Review of the effect of hydrogen on mechanical properties of low strength steels in oil and gas applications

Abstract

Hydrogen-Induced Cracking (HIC) is one of several related mechanisms whereby absorbed hydrogen atoms can compromise the integrity of components manufactured of low strength steels. A “low strength steel” is defined as having a maximum hardness of 22 HRC (249 HV). The corresponding maximum tensile strength is of the order of 800 MPa (116 ksi). Steels having localized areas with microhardness in excess of 22 HRC are particularly vulnerable to the development of HIC damage. HIC is a term applied to phenomena which occurs at low temperatures (typically less than about 90 °C), and must not be confused with high temperature hydrogen attack of low strength carbon-manganese and low alloy steels exposed to hot hydrogen gas-containing environments. This review will highlight main factors that affect HIC development or failure by considering the following: (i) Metallurgical factors (effect of materials and microstructures), and (ii) Environmental exposure factors.

1. Introduction

It is widely recognized that hydrogen can embrittle metallic materials, especially steels, where hydrogen is absorbed in the atomic form generated from adsorbed molecular hydrogen. The adsorbed hydrogen can then diffuse through the material, which can cause or contribute to the following forms of hydrogen damage: (i) Hydrogen embrittlement (HE)/or Hydrogen-induced cracking (HIC), (ii) Hydrogen stress cracking (HSC), and (iii) Sulphide stress cracking (SSC).

HIC occurs when H atoms diffusing through a pipeline steel becomes trapped as H₂ molecules at inhomogeneities in the steel. A planar, gas-filled defect is created, which grows parallel with the pipe surfaces as it traps more diffusing H atoms.

The H atom source is normally the cathodic reaction of an acid corrosion mechanism occurring at the internal pipe surface, i.e., the reduction of hydrogen ions, H⁺: $\begin{eqnarray}\displaystyle \begin{array}{@{}ll@{}}\text{Anodic reaction:} & \quad \quad \text{Fe}\rightarrow \text{Fe}^{2+}+2\text{e}^{-}\\ \text{Cathodic reaction:} & \quad \quad 2\text{H}^{+}+2\text{e}^{-}\rightarrow 2\text{H}_{\text{ads}}\end{array} & & \displaystyle \nonumber\end{eqnarray}$

HIC will occur in a steel if there is a susceptible microstructure (metallurgical factor) and sufficient diffusing hydrogen (H) atoms to initiate and propagate damage (environmental exposure factor). These two factors are inter-related. All steels have a unique critical or threshold value of H atom concentration for initiation of hydrogen damage. If the concentration of diffusing H atoms (C^H_O value) is measured at the charging surface, this threshold value is expressed as (C^H_O)_TH.

2. Introduction and definitions

Several different names have been used to identify the types of hydrogen-related damage which have been observed in low strength steels.

Definitions based on appearance:

Hydrogen-Induced Cracking: The development of internal cracks in low strength steels as a consequence of the trapping of absorbed hydrogen atoms (H atoms) as gas molecules (H₂ molecules) at inhomogeneities is termed HIC. The cracks most often (but not always) lie parallel to the rolling plane and the surfaces of the steel component. Residual or applied tensile stress is not needed for HIC development. See Fig. 1.

Fig. 1.

(a) Hydrogen-Induced Cracking in plate steel exposed to sour gas. (b) The image shows a microstructure containing a significant number of laminar inclusions which act as nucleation sites for HIC.

•

Hydrogen Blisters or Hydrogen-Induced Blister Cracking (HIBC): Blisters are a form of HIC in which the build-up of hydrogen gas pressure at the initiated cracks or pre-existing (mill) laminations results in localized deformation and bulging of the steel to the closest surface (or to both surfaces, if mid-wall). Hydrogen gas pressures as high as 2,700 psi (18.6 MPa) have been measured inside blisters. The resultant surface bulging has the appearance of a blister when viewed from inside or outside the vessel. Blisters often occur when the hydrogen-induced crack is unable to propagate further in the parallel-to-surface direction, and is unable to link up with HIC on adjacent planes in the steel. This may be because the hydrogen atom-trapping inhomogeneity which caused the blister to form has a finite length, or there are no other hydrogen-induced cracks in close proximity (same or adjacent planes). See Fig. 2.

•

Stepwise Cracking (SWC): SWC is a form of HIC in which adjacent cracks on different planes in the steel link-up in a stepwise fashion. This may lead to throughwall cracking and the loss of vessel integrity. Applied or residual stress is not necessary for SWC development. The hydrogen gas pressure inside HIC cracks or blisters can increase to hundreds of atmospheres. This pressure generates internal stress at the tips of the HIC cracks, resulting in localized plastic deformation of the steel (visible during metallographic examination). The cold-worked steel is susceptible to hydrogen embrittlement cracking (HEC), which is the mechanism of link-up. Note that EFC Publication Number 16 states: “The name SWC is given to surface blistering and cracking parallel to the rolling plane of the steel plate which may arise without any externally applied or residual stress”. SWC has also been called “Hydrogen-Induced Stepwise Cracking” and (in Japan) “Type 1 Sulfide Stress Corrosion Cracking”. See Fig. 3.

