Metallurgy and Corrosion Control in Oil and Gas Production

PREFACE

Were I to wait perfection, my book would never be finished. This quote from Tai K’ung was in the preface to the first edition, and it is worth repeating.

I had no idea that the second edition of a book would be so much work, but it is now time to submit my efforts to the publisher and thank those who have helped me get to this point. The first edition was dedicated to my daughter, Krista Heidersbach, who also helped me with this update. Over the years her advice and insights have been invaluable. In addition to the items that have been included in this book, she also advised on those subjects that were best omitted. No father could be prouder than I am of her, and I thank her for all the time she spent helping me on this book.

Most engineers can handle the routine; it’s the details that have caused problems in recent years. The Pareto principle, often called the 80/20 rule, suggests that most problems are associated with only some of the equipment in any installation or oilfield. In recent years I have been involved in working on both corrosion under insulation (CUI), which can cause surprising leaks in piping systems that are hard to inspect, and failure analysis on high‐strength steel fasteners. The problem with broken bolts of subsea equipment in the Gulf of Mexico is an ongoing technical challenge that has resulted in a number of new or revised industrial standards. These relatively small components on large structures can lead to catastrophic consequences. This second edition has increased discussion on these two subjects that reflects increased concern by industry, both in North America and worldwide. It is likely that updates and changes on these and other topics will continue in future years, and I have attempted to suggest where additional information on these, and other problem areas, may be found.

I suggest the reader should pay particular attention to the differences between corrosion monitoring, which is routine in many organizations, and the need for better inspection. Monitoring cannot replace inspection. No organization has the resources to inspect everything that can corrode or break, and the developments and implementation of risk‐based inspection, concentrating on equipment most likely to fail and have high consequences, cannot be overemphasized. It is unfortunate that the oil and gas industry continues to experience equipment failures that would have been prevented if management, and the technical experts who advise them, had understood that monitoring cannot replace, or substitute for, inspection.

In recent years the LinkedIn online discussion groups started by Riky Bernardo, currently with RasGas in Qatar but originally from Indonesia where I met him, have grown to over 50 000 members. Steve Jones, whom I have never met, often contributes ideas to this discussion group and is one of the international group of leaders that I have relied upon to help me develop this revised manuscript. I urge the reader to monitor the discussions at the LinkedIn Oil and Gas Corrosion and Materials Selection site. On the day that I am writing this, there is informed discussion of erosion corrosion comparing API RP14E and DNV RP‐0501, hydrogen embrittlement of high‐strength steels, post‐weld heat treatment of clad piping, and a variety of coatings topics. The reader can learn a lot from these discussions, because no international standard, let alone this book, can cover the entire range of technical questions that come up.

I am a metallurgical engineer by training, but materials selection and corrosion control is so multidisciplinary that many other fields must also be considered for safe and reliable oilfield operations. In addition to the people identified in the preface to the first edition of this book, I would also like to acknowledge the following people who have guided me on the preparation of this revised second edition:

Steve Jones, Ivan Gutierrez, Arun Soman, Reza Javaherdashti, and Roger Francis have provided many helpful discussions on Riky Bernardo’s LinkedIn sites and have taught me valuable lessons. I have never met these people, and probably never will, but their advice has been invaluable to me and to many others.

Herb Townsend, Tom Goin, Ian MacMoy, and Candi Hudson have been leaders in redefining how high‐strength fasteners can be used in marine and other environments. I urge the reader to keep up discussions on this still evolving technology.

Bob Gummow reviewed the 2011 discussions on cathodic protection and offered valuable advice on how it could be improved.

Travis Tonge, Ramesh Bapat, Chelsea LeHaye, and Juan Imamoto are examples of experts I met while teaching corrosion classes. I have learned far more from them than the little I was able to teach them.

Juan Imamoto and Tom Kuckertz offered me the opportunity to work on CUI projects in widely different situations. I learned a lot from them and from the work we did together. Bob Guise and Kash Prakash worked with me on one of these projects, and they taught me a lot.

Tim Bieri and coworkers at BP have remained valuable resources. I talked to Tim on several occasions, and he is one of my go‐to experts. The 2006 NACE BP article on inspection and monitoring is one of the best explanations of the advantages and limitations of both inspection and monitoring that I have found, and I have relied heavily on this and their other advice.

Chemists look at things differently than engineers, and both Mark Kolody and Luz Marina Calle are two of my favorite chemists. When I am confused in my understanding, they have the unique ability to explain things in ways that I can understand.

All writers need to have critical reviewers. Gurudas Saha, whom I have only met through the Internet, is one of these experts whose advice I value highly.

BOB HEIDERSBACH

Cape Canaveral, Florida, USA

1

INTRODUCTION TO OILFIELD METALLURGY AND CORROSION CONTROL

The American Petroleum Institute (API) divides the petroleum industry into the following categories:

Upstream

Downstream

Pipelines

Other organizations use terms like production, pipelining, transportation, and refining. This book will discuss upstream operations, with an emphasis on production, and pipelines, which are closely tied to upstream operations. Many pipelines could also be termed gathering lines or flowlines, and the technologies involved in materials selection and corrosion control are similar for all three categories of equipment.

Until the 1980s metals used in upstream production operations were primarily carbon steels. Developments of deep hot gas wells in the 1980s led to the use of corrosion‐resistant alloys (CRAs), and this trend continues as the industry becomes involved in deeper and more aggressive environments [1, 2]. Nonetheless, most metal used in oil and gas production is carbon or low‐alloy steel, and nonmetallic materials are used much less than metals.

Increased emphasis on reliability also contributes to the use of newer or more corrosion‐resistant materials. Many oilfields that were designed with anticipated operating lives of 20–30 years are still economically viable after more than 50 years. This life extension of oilfields is the result of increases in the market value of petroleum products and the development of enhanced recovery techniques that make possible the recovery of larger fractions of the hydrocarbons in downhole formations. Unfortunately, this tendency to prolong the life of oilfields creates corrosion and reliability problems in older fields when reductions in production and return on investment cause management to become reluctant to spend additional resources on maintenance and inspection.

These trends have all led to an industry that tends to design for much longer production lives and tries to use more reliable designs and materials. The previous tendency to rely on maintenance is being replaced by the trend to design more robust and reliable systems instead of relying on inspection and maintenance. The reduction in available trained labor for maintenance also drives this trend.

