Nace mr0175 certified user my reading 7 general

Understanding NACE MR0175-Carbon Steel Written Exam G General lR Reading. di

Reading 7

Fion Zhang/ Charlie Chong 3rd Nov 2017

Fion Zhang/ Charlie Chong

Oil And Gas Production Industry

Fion Zhang/ Charlie Chong

Oil And Gas Production Industry

Fion Zhang/ Charlie Chong

Oil And Gas Production Industry

Fion Zhang/ Charlie Chong

Fion Zhang/ Charlie Chong

过五关斩六将

Fion Zhang/ Charlie Chong

过五关斩六将

Fion Zhang/ Charlie Chong

NACE MR0175 MR0175-Carbon Carbon Steel Written Exam NACE-MR0175-Carbon Steel -001 Exam Preparation Guide May 2017

Fion Zhang/ Charlie Chong

NACE MR0175-Carbon Steel W i Written Exam E NACE-MR0175-Carbon Steel -001 Exam Preparation Guide May 2017

Introduction The MR0175-Carbon Steel written exam is designed g to assess whether a candidate has the requisite knowledge and skills that a minimally qualified MR0175 Certified User- Carbon Steel must possess. The exam comprises 50 multiple-choice questions that are based on the MR0175 Standard (Parts 1 and 2).

multiple-choice p Fion Zhang/ Charlie Chong

https://www.naceinstitute.org/uploadedFiles/Certification/Specialty_Program/MR0175-Carbon-Steel-EPG.pdf

EXAM BOK Suggested Study Material  NACE MR0175/ISO 15156 Standard (20171015-OK)  EFC Publication 17  NACE TM0177  NACE TM0198 NACE TM0316 Books  Introductory Handbook for NACE MR0175

Fion Zhang/ Charlie Chong

Fion Zhang/ Charlie Chong

Fion Zhang/ Charlie Chong

Reading 1:

Fion Zhang/ Charlie Chong

http://www.emerson.com/documents/automation/140706.pdf

Reading 1: Sulfide Stress Cracking --NACE NACE MR0175 MR0175-2002, 2002 MR0175/ISO 15156 Emerson Experiences

Fion Zhang/ Charlie Chong

http://www.emerson.com/documents/automation/140706.pdf

The Details NACE MR0175, MR0175 “Sulfide Sulfide Stress Corrosion Cracking Resistant Metallic Materials for Oil Field Equipment” is widely used throughout the world. In late 2003, it became NACE MR0175/ ISO 15156, “Petroleum and Natural Gas Industries - Materials for Use in H2S-Containing Environments in Oil and Gas Production.”

These standards specify the proper materials, heat treat conditions and strength levels required to provide good service life in sour gas and oil environments. NACE International (formerly the National Association of Corrosion Engineers) is a worldwide technical organization which studies various aspects of corrosion and the damage that may result in refineries, chemical plants, water systems and other types of industrial equipment.

MR0175 was first issued in 1975, but the origin of the document dates to 1959 when h a group off engineers i iin W Western t C Canada d pooled l d th their i experience i iin successful handling of sour gas. The group organized as a NACE committee p 1B163,, “Recommendations of Materials or and in 1963 issued specification Sour Service.” In 1965, NACE organized a nationwide committee, which issued 1F166 in 1966 and MR0175 in 1975.

Fion Zhang/ Charlie Chong

Revisions were issued on an annual basis as new materials and processes were added. added Revisions had to receive unanimous approval from the responsible NACE committee.

Fion Zhang/ Charlie Chong

In the mid-1990’s, the European Federation of Corrosion (EFC) issued 2 reports closely related to MR0175; Publication 16 16, “Guidelines on Materials Requirements for Carbon and Low Alloy Steels for H2S H2S-Containing Containing Environments in Oil and Gas Production” Production and Publication 17, “Corrosion Resistant Alloys for Oil and Gas Production: Guidance on General Requirements and Test Methods for H2S Service.” EFC is located in London, England.

The IInternational Th t ti lO Organization i ti ffor Standardization St d di ti (ISO) is i a worldwide federation of national standards bodies from more than 140 countries countries. One organization from each country acts as the representative for all organizations in that country. The American National Standards Institute (ANSI) is the USA representative in ISO. Technical Committee 67, “Materials, Equipment and Offshore Structures for Petroleum, Petrochemical and Natural Gas I d t i ” requested Industries,” t d th thatt NACE bl blend d th the diff differentt sour service i documents into a single global standard.

Fion Zhang/ Charlie Chong

This task was completed in late 2003 and the document was issued as ISO standard, standard NACE MR0175/ISO 15156 15156. It is now maintained by ISO/TC 67, Work Group 7, a 12-member “Maintenance Maintenance Panel�

Panel and a 40-member 40 member Oversight Committee under combined NACE/ISO control. The three committees are an international group of users, manufacturers and service providers.

Membership is approved by NACE and ISO based on technical knowledge and experience. Terms are limited. Previously, some members b on th the NACE T Task kG Group had h d served d ffor over 25 years.

Fion Zhang/ Charlie Chong

NACE MR0175/ISO 15156 is published in 3 volumes.  Part 1: General Principles for Selection of Cracking Cracking-Resistant Resistant Materials  Part 2: Cracking Cracking-Resistant Resistant Carbon and Low Alloy Steels Steels, and the Use of Cast Irons  Part 3: Cracking-Resistant CRA’s (Corrosion-Resistant Alloys) and Other Alloys

Fion Zhang/ Charlie Chong

NACE MR0175/ISO 15156 applies only to petroleum production, drilling gathering and flow line equipment and field processing drilling, facilities to be used in H2S bearing hydrocarbon service. In the past MR0175 only addressed sulfide stress cracking (SSC) past, (SSC). In NACE MR0175/ISO 15156, however, but both SSC and chloride stress corrosion cracking (SCC) are considered.

Question?

D Does SCC always l d due tto th the presence off Cl- ?

See http://oilfieldwiki.com/wiki/Stress_corrosion_cracking

Fion Zhang/ Charlie Chong

While clearly intended to be used only for oil field equipment, industry has applied MR0175 in to many other areas including refineries refineries, LNG plants, pipelines and natural gas systems. The judicious use of the document in these applications is constructive and can help prevent SSC failures wherever H2S is present. Saltwater wells and saltwater handling facilities are not covered by NACE MR0175/ ISO 15156. These are covered by NACE Standard RP0475, “Selection of Metallic Materials to Be Used in All Phases of Water Handling for I j ti into Injection i t Oil-Bearing Oil B i F Formations.” ti ”

Fion Zhang/ Charlie Chong

When new restrictions are placed on materials in NACE MR0175/ ISO 15156 or when materials are deleted from this standard, standard materials in use at that time are in compliance. This includes materials listed in MR0175-2002 MR0175 2002, but not listed in NACE MR0175/ISO 15156. However, if this equipment is moved to a different location and exposed to different conditions, the materials must be listed in the current revision. Alternatively, successful use of materials outside the limitations of NACE MR0175/ISO 15156 may be b perpetuated t t d by b qualification lifi ti ttesting ti per th the standard. t d d Th The user may replace materials in kind for existing wells or for new wells within a given field if the environmental conditions of the field have not changed.

Fion Zhang/ Charlie Chong

New Sulfide Stress Cracking Standard for Refineries Don B D Bush, h P Principal i i lE Engineer i - Materials, M t i l att Emerson E P Process Management Fisher Valves, is a member and former chair of a NACE task group that has written a document for refinery applications, NACE MR0103. The title is “Materials Resistant to Sulfide Stress Cracking g in Corrosive Petroleum Refining g Environments.� The requirements of this standard are very similar to the pre-2003 MR0175 for many materials. When applying this standard, t d d th there are changes h tto certain t i kkey materials t i l compared d with ith NACE MR0175-2002.

Fion Zhang/ Charlie Chong

Responsibility It has always been the responsibility of the end user to determine the operating conditions and to specify when NACE MR0175 applies This is now emphasized more strongly than ever in NACE applies.

MR0175/ISO 15156.  The manufacturer is responsible for meeting the metallurgical requirements of NACE MR0175/ISO 15156.  It is the end user’s responsibility to ensure that a material will be satisfactory ti f t in i the th intended i t d d environment. i t

Fion Zhang/ Charlie Chong

Some of the operating conditions which must be considered include pressure temperature, pressure, temperature corrosiveness corrosiveness, fluid properties properties, etc etc. When bolting components are selected, the pressure rating of flanges could be affected affected.

It is always the responsibility of the equipment user to convey the environmental conditions to the equipment supplier, particularly if the equipment will be used in sour service.

Th various The i sections ti off NACE MR0175/ISO 15156 cover th the commonly available forms of materials and alloy systems. The requirements for heat treatment treatment, hardness levels levels, conditions of mechanical work and post-weld heat treatment are addressed for each form of material. Fabrication techniques, bolting, platings and coatings are also addressed.

Fion Zhang/ Charlie Chong

Applicability of NACE MR0175/ISO 15156 Low concentrations L t ti off H2S (<0.05 ( 0 05 psii (0 (0,3 3 kP kPa)) H2S partial ti l pressure) and low pressures (<65 psia or 450 kPa) are considered outside the scope of NACE MR0175/ISO 15156 15156. The low stress levels at low pressures or the inhibitive effects of oil may give satisfactoryy p performance with standard commercial equipment. q p Many users, however, have elected to take a conservative approach and specify compliance to either NACE MR0175 or NACE MR0175/ISO 15156 any titime a measurable bl amountt off H2S iis present. The decision to follow these specifications must be made by the user based on economic impact impact, the safety aspects should a failure occur and past field experience. Legislation can impact p the decision as well. Such jjurisdictions include;; the Texas Railroad Commission and the U.S. Minerals Management Service (offshore). The Alberta, Canada Energy Conservation Board recommends d use off the th specifications. ifi ti

Fion Zhang/ Charlie Chong

Facilities operating at a total absolute pressure below 0 0.45 45 MPa (65 psi)

Fion Zhang/ Charlie Chong

Figure 1. Photomicrograph Showing Stress Corrosion Cracking

Fion Zhang/ Charlie Chong

Basics of Sulfide Stress Cracking (SSC) and Stress C Corrosion i Cracking C ki (SCC) SSC and SCC are cracking processes that develop in the presence of water, corrosion and surface tensile stress. stress It is a progressive type of failure that produces cracking at stress levels that are well below the materialâ&#x20AC;&#x2122;s tensile strength. The break or fracture appears brittle, with no localized yielding, plastic l ti d deformation f ti or elongation. l ti R Rather th th than a single i l crack, k a network t k off fifine, feathery, branched cracks will form (see Figure 1). Pitting is frequently seen, an will serve as a stress concentrator to initiate cracking. With SSC, hydrogen ions are a product of the corrosion process (Figure 2). These ions pick up electrons from the base material producing hydrogen atoms. At that point, i t two t hydrogen h d atoms t may combine bi tto fform a hydrogen h d molecule. l l M Mostt molecules will eventually collect, form hydrogen bubbles and float away y However,, some percentage p g of the hydrogen y g atoms will diffuse into harmlessly. the base metal and embrittle the crystalline structure. When a certain critical concentration of hydrogen is reached and combined with a tensile stress exceeding a threshold level level, SSC will occur occur. H2S does not actively participate in the SSC reaction; however, sulfides act to promote the entry of the hydrogen atoms into the base material.

Fion Zhang/ Charlie Chong

Figure 2. Schematic Showing the Generation of Hydrogen Producing SSC

Fion Zhang/ Charlie Chong

As little as 0.05 psi (0,3 kPa) H2S partial pressure in 65 psia (450 kPa) hydrocarbon gas can cause SSC of carbon and low alloy steels steels.

Sulfide stress cracking is most severe at ambient temperature, particularly in the range of 20° 20 to 120 120°F F ((-6° 6 to 49 49°C) C). Below 20°F 20 F ((6°C) the diffusion rate of the hydrogen is so slo that the critical concentration is never reached. Above 120°F (49°C), the diffusion rate is so fast that the hydrogen atoms pass through the material in such a rapid manner that the critical concentration is not reached.

Key-Numbers: Key Numbers: ¤ 65 psia absolute pressure ¤ as litle as 0.05 psi (0,3 kPa) H2S partial pressure

Fion Zhang/ Charlie Chong

Chloride SCC is widely encountered and has been extensively studied Much is still unknown studied. unknown, however however, about its mechanism. mechanism 1 One theory says that hydrogen, 1. hydrogen generated by the corrosion process, diffuses into the base metal in the atomic form and embrittles the lattice structure.

2. A second, more widely accepted theory proposes an electrochemical l t h i l mechanism. h i St Stainless i l steels t l are covered d with ith a protective, chromium oxide film. The chloride ions rupture the film at weak spots spots, resulting in anodic (bare) and cathodic (film covered) sites. The galvanic cell produces accelerated attack at the anodic sites, which when combined with tensile stresses produces cracking.

A minimum i i iion concentration t ti iis required i d tto produce d SCC SCC. A As th the concentration increases, the environment becomes more severe, reducing the time to failure failure.

Fion Zhang/ Charlie Chong

Temperature also is a factor in SCC. In general, the likelihood of SCC increases with increasing temperature temperature. A minimum threshold temperature exists for most systems, below which SCC is rare. Across industry, the generally accepted minimum temperature for chloride SCC of the 300 SST’s is about 160°F (71°C). NACE MR0175/ISO 15156 has set a very conservative limit of 140°F (60°C) due to the synergistic effects of the chlorides, H2S and low pH values values. As the temperature increases above these values values, the time to failure will typically decrease.

Resistance to chloride SCC increases with higher alloy materials. This is reflected in the environmental limits set by NACE MR0175/ISO 15156.

Environmental limits progressively increase from 400 Series SST and ferritic SST to 300 Series, highly alloy austenitic SST, duplex SST, nickel and cobalt base alloys.

Fion Zhang/ Charlie Chong

Carbon Steel Carbon C b and d llow-alloy ll steels t l h have acceptable t bl resistance i t tto SSC and d SCC however; their application is often limited by their low resistance to general processing g of carbon and low alloy y steels must be carefully y corrosion. The p controlled for good resistance to SSC and SCC. The hardness must be less than 22 HRC. If welding or significant cold working is done, stress relief is required Although the base metal hardness of a carbon or alloy steel is less required. than 22 HRC, areas of the heat affected zone (HAZ) will be harder. PWHT will eliminate these excessively hard areas.

ASME SA216 Grades WCB and WCC and SAME SA105 are the most commonly used body materials. materials It is Fisher Fisherâ&#x20AC;&#x2122;ss policy to stress relieve all welded carbon steels that are supplied to NACE MR0175/ISO 15156.

Fion Zhang/ Charlie Chong

All carbon steel castings sold to NACE MR0175/ISO 15156 requirements are produced using one of the following processes: 1. In particular product lines where a large percentage of carbon steel assemblies are sold as NACE MR0175/ISO 15156 compliant, castings are ordered from the foundry with a requirement that the castings be either normalized or stress relieved following all weld repairs repairs, major or minor minor.

Any weld repairs performed, either major or minor, are subsequently stress relieved.

2. In product lines where only a small percentage of carbon steel products are ordered NACE MR0175/ISO 15156 compliant, stock castings are stress relieved whether they are weld repaired by Emerson Process Management or not. This eliminates the chance of a minor foundry weld repair going undetected and not being stress relieved.

Fion Zhang/ Charlie Chong

ASME SA352 grades LCB and LCC have the same composition as WCB and WCC respectively.

WCC, respectively They are heat treated differently and impact tested at 50°F (-46°C) to ensure good toughness in low temperature service. LCB and LCC are used in locations where temperatures commonly drop below the 20°F (-29°C) permitted for WCB and WCC. LCB and LCC castings are processed in the same manner as WCB and WCC when required to meet NACE MR0175/ ISO 15156.

15156 For carbon and low-alloy steels NACE MR0175/ISO 15156 imposes some changes in the requirements for the weld procedure qualification report (PQR).

All new PQR’s will meet these requirements; however, it will take several years for Emerson Process Management and our suppliers to complete this work. At this time, we will require user approval to use HRC.

Fion Zhang/ Charlie Chong

Carbon and Low-Alloy Steel Welding Hardness R Requirements i t  HV-10, HV-5 or Rockwell 15N.  HRC testing is acceptable if the design stresses are less than 67% of the minimum specified yield strength and the PQR includes PWHT.  Other methods require user approval.  250 HV or 70 70.6 6 HR15N maximum. i  22 HRC maximum if approved by user.

Fion Zhang/ Charlie Chong

Low-Alloy Steel Welding Hardness Requirements  All of the above apply with the additional requirement of stress relieve at 1150°F ((621°C)) minimum after welding. g All new PQR’s at Emerson Process Management and our foundries will require hardness testing with HV HV-10, 10 HV HV-5 5 or Rockwell 15N and HRC HRC. The acceptable maximum hardness values will be 250 HV or 70.6 HR15N and 22 HRC. Hardness traverse locations are specified in NACE MR0175/ISO 15156 part 2 as a function of thickness and weld configuration. The number and locations of production hardness tests are still outside the scope of the standard The maximum allowable nickel content for carbon and low standard. low-alloy alloy steels and their weld deposits is 1%.

Note: A.2.1.6 Cold deformation and thermal stress relief Carbon and low-alloy steels shall be thermally stress-relieved following any cold deforming by rolling, cold forging or other manufacturing process that results in a permanent outer fibre deformation greater than 5 %. Thermal stress relief shall be performed in accordance with an appropriate code or standard. The minimum stress-relief temperature shall be 595 °C (1100 °F).

The final maximum hardness shall be 22 HRC except for pipe fittings made from ASTM A234 grade WPB or WPC, for which the final hardness shall not exceed 197 HBW.

Fion Zhang/ Charlie Chong

Table A.1 â&#x20AC;&#x201D; Maximum acceptable hardness values for carbon steel, carbon manganese steel and carbon-manganese low-alloy steel welds

Fion Zhang/ Charlie Chong

Low alloy steels like WC6, WC9, and C5 are acceptable to NACE MR0175/ISO 15156 to a maximum hardness of 22 HRC HRC. These castings must all be stress relieved to FMS 20B52 (?) .

The compositions of C12, C12a, F9 and F91 materials do not fall within the definition of “low alloy steel” in NACE MR0175/ISO 15156, therefore, these materials are not acceptable. acceptable A few customers have specified a maximum carbon equivalent (CE) for carbon steel. The primary driver for this requirement is to improve the SSC resistance in the as-welded condition.

Fisher’s practice of stress relieving all carbon steel negates this need.

Decreasing the CE reduces the hardenability of the steel and presumably i improves resistance i t tto sulfide lfid stress t cracking ki (SSC) (SSC).

Because reducing the CE decreases the strength of the steel, there is a limit to how far the CE can be reduced.

Fion Zhang/ Charlie Chong

ANSI/NACE MR0175/ISO 15156-1 3.14 low-alloy steel steel with a total alloying element content of less than about 5 % mass fraction, but more than specified for carbon steel (3.3) 3.3 carbon steel alloy of carbon and iron containing up to: ď ľ 2 % mass fraction carbon and ď ľ up to 1.65 % mass fraction manganese and residual quantities of other elements, except those intentionally added in specific quantities for deoxidation (usually silicon ili and/or d/ aluminium) l i i )N Note t 1 tto entry: t C Carbon b steels t l used d iin th the petroleum t l industry usually contain less than 0.8 % mass fraction carbon.

Fion Zhang/ Charlie Chong

Cast Iron Gray, austenitic G t iti and d white hit castt iirons cannott b be used d ffor any pressureretaining parts, due to low ductility. Ferritic ductile iron to ASTM A395 is p when p permitted by y ANSI,, API or other industry y standards. acceptable

Fion Zhang/ Charlie Chong

Gray Cast Iron

Fion Zhang/ Charlie Chong

Gray Cast Iron

Fion Zhang/ Charlie Chong

White Cast Iron

Fion Zhang/ Charlie Chong

White Cast Iron

Fion Zhang/ Charlie Chong

Austenitic Cast Iron

Fion Zhang/ Charlie Chong

Austenitic Cast Iron

Fion Zhang/ Charlie Chong

Austenitic Cast Iron

Fion Zhang/ Charlie Chong

Meallable Cast Iron

Fion Zhang/ Charlie Chong

Meallable Cast Iron

Fion Zhang/ Charlie Chong

Meallable Cast Iron

Fion Zhang/ Charlie Chong

Fe-C Equillibrium Diagram

Fion Zhang/ Charlie Chong

Optional Reading - Non CS Scope

Fion Zhang/ Charlie Chong

Stainless Steel 400 Series S i Stainless St i l Steel St l UNS 410 (410 SST), CA15 (cast 410), 420 (420 SST) and several other martensitic grades must be double tempered to a maximum hardness of 22 HRC. PWHT is also required. An environmental limit now applies to the martensitic grades; 1.5 psi (10 kPa) H2S partial ti l pressure and d pH H greater t than th or equall to t 3.5, 3 5 416 (416 SST) is similar to 410 (410) with the exception of a sulfur addition to produce free machining characteristics. Use of 416 and other free machining steels is not permitted by NACE MR0175/ISO 15156.

CA6NM is a modified version of the cast 410 stainless steel.

NACE MR0175/ISO 15156 allows ll its it use, but b t specifies ifi the th exact heat treatment required. Generally, the carbon content must be restricted to 0.03% maximum to meet the 23 HRC maximum hardness. PWHT is required for CA6NM. The same environmental limit applies; 1.5 psi (10 kPa) H2S partial pressure and pH greater than or equal to 3 3.5.

5

Fion Zhang/ Charlie Chong

300 Series Stainless Steel Severall changes S h h have b been made d with ith th the requirements i t off th the austenitic (300 Series) stainless steels. Individual alloys are no g listed. All alloys y with the following g elemental ranges g longer are acceptable: C 0.08% maximum, Cr 16% minimum, Ni 8% minimum, P 0.045% maximum, S 0.04% maximum, Mn 2.0% maximum and Si 2.0% maximum, 2 0% maximum. maximum Other alloying elements are permitted. The other requirements remain; solution heat treated condition, 22 HRC maximum and free of cold work designed to improve mechanical properties. The cast and wrought equivalents of 302, 304 (CF8), S30403 (CF3), 310 (CK20), 316 (CF8M), S31603 (CF3M) 317 (CG8M) (CF3M), (CG8M), S31703 (CG3M) (CG3M), 321, 321 347 (CF8C) and N08020 (CN7M) are all acceptable per NACE MR0175/ISO 15156.

Fion Zhang/ Charlie Chong

Environmental restrictions now apply to the 300 Series SST. The limits are 15 psia (100 kPa) H2S partial pressure pressure, a maximum temperature of 140°F (60°C), and no elemental sulfur. If the chloride content is less than 50 mg/L (50 ppm), the H2S partial pressure must be less than 50 psia (350 kPa) but there is no temperature limit.

There is less of a restriction on 300 Series SST in oil and gas processing and injection facilities. If the chloride content in aqueous solutions is low (typically less than 50 mg/L or 50 ppm chloride) in operations after separation, there are no limits for austenitic stainless steels, highly alloyed austenitic stainless steels, duplex stainless steels, or nickel-based alloys.

--------------- more reading di

Fion Zhang/ Charlie Chong

Reading 2 Stress Corrosion Cracking

http://oilfieldwiki com/wiki/Stress corrosion cracking http://oilfieldwiki.com/wiki/Stress_corrosion_cracking

Fion Zhang/ Charlie Chong

SCC Stress corrosion St i cracking ki (SCC) iis th the growth th off cracks k iin a corrosive i environment. It can lead to unexpected sudden failure of normally ductile j to a tensile stress,, especially p y at elevated temperature p in the metals subjected case of metals. SCC is highly chemically specific in that certain alloys are likely to undergo SCC only when exposed to a small number of chemical environments The chemical environment that causes SCC for a given alloy is environments. often one which is only mildly corrosive to the metal otherwise. Hence, metal parts with severe SCC can appear bright and shiny, while being filled with microscopic cracks. This factor makes it common for SCC to go undetected prior to failure. SCC often progresses rapidly, and is more common among alloys than pure metals metals. The specific environment is of crucial importance importance, and only very small concentrations of certain highly active chemicals are needed to produce catastrophic cracking, often leading to devastating and unexpected failure.[1] The stresses can be the result of the crevice loads due to stress concentration, or can be caused by the type of assembly or residual stresses from fabrication (e.g. cold working); the residual stresses can be relieved by annealing. li Fion Zhang/ Charlie Chong

Metals attacked  C Certain t i austenitic t iti stainless t i l steels t l and d aluminium l i i alloys ll crack k iin th the presence of chlorides, presence of alkali ((boiler cracking) g) and nitrates,,  mild steel cracks in the p  copper alloys crack in ammoniacal solutions (season cracking).

This limits the usefulness of austenitic stainless steel for containing water with higher than few ppm content of chlorides at temperatures above 50 °C. Worse still, high-tensile structural steels crack in an unexpectedly brittle manner in a whole variety of aqueous environments, especially containing chlorides. With the possible exception of the latter, which is a special example of hydrogen cracking all the others display the phenomenon of subcritical crack growth cracking, growth, i.e. small surface flaws propagate (usually smoothly) under conditions where fracture mechanics predicts that failure should not occur. That is, in the presence of a corrodent, cracks develop and propagate well below KIc. In fact, the subcritical value of the stress intensity, designated as KIscc, may be less than 1% of KIc I , as the following table shows:

Fion Zhang/ Charlie Chong

KIC & KISCC

Fion Zhang/ Charlie Chong

KIC & KISCC

Fion Zhang/ Charlie Chong

Crack growth The subcritical Th b iti l nature t off propagation ti may be b attributed tt ib t d tto th the chemical h i l energy released as the crack propagates. That is, elastic energy released + chemical gy = surface energy gy + deformation energy. gy energy The crack initiates at KIscc and thereafter propagates at a rate governed by the slowest process, process which most of the time is the rate at which corrosive ions can diffuse to the crack tip.

As the crack advances so K rises (because crack length appears in the calculation of stress intensity). Finally it reaches KIc , whereupon fast fracture ensues and the component fails fails.

One of the practical difficulties with SCC is its unexpected nature. Stainless steels, for example, are employed because under most conditions they are "passive", i.e. effectively inert. Very often one finds a single crack has propagated while the rest of the metal surface stays apparently unaffected unaffected.

The crack propagates perpendicular to the applied stress.

Fion Zhang/ Charlie Chong

Prevention SCC is i the th resultlt off a combination bi ti off three th factors f t –  a susceptible material, p to a corrosive environment,, and  exposure  tensile stresses above a threshold.

If you eliminate any one of these factors SCC initiation becomes impossible impossible.

The conventional approach to controlling the problem has been to develop new alloys that are more resistant to SCC. This is a costly proposition and can require a massive time investment to achieve only marginal success.

Fion Zhang/ Charlie Chong

Examples A 32 iinch h di diameter t gas ttransmission i i pipeline, i li north th off N Natchitoches, t hit h Louisiana, belonging to the Tennessee Gas Pipeline exploded and burned g 17 people. p p At least 9 others were injured, j , from SCC on March 4,, 1965,, killing and 7 homes 450 feet from the rupture were destroyed.[2][3] SCC caused the catastrophic collapse of the Silver Bridge in December 1967 1967, when an eyebar suspension bridge across the Ohio river at Point Pleasant, West Virginia, suddenly failed. The main chain joint failed and the whole structure fell into the river, killing 46 people in vehicles on the bridge at the time. Rust in the eyebar joint had caused a stress corrosion crack, which went critical as a result of high bridge loading and low temperature temperature. The failure was exacerbated (to increase the severity, bitterness, or violence of disease, ill feeling, etc.); aggravate ) by a high level of residual stress in the eyebar. The disaster led to a nationwide reappraisal of bridges.[4]

Fion Zhang/ Charlie Chong

A classic example of SCC is season cracking of brass cartridge cases, a problem experienced by the British army in India in the early 19th century century. It was initiated by ammonia from dung and horse manure decomposing at the higher temperatures of the spring and summer. There was substantial residual stress in the cartridge shells as a result of cold forming. The problem was solved by annealing the shells to ameliorate the stress.

Fion Zhang/ Charlie Chong

The collapsed Silver Bridge, as seen from the Ohio side

Fion Zhang/ Charlie Chong

The collapsed Silver Bridge, as seen from the Ohio side

Fion Zhang/ Charlie Chong

The collapsed Silver Bridge, as seen from the Ohio side

Fion Zhang/ Charlie Chong

SCC Aluminum- Ammonia NH3

Fion Zhang/ Charlie Chong

SCC Aluminum- Ammonia NH3

Fion Zhang/ Charlie Chong

Reading 3 Corrosion problems in production The WikiPetrol

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Corrosion problems in production Corrosion C i off metal t l iin th the presence off water t iis a common problem bl across many industries. The fact that most oil and gas production includes coproduced water makes corrosion a p p pervasive issue across the industry. y Age g and presence of corrosive materials such as carbon dioxide (CO2) and hydrogen sulfide (H2S) exacerbate the problem.

Corrosion control in oil and gas production is reviewed in depth in Treseder and Tuttle,[1] Brondel, et al.,[2] and NACE,[3] from which some of the following material is abstracted.

Fion Zhang/ Charlie Chong

Corrosion chemistry of steels IIron is i iinherently h tl (th (thermodynamically) d i ll ) sufficiently ffi i tl active ti tto reactt spontaneously with water (corrosion), generating soluble iron ions and y g g gas. The utility y of iron alloys y depends p on minimizing g the corrosion hydrogen rate. Corrosion of steel is an “electrochemical process,” involving the transfer of electrons from iron atoms in the metal to hydrogen ions or oxygen in water.

The corrosion reaction of iron with H2S under anearobic condition is described by the equation Corrosion under anaerobic condition commonly found in oilfield production: Anode: Cathode:

Fion Zhang/ Charlie Chong

Fe + H2O → Fe22+ + 2e- + H2O H2S + H2O → H+ + HS- + H2O HS- + H2O → H+ + S2- + H2O

This separation of the overall corrosion process into two reactions is not an electrochemical nuance; these processes generally do take place at separate locations on the same piece of metal. This separation requires the presence of a medium to complete the electrical circuit between anode (site of iron dissolution) and cathode (site for corrodant reduction). Electrons travel in the metal phase, but the ions involved in the corrosion process cannot. Ions require the presence of water; hence hence, corrosion requires the presence of water. This overall process is shown schematically in Fig. 1.[3] The space between the anode and cathode may be small or large depending on a number of factors.

Fion Zhang/ Charlie Chong

Acid is not the only corrodant possible. Another common cathodic process is the reduction of oxygen, oxygen which is written as : O2 + 4H+ + 4e- â&#x2020;&#x2019; 2H2O This reaction can also take place at a location different from that of iron dissolution.

The other chemical constituents in the vicinity of the anodic sites determine the ultimate chemical fate of the Fe++ ion, such as the precipitation of ironcontaining solids on or near the corroding surface.

The net rate of corrosion is determined by how fast the corrodant arrives at the iron-atom/water interface, how much corrodant is present, the electrical potential t ti l ((energy)) off the th corrodant d t (oxygen ( has h a higher hi h potential t ti l th than d do protons), and the intrinsic rate of the cathodic reactionsâ&#x20AC;&#x201D;electron transfer processes involving p gp protons and oxygen yg are not instantaneous and depend p on the nature of the solid surface on which they occur.

Fion Zhang/ Charlie Chong

“How fast the corrodants arrive” has two aspects:  Mass transport in the corroding fluid  Permeating surface barriers between the iron metal and the water phase Surface barriers are placed barriers, such as:  Paint or plastic coatings  Passivating oxide films inherent to the metal (discussed later)  Low-permeability corrosion products (e.g., siderite, as formed in the presence of certain oils and/or inhibitors) http://petrowiki.org/Corrosion_problems_in _production

Fion Zhang/ Charlie Chong

Nature of steels Alloying All i iron i with ith carbon b ((usually ll 0 0.2 2 tto 1%) fforms steel t l (l (low-alloy ll steel)â&#x20AC;&#x201D;a t l) far stronger metal than iron, hence, suitable for oilfield use. Other components p p can be added to iron to enhance corrosion-resistance properties.

Some of the carbon added is insoluble, forming iron carbide (Fe3C), which accelerates the cathodic processes necessary for corrosion to take place, accelerating the corrosion rate. rate One of the major, major ubiquitous impurities in steel is sulfur, and it is a major source of corrosion instability. This element is highly insoluble in iron and precipitates in the form of insoluble sulfide inclusions, in particular MnS and (Mn, Fe)S. These inclusions are generally the sites of pitting (discussed later).[4] Grain boundaries are also areas that are chemically active active.[3] When iron solidifies during casting, the atoms, which are randomly distributed in the liquid state, arrange themselves in a crystalline array. This ordering usually begins simultaneously at many points in the liquid, and as these blocks of crystals and grains meet, there is a mismatch in the boundaries. There are areas of higher energy energy. Chemical impurities in the melt tend to accumulate at these grain boundaries and are more susceptible to chemical attack than the iron surface itself.

Fion Zhang/ Charlie Chong

Plain carbon steels are processed by one of four heat treatments:  Annealing  Normalizing  Spherodizing  Quench and tempering These treatments determine, determine in part, part the physical and corrosion properties of the metal. Annealing or normalizing results in greater corrosion resistance than spherodizing or quench and tempering. The logic is that these treatments determine, in large, part of the physical dimensions and distribution of the impurities and inclusions in the metal.

Fion Zhang/ Charlie Chong

The corrosion products formed in oxygen-containing water on mild steel are FeOOH likely amorphous, FeOOH, amorphous and magnetite magnetite.[4] Below 200 200°C C, these oxides oxides, in the absence of reactive inclusions, are protective. In the presence of dissolved CO2, FeCO3 films form, which can sometimes be protecting (discussed later).

The compositions of corrosion-resistant alloys (CRAs) are chosen to spontaneously generate surface oxide films that will be stable and impermeable in the presence of the more aggressive corrodants. In oilfield use, it is also required that these films spontaneously reform if ruptured, as, for example, during and after erosion by sand or scratching by wireline/caliper tools. CRAs include the ferrous stainless steels and nonferrous nickel and cobalt alloys. Stainless steels contain at least 12% chromium. These alloys passivate in oxidizing environments through the formation of a thin layer of chromium oxide—containing film on the surface of the alloy. The crystallinity of thi fil this film d decreases with ith iincreasing i C Cr content t t iin th the steel, t l b becoming i more glass-like and more protective.[5] Again, various inclusions can be weak points passivating g film. The surfaces of nickel-based CRAs,, such as Incoloyy in the p 800™, are a passivating nickel ferrite (Ni0.8Fe2.2O4).

Fion Zhang/ Charlie Chong

There are four classes of stainless steels that are based on chemical content, metallurgical structure structure, and mechanical properties. properties These classes are:  Martensitic  Ferritic  Austenitic  Duplex  PH (?)

The manufacturing processes for CRAs are more complex than those producing low-alloy steels. Stainless steels are less costly than the nickel and cobalt alloys, though they are 1.5 to 20 times more expensive than low-alloy steels.

Fion Zhang/ Charlie Chong

Oilfield corrosion Oilfield Oilfi ld corrosion i can be b divided di id d iinto t corrosions i d due tto oxygen, ""sweet" t" corrosion, and "sour" corrosion.

Corrosion because of oxygen is found with surface equipment and can be found downhole with the oxygen introduced by waterflooding, pressure maintenance gas lifting, maintenance, lifting or completion and/or workover fluids fluids. It is the major corrodant of offshore platforms, at and below the tide line. The chemistry of this process follows the equations given below; Oxidation of iron at points of stress in the crystal lattice: 2Fe(s) 2Fe22+(aq) + 4eReduction of water at the site of carbon impurities: O2(g) + H2O(l) + 4e- 4OH-(aq) Overall equation: 2Fe(s) + O2(g) + H2O(l) Fe(OH)2 .

Fion Zhang/ Charlie Chong

http://www.chemicalformula.org/chemistry-help/corrosion

The iron(II) hydroxide is converted to rust through a serious of reactions.

The iron(II) hydroxide firstly oxides to iron(III) oxide.

1.

Fe(OH)2(s)

→ oxidation

Fe(OH)3

The iron(III) oxide then changes to rust through a dehydration reaction.

2.

Fe(OH)3(s)

→ dehydration

Fe2O3.nH2O(s) or rust

(2Fe(OH)3 → Fe2O3 + 3H2O) Rustt adheres R dh lloosely l tto th the surface f off th the metal. t l Thi This exposes th the metal t l tto more and more water and oxygen allowing the process of rusting to continue.

Fion Zhang/ Charlie Chong

http://www.chemicalformula.org/chemistry-help/corrosion

“Sweet” corrosion is generally characterized first by simple metal dissolution followed by pitting pitting. The corrodant is H+, H+ derived from carbonic acid (H2CO3) and the dissolution of CO2 in the produced brine. The pitting leaves distinctive patterns (e.g., “mesa” corrosion), attributable to the metallurgical processing used in manufacturing the tubing. “Ringworm” corrosion is caused when welding is not followed by full-length normalizing of the tubular after processing Corrosion inhibitors and CRAs are effective in mitigating sweet processing. corrosion. Naphthenic acids and simple organic acids indigenous to crude oil also contribute to corrosion

Fion Zhang/ Charlie Chong

Sweet Corrosion The deterioration of metal due to contact with carbon dioxide or similar corrosive agents, but excluding hydrogen sulfide [H2S]. Sweet corrosion typically results in pitting or material loss and occurs where steel is exposed to carbon dioxide and moisture http://www.glossary.oilfield.slb.com/en/Terms/s/sweet_corrosion.aspx

Fion Zhang/ Charlie Chong

Sweet The terms sweet and sour are a reference f to t the th sulfur lf content t t of crude oil. Early yp prospectors p would taste oil to determine its quality with low sulfur oil quality, actually tasting sweet. Crude is currently considered sweet if it contains less than 0.5% 0 5% sulfur. sulfur

Fion Zhang/ Charlie Chong

â&#x20AC;&#x153;Sourâ&#x20AC;? corrosion (H2S) results in the formation of various insoluble iron sulfides on the metal surface surface. Not only is H2S an acidic corrodant corrodant, it also acts as a catalyst for both the anodic and cathodic halves of the corrosion reaction.

Galvanic corrosion (bimetallic corrosion) is caused by the coupling of a corrosive and noncorrosive metal in the presence of a corrodant. Erosion is yet another category of corrosion corrosion. Erosion corrosion is the acceleration of corrosion because of the abrasion of metal surfaces by particulates (e.g., sand). Finally, there is corrosion caused by acidsâ&#x20AC;&#x201D;those used to stimulate wells (HCl and HF).

Fion Zhang/ Charlie Chong

Sweet The terms sweet and sour are a reference to the sulfur content of crude oil oil.

Early prospectors would taste oil to determine its quality, with low sulfur oil actually tasting sweet. Crude is currently considered sweet if it contains less than 0.5% sulfur.

Sweet crude is easier to refine and safer to extract and transport than sour crude. Because sulfur is corrosive, light crude also causes less damage to refineries and thus results in lower maintenance costs over time. Due to all these factors, sweet crude commands up to a $15 $ dollar premium per barrel over sour.

Major locations where sweet crude is found include the Appalachian Basin in Eastern North America, Western Texas, the Bakken Formation of North D k t and Dakota dS Saskatchewan, k t h th the N North th S Sea off E Europe, N North th Af Africa, i A Australia, t li and the Far East including Indonesia.

Fion Zhang/ Charlie Chong

http://www.petroleum.co.uk/sweet-vs-sour

Sour Sour crude oil will have greater than 0 0.5% 5% sulfur and some of this will be in the form of hydrogen sulfide. Sour crude also contains more carbon dioxide. Most sulfur in crude is actually bonded to carbon atoms, nevertheless, high quantities of hydrogen sulfide in sour crude can pose serious health problems or even be fatal.