•

Stress-Oriented Hydrogen-Induced Cracking (SOHIC): The presence of tensile stress in the component may cause individual ligaments of HIC to form in a stacked, through-thickness array. This is a necessary precursor to what is called SOHIC. This array is oriented perpendicular to the principal applied stress. The HIC may subsequently completely link up to cause throughwall cracking and loss of vessel integrity, i.e., SOHIC. SOHIC, like SWC, is probably a combination of HIC and either hydrogen embrittlement cracking (HEC) or sulfide stress cracking (SSC). HEC and SSC are possible in low strength steels if the material is highly stressed, especially if it is plastically-deformed and there is a stress-concentrating mechanism, and there is a high enough hydrogen atom charging rate. This can occur even if the steels have macrohardness less than 22 HRC and meet the other requirements of MR0175. SOHIC is most common in the heat-affected zones (HAZs) of welds in low strength steels, though can form in the base material. See Fig. 4.

The detailed mechanism of HIC, blistering, SWC and SOHIC will be discussed in a subsequent section. It is common to see two or more of these mechanisms active together in a hydrogen-damaged steel component.

Fig. 2.

(a) Hydrogen blister in NPS 6 sour gas pipeline (Canada 1997 failure). (b) HIC-blister (lab testing - NACE solution TM0284).

Fig. 3.

(a) Stepwise Cracking (SWC) in plate steel exposed to sour gas. (b) Interaction and coalescence of hydrogen-induced cracks mechanism.

Fig. 4.

SOHIC in linepipe steel. (a) Lab. C-ring test and (b) SOHIC failure of NPS 16 spiral-welded.

3. Mechanism of hydrogen embrittlement

HIC occurs when H atoms diffusing through a pipeline steel become trapped as H₂ molecules at inhomogeneities in the steel. A planar, gas-filled defect is created, which grows parallel with the pipeline surfaces as it traps more diffusing H atoms. If the defect grows sufficiently large, it may cause blister formation. HIC failure occurs if a mechanism exists for linkage of one or (normally) more nearby defects or blisters with the internal and external pipeline surfaces. Possible linkage mechanisms are SWC and SOHIC.

The H atom source is normally the cathodic reaction of an acid corrosion mechanism occurring at the internal linepipe surface, i.e., the reduction of hydrogen ions, H⁺: $\begin{eqnarray}\displaystyle \begin{array}{@{}ll@{}}\text{Anodic reaction}: & \quad \quad \text{Fe}\rightarrow \text{Fe}^{2+}+2\text{e}^{-}\\ \text{Cathodic reaction}: & \quad \quad 2\text{H}^{+}+2\text{e}^{-}\rightarrow 2\text{H}_{\text{ads}}\end{array} & & \displaystyle \nonumber\end{eqnarray}$

HIC will occur in a steel if there is a susceptible microstructure (metallurgical factor) and sufficient diffusing hydrogen (H) atoms to initiate and propagate damage (environmental exposure factor). These two factors are inter-related. All steels have a unique critical or threshold value of H atom concentration for initiation of hydrogen damage. If the concentration of diffusing H atoms (C^H_O value) is measured at the charging surface, this threshold value is expressed as (C^H_O)_TH.

A model for estimating the likelihood or probability that HIC development or failure might occur can be developed by considering the following metallurgical and environmental exposure factors.

4. Metallurgical factors

Segregation during casting is one of the most important metallurgical factors affecting the development of HIC in finished plate or pipe. The segregation of impurities results in the localized concentration of non-metallic inclusions, often as elongated single particles or as stringers of many particles. Voids may be associated with these individual particles or stringers. Both alloying and tramp elements may segregate during casting, resulting in the formation of “anomalous microstructures”. Many non-metallic particles and stringers, and the anomalous microstructures are excellent trapping sites for diffusing H atoms and the initiation and growth of HIC.

The centre (pipe or core) is typically the most highly segregated region of ingot cast steel. However, there is significant segregation in the ingot over approximately 80% of its thickness - “A - V” segregates. Only the outer 10% of the ingot thickness (two sides) is relatively free of segregation. Maximum susceptibility to HIC will occur in plates or hot-rolled skelp made from steel from locations corresponding with the centre of the ingot. Minimum susceptibility to HIC is expected at the edge of the plates and skelp and throughout the width of products made from steel taken from approximately the bottom 15% of the ingot. It is important to keep track of the location of the area of maximum segregation when skelp is cut into strip for fabrication of welded pipe. Quality assurance HIC testing must be performed on specimens taken from a plate or a joint of pipe made from the most segregated part of the ingot.

The HIC performance of products will be different between those made from ingot cast steel versus continuously cast steel as a consequence of the contrasting segregation behaviours. Products made from ingot cast steel are more likely to suffer SWC than those made from continuously cast steel. Products made from continuously cast steel are more likely to suffer centre wall blistering. From a pressure vessel/pipeline integrity perspective, blistering is preferable to SWC. However, there are several examples of linepipe failures which were associated with blistering plus SOHIC, see Fig. 5.