COSTS

A US government report estimated that the cost of corrosion in upstream operations and pipelines was $1372 billion per year, with the largest expenses associated with pipelines followed by downhole tubing and increased capital expenditures (primarily the use of CRAs). The most important opportunity for savings is the prevention of failures that lead to lost production. The same report suggested that the lack of corrosion problems in existing systems does not justify reduced maintenance budgets, which is a recognition that, as oilfields age, they become more corrosive at times when reduced returns on investment are occurring [3]. The 2013 environmental cracking problems with offshore pipelines in the Caspian Sea Kashagan oilfield are estimated to have cost billions of dollars for pipeline replacement costs plus lost production [4]. It is estimated that corrosion costs are approximately equal to mechanical breakdowns in maintenance costs.

SAFETY

While proper equipment design, materials selection, and corrosion control can result in monetary savings, a perhaps more important reason for corrosion control is safety. Hydrogen sulfide, H2S, is a common component of many produced fluids. It is poisonous to humans, and it also causes a variety of environmental cracking problems. The proper selection of H2S‐resistant materials is a subject of continuing efforts, and new industrial standards related to defining metals and other materials that can safely be used in H2S‐containing (often called sour) environments are being developed and revised due to research and field investigations [2].

Pipelines and other oilfield equipment frequently operate at high fluid pressures. Crude oil pipelines can leak and cause environmental damage, but natural gas pipeline leaks, like the corrosion‐related rupture in Carlsbad, New Mexico, shown in Figure 1.1, can lead to explosions and are sometimes fatal [5]. High‐pressure gas releases can also cause expansive cooling leading to brittle behavior on otherwise ductile pipelines. API standards for line pipe were revised in 2000 to recognize this possibility. Older pipelines, constructed before implementation of these revised standards, are usually made from steel with no controls on low‐temperature brittle behavior and may develop brittle problems if they leak. Gas pipelines are more dangerous than liquid pipelines, because of the stored energy associated with compression of enclosed fluid.

Figure 1.1 Natural gas pipeline rupture near Carlsbad, New Mexico, in 2000.

Source: From Pipeline Accident Report [5].

ENVIRONMENTAL DAMAGE

Environmental concerns are also a reason for corrosion control [6]. Figure 1.2 shows oil leaking from a pipeline that suffered internal corrosion followed by subsequent splitting along a longitudinal weld seam. The damages due to this leak are minimal compared with the environmental damages that would have resulted if the leak had been on a submerged pipeline. Figure 1.3 shows an oil containment boom on a river where a submerged crude oil pipeline was leaking due to external corrosion caused by nonadherent protective coatings that shielded the exposed metal surfaces from protective cathodic protection currents.

Figure 1.2 Aboveground leak from an internally corroded crude oil pipeline.

Figure 1.3 An oil containment boom to minimize the spread of crude oil from an external corrosion leak on a submerged pipeline.

In the 1990s, the entire downtown area of Avila Beach, California, was closed because of leaking underground oil pipelines. The cleanup from these corroded pipelines took years and cost millions of dollars.

CORROSION CONTROL

The environmental factors that influence corrosion are:

CO2 partial pressure

H2S partial pressure

Fluid temperature

Water salinity

Water cut

Fluid dynamics

pH

Corrosion is normally controlled by one or more of the following:

Material choice

Protective coatings

Cathodic protection

Inhibition

Treatment of environment

Structural design including corrosion allowances

Scheduled maintenance and inspection

Figure 1.4 shows an offshore platform leg in relatively shallow water, approximately 30 m (100 ft) deep, in Cook Inlet, Alaska. The leg is made from carbon steel, which would corrode in this service. Corrosion control is provided by an impressed current cathodic protection system. The bottom of the leg is 2½ cm (1 in.) thicker than the rest of the leg, and this is intended as a corrosion allowance for the submerged portions of the platform legs. Note that the water level goes above the corrosion allowance twice a day during high tides, because the platform is located in water 3 m (10 ft) deeper than was intended during design and construction. Fortunately the cathodic protection system was able to provide enough current, even in the fast‐flowing abrasive tidal waters of Cook Inlet, to control corrosion. This platform was obsolete when the picture was taken, but it was less expensive to operate and maintain the platform than it was to remove it. Thirty‐five years later oil prices had increased, recovery methods had improved, and the platform was economically profitable. Robust designs, adequate safety margins, and continuous reevaluation of corrosion control methods are important, not just for marine structures but for all oilfield equipment.

Figure 1.4 Offshore platform leg in Cook Inlet, Alaska. The extra metal for the corrosion allowance is submerged twice a day during high tides.

While it might seem desirable to stop all corrosion, this is not necessarily cost effective. An 80 : 20 Pareto‐type rule probably applies: 80% of corrosion can be prevented for relatively modest cost, but the increased cost of the remaining corrosion would not be justified [7]. The British ALARP (as low as reasonably practicable) terminology is a similar concept discussed in many recent corrosion‐related documents and standards [8].

REFERENCES

1 Kane, R. (2006). Corrosion in petroleum production operations. In: Metals Handbook, Volume 13C – Corrosion: Corrosion in Specific Industries, 922–966. Materials Park, OH: ASM International.

2 Iannuzzi, M. (2011). Chapter 15: Environmentally‐assisted cracking in oil and gas production. In: Stress Corrosion Cracking: Theory and Practice (ed. V. Raja and T. Shoji), 570–607. Oxford: Woodhead Publishing, Ltd.

3 Ruschau, G. and Al‐Anezi, M. (September 2001). Appendix S: Oil and gas exploration and production. In: Corrosion Costs and Preventive Strategies in the United States, Report FHWA‐RD‐01‐156. Washington, DC: US Government Federal Highway Administration.

4 Nurshayeva, R. (2014). Update 1 – new pipelines to cost Kashagan oil project up to $3.6 bn. Reuters (10 October). http://www.reuters.com/article/oil‐kashagan‐idUSL6N0S52P420141010 (accessed 17 May 2017).

5 National Transportation Safety Board (2003). Pipeline Accident Report, Natural Gas Pipeline Rupture and Fire Near Carlsbad, New Mexico, 19 August 2000, NTSB/PAR‐03/01 (11 February 2003). Washington, DC: National Transportation Safety Board.

