC cracking. Optimum alloying additions that can minimize H 2 S corrosion and inclusion morphology and size will have a positive effect on SSC resistance.  •  Threshold stress intensity factors (Kissc) for new generation HSLA steels. Kissc examinations for new HSLA steels are needed to determine SSC susceptibility in the presence of existing cracks or corrosion pits. Kissc determination will also facilitate maximum stress calculations of corroded pipeline materials.  •  Effect of test variables on Kissc values. Kissc determination using double cantilever beam (DCB) testing is sensitive to various experimental factors such as temperature, applied load, wedge thickness and even test vessel volumes. Research should be conducted to verify the exact effect of these variables.  •  Effect of temperature on H 2 S corrosion for new HSLA steels. Even though SSC is most severe at ambient temperatures, determining the effect of temperature for new HSLA steels is practically useful in the industry. Many pipelines are operated at temperatures below or above ambient temperatures and this will affect the materials response in SSC environments.  103  REFERENCES:  1. ASM Handbook. Volume 1: Properties and Selection: Iron Steels and HighPerformance Alloys. ASM international, Materials Park OH, 1993. 2. H. G. Hillenbrand et al."Development and Production of High Strength Pipeline Steels". Niobium 2001, December 02-05. Orlando Florida, 2001. 3. Alan F. Liu. Structural Life Assessment Methods. 2' nd edition. ASM international, Materials Park OH, 1998. 4. NACE Test Method TM177-96. "Laboratory Testing of Materials for Resistance to Specific Forms of Environmental Cracking in H2S Environments". NACE Houston TX, USA, 1996. 5. R. Denys. Pipeline Technology: Volume I. 1 ' t edition. Elsevier Science B.V, Amsterdam, 2000. 6. F. Borik et al. "SSC Resistance of Electron Beam Welded, Q&T, X60, X80 and X100 Line Pipe Steels". Corrosion 82. March 22-26, Houston TX, 1982. 7. W. F. Smith. Structure and Properties of Engineering Alloys. 2' nd Edition, McGraw Hill, New York NY, 1993. 8. A. A. Groni et al."Austenite Transformation and Age Hardnening of HSLA-80 and ULCB Steels". Journal of Materials Processing Technology, 155, 2004. pp: 1513-1518. 9. N. Ishikawa et al." High Performance UOE Linepipes". JFE Technical Report No. 7, January 2006, pp: 20-26.  104  10. Y. I. Komizo. "Progress in Structural Steels for Bridge and Linepipes". Transactions of JWRI. vol. 35. no. 1. 2006, pp: 1-7. 11. J. Bauer et al." Microstructure and Properties of Thermomechanical Controlled Processing Steels for Linepipe Applications". Iron Making and Steel Making, vol. 32, no. 4, 2005, pp: 325:330 12. S. K. Das et al." The Effect of Variation of Microstructure on Fracture Mechanics Parameters of HSLA-100 steel". Materials Science and Engineering A431, 2006, pp: 68-79. 13. F. Xiao et al." Acicular Ferrite Microstructure of a Low Carbon Mn-Mo-Nb Microalloyed Pipeline Steel". Materials Characterization, vol. 54, 2005, pp: 305314. 14. M. Zhao et al." Continuous Cooling Transformation of Undeformed and Deformed Low Carbon Pipeline Steels". Materials Science and Engineering A355, 2003, pp: 126-136. 15. H. G. Hillenbrand et al." Development of Line Pipe in Grade Up to X100". Paper 6, 1 1 'th EPRG/PRCI Biannual Joint Technical Meeting On Pipeline Research, Arlington, 1997. 16. M. Okatsu et al." Metallurgical and Mechanical Features of X100 Linepipe Steel". Proceeding of OMAE 1997, Yokohama, Japan, April 13-17, 1997, pp: 119-124. 