Fig. 5.

(a) Continuous casting and (b) ingots are manufactured by the freezing of a molten liquid in a mold (Google search).

4.1. Chemical composition factor

Many steels which have not been made to HIC-resistant requirements have good performance as sour gas pipelines or sour gas pressure vessels because they happen to have low concentrations of the inhomogeneities that trap diffusing H atoms and result in HIC and blistering. The relative susceptibility of a steel to HIC development can be estimated from a knowledge of the concentrations in the steel of the elements present in these inhomogeneities.

Metallographic investigations of actual HIC failures using light optical and scanning electron microscope (SEM) techniques have identified many types of non-metallic inclusions and other inhomogeneities in steel which trap diffusing H atoms as hydrogen gas molecules (H₂), leading to the development of HIC damage.

4.1.1. Type II manganese sulfide (MnS) particles

Manganese is added to steel as a solid solution strengthening agent and also to prevent the formation of FeS particles (which lead to problems in forging “hot shortness”). The sulfur preferentially combines with the Mn rather than the Fe. Unfortunately, in Al-Si fully-killed steels, the high levels of deoxidation results in high sulfur solubility in the molten steel. MnS solidifies last in the form of thin films in austenite grain boundaries and between dendrites. The “type II” MnS particles (which are formed coherently with the matrix) are quite soft, and are spread out parallel with the steel surfaces in working operations, e.g., during hot rolling to produce pressure vessel plate and skelp for welded linepipe. They assume a “pancake” shape, which is more elongated in the direction of working. The type II MnS particles are therefore oriented at right angles to the diffusing H atoms, and act as a barrier to their passage from the charging (entrance) surface to the exit surface. Type II MnS particles are probably the most common site for initiation of hydrogen damage. They may form at any location throughout the plate/pipe wall thickness. Note that if the steel has only been Si-semi-killed steels, a different MnS particle is formed, “type I”. Type I MnS particles solidify before the rest of the molten steel, assuming a globular shape. They are not as deformable during hot rolling operations as the type II particles, and are far less effective at trapping diffusing H atoms. Plate and pipe made from semi-killed steels have lower frequencies of HIC damage development, both in laboratory tests and in actual service.

4.1.2. Alumina (Al₂O₃) particles

Alumina particles may form in fully-killed steels if the total aluminum content exceeds about 0.035%. This is approximately the solubility limit for Al in steel. Aluminum is added to the steel to combine with the residual oxygen present. Low oxygen content favours a finer grain size and better impact toughness properties. Alumina particles may not be totally removed from the molten steel during the refining and casting process. Aluminum may be oxidized in the tundish of continuous casters if they are not effectively inert gas-shrouded. The resultant alumina particles may not have sufficient time to leave the steel before it solidifies. Alumina particles are hard and brittle and do not form coherently with the steel matrix. They shatter during working operations, forming particles with crystallographic shapes surrounded by voids. The alumina particles and associated voids are excellent trapping sites for diffusing H atoms.

4.1.3. Anomalous microstructure

This is the name given by Japanese researchers to the very thin (about 50 μm) layers of microhard material that form in the pearlite bands which lie parallel to the working direction of plate and skelp having “conventional” levels of carbon, i.e., about 0.20%. The anomalous microstructures develop during plate/skelp cooling after hot rolling. SEM-EDS analysis has shown that the anomalous microstructures contain higher levels of manganese and phosphorus (especially), and carbon (to a lesser extent) than the matrix. Concentrations of Mn may reach 1.3 to 2.0 times the bulk concentration as given in the steel product analysis, whilst P may reach 5 to 15 times the steel product analysis. The consequence of this segregation of high concentrations of Mn and P is that martensite/bainite forms locally when the steel is cooled, rather than the expected ferrite plus pearlite (matrix). The anomalous microstructure can be quite hard. The hard microstructure is susceptible to H damage, more strictly as a consequence of sulfide stress cracking (SSC) or hydrogen embrittlement cracking (HEC). However, since it occurs in otherwise soft material, it is included as a mechanism of HIC failure. This is especially of concern in electric resistance-welded (ERW) linepipe and downhole tubulars because the bands containing the anomalous microstructure become re-oriented during the forging operation accompanying high frequency welding. The bands may end up being almost at right angles to the internal and external pipe or tube surfaces on either side of the ERW. These results in there being two potential through-thickness crack paths if the H atom charging rate is sufficiently high to initiate HIC in the anomalous microstructure (see Fig. 6).

The presence of the anomalous microstructure has been a major contributing factor to the failure of many ERW linepipes: see Figs 7 and 8, and References to HIC Damage and Failures in Pipelines [5–10].

Fig. 6.

SOHIC along anomalous microstructure re-oriented at ERW (lab. HIC test).

Fig. 7.

HIC along mid-wall segregation zone in ERW linepipe (Canada 1997 failure).