6 Javaherdashti, R. and Nikraz, H. (2010). A Global Warning on Corrosion and Environment: A New Look at Existing Technical and Managerial Strategies and Tactics. Saarbrucken, Germany: VDM Verlag.

7 Palmer, A.C. and King, R.A. (2008). Subsea Pipeline Engineering, 2e. Tulsa, OK: Pennwell.

8 Health and Safety Executive (HSE‐UK). ALARP at a glance. http://www.hse.gov.uk/risk/theory/alarpglance.htm (accessed 22 May 2017).

2

CHEMISTRY OF CORROSION

Corrosion, the degradation of a material due to reaction(s) with the environment, is usually, but not always, electrochemical in nature. For this reason, an understanding of basic electrochemistry is necessary to the understanding of corrosion. More detailed descriptions of all phenomena discussed in this chapter are available in many general corrosion textbooks [1–8].

ELECTROCHEMISTRY OF CORROSION

Most corrosion involves the oxidation of a metal that is accompanied by equivalent reduction reactions, which consume the electrons associated with the corrosion reaction. The overall corrosion reactions are often referred to separately as half‐cell reactions, but both oxidation and reduction are interrelated, and the electrical current of both anodes, where oxidation is prevalent, and cathodes, where reduction predominates, must be equal in order to conserve electrical charges in the overall system.

Electrochemical Reactions

A typical oxidation reaction for carbon steel would be

(2.1)

Common reduction reactions associated with corrosion include

(2.2)

Oxygen reduction

(2.3)

(2.4)

Metal ion reduction or deposition is also possible:

(2.5)

(2.6)

The reduction reaction is usually corrosion rate controlling, because of the low concentrations of the reducible species in most environments compared with the high concentration (essentially 100%) of the metal. As one example, the dissolved oxygen concentration in most air‐exposed surface waters is slightly lower than 10 ppm (parts per million). This relatively low dissolved oxygen concentration is usually much higher than the concentration of any other reducible species, and the control of air leakage into surface facilities is a primary means of controlling internal corrosion in topside equipment and piping.

More than one oxidation or reduction reaction may be occurring on a metal surface, e.g. if an alloy is corroding or if an aerated acid has high levels of dissolved oxygen in addition to the hydrogen ions of the acid.

Electrochemical reactions occur at anodes, locations of net oxidation reactions, and at cathodes, locations of net reduction reactions. These anodes and cathodes can be very close, for example, different metallurgical phases on a metal surface, or they can have wide separations, e.g. in electrochemical cells caused by differences in environment or galvanic cells between anodes and cathodes made of different materials.

Electrolyte Conductivity

The electrical conductivity of an environment is determined by the concentration of ions in the environment, and the resulting changes in corrosivity can be understood by considering Ohm’s law:

(2.7)

where

E = the potential difference between anode and cathode, measured in volts.

I = the electrical current, measured in amperes.

R = the resistance of the electrical circuit, determined by the distances between anode and cathode and by ρ, the resistivity of the electrolyte, which is usually expressed in ohm‐centimeters (Ω‐cm). In most cases the distance between anode and cathode is not known, but the changes in the corrosion rate can be monitored and correlated in changes in resistivity, e.g. the changes in resistivity of soils caused by changes in moisture content, which alter the ionic content of the soil electrolyte.

The resistivity (inverse of conductivity) of liquids and solids is determined by the ions dissolved in the bulk solution. Hydrocarbons such as crude oil, natural gas, and natural gas condensates are covalent in nature and very poor electrolytes, because they have very high resistivities. Oilfield corrosion is usually caused by chemicals in the water phase that, among other things, lower the natural resistivity of water, which is also mostly covalent. Water is a very efficient solvent for many chemicals, and most oilfield corrosion occurs when metal surfaces become wetted by continuous water phases having dissolved chemicals, which lower the natural high resistivity (low conductivity) of water.

Faraday’s Law of Electrolysis

The mass of metal lost due to anodic corrosion currents can be determined from Faraday’s law for electrolysis, Equation (2.8), which is also used by the electroplating industry:

(2.8)

where

Wcorroded = mass (weight) of corroded/electrodeposited metal.

F = Faraday’s constant.

i = current in amps.

t = time of current passage.

M = molar mass of the element in question.

n = ionic charge of the metal in question.

The amount of a substance consumed or produced at one of the electrodes in an electrolytic cell is directly proportional to the amount of electricity that passes through the cell. Methods of measuring the corrosion current are difficult and are discussed in Chapter 7.

Electrode Potentials and Current

The electromotive force series (EMF series) is an orderly arrangement of the relative standard potentials for pure metals in standard, unit activity (one normal, 1 N), solutions of their own ions (Table 2.1). The more active metals on this list tend to be corrosion susceptible, and the less active, or noble metals, will resist corrosion in many environments.

TABLE 2.1 The Electromotive Force Series for Selected Metals

Source: Adapted from Parker and Peattie [9].

It should be noted that two sign conventions are followed in publishing the EMF series. This can cause confusion, which can be avoided if the reader understands that active metals like magnesium and aluminum will always be anodic to carbon steel and corrosion‐resistant metals like silver and palladium will be cathodic.

The EMF series shows equilibrium potentials for pure metals in 1 N (one normal or unit activity of ions) solutions of their own ions. While this is the basis for much theoretical work in corrosion and other areas of electrochemistry, pure metals are seldom used in industry, and oilfield corrosive environments never have 1 N metal ion concentrations. The more practical galvanic series (Figure 2.1), which shows the relative corrosion potentials of many practical metals, is often used in corrosion control. This is based on experimental work in seawater and serves as the basis for many corrosion‐related designs [10].

Figure 2.1 Galvanic series in seawater.

Source: Reproduced with permission of John Wiley & Sons.

The galvanic series in seawater shown in Figure 2.1 is widely used for engineering designs. Some authorities claim that the relationships between various alloys must be determined for each environment, but this is seldom done. The reason for this precaution is that zinc and carbon steel undergo a polarity reversal in some fresh waters at approximately 60 °C (140 °F). The only other polarity reversal that has been reported is when tin, which would normally be cathodic to carbon steel, becomes anodic to carbon steel in deaerated organic acids, such as are found in the common tin cans used for food storage. It is unlikely that any other polarity reversals will be found in oilfield environments, and designers should assume that the relationships shown in Figure 2.1 are valid. Revie and Uhlig offer a brief review of polarity reversals [2].

