Paper No.
06121 MATERIALS DESIGN STRATEGY: EFFECTS OF H2S/CO2 CORROSION ON MATERIALS SELECTION Bijan Kermani, KeyTech, Camberley, UK John Martin, BP Exploration, Sunbury on Thames, UK Khlefa Esaklul, BP Kuwait, Kuwait
ABSTRACT Corrosion remains a key obstacle to sustaining operational success in hydrocarbon production. Its continued occurrence affects the economy and has consequences for the safety of people and integrity of facilities. A central element in the design of facilities and corrosion mitigation is the correct choice and deployment of materials which are both economical and suitable to provide satisfactory performance over the design life. This paper captures the current understanding of corrosion mechanisms in the combined presence of H2S and CO2 acidic gases and discusses a systematic approach to materials design strategy for hydrocarbon production systems. The paper does not deal with the important environmental cracking aspects associated with sour service, but rather concentrates purely on metal loss degradation process. The combination of H2S and CO2 modifies the corrosion characteristics significantly as compared to damage caused in the sole presence of CO2 or H2S. An H2S/CO2 ratio is introduced to indicate the trends governing corrosion mechanism, i.e. dominated by CO2, H2S or a mixed mode of damage. A simple guideline has been produced offering a rule of thumb in addressing respective corrosion damages. Keywords:
Carbon and Low Alloy Steels, Corrosion, Integrity Management, Materials Selection Strategy, Production, Sour Service, Sweet Corrosion. INTRODUCTION
Corrosion in hydrocarbon systems manifests itself in several forms amongst which CO2 corrosion (sweet corrosion), H2S corrosion (sour corrosion) in the production systems and oxygen corrosion in water injection systems are by far the most prevalent forms of attack [1]. The environmental sensitive cracking damage caused by H2S and consequent materials optimisation are other very important aspects in these systems, but these are already covered in detail elsewhere [2, 3]. Corrosion in water injection systems is also outside the scope of the present overview. An Copyright ©2006 NACE International. Requests for permission to publish this manuscript in any form, in part or in whole must be in writing to NACE International, Conferences Division, 1440 South Creek Drive, Houston, Texas 77084. The material presented and the views expressed in this paper are solely those of the author(s) and are not necessarily endorsed by the Association. Printed in the U.S.A.
1
additional key element affecting corrosion is the presence of elemental sulfur in the production stream, which is again beyond the scope of the present paper. The majority of oilfield failures result from CO2 corrosion of carbon and low alloy steels (CLASs), primarily due to inadequate knowledge/predictive capability and the poor resistance of carbon steels to this type of attack [1]. Its understanding, prediction and control are key challenges to sound facilities design, operation and subsequent integrity assurance. Extensive research over the past five decades has focused on the mechanistic and engineering understanding of CO2 corrosion of CLASs, with a view to develop a realistic model to predict its occurrence. These are broadly covered in a review elsewhere [1]. Despite this, the majority of existing quantitative models remain unreliable in predicting the actual long-term CO2 corrosion rate of CLASs [1]. The anomalies are attributed to "field artefacts" with no clear indication of the cause. One key cause of the difference is now attributed to the effects of organic acid [4,5] a chemical normally ignored by many. An added complication is the presence of H2S which in turn affects potential corrosivity, insitu pH and interferes with the formation of corrosion product. This review article captures the current understanding and means of dealing with H2S in CO2 corrosion evaluations for CLASs in hydrocarbon production. It provides information on the mechanisms, highlights key parameters affecting the complementary influence of the two acid gases and draws attention to areas requiring further research. The primary focus has been placed on two key parameters affecting CO2 corrosion in the presence of H2S including (i) the nature of the surface film and (ii) development of an engineering guide for dealing with the risk of H2S-CO2 corrosion in production conditions. A brief overview of the specific material choice route for different production areas is provided within the context of a materials design strategy. The review has highlighted key areas of progress and has drawn attention to the future direction of research and development to enable improved and economical design of facilities for oil and a gas production. Background Both CO2 and H2S are acid gases that when produced with the hydrocarbon phase can render the associated water (condensed or formation) corrosive and lead to severe degradation. Corrosion resulting from each of these two acidic gases has its unique characteristics and, as a result, has received considerable industry attention, both to understand the corrosion mechanisms associated with the particular acid gas and the options available to mitigate the resulting corrosion [1-9]. Each of these gases occurs naturally in some of the producing reservoirs