Hydrogen sulfide is famous for its â&#x20AC;&#x153;rotten eggâ&#x20AC;? smell, which is only noticed at low concentrations. At moderate concentrations, hydrogen sulfide can cause respiratory and nerve damage. At high concentrations, it is instantly fatal.

Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome, which is characterized by chronic fatigue, headaches, dizziness, memory problems, serious breathing problems, and even birth defects. Hydrogen sulfide is so much of a risk that sour crude has to be stabilized t bili d via i removall off h hydrogen d sulfide lfid b before f it can b be ttransported t db by oilil tankers.

Sour crude is more common in the Gulf of Mexico,, Mexico,, South America,, and Canada. Crude produced by OPEC Member Nations also tends to be relatively sour, with an average sulfur content of 1.77%.

Fion Zhang/ Charlie Chong

http://www.petroleum.co.uk/sweet-vs-sour

Gulf War Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome

Fion Zhang/ Charlie Chong

Gulf War Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome

Fion Zhang/ Charlie Chong

Gulf War Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome

Fion Zhang/ Charlie Chong

Gulf War Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome

Fion Zhang/ Charlie Chong

Gulf War Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome

Fion Zhang/ Charlie Chong

Gulf War Exposure to high levels of hydrogen sulfide is thought to be in part responsible for Gulf War Syndrome

Fion Zhang/ Charlie Chong

Oilfield corrosion can take specific forms: 1.

2.

3.

4.

5 5.

6.

7.

8.

9.

Metal wastage Pitting Crevice corrosion Intergranular corrosion Stress corrosion cracking (SCC) Blistering (HIC) Embrittlement (SCC/SSC...?)

Sulfide stress cracking (SSC) Corrosion fatigue

The first five forms involve primarily carbonic acid and/or dissolved oxygen as corrodants. Items 6 through 8 are induced primarily by H2S.

Fion Zhang/ Charlie Chong

Corrosive failure by uniform loss of metal is only infrequently seen during the production of oil and gas. gas It is, is however, however the first step in corrosive failure of steels by means of localized corrosion. A circumstance for severe metal wastage is the pumping of poorly inhibited matrix stimulation acids.

Pitting is the common failure mode of sweet corrosion and corrosion because of dissolved oxygen oxygen. All passivating/protecting films on steel contain weak spots that will preferentially dissolve and form pits. As mentioned, these areas are generally the sulfide inclusions. Chloride ion weakens the repassivating film, allowing continued dissolution. The decreasing pH within the pit also enhances continued corrosion. The driver for theses processes is the large cathodic area of the metal oxide surface vs. the small anodic pit. Pitting is particularly dangerous because penetration through a tubular can occur relatively fast. Other corrosion mechanisms, such as SCC, frequently start at pits. it Oxygen O scavengers are typically t i ll used d tto remove thi this gas iin an attempt tt t tto minimize the pitting problem. However, small amounts may remain (e.g., 20 ppb), pp ), and these can be sufficient to induce corrosion.

Fion Zhang/ Charlie Chong

Carbonic acid, the driver for sweet corrosion, is a weak acid. The pH of the formation water depends on the CO2 partial pressure pressure, temperature temperature, and alkalinity (controlled primarily, but not exclusively, by the presence or absence of carbonate minerals in the formation). Shown in Fig. 2, as a function of CO2 partial pressure, are computed pH values for a seawater brine (containing 140 ppm alkalinity) and a seawater brine saturated in calcite at 50 and 150°C (substantially higher alkalinities). alkalinities) For the common case of carbonatecarbonate containing reservoirs and moderate temperatures, produced waters should have pH values of 6 or greater. Waters exposed to greater amounts of CO2 in noncarbonate-containing reservoirs can have pH values of 4 or less.

Fion Zhang/ Charlie Chong

Fig. 2â&#x20AC;&#x201D;Computed pH vs. pressure for a seawater brine exposed to a gas phase containing CO2; data are shown for seawater alone at 50oC and for calcite-saturated seawater at 50 and 150oC.

Fion Zhang/ Charlie Chong

Such corrosion induced by CO2 is a function not only of CO2 partial pressure and temperature but also of the crude oil. oil Crude oil contains surface-active surface active chemicalsâ&#x20AC;&#x201D;some oils contain more than others. These chemicals (e.g., resins and asphaltenes) can impact the corrosion process, at least for low-alloy steels. For a fixed brine composition, WOR(water to oil ratio) , temperature, and pressure, corrosion in the presence of some crudes can be negligible, while in the presence of others, others it can be extreme under identical environmental conditions.[6][7] Sweet corrosion generally results in the deposition of insoluble FeCO3 (siderite) on the steel surface. It has been suggested that this selectivity to oil composition relates to the physical morphology of the FeCO3 corrosion productâ&#x20AC;&#x201D;a compact, tight film can protect the steel; a loose, poorly adherent film does not.[7] An example is shown in Fig. 3. The average uniform corrosion rate for steel in Crude B was 0.6 mil/yr; the corrosion rate in Crude E was 26 mil/yr. Many corrosion inhibitors apparently tl actt by b the th same mechanism h i (i.e., (i th the generation ti off siderite id it films fil similar, and/or more compact than those formed from Crude B).

Fion Zhang/ Charlie Chong

Fig. 3â&#x20AC;&#x201D;Scanning-electron-microscope micrographs (X10K) of the surface of the N-80 N 80 steel coupons after a 24 24-hour hour exposure at 186oF to brine and 760 760-psi psi CO2, without crude oil (upper left), with 95 vol% crude oil E (upper right), with 95 vol% crude oil F (lower left), and with 95 vol% crude oil B (lower right); all deposits are siderite (courtesy of the Electrochemical Society).

Fion Zhang/ Charlie Chong

Fion Zhang/ Charlie Chong

Alternatively, it has been suggested that wettability plays the dominant role, whereby the surface surface-active active components in the crude oil provide for a waterwater wet surface (high corrosion rates) or an oil-wet surface (low corrosion rates).[8] Regardless of the mechanism, crude oil can modify the corrosion rate. The penalty for ignoring the effect of crude-oil chemistry is the cost of overtreating or using more expensive alloys than are required.

A crevice, crevice such as the junction space under a bolt or the physical junction of two metal parts, is in effect a pit. Uniform corrosion can initiate (in the presence of a corrodant) within the crevice and continue, driven by the large cathodic area outside the pit or crevice.

Fion Zhang/ Charlie Chong

Stress corrosion cracking is intergranular corrosion, but it takes place only when the metal is under stress and in the presence of a corrodant corrodant. The corrodant can be specificâ&#x20AC;&#x201D;not all corrodants induce SCC on all alloys. Metal wastage is generally small; SCC is often preceded by pitting. High-strength steels are more susceptible to SCC than low-strength alloys. The severity of intergranular corrosion generally depends on the metallurgical history of the steel Austenitic steels (common stainless steels) are particularly susceptible steel. to intergranular attack.

Blistering, as well as embrittlement and sulfide stress cracking, a subclass of SCC, all stem from the same cause: the presence of H2S in the system and at the metal surface (?) . The roots of the problem are in the mechanism for the cathodic discharge of hydrogen. The mechanism already discussed for the cathodic portion of the acid-induced corrosion process itself, involves two steps. t H+ + eH+H

Fion Zhang/ Charlie Chong

â&#x2020;&#x2019; â&#x2020;&#x2019;

H H2

((1)) (2)

(i.e., the proton is first reduced to a hydrogen atom on the metal surface (Hâ&#x20AC;˘), followed by the combination of two hydrogen atoms to yield hydrogen gas) gas).

Hydrogen sulfide inhibits the combination of hydrogen atoms (as does arsenic HCN and some other corrosion inhibitors). Accordingly, the hydrogen atoms can penetrate into the metal where they cause the corrosion problems that were already listed. This is shown schematically in Fig. 4

Fion Zhang/ Charlie Chong

Fig. 4 â&#x20AC;&#x201D;The alternatives for hydrogen atoms formed by the corrosion process: combination in the water phase to make gas gas, diffusion into the metal to make gas or embrittle steel, penetration through the metal, recombining to make gas (a phenomenon also used to measure corrosion)(after Schlumberger Oilfield Review).

Fion Zhang/ Charlie Chong

This hydrogen entry into low-strength steels can result in hydrogen blisters, if there is a macroscopic defect in the steel such as an inclusion inclusion. Such a void can provide a space for the hydrogen atoms to form hydrogen gas. Pressure builds and blisters form resulting in rupture and leakage.

Embrittlement (hydrogen-induced cracking and hydrogen embrittlement cracking) causes failure at stresses well below the yield strength . This phenomenon usually occurs only with high-strength, hard steels, generally those having yield strengths of 90,000 psi or higher. Tubing and line pipe (electric welded and seamless) are susceptible to this effect. The dominating factor is the metallurgical structure of the steel relating to its method of manufacture.

Fion Zhang/ Charlie Chong

Hydrogen Blister- Low strength steel

Fion Zhang/ Charlie Chong

Hydrogen Blister- Low strength steel

Fion Zhang/ Charlie Chong

Hydrogen Blister- Low strength steel

Fion Zhang/ Charlie Chong

Hydrogen Blister- Low strength steel Mi Micrography h showing h i h hydrogen d embrittlement b ittl t

Fion Zhang/ Charlie Chong

https://eduardodseng.wordpress.com/2013/11/10/different-types-of-corrosion/

SSC cracking failure requires only low concentrations of H2S. The time to failure decreases as stress increases increases. Cracking tendency increases as pH decreases. SSC can be thought of in the same language as that used in describing hydraulic fracturing. There is a critical “stress intensity factor” below that at which a fracture (crack) will not propagate. This factor is related linearly to tensile strength. Some of this problem has been attributed to the effects of cold working on the alloys alloys. Alloys that were stress relieved were found to increase in resistance to SSC.[9] Wells producing hydrocarbon liquids, with the hydrogen sulfide, are less susceptible to SSC, pitting, and weight loss. For example, certain Canadian condensate wells have produced fluids with 40 mol% H2S and 10% CO2 for 30 years without serious corrosion problems. Stability is associated with a protective iron sulfide film, wetted by the oil/liquid hydrocarbon. These wells also l h had d a BHT off 90°C 90°C; iiron sulfide lfid fil films are lless effective ff ti iin preventing ti corrosion above 110°C.

Fion Zhang/ Charlie Chong

Canadian Condensate Wells

Fion Zhang/ Charlie Chong

Steels, repeatedly stressed in a cyclical manner, may fail in time (corrosion fatigue) It is required for failure that the stress be above a critical value called fatigue). the â&#x20AC;&#x153;endurance limitâ&#x20AC;? (nominally 40 to 60% less than the tensile strength). The presence of a corrodant substantially reduces the fatigue life of a metal. Cyclic stress can be looked upon as a method of accelerating failure because of the other mechanisms previously described.

Bimetallic corrosion/galvanic corrosion can occur when two metals are coupled (in electrical contact) and a corrodant is present. The more reactive metal corrodes faster, while the less-reactive metal shows little or no corrosion.

The more-reactive metal cathodically protects the less-reactive metal (exploiting cathodic protection to prevent corrosion is discussed later). In general, the total corrosion of the anodic material is proportional to the exposed area of the cathodic material. Thus, steel rivets in monel corrode very rapidly, idl while hil monell rivets i t iin steel t l cause littl little d damage.

Fion Zhang/ Charlie Chong

Anode to Cathode Ratio

Fion Zhang/ Charlie Chong

Weld-related corrosion is a variant of galvanic corrosion. When a metal is welded the welding process can generate a microstructure different from that welded, of the parent metal. As a result, the weld may be anodic vs. the parent metal and may corrode more rapidly. This corrosion may take the form of localized metal wastage; if H2S is present, there is SSC cracking of hard zones in the metal or in the heat-affected zone. Similar problems can arise with electricresistance-welded resistance welded pipe. pipe Metal wastage in sweet systems is avoided by using weld consumable with a higher alloy content than that of the base metal; recourse is made to laboratory measurements to achieve the proper weld-metal/base-metal combination. Welding procedure standards are available to avoid hard zone SSC. Chemical inhibition is also effective in protecting welded pipe.

Fion Zhang/ Charlie Chong

Preventing corrosion The paths Th th to t obviating b i ti corrosion i problems bl are conceptually t ll straightforward: t i htf d  Isolate the metal from the corrodant p y a metal alloy y that is inherently y resistant to corrosion in the  Employ corrosive medium  Chemically inhibit the corrosion process  Move the electrical potential of the metal into a region where the corrosion rate is infinitesimally small (“cathodic protection”) An alternative is to live with the corrosion and replace the corroded component after failure. failure

Fion Zhang/ Charlie Chong

Isolation IIsolation l ti is i the th regime i off paints, i t coatings, ti and d liners. li A An iintroduction t d ti tto th the subject is given in NACE,[3] from which some of the following discussion is j is in NACE.[10] For anyy abstracted;; a detailed discussion of these subjects coating to be effective, it must be sufficiently thick to completely isolate the item being protected from the environment. Small holes in the coating ((â&#x20AC;&#x153;holidaysâ&#x20AC;?) holidays ) result in the rapid formation of pits. pits Considerable care and quality control is required to guarantee the generation of holidays during service.

Organic coatings, such as asphalt enamel and coal tar enamel, are used to protect equipment concerned with the handling of oil and gas. Baked thin-film coatings, such as thermosetting phenolics and epoxies (applied in multiple coats) can be used to protect tubular goods coats), goods. External protection of pipelines frequently involves use of adhesive tapes made of polyethylene or similar materials. Fusion bonded epoxy has been used successfully to protect a 150km seawater-injection line (oxygen was the corrodant, much of which, but not all, was removed by scavenging chemicals).[

Fion Zhang/ Charlie Chong

Inorganic coatings include both sacrificial coatings, which furnish cathodic protection (see below for mechanism) at small breaks in the coating coating, and nonsacrificial coatings, which protect only the substrates actually coated.

Sacrificial coatings include galvanizing or coating with other metals anodic to the substrate and heavy suspensions of anodic metals (e.g., zinc particles, in silicates or organic vehicles). Zinc-silicate coatings (paints) are often used to coat the splash zone of drilling and production platforms platforms. The zinc metal provides for cathodic protection of the steel substrate. Below the water line, the most economical approach to corrosion control is cathodic protection (see below). The pH of the environment is importantâ&#x20AC;&#x201D; highly basic or acidic environments can remove coatings.

Fion Zhang/ Charlie Chong

Zinc-silicate coatings

Fion Zhang/ Charlie Chong

Nonsacrificial inorganic coatings include metal platings, such as nickel and nonmetallic coatings such as ceramics ceramics. Nickel can be applied by electroplating or electroless plating. Ceramic coatings, when properly applied, are highly effective, but they are also costly and fragile. Other systems, while not truly coatings, perform the same function (e.g., Portland cement and plastic liners).

Plastic liners have been used for internal protection of tubing and lined pipe.

Some liners are sealed into individual joints of pipe and tubing; some are fused into one continuous close-fitting liner through the entire pipe. Both cement and plastic liners are suitable for water lines.

The proper application of coatings is, in large part, an art form. Accordingly, it is also not possible to overemphasize the need for close inspection of the coating process, good quality control, and testing that the coating has been complete.

Fion Zhang/ Charlie Chong

Pipeline Wrapping

Fion Zhang/ Charlie Chong

Corrosion resistant alloys (CRA) From a costt point F i t off view, i low-alloy l ll steels t l are preferred. f d IIn certain t i cases, â&#x20AC;&#x153;minorâ&#x20AC;? alterations in alloy composition can minimize corrosion. For example, p martensitic structure and a chromium content L-80 steel with a tempered > 0.5% has been used without problems in 20-ppb oxygen-containing environments, while a similar steel with < 0.1% Cr has shown serious corrosion. corrosion The choice of using CRAs or chemical means to solve the more severe corrosion problem comes down to economics (available capital vs. long-term operating costs). Remoteness of operation becomes an important consideration in determining operating costs costs, as does downtime and deferred/lost oil because of repeated intervention for inhibitor application.

Availability and cost of platform space is a consideration for offshore facilities.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

The corrosion-control effectiveness of CRAs depends on the chemical severity of the environment environment. Crevice corrosion corrosion, pitting attack attack, and SCC are the primary concerns. The corrosion resistance of annealed austenitic stainless steels, such as 304 and 316, is affected by the presence of chlorides and temperature; type 304 is less corrosion resistant than type 316. Both materials are susceptible to SCC when the temperature is above 150°F. Both alloys are also low-strength low strength steels. steels Alloys 654 SMo and AL6XN can be manufactured to higher strengths and are more resistant to SCC. Austenitic stainless steels are probably the most susceptible of all ferrous alloys to pitting.

Martensitic stainless steels have had the widest range of use of any of the available CRAs. Such steels may be manufactured through heat treatment into tubular products with acceptable yield strengths for downhole tubing.

Many millions of feet of tubing type (grade L-80) 13Cr are in corrosive well service; i it iis considered id d th the material t i l off choice h i ffor d deep sweet-gas t wells ll with ith temperatures less than 150°C. About 35% of the L-80 13Cr usage was for oil wells. The p passivity y of 13Cr is destroyed y by y high g chloride levels,, p particularly y at high temperature, which can lead to pitting and crevice corrosion.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Duplex stainless steels are high-strength alloys achieved by means of cold working Such steels are more corrosion resistant than martensitic steels but working. are similar in resistance to SSC. Cold-worked duplex has been used to 0.3 psi H2S. Annealed duplex is more resistant to H2S and SSC than the cold-worked versions. Annealed duplex line pipe has been used in wet CO2 service (99%) without problems. 22Cr duplex steel has been used where pH2S was between 0 5 and 1 psi 0.5 psi. Such steels have been used successfully in HT/HP wells (e (e.g., g 350°F and 14,000 psi), producing no H2S. However the copresence of chloride, stress, and dissolved oxygen can induce SSC. Wells not exposed to even small amounts of oxygen have operated successfully.[13] The material most commonly used for sour service is AISI Type 4130 steel, modified by microalloy additions with a quenched and tempered microstructure (martensite).[14] C-110 steel has been used as casing in North S wells Sea ll (30 tto 60 b bar CO2 and d 30 tto 50 millibar illib H2S).

S) [15] An A overview i off CRAs and their use in sour service is given in Treseder and Tuttle.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

North Sea wells

Fion Zhang/ Charlie Chong

North Sea wells

Fion Zhang/ Charlie Chong

North Sea wells

Fion Zhang/ Charlie Chong

North Sea wells

Fion Zhang/ Charlie Chong

Nickel and cobalt alloys are used in the most severely corrosive conditions (high pressure, high temperature, and high H2S contents). C-276, a nickelbased alloy, can be used to 8,000 psi H2S and 400°F. Nickel alloys have found extensive use in the Mobile Bay fields. They are less expensive than the cobalt alloy MP35N previously used for such extreme conditions. Nickel alloys are also used as weld cladding for wellhead and valve equipment.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Chemical inhibitors As with scale problems problems, the appropriate addition of chemicals can often inhibit corrosion problems, including some effects of H2S. The delivery techniques are often the same, but the inhibition mechanisms and types of chemicals are different.

Neutralizing inhibitors reduce the hydrogen ion in the environment. Typically, they are:  Amines  Ammonia  Morpholine They are effective in weak acid systems but are stoichiometric reactants: one molecule equivalent of inhibitor per molecule of acid. They have found minimal use in the oil field.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Scavenging inhibitors are compounds that also remove the corrodant. Oxygen scavengers are commonly used in the oil field (e.g., (e g in removing oxygen during water injection).

The majority of the corrosion inhibitors employed during production form thin barrier layers between the steel surface and the corroding fluid. The concept is that the organic inhibitor will strongly adsorb on the metal wall to form a barrier barrier, possibly only a few molecules thick, which will prevent access to the corrodant and possibly leave the surface oil-wet (further retarding access of the corrodant). The generic name given to these compounds is â&#x20AC;&#x153;filming amines.â&#x20AC;?

This name is qualitatively correct in that most inhibitors are indeed nitrogencontaining, and the inhibitor does finally reside on the surface. The specific mechanism can be more complicated. For example, the inhibitor can interact with the corrosion product to increase its adherence and to lower its permeability. bilit Such S h layers l are likely lik l to t be b far f thicker thi k th than a ffew molecules. l l [7]

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Regardless of the specific mechanisms involved, the inhibitor must contact the metal substrate. substrate The general procedures are:  Tubing displacement  Displacement from the annulus  Continuous injection  Squeeze into the reservoir as liquid or gas  Weighed liquids/capsules/sticks  Vapor-phase inhibitors

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

The first two batch treatments are operated by pushing the inhibitor-containing fluid across the face of the production tubulars top top-down down (Item 1) or bottom bottom-up up (Item 2). The inhibitor film then persists on the metal surface for some period of time ranging from days to months, depending on the specific environment and materials.

Continuous injection is done done, if the well completion allows for a â&#x20AC;&#x153;macaroni macaroni stringâ&#x20AC;? reaching to the perforations. This technique often includes a simple-tocomplicated valving system; it should be remembered that valves can plug.

Injection through the annulus has also been used.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Inhibitor squeezing into the formation is an alternative. Here, the mechanism is different than that of scale inhibitor squeezes. squeezes The large amount of inhibitor that returns initially is not wasted, but is intended to coat the tubular and production equipment with an adsorbed, persistent film of inhibitor. The small amounts of inhibitor that subsequently desorb from the formation are intended to repair holes that are generated in the initial film.

Weighed liquids/capsules/sticks are all variations on the theme of placing inhibitor in the rathole where it is slowly released into the wellbore fluid, continuously depositing and/or repairing the protective film.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Vapor-phase corrosion inhibitors are organic compounds that have a high vapor pressure pressure, generating volatile corrosion inhibitors (such as some amines) that allow this inhibitor material to migrate to distant, and often otherwise inaccessible, metal surfaces within the container. Such inhibitors have been used on the Trans-Alaska pipeline to protect low-flow areas, dead legs, and the annular space in road casings and contingency equipment. The concept has also been applied to storage tank protection protection.[[16]]

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Filming-amine inhibitors are intended to protect steels from the action of natural corrodants in the produced hydrocarbon and water phases. phases They are â&#x20AC;&#x153;naturalâ&#x20AC;? generally not effective in protecting the steels from the acids used to stimulate wells or from the partially spent acids returning from such treatments. These tasks are accomplished by the inclusion of large dosages of different inhibiting chemicals with the stimulation acids. Such inhibitor systems are also available to handle low-alloy low alloy steels and CRAs in HT/HP conditions. conditions [[17]] Concern for stability of CRAs during matrix stimulation of deep hot wells has resulted in the use of organic acids such as acetic acid and formic acid rather than HCl. Inhibitor systems have been developed for these chemicals as well.

Note: Matrix Stimulation

http://www.glossary.oilfield.slb.com/Terms/m/matrix_stimulation.aspx

A treatment designed to treat the near near-wellbore wellbore reservoir formation rather than other areas of the production conduit, such as the casing across the production interval, production tubulars or the perforations. Matrix stimulation treatments include acid, solvent and chemical treatments to improve the permeability of the near-wellbore formation, formation enhancing the productivity of a well well.

Matrix stimulation is a process of injecting a fluid into the formation, either an acid or solvent at pressures below the fracturing pressure, to improve the production or injection flow capacity of a well.

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Cathodic protection This technology Thi t h l is i used d to t protect t t pipelines, i li offshore ff h platforms, l tf and d surface f equipment. Corrosion is an electrochemical process: give up p electrons  Iron atoms g  Electrons flow through the metal to the corrodant  Ion movement in the water film contacting both corrodant and iron metal completes the electrical circuit

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

In certain important cases, it is possible to reverse this current flow out of the steel surface by the application of an external power supply (i (i.e., e make the surface to be protected cathodic rather than anodic). The technology involved in employing cathodic protection must take into account:  Quantity of current required  Composition and configuration of the impressed current anode  Resistivity of the corroding medium  Size of the item being protected  Accessibility of the surface being protected  Length of the item being protected

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Fion Zhang/ Charlie Chong

http://petrowiki.org/Corrosion_problems_in_production

Fion Zhang Xitang Fion Zhang/ Charlie Chong

Fion Zhang/ Charlie Chong

Read More

Fion Zhang/ Charlie Chong

SULFIDE STRESS CRACKING – PRACTICAL APPLICATION TO THE OIL AND GAS INDUSTRY Becky L. Ogden Southwest Petroleum Short Course Texas Tech University, 2005

ABSTRACT The phenomenon of sulfide stress cracking (SSC) can result in catastrophic failures of pressurized equipment and piping, resulting in extensive damage, injuries and possible fatalities. Sulfide stress cracking was first identified as a serious problem in the oil industry in the late 1950’s with the development of deeper sour reservoirs. The high strength materials required for these wells began to fail as a result of brittle fracture that was later identified as SSC. Research began on this phenomenon and a task group was formed, which later became associated with the National Association of Corrosion Engineers (NACE), now known as NACE International. The T-1B committee of NACE published a recommended practice addressing the metallic material requirements for protection against SSC. This recommended practice was later issued in 1975 as the Materials Requirement MR-0175, known today as “Metals for Sulfide Stress Cracking and Stress Corrosion Cracking Resistance in Sour Oilfield Environments”. Recently, NACE International has issued the International Standard MR0175/ISO 15156 addressing multiple forms of cracking associated with the presence of aqueous hydrogen sulfide.

This paper will concentrate on, and identify, the requirements for SSC to occur and give designers and operators practical options for the prevention of SSC in equipment operating in an aqueous H2S environment. While this paper will primarily discuss SSC, some insight will be given to address the concerns of other forms of cracking.

Key Words: hydrogen sulfide, cracking, sulfide stress cracking (SSC), partial pressure, heat affected zone, post weld heat treatment, hardness, sour environment, metallurgy INTRODUCTION Aqueous hydrogen sulfide (H2S) in oil and gas production operations can result in many challenges.

H2S is a poisonous gas that can result in severe metal loss corrosion as well as catastrophic brittle fractures of pressurized equipment and piping. These brittle fractures to metallic structures can happen quickly, with little to no warning, or may take years of exposure to occur. Several variables can influence a material’s likelihood or its resistance to cracking from exposure to hydrogen sulfide. The physical properties of the material, the chemical properties of the material, and the environment to which it is exposed all play an important role in determining whether a material is susceptible to SSC. Sulfide stress cracking, or SSC, is defined by NACE as the “Cracking of a metal under the combined action of tensile stress and corrosion in the presence of water and H2S (a form of hydrogen stress cracking).” Through the review of this definition, several factors must be present for SSC to occur.

These factors are 1) a susceptible material, 2) tensile stress, 3) hydrogen sulfide, and 4) water. If any one of these factors is missing, sulfide stress cracking will not occur.

1

MATERIAL PROPERTIES Steel is essentially a combination of iron and carbon with minor amounts of alloying elements added that enable that iron/carbon combination to perform the mechanical and chemical requirements of a particular grade of steel, or alloy. The primary elements added include manganese, silicon, phosphorus, chromium, nickel and molybdenum. Each of these elements is added in varying concentrations so as to enhance the steel’s properties. However, with regards to sulfide stress cracking, or other forms of cracking, these alloying elements must be reviewed, and in some cases, minimized.

Materials have to be strong enough to perform under the conditions we require for our production conditions and designs. However, generally with strength comes brittleness. Steels must be strong to perform, yet ductile enough to prevent brittle fractures. A delicate balance must be obtained. As a result of laboratory testing and field experience, NACE MR-0175:2003 details the parameters of acceptable chemical composition, physical properties, manufacturing processes, and fabrication processes that will yield a material acceptable for use in a NACE defined sour environment. These parameters, as they pertain to carbon steel materials, will be detailed in a later section of this paper.

FACTORS AFFECTING SULFIDE STRESS CRACKING Generally speaking, an environment that produces hydrogen sulfide is considered “sour”. However, for the environment to be defined as NACE sour, it must exhibit characteristics that are favorable for the initiation of sulfide stress cracking. NACE MR0175:2003 defines the conditions in which SSC can occur. For the purpose of this discussion, a “sour” environment shall be one in which the conditions are conducive to cracking by hydrogen sulfide.

It is important to understand that the environments which can result in the SSC of materials are very specific in their compositions. SSC does not occur under all operating conditions. Several factors affect whether or not SSC will occur to a particular metallic structure. These factors include the alloy composition, the material’s yield strength and hardness properties, heat treatment, microstructure, fluid pH, partial pressure of H2S, total applied tensile stress and cold work, temperature, and time. How each of these impacts the SSC potential is discussed below.

Alloy Composition The composition of a metallic material determines its susceptibility or resistance to various forms of cracking when exposed to particular environments. Generally speaking, iron based materials, or ferrous metals, are more susceptible to SSC than nickel based alloys, or non-ferrous materials. Additionally, various levels of resistance/susceptibility to SSC can be found within a given family of alloys due to chemical compositional differences. Therefore, each material should be reviewed prior to use to ensure it is acceptable for the intended use.

Yield Strength and Hardness Properties In general, the higher the strength of an alloy, the harder the material and the more susceptible it is to sulfide stress cracking. Although yield strength is a true material property, hardness is not. However, 2

there is a correlation between the two measurements. Generally, the higher the yield strength, the higher the hardness value. In most commercial grades of ferrous alloys, the maximum strength level suitable for sour service use is 90,000 psi yield strength. This roughly correlates to 22 Rockwell C (HRC) or 235 Brinell (BHN) hardness. This is often quoted for ferrous steels used in a NACE sour service. However, through controlling steel chemistry and using special mill processing, this upper limit can be increased.

Testing and qualification of materials can be performed to determine its suitability for use in sour systems.

Although hardness is not a true material property, it is the preferred method of testing because it is simple and easy to perform, relatively non-destructive, and in most cases, portable. Hardness values can be utilized by manufacturers and procurement agents as a quality control method during the fabrication process or by the field personnel as a field inspection technique. Additional discussions on hardness determinations are discussed in the next section.

Heat Treatment The type of heat treatment applied to a particular alloy can affect the material’s microstructure and ultimately its susceptibility to sulfide stress cracking. A microstructure comprised of tempered martensite with fine grains will result in materials of superior resistance to SSC. Carbon and low alloy steels are acceptable in the as-milled condition as long as they contain less than 1% nickel, meet the hardness requirements, and are in one of the following heat-treatment conditions: hot-rolled (carbon steels only); annealed; normalized; normalized and tempered; normalized, austenitized, quenched and tempered; or austenitized, quenched and tempered.

It should be noted that field fabrication, cold working, and welding of “approved” materials can alter the microstructure, making the material susceptible to SSC. It may be necessary to thermally stress relieve the materials following these processes to “reinstate” their resistance to SSC. Microstructure Although susceptibility to SSC increases with increasing hardness, some microstructures are more susceptible to cracking than others at the same hardness levels. As stated above, the tempered martensite is more resistant to SSC than the tempered bainite or mixed structures of the same hardness.

Additionally, the degree of segregation and the type, size, shape and distribution of inclusions are other microstructural variables that can influence the resistance to sulfide stress cracking.

Fluid pH The higher the fluid pH, the more resistant materials are to SSC. This tendency enables drilling operations to utilize high strength materials in zones known to produce H2S. Although pH control is acceptable and manageable in drilling operations, it is not readily utilized in production scenarios.

Maintaining a constant pH in production would prove troublesome and impractical. Therefore, hardness limitations and alloy selections are the preferred method for controlling SSC.

3

Partial Pressure of H2S As the partial pressure of H2S increases, the susceptibility of a material to SSC increases. The partial pressure of H2S is defined as the portion of the total pressure associated with the specific component of interest, in this case, H2S. The partial pressure is calculated by multiplying the total system pressure by the mole fraction of H2S in the gas phase. If the calculated partial pressure of H2S is above 0.05 psia, in a gas system, SSC is possible. Figures 1 and 2 show the relationship to H2S, system pressure and partial pressure for gas and multiphase systems as illustrated by NACE MR0175:2003 edition. It should be noted that these limits are a “practical” limit; due to other factors affecting SSC, materials have failed at partial pressures below 0.05 psia. Therefore, care should be taken to review all factors involved in the material selection.

Total Applied Tensile Stress and Cold Work Different alloys possess different threshold levels at which SSC will occur. Understanding this threshold level will enable the designer to ensure that the stresses applied to a material will not result in cracking. It needs to be understood that the total stresses working on a material are the combination of both the applied stress (i.e., pressure) and the residual stress (fabrication/manufacturing stresses). The higher the applied stress on a material, the more susceptible to SSC it becomes.

Cold work or cold formed materials may be susceptible to SSC at hardness levels below HRC 22 (BHN 235). Cold working will alter the microstructure and increase the residual surface tensile stresses. For this reason, heat treatment is recommended for cold worked or cold formed low alloy steels before they are used in a sour environment. An annealing or normalizing heat treatment will return the material to its original SSC resistance following cold working.

Temperature The potential for SSC decreases as temperatures increase. Therefore, additional high strength tubing and casing materials can be utilized above threshold temperatures. However, if a well is to be completed, or operated, in a sour zone with a temperature above a threshold temperature for a particular material, the engineer must confirm that the environment in contact with the material does not drop below that critical temperature. Below this temperature, these high strength materials are susceptible to SSC and cannot be utilized. Table 1, an excerpt from the NACE MR0175:2003 standard, illustrates the temperature dependence of tubing and casing materials in oil and gas wells.

Time The general rule of thumb is “the longer the time of exposure, at a constant stress level, the greater the danger of SSC for susceptible alloys”. Under laboratory controlled conditions, it is possible to determine the time to failure of a given alloy under a particular set of conditions. However, in actual field conditions, projecting a time to failure is extremely difficult. The time it takes for a material to fail due to SSC is dependent on the aggressiveness of the environment and the degree of susceptibility of the material. SSC can happen quickly, or may take years to develop. Therefore, it is critical that a review of the materials and environment be conducted prior to specifying the completion equipment. SSC resistant materials should be utilized.

4

HARDNESS TESTING OF MATERIALS While hardness testing is a simple procedure, it must be performed correctly and must represent the material in the as-received or as-fabricated condition. Hardness, by definition, is the resistance of a metal to plastic deformation, usually by an indention. Hardness testers utilize an indenter which is forced into the metal surface by a known loading. The relationship to the area, or depth, of the indentions to the load applied is known as the hardness of the material. Hardness can be measured on multiple “scales”. NACE MR0175 utilizes the Rockwell C scale (HRC) or the Brinell scale (BHN). If hardness values are specified for the parent metal and any heat affected zones left in the as-welded or asmilled condition, there shall be a sufficient number of hardness tests performed to ensure the readings are below the specified value as noted within NACE MR0175/ISO 15156:2003 for that particular material. Controlling hardness is an acceptable method for preventing SSC. MR0175/ISO 15156:2003 does not specify the number or locations of hardness tests on the parent material. However, if hardness control is to be utilized for approving a welding procedure for use in sour services, specific locations and numbers of tests must be performed. These are noted within the International NACE MR0175/ISO 15156:2003 standard for Vickers and Rockwell Hardness measurements for fillet welds, butt welds and repair/partial penetration welds. These illustrations must be followed for weld procedure qualifications.

Figures 3 and 4 illustrate the survey method requirements for hardness measurements on butt welds.

WELDING AND ITS IMPACT ON SSC Welding is a “necessary evil” in sour systems in the oil and gas industry; however, steps can be taken to minimize its negative impacts. When steel is welded, the parent material and consumables are variables that must be reviewed and controlled. But these are not the only variables that need to be considered when welding in sour service.

The effects of rapid cooling in the heat affected zone (HAZ) of a weld can result in areas of localized hardness. The HAZ is that area around the actual weldment that has been exposed to high temperatures, but not high enough to actually liquefy the material. However, as a result of this heat input, phase transformations do occur, resulting in a microstructure has been “partially melted” and altered due to the heat of welding. This altered HAZ is now more susceptible to SSC due to its increased hardness. A parent material that was suitable and acceptable in regard to SSC in its as-received, as-milled condition may be susceptible to SSC following fabrication that involves welding. Therefore, fabrication processes involving welding must be reviewed for their potential impact on the SSC potential of the parent material.

Weld procedures can be written and qualified as being in compliance with NACE with regard to both SSC and other forms of cracking associated with the presence of aqueous hydrogen sulfide. The qualification requirements for hardness measurements traverse across the weld are detailed in the NACE MR0175/ISO 15156:2003 Standard as stated in the previous section on hardness measurements (again, see Figures 3 and 4).

However, in lieu of qualifying a weld procedure to NACE, one also has the option to post weld heat treat (PWHT) following the completion of the welding. The use of a PWHT technique tempers the welded material. This reduces the residual internal stresses created when the weld metal solidified, and tempers any martensite that may be present into a configuration of lower internal strain. The process of 5

PWHT will be specific for each type and thickness of material and the procedures are described within ASME Boiler and Pressure Vessel Code, Section 8, Division 1.

When specifying line pipe for sour service, it has been this author’s experience to prefer the seamless line pipe over Electric Resistance Weld (ERW) pipe. When ERW pipe was utilized, the specification always called for full body normalizing following manufacturing, verses only seam annealing following manufacturing. In my experience, this proved to provide better resistance to SSC when exposed to the severe H2S environments in the Permian Basin area of West Texas and Southeastern New Mexico.

USE OF PLATINGS AND COATINGS While platings and coatings are an acceptable barrier for generalized corrosion, they are not acceptable for use in the prevention of SSC, as per NACE MR0175-2003.

DETERMINING A SOUR ENVIRONMENT Hydrogen sulfide is one of the most serious corrosion agents encountered in the oil and gas industry. In addition to its ability to crack metals, it can also result in pitting corrosion with subsequent failures. The release of H2S as a result of corrosion or cracking can endanger the lives of people working around, or in near proximity to, the release point. H2S can be fatal at concentrations as low as 500 ppm. Therefore, designing equipment resistant to H2S cracking is critical. Additionally, prevention of corrosion by H2S is also highly recommended. Inhibition of corrosion can be obtained through material selection, internal coatings, or the application of corrosion inhibitor. However, to prevent cracking, the NACE standard must be strictly adhered to and followed.

It is the responsibility of the owner/user to determine whether a given environment falls within the parameters of a sour environment, thus requiring SSC resistant materials. Information concerning the environment’s operating pressure, H2S content, water content, pH and temperature all play a role in making this determination. Designing for SSC resistance is not only a prudent and good engineering practice, but it is a requirement of many regulatory agencies. Specifically, the Texas Railroad Commission Rule 36, the BLM On-Shore Order #6 and the New Mexico Statewide Rule #118 all specify that SSC resistant materials must be utilized in an H2S environment. Therefore, this is both critical for safety and regulatory compliance!

Note: Currently, the regulations still specify NACE MR0175, latest edition, as the standard for compliance. It is unknown as of the writing of this paper as to whether the agencies will adopt the new International MR0175/ISO 15156 Standard. However, producers should be aware of the changes published in the new standard and be prepared to make appropriate modifications to fabrication and engineering specifications.

Referring again to Figures 1 and 2 will enable the user/owner to evaluate his/her system based on H2S content and pressure, assuming the presence of free water.

It should be noted that hydrogen sulfide can be present naturally in produced fluids or can be introduced as a result of contamination by incompatible waters or sulfate reducing bacteria. Frequent surveys of non-sour, or “sweet”, fluids should be conducted to determine if hydrogen sulfide generation is 6

occurring over the life of a well or producing field. All safety precautions should be exercised when determining the concentration of H2S in production fluids. Because of the dangers associated with low concentrations of H2S, it is recommended to always assume H2S is present, regardless of the past history of a field, lease, or individual well.