The Nernst equation, first published in 1888 by the German chemist who later won the 1920 Nobel Prize in chemistry, explains how potentials of both anodic and cathodic reactions can be influenced by changes in the temperature and chemical compositions of the environment. The reduction potential can be expressed as

(2.9)

where

E =the electrochemical potential of the reaction in question.

E° =the standard electrode potential at 25 °C in a 1 N (normal) solution of the ion formed by oxidation of the reactants in question.

R =the Boltzmann distribution constant, normally referred to as the universal gas constant = 8.31(15) J K−1 mol−1.

T =the absolute temperature, °K.

n =the charge on the ion being reduced.

F =Faraday’s constant, the number of coulombs per gram‐mole of electrons = 9.63 × 10⁴ C mol−1.

At standard temperature conditions this equation can be simplified to

The details of this relationship are described in many general corrosion textbooks [1–7]. What is important to understand for oilfield corrosion control is that electrochemical cells (corrosion cells) can be caused by changes in:

Temperature

Chemical concentrations in the environment

Both types of electrochemical cells are important in oilfield corrosion and will be discussed further in Chapter 5.

It is simplistic to describe a chemical reaction as either oxidation or reduction. In actuality the reversible chemical reactions are happening in both directions simultaneously. The equilibrium potential, determined by the Nernst equation, is the potential where the oxidation and reduction currents, measured in current density on an electrode surface, are equal. The current density at this point is called the exchange current density. Some metals, e.g. the platinum and palladium used in impressed current anodes, have very high exchange current densities. This means that a small surface area of these materials can support much higher anodic currents than other anode materials such as high‐silicon cast iron or graphite. Figure 2.2 shows the idea of exchange current densities for hydrogen oxidation/reduction reactions. A platinum surface can support 10 000 times the current density of an iron anode for the same reaction. This increase in efficiency is used in the cathodic protection industry to justify the use of relatively expensive precious metal surfaces to replace much heavier, and therefore harder to install, high‐silicon cast iron anodes.

Figure 2.2 Hydrogen–hydrogen ion exchange current densities.

As potentials change from the equilibrium potential, the electrode surface becomes either an anode or a cathode. It is common to plot the shifts in potential on linear‐logarithmic plots, because in many cases these plots show a region of activation‐controlled electrode behavior where the voltage of anodes and cathodes follows a log‐linear pattern, called the Tafel slope after the German scientist who first explained this behavior in 1905.

On an anode, the Tafel equation can be stated as

(2.10)

where

ηa = the overpotential or change between the measured potential and the potential at the current density of interest. The subscript a indicates that this polarization is activation polarization, which occurs at low current densities near the equilibrium potential.

βa = the so‐called Tafel slope.

i = the current density, A m−2.

i0 = the exchange current density, A m−2.

At low electrode current densities, the change in potential can be plotted as shown in Figure 2.3. These plots of potential vs. logarithm of current are often termed Evans diagrams, after Professor U.R. Evans, of Cambridge University, who popularized their use [5].

Figure 2.3 Activation polarization of an iron electrode.

As stated above, most oilfield corrosion rates are controlled by the low concentrations of reducible species in the environment. These species must migrate, or diffuse, to the metal surface in order to react. The rate of this diffusion is controlled by the concentration of the diffusing species in the environment, the thickness of the boundary layer where this diffusion is occurring (largely determined by fluid flow or the lack thereof), temperature, and other considerations. The resulting concentration polarization can be written as

(2.11)

where

ɳc is the overpotential, or polarization, caused by the diffusion of reducible species to the metal surface.

F is the Faraday’s constant.

i is the current on the electrode.

iL is the limiting current density determined by the diffusivity of the reducible species; this is the maximum rate of reduction possible for a given corrosion system.

The other terms are the same as described above in discussions of the Nernst equation and activation polarization (Tafel slope) behavior.

Concentration polarization is shown in Figure 2.4. In corrosion, the limited concentrations of reducible species produce concentration polarization only at cathodes. At low current densities, the concentration polarization is negligible, and as the reduction current density approaches the limiting current, the slope quickly becomes a vertical downward line.

Figure 2.4 Concentration polarization curve for a reduction reaction.

The total polarization of an electrode is the sum of both the activation and concentration polarization. The combined polarization for a reduction reaction on a cathode is

(2.12)

This is shown in Figure 2.5.

Figure 2.5 Combined polarization curve for activation and concentration polarization on a cathode.

As stated earlier, most oilfield corrosion rates are determined by the concentration of the reducible chemicals in the environment. Figure 2.6a shows how the polarization of both the oxidation of a metal and the reduction of hydrogen ions determines the corrosion rate, icorr, and the corrosion potential, Ecorr, for a generic metal.

Figure 2.6 (a) Corrosion current and potential of iron determined by the polarization of iron and the hydrogen reduction reaction. (b) Corrosion current and potential of iron determined by the concentration polarization of oxygen.

Source: Beavers [11]. Reproduced with permission of NACE International.

For surface equipment, most corrosion rates are determined by the concentration of dissolved oxygen in whatever water is available. This is shown in Figure 2.6b, where the oxidation line showing Tafel behavior intersects the vertical (concentration limited) portion of the reduction reaction.

The importance of potential in determining corrosion rates is apparent from the above discussions. Academic chemistry reports tend to describe potentials relative to the standard hydrogen electrode (SHE), which has been arbitrarily set to a potential of zero. In field applications, it is common to use other reference electrodes. The most common reference electrodes used in oilfield work are the saturated copper–copper sulfate electrode (CSE), used in onshore applications, and the silver–silver chloride electrode used for offshore measurements, where contamination of the CSE electrode would produce variable readings. Table 2.2 shows conversion factors for these electrodes and other commonly used reference electrodes compared with the SHE. As an example, an electrode that measures −0.300 V vs. CSE would measure +0.018 V vs. SHE. Figure 2.7 shows a standard copper–CSE.

TABLE 2.2 Potential Values for Common Reference Electrodes

Source: Adapted from Jones [3].

Figure 2.7 Saturated copper–copper sulfate reference electrode.