or may result from external contamination of the reservoir, such as the case of reservoir souring that may result when seawater is injected for secondary recovery or the use of gas injection for reservoir pressure maintenance. Selection of materials to combat corrosion relies mainly on the type of corrosion anticipated (e.g. whether general or localised [pitting]), the confidence in predicting the rate and type of corrosion, risk of failure and life cycle cost. While the primary concern in selection of materials in H2S containing systems is the sulfide stress cracking (SSC), the issue of corrosion should not be underestimated. SSC and other forms of cracking in H2S containing environments are well understood [2,3] and are not covered in this review. The focus of this paper is on the wastage corrosion in the combined presence of H2S and CO2. Relative Corrosiveness of CO2 and H2S and O2 While the respective corrosion mechanisms of the two acid gases prevailing in hydrocarbon systems, plus oxygen that can occur as a contaminant, are vastly different, a simple comparison under specific conditions was presented by Jones [10]. This is shown in Figure 1. The data are based on corrosion rates measured and computed by exposing clean carbon steel samples to water solutions containing various concentrations of each gas at 25oC. It has been claimed that these rates compare favourably with field data [10]. It is important to note that the synergistic effects of these gases are extremely influential in materials design and a point of consideration. A simple addition of respective damage rates does not necessarily lead to the overall damage as the complementary process is very complex. Furthermore, such a simple correlation does not bear in mind localised type of attack wherein the damage rate can be significantly higher than the overall
2
rate. However, the information provides a general idea of comparative corrosiveness of the three important gases at low temperature. TYPES OF CORROSION DAMAGE This section refers to the types of damage encountered in hydrocarbon production systems in CO2 only, H2S only and mixed CO2-H2S containing conditions. CO2 Containing Streams CO2 is usually present in produced fluids and although it does not generally by itself cause the catastrophic failure mode of cracking associated with H2S [2,3] its presence in contact with an aqueous phase can nevertheless result in very high corrosion rates, especially where the mode of attack is localised (e.g. mesa corrosion) [1]. CO2 Corrosion occurs primarily in the form of general corrosion and three variants of localised corrosion, i.e. pitting, mesa attack and flow-induced localised corrosion [1,11]: x x
x
Pitting corrosion normally occurs under relatively stagnant conditions (around the dew point for gas systems) – there are no certain rules to predict when such attack will occur. Mesa attack is a form of localised corrosion occurring under low to medium flow conditions (resulting from the localized removal of the protective carbonate film) – attack showing a large flat attack bottom steps with sharp edges– excessive rates at these areas occurs around temperature where carbonate can form but not stable Flow induced localised corrosion occurring at high flow conditions - corrosion takes the form of pits at sites of highly turbulent flow (often considered a form of erosion-corrosion)
CO2 corrosion is influenced by a number of parameters including environmental, physical and metallurgical variables [1]. The majority of these have been extensively covered by a number of authors and captured elsewhere [1,11]. Notable parameters affecting CO2 corrosion include: x
Fluid make-up as affected by water chemistry, organic acids, pH, water wetting, hydrocarbon characteristics and phase ratios
x
CO2 and H2S content (and possible oxygen contaminants)
x
Temperature
x
Steel surface including corrosion film morphology, presence of wax and ashphaltene
x
Fluid dynamics
x
Steel chemistry
All parameters are interdependent and can interact in many ways to influence CO2 corrosion as described elsewhere [1]. H2S Containing Streams; The Mechanism H2S results in a weak acid when dissolved in water. It affects CLASs in a similar manner to that of CO2 with all influential parameters outlined earlier for CO2 corrosion affecting its process and mechanism. The type of damage caused by H2S appears in the form of localised corrosion or general corrosion, depending upon the type and nature of corrosion product formed. H2S corrosion has been claimed to be strongly dependent on chloride ion concentration with severe damage rate in some situations, although the presence of other corrosive agents and fluid chemistry on this rate of degradation is unknown [12-15]. The corrosion reaction often leads to the formation of iron sulfide (FeS) scales, which under certain conditions are highly protective. However, their breakdown (i.e. under turbulent flow conditions) can lead to very severe localised corrosion in a similar manner to that for FeCO3 breakdown in the case of CO2 corrosion [1]. The kinetics and nature of FeS film formation, stability and its contribution to reducing corrosion are key to affording protection. Also, like CO2 corrosion, the corrosion rate is affected by fluid chemistry, organic acids and flow velocity in addition to the presence of elemental sulfur [13].