Fig. 8.

Linepipe ruptured during hydrotest due to HIC at ERW (Canada 1981 failure).

4.1.4. Other non-metallic particles

“Massive” niobium carbonitride particles have been associated with HIC initiation. Early attempts to control the shape of sulfide particles by the addition of rare earth metals (REMs) were not always successful. Cerium sulfide (CeS) and other REM-sulfide inclusions acted as HIC initiation sites if incomplete globularization occurred, or if there was clustering of the inclusions.

4.2. Plate/skelp manufacturing practices

It is common practice to use controlled rolling to manufacture plate and skelp for welded pipe. Type II MnS particles are more easily deformed (elongated) at lower temperatures. A significant improvement in HIC resistance was demonstrated for skelp rolled at 900 °C compared with the same chemistry skelp rolled at 790 °C. The optimum minimum temperature for controlled rolling of steels containing type II MnS inclusions is 1,000 °C. The deformability of the steel matrix is reported to be equal to that of the MnS inclusions at this temperature. A low controlled rolling finishing temperature (678 °C) was implicated in many oil industries and just recently in the Arabian Gulf pipeline.

Cold working operations (e.g., heavy cold rolling) increase susceptibility to HIC. Strains of about 2% maximum appear not to affect HIC susceptibility whilst higher strains may increase susceptibility. For this reason, specimens for HIC testing should not be obtained by flattening curved samples taken from smaller diameter, thin wall pipe. As-received, curved specimens should be employed in the HIC testing procedure.

4.3. Heat treatment

Quenching and tempering (Q&T) and other heat treatments (e.g., tempering or normalizing) have been shown to improve the resistance of some steels to HIC. However, complete immunity to HIC cannot be guaranteed by heat treatment alone. Q & T can remove some of the chemical segregation/anomalous microstructures that result from casting and pearlite banding, but cannot eliminate deleterious non-metallic inclusions. Tempering at 650 °C for 30 minutes has been shown to decompose martensite in acicular ferritic steels to cementite plus pearlite, thereby reducing susceptibility to HIC [11–18].

5. Environmental exposure factors

See Figs 9, 10, and 11 for an overview of the effects of environmental factors on hydrogen atom entry into, and permeation through, steel.

Fig. 9.

Factors affecting H Atom permeation transient immediately after breakthrough.

Fig. 10.

Effect of pH and other factors on C^H_O value.

Fig. 11.

Effect of the presence/absence of passive FeS scale on C^H_O value.

5.1. pH and chloride ion concentration

The most HIC-active environments are water with a low pH and a high dissolved H₂S content, and moist gas with a high H₂S partial pressure. The lower the pH, the higher the corrosion rate and the resultant surface concentration of diffusing H atoms. Since the solubilities of iron sulfides (FeS) increase with decrease in pH, any benefit from their presence and partially protective nature will be lost at lower pH. The presence of CO₂ in the sour gas will be harmful, even at low partial pressures of H₂S, because of its contribution to the lowering of pH.

Laboratory tests by numerous researchers have shown that:

•
Susceptibility to HIC increases with decrease in pH over the range 6 to 1, and
•
The surface concentration of diffusing H atoms (C^H_O value) increases with decrease in pH over the range 5.4 to 2.7.
•
The increase in C^H_O value can be as high as two times when the pH is reduced from 5.4 to 2.7.

At constant pH, the presence of chloride ions causes greater corrosion and higher H atom absorption into steel than pure water. This would be expected to result in higher material susceptibility to HIC. It has also been shown that synthetic seawater saturated with H₂S causes more corrosion of, and hydrogen absorption into steel than pure water saturated with H₂S at ambient conditions. This occurred despite the fact that the latter solution has lower pH.
5.2. Temperature

The temperature range of maximum susceptibility to HIC is between 15 °C and 35 °C. Susceptibility decreases below 15 °C, probably as a consequence of lowered corrosion rate and H atom production, and above 35 °C.

Several factors may contribute to decreased HIC susceptibility at temperatures above 35 °C:

•
Decreased solubility of H₂S in the aqueous phase, resulting in higher pH.
•
Lowered corrosion rate and absorbed H atom concentration as a consequence of the formation of more protective iron sulfide scales on the steel.
•
The diffusivity (diffusion coefficient) of H atoms in steel increases with rising temperature, reducing their residence time and the probability that they will become trapped as H₂ molecules at inhomogeneities.
•
Laboratory testing has shown that threshold H atom concentrations for HIC initiation [(C^H_O)_TH values] increase with temperature increase.