OTHER FORMS OF HYDROGEN DAMAGE In addition to SSC, there are other forms of hydrogen damage and cracking that can occur in aqueous hydrogen sulfide environments. Because the potential for catastrophic failures associated with these forms of cracking also exists, NACE has recently published a joint international standard NACE MR0175/ISO 15156:2003. This standard addresses the concerns for all types of cracking associated with sour production and makes recommendations for materials and operating conditions to prevent such failures.

Hydrogen Induced Cracking (HIC) Hydrogen Induced Cracking (HIC) is defined as a “hydrogen attack induced by decarburization”. This type of attack occurs at elevated temperatures and is caused by atomic hydrogen permeating through the steel and reacting to form other gases. Hydrogen reacts with the carbon in the steel to form methane gas which cannot diffuse out of the steel’s matrix. Accumulation of this methane at grain boundaries and other steel discontinuities results in localized high stresses from which cracks can occur. HIC attack generally occurs at temperatures greater than 5000 F and is dependant on the hydrogen partial pressure.

Consult the API Publication 941 Steels for Hydrogen Service at Elevated Temperatures and Pressures in Petroleum Refineries and Petrochemical Plants for additional information on this phenomenon.

Step Wise Cracking (SWC) Step Wise Cracking (SWC) is defined as “hydrogen cracks which lie parallel to each other and are connected by cracks between them”. This type of cracking can be either discrete cracks or an array of cracks. The cracks that connect the “main cracks” and lead to SWC are caused by the shear stresses between the main cracks. This type of cracking can lead to catastrophic failures due to the potential for the cracks to propagate through the thickness of the material, resulting in a considerable loss of strength and ultimate failure.

Stress Oriented Hydrogen Induced Cracking (SOHIC) Stress Oriented Hydrogen Induced Cracking (SOHIC) is defined as “hydrogen induced cracking propagated by high internal stresses (typically hoop stress)”. This type of cracking is similar to HIC and SSC, but the cracking is transgranular, or across the grains, in the through thickness direction.

These cracks initiate and propagate in the direction normal to the applied stress, and are typically observed in the HAZ of relatively high hardness microstructures. The application of a high external stress (i.e., pressure) typically contributes to the failure.

7

Hydrogen Blistering Hydrogen Blistering is defined as the “subsurface cracking from absorption and concentration of hydrogen”. Blistering occurs when hydrogen enters the steel and combines into molecular hydrogen at defects present in the steel plate, typically non-metallic inclusions such as sulfides. Hydrogen blistering generally occurs in low pressure equipment such as tanks and pipeline equipment that are exposed to a corrosive environment that contains hydrogen sulfide.

Because of the nature of the manufacturing process of rolled plates, inclusions present in the steels become elongated, with the larger inclusions present and aligned along the centerline of the plate. It is at these larger inclusions that hydrogen blistering tends to occur. The high internal pressures present with the formation of the molecular hydrogen create high internal stresses within the steel that can greatly exceed the yield strength of a material and result in the formation of blisters. These blisters can often be visually observed on the exterior surface of steels in the form of an area of localized “swelling”.

They often resemble a “paint blister”, but they are indeed a blister in the steel plate.

The best prevention for this type of cracking is to specify “HIC resistant material” when procuring steel plates. This material has substantially lower sulfur content (usually 0.005% maximum sulfur) and is commonly calcium-treated for sulfur shape control. This lower sulfur content reduces the amount of inclusions present in the steel, thereby reducing the number of available sites for hydrogen to accumulate and form molecular hydrogen. The calcium treatments help to “round” any inclusions that may be present in the low sulfur steel, thereby making it more difficult for hydrogen to enter.

CONCLUSIONS It is the goal of design engineers and operators to prevent failures, whether they are annoying seeps or catastrophic failures. It is a considerable economic benefit for users of steels to understand the environment in which that steel will be placed and the hazards associated with its exposure to that environment. By understanding the various forms of cracking that can occur to steels, the designer/operator can implement specification changes, or modify the environment to eliminate that mechanism. This will extend the effective life of the equipment, reduce the potential for failures, reduce downtime associated with equipment repairs and make the production area safer with respect to equipment failure incidents.

It is vital when selecting a manufacturing/fabrication process to consider the preventative measures for reducing the susceptibility of steel to various forms of cracking. Some Regulatory agencies require that equipment be manufactured, fabricated, and maintained in a condition that is resistant to Sulfide Stress Cracking (SSC). Therefore, it is critical during the early stages of a project to specify “NACE” trim and “NACE” compliance on all equipment in a sour environment. However, it must be noted that any modifications made to such equipment following its installation must be made such that this “NACE” condition is not negated.

By following the Standards published by NACE, the designer/operator can be confident that the specified equipment is acceptable for use in a sour environment and is resistant to SSC. By applying the details found in the new International Standard NACE MR0175/ISO 15156: 2003, additional protection from other forms of cracking can be integrated into the design specifications.

8

Table 1: Acceptable API and ASTM Specifications for Tubular Goods (Copyright NACE International 2003, Used with Permission)

28

Table 2 â&#x20AC;&#x201D; List of equipment (Copyright NACE International 2003, Used with Permission)

NACE MR0175/ISO 15156 is applicable to materials used for the following equipment Drilling, well construction and well-servicing equipment

Permitted exclusions Equipment only exposed to drilling fluids of controlled composition a Drill bits Blowout Preventer (BOP) shear bladesb Drilling riser systems Work strings Wireline and wireline equipmentc Surface and intermediate casing

Wells, including subsurface equipment, gas lift equipment, wellheads and christmas trees

Sucker rod pumps and sucker rodsd Electric submersible pumps Other artificial lift equipment Slips

Flow-lines, gathering lines, field facilities and field processing plants

Crude oil storage and handling facilities operating at a total absolute pressure below 0,45 MPa (65 psi)

Water-handling equipment

Water-handling facilities operating at a total absolute pressure below 0,45 MPa (65 psi)

Natural gas treatment plants Transportation pipelines for liquids, gases and multiphase fluids

Lines handling gas prepared for general commercial and domestic use

For all equipment above

Components loaded only in compression

a

See A.2.3.2.3 for more information.

b

See A.2.3.2.1 for more information.

c

Wireline lubricators and lubricator connecting devices are not permitted exclusions.

d

For sucker rod pumps and sucker rods, reference can be made to NACE MR0176.

Key X H2S partial pressure, kPa Y in situ pH 0 Region 0 1 SSC Region 1 2 SSC Region 2 3 SSC Region 3 In defining the severity of the H2S-containing environment, the possibility of exposure to unbuffered condensed aqueous phases of low pH during upset operating conditions or downtime, or to acids used for well stimulation and/or the backflow of stimulation acid, after reaction should be considered.

29

Figure 1: H2S Partial Pressure Relationship for a Gas System (Copyright NACE International 2003, Used with Permission)

30

Figure 2: H2S Partial Pressure Relationship for a Multi-Phase System (Copyright NACE International 2003, Used with Permission)

31

Figure 3: Butt weld survey method for Vickers hardness measurement (Copyright NACE International 2003, Used with Permission)

32

Figure 4: Repair and Partial Penetration Welds (Copyright NACE International 2003, Used with Permission)

33

List of References:

NACE MR0175-2003, “Metals for Sulfide Stress Cracking and Stress Corrosion Cracking Resistance in Sour Oilfield Environments,” NACE International, (Houston, TX:NACE) NACE MR0175/ISO 15156:2003, “Petroleum and Natural Gas Industries – Materials for use in H2S Containing Environments in oil and gas production,” NACE International, (Houston, TX:NACE) McIntyre, D. R., Moore, E. M. Jr. : “Specified Pipe Fittings Susceptible to Sulfide Stress Cracking”, Materials Performance (January 1996), pp 64 – 66 Tsukano, T., et al, “Development of Sour Service Drillstring with 110-ksi Yield Strength”, SPE/IADC Drilling Conference, 1991, No. 22004 Tuttle, R. N., “What Is a Sour Environment?”, Journal of Petroleum Technology, March 1990, pp 260 262 Texas Administrative Code, Title 16: Economic Regulation, Part 1; Railroad Commission of Texas, Chapter 3: Oil and Gas Division, Rule 3.36; Oil, Gas or Geothermal Resource Operation in Hydrogen Sulfide Areas, 2004 Bureau of Land Management, Onshore Oil and Gas Order No. 6, 43 CFR 3160, Federal Register/ Volume 55, No. 226 New Mexico State Wide Rule 118: 19,15,3.118, Hydrogen Sulfide Gas (Hydrogen Sulfide)

34

Revision 3 Verfasser/Dokument PV Plates for sour Service

Requirements for steel plates in sour service

1

Sour Service

Revision 3 Verfasser/Dokument PV Plates for sour Service

an introduction in the world of hydrogen induced corrosion

2

Sour service damage is not a new issue !

The oldest reports about sour service steel damage more than 60 years old Many organisations (like NACE or EFC), oil- and gas companies, engineering companies are still improving regulations

The importance of hydrogen damage due to sour service is more and more recognised.

Verfasser/Dokument

The exploitation of sour gases and out of sour oil sources is rising. Often sweet sources get more and more sour.

Requirements for steel plates in sour service

3 3

Why this sensitivity to sour service damage?

Sour media are aggressive to steel structures, damages not easy to detect.

Health and safety of personnel and the public are in danger if precautions in survey of equipment and a right material selection are not adjusted.

Severe environmental pollution could be the consequence out of such damages.

Shutdowns due to material failures and the replacement of pressure vessels can cause dramatic economical loss.

Accident at Chicago refinery in 1984; 17 people killed.

Verfasser/Dokument

A really bad example:

Many good reasons for our full attention

Requirements for steel plates in sour service

4 4

Verfasser/Dokument

Union Oil absorber vessel failure resulting from cracks growing in HAZ with no PWHT

Requirements for steel plates in sour service

5 5

The view of the steel plate manufacturer Steel plate requisitions reflect an increasing demand for plates with improved properties for sour service Large variety of customer requests: - many specifications based on published recommendations or test methods (e.g. NACE MR 0175, TM0284...)

- in combination with the “in house”-experience and -prescriptions Aim of this paper: - general overview over the damaging mechanisms

Verfasser/Dokument

- general survey about the current specified requisitions for plate orders - Dillinger Hütte GTS possibilities to supply improved steel plates

Requirements for steel plates in sour service

6 6

Revision 3 Verfasser/Dokument PV Plates for sour Service

Damaging Mechanisms and Test Methods

7

What are the sour service corrosion mechanisms?

Hydrogen-Induced Cracking (HIC) & Hydrogen Blistering

Sulfide Stress Cracking (SSC)

probably to be taken into consideration:

Verfasser/Dokument

Stress-Oriented Hydrogen-Induced Cracking (SOHIC)

Requirements for steel plates in sour service

8 8

Cracking mechanism in the steel during H2S corrosion process

Acidic, H2S -containing medium Sulfide Ionics Hydrogen Sulfide Proton Hydrogen Atom

Verfasser/Dokument

Electrons

Molecular Hydrogen

Steel with typical small imperfections Requirements for steel plates in sour service

9 9

Corrosion reaction

H2S → 2 H+ + S2Fe + 2 H+ → Fe2+ + 2 Had Fe2+ + S2- → FeS H2S + Fe → FeS + 2 Hab

Verfasser/Dokument

2 Hab → H2

Requirements for steel plates in sour service

1

10 0

Schematical appearance of damage mechanisms in sour service

SSC

Requirements for steel plates in sour service

Blistering

Verfasser/Dokument

HIC / SWC

SOHIC

1

11 1

Corrosion at stress free prismatic specimens

HIC Hydrogen Induced Cracking

Definition as per NACE MR0175/ISO 15156:

Requirements for steel plates in sour service

Verfasser/Dokument

Planar cracking that occurs in carbon and low alloy steels when atomic hydrogen diffuses into the steel and then combines to form molecular hydrogen at trap sites.

1

12 2

NACE TM 0284-2003 “Evaluation of Pipeline and Pressure Vessel Steels for Resistance to Hydrogen-Induced Cracking”

HIC:

Stepwise internal cracking on different planes of the metal;

origin:

1984, for evaluation and comparison of test result

test solution:

pH 3 (sol. A) and pH 5 (sol. B) saturated with H2S

test specimens:

position (one end/mid width) , preparation, dimensions

duration:

96 h

evaluation:

metallographic examination of cross sections

acceptance crit.:

to be agreed between purchaser and supplier

documentation:

CLR, CTR, CSR values for each section, specimen, test

Requirements for steel plates in sour service

Verfasser/Dokument

no external stress

1

13 3

Verfasser/Dokument

Test specimen location acc. to NACE TM 0284

Requirements for steel plates in sour service

1

14 4

HIC test method acc. to NACE TM 0284 Test specimens 0 10

test solution Solution A

test duration: 96h test solution: saturated with H2S

- pH 3 - 5% NaCl, 0.5% CH3COOH - identical to Solution A of NACE TM 0177

Solution B - pH 5 - synthetic seawater acc. ASTM D1141

Requirements for steel plates in sour service

1

15 5

Verfasser/Dokument

20

H2S

Verfasser/Dokument

In Detail

Requirements for steel plates in sour service

1

16 6

test specimens during HIC-test

Requirements for steel plates in sour service

1

17 7

Verfasser/Dokument

HIC test vessel

sectioning of test

Examination of the polished sections:

specimens

b a T

faces to be examined

b

20 mm

25 m m

e 25 m dir m ing ll ro

on i t c

W CLR =

∑ a ⋅100%

CSR =

∑ (a ⋅ b) ⋅100%

CTR =

W

∑ b ⋅100 % T

W ⋅T

a = crack length

b = crack width

W = specimen length T = specimen thickness Crack distance < 0.5 mm => single crack

Requirements for steel plates in sour service

1

18 8

Verfasser/Dokument

2

a

25 m m m 5m

Verfasser/Dokument

HIC or SWC damage

Requirements for steel plates in sour service

1

19 9

A516 GR70 Amine Contactor1 Requirements for steel plates in sour service

Verfasser/Dokument

Hydrogen Blistering

1: NACE RP0296 2

20 0

Verfasser/Dokument

Hydrogen Blistering

A516 GR70 Amine Contactor1

Requirements for steel plates in sour service

1: NACE RP0296 2

21 1

Amine Contactor/Water Wash Tower1 Requirements for steel plates in sour service

Verfasser/Dokument

Blister Cracking

1: NACE RP0296 2

22 2

Corrosion at specimens under stress

SSC Sulfide Stress Cracking

Verfasser/Dokument

Definition as per NACE MR0175/ISO15156:

Cracking of metal involving corrosion and tensile stress (residual and/or applied) in the presence of water and H2S

Requirements for steel plates in sour service

2

23 3

NACE TM0177 â&#x20AC;&#x17E;Laboratory Testing of Metals for Resistance to Specific Forms of Environmental Cracking in H2S Environmentsâ&#x20AC;?

1977, revised 1986, 1990 and 1996 tensile test (sol.A); preferred by DH-GTS1 Bent-Beam Test (sol. B) C-Ring test (sol. A) Double-Cantilever-Beam test (DCB) (sol.A) 2 test solutions: A: pH: 2.7; B: pH: 3.5, H2S saturated test duration: 720 h or until failure, whichever occurs first results report: applied stress over log time (stress level of no fail. after 720h) remark DH-GTS: acceptable only if PWHT plus DICREST route! no microalloying elements 1: also 4 point bend test acc. ASTM G39, sol.A (typ. linepipe)

Requirements for steel plates in sour service

Verfasser/Dokument

origin: 4 test methods:

2

24 4

Verfasser/Dokument

Sulfide Stress Cracking

SSC in HAZ of head to shell weld of FCC absorber tower.

Requirements for steel plates in sour service

2

25 5

Verfasser/Dokument

Sulfide Stress Cracking

Requirements for steel plates in sour service

2

26 6

Verfasser/Dokument

SSC four-point bend test

Requirements for steel plates in sour service

2

27 7

Verfasser/Dokument

SSC tensile test

Requirements for steel plates in sour service

2

28 8

Corrosion at notched specimens under stress

SOHIC Stress Orientated Stress Cracking

Verfasser/Dokument

Definition as per NACE MR0175/ISO15156:

Staggered small cracks formed approximately perpendicular to the principle stress (residual or applied) resulting in a â&#x20AC;&#x17E;ladder-likeâ&#x20AC;&#x153; crack array linking (sometimes small) pre-existing HIC cracks.

Requirements for steel plates in sour service

2

29 9

Stress-Oriented Hydrogen Induced Cracking (SOHIC) New phenomenon in the field of sour gas corrosion %

Sporadic documentation at spiral welded pipes and flaws in pressure vessels.

Combination of rectangular (SSC type) and parallel cracks (HIC type) in the area of a multi dimensional tension field.

Typical SOHIC crack below a flaw.

Created in a double beam bend test.

Requirements for steel plates in sour service

3

30 0

Verfasser/Dokument

%

Verfasser/Dokument

Stress-Oriented Hydrogen Induced Cracking (SOHIC)

SOHIC-Crack at a non PWHT repair weld of a primary absorber (deethanizer)1.

1: NACE RP0296

Requirements for steel plates in sour service

3

31 1

Stress-Oriented Hydrogen Induced Cracking (SOHIC)

•issue - still under large discussion •mechanism not fully understood •mixture of SSC and HIC type cracking

Verfasser/Dokument

•location close to the welds

Requirements for steel plates in sour service

3

32 2

SOHIC test as per NACE TM0103 / 2003 • SOHIC testing • 4 point bent double beam tests • test duration 168 h • metallographic examination of the cross sections • Reasonable acceptance criteria for CCL (Continuous Crack Length), DCL (Discontinuous Crack Length) and TCL (Total Crack Length)

Verfasser/Dokument

are not yet reported

Requirements for steel plates in sour service

3

33 3

Verfasser/Dokument

SOHIC test arrangement as per NACE TM0103 / 2003

NACE TM0103 â&#x20AC;&#x201C; Full Size Double-Beam Test Specimen Design Requirements for steel plates in sour service

3

34 4

SOHIC test specimens as per NACE TM0103 / 2003

Dimensions of the notch: Depth = 2mm, r = 0.13mm

cut line

5c

notch

m

faces to be examined

n1 2 o i t c n Se tio c Se rop D Verfasser/Dokument

Sectioning across the notch into two cross sections.

(centred)

Requirements for steel plates in sour service

3

35 5

SOHIC evalutation of the cross sections from the double beam specimens

CCL - continuous cracks (perpendicular) in the most stressed area near to the bottom of the notch.

DCL - discontinuous (parallel) cracks below the continuous crack area, with lower stresses.

Verfasser/Dokument

TCL - length of the whole cracked area.

Requirements for steel plates in sour service

3

36 6

Results of the SOHIC tests at Dillinger HĂźtte GTS (1)

Although the tests were performed with HIC resistant DICREST material, at a load of less than 50% yield in pH3 solution first SOHIC type cracks appeared.

Rising the load increases the appearance of these cracks

Testing in pH5 solution no SOHIC cracks are detected.

Verfasser/Dokument

The notch of specimens generates a very (too ?) harmful stressed area.

Requirements for steel plates in sour service

3

37 7

Results of the SOHIC tests at Dillinger HĂźtte GTS (2) It should be taken into consideration, whether a notch like this is permitted generally at pressure vessels.

This could explain why even HIC and SSC resistant steels (DICREST) show big amounts of SOHIC cracks with the proposed test method.

Acc. to DHâ&#x20AC;&#x2122;s opinion this test method is not appropriate as SOHIC test.

SOHIC resistant material (acc. to this test method) can not be produced with

Verfasser/Dokument

normalised steels. It seems to be that Q+T material will reach this aim.

Requirements for steel plates in sour service

3

38 8

Standards

Verfasser/Dokument

SSC + HIC

Requirements for steel plates in sour service

3

39 9

NACE MR0175/ISO 15156 - 2003 “Petroleum and natural gas industries—Materials for use in H2S- containing environments in oil and gas production” • By the end of 2003 NACE0175/ISO15156 was published giving requirements and recommendations for the selection and qualification of carbon and lowalloy steels, corrosion-resistant alloys, and other alloys for service in equipment used in oil and natural gas production and natural gas treatment plants in H2Scontaining environments • 3 parts: - Part 1: General principles for selection of cracking-resistant materials - Part 2: Cracking-resistant carbon and low alloy steels, and the Verfasser/Dokument

use of cast irons - Part 3: Cracking-resistant CRAs (corrosion-resistant alloys) and other alloys

• Qualification route for steels not yet proved to be suitable for H2S service Requirements for steel plates in sour service

4

40 0

SSC in NACE MR0175/ISO 15156 - 2003 SSC:

Metal cracking under corrosion in presence of H2S and stress; same time hydrogen embrittlement especially in steel with high hardness or high strength

SSC and SCC susceptibility depends on e. g.: - steel: chemical composition, heat treatment, microstructure, cold deformation - hydrogen activity (pH-value) - total tensile stress (including residual stress) - temperature, duration, ...

Definition of SSC severity levels from 0 to 3 with increasing severity Verfasser/Dokument

severity level 1starting from H2S partial pressure ≥ 0.0003 MPa No absolute resistance, material can fail in SSC-tests!

Requirements for steel plates in sour service

4

41 1

SSC in NACE MR0175/ISO 15156 – 2003 Requirements:

Carbon & low alloy steels: - heat treated (contr. Rolled, N, N+T, Q+T); - Ni < 1% wt - Hardness < 22 HRC (average) < 24 HRC (individual) fabrication conditions: * welding and PWHT have to respect 22HRC limitation also in HAZ and WM * > 5% cold deformation Ö SR to be applied Remark of the steel producer: NACE MR0175 shall prevent SSC-Cracking, but there is very few influence on steel making practice (Ö no influence on HIC-resistance!!)

Requirements for steel plates in sour service

4

42 2

Verfasser/Dokument

Pressure vessel steels classified as P-No 1, group 1 or 2 in Section IX of the ASME Boiler and Pressure Vessel Code are acceptable without testing

Listing of Section IX of the ASME Boiler & Pressure Vessel Code

Spec

Grade

SA-283 A, B, C, D

P-No.1, Group 2

UNS

Spec

Grade

UNS

-

SA-299

...

K02803

SA-285

C

K02801

SA-455

...

K03300

SA-285

A

K01700

SA-515

70

K03101

SA-285

B

K02200

SA-516

70

K02700

SA-36

...

K02600

SA-537

Cl. 1

K12437

SA-515

65

K02800

SA-662

C

K02007

SA-515

60

K02401

SA-737

B

K12001

SA-516

55

K01800

SA-738

A

K12447

SA-516

60

K02100

SA-516

65

K02403

SA-562

...

K11224

SA-662

A

K01701

SA-662

B

K02203

Requirements for steel plates in sour service

Verfasser/Dokument

P-No.1, Group 1

4

43 3

HIC in NACE MR0175/ISO 15156 – 2003 - The user shall consider HIC and HIC testing even if there are only trace amounts of H2S present - HIC susceptibility is influenced by chemistry and manufacturing route Requirements - low Sulphur content ( < 0,003 %) - test acc. to NACE TM0284

Verfasser/Dokument

- acceptance criteria (solution A: CLR ≤ 15%, CTR ≤ 5%, CSR ≤ 2%) - other conditions may be defined as per table B.3 for specific or less severe duty

Requirements for steel plates in sour service

4

44 4

SOHIC in NACE MR0175/ISO 15156 â&#x20AC;&#x201C; 2003 User should consider SOHIC when evaluating carbon steels - Pre-qualification to SSC prior to SOHIC/SZC evaluation - Small-scale tests: unfailed uniaxial tensile (UT) & four point bend (FPB) specimen are metallographicly examined - UT-specimen : - no ladderlike HIC indications or cracks exceeding 0,5mm in through thickness direction allowed - after hydrogen effusion the tensile strength shall not be less than 80% of the tensile strength of unused specimens

- Full pipe ring tests may be used, test method and acceptance criteria described in HSE OTI-95-635

Requirements for steel plates in sour service

4

45 5

Verfasser/Dokument

- FPB-specimen: - no ladderlike HIC indications or cracks exceeding 0,5mm in through thickness direction allowed - blisters less than 1mm below the surface and blisters due to compression regardless of the depth shall be disregarded

EFC 16 “Guidelines on Materials Requirements for Carbon and Low alloy Steels for H2S-Containing Environments in Oil and Gas Production Combined specification for test methods of HIC and SSC” concerns: published:

C- and low alloy steels in oil and gas production (not in refinery service); conclusion of NACE-test methods in 1995, rev. 2 in 2002

1. HIC - low S, shape control, low segregation, low CEQ

Verfasser/Dokument

- test acc. to NACE TM0284, Solution A - acceptance criteria: CLR ≤ 15%, CTR ≤ 5%, CSR ≤ 1.5%

Requirements for steel plates in sour service

4

46 6

EFC 16 (2) 2.

SSC - f(pH-value/ H2S-p.pressure): Non sour, transition region, sour service - in case of sour service: see guidelines * limited hardness in HAZ to max. 250 HV30 except cap´s cap layer up to 275 HV30 (t < 9,5mm) or 300 HV30 (t > 9,5mm) * limited cold deformation (5% for PV) or PWHT > 620°/650°C - various test methods for the evaluation of SSC resistance (uniaxial, 4pointbend, C-ring,....); pH= 3; DH recommend the tensile test and 4 point bend test - load and duration of the test to be agreed; proposals are made recommendation of DH-GTS e.g.: load: 0.72 SMYS; duration: 720 h SOHIC/ SZC (Soft zone cracking)

Verfasser/Dokument

3.

- PWHT recommended - testing the susceptibility by 4 point bend test as an option, however no acceptance criteria defined Requirements for steel plates in sour service

4

47 7

Guidelines RP + MR

Verfasser/Dokument

HIC + SSC

Requirements for steel plates in sour service

4

48 8

NACE RP0472 – 2000 “Guidelines for Detection, Repair and Mitigation of Cracking of Existing Petroleum Refinery Pressure Vessels in Wet H2S Environments”

Concerns HIC, SSC, SOHIC, ASCC (Alkaline Stress C.C.)

applicable for existing equipment in refineries made of carbon steel

valid if H2S concentration ≥ 50 ppm (but no threshold concentration defined)

reports about the parameters for each damage mechanism

reports about a large survey (in 1990) of 5000 (!) inspected pressure vessels

26% of all vessels showed cracking incidence (crack depth from 1.6 mm to more than 25 mm) recommendations for inspection

Requirements for steel plates in sour service

Verfasser/Dokument

4

49 9

NACE RP0472 – 2000 (2) “Guidelines for Detection, Repair and Mitigation of Cracking of Existing Petroleum Refinery Pressure Vessels in Wet H2S Environments” Definition of environment to be more susceptible to HIC, SOHIC or blistering - process temp.: Ambient to 150 °C - H2S: > 2000ppm + ph > 7.8 - H2S: > 50 ppm + ph < 5 - presence of HCN + others

Recommendations for repair: - Hardness of production welds < 200 HB - Welding procedure qualification hardness < 248 HV10 for HAZ and WELD - PWHT to be considered Verfasser/Dokument

Requirements for steel plates in sour service

5

50 0

NACE MR0103 – 2003 „Materials Resistant to Sulfide Stress Cracking in Corrosive Petroleum Refining environments“ NACE MR0175: for oil- and gas handling systems NACE MR0103: for refinery service; it based on the experience with MR0175 and other NACE publications.

Specific Process Conditions: > 50 ppm H2S dissolved in H2O or if pH < 4 + some H2S or if pH > 7.6 + 20 ppm HCN > 0.05 PSIA H2S in gas phase Also reference to NACE RP0472 requirements

Requirements for steel plates in sour service

+

some H2S or if Verfasser/Dokument

• • • • •

5

51 1

NACE MR0103 – 2003 (2) „Materials Resistant to Sulfide Stress Cracking in Corrosive Petroleum Refining environments“ Responsibility of the user: - HAZ - hardness - Residual stresses - Rm increase Ö risk increase

Verfasser/Dokument

Hardness Base Metal < 22 HRC (or also 248 HV 10) Cold deformation < 5% otherwise stress relieved

Requirements for steel plates in sour service

5

52 2

Revision 3 Verfasser/Dokument PV Plates for sour Service

Production of HIC-resistant steels

53

How to produce HIC and SSC-resistant steel plates?

Basis: Well developed know how (Dillinger HĂźtte GTS has been engaged in this field for more than 20 years) Adequate production installations Permanent exchange with the endusers

Verfasser/Dokument

Follow up in international research projects

Requirements for steel plates in sour service

5

54 4

Requirements for homogeneous Dillinger Crack Resistant Steel plates

hot metal desulphurisation

deep vacuum degassing

special chemical composition (C, Mn, S, P)

cleanliness stirring by Argon

special casting parameter (no bulging, adapted superheating)

intensified QA-process

special care to avoid unacceptable segregations

high shape factor rolling (strong reduction in thickness per rolling pass)

Requirements for steel plates in sour service

Verfasser/Dokument

DICREST-route

5

55 5

Production route in the steel plant Hot metal desulphurisation

BOF converter

Argon stirring process

Heating

Degassing process

Casting

O2 CaC2 Mg

Ar O2

Ar Ar

Ar 2

objective: hot metal dephosphorisation slag desulphur- decarburisation conditioning, isation denitrogenisation steel desulphurisation

temperature adjustment

removal of: Carbon Sulphur Nitrogen Hydrogen

cleanliness avoiding: - reoxidation - resulphurisation

analysis adjustment

Requirements for steel plates in sour service

5

56 6

Verfasser/Dokument

Ar/N

Verfasser/Dokument

Requirements for steel plates in sour service

5

57 7

Inclusion distribution for different caster types Curved caster

90

Curved caster

Curved caster r = 5.0 m; vc = 1.0 m/min

80

Vertical caster vc = 0.5 m/min

Total oxygen in ppm

70 60 50 40 30 20 10 0 0

20

40

60

80

100

Distance from the fixed side in % Requirements for steel plates in sour service

5

58 8

Verfasser/Dokument

Vertical caster Vertcal caster Vertical caster

Aspects of quality assurance: HIC properties and cast length

HIC resistant steel

80

60

40

~ 75% of cast length

20

0

non sour gas

non sour gas

Verfasser/Dokument

Frequency for CLR (av. of 9 sections) < 15%

100 % > 96%

cast length begin

end

Results from HIC test, according to NACE TM 0284-96, for one single heat in dependence of the cast length of DICREST 15 pressure vessel steel; test solution acc. to TM 0284-96: A (pH3).

Requirements for steel plates in sour service

5

59 9

Verfasser/Dokument

Influence of High Shape Factor Rolling

Requirements for steel plates in sour service

6

60 0

Verfasser/Dokument

Optimized production steps for DICREST plates in the heavy plate mill

Requirements for steel plates in sour service

6

61 1

Aspects of quality assurance: casting incidents

additional additional testing prohibited from release testing

acceptance criteria

Verfasser/Dokument

cracking extend in HIC test

incident risk range

cast strand length position

Example of deviation in casting parameter combination Requirements for steel plates in sour service

6

62 2

Test laboratory of Dillinger Hütte GTS to measure sour gas susceptibility:

8 laboratory fume hoods (7 for tests, 1 for cleaning)

overall 39 connections for tests vessels

12 connections for SSC tensile tests (CorTest rings) equipped with computer aided monitoring of specimen failure

3 independent gas supply systems for parallel use of 3 different types of test gases

temperature adjustment and control system Verfasser/Dokument

Equipment:

Additionally health and safety-installations: gas detection systems, flame guard system to maintain H2S combustion, activated carbon filters in the exhaust air conduit, collecting tanks for all waste waters from the process Requirements for steel plates in sour service

6

63 3

Sour service

Verfasser/Dokument

What can DH offer?

Requirements for steel plates in sour service

6

64 4

Actual statistics of requested standards and thicknesses Requested thicknesses

> 80mm

< 40mm

About 5% of the overall DICREST tonnage is requested in grades other than SA 516

Requirements for steel plates in sour service

6

65 5

Verfasser/Dokument

40 - 80mm

Dillinger Hütte´s standardised offer for HIC resistant plates: DICREST

DICREST 5 DICREST 10

DICREST 15 1)

1)

acceptance criteria CLR

CTR

CSR

≤5

≤ 1.5

≤ 0.5

≤ 10

≤3

≤1

≤ 15

≤5

≤2

≤ 0.5

≤ 0.1

≤ 0.05

The requested test solution must be stated in the order in case of DICREST 15

CLR = Note: ETC = ELC =

∑ a ⋅100% W

CTR =

∑ b ⋅100% T

CSR =

∑ (a ⋅ b) ⋅100% W ⋅T

Acceptance criteria are defined as the average of all sections of all specimens per plate Extent of transverse cracking = bmax Extent of longitudinal cracking = a max

Requirements for steel plates in sour service

6

66 6

Verfasser/Dokument

grade

test solution max. acc. plate thickness TM 0284-96 A 80 mm (pH 3) A 80 mm (pH 3) A (pH 3) 150 mm B (pH 5)

No. of sections

plate thickness

CLR <

CTR <

CSR <

15%

5%

0.5%

30mm < t ≤ 40mm

15%

3%

0.5%

40mm < t ≤ 110mm

15%

3%

0.1%

10%

3%

0.5%

10%

2%

0.1%

t ≤ 15mm

5%

1.5%

0.5%

15mm < t ≤ 30mm

5%

1.5%

0.5%

30mm < t ≤ 110mm

5%

1%

0.1%

t ≤ 30mm 1

3

9 resp. 15*

t ≤ 30mm 30mm < t ≤ 110mm

Verfasser/Dokument

Actual acceptance levels for DICREST 5 plates in pH3 solution HIC tested in dependence on averaging the values for a certain no. of section

* No. of Sections acc. to NACE TM0284-03 for t > 88mm = 15 Remark: All other requirements on request Requirements for steel plates in sour service

6

67 7

Risk assessment on real HIC and Pseudo-HIC plates 90 80

HIC-resistant Pseudo-HIC

Percentage [50%]

70 60 50 40 30 20

0

<2

>2≤4

>4≤6

>6≤8

> 8 ≤ 10 > 10 ≤ 20 > 20 ≤ 40

> 40

Optimisation of CLR-values in NACE TM 0284-96, solution A through application of special DICREST-production route (steel grades: A 516 Gr. 60, 65 and 70; plate thickness 6-80 mm) compared to Pseudo HIC-plates with a package of certain Pseudo-HIC measures.

Requirements for steel plates in sour service

6

68 8

Verfasser/Dokument

10

Verfasser/Dokument

Requirements for steel plates in sour service

6

69 9

DICREST ex mill and ex stock

Verfasser/Dokument

www.ancoferwaldram.nl

Requirements for steel plates in sour service

7

70 0

Specification details of DICREST stock plates (AWS) thickness: 8 -80 mm grades SA 516 grade 60, 65 or 70 delivery condition: normalised toughness requirements acc. SA20-S5 HIC testing frequency: per heat on the thinnest and thickest plate HIC test per NACE TM0284-2003, solution A (pH3) hot tensile test at 400°C ultrasonic testing: acc. A578 (ed. 2001) S 2.2 Verfasser/Dokument

additionally: - conformity in harness and Ni-content to NACE MR0175 - banding check acc. to E 1268 once per heat for information

DiME specification is more customized especially for Middle Eastern market in thickness range from 10 to 50 mm Requirements for steel plates in sour service

7

71 1

Conclusion sour service becomes more and more important; research and standardising efforts further ongoing 2 (3) major failure mechanisms are important (HIC, SSC and probably SOHIC) SSC rules have low influence on steel making practice. The phenomenon is mostly seen at hard HAZ or hard base metal. DH-GTS applies DICREST route.

Verfasser/Dokument

SOHIC is not quite fully understood. Most appearances are related to failures in HAZ; no proper test method; research is going on. Q+T steels show advantages.

HIC resistant steels need a special manufacturing route and require a lot of experience & know how

Requirements for steel plates in sour service

7

72 2

Dillinger HĂźtte GTS is prepared for the needs of sour service

Verfasser/Dokument

We contribute with 450.000 t of HIC resistant1 linepipe and pv-plates per year

1

Requirements for steel plates in sour service

with certified HIC-resistance 7

73 3

... but we can help you take it with a smile!

Requirements for steel plates in sour service

7

74 4

Verfasser/Dokument

We can not transform sour to sweet...

Paper No.

11115

2011

NACE CORROSION 2011 STG 32 – Advances in Materials for Oil & Gas Production Mitigation of Sulfide Stress Cracking in Down Hole P110 Components via Low Plasticity Burnishing

Jeremy Scheel, Doug Hornbach, Paul Prevéy Lambda Technologies 5521 Fair Lane Cincinnati, OH, 45227-3401 USA Darrel Chelette, Peter Moore U.S. Steel Tubular Products, Inc. 10343 North Sam Houston Park Drive #120 Houston, TX 77064 USA

ABSTRACT Sulfide stress cracking (SSC) along with hydrogen embrittlement (HE) prevents the use of less expensive high strength carbon steel alloys in the recovery of fossil fuels in H2S containing ‘sour’ service environments that are commonly seen in deep well fossil fuel recovery efforts. High magnitude tensile stresses are generated by connection interferences created during power make-up of down hole tubular components. When subject to service loads the stresses are increased further providing the high tensile stresses necessary for SSC initiation. Because these alloys processed into high strength grades are not suited for fully saturated sour service environments, the current solution is to use or develop much more expensive alloys with increased corrosion-cracking resistance or limit their use to significantly weaker sour environments or higher operating temperatures.

©2011 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole, must be in writing to NACE International, Publications Division, 1440 South Creek Drive, Houston, Texas 77084. The material presented and the views expressed in this paper are solely those of the author(s) and are not necessarily endorsed by the Association.

1

Introduction of stable, high magnitude compressive residual stresses into less expensive carbon steel alloys alleviates the tensile stresses and mitigates SSC while also improving fatigue strength. This could allow the potential of using less expensive alloys in sour environments. Low plasticity burnishing (LPB) is highly effective when applied to metallic components using a proven reproducible process of producing deep, high magnitude compressive residual stresses in complex geometric components without altering the geometry, design or chemistry.

The LPB process, applied with advanced control systems, is presently being employed to treat components resulting in a substantial increase in service life through SSC mitigation and improved fatigue life. The benefits of LPB have been evaluated on full size specimens of uni-axial hoop stress loaded coupling blanks and C-ring specimens manufactured from quench and tempered API P110 grade steel with a yield strength of 132 ksi (910 MPa). Specimens were exposed to 100% NACE TM0177-Solution A at 1 bara H2S in both the LPB treated and untreated condition. The time to failure was documented along with the increase in life resulting from LPB treatment. LPB was successful in completely mitigating SSC in each test specimen up to 85% SMYS hoop tension; and in each case met or exceeded the 720-hour exposure time defined in NACE TM0177. At an applied fiber stress of 90% SMYS, the C-ring samples have exceeded exposures of 840 hours without failure. The initial results indicate that LPB processing of down hole tubular components may provide an alternative economical means of SSC mitigation and greatly reducing risk of component failure in sour environments.

Key Words: low plasticity burnishing, sulfide stress cracking, fatigue, residual stress, sour service

INTRODUCTION Surface enhancement of metals, inducing a layer of surface compressive residual stresses in metallic components, has long been recognized1-4 to enhance fatigue strength and mitigate stress cracking. The fatigue strength of many engineering components is often improved by methods including rolling or shot peening. Modern surface enhancement treatments such as low plasticity burnishing (LPB),5 laser shock peening (LSP),6 and ultrasonic peening,7 have emerged that in varying degrees benefit fatigue and stress corrosion prone components.

Maximum benefits are obtained when deep compression is achieved with minimal cold working of the surface.