3
The ability of H2S to affect acidity is indicated by its ionisation as follows [11]: H2S ļ H+ + HS-
(1)
+
As the H is removed through cathodic reaction of hydrogen reduction, more is formed and hydrogen gas readily appears on steels exposed to oxygen free water containing H2S as follows: 2H+ + 2e ĺ 2H (atomic hydrogen) ĺ H2 (molecular hydrogen) (2) The anion (HS-) dissociates further to S2- and H+. The S2- ion reacts with iron to form the black FeS corrosion product commonly found in service. H2 may not be present in the bulk solution, but it forms locally within the corrosion layer as a cathodic corrosion product diffusing from its electrochemical production at the metal surface to its final dispersion in the bulk at the outer surface [16]. Mixed H2S/CO2 Containing Streams Ignoring the environmental sensitive cracking aspects of corrosion problems associated with sour service, low levels of hydrogen sulfide can affect CO2 corrosion in different ways. H2S can either increase CO2 corrosion by acting as a promoter of anodic dissolution through sulfide adsorption and affecting the pH or decrease sweet corrosion through the formation of a protective sulfide scale. The exact interaction of H2S on the anodic dissolution reactions in the presence of CO2 is not fully understood [1]. For similar conditions, oil and gas installations could experience lower corrosion rates in sour conditions compared to completely sweet systems. This is due to the fact that the acid created by the dissolution of hydrogen sulfide is about 3 times weaker than that of carbonic acids, but H2S gas is about 3 times more soluble in hydrocarbon phase than CO2 gas. As a result the effect of both CO2 and H2S gases on lowering the solution pH and potentially increasing corrosion rate are fundamentally the same. In addition, hydrogen sulfide may play a significant role on the type and properties of the corrosion films, improving or undermining them [1,11]. Many papers have been published on the interaction of H2S with CLASs. However, literature data on the interaction of H2S and CO2 is still limited since the nature of the interaction is highly complex. The majority of open literature indicates that CO2 corrosion rate is reduced in the presence of H2S at ambient temperatures. Nevertheless, it must be emphasised that H2S may also form a non-protective layer and that it may catalyse the anodic dissolution of bare steel [45]. Steels may experience some form of rapid, localised corrosion in the presence of H2S, although very little information is available. Published laboratory work has proved inconclusive, indicating that there is a need to carry out further studies in order to clarify the mechanism. In spite of the work on H2S corrosion of steels, no equations or models are available to predict corrosion, as is the case for CO2 corrosion of steels [1,11]. As a general rule in CO2 containing environments the presence of H2S can [1,16,45]: x
x
Increase the corrosion risk by either: o
facilitating localised corrosion, at a rate greater than the general metal loss or localised rate expected from CO2 corrosion, or
o
preferentially forming an FeS corrosion product that is less protective than an iron carbonate corrosion product
Decrease the corrosion risk by promoting the formation of an FeS corrosion product film through either o
replacing a less protective iron carbonate film, or
o
forming a combined protective layer of iron sulphide and iron carbonate
In the presence of both acid gases the corrosion process is governed by the dominant acid gas. The presence of H2S in CO2 containing producing environments has been reviewed by Pots et al [14]. They have introduced a notion of CO2/H2S ratio and considered three different corrosion domains based on the dominance of corrosion mechanism as affected by the dominating acid gas. These are tabulated in Table 1 and shown in Figure 2 [14] as follows:
4
x
CO2/H2S < 20 o
Corrosion dominated by H2S
x
o
Mixed CO2/H2S corrosion dominance
x
FeS as the main corrosion product
20 < CO2/H2S < 500 A mixture of FeS and FeCO3 as the main corrosion products
CO2/H2S > 500 o
CO2 corrosion dominates
FeCO3 as the main corrosion product.
These limits will be subject to environmental conditions highlighted in CO2 containing streams and described in a later section. This is in support of other investigations [17,18] in which it has been concluded that the CO2/H2S ratio determines the nature of scale and in turn the corrosion mechanism. Dunlop [17] proposed that as a general rule for CO2/H2S > 500, corrosion is dominated by CO2 and FeCO3 will form. When the ratio is < 20, FeS scale will form and H2S corrosion dominates as outlines in Figure 2 [14]. Smith [19] identified that the corrosion rate limiting step is determined by the type of corrosion product that forms as a result of the chemical reactions of CO2 and H2S. He goes on and describes the protectiveness or lack of it based on the thermodynamic stability of different compounds. Smith also developed a relationship that determines the boundaries of the FeCO3 and mackinawite scales by extending the Dunlop correlation and proposed the equilibrium boundary between mackinawite and FeCO3. It is difficult to extrapolate laboratory based data generated over a short period to real life corrosion reactions and describe kinetically driven corrosion processes in terms of thermodynamic data. It is worth noting that protective layers are always thin as they progressively reduce ionic transport and corrosion reaction, whereas non-protective layers are normally thick or even profuse [4,45]. Therefore, regarding a threshold in term of corrosion rate, the simple corrosion product layer thickness could be an indicator of the nature of sulfide formed, apart from thermodynamic data in potential-pH diagrams. The explanation for the disparity in protectiveness of corrosion product in CO2 and H2S containing media has been explained in terms of