The quantity of H atoms absorbed by HIC test specimens has been measured to be a maximum at 25 °C, hence the choice of this temperature for the standardized HIC test method. HIC damage is not expected at temperatures in excess of about 90 °C, even in very susceptible materials.
5.3. Protective scales

The presence of protective (passivating) iron sulfide scales is a major factor in preventing HIC failures of sour gas pipelines. Hydrogen permeation monitoring has shown that C^H_O values are very low if the sour gas pipeline has been in service for more than a month or so, even if the internal environment is potentially quite corrosive. The shape of the C^H_O value versus time plot for a new pipeline is very similar to that of an unused laboratory specimen when first exposed to a sour environment:

•
There is an initial period of time after wet sour gas or H₂S-saturated water has first contacted the internal pipe surface when no increase in C^H_O (above background) is apparent. During this time, corrosion of the steel is occurring and H atoms are being absorbed into the steel. There is a time delay whilst these H atoms diffuse through the pipe wall. The length of this time delay depends on the diffusivity (or diffusion coefficient) of H atoms in the steel at the operating temperature and the wall thickness. The approximate time delay is given by the formula: $\begin{eqnarray}\displaystyle \text{t}_{\text{b}}=\frac{L^{2}}{15.3D} & & \displaystyle \nonumber\end{eqnarray}$ where t_b = time delay (s) $\begin{eqnarray}\displaystyle \text{L} & = & \displaystyle \text{wall thickness}(\text{cm})\nonumber\\ \displaystyle \text{D} & = & \displaystyle \text{diffusivity of H atoms in the steel}(\text{cm}^{2}/\text{s})\nonumber\end{eqnarray}$ The presence of mill scale does not appear to act as a barrier to sour gas corrosion of the underlying steel, i.e., iron oxides are not effective (like iron sulfides) in passivating the steel surface.
•
Eventually, the first H atom breaks through to the external pipe surface and is detected by the permeation monitoring equipment. The C^H_O value builds up to a peak value as more and more H atoms arrive at the external surface. The peak C^H_O value is normally achieved within few hours. Its magnitude depends on several factors, including:

–
The corrosivity of the environment.
–
The susceptibility of the pipe material to corrosion.
–
The fraction of the H atoms produced by the cathodic corrosion reaction which become absorbed into the steel and successfully diffuse through the wall thickness.
–
The H atom diffusivity in steel.
–
The temperature.
–
The rate at which passivating iron sulfide scales are formed as a consequence of corrosion.
–
The pipe wall thickness.

•
The C^H_O value then falls exponentially with time, typically reaching a low, steady reading after two or three weeks, depending on the system.
•
After several months or so, the C^H_O value may fall to zero.

The reason for the fall in corrosivity and C^H_O value is the gradual development of a partially protective (or passivating) iron sulfide scale on the pipe internal surface. This also contributes to the fall in C^H_O value observed in laboratory permeation experiments. An iron sulfide (FeS) scale is formed whenever steel is exposed to an aqueous sour environment or a moist, sour gaseous environment. Depending on the conditions of formation, these scales can have greater or lesser protective properties, i.e., ability to resist corrosion of, and H atom entry into, the underlying steel. Scales formed in a sour aqueous environment are generally more protective than those formed in moist sour gas, probably because they are more dense and adherent and less friable.

At least eight different iron sulfides have been identified in H₂S (sour gas)/steel systems. Several of these compounds are nonstoichiometric. Some have similar or identical chemical formulae, differing only in their crystallinity. Irrespective of the pH of the aqueous system, or the presence or absence of oxidants such as O₂ and elemental sulfur, it is believed that the iron sulfide mackinawite, with approximate formula Fe₉S₈, forms first. Other iron sulfides, their chemical formulae, and their condition of formation (anaerobic versus aerobic) are as follows:

Mackinawite is the most soluble iron sulfide in acid systems, and will continually be dissolved in the aqueous sour gas environment (thermodynamics). The solubilities of iron sulfides decrease as the S to Fe atom ratio increases. Pyrite and marcasite are the most insoluble iron sulfides, having approximate solubilities of 1 ×10⁻⁷ mol/kg in a sour aqueous environment, pH 3.5–4.0, 1–18 atmospheres partial pressure of H₂S. Solubility ratios (Pyrite = 1) for iron sulfides formed under similar anaerobic conditions are:

Solubilities of iron sulfides increase as the pH is decreased.

Mackinawite is also the most rapidly dissolved iron sulfide (kinetics). Pyrite is dissolved most slowly in sour aqueous environments, at a rate of approximately 8 ×10⁻¹ mol/m²/s at pH 3.5, 18 atmospheres H₂S, 25 °C. Dissolution rate ratios (Pyrite = 1) for iron sulfides under these conditions are as follows:

Dissolution rates of iron sulfides increase as the pH is decreased.

The formation of the iron sulfide mackinawite is not favourable for the control of corrosion and H atom charging of the steel from both the thermodynamic and kinetic considerations (compared with the formation of the more sulfur-rich or ‘higher’ iron sulfides). Under heavy water plant environmental conditions, both corrosion and H atom permeation rates are controlled by the dissolution of mackinawite. It is preferable that more thermodynamically- and kinetically-stable iron sulfides form on the steel in the sour gas system, providing, of course, that they remain adherent to the steel substrate. Scales comprised of the higher iron sulfides can form in two ways:

•
Transformation from the lower (iron-rich) to the higher iron sulfides. This is promoted by higher temperature, higher pH, inhibitors and time, and prevented by high CO₂ partial pressures, high fluid flow rates and scale dissolution.
•
Deposition from solution of the higher (more insoluble) iron sulfides. This may occur when the ionic concentrations of ferrous and sulfide ions exceed the solubility product of the particular iron sulfide in the aqueous environment.