Environmentally assisted cracking (EAC) in the form of Sulfide Stress Cracking (SSC), Stress Corrosion Cracking (SCC) and Hydrogen Embrittlement (HE) prevent the use of less expensive high strength carbon steel alloys in the recovery of fossil fuels in corrosive-cracking environments commonly seen in offshore and deep well recovery efforts. Tensile residual stresses generated from straightening, machining and connection make-up when added to applied stresses during down hole operations in high-pressure environments are significant contributors to EAC and fatigue failure. Because these alloys at high strength levels are not suited for sulfide or chloride environments, the current solution is to use or create much more expensive alloys with increased corrosion-cracking resistance to mitigate the problems or limit their use to significantly weaker sour environments.

Introducing compressive residual stresses into less expensive carbon steel alloys can dramatically reduce the risk of failure, mitigate SSC, HE and SCC, and improve fatigue

2

strength.8 This could allow the potential of using less expensive alloys in harsh environments where they currently are unable to be used. Low plasticity burnishing (LPB) is highly effective, reliable and reproducible method of producing deep compressive residual stresses in complex geometric components. With advanced control systems LPB can be applied using a closed loop feedback surface enhancement method capable of introducing a customized compressive residual stress field specifically tailored for each application. LPB so applied produces a very smooth surface finish, which aids in nondestructive inspection and examination. LPB tooling can be integrated with existing equipment used for manufacture and repair of down hole tubular products.

SSC susceptibility in high strength API P110 grade tubular products prevents their use in 100% H2S sour environments at temperatures less than 79°C (175 °F).9,10 As more deep wells and offshore resources are probed and recovered it is imperative to mitigate the problem of EAC in a cost effective manner. Laboratory testing in standard 100% NACE TM0177-Solution A 11 at 1 bara H2S liquid environment has shown that the LPB process can be employed to treat components with the effect of providing a substantial increase in service life, and SSC mitigation. The LPB technology is now being evaluated to determine if it can play a pivotal role in creating more reliable and efficient fossil fuel recovery systems that are capable of safely and reliably operating in aggressive environments. LPB processing has successfully been used to mitigate EAC in high strength steel, stainless steels and aluminums used in both the aerospace and nuclear industries. LPB technology was developed in conjunction with NASA’s SBIR program and is currently used in production of parts used in the aerospace, medical, and nuclear industries and on many different metal 12-16 alloys.

EXPERIMENTAL PROCEDURE Material C- ring specimens and full size coupling blank specimens were sectioned from a length of API P110 grade quench and tempered coupling stock. Figure 1 shows the coupling stock used to manufacture the specimens. Figure 2 shows an example of each of the 2 geometries tested in this investigation.

Figure 1: API P110 Quench + Temper coupling stock.

3

A

B

Figure 2: Examples of tested geometries: (A) C-ring specimen, and (B) full sized coupling blank in test fixture.

Specimen Processing LPB process parameters were developed to achieve nominally 0.040 in. (1 mm) depth of compression. Samples were processed on a CNC mill or lathe to allow positioning of the LPB tool in a series of passes along the region to be processed while controlling the burnishing pressure to develop the pre-determined magnitude of compressive stress with controlled low cold working. The full lengths of the outer diameter of the full sized coupling blanks were LPB processed. The C-ring specimens were processed on the exposed section of the outside diameter. The LPB process has been previously documented in detail.17 X-ray Diffraction Residual Stress Analysis X-ray diffraction residual stress measurements were made at the surface and at several depths below the surface on the outside diameter of both LPB and untreated specimens to characterize the residual stress distributions. Measurements were made in the axial direction employing a sin2Ď&#x2C6; technique and the diffraction of chromium KÎą1 radiation from the (211) planes of steel.

Material was removed electrolytically for subsurface measurement in order to minimize possible alteration of the subsurface residual stress distribution. The measurements were corrected for both the penetration of the radiation into the subsurface stress gradient and for stress relaxation caused by layer removal. The value of the x-ray elastic constants required to calculate the macroscopic residual stress from the strain normal to the (211) planes of steel were determined in accordance with ASTM E1426-9.18,19 Systematic errors were monitored per ASTM specification E915.20 Surface Roughness The improvement in surface roughness was documented for LPB vs. un-treated coupling material. Surface roughness measurements were performed on both the untreated and LPB treated coupling blanks using a standard surface roughness tester. The Ra surface roughness was calculated over a 0.50 in. (12.7 mm) evaluation length in the axial direction.

4

SSC Testing SSC Testing was conducted on 4 ½ in. (114.3 mm) API P110 quench and temper coupling stock (5 in. (127 mm) outside diameter). The coupling stock was sectioned into C-Ring specimens per NACE TM0177-Method C for testing and the full sized coupling stock blanks were machined to create an inside diameter of 4.375 in (111.1 mm) and provide a sealing surface for the internal pressure seals. Specimens were tested in both the un-treated condition as well as after LPB processing to determine the differential effects resulting from LPB treatment.

C-ring Testing: Testing was performed on LPB treated and un-treated C-ring specimens per NACE TM0177Method C. All testing was performed in 100% NACE TM0177-Solution A at 1 bara H2S at 25° C, the pH was monitored continuously throughout testing to ensure conformance to NACE TM0177. Specimens were sectioned from a length of API P110 coupling stock. The specimens were loaded initially to nominally 45% of SMYS. After exposure to at least 720 hours the specimens were tested at 80%, 85% and 90% of SMYS. Stress on the specimens was monitored continuously using strain gage rosettes placed on the inner diameter opposite the exposed location of maximum applied tension. The entire specimen except the outer gage region was coated in a polymer based stop off coating after loading and prior to immersion in solution. Figure 3 shows a C-ring specimen ready for testing.

Figure 3: C-ring specimen prior to testing.

Full Sized Pressurized Coupling blank Testing: Testing of full size coupling blanks was performed using a custom made holding fixture connected to a pressurizing test station. Specimens were tested in both the LPB treated and un-treated conditions. All tests were performed in 100% NACE TM0177-Solution A at 1 bara H2S at room temperature with the pH continuously monitored. The full sized coupling blanks were internally pressurized hydraulically to impart the desired amount of applied hoop stress.

Test solution was monitored for pH and refreshed as needed to conform to the NACE TM0177 standard. Testing was conducted until specimen failure or a run out life of 720 hours (30 days) or more was achieved per NACE TM0177 standard. Specimens were tested at 45%, 80%, and

5

85% of SMYS. Pressure was monitored continuously throughout the test and a timer was placed in the circuit to trip upon sample failure.

Figure 4: Full size coupling blank pressurized test apparatus and setup.

RESULTS AND DISCUSSION

X-ray Diffraction Residual Stress Analysis X-ray diffraction residual stress vs. depth results for untreated and LPB processed API P110 coupling blanks are presented graphically in Figure 5. Compressive stresses are shown as negative values, and tensile stresses as positive, in units of ksi (103 psi) and MPa (106 N/m2).

Compared to the untreated condition, LPB produced a compressive residual stress field with a much greater magnitude of compression (>10X) and over 2X the depth of compression. The magnitude of compression is near the SMYS of 110 ksi (759 MPa) at the surface. LPB produces much less cold working than conventional processes and ensures the deep fiber layers remain in stable compression, even at high temperature or in the case of mechanical overload as has been demonstrated in prior work.21

6

-3

20

0

200

Depth (x 10 mm) 400 600 800

1000

1200 100 0

0

-100

Residual Stress (ksi)

Quench + Temper Untreated

-40

LPB

-300 -400

-60

-500

-80

-600

-100 -120

-200

Residual Stress (MPa)

-20

-700 -800 0

10

20

30

40

50

-3

Depth (x 10 in.)

Untreated (Quench + Temper)

LPB

Figure 5: Residual stress comparison for LPB processed and un-treated material.

Surface Roughness The improvement in surface roughness after LPB processing was quantified using the Ra surface roughness. LPB improved the surface finish by a factor of 2.6X. This can aid in NDI examination as well as reduce friction in service. Figure 6 displays the results graphically.

Each value is an average of three repeat measurements.

7

300

AVERAGED SURFACE ROUGHNESS API P110 Steel Coupling Average of 3 Measurements

7

Surface Roughness - Ra (Îź in)

241.7 200

6 5 4

150

3 100 2

92.0 50

Surface Roughness - Ra (Îźm)

250

1

0

0

UNTREATED

LPB PROCESSED

Figure 6: Surface roughness for LPB processed and un-treated API P110 coupling blank.

SSC Testing The SSC testing data is presented graphically below in Figures 7 & 8. The un-treated C-ring specimen with an OD exposed surface failed in 10 hours at a stress of 45% SMYS, The LPB processed specimens exceeded the run-out life of 720 hours at 45%, 80%, 85% and 90% of SMYS exceeding typical hold-time requirements for testing in a sour service environment. The full sized coupling blank test results are very similar with the un-treated coupling blank failing after the entire OD surface was exposed for 37.5 hours. The LPB processed specimens exceeded 720 hours at 45%, 80% and 85% SMYS stress levels while surpassing typical holdtime requirements for sour service testing. The second full sized LPB coupling blank ran for a total of 1454.75 hours in solution before testing was terminated and the specimen was removed from solution for dye inspection, which revealed no cracking. These test results demonstrate the dramatic improvement achieved by the LPB treatment compared to the untreated P110 material.

A macro photo comparison of a failed untreated C-ring specimen and a run out LPB C-ring specimen is shown in Figure 9. Dye penetrent was used to reveal the axial SSC failure in the un-treated coupling blank shown in Figure 10. Figure 11 shows the LPB coupling blank after timed run out at 85% SMYS, with and without FDI developer, documenting that there are no cracks of any size beginning to initiate on the specimen. The SSC testing results show definitively that LPB is able to mitigate SSC cracking in common API P110 steel and dramatically increase the life. The 85% SMYS stress level is regarded as an aggressive performance test for metal that is in direct contact with a 100% H2S saturated environment.

API Specification 5CT uses 80% SMYS stress levels for C90 and T95 tensile specimens and the next edition will likely add 85% SMYS stress level for a new C110 grade.

8

API P110 STEEL C-RING TESTING NACE A Solution, 25° C LPB RUN-OUT PROCESSED EXCEEDED NACE TM0177 841 Hours 90% SMYS LPB RUN-OUT PROCESSED EXCEEDED NACE TM0177 820 Hours 85% SMYS LPB RUN-OUT PROCESSED EXCEEDED NACE TM0177 822 Hours 80% SMYS LPB PROCESSED 45% SMYS

RUN-OUT (EXCEEDED NACE TM0177 by > 2X) 1,719 Hours Test Stopped

UNTREATED FAILED (10 Hrs) (Quench +Temper) 45% SMYS 0

200

400

NACE TM0177 RUN-OUT 720 Hrs

600

800

1000

1200

1400

1600

1800

TIME (Hours)

Figure 7: C-ring testing results.

API P110 STEEL COUPLING PRESSURE TEST NACE A Solution, 25° C LPB PROCESSED 85% SMYS

RUN - OUT (734.5 hrs) 1454.75 hrs = Total Time Exposed

LPB PROCESSED 80% SMYS

RUN - OUT (720.25 hrs)

UNTREATED (Quench+Temper) 45% SMYS

FAILED (37.5 Hrs) NACE TM0177 RUN-OUT 720 Hrs 0

100

200

300

400

500

600

700

TIME (Hours)

Figure 8: Full Sized Coupling blank Test Results.

9

800

UNTREATED (FAILED)

LPB

Figure 9: Comparison of LPB treated and un-treated C-ring specimens after testing. The untreated specimen failed in 10 hours at 45% of SMYS. The LPB specimens ran-out at 45%, 80%, 85%, and 90% SMYS with no cracking.

AXIAL FAILURE

AXIAL FAILURE

Figure 10: FDI inspection of failed un-treated coupling blank revealing thru wall axial SSC.

Figure 11: LPB processed coupling blank and test fixture after run-out at 85% SMYS. FDI developer showing no signs of crack initiation.

10

CONCLUSIONS • • • • •

LPB imparted a deep compressive layer of stable residual compression over 2X deeper and 10X greater in magnitude than the untreated coupling blanks.

LPB was able to completely mitigate SSC failure in all tested specimens. The full sized coupling blank test exceeded the NACE TM0177 720 hour NACE A exposure time requirement at 45%, 80%, and 85% of the SMYS of 110 ksi (759 MPa).

The LPB processed C-ring tests exceeded NACE TM0177 time requirements at stresses levels equal to 45%, 80%, 85%, and 90% of SMYS. The untreated coupling blanks and c-ring specimens failed in 33 hrs and 10 hrs respectively at a stress load of 45% SMYS. Use of an engineered deep compressive stress field using LPB to mitigate SSC was successful on API P110 quench and temper coupling blank specimens.

REFERENCES 1. Frost, N.E. Marsh, K.J. Pook, L.P., (1974), Metal Fatigue, Oxford University Press.

2. Fuchs, H.O. and Stephens, R.I., (1980), Metal Fatigue In Engineering, John Wiley & Sons.

3. Berns, H. and Weber, L., (1984), "Influence of Residual Stresses on Crack Growth," Impact Surface Treatment, edited by S.A. Meguid, Elsevier, 33-44.

4.

Ferreira, J.A.M., Boorrego, L.F.P., and Costa, J.D.M., (1996), "Effects of Surface Treatments on the Fatigue of Notched Bend Specimens," Fatigue, Fract. Engng. Mater., Struct., Vol. 19 No.1, pp 111-117.

5. Prevéy, P.S. Telesman, J. Gabb, T. and Kantzos, P., (2000), “FOD Resistance and Fatigue Crack Arrest in Low Plasticity Burnished IN718,” Proc of the 5th National High Cycle Fatigue Conference, Chandler, AZ. March 7-9.

6. Clauer, A.H., (1996), "Laser Shock Peening for Fatigue Resistance," Surface Performance of Titanium, J.K. Gregory, et al, Editors, TMS Warrendale, PA, pp 217-230.

7. T. Watanabe, K. Hattori, et al,, (2002), “Effect of Ultrasonic Shot Peening on Fatigue Strength of High Strength Steel,” Proc. ICSP8, Garmisch-Partenkirchen, Germany, Ed. L. Wagner, pg 305310.

8. Paul S. Prevéy, N Jayaraman "Overview of Low Plasticity Burnishing for Mitigation of Fatigue Damage Mechanisms," Proceedings of ICSP 9, Paris, Marne la Vallee, France, Sept. 6-9,2005.

9. Snape, E.: “Sulfide Stress Corrosion of Some Medium and Low Alloy Steels,” Corrosion (June 1967) 23, 326-332.

10. Carter, C.S. and Hyatt, M.V.: “Review of Stress Corrosion Cracking in Low Alloy Steels With Yield Strength Below 150 ksi,” SCC and Hydrogen Embrittlement of Iron Base Alloy, NACE Reference Book No. 5 (1977) 524-600.

11. NACE Standard TM0177-2005: Laboratory Testing of Metals to Specific Forms of Environmental Cracking, NACE International.

12. J. Scheel, D. Hornbach, P. Prevey, “Mitigation of Stress Corrosion Cracking in Nuclear Weldments Using Low Plasticity Burnishing,” Proceedings of the 16th International Conference on Nuclear Engineering (ICONE16), May 11-15, 2008, Orlando, FL. 13. N. Jayaraman, P. Prevéy, “An Overview of the use of Engineered Compressive Residual Stresses to Mitigate SCC and Corrosion Fatigue,” Proceedings of 2005 Tri-Service Corrosion Conference, Orlando, FL, Nov. 14-18, 2005.

14. D.H. Hornbach and P.S. Prevéy, “Tensile Residual Stress Fields Produced in Austenitic Alloy Weldments,” Proceedings: Energy Week Conference Book IV, Jan. 28-30, Houston, TX, ASME International, 1997.

11

15. P.S. Prevey, et al. “Effect of Prior Machining Deformation on the Development of Tensile residual Stresses in Weld Fabricated Nuclear Components” Journal of Materials Engineering and Performance, vol. 5(1), Materials Park, OH; ASM International, 1996 pp. 51-56.

16. D. Hornbach, P. Prevéy, “Reducing Corrosion Fatigue and SCC Failures in 300M Steel Landing Gear Using Low Plasticity Burnishing,” Proceedings of 2007 SAE AeroTech Congress & Exhibition, Los Angeles, CA, September 17-20, 2007.

17. P. Prevey., “Burnishing Method and Apparatus for Providing a Layer of Compressive Residual Stress in the Surface of a Workpiece.” US Patent # 5,826,453, Oct. 27, 1998.

18. Cullity, B.D., (1978) Elements of X-ray Diffraction, 2nd ed., (Reading, MA: Addison-Wesley), pp.

447-476.

19. Prevéy, P.S., (1986), “X-Ray Diffraction Residual Stress Techniques,” Metals Handbook, 10, (Metals Park, OH: ASM), pp 380-392.

20. ASTM Standard E915, 2010, "Standard Test Method for Verifying the Alignment of X-Ray Diffraction Instrumentation for Residual Stress Measurement," ASTM International, West Conshohocken, PA, 2003, DOI: 10.1520/E0915-10, www.astm.org.

21. Paul S. Prevéy, “The Effect of Cold Work on the Thermal Stability of Residual Compression in Surface Enhanced IN718”, Proceedings of the 20th ASM Materials Solutions Conference and Exposition, St. Louis, MO, Oct. 10-12, 2000.

12

Hydrogen Induced Damage in Pipeline Steels

by Garrett R. Angus

Copyright by Garrett R. Angus 2014 All Rights Reserved

A thesis submitted to the Faculty and the Board of Trustees of the Colorado School of Mines in partial fulfillment of the requirements for the degree of Masters of Science (Metallurgical and Materials Engineering).

Golden, Colorado

Date: ______________

Signed: _________________________ Garrett R. Angus

Signed: _________________________

Dr. Kip O Findley Thesis Advisor

Golden, Colorado

Date: ______________

Signed: _________________________

Dr. Chester J. Van Tyne FIERF Professor and Interim Department Head Department of Metallurgical and Materials Engineering

ii

Abstract

The hydrogen induced cracking (HIC) resistance of several grades of plate steels was investigated using electrolytic hydrogen charging. HIC generated by electrolytic charging was also compared to the industrial standard test for HIC, the NACE standard TM0284. The electrolytic charging (EC) apparatus was designed to optimize the reproducibility of the HIC results and the robustness of the components during long charging times.

A characterization study on the EC apparatus was undertaken. Alterations to applied current density and charging time were conducted on a highly susceptible plate steel, 100XF, to assess HIC damage as a function of charging conditions. Intermediate current densities of 10 to 15 mA/cm2 produced the greatest extent of cracking without significant corrosion related surface damage. The hydrogen charging time did not greatly affect the extent and depth of cracking for test times between 24 to 48 hours. Thus, for subsequent experiments, the applied current density was set to 15 mA/cm2 and the charging time was set to 24 hours.

Plate steel grades X52, X60, X70, and 100XF were prestrained in tension to various levels and then electrolytically charged with hydrogen or tested with the NACE standard TM0284 test (solution A) saturated with H2S(g) to induce HIC. Prestrain was introduced to assess its impact on HIC. Hydrogen damage was quantified with the crack ratios defined in the NACE Standard TM0284. The results from the EC and NACE methods were very comparable to one, with respect to the magnitude of cracking and the trends between alloy and pre-strain conditions observed. Both methods showed that HIC substantially increased for the high strength 100XF steel compared to the lower strength alloys. This is consistent with NACE recommendations for HIC resistance steels, which suggests that alloy strength should be less than 116 ksi (800 MPa) or 248 HV (22 HRC). The HIC results were largely independent of the pre-strain levels imposed within the statistical accuracy of the evaluation method employed.

The total, irreversibly trapped, and diffusible hydrogen amounts were measured or estimated for each condition using a LECO interstitial analyzer and the American Welding Society method for measuring diffusible hydrogen concentrations. The total amount of diffusible hydrogen was highest for the 100XFalloy and lowest for the X52 alloy. The amount of trapped hydrogen was similar for all the alloys, implying that the number of irreversible trap sites were comparable. However, the diffusible hydrogen content was greatest for the 100XF alloy and lowest for the X52 alloy, which is believed to be related to the relatively high amount of grain boundary area and high dislocation density of the 100XF alloy.

A qualitative analysis on the effect of microstructure and nonmetallic inclusions on HIC was performed and produced results that confirmed findings from literature. Cracking was observed around nonmetallic inclusions such as sulfides and oxides in the metal matrix. For materials in which both inclusion types were present, X60 and X70, HIC originated and was observed most often around sulfide type inclusions.

iii

Table of Contents

Abstract ........................................................................................................................................................ iii Table of Contents ......................................................................................................................................... iv LIST OF FIGURES .......................................................................................................................................... vi LIST OF TABLES.......................................................................................................................................... xiii Acknowledgements .................................................................................................................................... xiv CHAPTER 1: 1.1

Introduction ....................................................................................................................... 1

Research Objectives ...................................................................................................................... 1

CHAPTER 2:

Background ....................................................................................................................... 3

2.1

Hydrogen Entry into Steel and Hydrogen-Assisted Cracking Mechanisms ................................. 3

2.2

Metallurgical Variables ................................................................................................................. 5

2.2.1

Deoxidation and Casting Techniques.................................................................................... 6

2.2.2

Microstructural Variables ..................................................................................................... 8

2.2.3

Mechanical Processing Effects ........................................................................................... 11

2.3

Experimental Methods to Test Plate Steels for HIC Susceptibility ............................................ 13

2.3.1

NACE Standard Test TM0284 ............................................................................................ 13

2.3.2

Electrolytic Charging .......................................................................................................... 15

CHAPTER 3:

Experimental Design and Methods ................................................................................. 17

3.1

Experimental Design ................................................................................................................... 17

3.2

Experimental Materials and Methods ......................................................................................... 17

3.2.1

Materials ............................................................................................................................. 17

3.2.2

Mechanical Properties ......................................................................................................... 18

3.2.3

Microstructural Analysis ..................................................................................................... 18

3.2.4

Introduction of Prestrain ..................................................................................................... 19

3.2.5

Hardness Traverse and Determination of Microstructural Dependence on Plate Thickness ……………………………………………………………………………………………..21

3.3

HIC Sample Generation and Preparation for HIC Testing ......................................................... 21

3.4

NACE Standard TM0284 (H2S Method) .................................................................................... 22

3.5

Electrolytic Charging Methodology (EC) ................................................................................... 26

3.5.1

Initial Design and Fabrication of New Charging Apparatus ............................................... 27

3.5.2

Characterization Study and Experimental Procedure for EC .............................................. 30 iv

3.6

Assessment of Hydrogen Damage .............................................................................................. 32

3.7

LECO® Hydrogen Analysis ....................................................................................................... 35

3.7.1

LECO® Hydrogen Analysis Sample Generation................................................................ 35

3.7.2

LECO® Hydrogen Analysis Experimental Procedure........................................................ 37

3.8

Diffusible Hydrogen Content Determined by Mercury Displacement ....................................... 37

3.8.1

Mercury Displacement Sample Generation ........................................................................ 37

3.8.2

Mercury Displacement Experimental Procedure ................................................................ 38

CHAPTER 4:

Results and Discussion.................................................................................................... 43

4.1

Microstructure and Nonmetallic Inclusions ................................................................................ 43

4.2

Mechanical Properties ................................................................................................................. 49

4.3

Characterization Study ................................................................................................................ 54

4.4

HIC Susceptibility Measurements .............................................................................................. 58

4.4.1

Influence of Prestrain on Calculated Critical Crack Ratios ................................................ 59

4.4.2

Influence of Mechanical Properties on HIC ........................................................................ 62

4.5

Hydrogen Analysis...................................................................................................................... 63

4.5.1

Average Hydrogen Contents Total and Trapped from LECO® Analysis .......................... 64

4.5.2

Average Diffusible Hydrogen Content from LECO® and Mercury Displacement Analysis ……………………………………………………………………………………………..66

4.5.3

Influence of Sample Location on Diffusible Hydrogen Results (Sample Size Effect) ....... 67

4.5.4

HIC Susceptibility and Diffusible and Trapped Hydrogen Contents from LECO® Analysis ……………………………………………………………………………………………..70

4.6

Evaluation of HIC with respect to Microstructure and Nonmetallic Inclusions ......................... 71

CHAPTER 5:

Summary and Conclusions.............................................................................................. 79

5.1

Electrolytic Charging .................................................................................................................. 79

5.2

HIC Susceptibility Measurements .............................................................................................. 79

5.2.1

Influence of Mechanical Properties .................................................................................... 79

5.2.2

Influence of Uniaxial Prestrain ........................................................................................... 80

5.3

Hydrogen Analysis...................................................................................................................... 80

5.4

Evaluation of HIC with respect to Microstructure and Nonmetallic Inclusions ......................... 80

CHAPTER 6:

Future Work .................................................................................................................... 81

References Cited ......................................................................................................................................... 82

v

LIST OF FIGURES Figure 2.1

Schematic representation of the dissociation of hydrogen sulfide gas at the metal/solution interface and subsequent diffusion into the metal to areas where hydrogen-assisted cracking can occur [5]. ......4

Figure 2.2

Schematic representation of enhanced plastic flow mechanism for hydrogen embrittlement. Parts A, B, and C show various configurations of inclusion size/elongation, distribution, and amount and the resulting morphology of cracks [3]. ........................................................................................................5

Figure 2.3

Schematic representation of the influence of ingot material location on HIC susceptibility of large sized ingots used to produce hot-rolled product [3]. ...............................................................................7

Figure 2.4

Schematic diagram illustrating the hydrogen penetration through two varying degrees of banding in a ferrite/pearlite microstructure: a) lower degree of banding and b) higher degree of banding. In a), the hydrogen penetration is hindered by pearlite along the ferrite regions. Adapted from [19]. ..................9

Figure 2.5

Examples of nonmetallic inclusions that form in fully killed steels. Shown are the changes of morphology due to rolling operations after casting. Adapted from [25]. .............................................. 10

Figure 2.6

HIC susceptibility (crack length ratio ratio) as a function of cold rolling reduction. Lattice strain was determined through X-ray diffraction after cold rolling operations denoted as line broadening in the above figure. HIC testing was performed with respect to the NACE standard test TM0284. Steel used for HIC testing was a 0.07-C, 1.22-Mn, 0.006-S (all values in wt pct) micro-alloyed with Nb and V heavily controlled rolled steel. Adapted from [20]. .............................................................................. 12

Figure 2.7

Effects of cold reduction at levels greater than 10 pct. Steels A and B have similar carbon equivalents, 0.31, and were produced from an ingot whereas, steel C was produced by continuous casting and has a higher carbon equivalent of 0.39. Steels A and C would be categorized are HIC-Resistant steels and steel B is a low sulfur grade steel. See text for more detailed chemical compositions of Steels A, B, and C. Adapted from [20]. .................................................................................................................... 13

Figure 2.8

Schematic illustration of the two reactions for hydrogen that occur at the metal/solution interface. The reaction (a) occurs at a much higher rate than (b). Reaction kinetics of (a) are suppressed in sour service applications due to the presence of sulfur ions, Equation 2.3. Reaction (b) causes subsequent HIC within the material through hydrogen interactions with features in the metal matrix [36]. .......... 14

Figure 3.1

Schematic of strips that were machined in the transverse direction with respect to the rolling direction and then subsequently prestrained to target levels. The prestrain and rolling directions are labeled.

Strip length was dependent on the plate material that was machined. The length ranged from 450 mm for the X52 to 700 mm for the 100XF. ................................................................................................. 20

Figure 3.2

Example of a strip being prestrained using uniaxial tension. The entire length between grips was considered the gauge length. This gauge length was used in correspondence with target prestrain levels to set a predetermined crosshead displacement to achieve the desired prestain amount. ........... 20

vi

Figure 3.3

Test specimen geometry and orientation used in the NACE Standard and electrolytic charging studies; all dimensions are in millimeters. For specimens that were subjected to electrolytic charging, a taper hole was drilled and reamed in the transverse top face of the test specimen to allow for electrical conductivity. ......................................................................................................................................... 22

Figure 3.4

a) NACE Standard TM0284 test vessel used to evaluate materials suscpetibilty to HIC, b) carousel used to hold test specimens allowing for adequate space for gas purge during testing. ....................... 26

Figure 3.5

Prior generation cell used for electrolytic charging by B. Rosner [7] and G. Angus [44]. ................... 28

Figure 3.6

From left to right, the images display the salient features for (a) the construction of the cathode electrode connection to be used in correspondence with (b) the steel specimens. The cathode electrode connection in (a) employs a taper pin fastened to a 316 SS rod. A corresponding taper hole is reamed into the top transverse face of the steel specimen. The cathode electrode connection is then inserted into the hole and frictionally locked with the test specimen. To protect this electrical connection, the region around the connection is covered in epoxy. ............................................................................... 29

Figure 3.7

Schematic diagram for dual cell used for electrolytic charging experiments. ...................................... 29

Figure 3.8

Electrolytic hydrogen charging (EC) test apparatus with components labeled. Electrolytic charging (EC) was accomplished through electrochemical polarization using this test apparatus. Tests apparatus was designed to optimize the reproducibility of the HIC results and the stability of the components during long charging times ................................................................................................................... 31

Figure 3.9

Test specimen geometry and orientation used in the NACE Standard and electrolytic charging studies; all dimensions are in millimeters, (a) Faces 2, 4 and 6 are used in the NACE Standard evaluation while Faces 1 - 6 are used for the electrolytic charging method. Face 1 is the furthest from the electrode link. (b) Schematic showing how cracks are measured on the evaluated faces [34]. ............ 32

Figure 3.10

ImageJ interface that allows the user to import selected images into the software for analysis. Once imported, the plate thickness is measured in pixels, shown by the white line. Knowing this distance in pixels, the “set scale” feature in the software allows the user to set the distance in pixels to a known distance. For this image and cross-section, 502 pixels equaled 12.42 mm. The “set scale” feature also displays the resolution of the image relative to the scale. This resolution was used to determine the accuracy of the measurements. ............................................................................................................. 33

Figure 3.11

Example cracked face showing how images were used to (a) identify cracks that are offset from one another. (b) Example cracked face showing the measurement offset to determine if multiple cracks should be considered a single (cracks separated by less than 0.5 mm were considered as a single crack [34]). (c) The final measurement of the single cracks. Each crack in (b) was enhanced using the “zoom” feature to produce the images in (c) to aid in the precise measurement of the single crack. The white lines in (c) represent the length, a, and thickness, b, measurements used to calculate critical crack ratios. ........................................................................................................................................... 36

vii

Figure 3.12

“Daisy Chain” assembly components. From left to right, the images display the salient features (the five sub-sized specimens and copper wire) of the component assembly leading to the assembly of the “Daisy Chain” to allow for simultaneous charging of multiple specimens for hydrogen content measurements. ...................................................................................................................................... 35

Figure 3.13

(a) Schematic representation with specimen geometry and the locations where test specimens were sectioned. The black areas represent material that was left after the sectioning process was complete. (b) Visual representation of samples fabricated for diffusible hydrogen analysis using EC. The sectioned areas allow for easy detachment after EC has been conducted. ............................................ 38

Figure 3.14

(a) Dimensions of the eudiometer used in the mercury displacement method to determine diffusible hydrogen amount, (b) schematic representation of the use of a mercury-filled eudiometer to capture and measure the amount of diffusible hydrogen in the sample that is placed into the assembly.

Adapted from [45]. ............................................................................................................................... 40

Figure 3.15

Test setup for the mecury displacement method with the compontents labeled. .................................. 41

Figure 4.1

Secondary electron micrographs taken with the FESEM of X52 plate steel. Images from three different planes (a) transverse, (b) longitudinal, and (c) normal plane are shown. 2 pct Nital etch. ..... 44

Figure 4.2

Presence of degenerated pearlite in the X52 plate material. Image taken with the FESEM. 2 pct nital etch. ...................................................................................................................................................... 44

Figure 4.3

Secondary electron micrographs taken with the FESEM of X60 plate steel. Images from three different (a) transverse, (b) longitudinal, and (c) normal plane planes are shown. 2 pct nital etch.

Black circles on each micrograph indicate the presence of a secondary microconstituent. .................. 45

Figure 4.4

Secondary electron micrographs taken with the FESEM of X70 plate steel. Images from three different planes (a) transverse, (b) longitudinal, and (c) normal plane are shown. 2 pct nital etch.

Black circles on each micrograph indicate the presence of a secondary microconstituent. .................. 45

Figure 4.5

Secondary electron micrographs taken with the FESEM of 100XF plate steel. Images from three different planes (a) transverse, (b) longitudinal, and (c) normal plane are shown. 2 pct nital etch.

Black circles on each micrograph indicate the presence of a secondary microconstituent. .................. 45

Figure 4.6

Nonmetallic inclusions observed in the four plate steels. Elements present in each image were confirmed by EDS mapping of the image shown. (a) Al-Ca-O X52, (b) Al-O X60, (c) Al-Mg-O-Ca-S X70, (d) Al-Mg-O 100XF (see pdf version for color). ......................................................................... 46

Figure 4.7

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the X52 plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. ............................................................................................................................ 47

viii

Figure 4.8

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the X60 plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. Regions 3 in (b) indicate the presence of MnS type inclusions identified through AFA. ........................................................................................................................................ 48

Figure 4.9

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the X70 plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. Regions 3 in (b) show the presence of MnS type inclusions identified through AFA. ..................................................................................................................................................... 48

Figure 4.10

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the 100XF plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. Regions 3 in (a) show a region where the types of inclusions within the area if evaluated with other elements i.e. Fe and O the composition of the inclusion would be found to be different. Speculated that these inclusions would be Fe-S and Al-O type inclusions. .......................... 49

Figure 4.11

(a) Microhardness data from each rolled face to the center of the X52 plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Cross-sections of the full plate thickness of the transverse plane with respect to rolling direction were evaluated. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time. .................. 50

Figure 4.12

(a) Microhardness data from each edge to the center of the X60 plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Cross-sections of the full plate thickness of the transverse plane with respect to rolling direction were evaluated. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time. .................. 51

Figure 4.13

(a) Microhardness data from each edge to the center of the X70 plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Cross-sections of the full plate thickness of the transverse plane with respect to rolling direction were evaluated. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time. .................. 52

ix

Figure 4.14

(a) Microhardness data from each edge to the center of the 100XF plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time. ................................................... 53

Figure 4.15

Surface conditions of 100XF after cathodic charging at various applied current densities while keeping the duration of the test constant at 24 hours. (a) 0.80, (b) 1.50, (c) 5.0, (d) 10.0, (e) 15, and (f) 25 mA/cm2. Image taken using light optical flash photography. .......................................................... 55

Figure 4.16

Hydrogen Induced Cracks produced through electrolytic charging of 100XF at a current density of 15 mA/cm2 for 24 hours, (a)Face #2 (b) Face #4. ..................................................................................... 57

Figure 4.17

Variation of calculated crack ratios versus applied current density for 100XF that was cathodically charged for 24 hours. CSR – Critical Size Ratio, CLR – Critical Length Ratio, and CTR – Critical Thickness Ratio. ................................................................................................................................... 57

Figure 4.18

Variation of average calculated crack ratios versus charging time for 100XF cathodically charged at an applied current density of 15 mA/cm2.............................................................................................. 58

Figure 4.19

Critical Crack Ratio values as a function of prestrain for the X52 material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio. ..................................................................................................... 59

Figure 4.20

Critical Crack Ratio values as a function of prestrain for the X60 material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio. ..................................................................................................... 60

Figure 4.21

Critical Crack Ratio values as a function of prestrain for the X70 material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio. ..................................................................................................... 61

Figure 4.22

Critical Crack Ratio values as a function of prestrain for the 100XF material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio. .................................................................................... 61

Figure 4.23

a) Crack length ratio dependence on tensile strength for the 0 pct prestrain condition: Electrolytic method (EC) and NACE Standard method (H2S). b) Crack thickness ratio dependence on tensile strength for the 0 pct prestrain condition. The vertical dashed lines represent the suggested 116 ksi threshold value for hydrogen assisted-cracking phenomena [47 and 48]. The two horizontal dashed lines represent sour service requirements outlined by ISO 3183 [52]. ................................................. 62

Figure 4.24

(a) Crack length ratio and (b) crack thickness ratio dependence on hardness for the materials and prestrain conditions evaluated by EC. The vertical dashed lines represent the suggested 22 HRC (248 HV) threshold value for HIC susceptibility [47 and 48]. The two horizontal dashed lines represent sour service requirements outlined by ISO 3183 [52]. ......................................................................... 63

x

Figure 4.25

(a) Hydrogen concentration in as-received alloys. (b) Hydrogen concentration for each plate material as a function of prestrain. ...................................................................................................................... 64

Figure 4.26

Trapped hydrogen values after degassing of material. Materials were hydrogen charged with the EC method, degassed at 50 °C for 72 hours, sectioned, and evaluated using the LECO® hydrogen analyzer to determine the trapped hydrogen content. ........................................................................... 65

Figure 4.27

Diffusible hydrogen values determined by two different evaluation methods on the four materials and prestrain conditions investigated: (a) Mercury displacement and (b) LECO® analysis. LECO® diffusible hydrogen values were determined using Equation 4.2. ........................................................ 67

Figure 4.28

Observation of hydrogen bubbles forming preferentially on the lower bottom end of the steel specimens during EC experiments. Sample numbers are shown on the steel sample in the image. ..... 68

Figure 4.29

Dependence of location of sample (sample number) on the amount of diffusible hydrogen measured by the mercury displacement method. Sample 1 represents the sample furthest away from the electrode connection. Dimensions of the sample were the full thickness of the plate (100 XF and X70 - 12.7, X60 - 9.5, and X52 - 19 mm) x 20 ± 3 mm (width) x 25 ± 2 mm (length). ............................... 69

Figure 4.30

Dependence of calculated Crack Length Ratio on the location of the examined faces for each material and prestrain condition for EC experiments (a) X52, (b) X60, (c) X70, and (d) 100XF. ..................... 70

Figure 4.31

HIC (CLR) dependence on (a) LECO® diffusible and (b) trapped hydrogen contents from EC. The two horizontal dashed lines represent sour service requirements outlined by ISO 3183, CLR < 0.15 [52]. ...................................................................................................................................................... 71

Figure 4.32

Light optical macrographs taken on as-polished full width and thickness EC test specimens. Images are representative of the cracking behavior observed for materials at all prestrain levels. (a) X52 0 pct prestrain Face 1, (b) X60 3 pct prestrain Face 2, (c) X70 5 pct prestrain Face 1, and (d) 100XF 0 pct prestrain Face 2. .................................................................................................................................... 72

Figure 4.33

Secondary electron micrograph taken on the FESEM of the X52 EC 18 % prestrain Face 1 condition.