diffusion transport phenomena through the liquid phase within the corrosion product layer [16,45]. It is argued that depending on the dominant process, three possible types of corrosion control processes are plausible; soluble, insoluble cationic (IC) and insoluble anionic (IA) layers. He goes on to say that in H2S-containing media through the formation of IA-type deposit can explain the presence of a layer of highly soluble corrosion products, including FeCl2, between an outer layer of virtually insoluble FeS and a corroding steel substrate. It is said that this is also true for the very high corrosion rates observed beneath thick profuse mackinawite deposits. Furthermore, a protective IC layer formed in the presence of excess H2S can explain the formation of pyrite, FeS2, while a nonprotective IA layer, produced under a deficiency of H2S, can explain the formation of mackinawite, FeS1-e. Therefore, it is probably more correct to consider that it is the mechanism of protectiveness which determines the nature of the solid deposit, rather than the opposite [16,45]. The in-situ pH is also a key parameter governing corrosion in wet hydrocarbon production conditions affecting the formation and retainment of a protective layer. The in-situ pH is influenced by three controlling buffer systems [5]: x
CO2/HCO3- through reaction (3)
x
H2S/HS- through reaction (1)
x
HAc/Ac - (or other organic acids or other organic) through reaction (4) CO2 + H2O ļ HCO3- + H
+
(3)
5
HAc ļ Ac- + H+
(4)
The carbonic, sulfidic and acetic buffers are represented by the mass action laws of their respective dissociation equilibrium as outlined in reactions 1, 3 and 4. When several buffers are simultaneously present, they react together since H+ is a common species in their respective dissociation equilibriums (5) and (6): HAc + HCO3- ļ Ac- + CO2 + H2O -
(5)
-
H2S + HCO3 ļ HS + CO2 + H2O
(6)
These and the corresponding in-situ pH in turn influence the corrosion process. The synergistic interaction of these three buffering (and others) reactions govern corrosivity as influenced by the formation of protective scales and they should constitute the basis of any corrosion analysis in hydrocarbon production systems. KEY FACTORS INFLUENCING H2S/CO2 CORROSION The operational parameters affecting CO2-H2S corrosion include those outlined earlier for CO2 containing conditions, notably: x
CO2/H2S ratio
x
Temperature
x
Fluid Chemistry (water chemistry, pH, organic acids, water cut, oil wetability, phase ratios, etc.)
x
The hydrocarbon phase
x
Flow characteristics and fluid velocity
x
Steel surface, including corrosion products, scales, wax and asphaltene
x
Steel chemistry
These parameters, highlighted in Figure 3, are interdependent and can interact in many ways to influence the corrosion process and the exact influence of many of these is still unknown. In particular the effect of hydrocarbon phase on corrosion behaviour still remains unanswered, not only in the case of CO2 corrosion, but also when both H2S and CO2 are present [1,10,11]. However, an overview of the current understanding of these parameters is captured in this section. As explained earlier, the interaction of the buffering reactions is a key consideration in the corrosion process. Effect of Chlorides Bich [12] reported very high corrosion rates in the order of 30 mm/y, in a failed gas pipeline with high level of chlorides while other lines with low level of chlorides did not exhibit the high corrosion rates. The high corrosion rate was attributed to the effect of chlorides on breaking down of the protective FeS scale and initiation of pitting corrosion. However, this may have been related to the influence of other constituents of the solution affecting the in-situ pH or organic acids, the presence of which was not fully established. Foroulis [20] reported large increase in corrosion rate with an increase in chloride content in solutions saturated with H2S and suggested that the increase is due to the increase in conductivity and the interference of the chloride ions with the formation of FeS protective film. Agrawal et al [18] reported that there is strong correlation between the corrosion rate and the CO2/H2S ratio and the relation followed a bell-shape curve, with the peak corrosion rate occurring at an order of magnitude higher CO2/H2S ratio when the chlorides increased from 0.01 to 10% NaCl. This may suggest that chloride ions interfere with the formation of protective scales. Furthermore, they concluded that any damage in the protective film can lead to an accelerated corrosion unless and until the FeS protective film is reformed. Hence, the role that chloride content of the environment plays is not fully established. However, its influence on corrosion of carbon and low alloy steels in CO2-H2S containing media is often
6
considered insignificant at low chloride contents. Nevertheless, chloride may have a role in affecting the in-situ pH. Chloride also has some affinity to interfere with the formation of protective FeCO3 or pyrite and the tendency to influence its formation and growth on CLASs [4,16]. The observations reported on the effect of chloride [12,20] may have been related to the influence of other constituents of the media affecting the in-situ pH or organic acids the presence of which were not fully reported. A critical chloride concentration of 10,000 ppm has been conservatively proposed [14] based on field experience and laboratory testing. It has been concluded that above a concentration of 10,000 ppm, the chloride ion can destroy the protective FeS scales and can lead to increased corrosion rate. The role of chloride needs further clarification and a subject which should be taken in the context of solution chemistry and the nature of corrosion product. Effect of H2S Concentration Russ and Rainsford [21] reported that change in the CO2/H2S ratio from 3 mole % CO2/700 ppm H2S to 3.75 mole % CO2/350 ppm H2S resulted in significant increase in corrosion of oil pipelines. In-line inspection revealed that the section of the line with CO2/H2S ratio of 43 had low or non detectable corrosion while the one that had a ratio of 88 showed localized area representing 2.5% of the length of the line with pits 20-35% of the wall thickness. The difference in flow rate expressed as the shear stress was reported as 0.2 Pa for the segment with no corrosion and 1.0 Pa in the segment that showed corrosion. These shear stresses were low so that corrosion could not be due to the increase in fluid velocity. Again the full fluid chemistry has not been reported. Smith and de Waard [22] proposed an H2S reduction factor in their corrosion rate prediction model for CO2 corrosion. The factor is a function of the partial pressure ratio of H2S and CO2 as follows: F