CORROSION RATE EXPRESSIONS

Corrosion rates are measured in a number of ways:

Depth of penetration

Weight loss

Electrical current associated with corrosion

Time to failure

The simplest of these concepts to understand is depth of penetration. It can be expressed in mm yr−1 or mpy (mils or thousandths of an inch per year). The loss of wall thickness is often used to determine remaining equipment life or safe operating pressures for piping systems, storage tanks, etc. Table 2.3 shows a commonly used classification of relative corrosion rates. The US standard units, mpy, produce small numbers that are easy to understand, and corrosion rates in mpy are commonly used worldwide, although other expressions are also common [1].

TABLE 2.3 Relative Corrosion Resistance vs. Annual Penetration Rates

Source: Adapted from Fontana [1].

Weight loss measurements are commonly used on exposure samples used to monitor corrosion rates in oil and gas production. It is a simple matter to convert these weight loss measurements into average depths of penetration, although this can be very misleading, because most corrosion is localized in nature and the average penetration rate seldom gives an indication of the true condition of oilfield equipment.

The electrical current associated with anodic dissolution of a metal can be used to determine the corrosion rate using Faraday’s law. This calculation of mass loss can be converted into remaining thickness. Once again, the reader is cautioned that most corrosion is localized in nature and calculations assuming uniform loss of cross section are frequently misleading.

The time to failure, however defined, is the most common concern of managers and operators of equipment. For some forms of corrosion testing, e.g. stress corrosion cracking, the time to failure is used to screen alloys, environments, or other variables.

pH

The pH of an environment is one of the major factors determining if corrosion will occur. It also influences the type of corrosion that is experienced.

pH is defined as

(2.13)

where the [H+] expression shows the hydrogen ion activity of the environment. The [H+] depends on the ionization of water and varies with temperature. The pH of neutral water at standard temperature (25°C) is 7, but neutrality varies with temperature as shown in Figure 2.8. Downhole oilfield temperatures are usually elevated, and it is common to calculate the in situ pH of any fluids that might affect corrosion or scale deposition. There are many software packages available for this purpose. Figure 2.9 shows the effects of pH on the corrosion rates of iron in water. At low pH bare metal is exposed to the environment, and acid reduction on the surface controls corrosion rates. For intermediate pH a partially protective film of iron oxide reduces the corrosion rate and the diffusion of oxygen to cathodic locations on the metal surface controls. As the pH increases to even higher values, the surface becomes covered with mineral scales, and corrosion is reduced.

Figure 2.8 pH values of pure water at various temperatures [12].

Source: Reproduced with permission of John Wiley & Sons.

Figure 2.9 The effect of pH on the corrosion rate of iron in water at room temperature.

Source: Adapted from Revie and Uhlig [2].

PASSIVITY

Passivity is a phenomenon that is frequently misunderstood. Most metals form oxide films in most corrosive environments. These passive films can be protective and retard or even effectively stop corrosion, but they can also lead to fairly deep localized corrosion in situations where the protective films are removed or defective. Except in rare circumstances, the oxide films formed on carbon steel are not adequately protective, and other means of corrosion control are necessary. This is in contrast to stainless steels, titanium, and aluminum – oilfield metals that form protective passive films that are commonly the primary means of corrosion control for these alloys. On many corrosion‐resistant alloys such as stainless steels, the passive films may be only dozens of atoms thick. This means that they are very weak and subject to mechanical damage, and this can lead to localized corrosion at the damaged locations.

Potential‐pH (Pourbaix) Diagrams

Marcel Pourbaix developed a means of explaining the thermodynamics of corrosion systems by plotting regions of thermodynamic stability of metals and their reaction plots on potential vs. pH plots [12–14]. The regions of a Pourbaix diagram can be described as:

Immunity The metal cannot oxide or corrode (although it may still be subject to hydrogen embrittlement).

Corrosion Ions of the metal are thermodynamically stable and the metal will corrode.

Passivity Compounds of the metal and chemicals from the environment are thermodynamically stable, and the metal may be protected from corrosion if the passive film is adherent and protective.

Many users of Pourbaix diagrams miss the final point above. Thermodynamics alone cannot predict if passive films will be protective or not [2, 13–15].

Figure 2.10 shows the Pourbaix diagram for water. Water is thermodynamically stable over a potential region of 1.23 V, and the potentials at which oxidation and evolution (bubbling off) of oxygen from water at the top of the diagram or the evolution of hydrogen at the bottom of the diagram depend on the pH of the environment.

Figure 2.10 Potential‐pH (Pourbaix) diagram for water [13].

Source: Reproduced with permission of National Association of Corrosion Engineers.

The Pourbaix diagram of iron is superimposed on the diagram for water in Figure 2.11. Similar diagrams are available for most structural metals for which thermodynamic data are available [2, 13–15].

Figure 2.11 Potential‐pH (Pourbaix) diagram for iron [13].

Source: Reproduced with permission of National Association of Corrosion Engineers.

These diagrams make a number of important points useful for oilfield corrosion control:

Water is only stable over a potential range of slightly more than one volt. This is very important in cathodic protection.

Iron (carbon steel) is covered with iron oxides (passive films) in most aqueous environments. Unfortunately, these passive films are usually not sufficiently protective, and other means of corrosion control are necessary.

The potentials at which iron (carbon steel) is protected from corrosion do not coincide with the immunity regions on the Pourbaix diagram. This point is discussed in greater detail in Chapter 6.

The diagrams for zinc, aluminum, and cadmium, commonly referred to as the amphoteric coating metals, have passive regions in neutral environments. These metals also have low corrosion rates in neutral environments and higher corrosion rates in both acids and bases.

Pourbaix diagrams have limitations in addition to the inability of thermodynamics to predict the protectiveness of passive films. These include the idea that they cannot be calculated for alloys, although experimental Pourbaix diagrams have been reported [14, 15]. Revie and Uhlig list other limitations [2].

SUMMARY

The following ideas have been discussed in detail in this chapter:

Corrosion is electrochemical in nature.

Most metal surfaces have both oxidation and reduction occurring simultaneously.

If the predominant reaction is oxidation, the metal will corrode.

The most important reduction reaction is oxygen reduction for many oilfield systems. If no oxygen is available, the corrosion rate will often be very low.

Electrode potentials are determined by:

Metal chemistry

Chemicals in the environment

Temperature

These potentials are usually measured against either copper–copper sulfate or silver–silver chloride electrodes, depending on the environment.