Etched with 2 pct nital. Evidence of transgranular (1and 3) and intergranular crack propagation (2).. 73

Figure 4.34

EDS mapping of non-etched X52 EC 18 % prestrain Face 1condition. Primary crack propagation in close proximity to Al-O-Ca type inclusion. SEM image produced was taken in backscatter mode on the FESEM (see pdf version for color). ................................................................................................ 74

Figure 4.35

Secondary electron micrographs taken on the X60 EC 0 % prestrain Face 6 condition. Image captured using the ESEM. Etched with 2 pct nital. Features 1 and 2 show HIC around inclusions in the microstructure. ...................................................................................................................................... 75

Figure 4.36

Secondary electron micrograph taken on the X70 EC 5 % prestrain Face 1 condition. Image shows the presence of an inclusion in the primary crack area. Image was captured using the ESEM. Etched with 2 pct nital. ............................................................................................................................................. 75

Figure 4.37

EDS mapping of non-etched X60 EC 0 % prestrain Face 6 condition. SEM image was taken in backscatter mode on the FESEM (see pdf version for color). The SEM image and EDS maps show the presence of MnS type inclusions in and distributed around the primary crack region. ................... 76

xi

Figure 4.38

EDS mapping of non-etched X70 EC 5 % prestrain Face 1 condition. SEM image was taken in backscatter mode on the FESEM (see pdf version for color). The SEM image and EDS maps show the presence of MnS type inclusions in and distributed around the primary crack region. ........................ 76

Figure 4.39

Secondary electron images take on a) 100XF EC 0 % prestrain Face 2 (feature #1 shows transgranular cracking of acicular ferrite), b) 100XF EC 0 % prestrain Face 2 (feature #2 shows crack propagation across secondary microconstituent), and c) 100XF EC 2 % prestrain Face 1 (feature #3 shows nonmetallic inclusion). Image was captured using the FESEM. Etched with 2 pct nital. ..................... 77

Figure 4.40

EDS mapping of non-etched 100XF EC 0 % prestrain Face 2 condition. SEM image was taken in backscatter mode on the FESEM (see pdf version for color). SEM image reveals an globular shaped inclusion found outside the primary crack region. EDS mapping identifies the inclusion being of the mixed composition of Al-O-Ca-S. ........................................................................................................ 78

Figure 4.41

EDS mapping of etched 100XF EC 2 % prestrain Face 1 condition. SEM image was taken in backscatter mode on the FESEM. Etched with 2 pct nital (see pdf version for color). SEM image reveals an elongated inclusion found within the primary crack region. EDS mapping identifies the inclusion being of the mixed composition of Al-Mg-O. ....................................................................... 78

xii

LIST OF TABLES Table 2.1 Chemical Composition of the Materials used in [41] (wt pct) ..................................................................... 16 Table 2.2 NACE Standard Test Results and Results Using Electrochemical Charging on X70 Plate [7] ................... 16 Table 3.1 Chemical Composition of As-Received Plate Steels in wt pct .................................................................... 18 Table 3.2 Mechanical Properties in the Transverse Direction of As-Received Plate Steels ........................................ 18 Table 3.3 Selected Prestrain Amounts for the Plate Steels .......................................................................................... 21 Table 3.4 Test Specimen Geometry for the X52 Specimens used in each Hydrogen Exposure Method .................... 23 Table 3.5 Test Specimen Geometry for the X60 Specimens used in each Hydrogen Exposure Method .................... 24 Table 3.6 Test Specimen Geometry for the X70 Specimens used in each Hydrogen Exposure Method .................... 25 Table 3.7 Test Specimen Geometry for the 100XF Specimens used in each Hydrogen Exposure Method ................ 26 Table 3.8 Design Issues and Solutions taken for the Electrolytic Charging Methodology .......................................... 27 Table 3.9 Measured values from Manual Image J Measurements of length ‘a’ and thickness ‘b’ of Cracks in Figure 3.11. The Calculated Crack Ratio Values are also shown. ……………………………………………34 Table 3.10 Uncertainty from the Method used to Measure Variables for use in Crack Ratio Calculation .................. 35 Table 3.12 Uncertainty Values for the Instruments Used to Calculate Diffusible Hydrogen ...................................... 42 Table 3.13 Total Uncertainty Calculated using Equation 3.9 for the Mercury Displacement Method ........................ 42 Table 4.1 Microhardness Data Taken for each Material and Prestrain Condition ....................................................... 54 Table 4.2 Experimental Matrix for the Characterization Study Conducted on the 100XF alloy ................................. 54 Table 4.3 Charging conditions that produced cracking in 100XF that was cathodically charged for 24 hours ........... 57 Table 4.4 Average crack depth in 100XF specimens at the current densities explored during EC experiments run for 24 hours: Extremes represent the minimum and maximum crack depth observed ………………58 Table 4.5 Average crack depth in 100XF specimens as a function of test duration during EC experiments, run at 15 mA/cm2: Extremes represent the minimum and maximum crack depth observed ………………58 Table 4.6 Hydrogen trap sites found in iron, table modified from [54] ....................................................................... 66 Table 4.7 Crack Depth Observed on all Sectioned Faces for each alloy after EC ....................................................... 72 Table 4.8 Summarized Cracking Behavior .................................................................................................................. 78

xiii

Acknowledgements

I would like to thank the sponsors of the Advanced Steel Processing and Products Research Center for their support, guidance, and encouragement. Particular thanks are given to Evraz for allowing me to visit their RND facility to conduct the NACE standard test for my thesis work (Thank you Laurie Collins and Shahrooz Nafisi for your support and guidance during my thesis). Thank you Nucor for also allowing me to visit and give you sample to conduct AFA analysis on (Thank you Dan Edelman for all your support and guidance). Dr. Martins thank you for all your knowledge and guidance in this project, without your expertise and communication the test cell would be nowhere close to where it is today. Thank you Lee and Jim Johnson in aiding me in the rebuild of the Bridgeport Knee mill as well as your knowledge you bestowed upon me. Additional thanks go to Matt Merwin (U.S. Steel), D. Bai and Rick Bodnar (SSAB). I am grateful for my advisor Kip O Findley, for supporting me through this program and increasing my knowledge and skills within the field of metallurgical and materials engineering. Finally I would like to thank my thesis committee for the knowledge and advice they gave to me during my thesis.

xiv

CHAPTER 1: Introduction

Plate steels used in pipeline applications are exposed to specific environmental challenges. One deleterious environment is the transportation of “sour gas” commodities. The presence of hydrogen sulfide gas promotes hydrogen-assisted cracking; specifically, sulfur inhibits the recombination of hydrogen into hydrogen gas, thus facilitating hydrogen diffusion into steel.

Steel processing has evolved over the years to produce steel products that exhibit higher resistance to hydrogen induced damage. Specifically, processing strategies have included lowering sulfur contents (higher cleanliness of steel produced) [1], minimizing hydrogen contents (controlled cooling and low moisture in raw materials) [1], controlling the formation of specific microstructural constituents (more homogenous) [2], controlling nonmetallic inclusion shape control (limiting the degree of elongation) [6], and decreasing corrosion potentials while lower the diffusivity of hydrogen into the steel matrix (alloying) [6]. Also, casting techniques (ingot versus continuous cast) [4], deoxidation processes (fully versus semi killed steels) [6], rolling upon hot slab reduction (angle of rolls) [1], rolling [5], and controlled cooling (minimization of ‘hard’ bands and alloy segregation) [1] have been shown to affect hydrogen induced cracking resistance. Furthermore, plate manufacturing results in bending strains in the material, which may affect resistance to hydrogen induced damage through local increases in strength and decreases in ductility as well as increases in dislocation density and imposed residual stresses.

The industrial standard for the evaluation of hydrogen induced cracking (HIC) is the NACE TM0284 standard. This method relies on naturally occurring corrosion (reduction) reactions at the interface between one of two different solutions, A or B, containing dissolved hydrogen-sulfide gas, and the iron-containing base-metal.

Thus, hydrogen adsorption at the metal/solution interface and subsequent diffusion into the metal is accomplished.

The reduction of the dissolved hydrogen sulfide occurs at cathodic sites on the metal surface and the adsorption of atomic hydrogen is facilitated by the accompanying sulfide anion reaction-product that functions as a hydrogen association-inhibitor (reduces the rate of formation of diatomic-hydrogen gas). However, there is inherent danger in handling hydrogen sulfide gas, and special procedures are required to use it for hydrogen-assisted cracking evaluation purposes.

1.1

Research Objectives The scope of the current project is to develop a reproducible methodology using electrolytic charging as a

means to introduce hydrogen into steels to assess their resistance to hydrogen-assisted cracking. Electrolytic charging has been commonly used to pre-charge steel specimens with hydrogen before mechanical testing or to charge hydrogen into specimens while applying a load; however, it has not been extensively employed to assess hydrogen-induced cracking resistance [6]. Electrolytic charging at prolonged test durations was explored by Rosner [7], who found that electrolytic charging produced similar results to the NACE Standard test mentioned above, a promising initial result.

1

The current project utilizes four low carbon plate steel grades, X52, X60, X70, and 100XF, to evaluate hydrogen susceptibility in relation to: 1.

As-received mechanical properties i.e. strength, hardness, and ductility,

2.

The introduction of cold work in the form of uniaxial prestrain,

3.

Diffusible hydrogen amounts along with analysis of trapped hydrogen after degassing,

4.

Parent microstructure along with second phase microconstituents, and

5.

Presence of nonmetallic inclusions i.e. type, morphology, and/or size.

2

CHAPTER 2: Background

In the following section, the various proposed mechanisms for hydrogen-assisted cracking are outlined and discussed. Variables such as selection of raw materials for steelmaking, fabrication processes that occur during the production of steel plates, and the final forming methods are presented in relation to hydrogen induced damage. Two experimental methods to introduce hydrogen into steel to study the effects of hydrogen are also discussed.

2.1

Hydrogen Entry into Steel and Hydrogen-Assisted Cracking Mechanisms Hydrogen can be introduced into steel though steelmaking processes and in-service environmental

exposure. Exposure to hydrogen from steelmaking comes from melting of the raw materials, casting of ingots or slabs, and subsequent processing techniques (e.g. electro-plating and pickling). In-service environmental hydrogen exposure has increased due to the increasing demand for oil and gas commodities, leading to the exploring and harvesting of reserves that contain hydrogen sulfide gas and/or sulfur as impurities. In the oil and gas industry, the terms ‘sour’ and ‘sweet’ are used to denote the concentrations of the two impurities found in the commodity. For gas commodities, a ‘sour’ gas is defined as when the concentration of hydrogen sulfide gas exceeds 5.7 milligrams per cubic meter of natural gas. Oil wells are considered ‘sour’ when sulfur levels are at concentrations that exceed 0.5 wt pct. At lower levels of hydrogen sulfide gas or sulfur, the gas or oil commodity is described as ‘sweet.’

The production, interaction, and subsequent damage from hydrogen in steels are schematically shown in Figure 2.1. The production of nascent hydrogen atoms is generated through the naturally occurring corrosion reaction →

( (2.1)

in the sour environment [5]. Hydrogen ions are attracted to the metal interface due to the presence of excess electrons. Once at the interface, the nascent hydrogen can either recombine into hydrogen gas and bubble off the surface or diffuse into the steel and cause internal cracking and/or blistering at the surface. The diffusion of hydrogen into the steel is assisted by the presence of sulfide ions. The sulfide ions slow the recombination reaction of hydrogen ions to form hydrogen gas within the system. Internal cracking is generated though interactions hydrogen has with specific features, such as inclusions in the metal matrix. Hydrogen-assisted cracking phenomena are caused by the following proposed mechanisms [8]: internal hydrogen pressure (blister or void formation), surface adsorption, decohesion, and enhanced plastic flow. Surface adsorption affects the energy needed (decreases) to form brittle cracking [8]. Decohesion occurs because the dissolved hydrogen weakens interatomic bonds therefore lowering the ductility of a material. These two mechanisms (surface adsorption and decohesion) require an applied stress, either external or residual, to observe their consequences. However, the other two mechanisms (internal hydrogen pressure and enhanced plastic flow) can occur without an applied external stress; they are related to an internal stress caused by increased levels of hydrogen in specific areas. These mechanisms lead to premature failures or costly repairs of the steels used in service.

3

Figure 2.1

Schematic representation of the dissociation of hydrogen sulfide gas at the metal/solution interface and subsequent diffusion into the metal to areas where hydrogen-assisted cracking can occur [5].

Once hydrogen is introduced into the steel, it can lead to different forms of cracking such as sulfide-stresscracking (SSC), stepwise cracking (SWC), hydrogen induced cracking (HIC), hydrogen blister cracking, and stressoriented-hydrogen-induced-cracking (SOHIC). SSC is a form of hydrogen-assisted cracking that occurs in high strength steels (> 100 ksi yield strength) or in areas of localized hardness, i.e. â&#x20AC;&#x2DC;hard bandsâ&#x20AC;&#x2122; produced though alloy segregation or heat affected zones found close to weldments. This type of cracking is produced through exposure to hydrogen while under tensile stress. SSC is generally intergranular and is primarily oriented perpendicular to the applied tensile stress [9]. SOHIC is a form of hydrogen-assisted cracking where the damage caused through the interaction with hydrogen is indicative of an applied stress. SOHIC is characterized by the interlinking of microscopic cracks, both on the surface and within the steel, that are oriented perpendicular to the applied stress as well as the plane defined by nonmetallic inclusions. Areas that are highly susceptible to SOHIC are the same as those described for SSC and material adjacent to the weld seam in pipeline steels.

HIC covers a broad range of cracking behavior that occurs through the interaction with hydrogen. Within HIC, terms such as SWC, hydrogen pressure cracking, and hydrogen blister cracking are interchanged frequently.

For the purposes of this document, all will be combined and referred to as HIC. HIC is caused through the internal interaction of hydrogen with the steel in the absence of an externally-applied stress. Hydrogen is believed to affect a material through its interaction with crystalline features such as dislocations and/or secondary micro-constituents such as carbides and non-metallic inclusions [8]. At these features, hydrogen becomes trapped and builds to pressures greater than the local strength of material, causing internal micro-cracks to form.

Enhanced plastic flow has been one of the more accepted mechanisms associated with HIC and is shown schematically in Figure 2.2. This proposed mechanism for enhanced plastic flow is as follows: 1) hydrogen diffuses through the metal matrix and becomes trapped at the metal/inclusion interface, 2) the presence of molecular hydrogen causes a separation between matrix metal and inclusions, 3) plastic deformation occurs around the crack tips; 4) plastic regions are embrittled by hydrogen ( stage 1); and 5) cracks propagate through the embrittled regions in a direction normal to blisters cracks (stage 2) [3]. Scenarios A, B, and C in figure 2.2 represent various configurations of inclusion size, amount, and spacing. Scenario A represents two elongated type inclusions that are separated, B represents a cluster of smaller inclusions that within close proximity to one another, and C shows a very

4

elongated inclusion in the presence of two minor inclusions. The crack morphology is altered from scenario A to C due the differences in the features of the inclusions.

Figure 2.2

2.2

Schematic representation of enhanced plastic flow mechanism for hydrogen embrittlement. Parts A, B, and C show various configurations of inclusion size/elongation, distribution, and amount and the resulting morphology of cracks [3].

Metallurgical Variables This section evaluates important metallurgical variables that are considered when producing plate steels.

Variables such as different casting techniques, microstructure, nonmetallic inclusions, and alloy chemistry are important to monitor to increase steel resistance to HIC. Additionally, the effect of thermomechanical processing route and forming steps required for final pipe production from plate product are also considered.

Before addressing these variables, it is important to distinguish the different classes of pipeline steels that are used within the oil and gas industry. These steels are as follows: 1) conventional steels, 2) low sulfur conventional steels, 3) “HIC-Resistant” Steels, and 4) ultra-low sulfur advanced steels [10].

The term “conventional steel” refers to commercially produced hot rolled or normalized steel. When the environment is not challenging, the steels generally have a moderate level of inclusions and higher amounts (compared to others mentioned here) of sulfur present in the final chemistry of the steel (≥ 0.010 wt pct sulfur) [10].

These steel grades, when placed into sour environments, exhibit high degrees of hydrogen induced damage, which is to be expected from the higher levels of sulfur and inclusions present in the steel.

Low sulfur conventional steels are produced in a similar manner to conventional steels but with more controlled sulfur levels. The sulfur levels range from 0.003 to 0.010 wt pct. Due to the decreased levels of sulfur, the amount of inclusions present is lower and these steels exhibit improved mechanical properties compared to conventional steels. When placed into a sour environment, low sulfur conventional steels perform better than conventional steels but may still exhibit a high susceptibility to cracking at moderate to severe service conditions [10].

5

HIC-Resistant steels denote steel that has been engineered at multiple levels of processing such that this material can be placed into moderate and severe environments and not fail due to hydrogen induced damage.

Processing during ladle metallurgy involves the reduction of sulfur to levels designated as ultra-low, i.e. ≤ 0.002 wt pct, and a possible addition of calcium to aid in nonmetallic inclusion shape control during subsequent processing [10]. After hot rolling, these steels undergo a normalizing heat treatment to modify the rolled microstructure to improve the resistance to hydrogen induced damage.

Ultra-low sulfur advanced steels have been specifically engineered to resist HIC under severe service conditions and improve resistance to SOHIC through the use of modern steelmaking and processing techniques.

These steels, like the HIC-Resistant steel, have ultra-low sulfur contents and lower carbon equivalents at similar tensile strengths as conventional steels (such as ASTM A516-70). Ultra-low sulfur advanced steels undergo thermomechanically controlled processing (TMCP) and/or accelerated cooling techniques to produce ferritic/bainitic microstructures where banding is minimal or not present [10]. Because of the TMCP, attention to sulfur contents, and nonmetallic inclusion shape control, these steels exhibit the highest resistance to hydrogen induced damage of the four grades of steel mentioned.

2.2.1

Deoxidation and Casting Techniques Resistance to hydrogen induced damage is altered by casting techniques, due to the control over alloy

segregation and regions of nonhomogeneous microstructure. Variables such as deoxidation process (‘fully’ versus ‘semi’ killed), casting technique (ingot versus continuous cast); super heat in the tundish, casting speed, and position of rolls can alter the how the final steel product solidifies and where nonmetallic inclusions are present.

Two types of deoxidation processes can be used before casting. The terms “fully-killed” and “semi-killed” are in reference to the deoxidation process before casting and the elements that are added to achieve deoxidation. In “fully-killed” steels, aluminum is added because its strong affinity for oxygen limits the gas evolution in the bath of liquid steel [11]. Aluminum oxides form in the liquid and float to the top but some oxide inclusions remain in the steel upon casting. In comparison, “semi-killed’ steels have ferrosilicon added to achieve deoxidation within the bath of liquid steel [12]. “Semi-killed” steels produce heats that have more homogenous chemistries as well as more uniform mechanical properties after rolling operations [12]. In a study conducted by Moore and Warga [12] on 14 candidate steels for pipeline, HIC susceptibility was greater in fully-killed steels than semi-killed steels. The susceptibility was independent of the varying levels of sulfur, manganese, and silicon in the alloys but did depend on the amount of aluminum. The higher amount of deoxidation that occurs in the fully-killed steels leads to higher solubility of sulfur in the melt. The increase in solubility results in more Type I ellipsoidal manganese sulfides (MnS) upon solidification, which are subsequently elongated during rolling, leading to Type II stringer-like MnS [12]. The Type II MnS inclusions increase the degree of cracking in plate steels subjected to sour environments. In the semi-killed steels, the sulfides present after solidification and rolling operations are more globular, increasing the resistance to hydrogen-assisted cracking phenomena.

Herbslab et al. [4] showed that continuous casting should be used instead of ingot casting because the steelmaker has better control over the position of the nonmetallic inclusions. Nieto et al. showed that continuous casting enabled them to control the final solidification of slabs to minimize centerline segregation. They also showed

6

that continuous casting, low superheats in the tundish (<250°C), and slower casting speeds produced plate steels (API X65) with high resistance to HIC [1] due to no centerline segregation, homogeneous chemistry though thickness, and microstructural uniformity.

In contrast, when plates are produced from large ingots, it has been shown that HIC susceptibility scales with the position in the ingot [3]. In a study conducted by Ikeda et al. [13], samples were taken from four different locations within a large ingot and tested for HIC susceptibility. The HIC susceptibility in relation to these regions is shown in the ingot schematic in Figure 2.3. The positions examined were1/6 of the width (W), 1/3 W, 5/12 W, and 11/12W from the surface, as well as the lower section (1/6 of the height) of the ingot away from the slag layer (Figure 2.3). This study showed that in the area labeled as ‘A’ (the outer edge and lower portion of the ingot), the HIC susceptibility of the final plate is not severe. Towards the center of the ingot, (i.e. positions of higher concentrations of impurities) denoted as areas ‘B and C’, the HIC susceptibility increases moderately compared to ‘A’. Samples taken from the inner third of the ingot, section D, produced the lowest resistance to HIC due to the large amounts of alloy segregation and high concentrations of nonmetallic inclusions [13]. The importance of this study showed that steels produced from ingots have varying degrees of susceptibility and those from sections B, C, and D should not be used in sour environments.

Figure 2.3

Schematic representation of the influence of ingot material location on HIC susceptibility of large sized ingots used to produce hot-rolled product [3].

7

2.2.2

Microstructural Variables This section highlights the different microstructural variables that are present in the final plate steel and

how they influence the susceptibility to hydrogen-assisted cracking. The parent microstructure, presence of secondary microconstituents, non-homogenous microstructure, chemical make-up of nonmetallic inclusions, size and shape of nonmetallic inclusions, and finally alloying influence susceptibility to hydrogen-assisted cracking. In general, it is recommended by the National Association of Corrosion Engineers (NACE) that plate steels for use in sour service environments should not exceed a hardness value of 22 – 25 Rockwell C (HRC) [3]. Based on this recommendation, steels with an ultimate tensile strength of 116 ksi (800 MPa) or greater should not be considered for use in hydrogen environments. It is accepted in literature [14 and 15] that as the strength of the material increases, the tolerance to hydrogen concentration decreases and its overall mechanical properties (toughness, fracture toughness, tensile strength) are altered by the interaction with hydrogen. “Harder” microstructures inherently are generally less ductile, and require lower concentrations of hydrogen to cause severe damage to the steel.

Acicular ferrite grains have a three-dimensional morphology of thin, lenticular plates, where grain size is characterized by an aspect ratio that ranges from 3:1 to 10:1 [16]. Acicular ferrite steels exhibit high strength coupled with high ductility, and have a higher resistance to hydrogen compared to polygonal ferrite or mixed ferrite/pearlite microstructures. The degree of brittle intergranular fracture from the presence of hydrogen is reduced in acicular ferrite compared to polygonal ferrite [14 and 15]. In one study, tempering acicular-ferritic steels at 650 °C for 30 minutes increased the resistance to HIC thorough the elimination of hard transformation products [3].

The tempering step allowed sufficient carbon diffusion out of martensitic islands, decreasing the susceptibly of cracking around these features. Increases in HIC resistance were also tied to increases in the uniformity of the microstructure, increases in fracture toughness (lower hardness), and internal stress relief.

Grain size effects have been shown to alter HIC susceptibility in polygonal ferritic steels. Grain coarsening in polygonal ferrite has pronounced adverse effects on hydrogen-assisted cracking due to the ease of transgranular crack propagation in larger grains. Grain boundaries have been shown to reduce crack propagation through crack tip shielding [17]. Therefore, increasing the amount grain boundaries and decreasing grain size increases the degree to which a crack interacts with grain boundaries, reducing hydrogen-assisted cracking. However, at very fine grain sizes, it has been hypothesized that the increased grain boundary surface area associated with small-grained structures increases the probability for monoatomic hydrogen to segregate to grain boundaries, increasing the susceptibility to hydrogen-assisted cracking [18].

Pearlite in steel interacts differently with hydrogen than ferrite. Cementite in the pearlite structure acts as a retardant for hydrogen diffusion through the microstructure, but hydrogen reaches similar concentration levels to purely ferritic steels after longer exposure times [19 and 20]. It has been reported that HIC susceptibility scales with the degree of banding present in the parent microstructure [3, 10, and 21]. Hydrogen diffusion through a mixed ferrite/pearlite microstructure is shown Figure 2.4. The arrows indicate the path by which hydrogen diffuses though microstructure. Figure 2.4a shows a lower degree of banding compared to figure 2.4b. As hydrogen moves through the microstructure, pearlitic colonies hinder its diffusion, and thus less banded steels inhibit hydrogen diffusion.

8

Thus, it was observed that an increase in the banding (Figure 2.4b) increased the amount of area affected by cracks [2]. Hydrogen is irreversibly trapped at the interface between the ferrite and cementite phases that make up lamellar pearlite. After long exposure times, the interface of the cementite becomes saturated by hydrogen, causing cracking.

a) Figure 2.4

b)

Schematic diagram illustrating the hydrogen penetration through two varying degrees of banding in a ferrite/pearlite microstructure: a) lower degree of banding and b) higher degree of banding. In a), the hydrogen penetration is hindered by pearlite along the ferrite regions. Adapted from [19].

Taira et al. [2], showed that by quenching (at 900 °C) and tempering (at 650 °C) (Q&T) controlled rolled steels, there was a reduction in the amount of cracking compared to the same material that was only controlled rolled. The increased resistance was attributed to the elimination of the pearlitic structure and increased uniformity of the microstructure compared to controlled rolled steels of the same grade. The quenching and tempering process produced tempered bainite, resulting in an overall increase resistance to HIC. Other studies [22 and 23] on Q&T steels have also reported an increase in the resistance to HIC. It becomes apparent that for pipeline steels that are subjected to sour/severe environments that the control of nonmetallic inclusions is paramount to ensure resistance to hydrogen-assisted cracking. Inclusions that form in the final plate product are dependent on liquid metal processing practices. Practices such as stirring (how long, when, how vigorous), timing of alloy additions, shrouding (gas layers above the steel), controlling slag, and many others have a very strong effect on inclusion counts in the final steel [24].

The region of discontinuity around the inclusion and matrix creates a trap site for hydrogen accumulation, favoring crack nucleation and propagation in those regions. The length of the inclusion is of greater importance than its thickness [15], with respect to the rolling direction or parallel to the banding direction. The interface area along the length of the inclusion is greater, making the material more susceptible to brittle fracture normal to these interfaces.

Figure 2.5 is a schematic representation of examples of inclusions that might be present in fully killed steels [25] after casting and rolling operations. Alumina inclusions during casting form a dendritic structure but upon rolling operations are broken up and distributed with respect to the rolling direction. At higher calcium oxide to alumina ratios, the inclusion tends to become more globular in shape and will have minimal shape change after rolling operations. In contrast, MnS inclusions elongate greatly during rolling, which is very detrimental to resistance to hydrogen assisted cracking [3, 5, 21, and 26]. During casting, MnS inclusions are globular, but the

9

inclusion becomes elongated after rolling operations. This elongation allows for an increase in surface area for hydrogen adsorption along a single plane, increasing the degree of cracking observed in the microstructure [3, 26].

Because nonmetallic inclusions cannot be eliminated, the final example found in Figure 2.5 is an example of the type of inclusion that would be more resistant to HIC. The 12 CaO â&#x20AC;˘ 7 Al 2O3 inclusion forms first during casting and then the sulfide ring (CaS/MnS) forms around it. The sulfide ring allows for better control of inclusion shape as well as size and distribution within the metal matrix. Controlling these variables with additions of calcium aids in enhancing resistance to HIC.

Figure 2.5

Examples of nonmetallic inclusions that form in fully killed steels. Shown are the changes of morphology due to rolling operations after casting. Adapted from [25].

Alloying elements and impurities may affect interactions of steel with hydrogen by changing the corrosion potential of the steel, poisoning surfaces such as grain boundaries, forming precipitates that act as traps, and forming protective layers [3, 6, 26, and 27]. It is shown in literature that in environments where protective surface films can form, i.e. not severe sour service, copper additions greater than 0.20 wt pct increase plate steels resistance to HIC [3, 21, 27 â&#x20AC;&#x201C; 29]. However, when copper is present with molybdenum, nickel, or tungsten in the steel, it has been documented that the resistance to HIC is decreased due to their interactions increasing corrosion rates along with hydrogen adsorption [22 and 30]. In chromium, nickel, and cobalt-bearing alloys, the addition of these elements lowers hydrogen adsorption during exposure, thus increasing the resistance to HIC. Certain irreversible traps (precipitates) have been shown to increase the resistance to hydrogen-assisted crack phenomena [19 and 31]. The overall effectiveness of precipitates to increase resistance to HIC is based on the size and the distribution of that precipitate in the matrix [19 and 31]. Precipitates have high enough binding energy (i.e. > 60 kJ/mole) to act as irreversible traps and when they are distributed homogenously, small compared to inclusions, and present in sufficient quantities, they increase the resistance to HIC [31]. When the precipitates interact with hydrogen, they minimize the amount of hydrogen that can diffuse elsewhere into the material and initiate cracks around other irreversible traps such as inclusions. Assuming the precipitates are distributed

10

homogenously and in sufficient quantities, they can minimize the amount of cracking that occurs by allowing hydrogen to be more evenly distributed through the microstructure instead of at areas of higher crack susceptibility.

2.2.3

Mechanical Processing Effects During the forming of plate steels from continuous cast slabs or ingots, parameters that may alter the

resistance to HIC are start temperature for accelerated cooling, cooling rate, and finishing temperature [3, 5, and 20].

Strain introduced through cold rolling operations has been reported to have both detrimental and beneficial effects on plate steel performance in hydrogen environments. Further processing of plate into seam-welded pipe introduces plastic deformation and residual stresses into the final product. This section highlights the important variables to consider when rolling, cooling, and forming plate steels into pipeline steel products.

During hot rolling of plate steels at low finishing temperatures, residual stresses are generated within the plate steel, and subsequent hot working does not relieve these stresses. Finishing temperatures of 900 °C produced steels that had higher resistance to HIC when compared to the same steels rolled at a low finishing temperature of 790 °C, because the higher finishing temperature allowed for higher amounts of residual stresses to be relieved [19].

Moore and Warga [12] showed that with decreasing finishing temperature, the elongation of Type I MnS inclusions increased. This increase in elongated MnS within the plate steel increased the overall susceptibly to HIC. In a study conducted by Shinohara and Hara et al. [5], when the start temperature of accelerated cooling was held above approximately 10 °C of the Ar3 temperature, HIC susceptibly was lowered. Similarly, they found that when the cooling rate of the steel is below 10 °C/s, the banding present in their X70 pipeline steel at mid-thickness was still present. The banding present at mid-thickness created areas of preferential cracking, increasing the susceptibility to HIC. A good approximation for the levels of residual stresses is the X-ray diffraction line broadening [3 and 20].

Line broadening also relates to increases in increased dislocation density. Figure 2.6 shows the line broadening parameter and crack length ratio (extent of cracking) in a low carbon (0.07 wt pct) micro-alloyed (Nb and V) plate steel. The HIC test used to evaluate this material was the NACE standard test TM0284. Crack length ratio, the definition of which is provided in section 3.5, was calculated as per the standard, upon optical microscopy inspection of sectioned faces. The HIC resistance decreases while the line broadening parameter increases as the amount of strain in the material increases. Cold reduction from 0 to 10 pct reduces the crack length ratio from a value of 60 pct to 20 pct. The largest change in susceptibility is between the values of 0 to 5 pct cold rolling. This is important to note because beneficial effects of cold rolling are observed at relatively low amounts of cold reduction, i.e. strain. The uniformity of strain through the thickness at these low reductions may also play in a role in HIC resistance.

The effect of larger amounts of cold reduction on HIC susceptibly is shown in Figure 2.7. Three steels A, B, and C had similar levels of cold reduction introduced and were exposed to the NACE standard test TM0284.

Compositions for the steels are as follows: Steel A - 0.04-C, 1.04-Mn, 0.004-S, Steel B – 0.05-C, 1.24-Mn, 0.005-S, Steel C – 0.09-C, 1.01-Mn, 0.008-S (all values are in wt pct); all 3 steels were micro-alloyed Nb and V plate steels.

The plate thickness was 9.52 mm for A, 19 mm for B, and 14.3 mm for C. The experimental methodologies for cold rolling operations, such as strain per pass were not described. Steels A and C were produced for sour service

11

applications and steel B was a low sulfur grade steel. Steel C had the lowest degree of HIC at the highest cold reduction. This was attributed to it being continuously cast, compared to steel A and B that were ingot cast. The increase in non-homogenous microstructure in steel A increased its susceptibility at amounts higher cold reduction compared to the other alloys. The overall trend observed for the three steels is very similar and can be described as an S-curve (Figure 2.7). At values of cold reduction greater than 50 pct, the crack length ratio reaches saturation, and thus increasing the value of cold reduction to greater values has minimal effect to increase the degree of cracking.

It is interesting to compare the results from Figure 2.6 and 2.7, where relatively small amounts of cold reduction have a beneficial effect on HIC resistance in Figure 2.6, while much larger reductions have a detrimental effect. Other studies [32 and 33] have reported that values of cold strain from 2 to 16 pct increase susceptibility to HIC. The increase in resistance at low strains was attributed to decreases in the permeability of hydrogen into the metal matrix and a more uniform distribution of hydrogen in the matrix [20]. When hydrogen is unevenly distributed through the metal matrix, areas susceptible to hydrogen such as nonmetallic inclusions, reach levels of hydrogen in sufficient amount to initiate HIC. Whereas if hydrogen is more evenly distributed through the metal matrix, due to an increased level of traps created by cold reduction, HIC susceptibility decreases [20].

Figure 2.6

HIC susceptibility (crack length ratio ratio) as a function of cold rolling reduction. Lattice strain was determined through X-ray diffraction after cold rolling operations denoted as line broadening in the above figure. HIC testing was performed with respect to the NACE standard test TM0284.

Steel used for HIC testing was a 0.07-C, 1.22-Mn, 0.006-S (all values in wt pct) micro-alloyed with Nb and V heavily controlled rolled steel. Adapted from [20].

12

Figure 2.7

2.3

Effects of cold reduction at levels greater than 10 pct. Steels A and B have similar carbon equivalents, 0.31, and were produced from an ingot whereas, steel C was produced by continuous casting and has a higher carbon equivalent of 0.39. Steels A and C would be categorized are HICResistant steels and steel B is a low sulfur grade steel. See text for more detailed chemical compositions of Steels A, B, and C. Adapted from [20].

Experimental Methods to Test Plate Steels for HIC Susceptibility In this section, two experimental methods to evaluate pipeline steel susceptibility to HIC are outlined and

discussed. The industrial standard developed by NACE involves the use of hydrogen sulfide gas to achieve hydrogen entry into the steel matrix. Electrolytic charging uses electrochemical reactions and polarization in an aqueous solution to introduce hydrogen to achieve similar effects as produced by the NACE standard test.

2.3.1

NACE Standard Test TM0284 For the introduction of hydrogen into the pipeline steels, the NACE standard [34] method uses naturally

occurring corrosion reactions within an aqueous system saturated with hydrogen sulfide gas. These principals are schematically shown in Figure 2.8. At the metal/solution interface, anodic sites, i.e. elemental iron, are oxidized through the following equation: ( (2.2) while the hydrogen sulfide is oxidized according to

13

↔

( (2.3)

due to the presence of water in the solution. Hydrogen is then reduced according to the following equation: ( (2.4) These reactions create an electrochemical cell of reactions within the system. The presence of sulfide ions in solution due to the oxidation of hydrogen sulfide gas is known to suppress the rate at which the reaction in Equation 2.4 occurs. Hydrogen ions are attracted to the metal/solution interface due to the presence of electrons at the surface from the oxidation of iron [35 and 36]. The interaction of hydrogen at the surface, allows for hydrogen to recombine into hydrogen gas and bubble off to the surface into solution (Figure 2.8a) or diffuse into the steel matrix and potentially cause damage in the form of HIC (Figure 2.8b).

Figure 2.8

Schematic illustration of the two reactions for hydrogen that occur at the metal/solution interface.

The reaction (a) occurs at a much higher rate than (b). Reaction kinetics of (a) are suppressed in sour service applications due to the presence of sulfur ions, Equation 2.3. Reaction (b) causes subsequent HIC within the material through hydrogen interactions with features in the metal matrix [36].

The NACE Standard TM0284 [34] is a standardized test method, which was established to enable consistent evaluation of pipeline steels and their performance under sour environments. This performance is evaluated by the amount of HIC-induced damage the test generates in the specimen. The test is not designed to evaluate other adverse effects from a sour environment such as pitting, weight loss from corrosion, or sulfide stress cracking. The conditions of the test are not designed to simulate any specific pipeline or process operation [34]. For the test, unstressed specimens are exposed to one of two test solutions, solution A – a sodium chloride, acetic acid (NaCl, CH3COOH) solution saturated with hydrogen sulfide gas at ambient temperature and pressure, or solution B – a synthetic seawater solution saturated with hydrogen sulfide gas at ambient temperature and pressure. The specimens are subject to a 96 hour test in either solution after which they are removed and evaluated. Tests are conducted in an airtight vessel with adequate space for the test specimen along with provisions for gas purging and introduction of hydrogen sulfide gas into the vessel. The ratio of volume of test solution to the total exposed surface area of the test specimen is a minimum of 3 mL/cm2.

The evaluation of HIC susceptibly is based on the calculation of three parameters, crack length ratio (CLR), crack thickness ratio (CTR), and crack size ratio (CSR). CLR is the sum of all the length of the cracks found on the evaluated faces normalized by the total width of the face that is examined. Higher values of CLR without inspection of microstructure could indicate higher degrees of cracking oriented with microstructural banding or elongation of inclusions. CTR is a measure of how much of the plate thickness is affected by HIC cracking. CTR values are a good indication of the linkage of cracks within the microstructure based on inclusion size and distribution as shown

14

in Figure 2.2b. CSR is a measure of the total area of the cross-sectioned face that is affected by cracking. Because this sums the area of the cracks in reference to total cross-section, high values of CSR would indicate large cracks were produced, both in length and thickness of the crack. The definitions of CLR, CTR, and CSR are presented in greater detail in section 3.5.

2.3.2

Electrolytic Charging Electrochemical polarization (Ć&#x17E;) is the measure of the potential change in a system relative to the

equilibrium half-cell reactions; the potential change is caused by changing the rate at which half-cell reactions occur at the interface of the metal/solution. Cathodic polarization causes an excess of electrons at the metal/solution interface and the surface potential becomes negative, which slows the anodic dissolution reaction rate. Therefore, by definition, the value of Ć&#x17E; for cathodic polarization is negative. For anodic polarization, the surface potential is positive and the value of Ć&#x17E; is positive. Another term used for cathodic polarization is overpotential [8]. These surface potentials and their effects on reaction rates are related to the kinetics of the system, not the thermodynamics. It should be noted that the reaction rates within the system do not correspond to the half-cell electrode potentials but rather to the current density associated with the reaction. The reactions are limited by kinetic surface reaction rates, mass transfer from both the material and aqueous solution, and potential concentration gradients found within the system [8].

Electrolytic charging is conducted by selectively polarizing a two electrode system by using a DC power supply that has negative and positive output terminals. The steel specimen serves as the cathode and another material such as graphite serves as the anode. The galvanic couple between the steel specimen and anode material creates the cathodic polarization mentioned in the previous paragraph, which allows hydrogen to be attracted to the steel specimen. Electrolytic charging is aided by the use of elemental additives such as arsenic or sulfuric species present in the electrolyte. The additives are introduced into the system as soluble compounds (e.g. As2O3, NaAsO2, CS2), which inhibit the formation of diatomic hydrogen from monatomic hydrogen [37]. By doing so, the uptake of hydrogen in the steel is enhanced, increasing the efficiency of the charging setup. In literature, the additives are referred to in a variety of terms such as hydrogenation promoter [37 and 38], hydrogen recombination inhibitor agents [38], or poison [38 - 40]. The effectiveness of the inhibitors can depend on their relative amounts in the electrolyte [40]. An electrolytic charging methodology has been commonly used to pre-charge steel specimens with hydrogen before mechanical testing or to charge hydrogen into specimens while applying a load. However, it has not been extensively employed to assess HIC resistance [6].

Perez Escobar et al. conducted a study that assessed the damage in high strength steels in relation to electrolytic charging conditions [41]. While this study has not designed to assess HIC resistance, it does provide some insight on charging conditions that might produce HIC damage. Four multi-phase high strength steels and pure iron were selected. The chemical compositions of the alloys used are displayed in Table 2.1. FB450 is a ferriticbainitic mixed microstructure steel, the TRIP700 is a multiphase steel containing ferrite, bainite, and retained austenite, the DP600 is a dual phase ferrite-martensite steel, and the S550MC is a high strength low alloy that has a mixed ferritie-pearlite microstructure with Ti-Nb precipitates [41]. The main objective of this study was to identify electrolytic charging conditions that produced blisters on the surface of the materials investigated. To achieve this,

15

electrochemical variables such as current density applied, test duration, aqueous solution (acidic or basic), and presence of additives (As2O3 or thiourea (CH4N2S)) were varied [41]. Blister formation increased at both higher current densities and longer test durations. For acidic aqueous solutions at similar charging times and applied current densities, the blister formation was greater. Pure iron was highly susceptible to blister formation in comparison to high strength ferrite bainite steel. In the absence of blisters on the surface, Perez Escobar et al. sectioned the high strength steels and examined for internal cracks. For the FB450, TRIP700, and DP600 cracks were generated at the center of the sample around elongated MnS inclusions [41].