17. L. Barsanti et al." Possible Use of New Materials for High Pressure Linepipe Construction: an Opening on X100 Grade Steel". Proceedings of IPC, 2004.  105  International Pipeline Conference, October 4-8, Calgary, Alberta, Canada, 2004, pp: 287-298. 18. G. Demofonti et al." Evaluation of the Suitability of X100 Steel Pipes for High pressure Gas Transportation Pipelines by Full Scale Tests". Proceedings of IPC 2004 International Pipeline Conference, October 4-8, Calgary, Alberta, Canada, 2004, pp: 1685-1692. 19. V. Yu Chernov et al. "On the Causes of Corrosion Fracture on Industrial Pipelines", Materials Science, vol. 38, no. 6, 2002, pp: 114-147. 20.0. I. radkevych et al. "Influence of Hydrogen Sulfide on Serviceability of Materials of Gas Field Equipment", Materials Science, vol. 37, no. 2, 2001, pp: 319-332. 21. S. U. Koh et al." Effect of Alloying Elements on the Susceptibility to Sulfide Stress Cracking", Corrosion Science, vol. 60, no. 3, 2004, pp: 262-274. 22. A. D. Brasunas et al. Corrosion Basics: An Introduction. 1 ' st Edition. NACE, Houston TX, 1984. 23. D. A. Jones. Principles and Prevention of Corrosion. 2' nd edition, Prentice-Hall , Upper Saddle River NJ, USA, 1996. 24. ASTM Standard G5-94. "Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements". ASTM International. West Conshohocken PA, USA, 1994. 25. J. L Albarran et al. "Potential Drop Measurements During Stress Corrosion of API-X80 Pipeline Microalloyed Steel". Paper 312, Corrosion 94, Feb. 27 - March 4, Baltimore MD, 1994.  106  26. ASM Handbook. Volume 13A: Corrosion. ASM international, Materials Park OH, 2003. 27. S. Serna et al. "Crack Growth in Microalloyed Pipeline Steels for Sour Gas Transport". Journal of Material Engineering Performance, vol. 14, no. 2, April 2005, pp: 224-228. 28. API Recommended Practice. Damage Mechanisms Affecting Fixed Equipment in the Refining Industry. l' st edition, American Petroleum Institute, Washington DC, 2003. 29. M. Elboujdaini et al."Studies on Inhibition of Hydrogen-Induced Cracking of Linepipe Steels". Corrosion, vol. 62, no. 1, 2006, pp: 29-34. 30. K. Ume et al. "Initiation and Propagation Morphology of Sulfide Stress Corrosion Cracking at Welds on Line Pipe Steels". paper 240, Corrosion 85, 25-29 March, Boston, Massachusetts,1985. 31. H. F. Lopez et al. "Microstructural Aspects of Sulfide Stress Cracking in an API X80 Pipeline Steel". Metallurgical and Materials Transactions A, vol. 27A, November 1996, pp: 3601-3611. 32. M. Elboujdaini et al. "Studies on Inhibition of Hydrogen-Induced Cracking of Linepipe Steels". Corrosion, vol. 62, no. 1, 2006, pp: 29-34. 33. W. D. Callister. Materials Science and Engineering: An Introduction. 3'rd edition, Willey, Hoboken NJ, 1994. 34. ASM Handbook. Volume 11: Failure analysis and Prevention, ASM international, Materials Park OH, 1992.  107  35. D. Broek. Elementary Engineering Fracture Mechanics. 4'th Edition, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1986. 36. H. Tada. The Stress Analysis of Cracks. Del Research Corp, Hellertown PA, 1973. 37. H. Asahi et al. "Sulfide Stress Cracking Resistance Evaluation Methods for Steels Used in Oil Field Environments: Features and Problems". Paper 29, Corrosion 91, March 11-15, Cincinnati, OH, 1991. 38. NACE Standard MR175. "Materials for Use in H2S Containing Environments in Oil and Gas Production", NACE, Houston TX, USA 2001. 39. J. L Albarran et al. "Corrosion and Cracking Behaviour in an API X-80 Steel Exposed to Sour Gas Environments". Corrosion, Vol. 58, no. 9, 2002, pp: 783792. 40. D. Hardie et al." Hydrogen Embrittlement of High Strength Pipeline Steels". Corrosion Science, vol. 48, 2006, pp. 4378-4385. 