Corrosion rates are often expressed by average depth of penetration, and this can be misleading because most oilfield corrosion is localized in nature.

The pH of the environment has a major effect on corrosivity.

Passive films may limit corrosion in many environments, but carbon steel, the most common oilfield metal, seldom forms adequately protective passive films, and other means of corrosion are often necessary.

REFERENCES

1 Fontana, M. (1986). Corrosion Engineering. New York: McGraw‐Hill.

2 Revie, W.R. and Uhlig, H.H. (2008). Corrosion and Corrosion Control. Hoboken, NJ: Wiley‐Interscience.

3 Jones, D.A. (1996). Principles and Prevention of Corrosion, 2. Upper Saddle River, NJ: Prentice‐Hall.

4 Perez, N. (2004). Electrochemistry and Corrosion Science. Dordrecht, The Netherlands: Kluwer Academic Publishers.

5 Ahmad, Z. (2006). Principles of Corrosion Engineering and Corrosion Control. Boston, MA: Elsevier.

6 Bradford, S.A. (2001). Corrosion Control, 2. Edmonton, Alberta, Canada: CASTI Publishing Inc.

7 Trethewey, K.R. and Chamberlin, J. (1995). Corrosion for Science and Engineering, 2. London, UK: Longman Scientific and Technical.

8 McCafferty, E. (2009). Introduction to Corrosion Science. Berlin: Springer.

9 Parker, M. and Peattie, E. (1984 [1995 reprinting]). Pipeline Corrosion and Cathodic Protection, 3. Houston, TX: Gulf Publishing.

10 LaQue, F.L. (1975). Marine Corrosion, Causes and Prevention. New York: Wiley.

11 Beavers, J. (2000). Chapter 15: Fundamentals of corrosion. In: Peabody’s Control of Pipeline Corrosion, 2 (ed. R. Bianchetti), 297–317. Houston, TX: NACE International.

12 Baboian, R. (ed.) (2002). Corrosion Engineer’s Reference Book, 3, 78. Houston, TX: NACE International.

13 Pourbaix, M. (1974). Atlas of Electrochemical Equilibria in Aqueous Solutions, 2e English. Houston, TX: NACE International, and Brussells: Centre Belge d’Etude de La Corrosion (CELBECOR).

14 Verink, E.D. Jr. (2000). Chapter 6: Simplified procedure for constructing Pourbaix diagrams. In: Uhlig’s Corrosion Handbook, 2 (ed. R.W. Revie), 111–124. New York: Wiley‐Interscience.

15 Thompson, W.T., Kaye, M.H., Bale, C.W., and Pelton, A.D. (2000). Chapter 7: Pourbaix diagrams for multielement systems. In: Uhlig’s Corrosion Handbook, 2 (ed. R.W. Revie), 125–136. New York: Wiley‐Interscience.

3

CORROSIVE ENVIRONMENTS

A very limited amount of oilfield corrosion is associated with very high‐temperature atmospheric exposures, common in flares, and with liquid metals, usually mercury found in natural gas and some crude oils. The great majority of oilfield corrosion requires liquid water. Downhole formation water that comes to the surface with oil and gas production can also include the following impurities that can affect corrosion rates [1–3]:

Oxygen – this is normally a problem only with surface equipment, because oxygen is unlikely to occur naturally in downhole formations.

Sulfur‐containing species.

Naturally occurring radioactive materials (NORM).

CO2.

H2S.

Fresh surface water is generally considered less corrosive than seawater or produced formation water. Designers will often assume that fresh water will require less stringent corrosion control efforts, but this can be a mistake, because surface water will usually have dissolved oxygen contents high enough to promote corrosion. Uncontaminated formation water, while usually very high in mineral content and very salty, becomes more corrosive if air (and oxygen) is allowed to enter. Oxygen scavengers are often used in produced water systems to limit the corrosion rates before produced water is reinjected downhole to maintain formation pressure.

For this reason, as an oilfield ages and the water cut increases, corrosion also increases. This is shown in Figure 3.1, which shows the effect of water cut on corrosion rates [4]. Many operators use rules of thumb such as the idea that corrosion is not a problem until the water cut reaches 40 or 50%. For some oilfields, this may take several years before corrosion becomes a problem. Unfortunately, this means that corrosion and other maintenance problems become more important at a time when maintenance funds, often related to production rates, decrease.

Figure 3.1 The effects of water cut on the corrosion rate of oil well tubing.

Source: Craig [4]. Reproduced with permission of NACE International.

Water has very limited solubility in hydrocarbons, and the presence of a separated water phase is necessary for corrosion. The low corrosion region in Figure 3.1 is where most of the metal surface is in contact with a water‐in‐oil emulsion. The small water droplets are not continuous, and most of the metal surface is in contact with nonconductive hydrocarbons. As the water cut increases, the amount of the metal surface in contact with water gradually increases until the emulsion reverses, and the liquid becomes continuous water with entrained hydrocarbon droplets. Production and fluid flow rates, along with temperature and pressure considerations, determine when this will happen. Figure 3.2 shows how water separates out on production tubing.

Figure 3.2 Water wetting producing corrosion on deviated oil wells.

Source: de Waard et al. [5]. Reproduced with permission of NACE International.

In contrast to oil wells, natural gas wells are corrosive from the beginning. This is due to the fact that all natural gas reservoirs will produce some water, and minor components of the natural gas, which condense from the gas stream as temperatures and pressures are reduced, dissolve in this water and make it corrosive. Condensed water lacks dissolved minerals, which could lower corrosion rates by buffering pH changes. Rainwater has enough dissolved CO2 to lower the pH to between 5 and 5.6, and the increased pressures in pipelines and downhole systems can lower the pH to even more acidic levels.

Most downhole hydrocarbon reservoirs have virtually no dissolved oxygen in the fluids, and this is fortunate, because the presence of oxygen at the parts‐per‐billion (ppb) level has been shown to promote corrosion. This is in contrast to carbon dioxide (CO2) and hydrogen sulfide (H2S), which may be present in varying quantities in both oil and gas fields. The relative effects of these three gases are shown in Figure 3.3. Oxygen is approximately 50 times more corrosive than CO2 and more than a hundred times more corrosive than H2S.

Figure 3.3 The effect of dissolved gases on the corrosion of carbon steel.

Source: From Shankardass [6].