Table 2.1 Chemical Composition of the Materials used in [41] (wt pct) Material/Element FB450 TRIP700 DP600 S550MC Pure Iron

C 0.07 0.17 0.07 0.07 0.0015

Mn 1.00 1.60 1.50 0.95 0.0003

Si 0.10 0.40 0.25 0.0 0.0

Other 0.5 – 1.0 Cr 1 – 2 Al 0.04 – 0.1 P 0.4 – 0.8 Cr + Mo 0.08 – 0.12 Ti + Nb < 0.02 Al, P

Most electrochemical charging times have been short, i.e. less than 6 hours, which does not allow for hydrogen to reach high enough levels of diffusible hydrogen to cause permanent internal damage in the form of HIC. In a recent ASPPRC study, X70 plate steel was charged with hydrogen by electrochemical charging in a 1 normal sulfuric acid with 20 mg/L of As2O3 additive at test durations up to 24 hours [7]. The same X70 plate was also exposed to the NACE standard test TM0284 methodology. After the exposure to each method, the specimens were sectioned and evaluated for internal cracks generated by hydrogen. At test durations of 24 hours in electrolytic charging, the critical cracking parameters were on the same order of magnitude as the results produced by the NACE standard method conducted by U.S. Steel [7]. Table 2.2 shows the results obtained from the NACE Standard test and the electrolytic testing methodology. In general, the CTR, CTR, and CSR results from electrolytic testing range were greater than or equal to results obtained from the NACE standard test.

Table 2.2 NACE Standard Test Results and Results Using Electrochemical Charging on X70 Plate [7] NACE standard test conducted by U.S. Steel CSR (%) CLR (%) CTR (%) 0.1 6.0 0.3 1.5 16.7 3.1

16

B. Rosner Electrochemical Testing Test No. CSR (%) CLR (%) CTR (%) 1 0.33 24.7 2.9 2 0.05 9.40 1.60 3 0.33 15.50 3.40 4 0.06 6.65 1.25 5 0.10 2.95 3.60 6 1.46 41.5 10.8 7 1.88 43.4 15.1

CHAPTER 3: Experimental Design and Methods

This chapter presents the experimental methodology to create an electrolytic charging apparatus and procedure to evaluate plate steel susceptibility to hydrogen damage.

3.1

Experimental Design Rosner [7] explored electrolytic charging to develop a reproducible methodology to investigate hydrogen

induced cracking and hydrogen damage in steels. In the study, he showed that electrolytic charging could serve as a viable substitute to evaluate hydrogen susceptibility if there were modifications to increase the stability of methodology at prolonged test durations [7]. These issues and solutions are presented and discussed in section 3.5.1.

A newly designed test cell for electrolytic charging was fabricated. Characterization and validation of the electrolytic charging apparatus was undertaken. Alterations to certain parameters of the test apparatus and sample were conducted on a highly susceptible material, 100XF, to optimize the setup for HIC susceptibility experiments.

Parameters such as current density (voltage), test duration, and surface finish of cathode were identified as variables that could alter the reproducibility of results obtained from this methodology. The validation of the proposed experimental procedure was evaluated by comparing results from the industrial NACE Standard for HIC evaluation.

The electrolytic charging apparatus was used to investigate HIC susceptibility of four different plate steel grades, which were also subjected to cold-working before charging (uniaxial prestrain). The alloys were all low carbon plate steels with a range of tensile strengths. Two of the alloys, X52 and X60, were designated as HIC resistant grades.

After exposure to the hydrogen environments, either wet hydrogen sulfide gas or electrolytic charging, through-thickness cross-sections were used to measure the extent of cracking. Further characterization of hydrogen damage was undertaken to explore effects of microstructure, nonmetallic inclusions associated with cracks, and diffusible and trapped hydrogen amounts. These methods are presented and discussed in subsequent sections of this chapter.

3.2

Experimental Materials and Methods

3.2.1

Materials The chemical compositions of the steel plates used in this study are shown in Table 3.1. The effects of

prestraining were investigated using 100XF plate steel from the M.S. thesis of B. Farber [42], X70 plate steel from the M.S thesis of H.M. Al-Jabr [43], X60 plate steel received from EVRAZ North America, and X52 plate steel received from SSAB. The X70 was provided by Essar Algoma Steel in 45 x 30 x 1.27 cm (18 x 12 x 0.50 in) plates.

The 100XF plate steel was received as 91 x 91 x 1.27 cm (36 x 36 x 0.5 in) plates. The X60 plate steel was received as 84 x 34 x 0.95 cm (33 x 13.5 x 0.375 in) plates. The X52 plate steel was received as 45 x 30 x 1.9 cm (18 x 12 x 0.75 in) plates; the plates came from the center of the plate stock. The X52 and X60 plate steels have been engineered for use in hydrogen environments and rated for sour service by the NACE Standard TM0284 test. All of the steels are fully killed, calcium treated continuous cast steels. The X52 has the lowest amount of sulfur

17

(0.0007 wt pct) and highest amount of copper (0.35 wt pct). All grades were microalloyed with titanium, niobium, and/or vanadium to increase strength and control microstructure and transformation products. The thermal mechanical processing and rolling operations were not provided by the steel suppliers.

Table 3.1 Chemical Composition of As-Received Plate Steels in wt pct wt pct X52 X60 X70 100XF

3.2.2

C Mn Si Ni Cr Mo Ti Nb V Al N S P Cu Ca 0.067 1.03 0.24 0.10 0.06 0.03 0.007 0.044 0.007 0.031 0.0082 0.0007 0.010 0.35 0.0027 0.055 1.42 - 0.08 0.032 Ti+Nb+V – 0.09 0.0028 0.050 1.59 0.30 0.01 0.26 0.09 0.013 0.066 0.005 0.026 0.0081 0.003 0.010 0.01 0.046 1.8 0.21 0.3 0.018 0.070 0.084 0.031 0.009 0.002 0.008 * “-“ in table represent values that were not reported from the steel supplier upon receiving plate steels

Mechanical Properties Mechanical properties of the plate steels are summarized in Table 3.2. The mechanical properties were

obtained from specimens oriented in the transverse direction with respect to rolling direction. The longitudinal and diagonal mechanical properties are not included because prestrain was performed in the transverse direction.

Mechanical property data for the 100XF and X70 materials were obtained from Farber [42] and Al Jabr [43] thesis work. X52 and X60 mechanical property data were obtained from the two plate producers, SSAB and EVRAZ North America, respectively.

Table 3.2 Mechanical Properties in the Transverse Direction of As-Received Plate Steels Material X52 X60 X70 100XF

3.2.3

0.2% Offset Yield Strength (MPa (ksi)) 405 (58.7) 515 (74.7) 465 (67.4) 724 (105)

Tensile Strength (MPa (ksi)) 480 (69.6) 585 (84.8) 595 (86.3) 804 (116.6)

YS/UTS 0.86 0.88 0.78 0.9

pct Elongation @ 2 in 41.2 34 32 18.8

Microstructural Analysis Microstructural analysis on the four plate grades was performed with field-emission scanning electron

microscopy (FESEM). Faces in the transverse direction (TD), rolling direction (RD), and the normal direction (ND) were polished and etched with 2 pct Nital. The three orientations were evaluated to assess grain shape with respect to the rolling direction, the presence of second phase microconstituents, and degree of banding. Analyses of polished and etched HIC samples were conducted. Using the energy-dispersive x-ray spectroscopy (EDS) capabilities of the FESEM, the presence and composition of nonmetallic inclusions was identified.

Automatic feature analysis (AFA) was conducted with the aid of Nucor Corporation at their Crawfordsville, IN facility. AFA was conducted on two inclusion families (conventional melt shop analysis for inclusion control): 1) Ca-Al-S and 2) Mn-Ca-S. Comparison between these two families is a typical melt shop evaluation of the nonmetallic inclusions formed during steelmaking [24]. Steels modified for use in hydrogen environments show higher amounts of inclusions in the Ca-Al-S family to increase resistance to HIC. HIC susceptibility is greater if greater amounts of MnS type inclusions are produced during steelmaking. The presence of

18

MnS is evaluated through the analysis of the Mn-Ca-S inclusion family [24].The purpose of the AFA analysis was to identify the relative composition of the inclusions that formed. The term relative composition is used because the ternary diagrams that were generated represent amounts of each element relative to the other two elements evaluated. Thus, when the amount of an element is observed at levels of 90 pct or greater in the ternary diagrams, it assumed that the inclusion is comprised of other elements i.e. Fe, Mg, O, etc.

3.2.4

Introduction of Prestrain The experimental plate alloys were pre-strained in tension rather than by rolling or bending in order to

avoid strain gradients and residual stresses. A schematic representation of the strip geometry, orientation, and the prestrain direction of the strips is shown in Figure 3.1. Prestrain was introduced using a 978.6 kN (220 kip) tensile machine located at NIST in Boulder, Colorado with the help of Mr. D. McColskey. Strips that were 20 mm (0.787 in) in width were cut so the specimen longitudinal axis was transverse to the rolling direction. Based on the as-received plates geometry, the strip length differed from between each material and was between 450 mm (17.7 in) and 700 mm (27.5 in). Prestrain was introduced in the transverse direction with respect to the rolling direction. The presence of LĂźder bands in the X52, X60, and X70 materials required specimens from these plates to be subjected to higher amounts of prestrain (to ensure uniform strain was introduced throughout the material) than specimens from the 100XF plate, which did not exhibit such behavior.

Strips were first loaded to failure and, using crosshead displacement and the grip-to-grip distance of the strip as the gauge length, engineering stress-strain curves were generated in order to select displacement/strain values outside the LĂźders band region. The prestrain levels for each material and are shown in Table 3.3.

Extensometers were not used due to the large grip to grip length of the strips, approximately 225 mm (9 in) to 305 mm (12 in). To introduce these prestrain levels, strips were placed into the load frame and the grip to grip distance was measured as shown in Figure 3.2. This distance was used as the gauge length, and the crosshead displacement was then set to a desired distance to obtain the target prestrain level. When the set value of crosshead displacement was reached for each material and prestrain condition, the final gauge length was measured to ensure the desired prestrain level had been reached. Prestrain levels for the X60 material were chosen based on recommendations by the steel producer. The X70 material were prestrained to higher amounts than the X60 to evaluate greater amounts of cold work at an increased strength. Prestrain was greatest for X52 material to test if higher amounts of cold work could initiate HIC. The 100XF specimens could not be pre-strained beyond 2 pct, because the specimens did not contain a reduced cross-section gauge length and stress concentrations near the gripped sections resulted in non-uniform deformation at relatively small strains. Therefore, only one prestrain condition was selected for 100XF.

19

Figure 3.1

Schematic of strips that were machined in the transverse direction with respect to the rolling direction and then subsequently prestrained to target levels. The prestrain and rolling directions are labeled. Strip length was dependent on the plate material that was machined. The length ranged from 450 mm for the X52 to 700 mm for the 100XF.

Figure 3.2

Example of a strip being prestrained using uniaxial tension. The entire length between grips was considered the gauge length. This gauge length was used in correspondence with target prestrain levels to set a predetermined crosshead displacement to achieve the desired prestain amount.

20

Table 3.3 Selected Prestrain Amounts for the Plate Steels Material

3.2.5

Prestrain Amount in pct

X52

0

12

18

X60

0

3

5

X70

0

5

7

100XF

0

2

Hardness Traverse and Determination of Microstructural Dependence on Plate Thickness Vickers microhardness measurements were conducted on all as-received plate materials. Cross-sections of

the full thickness of the transverse plane were sectioned and polished to 1 Âľm surface finish. Microhardness measurements were taken (every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time) from each rolled face edge to the middle thickness of the material. The microhardness traverse data for each as-received material were normalized with respect to plate thickness. For each hardness traverse, 50 pct plate thickness represents the midthickness of the plate with respect to each rolled face. Edge 1 and edge 2 represent traverses taken from each of the rolled surfaces towards the center of the material. Microhardness measurements were also taken on the prestrained plate materials. Sample size and orientation were kept constant for the hardness traverse procedure. Hardness measurements were taken from the center of the material towards the rolled surface. A minimum of five hardness measurements were performed for each material and prestrain condition; additionally, care was taken to ensure that no indent was within three diagonals of the previous indent.

In order to determine the microstructure as a function of plate thickness, micrographs were taken at 20X from the mid-thickness to the rolled face and subsequently pieced together. Micrographs were produced on both non-etched and 2 pct nital etched samples. The non-etched samples illuminate the distribution of nonmetallic inclusions from the rolled surface to the mid-thickness. It should be noted that these micrographs were not used to determine the actual size or composition of the non-metallic inclusions present. The 2 pct nital etched samples were used to identify areas of microstructural differences on a broad scale such as banding and differences in microstructural from the rolled surface to the mid-thickness of the plate.

3.3

HIC Sample Generation and Preparation for HIC Testing The geometry and orientation of the HIC test specimens, machined after pre-straining, are shown in Figure

3.3. After prestrain was introduced, strips were sectioned into shorter pieces approximately 110 mm long. Once these smaller test specimens were created, an abrasive saw was used to precisely cut the specimen to the final desired length of 100 mm. To remove the oxide layers on the rolled faces and rust that formed on the cut faces from machining the as-received plates into strips, the 100 mm long specimens were placed into the Bridgeport knee mill and face milled. Milling removed 125 hundredths of inch (0.3175 mm) from the selected face for each pass until the oxide layer and rust were removed. This procedure was repeated for each of the remaining three faces with respect to normal and longitudinal planes. Milling did not alter the total length of the sample. After milling, all faces were

21

ground and finished with 320 SiC grit paper to maintain a constant surface roughness between test specimens. For specimens that were subjected to electrolytic charging, a taper hole was drilled and reamed in the transverse top face of the test specimen to allow for electrical conductivity. After these procedures, each test specimen was 100 ± 3 mm (3.937 ± 0.118 in) long by 20 ± 3 mm wide (0.7874 ± 0.118 in), and the thickness is the full thickness of the plate [34]. Three samples of each material and prestrain were fabricated for each hydrogen exposure environment. The test specimen geometry of each specimen created from each material and prestrain condition is found in Tables 3.4 through 3.7.

Figure 3.3

3.4

Test specimen geometry and orientation used in the NACE Standard and electrolytic charging studies; all dimensions are in millimeters. For specimens that were subjected to electrolytic charging, a taper hole was drilled and reamed in the transverse top face of the test specimen to allow for electrical conductivity.

NACE Standard TM0284 (H2S Method) A total of 33 specimens, representing the full experimental matrix with 3 replicates for each material and

prestrain condition, were tested according to NACE standard TM0284 [34]. The test vessel for the NACE Standard test along with the carousel that the test specimens were placed into is shown in Figure 3.4. Figure 3.4a shows the airtight vessel the test specimens are placed into and the input for hydrogen sulfide gas and the off gas stream.

Figure 3.4b shows the carousel that was used to hold the specimens while being submerged in the test solution. The carousel allows for adequate space between the test specimens for equal exposure to the solution and hydrogen sulfide gas. The ratio of volume of test solution to the total exposed surface area of the test specimen was a minimum of 3 mL/cm2. Solution A, a sodium chloride, acetic acid (NaCl, CH 3COOH) solution, saturated with hydrogen sulfide gas at ambient temperature and pressure was used. The specimens were subject to a 96 hour test in this solution after which they were removed and evaluated. Results from the H2S method serve a dual purpose in the current study: 1) to validate the experimental procedure for electrolytic charging to produce HIC and 2) to

22

investigate the effects of prestrain on HIC susceptibility. This test was conducted with the help of EVRAZ North America at their Research and Development facilities located in Regina, Saskatchewan.

Table 3.4 Test Specimen Geometry for the X52 Specimens used in each Hydrogen Exposure Method NACE Standard test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

X52 0 %

1

19.17

18.38

101.48

2

19.17

18.31

101.78

3

19.23

18.4

101.19

1

18.13

17.26

100.72

2

18.45

17.22

101.02

3

18.34

17.23

101.04

1

17.67

16.84

100.98

2

17.52

16.82

100.85

3

17.48

16.93

100.17

X52 12 %

X52 18 %

EC test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

X52 0 %

1

19.23

18.31

99.89

2

19.37

18.21

101.21

3

19.41

25.58

102.99

1

17.93

17.78

102.77

2

17.91

17.42

102.92

3

17.81

17.32

98.54

1

17.93

16.84

102.33

2

17.81

16.99

100.82

3

17.78

16.69

102.76

X52 12 %

X52 18 %

23

Table 3.5 Test Specimen Geometry for the X60 Specimens used in each Hydrogen Exposure Method NACE Standard test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

X60 0 %

1

19.4

8.24

98.3

2

19.82

8.49

98.77

3

19.28

8.36

98.51

1

19.56

8.66

99.74

2

19.88

8.68

99.55

3

19.68

8.4

98.52

1

19.48

8.92

99.38

2

19.7

8.86

98.75

3

19.73

8.9

99.42

X60 3 %

X60 5 %

EC test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

X60 0 %

1

20.04

8.89

98.97

2

19.66

8.92

102.21

3

19.69

9.02

100.71

1

20.12

8.69

98.14

2

19.81

8.69

101.26

3

19.71

8.69

98.62

1

19.91

8.53

100.35

2

20.12

8.33

100.63

3

20.08

8.28

100.87

X60 3 %

X60 5 %

24

Table 3.6 Test Specimen Geometry for the X70 Specimens used in each Hydrogen Exposure Method NACE Standard test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

X70 0 %

1

19.48

12.1

100.01

2

18.96

12.13

99.91

3

19.01

12.06

99.35

1

19.05

11.83

100.01

2

18.97

11.84

99.74

3

19.07

11.91

99.78

1

18.8

11.89

99.75

2

18.97

11.87

100.69

3

18.57

11.8

99.7

X70 5 %

X70 7 %

EC test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

X70 0 %

1

19.58

12.12

100.73

2

19.42

12.17

100.49

3

19.61

12.07

100.44

1

19.1

11.84

100.22

2

19.08

11.58

100.61

3

19.23

11.89

99.84

1

19.08

11.74

99.62

2

19.03

11.68

102.73

3

19.03

11.71

98.63

X70 5 %

X70 7 %

25

Table 3.7 Test Specimen Geometry for the 100XF Specimens used in each Hydrogen Exposure Method NACE Standard test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

100XF 0 %

1

19.26

11.77

100.47

2

19.39

12.11

100.16

3

18.78

11.85

100.15

1

18.72

12.08

98.11

2

18.77

12.1

98.65

3

18.74

12.09

98.27

100XF 2 %

EC test specimen geometry Material & Prestrain

Specimen No.

Width (mm)

Thickness (mm)

Length (mm)

100XF 0 %

1

18.69

12.27

101.63

2

19.38

12.37

100.83

3

19.28

12.02

98.11

1

19.43

12.34

99.93

2

19.33

12.16

100.88

3

19.48

12.34

100.58

100XF 2 %

a) Figure 3.4

3.5

b)

a) NACE Standard TM0284 test vessel used to evaluate materials suscpetibilty to HIC, b) carousel used to hold test specimens allowing for adequate space for gas purge during testing.

Electrolytic Charging Methodology (EC) The following section highlights the research and development of the EC methodology. Design issues

encountered from previous attempts at EC charging for HIC resistance are presented. Then, solutions to address

26

these design issues are discussed as well as the fabrication of the new EC apparatus used in the present study. A characterization study of the new EC cell was undertaken to optimize the experiment for HIC assessment. Based on the characterization study, an experimental procedure was developed and implemented to investigate HIC using EC.

3.5.1

Initial Design and Fabrication of New Charging Apparatus Cathodic charging to introduce hydrogen to the test specimens has been selected as a substitute for NACE

standard test TM0284 [34]. Figure 3.5 shows the previous generation cell used for electrolytic charging. The cell is a simple beaker (referred to as a monocell) that holds the electrolyte and both electrodes (steel cathode and graphite anode). In previous attempts to validate this electrolytic charging setup for long tests (e.g. 24 hours or greater), it was observed that the current within the system tends to vary when the system is voltage controlled, or voltage varies if the applied current is controlled. The slight variation in the electrical response of the system is important to monitor in order to assess the reliability of the charging. Due to the long term nature of the charging, Rosner [7] and Angus [44] discovered issues related to the stability and repeatability of electrolytic charging experiments. The issues Rosner [7] and Angus [44] summarized and the solutions to them are summarized in Table 3.8.

Table 3.8 Design Issues and Solutions taken for the Electrolytic Charging Methodology Design Issue During electrolytic charging, the anode material (graphite) underwent mass loss and changes in geometry.

The mass loss of the graphite would then pollute the electrolyte, changing the nature of the charging conditions within the cell.

Alligator clips have been used as the main method to achieve an electrical connection to the cathode, which is the steel specimen. At long test durations, i.e. greater than 18 hours, the alligator clips would fail and the electrical connection to the steel specimen would be lost, limiting the possible test duration.

The initial trials for electrochemical hydrogen charging of test specimens in aqueous solutions used a single cell design. Both the anode and cathode were in the same compartment. This lead to cross-contamination of ion species at each anode/cathode interface. This crosscontamination causes unwanted reactions to occur at those interfaces, lowering the overall effectiveness of electrolytic charging. It was also observed that corrosion and mass loss of the steel specimen could occur in this cell due to increased concentration of ions within the system around the steel specimen, regardless of the cathodic overpotential that the steel specimen was experiencing.

Oxide scale from the steel specimen flaked off during testing and polluted the electrolyte. This pollution altered the nature of the electrolyte and changed to electrical response within the system.

27

Design Solution A titanium mesh that was coated in ruthenium oxide, provided by Dr. Martins, was selected as the anode. This material does not undergo the weight loss and changes in specimen geometry during testing as observed for the graphite.

A new taper pin plug in friction fitted electrode connection was devised. The taper pin fit to a predrilled hole into the steel specimen to create the electrical connection. To protect this connection further, the area of the taper pin exposed to the aqueous solution was coated in epoxy.

To minimize cross-contamination within the system, a newly designed dual cell was fabricated. The dual cell has separate anodic and cathodic compartments. Ion separation was achieved through a salt bridge between the compartments. This salt bridge allows for anodic species in the cathodic compartment to flow to the anodic compartment and vice versa for cathodic species. Purge gas systems was fabricated for each compartment to aid in the flow of ionic species from one compartment to another and minimize potential ionic concentration gradients.

Each steel specimen face was milled and subsequently ground to a final surface roughness of 320 grit to remove oxide scale.

Figure 3.5

Prior generation cell used for electrolytic charging by B. Rosner [7] and G. Angus [44].

The initial cell in Figure 3.5 was a very simple monocell design. In order to achieve ion separation, the anodic and cathodic compartments were separated to create a dual cell design, schematically shown in Figure 3.7.

To allow for ion species to travel from one compartment to another, the two compartments are connected to each other via a side branch with a salt bridge. An O-ring seal is used at the connection, which is fastened together with flanges. Each compartment is fitted with a Teflon cover that incorporates ports with Teflon compression-fittings for positioning of each electrode assembly as well as for gas effluent-discharge. Fitted aluminum flanges provide a means to attach the Teflon compression-fittings to each compartment. A side-connection on the bottom of the main body of each compartment allows inert-gas purging of the anodic and cathodic compartments, respectively. Altering the initial test apparatus to the proposed dual cell design aids in stabilizing EC experiments at prolonged test durations.

Using the design solutions in Table 3.8, a newly designed test cell for electrolytic charging was fabricated.

The anode material was changed from the graphite anode to the titanium mesh coated in ruthenium oxide, and this assembly is referred to as a dimensionally stable anode (DSA). The (anode) counter-electrode assembly was fabricated at Colorado School of Mines and consisted of a titanium rod welded to a titanium collar that is welded to the DSA. The composite material comprising the DSA is a diamond-pattern expanded titanium-metal, coated with ruthenium oxide (semi-conductor).

The electrical connector to the working-electrode test-specimen consisted of a hardened taper-pin welded to a 316SS rod. Fabrication of the cathodic electrical connector is shown in Figure 3.6a. A taper hole to accommodate this connector was drilled and reamed in the transverse top face of the test specimen shown in Figure 3.6b. Once

28

inserted (and friction locked) into the test specimen, the lower region of the connector was electrically insulated with a thin coating of epoxy resin.

a)

b)

Figure 3.6

From left to right, the images display the salient features for (a) the construction of the cathode electrode connection to be used in correspondence with (b) the steel specimens. The cathode electrode connection in (a) employs a taper pin fastened to a 316 SS rod. A corresponding taper hole is reamed into the top transverse face of the steel specimen. The cathode electrode connection is then inserted into the hole and frictionally locked with the test specimen. To protect this electrical connection, the region around the connection is covered in epoxy.

Figure 3.7

Schematic diagram for dual cell used for electrolytic charging experiments.

The fabricated EC cell is shown in Figure 3.8. The separate anodic and cathodic compartments were custom made Pyrex glass vessels, fabricated by Allen Scientific Glass Inc., of Boulder, Colorado. The fitted Teflon

29

compression fittings were fabricated by Dr. Martins. The two compartments, when assembled, are held by a customdesigned aluminum stand, fabricated by Stephen Tate. The gas purge ports include a Teflon up-leg connection to allow for inert-gas purging (argon gas), fabricated by Dr. Martins. The flow rate of the gas is controlled by a flow meter (found in Figure 3.8). A BK Precision 1735A 30V/3A DC power supply with high sensitivity (voltage ≤ 0.02% ± 3 mV, current ≤ 0.2% ± 3 mA) was used to supply the current (the Potentiostat found in Figure 3.8).

3.5.2

Characterization Study and Experimental Procedure for EC Electrolytic charging (EC) was accomplished through electrochemical polarization using an

electrochemical cell designed to optimize the reproducibility of the HIC results and the stability of the components during long charging times. Current was controlled using a Potentiostat, while voltage varied based on the electrochemical nature of reactions occurring between the two electrodes in the test setup.

A characterization study on the newly fabricated test cell (Figure 3.8) for EC was undertaken. The 100XF as-received plate material was chosen for this study, based on the assumption that it would show the highest susceptibility to HIC and facilitate identifying the parameters that would alter the reliability of EC. The parameters that were altered during this study were: 

The applied current (voltage) density



Test duration Applied current density and test duration were chosen as variables to alter within the system based on a

study by Perez Escobar et al. [41]. The results Perez Escobar et al. obtained showed increased levels of hydrogen damage were caused from increases in both applied current density and test duration. Changes in electrolyte chemistry were not explored in this study because these effects have been explored in previous theses [6 and 7].

Other variables such as flow rate of gas purge, cathodic surface exposure to gas purge bubble stream, temperature of electrolyte, and surface finish of steel specimen were held constant throughout the characterization study. Tests were conducted at room temperature (20 – 22 °C) and ambient barometric-pressure (~ 0.80 atm). The surface finish of the steel specimen was ground to 320 grit (as described in section 3.3) to avoid contamination of electrolyte during testing. The results from the characterization study are presented in section 4.3. The following experimental procedure for EC was established based on this study for the remaining investigation of HIC in the material and prestrain conditions outlined in Section 3.2: 

Current density of 15 mA/cm2



Test duration set at 24 hours



The total volume of the electrolyte maintained at 750 mL



The argon gas purge set to 25 cm3/min



The electrolyte is 1 normal H2SO4 solution with an addition of 20 mg/L of As2O3



Surface finish of steel specimen ground to 320 grit



No manual alteration to increase temperature during testing, tests were conducted at room temperature



No alteration to pressure within the test apparatus was undertaken, tests were conducted at ambient barometric pressure

30

Figure 3.8

Electrolytic hydrogen charging (EC) test apparatus with components labeled. Electrolytic charging (EC) was accomplished through electrochemical polarization using this test apparatus. Tests apparatus was designed to optimize the reproducibility of the HIC results and the stability of the components during long charging times

31

3.6

Assessment of Hydrogen Damage Upon completion of each hydrogen charging method (H 2S or EC), the test specimens were taken out of the

solution and cleaned, and the electrical connection was subsequently removed from the charged specimens. Each specimen was then sectioned as shown in Figure 3.9a; as per the NACE Standard Test, only faces 2, 4 and 6 were evaluated for specimens that were charged by the H2S method, whereas all 6 locations were evaluated for specimens charged by the EC method.

Light optical microscopy was used to determine the NACE crack ratios on the polished sections, as shown schematically by Figure 3.9b. The Crack Sensitivity Ratio (CSR), Crack Length Ratio (CLR), and Crack Thickness Ratio (CTR) were calculated with the following equations: ∑

⁄

(3.1)

∑ ⁄

(3.2)

∑ ⁄

(3.3)

where a is length of single crack, b is the thickness of a single crack including crack branching, and W and T are the width and thickness of the cross-section, respectively. Cracks separated by less than 0.5 mm were considered as a single crack. All cracks observable at magnifications as high as 100X are included in the three calculations. The average of each ratio was calculated for all of the examined faces of each test specimen [34].

a) Figure 3.9

b)

Test specimen geometry and orientation used in the NACE Standard and electrolytic charging studies; all dimensions are in millimeters, (a) Faces 2, 4 and 6 are used in the NACE Standard evaluation while Faces 1 - 6 are used for the electrolytic charging method. Face 1 is the furthest from the electrode link. (b) Schematic showing how cracks are measured on the evaluated faces [34].

32

The hydrogen charged specimens from both the EC method and NACE standard test were evaluated according to the procedures outlined in NACE standard test TM0284 [34]. The specimens from the NACE standard test conducted at EVRAZ were evaluated by a technician, and the results were provided in an Excel spreadsheet report. The method used to measure cracks on the EC cross-sections started by taking an image of the full crosssection of the examined face using the stereoscope; the image was imported into ImageJ, an image processing program. . The scale of the image was set in ImageJ. A screen shot of this process is shown in Figure 3.10.

Figure 3.10 shows the interface from the program along with the measurement of the plate thickness in order to set the scale. Once the plate thickness is measured and the corresponding value in pixels is found, the “set scale” feature in the program the amount of pixels can be set to the known plate thickness in mm. For the image shown in Figure 3.10, the value in pixels ‘502’ corresponds to 12.42 mm. ImageJ also displays the resolution of the image in reference to pixels, which was 40.419 pixels per mm in Figure 3.10. Taking the reciprocal of this value to get mm per pixel, it is assumed that calculations of crack lengths and thickness from the image are accurate out to 3 hundredths of an mm (30 µm).

Figure 3.10

ImageJ interface that allows the user to import selected images into the software for analysis. Once imported, the plate thickness is measured in pixels, shown by the white line. Knowing this distance in pixels, the “set scale” feature in the software allows the user to set the distance in pixels to a known distance. For this image and cross-section, 502 pixels equaled 12.42 mm. The “set scale” feature also displays the resolution of the image relative to the scale. This resolution was used to determine the accuracy of the measurements.

Cracks were identified and measured using the steps shown in Figure 3.12. A crack was distinguished from a scratch by the jagged features of the cracks compared to the straight line morphology of the scratches. According to the NACE standard, “cracks separated by less than 0.5 mm are considered as a single crack [34].” An example of

33

the measured separation in cracks is shown in Figure 3.12a. Distances between the cracks were measured as shown in Figure 3.12a. Because the offset of cracks identified in regions No. 1, 2, and 6 in Figure 3.12a were less than 0.5 mm (Figure 3.12b), the two cracks in each region are considered a single crack. Cracks identified as No. 3 – 5 (Figure 3.12b), have an offset distance greater than 0.5 mm, so crack was identified as a single crack. Using the “zoom” feature in ImageJ, each crack is enhanced to measurement length, a, and thickness, b, of the crack (Figure 3.12c). White bars represent the measurements of the 8 identified cracks shown in Figure 3.12c. The corresponding values of length and thickness for each crack are shown as an example of representative measurements in Table 3.9.

Using Equations 3.1 – 3.3, the crack ratios were calculated (Table 3.9).

Table 3.9 Measured values from Manual Image J Measurements of length ‘a’ and thickness ‘b’ of Cracks in Figure 3.12. The Calculated Crack Ratio Values are also shown.

Material & Prestrain condition Crack # a (mm) 1 1.93 2 11.93 3 2.23 4 1.65 5 2.05 6 4.11 7 2.59 8 1.04

100XF b (mm) 0.29 1.06 0.07 0.07 0.07 0.21 0.61 0.13

0 pct CLR (fractional) 1.1114

W = 24.77 mm CTR (fractional) 0.2013

T = 12.47 mm CSR (fractional) 0.0524

The uncertainties for the measurements used in equations 3.1 – 3.3 are shown in Table 3.10. The uncertainty for variables a and b are in reference to the image resolution imported into Image J used to measure each individual crack. Values of uncertainty for the W and T variables are from the calipers that were used to measure the cross-section of the selected face. The total uncertainty for equations 3.1 – 3.3 are calculated by the root sum of squares equation, ⁄

[ ∑

[ ∑

⁄

]

]

(3.4)

(3.5)

where wx is the uncertainty in the device used to obtain the measurement and δ(CTR, CLR, CSR)/δx is the partial derivative of the function with respect to the variable associated with the measuring device. Using Equations 3.4 and 3.5, values of total uncertainty were calculated for each plate material. The uncertainty values for each material and critical crack ratio are shown in Table 3.10. All uncertainty values for critical crack ratios are accurate out to approximately 0.0005 of a fractional value except the CTR value for the X60. Because the X60 is the thinnest plate material out of the four, the uncertainty is 0.001 of a fractional value.

34

Table 3.10 Uncertainty from the Method used to Measure Variables for use in Crack Ratio Calculation Method Variable Uncertainty (mm) Caliper W and T 0.005 Image J a and b 0.03

3.7

LECO® Hydrogen Analysis The LECO® RH-404 hydrogen analyzer induction melts a test specimen and reports the total hydrogen

content at the time of measurement in units of ppm (mass fraction). For accurate measurements of hydrogen content, the weight of the specimens that are placed into the hydrogen analyzer should be 1.0 ± 0.5 g. Total hydrogen levels in the current study were measured using sub-size test specimens that were placed into the LECO® RH-404. Asreceived plate materials were evaluated for hydrogen content to serve as a baseline of hydrogen present in the material before charging. Total hydrogen levels were also measured for each material and pre-strain condition.

3.7.1

LECO® Hydrogen Analysis Sample Generation A means to charge multiple sub-sized specimens was generated because the hydrogen analyzer is only able

to analyze small specimens. The specimen geometry is set to: plate thickness (which ranges from 0.95 to 1.9 cm) x 0.2 x 0.7 cm ((0.375 to 0.75 in) x 0.079 x 0.276 in). Figure 3.11 shows the fabrication of a “daisy chain” that allowed for multiple sub-sized specimens to be charged for hydrogen evaluation. For each material and prestrain condition, a total of 5 replicas were created. Samples of each material and prestrain condition were cut to the geometry mentioned above and, using copper wire, wound so that they were securely in contact with the wire to ensure electrical conductivity.

Figure 3.11

“Daisy Chain” assembly components. From left to right, the images display the salient features (the five sub-sized specimens and copper wire) of the component assembly leading to the assembly of the “Daisy Chain” to allow for simultaneous charging of multiple specimens for hydrogen content measurements.

35

a)

Figure 3.12

b)

c) Example cracked face showing how images were used to (a) identify cracks that are offset from one another. (b) Example cracked

face showing the measurement offset to determine if multiple cracks should be considered a single (cracks separated by less than 0.5 mm were considered as a single crack [34]). (c) The final measurement of the single cracks. Each crack in (b) was enhanced using the â&#x20AC;&#x153;zoomâ&#x20AC;? feature to produce the images in (c) to aid in the precise measurement of the single crack. The white lines in (c) represent the length, a, and thickness, b, measurements used to calculate critical crack ratios.

36

3.7.2

LECO® Hydrogen Analysis Experimental Procedure With “daisy-chains” fabricated for each material and prestrain condition (Figure 3.11) the total exposed

surface area of the samples and copper wire was determined to apply the correct current at the applied current density of 15 mA/cm2. The “daisy-chain” was placed into the EC apparatus (Figure 3.8) and charged with hydrogen for 24 hours, following the sample procedure outline for full size test specimens that were evaluated for HIC. After charging, the “daisy chains” were removed, cleaned with water, and placed into a liquid bath of nitrogen within 30 seconds in order to suppress the diffusion of hydrogen out of the samples. LECO® steel standards with known amounts of hydrogen were then used to calibrate the H analyzer before measuring the steel specimen hydrogen content. The resolution of the LECO® RH-404 is 0.05 ppm H. After calibrating the LECO® H analyzer, individual samples were taken from the bath of liquid nitrogen, removed from the “daisy chain,” rinsed, dried, and placed into the LECO® RH-404 for hydrogen analysis. For each sub-sized test specimen, the time spent at room temperature (22 °C) before the LECO® RH-404 induction melted and evaluated the hydrogen content was on average 90 seconds.

Diffusible hydrogen concentration for the LECO® method were calculated by, (3.6) where, HTotal is the hydrogen concentration determined by the LECO® immediately after EC, and HTrapped is the hydrogen concentration evaluated after degassing determined by the LECO® analyzer. Degassing was achieved by placing full size test specimens in a liquid water bath at an elevated temperature of 50 °C for 72 hours, as described in Section 3.8. Samples were sectioned from these degassed full size test specimens to create the sub-sized specimens needed for LECO® analysis. A total of 5 specimens were generated from each material and prestrain condition. The 5 values from each hydrogen concentration value, HTotal and HTrapped, were then used in combination with Equation 3.6 to produce LECODiffusible values.

3.8

Diffusible Hydrogen Content Determined by Mercury Displacement The American Welding Society (AWS) created a standard for diffusible hydrogen measurements based on

mercury displacement. This section presents the experimental procedure that was modified from this standard, AWS A4.3-86 [45], and implemented to measure diffusible hydrogen contents after exposure to the EC method.

3.8.1

Mercury Displacement Sample Generation Multiple specimens for diffusible hydrogen analysis were created from a single sample charged with

hydrogen by EC. A visual representation of the fabrication of these sectioned EC test specimens is shown in Figure 3.13. Full size test samples were sectioned so that after EC trials, 4 equal sized specimens (test samples labeled #1, #2, #3, and #4 in Figure 3.13) could be generated without further processing steps. After the taper hole was drilled into the top transverse section, the specimen was placed into the abrasive saw, where it was partially sectioned at the 25, 50, and 75 mm locations (Figure 3.13a). The black area in Figure 3.13a represents the material left after sectioning, visually shown in Figure 3.13b (designated as the areas of electrical connection by the arrows). The

37

overall dimensions of the cut were not recorded, but care was taken to leave enough material at each section location to ensure electrical conductivity but remove enough so sections could be separated from one another after EC trails.

The sectioning of these areas produced 6 added surfaces that were exposed to solution and EC. The surface area of these surfaces was factored into the current calculation, calculated to produce an applied current density of 15 mA/cm2. It was assumed that the material left from the sectioning process was minimal, such that the total width and plate thickness dimensions were used to calculate the additional exposed surface area. The 4 test specimens generated from this method allowed for an average diffusible hydrogen concentration to be calculated, as well as the distribution of diffusible hydrogen along the specimen i.e. from test sample #1 to sample #4.

a) Figure 3.13

3.8.2

b)

(a) Schematic representation with specimen geometry and the locations where test specimens were sectioned. The black areas represent material that was left after the sectioning process was complete. (b) Visual representation of samples fabricated for diffusible hydrogen analysis using EC. The sectioned areas allow for easy detachment after EC has been conducted.