H2S
= 1 / (1 + 1800 (pH2S/pCO2))
(7)
They state that this factor is speculative since the protective FeS layer can suffer breakdown. In cases where FeS breakdown occurs the corrosion rate can be an order of magnitude higher than the corresponding rate for pure CO2. This high corrosion rate in the presence of H2S is a result of drop in the pH due to the reduction of the dissolved iron ions that occurs with FeS precipitation and galvanic couple formed between the steel and corrosion scale. Brown, Parakala and Nesic [23] studied the effect of low levels of H2S on CO2 corrosion and showed that in the absence of protective iron carbonate and iron sulfide scales very small amounts of H2S < 10 ppm in the gas phase can lead to rapid and significant reduction in the CO2 corrosion rate. The trend is arrested and somewhat reversed at higher H2S concentration. Protective adherent films formed at 60º C with 25 ppm H2S and 7.9 bar pressure at pH of 6.0. Effects of Low H2S Partial Pressures on CO2 Corrosion Rates Recognising the importance of trace H2S on CO2 corrosion, a limited number of tests were carried out using an ambient pressure corrosion loop facility [24]. The programme assessed the effects of low H2S partial pressures on CO2 corrosion. The partial pressures of H2S used were 1.5x10-3, 1.5x10-2 and 1.5x10-1 psi (0.0001, 0.001 and 0.01 bar) corresponding to H2S concentrations of 100, 1,000 and 10,000 ppm (in the gas phase), respectively at an atmospheric pressure. Tests were carried out on linepipe steel grade X65 at 30, 50 and 75oC in the presence of 1 bar CO2 (14.5 psia) under two environmental conditions as shown in Table 1. Corrosion rates were determined by continuous LPR monitoring over a 24 hours’ period [24]. The results are summarised in Figures 4 and 5 for two conditions: x
Chloride containing fluid with no buffering agent at a pH range of 3.8 to 4.0
x
Formation water with strong buffering agents at a pH range of pH Range 5.5 to 5.8
Also included are the predicted corrosion rates for these conditions, using the de Waard and Milliams model [25]. The respective values under sweet conditions are somewhat higher than those predicted by the model, particularly at lower temperatures.
7
Despite the inability to extrapolate the outcome of this limited study to develop a trend of corrosion rate versus H2S, it can be seen that low levels of H2S has a strong influence on CO2 corrosion by lowering the general rate of attack irrespective the environmental conditions. The lowering of damage rate increased progressively with increasing test temperature, although the magnitude of the reduction in corrosion rates was smaller in the buffered solution. The reduction in corrosion rates were considered to be due to the formation of a semi-protective FeS film or by stabilising an iron carbonate film. It seems that the partial pressure range investigated was above that required for semi-protective films to form but below that necessary to cause increased corrosion rates. While the degree of damage by localised corrosion was examined, the results were not conclusive. Nevertheless, the rate of pit propagation was considered unlikely to exceed the predicted general surface corrosion rate predicted determined by CO2 corrosion rate prediction models. The results indicate that at low levels of H2S which are defined as levels generally below the ‘occurrence of SSC’ limits (i.e. Region 0 of ISO 15156/NACE MR0175 at <0.05 psi) can reduce corrosion by a factor of 3 to 4 [24]. Effect of Flow and Temperature Brown and Nesic [26] showed that mackinawite FeS scale is protective when the flow conditions do not disturb the FeS scale but becomes less effective under turbulence flow conditions. Under conditions of porous film formation, the scale becomes less effective barrier and the corrosion rate increases. The key parameters in determining the film morphology are the FeCO3 and FeS supersaturation values. It has been claimed that the mechanisms of corrosion product film growth can produce multilayer films with some leading to increased corrosion rate and probability of localised corrosion [26]. Omar et al. [27] showed that CO2-H2S corrosion is not dependent on temperature and flow velocity for test conditions of 25 – 80ºC and 1 – 5 m/sec and the measured corrosion rates were lower than the predicted corrosion rates from pure CO2. At the highest velocity 5 m/sec they reported a tendency toward localised corrosion attacks and development of pitting corrosion which they attributed to the local breakdown of the protective films due to the high shear stresses or flow eddies close to the film. The magnitude of shear stress at this flow velocity is near 77 Pa. This raises the issue of pitting corrosion under conditions where film breakdown may occur due to high flow rates. In addition, they reported that at 80º C multiple layers of corrosion product formed with the inner layer close to the steel contained substantial fractions of iron oxides. The oxides would typically form due to the limited sulfide ions indicating that the corrosion products act as diffusion barriers and