41. S. U. Koh et al." Effect of Line Pipe Steel Microstructure on Susceptibility to Sulfide Stress Cracking". Corrosion Science, vol. 60, no. 3, 2004, pp: 244-253. 42. G. Echanniz et al." The Effect of Microstructure on the Kissc Low Alloy Carbon Steels". Paper 120, Corrosion 98, March 22-27, San Diego CA, 1998. 43. M. Zhao et al." Role of Microstructure on Sulfide Stress Cracking of Oil and Gas Pipeline Steels". Metallurgical and Materials Transactions A, vol. 34A, May, 2003, pp: 1089-1096. 44. J.C. Charbonnier et al." SSC of High Strength Modified Cr-Mo Steels". Metallurgical and Materials Transactions A, vol. 16A, no. 935, 1985, pp. 935-44.  108  45. M. Diaz-fuentes et al." Analysis of Different Acicular ferrite Microstructures in Low Alloy-Carbon Steels by Electron Backscattered Diffraction: Study of Their Toughness Behaviour", Metallurgical and Materials Transactions A, vol. 34A, November, 2003, pp. 2505-2516. 46. M. Zhao et al." Investigation on the H2S-resistance behaviours of Acicular Ferrite and Ultrafine Ferrite". Materials Letters, vol. 57, November, 2002, pp. 141-145. 47. J. L Albarran et al. "Crack Growth in a Welded Microalloyed Steel under Sulfide Stress Cracking Conditions", Journal of Materials Engineering and Performance, vol. 7, no. 6, December, 1998, pp: 777-783. 48. J. L. Albarran et al. "Stress Corrosion Cracking Response of Microalloyed Pipeline Steels Exposed to Pressurized Sour Gas Environments". Paper 3447, Corrosion 2003, March 16-20, San Diego CA, 2003. 49. R. A. Carneiro et al."The influence of Chemical Composition and Microstructure of API Linepipe Steels on Hydrogen Induced Cracking and Sulfide Stress Cracking". Materials Science and Engineering A357, 2003, pp: 104-110. 50. ASTM Standard E8-04, Standard Test Method for Tension Testing of Metallic Materials, ASTM, West Conshohocken PA, USA, 2004 51. R. A. Farrar et al." Acicular Ferrite in carbon Manganese Weld Metals: An Overview", Journal of Materials Science, vol. 22, no. 22, 1987, pp: 3812-3820. 52. M. Nuspl et al." Qualitative and Quantitative Determination of Micro-Inclusions by Automated SEM/EDX Analysis", Annual Bional Chem., 379, 2004, pp. 640645.  109  53. L. E. K. Holappa et al." Inclusion Control in High Performance Steels", Journal of materials Science Technology, vol. 53, 1995, pp. 177-186. 54. H. V. Atkinson et al. - Characterization of Inclusion in Clean Steels: A Review Including the Statistics of Extreme Methods - , Progress in Materials Science, vol. 48, 2003, pp. 457-520. 55. A. V. Dub et al." Nonmetallic Inclusion in Low-Alloy Tube Steel", Metallurgist, vol. 49, no. 4, 2005, pp. 138-148. 56. X. L. Cheng el al." Corrosion of Iron in Acid Solutions with Hydrogen Sulfide". Corrosion, vol. 54, no. 5, May, 1998, pp: 369-376. 57.1. I. Reformatskaya et al." The Effect of Nonmetallic Inclusions and Microstructure on Local Corrosion of Carbon and Low-alloyed Steels", Protection of Metals, vol. 40, no. 5, 2004, pp: 498-504. 58. Y. Wang et al." Role of Nonmetallic Inclusions in Fatigue, Pitting and Corrosion Fatigue", Corrosion Science, vol. 52, no. 2, 1996, pp: 92-101. 59. J. Welty et al. Fundamentals of Momentum, Heat and Mass Transfer. 3'" 1 edition. John Willey and Sons, New York NY, 1984. 60. M. Pontremoli et al." Composition, Microstructure and Properties of Pipeline Steels with High HIC and SSCC Resistance". XXII International Metallurgy Congress: Innovation for Quality, 17-19 May, 1988, Bologna, Italy, pp: 15431561.  110  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0066205/manifest

Comment

Related Items