Mercury Displacement Experimental Procedure The procedure outlined by the AWS [46] covers correct handling, storage, and placement of specimens into

the mercury filled eudiometers for measurement, as well as incubation periods and the final measurement of diffusible hydrogen. The following section highlights important details from the AWS standard pertinent to setting up the procedure for the current study.

The specimen of interest is required to be placed into a low-temperature liquid bath (-60 °C (-76 °F)) within 60 seconds after removal from the hydrogen environment. After test specimens have been in this liquid bath for two minutes, they may be removed for further handling, i.e. cleaning, detaching specimens from one another, and/or being placed into the mercury displacement apparatus [45]. It is stated that when removed from the low-temperature bath, a specimen has to be placed back into the bath after one minute exposure to ambient conditions and only after

38

two minutes in the bath may it be removed again [45]. This procedure is used each time a specimen is removed from the liquid nitrogen bath to ensure that hydrogen remains trapped before further analysis.

Diffusible hydrogen analysis is performed with a mercury-filled eudiometer. Figure 3.14 shows schematically the dimensions of the eudiometer (Figure 3.14a), as well as the evolution of hydrogen from a sample placed into the mercury-filled eudiometer (Figure 3.14b). The test apparatus requires loading a test specimen into the incubation chamber located at the lower end of the eudiometer, allowing for hydrogen to diffuse out of the sample and recombine into molecular hydrogen. The process of removing the test specimen from the lowtemperature bath and placing into the incubation chamber is advised to be within 150 seconds from start to finish to ensure reliable results.

During the degassing of the sample in the eudiometer, hydrogen displaces the mercury away from the top valve (the Teflon stopcock) towards the mercury bath at the bottom. Because there is a physical height difference of mercury in the eudiometer and the lower mercury bath, there is a pressure difference for the evolved hydrogen compared to the ambient pressure on the mercury bath. When calculating the volume of hydrogen, this pressure difference is incorporated. The time required before evaluation is dependent on the liquid bath temperature. After the incubation period, the volume of hydrogen present and height of displacement are measured to calculate diffusible hydrogen value at STP. The volume of hydrogen gas is calculated as,

(3.7) where T is the temperature (°C) of the gas column at the time of measurement (room temperature), P is the barometric pressure (mm Hg) at the time of measurement, V is the measured volume (mL) taken from the eudiometer, H is the head height of mercury (mm) at time of measurement as shown in Figure 3.14b, and VH is the volume of hydrogen gas at STP in milliliters [45].

The same conditions used to generate HIC by EC were employed to test for diffusible hydrogen in the sectioned specimens for each alloy and prestrain conditions. After EC was completed, the steel specimen was removed from the EC test apparatus, rinsed with water, and placed into a low-temperature liquid nitrogen bath within 60 seconds. The specimens were transported to the displacement test apparatus. The mercury test setup is shown in its entirety in Figure 3.15; shown are the four mercury-filled eudiometer tubes, the temperature setting and thermometer for the water bath, and the ambient pressure gauge. When placing the samples into the eudiometers, the following procedure was implemented: 

Remove full-size test specimen from liquid nitrogen bath



Using pillars, break sample of interest from the full-size test specimen (placing remaining pieces back into the bath of liquid nitrogen)



Clean in liquid water bath to remove layer of ice created by exposure to the atmosphere



Dry sample so no moisture is present on the surface (failing to do so will give inflated readings of VH)



Place sample into the bottom end of the eudiometer; this assembly is then inverted and placed into the mercury

39

ď&#x201A;ˇ

The mercury is drawn up to the top of the eudiometer with a vacuum, after which the top valve (the Teflon stopcock) is closed to allow for hydrogen to be captured and displace the mercury

Figure 3.14

(a) Dimensions of the eudiometer used in the mercury displacement method to determine diffusible hydrogen amount, (b) schematic representation of the use of a mercury-filled eudiometer to capture and measure the amount of diffusible hydrogen in the sample that is placed into the assembly. Adapted from [45].

40

Figure 3.15

Test setup for the mecury displacement method with the compontents labeled.

For the current study, the temperature of the liquid bath was set to 50 °C, which corresponds to a 72 hour incubation period. After the incubation period is completed, the height, volume, and other required variables at time of measurement are recorded such that VH (Equation 3.7) can be calculated for each specimen. In order to compare the measured hydrogen contents with those obtained through other hydrogen analysis techniques, the VH value was converted to parts per million. The parts per million of diffusible hydrogen for each sample was calculated as, (3.8) where VH is the calculated value from Equation 3.7, msteel is the mass of the individual steel specimen, is the density of molecular hydrogen gas at STP, and 10-6 is the conversion factor to parts per million.

Uncertainty for the mercury displacement method was calculated by the root sum of squares equation

(3.9)

wppm =

where wH is the uncertainty for the height measurement taken by a tape ruler with a resolution of 1 cm, wM is the uncertainty from the analytical balance used to measure the mass of the individual samples, wP is the uncertainty in the ambient pressure gauge, wT is the uncertainty in the thermometer used to determine room temperature, wV is the uncertainty of the volume taken from the eudiometer, and ,

,

,

,

are the partial derivatives of Equation 3.8 with respect to each variable that was

41

measured. Uncertainty values for Equations 3.7 and 3.8 are shown in Table 3.11. Using a range of values of diffusible hydrogen from results presented in section 4.5, total uncertainty calculations were made for each material using Equation 3.9 and values of uncertainty in Table 3.11. The results for total uncertainty are shown in Table 3.12.

Lower amounts of diffusible hydrogen in the X52 produced lower amounts of uncertainty (0.02 ppm), and at the highest value of diffusible hydrogen measured in the 100XF, the uncertainty increased to 0.05 ppm.

Table 3.11 Uncertainty Values for the Instruments Used to Calculate Diffusible Hydrogen Method Eudiometer Pressure gauge Tape ruler Analytical balance Standard thermometer

Variable V P H msteel T

Uncertainty 0.01 mL 0.01 mm Hg 5 mm 0.01 grams 0.01 °C

Table 3.12 Total Uncertainty Calculated using Equation 3.9 for the Mercury Displacement Method Material

Method

Variable

Diffusible hydrogen (ppm)

X52 X60 X70 100XF

Mercury Displacement Mercury Displacement Mercury Displacement Mercury Displacement

ppm ppm ppm ppm

1.56 3.99 3.97 6.42

42

Total uncertainty (ppm) 0.02 0.04 0.04 0.05

CHAPTER 4: Results and Discussion

This chapter will present and discuss the results obtained from the different experimental methods discussed in the previous chapter.

4.1

Microstructure and Nonmetallic Inclusions The microstructure of the as-received materials was evaluated using standard metallographic techniques.

Faces in the transverse direction (TD), rolling direction (RD), and the normal direction (ND) were polished and etched (2 pct Nital).

The X52 plate material was produced by controlled rolling and then cooled using on-line accelerated cooling. FESEM micrographs of X52 are shown in Figure 4.1. Micrographs were taken on the three different planes to evaluate the uniformity of the microstructure. The micrographs exhibit a mixed ferrite/pearlite structure. The ferrite in the microstructure is polygonal or equiaxed. Figure 4.2 shows higher resolution image of the secondary phase that formed in the X52. The pearlite (Figure 4.2) in the microstructure is not fully developed and does not resemble the conventional lamellar morphology, caused by the low carbon levels (hypoeutectoid) and the cooling operations performed on the X52. It is described as degenerated pearlite, which forms due to insufficient carbon diffusion when the eutectoid reaction proceeds at a low temperature [46]. The degree of elongation with respect to rolling direction is minimal, i.e. minimal pancaking of grains, and the degree of banding is low.

The X60 and X70 material were both produced from continuous cast slabs that were subsequently hotrolled into plate. The micrographs of the three planes with respect to rolling direction are shown in Figures 4.3 and 4.4, for the X60 and X70 respectively. The difference in microstructures is minimal as shown in Figures 4.3 and 4.4. The microstructure is dominated by acicular ferrite with some quasi-polygonal ferrite. The presence of second phase microconstituents (martensitic islands) appears to be slightly greater in the X70, as shown by the circled regions in each micrograph. Rolling operations and cooling practices for these plate materials resulted in a small amount of elongation of grains with respect to the rolling direction.

The three micrographs from each plane with respect to rolling direction for the 100XF steel are shown in Figure 4.5. The plate was produced by controlled rolling followed by accelerated cooling, which is expected to result in an acicular ferrite microstructure [42]. The micrographs show elongated grains as well as some that are more equiaxed in shape. The microstructure can be described as acicular ferrite with minimal quasi-polygonal ferrite. The 100XF exhibits relatively higher amounts (visual comparison) of second phase microconstituents, indicated by the black circled regions in the micrographs, compared to the other alloys.

The plates selected in the current study had calcium additions to aid in controlling the types (composition) and morphologies of nonmetallic inclusions present. As mentioned in the materials selection section, the X52 and X60 materials were produced for use in hydrogen environments. A general guide for steels used in sour service is to have ratios of Ca:S at levels greater than 2 [24]. It is expected that the addition of calcium resulted in inclusions that are globular instead of elongated and the amount of MnS type inclusions is minimized. During steelmaking, the

43

aluminum and calcium react with the oxygen and sulfur before manganese or other elements in the melt, therefore minimizing the formation of MnS type inclusions [24].

a)

b)

c)

Figure 4.1

Secondary electron micrographs taken with the FESEM of X52 plate steel. Images from three different planes (a) transverse, (b) longitudinal, and (c) normal plane are shown. 2 pct Nital etch.

Figure 4.2

Presence of degenerated pearlite in the X52 plate material. Image taken with the FESEM. 2 pct nital etch.

44

Figure 4.3

Figure 4.4

a) b) c) Secondary electron micrographs taken with the FESEM of X60 plate steel. Images from three different (a) transverse, (b) longitudinal, and (c) normal plane planes are shown. 2 pct nital etch.

Black circles on each micrograph indicate the presence of a secondary microconstituent.

a) b) c) Secondary electron micrographs taken with the FESEM of X70 plate steel. Images from three different planes (a) transverse, (b) longitudinal, and (c) normal plane are shown. 2 pct nital etch.

Black circles on each micrograph indicate the presence of a secondary microconstituent.

a) Figure 4.5

b)

c)

Secondary electron micrographs taken with the FESEM of 100XF plate steel. Images from three different planes (a) transverse, (b) longitudinal, and (c) normal plane are shown. 2 pct nital etch.

Black circles on each micrograph indicate the presence of a secondary microconstituent.

45

Nonmetallic inclusions in each steel are shown in the images of Figure 4.6. The micrographs and their corresponding EDS maps represent the common shape, size, and compositions of the inclusions that were found.

Because an extensive statistical inclusion study was not conducted, the possibility of other types, sizes, and morphologies cannot be ruled out. Metallographic observations of the alloys showed that the inclusions were 6 Âą 1 Âľm2 on average and were largely globular in their morphology (Figure 4.6). Several inclusion types were found including: Al-Ca-O (Figure 4.6a), Al-O (Figure 4.6b), Al-Mg-O-Ca-S (Figure 4.6c), and Al-Mg-O (Figure 4.6d). The composition of inclusions presented here were present within each steel evaluated.

a)

Figure 4.6

b)

c) d) Nonmetallic inclusions observed in the four plate steels. Elements present in each image were confirmed by EDS mapping of the image shown. (a) Al-Ca-O X52, (b) Al-O X60, (c) Al-Mg-OCa-S X70, (d) Al-Mg-O 100XF (see pdf version for color).

The ternary diagrams produced for X52 from two inclusion families, 1) Al, Ca, S and 2) Ca, Mn, S, are shown in Figure 4.7a and b, respectively. The values of composition determined by the tie lines in the ternary diagram are relative to the other elements present, not necessarily the actual composition of inclusion evaluated. The grouping of compositions in the two ternary plots is an example of the qualitative behavior modified steel for use in sour service would display [24]. Region 1 in Figure 4.7a shows inclusions in the family of Al-Ca-S (on average) have relative levels of Ca at 40 to 80 wt pct, S at levels of 50 wt pct or less, and Al at a maximum of 30 wt pct .

Lower amounts of sulfur in the inclusion compositions are expected because of the ultra-low level of sulfur (0.0007 wt pct) that was achieved during plate production. Figure 4.7b shows that the amount of MnS inclusions formed in the X52 is minimal. Most of the inclusions in the Ca-Mn-S family appear to be CaS. This is displayed by the dashed region 2 in Figure 4.7b: Ca levels range from 30 to 85 wt pct, S from 15 to 70 wt pct, and Mn from 0 to 20 wt pct.

Figure 4.8 shows the ternary diagrams for the X60 inclusions in the Ca-Al-S and Ca-Mn-S families.

Dashed regions in Figure 4.8a and b represent the common range of inclusion compositions within each ternary

46

diagram. As shown in the dashed regions 1 and 2 in Figure 4.8a and b, the inclusions in the X60 plate material have a higher tendency compared to the X52 (Figure 4.7b) to be rich in sulfur, at least 30 wt pct or greater. The X60 also exhibits higher propensity to form aluminum rich inclusions, shown in the dashed region 1 of Figure 4.8a. There is also an increase in the presence of manganese in the inclusions in comparison to the X52 with a large number of values ranging from 0 â&#x20AC;&#x201C; 50 wt pct, shown in the dashed region 2 of Figure 4.8b. The level of manganese added to the X60 is 1.42 wt pct compared to 1.03 wt pct for the X52. The higher manganese content present in the X60 increases the likelihood for more MnS type inclusions to form with unreacted sulfur in the melt, causing the shift in Ca-Mn-S ternary plot towards the manganese sulfur tie line (Figure 4.8b). It was expected based on literature findings [3, 10, 12, 15, 22, and 23] that cracks generated in this material will originate around MnS inclusions.

a) Figure 4.7

b)

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the X52 plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram.

The analysis of the two inclusional families 1)Al-Ca-S, and 2) Ca-Mn-S for the X70 are shown in Figure 4.9. Dashed regions 1 and 2 in Figure 4.9a and b represent the common range of inclusion compositions within each ternary diagram produced from AFA. The bulk of inclusions present in the X70 material are greater in calcium (Figure 4.9a and b) than the levels found in the X60. Like the X60, there is more scatter in the compositions of the inclusions when compared to the X52 material, as indicated by the qualitative assessment of the areas encompassed by the two dashed regions. The relative amount of inclusions with the composition rich in manganese is comparable to the X60 (Figure 4.8b) but greater than the X52 (Figure 4.7b). This is expected from the amount of manganese present, 1.59 wt pct. The X70 compositions indicate the presence of MnS (regions 3 in Figure 4.9b). The formation of MnS is related to ladle metallurgy practices [24]. It was expected that cracks generated through hydrogen exposure in the X70 would originate around the sulfide type inclusions because of their ability to irreversibly trap hydrogen.

47

a) Figure 4.8

b)

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the X60 plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. Regions 3 in (b) indicate the presence of MnS type inclusions identified through AFA.

a) Figure 4.9

b)

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the X70 plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. Regions 3 in (b) show the presence of MnS type inclusions identified through AFA.

48

Figure 4.10 displays the ternary diagrams generated by AFA analysis. Within the two separate ternary plots, a and b in Figure 4.10, the dashed areas (regions 1 and 2) represent the common range of inclusion compositions. In region 1, the range inclusion compositions are Ca 25 – 75 wt pct, S 35 – 75 wt pct, and Al - < 35 wt pct (Figure 4.10 a). In region 2, the range of inclusion compositions are Ca < 65 wt pct, S > 35 wt pct, and Mn - < 30 wt pct (Figure 4.10b). The 100XF material had the highest amount of manganese (1.8 wt pct) of all the materials, but unlike the X60 and X70 (at a similar level of sulfur 0.003 wt pct), the inclusion compositions did not show a shift towards the manganese sulfur tie line. This is believed to be attributed to differences in steeling making operations that the 100XF underwent compared to the X60 and X70. Steel casting operations were not obtained from the steel producers. The 100XF exhibits low amounts of aluminum type nonmetallic inclusions in the steel matrix, shown in region 3 of Figure 4.10a. It is speculated that if the inclusions within this area (Figure 4.10a region 3) were evaluated with other elements, i.e. Fe and O, the composition of the inclusion would be different, i.e. Fe-S and Al-O type inclusions.

a) Figure 4.10

4.2

b)

Ternary diagrams (in wt pct) of two inclusions families, (a) Al-Ca-S and (b) Ca-Mn-S, present in the 100XF plate steel. Ternary diagrams were generated by automatic feature analysis at the Nucor Indiana facility. Dashed regions 1and 2 in (a) and (b), respectively, show the grouping of the relative compositional distribution of inclusions evaluated by AFA, in reference to the three elements found in each ternary diagram. Regions 3 in (a) show a region where the types of inclusions within the area if evaluated with other elements i.e. Fe and O the composition of the inclusion would be found to be different. Speculated that these inclusions would be Fe-S and Al-O type inclusions.

Mechanical Properties Figures 4.11 – 4.14 show microhardness data from each rolled face to the center of each plate thickness,

along with the average microhardness (solid line with 90 pct confidence interval calculated from the standard deviation) in part (a). Part (b) shows etched (2 pct Nital) macrographs taken from the rolled face to middle thickness.

Finally, part (c) shows non-etched macrographs taken from the rolled face to mid-thickness. This analysis of

49

hardness in part (a) was used to identify the presence and location of hard bands. The microstructure as a function of plate thickness was evaluated from the macrograph shown in part (b). The non-etched specimens in part (c) reveal inclusion location and distribution.

Figure 4.11shows the hardness traverse, microstructure, and inclusion distribution as a function of thickness for the X52steel. The hardness in the X52 decreases from each rolled face, stabilizing to a constant value after approximately 8 pct through-thickness (Figure 4.11a). There is also a slight rise in hardness from the average hardness (162 HV), towards the centerline of material, but otherwise there are no hard bands through the thickness..

The mostly consistent hardness through-thickness is consistent with the homogenous microstructure present from the edge to mid-thickness (Figure 4.11b). The inclusions in X52 also appear to be distributed homogeneously (Figure 4.11c).

a) Figure 4.11

b)

c)

(a) Microhardness data from each rolled face to the center of the X52 plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Cross-sections of the full plate thickness of the transverse plane with respect to rolling direction were evaluated.

Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time.

50

Figures 4.12 and 4.13 show the hardness traverse, microstructure, and inclusion distribution as a function of thickness for the X60 and X70 alloys. The X60 and X70 show similarities with respect to the hardness traverse taken from each rolled face (Figures 4.12a and 4.13a). Hardness is greater than average (X60 â&#x20AC;&#x201C; 202HV and X70 â&#x20AC;&#x201C; 210HV) from the rolled face to a thickness of 20 pct plate thickness; after the quarter positions in each material, the values are slightly lower than the average hardness. The X60 and X70 grade have very similar mechanical properties and microstructures that were produced during steelmaking and subsequent processing operations. There are higher amounts of banding and elongation towards the mid-thickness for the X70. The X60 and X70 exhibited no real observable hard bands based on the evaluation of (Figures 4.12b and 4.13b) or inhomogeneous inclusion distribution (Figures 4.12c and 4.13c). Because a comprehensive study on inclusion distribution was not conducted during the thesis work, areas of alloy segregation cannot be ruled out.

a) b) Figure 4.12

c)

(a) Microhardness data from each edge to the center of the X60 plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Cross-sections of the full plate thickness of the transverse plane with respect to rolling direction were evaluated. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time.

51

Figure 4.14 shows the hardness traverse, microstructure, and inclusion distribution as a function of thickness for the 100XF alloy. The 100XF edge 2 exhibited a higher hardness, relative to the average hardness (280 HV), for the first 20 pct through thickness of the plate (Figure 4.14a). The higher values of hardness produce areas that would be more susceptible to HIC, and thus, the first 20 pct through thickness of the material might be more vulnerable to HIC damage. It was inferred the edge experienced higher cooling rates, allowing for higher hardness to be achieved during rolling operations. The macrographs (Figure 4.14b) produced for the 100XF do not exhibit higher amounts of second phase microconstituents or drastic changes in microstructure due to plate thickness. Additionally, inclusions appear to be equally distributed throughout the matrix (Figure 4.14c).

a) b) Figure 4.13

c)

(a) Microhardness data from each edge to the center of the X70 plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Cross-sections of the full plate thickness of the transverse plane with respect to rolling direction were evaluated. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time.

52

a) b) Figure 4.14

c)

(a) Microhardness data from each edge to the center of the 100XF plate thickness, along with the average microhardness (solid line with 90 pct confidence interval), (b) etched (2 pct Nital), and (c) non-etched macrographs taken from the edge to middle thickness. Microhardness measurements were taken every 0.254 mm (0.01 in), 500 gmf load, and 10 s dwell time.

Hardness measurements after prestrain were conducted as a measure of the amount of work hardening that occurred due to prestrain. Results from the hardness measurements are shown in Table 4.1. The hardness for the X52 increases the most due to the higher amounts of prestrain introduced (increase from 157 to 200HV). At similar levels of prestrain, 5 pct, the X60 hardness increases more than the X70 (increase of 22HV and 15HV respectively), but the overall hardness at 5 pct prestrain is greater in the X70 (218 HV) than the X60 (211 HV). The low prestrain amount for the 100XF does not increase the hardness. Hardness for plate steels used for sour service are is advised to be below 248 HV [3]. Based on the hardness measurements, HIC susceptibility is expected to be greatest in the 100XF condition.

53

Table 4.1 Microhardness Data Taken for each Material and Prestrain Condition Material & Prestrain Condition X52 0% X52 12% X52 18% X60 0% X60 3% X60 5% X70 0 % X70 5 % X70 7% 100XF 0% 100XF 2 %

4.3

Microhard ness (HV) 157 200 200 190 206 211 203 218 222 271 272

90 % Confidence (HV) 2 4 1 1 3 3 1 3 4 2 2

Characterization Study A characterization study on the fabricated double cell (Figure 3.8) was conducted to optimize charging

conditions. This study tested the ability of the test method to produce cracks in the material. The 100XF as-received plate was selected for the characterization study, as it was assumed to have the highest HIC susceptibly. The high susceptibility allows for ease in identifying parameters that could alter charging conditions during the characterization study.

Two variables, 1) applied current density and 2) duration of test, were varied to observe their effects on HIC damage; the applied current density ranged from 0.80 to 25 mA/cm 2 (duration set at 24 hours), and charging times of 24, 36, and 48 hours were assessed (applied current density set at 15 mA/cm2). The experimental matrix is shown in Table 4.2. Current density was varied first, while holding test duration constant. A total of two tests were conducted for each condition, with the exception of the 25.00 mA/cm2 condition due to deleterious conditions that occurred during charging, which are described later. Once the optimal applied current density was identified, it was held constant while test duration was varied. Three tests for each condition were conducted during this round of experiments to identify the influence of test duration.

Table 4.2 Experimental Matrix for the Characterization Study Conducted on the 100XF alloy Applied Current Density (mA/cm2) Duration of Test 0.80 1.50 5.00 10.00 15.00 25.00 (hrs) 24 2 2 2 2 3 1 36 3 48 3 The numbers denote the number of tests conducted at each test duration and applied current density pairing After charging the specimens during applied current density experiments, the specimen surface conditions were observed. A visual comparison of the surface conditions for the range of applied current densities is shown in Figure 4.15. The luster of the charged specimens is affected the applied current density. The most pronounced

54

effects are observed in the 0.80 and 25 mA/cm2 cases. The luster of these two specimens is far less than the other four. It is speculated that the 0.80 mA/cm2 current density is associated with a lower cathodic-overpotential (compared to the higher current density), and as such, galvanic corrosion of iron can occur at the metal/solution interface (Figure 4.15a). A possible surface reaction at the highest current density, 25 mA/cm2, is also possible, although the applied current density was sufficient to provide cathodic protection. At this current density, the cell voltage and current changed significantly from the initial readings during testing. After 18 hours of EC charging, the cell potential changed from 20.85 Volts to 28.51 Volts, and the current dropped from 1.622 Amps to 0.725 Amps.

Deterioration and mass loss of the steel sample were observed, which subsequently polluted the electrolyte and is believed to be the cause of the change in potential (Figure 4.15f). At intermediate current densities of 5 to 15 mA/cm2, surface blisters appeared on the surface of the test specimens (Figure 4.15c â&#x20AC;&#x201C; e). The amount of blisters present on the surface of the 15 mA/cm2 condition (Figure 4.15e) was greater than the other two applied current densities.

a) Figure 4.15

b)

c)

d)

e)

f)

Surface conditions of 100XF after cathodic charging at various applied current densities while keeping the duration of the test constant at 24 hours. (a) 0.80, (b) 1.50, (c) 5.0, (d) 10.0, (e) 15, and (f) 25 mA/cm2. Image taken using light optical flash photography.

Upon completion of hydrogen charging, specimen faces 2, 4, and 6 (Figure 3.9a) were sectioned, polished, and evaluated for cracks. Face 2 is the furthest from the electrode connection as shown in Figure 3.6. The charging conditions that produced cracking for each of the faces and those that did not are shown in Table 4.3. For the lower current densities of 5.00 and 10 mA/cm2, cracking was not observed throughout the entire specimen. An applied current density of 15 mA/cm2 produced hydrogen cracks on each of the faces evaluated. The typical cracks observed are shown in Figure 4.16. The cracks generated were mainly long and straight, but there were some regions of stepwise cracking. The CLR, CTR, and CSR are plotted as a function of applied current density in Figure 4.17. For the values of current density that were evaluated, the crack ratios reached their maximum at 15 mA/cm2 and then

55

decreased at 25 mA/cm2. The values for the 25 mA/cm2 current density condition are likely affected by the surface effects that were observed during testing (Figure 4.15f). Table 4.4 shows the average crack depth (measured from the top surface) as a function of the applied current density. Most of the cracks occurred at depth of approximately 2 mm (Figure 4.16), and very few samples exhibited cracks at the centerline (~ 6.25 mm). The applied current density of 15 mA/cm2 produced the largest range of crack depths. Because an applied current density of 15 mA/cm2 produced the greatest extent of cracking it was chosen for the current density for further experiments.

The effect of charging time on hydrogen cracking was explored by performing experiments at a range of charging-times while controlling current density (15 mA/cm2). The diffusivity of hydrogen in BCC ferrite is reported to be 10-6 to 10-4 cm2/s at room temperature [47 and 48]. The hydrogen penetration for the different charging conditions can be estimated based on these diffusivity values. A simple estimate for diffusion distance (lattice diffusion) is [49]:

â&#x2C6;&#x161;

(4.1)

where x is the diffusion distance, Dz is the hydrogen diffusivity, and t is the elapsed time in seconds during which diffusion occurred. The equation is an approximation since it is based on unidirectional diffusion in a system where the dimension in the diffusion direction is infinite.

Cracks in the 100XF material were predominately within 3 mm of the outer exposed surface, perhaps implying that damage deeper in the specimens was diffusion limited after charging for 24 hours. Using a diffusivity of 10-6 cm2/s, a lower limit, the diffusion distance of hydrogen at room temperature is approximately 2.9 mm for 24 hours, which approximately correlates to the value of crack depth observed at an applied current density of 15 mA/cm2 at the test duration of 24 hours (Table 4.4). Table 4.5 shows crack depth data as charging time is changed. At longer charging times of 36 and 48 hours (at 15 mA/cm2), the depth of cracking did not change. Also, electrolytic charging of the 100XF occasionally produced cracking close to the centerline (6.06 mm at 15 mA/cm2 and 24 hours) and thus, it was concluded that hydrogen was able to diffuse through the thickness of the specimens within 24 hours. The crack depths observed for this material is believed to be related to the higher hardness in the first 20 pct through thickness of the plate (Figure 4.14a). It is also believed that cracks that are generated in the first 24 hours inhibit the growth of other cracks deeper in the specimen. In support of this argument, Griesche and Dabah et. al. [50], electrically charged duplex stainless steel samples and evaluated the presence of hydrogen though neutron imaging. The results from this study showed that sub-surface pores (generated from the blister/void mechanism) had high amounts of hydrogen, compared to other areas found in the cross-section of the material. The pores were likely produced by the volume expansion produced through the formation of H 2 at grain boundaries, producing internal stress that created the cavity. If hydrogen is trapped in the initially cracked regions, HIC deeper in the specimen could be inhibited.

The effect of test duration on the crack ratios is shown in Figure 4.18. The CLR at all test durations is greater than the CTR values. The scatter in the data for the critical crack ratios suggests that test duration does not significantly change the amount of cracking in the 100XF.

56

Table 4.3 Charging conditions that produced cracking in 100XF that was cathodically charged for 24 hours Applied Current Density (mA/cm2) Face 0.80 1.50 5.00 10.0 15.0 Examined 2 X X X 4 X 6 X X Marks Presence of Cracks on Examined Cross-sections

a)

25.0 X -

b)

Figure 4.16

Hydrogen Induced Cracks produced through electrolytic charging of 100XF at a current density of 15 mA/cm2 for 24 hours, (a)Face #2 (b) Face #4.

Figure 4.17

Variation of calculated crack ratios versus applied current density for 100XF that was cathodically charged for 24 hours. CSR – Critical Size Ratio, CLR – Critical Length Ratio, and CTR – Critical Thickness Ratio.

57

Table 4.4 Average crack depth in 100XF specimens at the current densities explored during EC experiments run for 24 hours: Extremes represent the minimum and maximum crack depth observed

2

mA/cm 0.8 1.5 5 10 15 25

Average Thickness = 12.5 mm Standard Deviation Depth (mm) (mm) 2.13 0.09 1.54 0.90 2.33 1.87 1.78 0.07

Extremes Min (mm)

Max (mm)

2.07 0.34 0.42 1.73

2.19 2.17 6.06 1.83

Table 4.5 Average crack depth in 100XF specimens as a function of test duration during EC experiments, run at 15 mA/cm2: Extremes represent the minimum and maximum crack depth observed Average Thickness = 12.5 mm Standard Deviation Test Duration (hrs.)

Depth (mm) (mm) 24 2.33 1.87 36 2.12 0.87 48 2.17 0.68

Figure 4.18

4.4

Extremes Min (mm)

Max (mm)

0.42 0.64 0.68

6.06 2.64 2.60

Variation of average calculated crack ratios versus charging time for 100XF cathodically charged at an applied current density of 15 mA/cm2.

HIC Susceptibility Measurements This section presents the degree of cracking that was observed for each alloy and prestrain condition and

discusses the results in light of differences in mechanical properties (tensile strength and hardness).

58

4.4.1

Influence of Prestrain HIC was observed in all but the X52 12 pct and X60 3 and 5 pct prestrain conditions exposed to the H2S

method. Influence of Prestrain on Calculated Critical Crack Ratios Each pre-strained test specimen was charged with hydrogen either by the EC method or H 2S method. The crack ratios for each of the two charging methods, alloy, and pre-strain condition are shown in Figure 4.19 - Figure 4.22. Error bars on the plots show the 90 pct confidence interval, calculated based on the standard deviation of the population as well as the sample size.

The results from the EC and H2S methods are comparable in magnitude to one another even though there is generally significant scatter in the data for both methods. The most significant difference is apparent for the X60 alloy (Figure 4.20), which exhibited significant cracking when charged by the EC method, but only minor cracking when charged by the H2S method. The differences in behavior are believed to be due to the statistical uncertainty of the evaluation method used. The statistical uncertainty of the evaluation method is possibly related to sectioning predetermined areas of the full specimen geometry. The sectioning process could miss crucial cracks that would contribute to the calculation of critical crack ratios. The critical crack ratios for the 100XF material and prestrain conditions are considerably greater compared to the other three alloys.

a) Figure 4.19

b)

Critical Crack Ratio values as a function of prestrain for the X52 material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio.

59

The HIC results are largely independent of pre-strain level within the statistical accuracy of the methods employed with the possible exceptions found in the X52 (Figure 4.19a) and X70 (Figure 4.21a). The EC CLR results for X52 and X70 possibly exhibit an effect of pre-strain, though HIC increases with increasing prestrain for the X70 while HIC decreases with pre-strain for the X52. It is speculated that dislocation density increases with increasing prestrain levels, which might be expected to mitigate HIC by trapping hydrogen at dislocations [51] rather than more detrimental sites such as inclusions in the material. The X52 results are consistent with this hypothesis. However, prestraining increases hardness and decreases ductility, which might make steel more vulnerable to hydrogen embrittlement, consistent with the X70 results. It is also worth noting that the prestrain levels of the X70 are lower than the pre-strain levels of the X52, and thus, the increase in trapping sites is less in X70. However, neither of the trends observed in the X52 and X70 alloys are observed in the other alloy conditions for either test method, so the statistical certainty based on the evaluation method is unclear. The scatter observed with the critical crack ratios and prestrain is believed to be a reflection of the large variance associated with, and intrinsic to, the evaluation method employed.

a) Figure 4.20

b)

Critical Crack Ratio values as a function of prestrain for the X60 material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio.

60

a) Figure 4.21

b)

Critical Crack Ratio values as a function of prestrain for the X70 material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio.

a) Figure 4.22

b)

Critical Crack Ratio values as a function of prestrain for the 100XF material for each charging method: (a) Electrolytic charging and (b) H2S Method. CLR – Crack Length Ratio, CTR – Crack Thickness Ratio, and CSR – Crack Size Ratio.

61

4.4.2

Influence of Mechanical Properties on HIC The HIC susceptibility of the non-strained specimens was compared based on the tensile strength. As the

strength of the material increases, the tolerance to hydrogen concentration generally decreases [14 and 15]. Figure 4.23 shows the dependence of CLR and CTR values on tensile strength for non-strained test specimens, exposed to both EC and H2S methodologies. Horizontal lines in Figure 4.23 represent the values for which steel plates are deemed fit for sour service in accordance with ISO 3183 [52]; this standard states that at values of CLR less than 0.15 (Figure 4.23a) and CTR less than 0.05 (Figure 4.23b), plate steel is qualified for sour service. The vertical line represents the NACE recommended maximum tensile strength (116 ksi) to avoid hydrogen-assisted cracking phenomena. Figure 4.23a shows the relationship between the CLR values from each method as a function of tensile strength for the non-strained conditions. The CLR values presented in Figure 4.23a are consistent with the expectation that HIC increases with increased strength [14 and 15]. The same trend is observed for the CTR values shown in Figure 4.23b. These results also show that the difference between the HIC resistance of the X52, X60, and X70 alloys from the 100XF alloy is approximately the same for samples charged by both the EC and H 2S methods.

Differentiation between the HIC resistance of the X52, X60, and X70 alloys is not evident with any statistical significance (scale of 90 pct confidence intervals found in Figures 4.19 â&#x20AC;&#x201C; 4.21) for results obtained from either method. Based on the results for both CLR and CTR, the X52, X60, and X70 materials could be considered for sour service.

a) Figure 4.23

b)

a) Crack length ratio dependence on tensile strength for the 0 pct prestrain condition: Electrolytic method (EC) and NACE Standard method (H2S). b) Crack thickness ratio dependence on tensile strength for the 0 pct prestrain condition. The vertical dashed lines represent the suggested 116 ksi threshold value for hydrogen assisted-cracking phenomena [47 and 48]. The two horizontal dashed lines represent sour service requirements outlined by ISO 3183 [52].

62

Hardness measurements after the introduction of prestrain were performed to relate HIC susceptibility to the hardness change of each condition. Figure 4.24a and b shows the CLR and CTR values as a function of the measured hardness for each condition after prestraining. Neither crack ratio is strongly affected by hardness in the X52, X60, and X70 alloys. The ISO 3183 crack ratio values are plotted as horizontal dotted lines on Figure 4.24, and the vertical dotted lines show the suggested threshold hardness (22 HRC (248 HV)) where materials become susceptible to hydrogen embrittlement. As expected (based on its higher strength and lower ductility), the 100XF at 0 and 2 pct prestrain has crack ratio values that disqualify it for use in sour service [3, 14, and 15]. All of the lower strength alloys have lower hardness values than this threshold, even after pre-straining, which is consistent with their low HIC susceptibility. Their relatively low hardness may also explain why there is no apparent and consistent effect of pre-strain on their HIC susceptibility.

a) Figure 4.24

4.5

b)

(a) Crack length ratio and (b) crack thickness ratio dependence on hardness for the materials and pre-strain conditions evaluated by EC. The vertical dashed lines represent the suggested 22 HRC (248 HV) threshold value for HIC susceptibility [47 and 48]. The two horizontal dashed lines represent sour service requirements outlined by ISO 3183 [52].

Hydrogen Analysis The following section presents the hydrogen contents found by both hydrogen analysis methodologies

presented in section 3.7 (LECOÂŽ) and 3.8 (mercury displacement). The results of total hydrogen concentration immediately after EC and values of trapped hydrogen (in as-received and degassed specimens) are presented in section 4.5.1. Values of diffusible hydrogen concentration directly calculated by the mercury displacement method and estimated using the LECO measurements and Equation 3.6 are presented in section 4.5.2. The LECO and mercury displacement method used different sample geometries, and the influence of sample size on diffusible hydrogen results is discussed in section 4.5.3. HIC susceptibilities are then related to the diffusible hydrogen values and the amount of hydrogen permanently trapped in the material after exposure to EC.

63

4.5.1

Average Hydrogen Contents Total and Trapped from LECOÂŽ Analysis Initial hydrogen concentrations in as-received plate steels were evaluated using the LECOÂŽ Hydrogen

Analyzer; the results are displayed in Figure 4.25a. The values represent the average of 5 measurements taken on each as-received plate steel. Error bars represent a 90 pct confidence interval calculated on the standard deviation of those 5 measurements. All of the alloys had levels of hydrogen less than 0.45 ppm (Figure 4.25a). These levels suggest that the as-received alloys did not have hydrogen in high enough levels to produce HIC damage.

Using the daisy chain setup described in section 0.1, total hydrogen concentration in all of the conditions was evaluated after exposure to the EC method; the results are shown in Figure 4.25b. The values represent the average of 5 measurements taken for each material and prestrain condition. Error bars represent a 90 pct confidence interval calculated on the standard deviation of those 5 measurements. The total hydrogen concentration consists of both trapped and diffusible hydrogen. The total hydrogen present after EC varied between 2 and 6 ppm. The hydrogen concentration in the X52 and X70 increases with increasing prestrain level. The hydrogen concentrations of the X60 and 100XF alloys are independent of prestrain level. Hydrogen concentration is greatest for the 100XF (~ 6 ppm) and lowest for the X52 (~ 3 ppm). The comparison of these two materials represents the extremes of the materials investigated. These two materials in the as-received and prestrain conditions have drastically different levels of hardness (157 HV for X52 and 270 HV for 100XF), and the microstructure was significantly different as well. Conversely, the similarities of chemical chemistry and microstructure for the X60 and X70 allow for similar levels of hydrogen to be achieved (~ 4 ppm). The effect of differences in microstructure on measured hydrogen contents is elaborated further in section 4.5.2.

a) Figure 4.25

b)

(a) Hydrogen concentration in as-received alloys. (b) Hydrogen concentration for each plate material as a function of prestrain.

64

Diffusible hydrogen can move through the metal matrix and encounter microstructural constituents with high binding energies (> 60 kJ/mole) and become irreversibly trapped [53]. Values of binding energies for some typical hydrogen traps in steel are listed in Table 4.6. Depending on the type of trap, HIC susceptibility has been shown to increase or decrease [12 and 54]. Dislocations and grain boundaries are believed to exert a transient trapping effect on hydrogen as it moves through the lattice, which was observed by Bouraoui et al. [55]. However, features such as inclusion and precipitate interfaces serve as irreversible trap sites.