slow the corrosion attack. The effect of iron oxide presence on the long term integrity of the FeS protective layer is a major concern, although its presence may be related to laboratory test conditions as it is not a field produced corrosion product in CO2 environments. Effect of Organic Acids The significance role of organic acids on the performance of CLASs in CO2 containing hydrocarbon production systems is gaining due considerations over the past few years [4,28-31]. This remains a challenging topic as various mechanisms prevails and the subject is complex and still under extensive scrutiny. While in gas producing wells, the free acetic acid (HAc) content is physically dissolved from the produced gas phase, in oil producing conditions free HAc is chemically produced from the reaction between dissolved CO2 and the acetate ions present in water. In general term, organic acids change the solubility of corrosion product (i.e. dissolved iron, Fesat, at the saturation of corrosion product in FeCO3) and hence interfere with the formation of FeCO3 protective layer. This has a strong influence on CO2 corrosion and hence affects corrosivity assessment and prediction [4]. The synergistic interaction of the buffering reactions described earlier governs corrosivity. In CO2H2S containing systems, the effect or organic acids is not yet fully known, although their role is
8
expected to be similar to the effect in CO2 systems and need to be accounted for in both chemical analyses as well as corrosion and predictions and mitigation. Effect of Microstructure There is limited data on the true effect of metal surfaces and steel microstructure on the corrosion behaviour of H2S-CO2 corrosion. Data reported by Perdomo et. al. [32] showed different behaviour of API 5L X52 steel from API 5L Grade B. The corrosion rate for Grade B increased with increase in H2S, reached a maximum then decreased where for the X 52 steel the corrosion rate continued to increase for the ranges of H2S tested. The difference has been attributed to the less compact and less uniform FeS layer on the X52 steel. However, new generation of low carbon microalloyed 3%Cr steel, has shown improved properties compared with conventional grades of CLASs with enhanced CO2, H2S and O2 corrosion resistance, satisfactory sulphide stress cracking (SSC) performance together with more than acceptable mechanical properties [33-34]. Effect of Corrosion Product Scales Smith and de Waard [35] reported the effectiveness of the corrosion product protective scales that form even with small amount of H2S. The corrosion is effectively mitigated with the protective scales when combined with corrosion inhibitors. They reported corrosion rates below 0.05 mm/y in the surface facilities of 25 gas fields that had H2S/CO2 ratio in the range of 1 to 750 where downhole inhibitors injection was used. The piping and vessels were coated with a sulfide rich film even in areas where inhibitors carryover was expected to be minimal. In support of other studies [14,17] Srinivasan and Kane [36] identified that the effect of H2S on CO2 corrosion depends on the ratio of CO2 to H2S. At low levels of H2S (< 0.01 psia) H2S has minimal or no influence on corrosion. At small amounts of H2S (CO2/H2S > 200) protective iron sulfide films form and reduce corrosion. Below 120º C, the dominant film is mackinawite and its formation depends on pH and temperature. In conditions where H2S is the dominant acid gas (partial pressure CO2/H2S < 200) meta-stable iron sulfide films form preferentially over FeCO3 scale in the range of 60 to 240º C. The protective film is initially mackinawite and at higher H2S concentration and temperature, the more stable pyrhotite iron sulfide forms which is more protective. At below 60 C and above 240º C the presence of H2S exacerbates corrosion since H2S prevents the formation of the protective FeCO3 scale and the FeS scale becomes unstable and porous. Pots et al [14] reported that testing at partial pressure ratio between 20 and 500 revealed that the highest pitting corrosion rate was never worse than the sweet corrosion rate. The corrosion product films were a mixture of FeCO3 and FeS. Irrespective of the CO2/H2S ratio, pyrite is the most stable ferrous sulfide in pure H2S or in mixed H2S/CO2.as seen in field removed samples. However, its stability may be locally jeopardized by generation of cathodically produced molecular hydrogen. Therefore, the degree of hydrogen evolution affects the nature of corrosion product in H2S containing conditions in which high H2 evolution results in high corrosion rate as it does not allow a protective layer to form readily. It is apparent that hydrogen transport needs to be taken on board when addressing protectiveness of FeS layer. This is due to the fact that transport in the surface layer governs the liquid surface state, which in turn governs both corrosion rate and the solid surface state present in-situ or observed afterwards. A similar analogy has been made on the possible multiple steady state corrosion rates in CO2 corrosion [37, 45]. Notable Remarks The data in the literature and field experience clearly indicate that, in certain conditions the presence of H2S leads to the formation and growth of FeS protective scale and decreases the corrosion rate. However, there is substantial evidence that this protective scale behaves in a similar manner to FeCO3 in CO2 containing fluids [37,45] wherein protectiveness is not universal and certain conditions render it ineffective resulting in severe localised corrosion. The corrosion rates under these conditions can be significantly higher than the corrosion rates for CO2 corrosion
9