Downhole corrosion, in the absence of oxygen, is largely determined by the concentrations of CO2 or H2S in the produced fluids. The terms sweet corrosion to describe corrosion caused by CO2 and sour corrosion to describe problems with H2S have been used for many years to differentiate which of these two gases is likely to predominate in a given field [1–3, 6–9]. Other considerations that affect corrosion rates include temperature and pressure, which determine the nature of the fluid (gas, liquid, etc.) on the metal surface, and minor constituents in the liquid water phase. Figure 3.4 shows how complex the determination of corrosivity can be.

Figure 3.4 A flowchart indicating the factors that determine the corrosion severity to be expected in an oil or gas field [10].

Source: Reproduced with permission of Corrosionsource.

EXTERNAL ENVIRONMENTS

The external environments discussed in this section are not unique to oil and gas production, but much of the information comes from oilfield experience with production platforms, buried or subsea pipelines, and similar equipment. External corrosion can affect all equipment, from the bottom of the well to the surface.

Atmospheric Corrosion

Like all other types of corrosion discussed in this book, this form of corrosion requires the presence of condensed water on the metal surface in order for corrosion to occur. The only exception to this general rule is at very elevated temperatures, e.g. those associated with flares and other combustion processes, where corrosion can occur without liquid water. Even here most corrosion occurs below the dew point, because these high‐temperature applications require special alloys to withstand corrosion at high temperatures, and this equipment often suffers the worst corrosion during shutdown, when acidic moisture can condense on rough surfaces and cause corrosion similar to that on automotive mufflers during times when the system is cold enough for condensation.

It would seem logical that atmospheric corrosion would not occur until the relative humidity is 100%, but this is not the case. Research dating back to the 1920s has shown that corrosion can occur once the humidity reaches a critical humidity of approximately 60–70% [11–13]. Many structures, especially on the away‐from‐the‐sun side (the north side in northern latitudes), stay above this critical humidity virtually all the time, at least whenever the temperature is above freezing [13]. It is important to realize that heat sinks, e.g. large structural members on offshore structures, can remain above the critical humidity long after the sun comes up and corrosion has diminished elsewhere on the same structure. The presence of deliquescent salts means that many surfaces remain wetted even in sunlight. Salt‐contaminated surfaces have been found to be wet below 20% relative humidity [14]. This is a very important consideration when painting structures, because flash rusting due to surface moisture can quickly form and severely degrade the adherence of primary coatings to painted structures (Figure 3.5).

Figure 3.5 Simplified diagram showing the effect of relative humidity and pollution on the corrosion of carbon steel.

Source: Revie and Uhlig [11]. Reproduced with permission of John Wiley & Sons.

Most oilfield metal exposed to atmospheric corrosion is carbon steel, and the most common method of corrosion control is by the use of protective coatings (painting) [11]. Some process equipment, storage tanks, and electronic control systems are protected by the use of inerting gases, heaters, deliquescing agents, or vapor phase inhibitors. Control lines, conduit, and similar tubing are often stainless steel on offshore structures.

The atmospheric corrosion exposure procedures described in international standards are of limited use for operating personnel, and they should only be used for potential coating systems evaluation [15–20]. Atmospheric corrosion is most severe in local areas on structures. The atmospheric exposure test panels shown in Figure 3.6 were intended to identify portions of locations a seaside petrochemical processing facility where corrosion damage inspections should be concentrated. The actual atmospheric corrosion on nearby equipment, as shown in Figure 3.7, was determined by the equipment and structure geometries and drainage patterns. The boldly exposed simple geometry exposure samples shown in Figure 3.6 cannot identify where the in‐plant corrosion inspection and maintenance should be concentrated. Accelerated tests in artificial atmospheres are useful for comparing prospective coating systems, but they cannot predict long‐term performance [21].

Figure 3.6 Atmospheric exposure panels at a large seaside petrochemical facility.

Figure 3.7 Localized atmospheric corrosion at the same facility shown in Figure 3.6.

NORSOK and other materials selection guidelines suggest that atmospheric corrosion of carbon steels should have protective coatings, but corrosion‐resistant alloys (CRAs) usually do not require protective coatings except under insulation or when submerged in seawater [21]. This NORSOK guidance is questioned by some operators having concerns with stress corrosion cracking (SCC) of stainless steels. The alternative to this practice is to use more alloys having better corrosion resistance than standard stainless steels, e.g. nickel‐based alloys, for atmospheric exposure temperatures above 60 °C.

Water as a Corrosive Environment

The effect of pH on corrosion of carbon steel was discussed in Chapter 2, Figure 2.9. Carbon steel, the most common metal used in oilfield systems, corrodes at unacceptable rates in many aqueous environments, and pH adjustment is a common means of controlling corrosion. The pH of natural surface waters is usually in the range between 4.5 and 8.5, and lower pH values, at which gaseous hydrogen evolution is the predominant reduction action, are not common in surface waters. Pure water is not corrosive in the absence of dissolved gases [22].

Most readers are familiar with the idea that salt water is more corrosive than fresh water. The combined effects of dissolved oxygen and salt concentration on the corrosivity of water are shown in Figure 3.8. As increasing amounts of salt are added to water, the electrical conductivity of the electrolyte increases and so does the corrosion rate. At the same time, the oxygen solubility decreases continuously with additional concentrations of salt, and this limits the corrosion rate because oxygen reduction is the rate‐controlling chemical (reduction) reaction [11]. The same phenomenon happens with all other salts. The maximum corrosion rate is at approximately 3% salt – the exact concentration depends on temperature and the salt involved [11, 23]. This explains why highly concentrated brines, such as those used in packer fluids, are noncorrosive, provided they are properly pH adjusted and have little or no dissolved oxygen.

Figure 3.8 The corrosion rate of iron in air‐exposed fresh water at varying salt (sodium chloride) concentrations.

Source: Adapted from Revie and Uhlig [11] and Uhlig [23].

Figure 3.8 shows that fresh water, low in salt, is less corrosive than salt water, but the most important point to be learned from this picture is that, even at its most corrosive, only about one‐third of the corrosion in salt water is due to salt – most of the corrosion would occur anyway due to the presence of oxygen. It should be noted that even deionized water can be corrosive if it is exposed to air [11, 24, 25]. Fresh surface water is generally considered less corrosive than seawater or produced formation water. Designers will often assume that fresh water will require less stringent corrosion control efforts, and one possible definition of fresh water will be water with less than 100 ppm chloride ions, although other definitions range from 50 to 500 ppm, dependent upon temperature.