The average values obtained from the LECOÂŽ trapped hydrogen analysis for the materials and prestrain conditions are shown in Figure 4.26. The average values represent the average of 5 values. Error bars represent a 90 pct confidence interval that was calculated from the standard deviation of the 5 independent trapped hydrogen values. Other than X52 0 pct and 12 pct prestrain conditions, values of trapped hydrogen were 1 ppm or greater (Figure 4.26). The X52 trapped hydrogen concentration increases with prestrain. A statistically significant monotonic trend was not observed with respect to prestrain for the X60, X70, and 100XFalloys. Overall, the trapped hydrogen in each condition is comparable, indicating that there are no significant differences in irreversible trap sites between the alloys. The trapped hydrogen content of the alloys is likely mostly independent of pre-strain level, because dislocations are reversible trap sites and thus, increasing dislocation density through pre-straining should have minimal effect.

Figure 4.26

Trapped hydrogen values after degassing of material. Materials were hydrogen charged with the EC method, degassed at 50 °C for 72 hours, sectioned, and evaluated using the LECOŽ hydrogen analyzer to determine the trapped hydrogen content.

65

Table 4.6 Hydrogen trap sites found in iron, table modified from [54] Trapping site H - dislocation H - grain boundary H - MnS H - Fe3C H - TiC interface

4.5.2

Binding/activation energy (kJ/mol) 26 49 72 84 95

Assessment Method* HTD analysis HTD analysis HTD analysis Permeation Permeation

Reference [56] [57] [58] [59] [60]

Average Diffusible Hydrogen Content from LECO® and Mercury Displacement Analysis The diffusible hydrogen concentration was measured using the mercury displacement method on full size

samples (outlined in section 3.8.1) that were subjected to the EC method. Diffusible hydrogen was determined by averaging values from each of the four samples, and a 90 pct confidence interval was calculated based on the standard deviation. The values for diffusible hydrogen determined from the mercury displacement method are shown in Figure 4.27a. The measured values of diffusible hydrogen for each material and prestrain condition showed high amounts of deviation. The greatest amount of diffusible hydrogen was observed in the 100XF material and prestrain condition (~ 7 ppm) and the least in the X52 material and prestrain conditions (~ 2.5 ppm) (Figure 4.27a). The amount of deviation was attributed to non-uniform charging over the sample geometry, which is discussed in the following section.

Figure 4.27b shows the values of diffusible hydrogen calculated (Equation 3.6) from the two values, HTotal and HTrapped, obtained by the LECO (Figure 4.25b and Figure 4.26). Error bars on the plot represent a 90 pct confidence interval calculated on the standard deviation of the 5 values. The levels of diffusible hydrogen determined by the LECO® method are, depending on the material, 1 to 3 ppm less than those measured with the mercury displacement method. This is likely due to the sub-sized specimens. The sub-sized specimens are believed to have more uniform charging during EC due to the smaller size, represented by the tighter confidence interval (Figure 4.27b). However, the total hydrogen concentration values used in the calculation could be reduced during the lag time between placing each sample into the analyzer and subsequent melting and analysis by the LECO® analyzer (~ 90 seconds). The time where the sub-sized specimens are at room temperature allows for a degree of hydrogen diffusion out of the sample. This reduction in measured total hydrogen level could also reduce the estimated diffusible hydrogen concentration.

In both methods, the diffusible hydrogen concentration is greater than 1 ppm for all conditions. Kittel et al.

[61] suggested that a critical diffusible hydrogen concentration level of 1 ppm is necessary to induce damage within a steel matrix. These levels indicate that hydrogen was present in sufficient amounts to initiate HIC in all of the conditions, Also, the average values of diffusible hydrogen increase with increasing values of prestrain, except for the X60 material. This trend agrees with literature [62 and 63] that at higher degrees of strains (i.e. increases in dislocation density) the diffusible hydrogen is increased. Bouraoui et al. [55] stated that dislocations exert a transient trapping effect on hydrogen atoms, decreasing the hydrogen permeability though high purity iron, and a similar effect is likely observed in the present study. The differences in the amount of diffusible hydrogen in each alloy could be related to the grain size and morphology present in the materials. The X52 is dominated by larger

66

polygonal ferrite grains, whereas the other alloys are dominated by finer acicular ferrite grains and probably a higher dislocation density. Of the acicular ferrite microstructures, the amount of second phase microconstituents is highest in the 100XF alloy. It is hypothesized that the finer grain size (larger grain boundary surface area), higher dislocation density, and higher surface area of boundaries at secondary microconstituents all serve to increase the transient trapping sites and result in the 100XF having the highest diffusible hydrogen content.

a) Figure 4.27

4.5.3

b)

Diffusible hydrogen values determined by two different evaluation methods on the four materials and prestrain conditions investigated: (a) Mercury displacement and (b) LECO® analysis.

LECO® diffusible hydrogen values were determined using Equation 4.2.

Influence of Sample Location on Diffusible Hydrogen Results (Sample Size Effect) During the EC tests, it was observed that gas bubbles nucleate, grow, and detach preferentially at the

bottom end of the test specimen (working electrode). This characteristic is displayed in the photograph shown in Figure 4.28. The evolution of bubbles was predominately within the first 25 mm of the specimen being charged (sample #1 area shown in Figure 4.28). The effects of preferential bubble formation were assessed based on differences in diffusible hydrogen content as a function of sample number (measured by mercury displacement) and CLR value as a function of location of the examined face.

Diffusible hydrogen measurements for samples 1 – 4 from each material and prestrain condition were compiled to evaluate possible non-uniform charging conditions. Figure 4.29 shows the dependence of diffusible hydrogen measured by the mercury displacement method, with respect to the sample number referenced to the location of the electrode connection (shown in Figure 4.28). The diffusible hydrogen amount in sample #1 for all materials is greater than samples #3 and #4. The amount of diffusible hydrogen decreases from sample 1 to sample

67

4, which is evidence of the non-uniform charging. The non-uniform charging of these samples is believed to be due to a slight potential gradient on the sample during charging, causing the evolution of hydrogen to be greater at the end of the sample. Because of the large difference in diffusible hydrogen content from sample #1 to #4, the standard deviation between the four values was also large, producing the wide confidence interval in the diffusible hydrogen results shown in Figure 4.27a. The sub-sized specimens used for the LECO analysis are believed to have more uniform charging conditions, which is indicated by the tighter confidence interval in the diffusible hydrogen estimates shown in Figure 4.27b.

The CLR values from faces 1, 2, 4, and 6 of full-sized EC specimens were averaged together for each face to produce a single CLR value at the specific sectioning location; this was done for each alloy and pre-strain condition. The CLR dependence on the examined face location is presented in Figure 4.30. Faces 1 and 2 are within areas of preferential bubble-formation and face 6 is farthest from it. The X70 (Figure 4.30c) and 100XF (Figure 4.30d) samples exhibit higher CLR values near the preferential hydrogen bubble formation region. The decrease in CLR value towards the top of the electrode, i.e. from face 1 and 2 to the other faces (face 4 and 6), can be interpreted as an indication that the presence of these bubbles at the bottom of the specimen corresponds to higher amounts of hydrogen damage in these regions. The measured crack ratios do not depend as greatly on location in the X52 (Figure 4.30a) and X60 (Figure 4.30b) samples, which may be a reflection of the large variance in CLR associated with, and intrinsic to, the evaluation method employed.

Figure 4.28

Observation of hydrogen bubbles forming preferentially on the lower bottom end of the steel specimens during EC experiments. Sample numbers are shown on the steel sample in the image.

68

Figure 4.29

Dependence of location of sample (sample number) on the amount of diffusible hydrogen measured by the mercury displacement method. Sample 1 represents the sample furthest away from the electrode connection. Dimensions of the sample were the full thickness of the plate (100 XF and X70 - 12.7, X60 - 9.5, and X52 - 19 mm) x 20 Âą 3 mm (width) x 25 Âą 2 mm (length).

69

Figure 4.30

4.5.4

a)

b)

c)

d)

Dependence of calculated Crack Length Ratio on the location of the examined faces for each material and prestrain condition for EC experiments (a) X52, (b) X60, (c) X70, and (d) 100XF.

HIC Susceptibility and Diffusible and Trapped Hydrogen Contents from LECOÂŽ Analysis Figure 4.31 shows the dependence of CLR as function of (a) the diffusible hydrogen content determined by

the LECOÂŽ analysis method and (b) the amount of trapped hydrogen after degassing. HIC appears to be more dependent on the diffusible hydrogen concentration than the total trapped hydrogen concentration (Figure 4.31a).

70

Two key observations can be made from Figure 4.31a. The first observation is that for the highest strength material, 100XF, both CLR and diffusible hydrogen were the greatest. The CLR is believed to be associated with the higher strength and hardness of the material leading to higher amounts of cracking produced through the interaction with hydrogen; whereas, the higher diffusible hydrogen value is believed to be associated with increased surface area of grain boundaries and dislocation density to increase the overall transient trapping and therefore increasing the diffusible hydrogen. The second observation is that even with increases in the amount of diffusible hydrogen in the X60 and X70 (acicular ferrite steels) compared to the X52 alloy, the degree of cracking in the 3 alloys is approximately equal. It would be expected from interpretation of literature [61] that at higher amounts of diffusible hydrogen, HIC would increase. This trend is not observed and thus it can be inferred that for moderate and low strength steels, changes in microstructure will not greatly impact HIC susceptibility. Figure 4.31b shows that HIC is independent of the trapped hydrogen amount in the conditions assessed. At similar levels of trapped hydrogen, the HIC susceptibility of the 100XF is much greater than the other alloys.

a) Figure 4.31

4.6

b)

HIC (CLR) dependence on (a) LECOÂŽ diffusible and (b) trapped hydrogen contents from EC. The two horizontal dashed lines represent sour service requirements outlined by ISO 3183, CLR < 0.15 [52].

Evaluation of HIC with respect to Microstructure and Nonmetallic Inclusions Figure 4.32 shows representative images of the cracking behavior in each of the four alloys. Centerline

cracking was observed for some of the sectioned faces for each material. Centerline cracking was observed more predominantly in the X52 (Figure 4.32a), X60 (Figure 4.32b), and X70 (Figure 4.32c) conditions, whereas the 100XF material produced cracking near the rolled face ( average depth of 1.65 mm) for all prestrain conditions (Figure 4.32d), consistent with the results obtained during the characterization study. The location of cracking within the 100XF alloy is believed to be due to higher levels of hardness, compared to other areas within the plate thickness. The large range of crack depths, shown in Table 4.7, observed for the X52, X60, and X70 implies that

71

crack nucleation sites i.e. nonmetallic inclusions, are distributed throughout the plate thickness, as presented by the non-etched micrographs shown in Figures 4.11 â&#x20AC;&#x201C; 4.13. The average crack depth shown in Table 4.7, represents the average of all cracks observed for the material and prestrain conditions evaluated. Each material had cracks at the centerline of the plate, as well as cracks near the rolled surface.

Table 4.7 Crack Depth Observed on all Sectioned Faces for each alloy after EC Material

Half Thickness (mm)

Average Depth of Cracking (mm)

Standard Deviation (mm)

X52 X60 X70 100XF

9.50 4.75 6.50 6.50

8.02 3.02 4.09 1.65

2.03 1.27 1.77 1.45

b)

a)

d)

c) Figure 4.32

Extremes Max Depth Min Depth (mm) (mm) 9.19 2.05 4.54 0.59 6.23 0.54 6.38 0.27

Light optical macrographs taken on as-polished full width and thickness EC test specimens.

Images are representative of the cracking behavior observed for materials at all prestrain levels. (a) X52 0 pct prestrain Face 1, (b) X60 3 pct prestrain Face 2, (c) X70 5 pct prestrain Face 1, and (d) 100XF 0 pct prestrain Face 2.

72

ESEM, FESEM, and EDS mapping were used to investigate the interactions of HIC with microstructural features such as parent microstructure, secondary microconstituents/phases, and nonmetallic inclusions. The results presented in the following section were obtained from 5 faces that showed the highest amount of cracking, i.e. greatest CLR and CTR for each material. Primary cracking is denoted as the largest cracks; whereas, secondary cracking is the qualitative observation of cracking that branches from the largest cracks.

Sectioned faces with the highest degree of cracking in X52 were evaluated after etching as shown in Figure 4.33. The cracking behavior had evidence of transgranular propagation though both the ferrite and pearlite (Figure 4.33 features 1 and 3), as well as intergranular cracking of ferrite (Figure 4.33 feature 2).

Samples were back polished to 1 Âľm surface finish after being etched to conduct EDS on non-etched samples. Cracks on the polished surface are shown in Figure 4.34. Cracking was not observed around inclusions on the plane of view for the faces evaluated for HIC in the X52; this does not rule out the possibility of inclusions being present in the cracks out of the plane of view that was examined. If inclusions were observed, they were outside the primary crack area (Figure 4.34). The composition of the inclusions found within close proximity of the primary crack, identified though EDS mapping, consistently contained Al, Ca, and O. These are likely alumina inclusions encapsulated by a calcium sulfide shell as shown in Figure 2.5.

Figure 4.33

Secondary electron micrograph taken on the FESEM of the X52 EC 18 % prestrain Face 1 condition. Etched with 2 pct nital. Evidence of transgranular (1and 3) and intergranular crack propagation (2).

73

Figure 4.34

EDS mapping of non-etched X52 EC 18 % prestrain Face 1condition. Primary crack propagation in close proximity to Al-O-Ca type inclusion. SEM image produced was taken in backscatter mode on the FESEM (see pdf version for color).

The X60 and X70 steels were first etched with 2 pct nital to reveal the microstructure and then placed into the SEM to analyze the crack interaction with the microstructure. Figure 4.35 and Figure 4.36 show secondary electron images on the etched surfaces around the primary crack region for the X60 and X70, respectively. Images taken in secondary electron mode on the FESEM revealed that in the X60 (Figure 4.35) and X70 (Figure 4.36), HIC was associated with inclusions. It was difficult to determine whether the crack growth was transgranular or intergranular in nature, but cracking associated with secondary microconstituents was occasionally observed.

Figure 4.35 shows evidence of the primary crack propagating directly through two nonmetallic inclusions (feature #1), and cracking that originates around the interface of an elongated nonmetallic inclusion (feature #2) in the X60.

Figure 4.36 also shows an example of HIC that propagates directly though a globular type nonmetallic inclusion in X70.

After the microstructure was evaluated, samples were polished to a 1 Âľm to specifically examine how inclusions interact with HIC. Using backscatter mode on the SEM (gives composition contrast), areas where nonmetallic inclusions were present around the primary crack region were identified. Once identified, the areas were evaluated by EDS mapping to identify the type of inclusion present.

Figures 4.37 and 4.38 show EDS mapping of primary crack regions where nonmetallic inclusions are present. In both SEM images, there is a cluster of nonmetallic inclusions present in and around the primary crack area. EDS showed these clusters as being high in sulfur and manganese. MnS type inclusions were found by both EDS mapping and the AFA analysis presented in section 4.1. It is well documented in literature that these types of inclusions are detrimental to HIC resistance [3, 5, 8, and 18]. This behavior is confirmed and observed for the X60 and X70.

74

Figure 4.35

Secondary electron micrographs taken on the X60 EC 0 % prestrain Face 6 condition. Image captured using the ESEM. Etched with 2 pct nital. Features 1 and 2 show HIC around inclusions in the microstructure.

Figure 4.36

Secondary electron micrograph taken on the X70 EC 5 % prestrain Face 1 condition. Image shows the presence of an inclusion in the primary crack area. Image was captured using the ESEM. Etched with 2 pct nital.

75

Figure 4.37

EDS mapping of non-etched X60 EC 0 % prestrain Face 6 condition. SEM image was taken in backscatter mode on the FESEM (see pdf version for color). The SEM image and EDS maps show the presence of MnS type inclusions in and distributed around the primary crack region.

Figure 4.38

EDS mapping of non-etched X70 EC 5 % prestrain Face 1 condition. SEM image was taken in backscatter mode on the FESEM (see pdf version for color). The SEM image and EDS maps show the presence of MnS type inclusions in and distributed around the primary crack region.

Figure 4.39 shows secondary electron images from the 100XF alloy etched with 2 pct nital; the images were taken using the FESEM. Cracking behavior in the 100XF showed: transgranular cracking of acicular ferrite grains (Figure 4.39a feature #1), cracking of secondary microconstituents (Figure 4.39b feature #2), and cracking around nonmetallic inclusions (Figure 4.39c feature #3). At areas where nonmetallic inclusions were present, the degree of secondary cracking decreases. In the absence of nonmetallic inclusions, smaller micro-cracks branch off from the main crack increasing the degree of secondary cracking (Figure 4.39a and b). It is speculated that the areas in the immediate vicinity of inclusions contains a high concentration of hydrogen, so cracking is localized in these regions. Hydrogen is more uniformly distributed in regions without inclusions and therefore secondary cracking can occur more easily.

76

Cracking behavior associated with inclusions in the 100XF is further shown in Figures 4.40 and 4.41. EDS mapping was conducted on both conditions to assess the type of inclusion present. Figure 4.40 exhibits a primary crack that propagates near but not through a spherical nonmetallic inclusion (Al-O-Ca-S). Conversely, in Figure 4.41, the primary crack interacts directly with the nonmetallic inclusion (Mg-Al-O). It was observed that when the inclusions are more spherical in shape and are of the composition like the inclusion present in Figure 4.40 (Al-O-CaS), cracks do not interact with the inclusion. The slight elongation of the inclusion in Figure 4.41is likely more detrimental for HIC, since there is a larger surface area in one plane for hydrogen accumulation and crack nucleation. The observed effects of inclusion shape as well as type of inclusion correlates with literature findings [3, 18, 21, and 31]. Introducing CaS type inclusions or having a higher degree of spherical shaped nonmetallic inclusions from calcium treatment increases the overall HIC resistance with respect to inclusion susceptibility.

a) Figure 4.39

b)

c)

Secondary electron images take on a) 100XF EC 0 % prestrain Face 2 (feature #1 shows transgranular cracking of acicular ferrite), b) 100XF EC 0 % prestrain Face 2 (feature #2 shows crack propagation across secondary microconstituent), and c) 100XF EC 2 % prestrain Face 1 (feature #3 shows nonmetallic inclusion). Image was captured using the FESEM. Etched with 2 pct nital.

The interaction with microstructural features for each alloy is summarized in Table 4.8. The X60 and X70 interacted with hydrogen similarly. HIC appeared to be related to the presence of nonmetallic inclusions (MnS) in these alloys. The X52 and 100XF were the only alloys that showed signs of both intergranular/transgranular cracking within the fields of view observed, as well as consistently observed cracking around secondary microconstituents (pearlite in the X52 and martensite in the 100 XF). The X52 was the only material that did not show significant evidence of cracking around nonmetallic inclusions, which is believed to be connected to the low sulfur content achieved during steelmaking, lower manganese level, and effective shape control over nonmetallic inclusions due to Ca additions. The X60 and X70 had higher amounts of manganese and sulfur (1.03 and 0.0007 wt pct respectively), allowing for the formation of MnS, and leading to preferential cracking around those features. Further characterization and analysis is needed on the X60 and X70 alloys to evaluate for the intergranular and transgranular cracking behavior. The 100XF had signs of multiple types of cracking: intergranular and transgranular cracking through ferrite grains, cracking around secondary microconstituents (martensite), secondary cracking, and cracking along nonmetallic inclusions (oxides).

77

Figure 4.40

EDS mapping of non-etched 100XF EC 0 % prestrain Face 2 condition. SEM image was taken in backscatter mode on the FESEM (see pdf version for color). SEM image reveals an globular shaped inclusion found outside the primary crack region. EDS mapping identifies the inclusion being of the mixed composition of Al-O-Ca-S.

Figure 4.41

EDS mapping of etched 100XF EC 2 % prestrain Face 1 condition. SEM image was taken in backscatter mode on the FESEM. Etched with 2 pct nital (see pdf version for color). SEM image reveals an elongated inclusion found within the primary crack region. EDS mapping identifies the inclusion being of the mixed composition of Al-Mg-O.

Table 4.8 Summarized Cracking Behavior Ferrite Material

Intergranular

Transgranular

Secondary Microconstituents

Secondary Cracking

Nonmetallic Inclusions

X52

X

X

X (pearlite)

-

-.

X60

-

-

X

-

X (sulfides)

X70

-

-

X

-

X (sulfides)

100XF

X

X

X (martensite)

X

X (oxides)

X - Presence of behavior â&#x20AC;&#x153;-â&#x20AC;&#x153; denotes that further analysis is needed to determine presence of behavior

78

CHAPTER 5: Summary and Conclusions

Conclusions for the study of hydrogen induced damage in plate steels are discussed below in reference to the research objectives outlined in Chapter 1.

5.1

Electrolytic Charging A new dual cell electrolytic charging cell was designed for HIC experiments. The design addressed issues

discovered in the previously used monocell design. The important features of the new electrolytic cell are the salt bridge to avoid cross contamination of the ion species from the cathode and anode, a gas purge system to help agitate the solutions, removal of oxide scale on steel specimens to inhibit oxide particles from contaminating the solution, and an improved electrode connection. The new design was successfully used for stable long duration (e.g.

24 hours) experiments.

A characterization study on the fabricated dual cell using the 100XF alloy studied the influences of applied current density and test duration to on HIC damage. It was found that at an applied current density 15 mA/cm 2 and a test duration 24 hours provided consistent HIC results.

These testing parameters were then used for an investigation of HIC on non-strained and prestrained plate steels (X52, X60, X70, and 100XF). The results were compared to NACE Standard TM0284 tests (solution A) on the same conditions. The critical crack ratios calculated from the EC method were similar in magnitude to the critical crack ratio values from the NACE test. Thus, the EC method is possibly a viable substitute for the NACE standard test to investigate HIC. However, further experimental refinement is necessary. The diffusible hydrogen values in the lower section of the specimen (furthest section from electrode connection) were significantly larger than the rest of the specimen; there was approximately a 6 ppm difference in the 100XF and 3 ppm difference in the X52 alloy. The cause of the difference in hydrogen content is believed to be associated with non-uniform hydrogen charging over the specimen geometry.

5.2

HIC Susceptibility Measurements

5.2.1

Influence of Mechanical Properties The lower strength steels, X52, X60, and X70, always have HIC susceptibility less than the ISO standard

guidelines for qualification of sour service plates: values of CLR < 0.15 and CTR < 0.05. The 100XF, regardless of prestrain level, had the highest values of crack parameters: CLR = 0.58 ± 0.06, CTR = 0.07 ± 0.01, CSR = 0.03 ± 0.005. The high susceptibility was attributed to higher tensile strength, hardness, and lower ductility, leading to high amounts of HIC. These observations are consistent with NACE recommendations that HIC resistant alloys have tensile strength and hardness less than 116 ksi (800 MPa) and 248 HV (22 HRC).

79

5.2.2

Influence of Uniaxial Prestrain The HIC results are largely independent of the pre-strain levels imposed within the statistical accuracy of

the evaluation methods employed. HIC resistance was largely independent of the effects of prestrain including increases in hardness, dislocation density, and reductions in ductility. Further, prestrain did not increase the level of HIC in the X52, X60, and X70 alloys to a point where it would be disqualified for sour service according to the ISO standard guidelines.

5.3

Hydrogen Analysis The total amount of hydrogen, which includes both the diffusible and trapped hydrogen, was greatest in the

100XF alloy conditions and least in the X52 alloy conditions. The level of trapped hydrogen in the non-strained conditions did not vary greatly: X52 â&#x20AC;&#x201C; 0.95, X60 1.07, X70 â&#x20AC;&#x201C; 1.25, and 100XF - 1.18 ppm. This observation implies that the number of irreversible trap sites in the alloys is comparable. The differences in total hydrogen content are attributed to differences in the amount of diffusible hydrogen in the alloys. The average values for diffusible hydrogen are greatest in the 100XF ~ 7 ppm and lowest in the X52 ~ 2.5 ppm (values from mercury displacement).

The large amount of diffusible hydrogen in the 100XF alloy compared to the X52 alloy is likely due to the relatively high grain boundary area and dislocation density.

5.4

Evaluation of HIC with respect to Microstructure and Nonmetallic Inclusions The qualitative analysis of microstructure and nonmetallic inclusions produced results that confirmed

findings from literature. Cracking is observed around nonmetallic inclusions such as sulfides and oxides in the metal matrix. For materials in which both types are present, X60 and X70, HIC is linked to areas of sulfide type inclusions. Elongated sulfide inclusions show the propensity to increase the degree of cracking.

80

CHAPTER 6: Future Work

This section outlines proposed areas of future work to further optimize the EC test apparatus and to further evaluate HIC in plate steels.

It was observed that the overall size of the sample used in EC influenced the uniformity of charging that occurred during testing. It is proposed to test the effects of size/volume of EC specimens. The goal would be to optimize EC specimens for hydrogen charging uniformity.

Because this study focused on the effects of uniaxial prestrain on HIC susceptibility, a more comprehensive study involving different levels of residual stresses/strains to alter HIC behavior is suggested. This study should be applicable to actual industry practices used for plate forming operations that may alter plate steel resistance to HIC though the formation of residual stresses/strains.

Conducting a study focusing on the microstructure influence on HIC susceptibility would be of interest.

Targeted microstructures could be created by varying heat treatments, cooling techniques, and subsequent processing operations. The targeted strengths of the produced microstructures would have to be similar to be able to ignore the effects of increased strength to alter HIC susceptibility.

The current evaluation method for HIC used in this study produced large variations in the cracking parameters calculated. The large variation made it difficult to discern differences between alloy and prestrain conditions. Similar concerns have been stated in the literature about the NACE standard H 2S test. The sectioning procedure may miss cracks that could significantly change the test results. It is suggested that a study involving different evaluation methods be considered to increase the reliability of the results from HIC. The evaluation methods of interest could be Ultra-Sonic evaluation or 3-D X-Ray Tomography. These methods incorporate taking scans over the full size of the sample to identify defects such as internal cracks. It is thought that this type of evaluation may reduce the statistical error of the crack measurements.

Inclusions in the low strength steels, X60 and X70, provided the weakest link within the microstructure with respect to HIC. It was suggested by industry members that a more in depth inclusion analysis would be worth pursuing. This study would incorporate quantifying relative inclusion content found in the steel, size distribution in the steel, morphology produced, and compositional mapping. Quantifying these variables would give rise to a rating scale of HIC susceptibility based on the features of the inclusion present within the steel.

81

References Cited

1. J. Nieto, T. Elias, G. Lopez, G. Campos, F. Lopez, R. Garcia and A.K. De, “Development of Technology for the Production of HIC Resistant Slabs for Sour Service Applications at ArcelorMittal Lazaro Cardenas, Mexico,” Materials Science and Technology (MS&T), 2012.

2. T. Taira, Y. Kobayashi, H. Inagaki, and T, Watanabe, “Sulfide Corrosion Cracking of Line-Plate or Sour Gas Service,” NACE National Conference, unpublished paper No. 171. Atlanta, 1979.

3. G. Biefer, “The Stepwise Cracking of Line-Plate Steels in Sour Envionments,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp. 479 - 494.

NACE International, 1996.

4. G. Herbslab, R. Popperling, and W. Schwenk, “Occurrence and Prevention of Hydrogen-Induced Stepwise Cracking and Stress Corrosion Cracking of Low Alloy Pipeline Steels,” NACE National Conference, unpublished paper No. 9. Chicago, 1980.

5. Y. Shinohara and T. Hara, “Metallurgical Design of UOE Line Plate for Sour Service,” Microalloyed Steels for Sour Service International Seminar, São Paulo, Brazil, August 2012.

6. J. A. Ronevich, “Hydrogen Embrittlement in Advanced High Strength Steels,” M.S. Thesis, Metallurgical and Materials Engineering, Colorado School of Mines, Golden, CO, 2009.

7. B. Rosner, “Exploration of a Hydrogen Charging Method to Understand Performance of Energy Tubular Products,” M.S. Thesis, Metallurgical and Materials Engineering, Colorado School of Mines, Golden, CO, 2011.

8. D. Jones, Principals And Prevention Of Corrosion, 2nd ed., New Jersey: Prentice-Hall, Inc, pp. 5-18, 40-50, 75-96, 143-148, 249-254, 279-286, 335-342, 439-444. 1996.

9. R. Kane, S. Wilhelm, and J. Oldfield, “Review of Hydrogen Induced Cracking of Steels in Wet H2S Refinery Service,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp. 173 - 185. NACE International, 1996.

10. M. Cayard and R. Kane, “Characterization and Monitoring of Cracking of Steel Equipment in Wet H2S Service,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp. 633 - 651. NACE International, 1996.

11. G. Krauss, Steels: Processing, Structure, And Performance, 4th ed., Ohio: ASM International, pp. 101-116. 2005.

12. E. Moore and J. Warga, Materials Performance, vol. 15, No. 6, pp. 176. 1976.

13. A. Ikeda, T. Kaneko, and F. Terasaki, “Influence of Environmental Conditions and Metallurgical Factors on Hydrogen Induced Cracking of Line Plate Steel,” NACE National Conference, unpublished paper No. 7. Chicago, 1980.

14. S.U. Koh, J.S. Kim, B.Y. Yang, and K.Y. Kim. “Effect of Line Plate Steel Microstructure on Susceptibility to Sulfide Stress Cracking,” Corrosion Science, vol. 60, No. 3. 2004.

82

15. G.T. Park, S.U. Koh, H.G. Jung, and K.Y. Kim, “Effect of microstructure on the hydrogen trapping efficiency and hydrogen induced cracking of linepipe steel,” Corrosion Science, vol. 50, pp. 1865-1871. 2008.

16. H.K.D.H. Bhadeshia, “Acicular Ferrite,” In Bainite in Steels (Transformation, Microstructure and Properties), The Institute of Materials, London, Great Britain, Cambridge University Press, pp. 245-248. 1992.

17. Z. Shabir, E. Van Der Giessen, C. A. Duarte, and A. Simone, “The role of cohesive properties on intergranular crack propagation in brittle polycrystals,”In Modelling and Simulation in Materials Science and Engineering, vol. 19, IOP Publishing Ltd. 2011.

18. T. Michler and J. Naumann. “Microstructural aspects upon hydrogen environment embrittlement of various bcc steels,” International Journal of Hydrogen Energy 65, pp. 821 – 832. 2010.

19. H. Lee and S.L. Chan, “Hydrogen embrittlement of AISI 4130 steel with an alternate ferrite/pearlite banded structure,” Materials Science and Engineering, A142, pp. 193-201. 1991.

20. E. Miyoshi, T. Tanaka, F. Terasaki, and A. Ikeda, “Hydrogen-Induced Cracking of Steels Under Wet hydrogen Sulfide Environment,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp. 461 – 470. NACE International, 1996.

21. E. Moore, J. Warga, “Factors Influencing the Hydrogen Cracking Sensitivity of Pipeline Steels,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp. 725 - 744. NACE International, 1996.

22. T. Nishimura, H. Inagaki, and M. Tanimura, “Hydrogen Cracking in Sour Gas Pipeline Steels,” Second Intl. Cong. On Hydrogen in Metals, Paper Number 3E9, pp. 1 – 8, Paris, June 1977.

23. H. Nakasugi, S. Sugimura, H. Matsuda, and T. Murata, “Development of New Line-Plate Steels for Sour Gas Service,” Nippon steel Technical Report, vol. 14, No. 12. pp 66, 1979.

24. Email from D. Edelman, UCS Development Metallurgist, Nucor, Crawfordsville, Indiana, Dec. 19, 2013.

25. A. Ghosh, Secondary Steelmaking: Principals and Applications. Boca Raton, FL, USA: CEC Press, 1999.

26. R. Kane and M. Cayard, “Evaluation of Advanced Plate Steels For Resistance to HIC and SOHIC in Wet H2S Environments,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp 615 – 631. NACE International, 1996.

27. T. Taira, Y. Kokayashi, K. Matsumoto, and T. Tsukada, “Resistance of Line Pipe Steels to Wet Sour Gas,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp 509 -518. NACE International, 1996.

28. R. Merrick and M. Bullen, “Prevention of Cracking in Wet H2S Environments,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp. 149 – 172.

NACE International, 1996.

83

29. R. Kane and D. Abayarathna, “Characterization of Hydrogen Charging Severity in Simulated Wet H2S Refinery Environments,” In R. Kane, R. Horvath, and M. Cayard, Wet H2S Cracking of Carbon Steels and Weldments, pp 603 – 613. NACE International, 1996.

30. A. Ikeda, T. Kaneko, and F. Teraski, “Influence of Environmental conditions and Metallurgical Factors on Hydrogen Induced Cracking of Line Plate Steel,” NACE National Conference, unpublished paper No. 8. Chicago, 1980.

31. G. Domizzi, G. Amteri, and J. Ovejero-Garcia, “Influence of sulfur content and inclusion distribution on the hydrogen induced blister cracking in pressure vessel and pipeline steels,” Corrosion Science, vol. 43, pp. 325-339. 2001.

32. C Parrini and A. DeVito, “High Strength Microalloyed Plate Steels Resistant to Hydrogen Induced Failures,” Proceedings of. MICON ’78, Houston , TX, April1978, 33. A. Ikeda, Y. Morita, F. Terasaki, and M. Takeyama, “On the Hydrogen Induced Cracking of Line-Plate Steel Under Wet Hydrogen Sulfide Environment,” Second Intl. Cong. On Hydrogen in Metals, Paper No. 4A7, pp. 1 – 8. Paris, June 1977.

34. Evaluation of Pipeline and Pressure Vessel Steels for Resistance to Hydrogen-Induced Cracking, NACE Standard TM0284-2003, 21215, NACE International, 2003.

35. B. Berkowitz, J. Burton, C. Helms, and R. Polizzotti, “Hydrogen dissociation poisons and hydrogen embrittlement,” Scripta Metallurgica, vol. 10, pp. 481 – 484, 1996 36. I. Bernstein and A. Thompson, “The Use of Electrical Permeation Technique to Study Hydrogen Transport and Trapping of Iron,” Presented at the Conference Proceedings of the Metallurgical Society of AIME, 1982.

37. L. Céline, “Cathodic Hydrogenation of Very High Strength Steels - Influence of Arsenic as a Hydrogenation Promoter”, Effects of Hydrogen on Materials, Proceedings of the 2008 International Hydrogen Conference, ASM International, pp. 300-307. 2009.

38. S. M. Charca, O.N.C. Uwakweh, B. Shafiq, and V.S. Agarwala, “Characterization of Hydrogen Permeation in Armco-Fe during Cathodic Polarization in Aqueous Electrolytic Media,” ASM International, pp. 127-133. 2008.

39. S.Y. Qian, B.E. Conway, and G. Jerkiewicz, “Kinetic Rationalization of Catalyst Poison Effects on Cathodic H Sorption into Metals: Relation of Enhancement and Inhabitation to H Coverage,” Journal of the Chemical Society, Faraday Transactions, vol. 94, pp. 2945-2954.1998.

40. S. M. Charca, O.N.C. Uwakweh, and V.S. Agarwala, “Hydrogen Transport Conditions and Effects in Catholically Polarized AF1410 Steel,” The Minerals, Metals & Materials Society and ASM International, 2007.

41. D.M. Perez Escobar, L. Duprez, K. Verbeken, C. Miñambres, M. Verhaege, “Identification of the damage in high strength steels after electrochemical hydrogen charging,” Metallurgists 49th Conference Proceedings, 2010.

84

42. B.G. Farber, “Influence of Prestrain on the Impact and Fracture Toughness Properties of High Strength Plate Steels with regard to Various Microstructures,” M.S. Thesis, Metallurgical and Materials Engineering, Colorado School of Mines, Golden, CO, 2009.

43. H. M. Al-Jabr, “Anisotropy of Mechanical Properties of API-X70 Spiral Welded Plate Steels,” M.S. Thesis, Metallurgical and Materials Engineering, Colorado School of Mines, Golden, CO, 2011.

44. G. Angus and K. Findley, “Investigation of Hydrogen-Induced Damage in Pipeline Steels,” Research Report, MT-SRC-012-003, Plate and Hot Rolled Steels, ASPPRC, Golden, CO, 2012.

45. Standard Methods for the Determination of Diffusible Hydrogen Content of Martensitic, Bainitic, and Ferritic Steel Weld Metal Produced by Arc Welding, ANSI/AWS A4.3-86, American Welding Society Inc., December 1985.

46. J. Kang,H. Lee, and S.H. Han, “Effect of Al and Mo on the textures and microstructures of dual phase steels,” Materials Science and Engineering: A, vol. 530, pp. 183 – 190. Dec. 2011.

47. I. M. Bernstein, Materials Science and Engineering, vol. 6, pp. 1-19, 1970.

48. J.P. Hirth, “Effects of hydrogen on the properties of iron and steel,” Metallurgical and Materials Transactions A, vol. 129, pp. 99-107, 1990.

49. R. E. Reed-Hill and R. Abbaschian, Physical Metallurgy Principals, 3rd ed., Massachusetts: PWS Publishing Company, pp. 382. 1994.

50. A. Griesche, E. Dabah, A. Hilger, N. Kardjilov, I. Manke, and T. Kannengiesser, “Neutron Imaging of Hydrogen in Steels,” Materials Science and Technology (MS&T) 2013, Montreal, Quebec, Canada. Oct. 27-37, 2013.

51. H. Huang and W.J.D Shaw, “Hydrogen Embrittlement Interactions in Cold-Worked Steel,” Corrosion Science, pp. 30-36. January 1995.

52. Petroleum and natural gas industries-Steel plate for pipeline transportation systems, ISO 3183, 2nd edition, 2007, Appendix H. 53. M.D. Herynk, S. Kyriakides & A.Onoufriou, "Effects of the UOE/UOC plate manufacturing processes on pipe collapse pressure," Int. J. Mechanical Sciences, vol. 49, pp. 533 – 553, Dec. 2006.

54. I. Maroef, D.L. Olson, M. Eberhart, and G. R. Edwards, “Hydrogen Trapping in Ferritic Steel Weld Metal,” International Materials Reviews, Vol. 47, No. 4, pp. 191-223. 2002.

55. R. Bouraoui, Hydrogen Diffusion in Metals: A 30-Year Retrospective, Scitec Publications, pp. 65.

1999.

56. W. Y. Choo, J. Y. Lee, “Thermal Analysis of Trapped Hydrogen in Pure Iron,” Metallurgical Transactions A, vol. 13A, pp. 135-140. 1982.

85

57. K. Ono, M. Meshi, “Hydrogen Detrapping from Grain Boundaries and Dislocations in High Purity Iron,” Acta Metallurgica et Materialia, vol. 40, pp. 1357-1364. 1992.

58. J. L. Lee and J. Y. Lee, “Hydrogen Trapping in AISI 4340 Steel,” Metal Science, vol. 17, pp.

426-432. 1983.

59. T. Asaoka, C. Dagbert, M. Aucouturier, and J. Galland, “Quantitative Study of Trapping Characteristics of H in a Fe-0.15% Ti Ferrite by High-Resolution Autoradiography and during Degassing at High Temperatures,” Scripta Metallurgica, vol. 11, pp. 467-472. 1977.

60. G. M. Pressouyre and I. M. Bernstein, “A Quantitative Analysis of Hydrogen Trapping,” Metallurgical Transactions A, vol. 9, pp. 1571-1580. 1978.

61. J. Kittel, J.W. Martin, T. Cassagne and C. Bosch, “Hydrogen Induced Cracking (HIC) – Laboratory Testing Assessment of Low Alloy Steel Linepipe,” Corrosion Conference & Expo, Paper No. 08110. NACE International, 2008.

62. D. Hejazi, A.J. Haq, N. Yazdipou, D.P.Dunne, F. Barbaro, and E.V. Pereloma, “Role of Microstructure in Susceptibility of X70 Plateline Steel to Hydrogen Embrittlement,” Materials Science Forum Cols, Trams Tech Publications, Switzerland, pp. 162 – 165, 2010.

63. K. Easterling, Introduction to the Physical Metallurgy of Welding, 2nd edition, ButterworthHeinemann, pp219 -233. 1992.

64. Personal Communications with ArcelorMittal, Plate/HR Tubular Products Group, ArcelorMittal, December 13, 2013.

86