either measured or predicted [12,13,41-45]. The primary concern with the presence of H2S is the potential failure of the FeS protective scale and the risk of high pitting corrosion rate, as described earlier. The evidence also suggests that the data generated to date is still not sufficient to characterise the issues related to H2S-CO2 corrosion and the general understanding that in the presence of H2S, corrosion increases initially until the FeS scale forms and then the corrosion rate decreases significantly. This behaviour is not well defined such that it cannot be relied on for the selection of materials since conditions may vary within the facility or during its operating life. Nevertheless, the data in the literature clearly shows that corrosion inhibitors can be effective in controlling H2S-CO2 corrosion if/when the correct inhibitor is selected to match the operating conditions and is effectively deployed. Therefore, H2S-CO2 corrosion can be effectively controlled with inhibition but requires that inhibitor selection and application be specific for the environment of concern and not be based on pure CO2 or H2S. MATERIALS SELECTION STRATEGY There is a growing desire to have a corrosion design philosophy for production facilities to transport wet hydrocarbons. Such an approach can be used in the technical/commercial assessment of new field development and in prospect evaluation and for handling of sour fluids by facilities not normally designed for sour service [39]. A universal method of preventing oilfield corrosion is through selection of the most appropriate material for a specific application. Optimum choice of materials is governed by a number of key parameters including adequate mechanical properties, corrosion performance, weldability (where appropriate), availability and cost. The choice of material is governed by the nature of its application and generally falls into two categories of production and injection. A simple chart outlining the necessary steps in selection of materials is shown in Figure 6 [39]. This methodology captures a systematic approach to determining potential deployment of CLASs as the first reference point. Having established the degradation rate of CLASs, the second, although key parameter, is its resistance to SSC in the presence of H2S. The majority of these steps are covered elsewhere [1,2,39] and this section briefly outlines specific measures required to allow selection of the most appropriate materials for a particular duty. In general there is no consensus on applicability of corrosion models to H2S-CO2 corrosion [38]. Some of the data gathered both in the field and in the laboratory indicate that the corrosion rate for H2S-CO2 corrosion is lower than the predicted corrosion rate for CO2 alone with localised corrosion rate rarely exceeding CO2 corrosion predicted rates. This suggests that use of CO2 corrosion prediction models, although conservative, may provide good estimate for the maximum corrosion rate expected [12,22,24,36] when FeS scale is formed. Bearing these in mind, simple rules for the prediction of corrosion damage rate in H2S-CO2 containing streams are included in Table 2. It should be noted that a major consideration in materials design strategy is careful attention to acid flow back. In these conditions, the effluents’ property may render it highly corrosive if not neutralised. It can contain very high chloride brine with low pH fluids and additionally unknown fluid chemistry. The prevailing FeS or FeCO3 scales might not be stable under such fluid chemistry. MATERIALS ROUTES A brief summary of materials selection route for different aspects of production is given in this section acting as an overview guide in a holistic materials selection strategy. It is important to note that while CLASs are chosen primarily based on their corrosion resistance with adequate resistance against SSC, CRAs are normally selected based on their resistance to environmental cracking with secondary consideration to their general corrosion behaviour. These include SSC and Cl-SCC (chloride stress corrosion cracking) or a combination of these types of damage as affected by the operating temperatures [2,3]. The exception to this overview for CRAs is under extreme
10
conditions (a combination of high temperature, low pH, high CO2 and H2S) where general corrosion may have to be considered in the overall selection strategy. Downhole and Well Heads Selection of materials for these applications are typically controlled by the need for resistance to both corrosion and to SSC. The latter is important even at low levels of H2S due to the high pressure and the need for long term reliability to avoid potential safety risks and unnecessary workover costs. Uses of low alloy steels with continuous inhibition for low to moderate corrosive conditions have proven to be successful in certain conditions. However, such systems can often prove impractical or too costly (e.g. deepwater subsea developments) such that they are not always the best approach. For highly corrosive conditions CRAs remain the most effective and economic option. Flowlines and Unprocessed Fluids Pipelines Two scenarios are considered here: i.
Highly Corrosive or High Risk applications For these applications CRAs often remain the most cost effective option since the risk of corrosion failure is high and use of corrosion inhibition with carbon and alloy steels is often either impractical, costly or poses too high a risk.
ii.
Low to Moderate Corrosiveness or Low Risk applications CLAS with corrosion inhibition or pH stabilisation is an effective option. The corrosion inhibitor must be selected appropriately in accordance with field conditions and operating parameters. On-line corrosion monitoring and frequent pigging may be required to ensure effective inhibition and inhibitor replenishments particularly where deposits may drop or accumulate within the flowlines (velocity of the fluids below the entrained velocity). Nonmetallic liners have been used successfully to reduce failures and inhibition cost with HDPE and special grades of Nylon (Rilsan), although the limits of applicability for such liners needs to be taken into account.