Many reports on corrosion ascribe corrosion damage to the presence of chlorides, the most common anions found in seawater and often found in fresh water as well. This dates back to analytical chemistry practices in the early twentieth century, when qualitative analysis techniques (methods of determining the presence of various chemicals in the environment) were relatively new. The field methods for identifying chloride were relatively easy, and many authors started blaming chlorides for damage caused by salts. It was unnecessary to identify the other components of the salt, as there will always be cations (positively charged ions) present to balance the charge of the negatively charged anions. This practice continues. Any highly ionic salt would result in similar damage, but chloride salts are the most common in most natural environments. It is important to remember, as shown in Figure 3.8, that most of the corrosion in any location is due to the presence of dissolved oxygen or some other chemically reducible species (oxidizer). Salt cannot cause corrosion – it can only increase the corrosion rate by increasing the conductivity of the electrolyte.

Figure 3.9 shows the corrosion rates of piling in seawater at various elevations. The highest corrosion rates are in the splash zone, where the metal is frequently covered with air‐saturated water. The relatively low corrosion rates in the tidal region are due to the oxygen concentration cells between the highly aerated tidal zone and the fully submerged zone just below. The tidal zone, having high oxygen concentrations, is cathodic to the fully submerged zone just below, which is anodic. As the water deepens, the oxygen concentrations lower and corrosion decreases. NORSOK suggests an additional corrosion allowance (thicker metal) and the use of thick‐film protective coatings in the splash zone for both carbon steel and for martensitic (13Cr) stainless steels [21]. The same guidelines suggest Alloy 625 (UNS. N06625) and other nickel alloys with equal or higher pitting resistance, titanium alloys, or glass‐reinforced polymer composites for submerged service [21].

Figure 3.9 Zones of corrosion for steel piling in seawater.

Source: Adapted from Baboian and Treseder [26] and LaQue [27].

In locations with no tidal flows, the most corrosive location is at the air–water interface. This problem occurs on pilings (Figure 3.10) and in storage tanks and other equipment (Figure 3.11).

Figure 3.10 Waterline corrosion.

Figure 3.11 Leaking seawater filter vessel on an offshore platform due to waterline corrosion.

Produced water can vary from very salty, which is common in oil wells, to almost pure, the condensate associated with some gas wells. This pure water can become very corrosive, because the dissolved gases, CO2 and H2S, plus acidic hydrocarbons can drastically lower the pH, especially at downhole temperatures and pressures.

Most oilfield metal exposed to corrosive waters is carbon steel, and the most common method of corrosion control is by the use of protective coatings (painting), which is often supplemented by cathodic protection. Corrosion inhibitors and CRAs are also used, especially in downhole environments.

Soils as Corrosive Environments

Much has been written on corrosion in soils. The definitive work on this subject was published by M. Romanoff of the National Bureau of Standards (now the National Institute of Standards and Technology) in 1957, and, when the government report went out of print, it was republished by NACE [28]. Many advances have been made in the understanding of corrosion and cathodic protection since the original publication, but the data in this report represents one of the most extensive sources of corrosion in soil data that is available.

Water and gas occupy much of the space between the solid particles of soil, and these are very important in determining the corrosivity of soils. The air–water interface, wherever located, is the most corrosive location for buried structures, and this location often varies with seasonal rainfall patterns. The minerals in soil dissolve in water and affect the soil resistivity. This directly affects corrosivity, as shown in Table 3.1.

TABLE 3.1 Corrosivity Ratings Based on Soil Resistivity

Source: Bianchetti [29]. Reproduced with permission of NACE International.

Sandy soils drain well and tend to have the highest resistivities and lowest corrosion rates. Clays, that can swell when wetted, sometimes produce situations where drainage is prevented and buried structures remain wet and corrode.

Soil pH can also affect corrosion. Table 3.2 shows the effect of soil pH on corrosivity. Acidic soils are encountered in swampy locations, volcanic regions, and areas with silicate rocks and high moisture.

TABLE 3.2 Soil pH vs. Corrosivity

Source: Roberge [30]. Reproduced with permission of NACE International.

Some dry soil, especially clay‐rich soil, contracts during dry seasons as shown in Figure 3.12. This can lead to air ingress down to the buried structure, usually a pipeline, and lead to corrosion when rainy weather returns. Soil expansion and contraction can also cause movement of the buried structure. This produces stresses that can lead to SCC. More common is coating damage due to motion of the coated pipeline against rocks and other hard features in the trench.

Figure 3.12 Cracked soil due to drying after the rainy season.

Peabody, in his classic book on pipeline corrosion, cautioned against galvanic cells between new pipe and old pipe [31].There are a number of possibilities for galvanic cells to form when a new structure is placed in the soil adjacent to already‐buried structures. The most obvious reason for this corrosion is that the lack of soil compaction over the recently disturbed soil is more likely to leave void spaces and locations for enhanced air and moisture ingress. The new structure usually acts as an anode, indicating that increased moisture permeability is corrosion rate controlling.

Buried pipelines are in disturbed soil near soil that has been in place for many years. The differences in aeration and moisture are evident in the vegetation patterns over many pipelines. This is shown in Figure 3.13, which shows two obvious right‐of‐way locations, each of which has several parallel buried pipelines.

Figure 3.13 Differences in vegetation over two parallel pipeline rights of way.

Figure 3.14 shows how the corrosion rate of buried steel decreases with time. This is due to soil compaction and other poorly documented factors.

Figure 3.14 Variation of corrosion penetration rate as a function of buried steel exposure duration.

Source: Adapted from Ricker [32] and Logan [33].

The most corrosive location in any buried structure is usually where the structure crosses the air‐to‐soil interface. This is shown in Figure 3.15. It is important to concentrate inspections in these locations, because cathodic protection, which protects buried structures, cannot be effective in the loosely compacted soil at these locations. Abrasion, motion due to solar‐induced expansion and contraction, and a variety of other factors are likely to cause coating damage at these locations.

Figure 3.15 Corroded pipeline at the air‐to‐soil interface.

Virtually all oilfield equipment buried in soil is protected by a combination