Iron sulfide scales rich in pyrrhotite and/or pyrite are normally very protective for the underlying steel.
6. Avoidance of HIC – new linepipe

Two measures can be taken to increase the HIC resistance of new linepipe:

•
Lower C^H_O, the surface concentration of diffusing H atoms.
•
Raise (C^H_O)_TH, the threshold surface concentration of diffusing H atoms to initiate HIC.

To a limited extent, steps can be taken to increase the resistance of the steel to crack propagation once HIC has initiated. This often occurs as a consequence of the processing undertaken to raise (C^H_O)_TH.
6.1. Lowering the C^H_O value

The C^H_O value can be lowered by such measures as chemical corrosion inhibition and encouraging the development of passivating iron sulfide films. However, these measures are normally taken when avoiding HIC in existing steels not made to a HIC-resistant specification. Materials should be selected which are inherently resistant to all known cracking phenomena.

The C^H_O value may be reduced in some sour environments by alloying the steel with certain elements:

•
Copper. Alloying the steel with a small amount of copper (Cu) can improve HIC resistance in mildly sour environments. Much work has been performed to determine the optimum concentration of Cu for reducing the C^H_O value, and the environmental conditions under which this reduction occurs. Hydrogen desorption measurements have been conducted on linepipe steels with Cu contents from 0.02% up to 0.40%.

Addition of Cu was ineffective if the sour environment had a pH below about 4.5. Addition of Cu does not affect the (C^H_O)_TH value. It merely prevents HIC in susceptible materials by keeping C^H_O < (C^H_O)_TH in sour environments with pH values in excess of about 4.5. Weight loss determinations have shown that Cu-containing steels have lower corrosion rates in pH 4.5 and higher sour environments. The presence of Cu is believed to stabilize the iron sulfide (FeS) scale formed on the steel.

Copper has a solid solution strengthening effect, so it is commonly added to HIC-resistant steels, even if they are intended for use in sour environments which have pH values below 4.5. This effect is highly desirable because the concentrations of other steel strengthening elements such as carbon and manganese must be lowered to improve HIC resistance [raise (C^H_O)_TH].

•
Cobalt. About 0.6 – 1.0% cobalt will significantly reduce the C^H_O value of Cu-free steels in sour environments, including those with low pH. Enriched levels of cobalt have been found in the protective iron sulfide scale. The beneficial effect of copper additions is enhanced by cobalt additions. Because of the cost, cobalt-alloyed steels have not been produced commercially for sour gas linepipe fabrication.
•
Nickel. Several researchers have reported that the addition of Ni reduces the corrosion rate of Cu-free steels in sour environments. However, some researchers reported a reduction of H atom absorption (at 0.2% Ni and at >0.6% Ni), whilst another reported no effect on the C^H_O value at 0.1–0.5% Ni.
•
Molybdenum. Addition of from 0.1–0.5% Mo neither reduced the corrosion rate nor the C^H_O value of Cu-free steels exposed to sour environments.
•
Chromium. One researcher reported that 0.1–0.5% Cr had no effect on the corrosion rate and C^H_O value of Cu-free steels exposed to sour environments. Another said that 0.6% Cr was effective at reducing H atom absorption.

Note that no further steps need be taken to raise the (C^H_O)_TH value if the C^H_O value is lowered by alloying, if the sour environment pH is appropriate.
6.2. Raising the (C^H_O)_TH value

The (C^H_O)_TH value will be increased by:

•
Eliminating type II MnS inclusions, Al₂O₃ particles, the anomalous micro-structure, and other non-metallic inclusions, e.g., silicates, slag, iron oxides, massive niobium carbonitride particles, etc.
•
Reducing the segregation of trace/impurity elements in the steel.
•
Increasing the homogeneity of the microstructure.

Three steps are typically taken during steelmaking to eliminate type II MnS inclusions:

•
The sulfur content is reduced to as little as possible, normally 0.003% maximum and often less than 0.001%. This is typically achieved through the use of a high sulfur capacity slag during steelmaking.
•
The manganese content is controlled to 1.00%, maximum, and ideally 0.80% maximum.
•
The remaining metal sulfide inclusions are ‘shape-controlled’.

‘Shape-controlling’ involves adding a metal to the molten steel which has a higher affinity for sulfur (S) than does manganese (Mn), producing a metal sulfide that does not deform on rolling, remains coherent with the steel matrix, and does not trap diffusing H atoms as H₂ molecules. The remaining metal sulfide inclusions are ‘shape-controlled’. Early attempts to control the shape of sulfide inclusions by the addition of rare earth metals (REM) were not always successful in reducing susceptibility to HIC. Cerium and other REM-sulfide and oxide inclusions acted as HIC initiation sites if incomplete globularization occurred, or if there was clustering of the inclusions. The author has tested a commercially-made, REM-treated, ‘HIC-resistant’ linepipe steel and found it to have very high relative susceptibility to HIC. Severe HIC damage was found where REM sulfides had clustered. REM treatment is no longer used for sulfide shape control in HIC-resistant steels.

The addition of calcium (Ca) to form spherical, non-deformable calcium sulfide (CaS) inclusions is now the favoured method of shape controlling steels for HIC-resistant plate and linepipe. Calcium is typically lance-injected (often as a Ca-Si compound) during the ladle refining stage of steelmaking. This occurs subsequent to desulfurization and deoxidation, and normally concurrent with argon-stirring. Calcium has also been added to the tundish (argon-shrouded) during continuous casting, and to the teeming spout when the molten steel is bottom-poured to ingots. The element has also been added in iron-clad form (lumps) and as Ca-Al wire. It is important that the oxygen content of the molten steel is low prior to calcium treating, and remains low during casting to prevent oxidation of the calcium (and excess Al used to kill the steel or present through the use of Ca-Al wire).

It is also important to have an optimum atomic ratio of Ca to S for perfect shape control and high HIC resistance. For steels with 0.003% S maximum, these ratios should be a minimum of about 1.8 for plate and welded linepipe, and a minimum of 1.0 for seamless linepipe. A ratio greater than 1.0 is required for plate and skelp for welded linepipe because some of the Ca reacts with alumina particles to form spherical calcium sulfide-aluminate inclusions (Ca-Al-O-S). It is also important to limit the maximum Ca to S ratio. If the ratio is too high, deleterious calcium oxy-sulfide and calcium sulfide-aluminate inclusions form as clusters. These are excellent H atom trapping sites, increasing HIC susceptibility. For plates and welded linepipe, a maximum Ca to S ratio of 3.5 is recommended. A Ca to S ratio up to 2.5 does not increase the HIC susceptibility of seamless linepipe.

The Ca to S ratio must be controlled within a narrower range if the S content is higher than 0.003%. For a 0.004% S content steel, the maximum Ca to S ratio should be 2.5, whilst it should be 2.0 for a 0.005% S content steel. Shape control is not recommended if the sulfur content exceeds 0.005%.

Additional benefits of effective sulfide shape control measures include:

•
A marked improvement in impact toughness.
•
Improved sulfide stress cracking resistance.

It has been reported that 100% sulfide shape control by Ca addition is more easily achieved if the steel is continuously cast versus ingot cast.

Alumina (Al₂O₃) particle stringers, or the voids associated with them, are very effective H atom trapping sites and HIC damage initiators. For both seamless and welded pipe, and plate, the presence of Al in ‘total’ concentrations above about 0.030% may indicate the presence of Al₂O₃ particle stringers and aluminates. (Note that the ‘total’ Al is the sum of the Al dissolved in the steel and the Al in alumina, aluminates and aluminum nitrides. Mill certificates may report either ‘Total’ or ‘Sol’ Al. Sol (Soluble) Al is the Al soluble in an acid, typically H₂SO₄, i.e., the Al dissolved in the steel plus the Al in the aluminates and nitrides. The Al as alumina can be estimated by subtracting Sol. Al from Total Al, if they are both known.)
7. Conclusions

•

Multiple types of crack initiation sites exist for hydrogen-induced cracking.

•

The elongated MnS and planar array of other inclusions are primarily responsible for HIC.

•

Lower volume fractions of inclusions corresponded to higher resistance to HIC.

•

However, the microstructure may also play a role in HIC, in particular, heavily banded microstructures could enhance HIC by providing low fracture resistance paths for cracks to propagate more easily. The balance of C and Mn has an overriding effect on resistance to HIC.

•

All steels have a unique critical or threshold value of H atom concentration for initiation of hydrogen damage. If the concentration of diffusing H atoms (C^H_O value) is measured at the charging surface, this threshold value is expressed as (C^H_O)_TH.

•

H damage will only initiate in a steel if the C^H_O value exceeds the steel’s (C^H_O)_TH value. The steel may be highly susceptible to HIC, but may never suffer damage because of exposure to a benign environment [C^H_O < (C^H_O)_TH]. Conversely, the steel may have very good resistance to HIC (high (C^H_O)_TH value), but because it is exposed to a very aggressive environment (high C^H_O value), may still suffer HIC initiation; C^H_O > (C^H_O)_TH.

Conflict of interest

None to report.

References

Elboujdaini

, Research in Progress, NACE Corrosion/2009.

Uhlig’s Corrosion Handbook: Chapter#12: M. Elboujdaini “Hydrogen-Induced Cracking and Sulphide Stress Cracking”; Second Edition, Edited by R. Winston Revie; John Wiley & Sons Inc, New York, 2000.

REVIEW - Metallurgical factors in stress corrosion cracking (SCC) and hydrogen-induced cracking (HIC), J Solid State Electrochemistry (2009). (Dedicated to the 85th birthday of John O.M. Bockris).

Paredes

and Mize

W.W.

, Hydrogen blisters - their causes and methods of prevention, Gas40(12) (1954), 89.

Paredes

and Mize

W.W.

, Unusual pipeline failures traced to hydrogen blisters, Oil and Gas Journal52(12) (1954), 99.

Williams

D.N.

, Correlation of the Results of Stepwise Cracking Tests with Service Experience in Sour Gas. Battelle report to NG-18 Line Pipe Research Supervisory Committee of the American Gas Association, October 1983. NG-18 Report#138.

Class

, Report on Investigation of Sulfide Stress Corrosion Cracking of Steels, Particularly of Steels of Comparatively Low Tensile Strength. Second International Conference on Metallic Corrosion, New York, U.S.A., 1963-03-11/15, pp 342-355 (NACE 1966).

Leslie

and King

J.A.

, Sour Gas Pipelines. Paper OTC 5741, Volume 2, pp 499, 20th Annual Offshore Technology Conference, Houston, Texas, USA, 1988-05-02/05.

Nippon Steel Corporation. The Investigation of the Cause of Gas Pipe Failure. NSC Report 1975-10-20.

10.

Moore

E.M.

and Hansen

D.A.

, Specifying Linepipe Suitable for Safe Operation in Sour, Wet Service, Transactions ASME, Journal of Energy Resources Technology104(134) (1982), and ASME 81-PET-1.

11.

Moore

E.M.

and Warga

J.J.

, Factors Influencing the Hydrogen Cracking Sensitivity of Pipeline Steels. NACE Corrosion/76 Paper 144, Houston, Texas, USA, 1976-03-22/26, and Materials Performance 15(6), 17, (1976).

12.

El-Jundi

I.M.

, Qatar NGL-2 pipeline problems, Society of Petroleum Engineers Production Engineering1(6) (1986), 478.

13.

Alberta

Canada

, Energy Resources Conservation Board data, including Pipeline Integrity Workshop, 1992-10-08.

14.

Moore

E. M.

, Hydrocarbon-Induced Damage in Sour, Wet Crude Pipelines. SPE 11514, Middle East Oil Technical Conference of the Society of Petroleum Engineers, Manama, Bahrain, 1983-03-14/17.

15.

Bruckhoff

Geier

Hofbauer

Schmitt

and Steinmetz

, Rupture of a Sour Gas Line Due to Stress Oriented Hydrogen Induced Cracking - Failure Analyses, Experimental Results and Corrosion Prevention. NACE Corrosion/85, Paper 389, Boston, USA, 1985-03-25/29.

16.

Kurdi

A.M.

Abougfeefa

M.S.

and Denney

A.K.

, Rehab permits desert line to run at original pressures, Oil and Gas Journal91(30) (1993), 76 (Fourth European and Middle Eastern Pipeline Rehabilitation Seminar, Abu Dhabi, United Arab Emirates, 1993-04-24/27).

17.

Punter

Fikkers

A.T.

and Vanstaen

, Hydrogen-induced stress corrosion cracking on a pipeline, Materials Performance31(6) (1992), 24 (Proceedings of the Pipe Protection Conference, Cannes, France, 1991).

18.

Gajam

S.Y.

and El-Amari

, Failure of a gas/condensate line, Materials Performance31(10) (1992), 55.

19.

Hay

M.G.

and Stead

M.D.

, The hydrogen-induced cracking failure of a seamless sour gas pipeline. NACE Canadian Region Western Conference, Calgary, Canada, 1994-02-07/10.

20.

Hay

M.G.

and Rider

D.W.

, Integrity management of a HIC-damaged pipeline and refinery pressure vessel through hydrogen permeation measurements. NACE Corrosion/98 Paper 395, San Diego, California, USA, 1998-03-22/27.

21.

Commodity Pipeline Occurrence Report - Natural Gas Pipeline Rupture. Transportation Safety Board of Canada Report Number P94H0003, 1995-08-23.

22.

Jones

and Hetu

, A Case Study of Hydrogen Induced Cracking Failure - Investigation, Remedial Action and Maintenance. NACE International, Calgary Section, One Day Seminar on Failure Mechanisms, Calgary, Alberta, Canada, 1998-04-07.

23.

Craig

B.D.

Bruno

T.V.

and Buehler

W.M.

, Complexities of Failure Analyses in Sour Systems. Canadian Welding Society/American Society for Metals conference “Materials and Welding for Sour Service and Low Temperature Service”. Calgary, Alberta, Canada, 1997-11-20/21.

Review of the effect of hydrogen on mechanical properties of low strength steels in oil and gas applications

Abstract

1. Introduction

2. Introduction and definitions

4. Metallurgical factors

4.1.1. Type II manganese sulfide (MnS) particles

4.1.2. Alumina (Al2O3) particles

4.1.3. Anomalous microstructure

4.2. Plate/skelp manufacturing practices

4.3. Heat treatment

5. Environmental exposure factors

Conflict of interest

References

4.1.2. Alumina (Al₂O₃) particles