Process Facilities In general, process facilities will require the same materials selection and corrosion mitigation strategy as pure H2S with vessels requiring internal corrosion barriers to prevent under deposit corrosion and corrosion inhibition for carbon steel components. Selection of organic coatings vs. CRA cladding will depend on the corrosivity of the processed fluids and the potential impact of H2S presence on degradation of organic coatings such as blistering, etc. Heat exchangers particularly gas coolers will require CRA material at minimum for the heat exchanger tubes and heads or shell depending on the design of the unit. Process piping can be either CRAs or CLASs with corrosion inhibition (depending upon the piping configuration, quantities, etc,) except for high temperature areas (> 100oC) where CRAs are more suitable. Non-metallic materials such as HDPE, FRP and lined piping are practical alternatives for produced water handling and are becoming more widely used. Gas Treating Plants Except for the high corrosive sections of the plants such as amine towers, glycol reboilers and gas coolers where cladding or solid CRAs is required, carbon and alloy steels with corrosion inhibition are acceptable options. Export Pipelines Export pipelines for systems with significant level of H2S and CO2 requires the use of inhibited CLASs to mitigate any corrosion that may occur.
11
Seals and Elastomers H2S even at low concentration can cause severe degradation to elastomers and increases the propensity for elastomers embrittlement and explosive decompression. Selection of seals and elastomers for H2S/CO2 environments must follow the same selection guidelines for pure H2S environments. CONCLUSIONS The review and analyses captured in the present paper demonstrate that the current understanding of combined H2S-CO2 corrosion is far from complete and many of the issues affecting its occurrence remain unresolved. The current state of knowledge points to the following conclusions: 1.
In the combined presence of CO2 and H2S, there is a competitive interaction between FeCO3 and FeS corrosion products that may lead to affording protection or breaching the layers with resultant progressive localised corrosion
2.
Subject to the type and nature of the corrosion product, H2S may lead to an increase in CO2 corrosion until certain concentration threshold after which weight loss corrosion may be reduced and in many cases results in a significant reduction. The integrity of the FeS protective layer may be affected by the operating conditions. In this, it is more accurate to consider that it is the mechanism of protectiveness which determines the nature of the solid deposit, rather than the opposite
3.
The CO2/H2S ratio is an acceptable means of categorising metal loss corrosion damage caused in the combined presence of H2S and CO2 – this ratio affects the nature of corrosion product and together with other key operational parameters can be considered in corrosion prediction models
4.
A systematic materials optimisation strategy has been introduced integrating key parameters of past successes, present understanding of corrosion processes in hydrocarbon production, whole life costing and application regime of conventional as well as proprietary grades hence allowing the selection of the most suitable, safe and economical material option and corrosion control procedures
5.
H2S-CO2 corrosion damage can be mitigated with a correct materials selection strategy and implementation of corrosion control measures. In this, appropriate corrosion inhibitors have shown effective mitigating measures
6.
While CLASs are chosen primarily based on their corrosion resistance with adequate resistance against SSC, CRAs are normally selected based on their resistance to environmental cracking with secondary considerations to their general corrosion behaviour
7.
There remains a need to develop clear understanding of H2S-CO2 corrosion process and the interaction of other key environmental, metallurgical and hydrodynamic parameters affecting the phenomenon and the formation of corrosion products. The interaction between these parameters is a key to determining their accumulative effect and means of mitigation through effective materials selection and corrosion control strategy. ACKNOWLEDGEMENTS
Thanks are due to Mr Dominic Paisley (BP) for his contribution to the development of this paper. Valuable comments and contributions from Dr Jean Louis Crolet (Consultant) and Mr Don Harrop (BP) are highly appreciated. REFERENCES 1. 2.
M B Kermani and A Morshed, Corrosion Vol. 59 No. 8, 2003, p.659 – 683. M B Kermani, Paper No 00156, NACE, Orlando, March 2000
12
3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.
Petroleum and natural gas industries – Materials for use in H2S containing environments in oil and gas production ISO 15156, Parts 1-3, 2003 M R Bonis and J L Crolet, NACE Annual Corrosion Conference, Paper 05272, 2005 J L Crolet, Eurocorr 2004, (London UK; The Institute of Materials, 2004) Metals Handbook, “Corrosion”, Volume 13 (ASM International, Materials Park, Ohio, 1987), p. 1233. B Craig, SPE Monograph Volume 15, Society of petroleum Engineers, Richardson, Texas, p. 22-29, 1993 R H Hausler, NACE Annual Corrosion Conference Paper No. 04732, 2004 R.N. Tuttle, Journal of Petroleum Technology, p 756-762, 1987 L W Jones, Corrosion and Water Technology for Petroleum Producers; OGCI, Tulsa, 1988 CO2 Corrosion Control in Oil and Gas Production - Design Considerations, eds. M B Kermani and L M Smith, European Federation of Corrosion Publication No 23, 1997. N N Bich and K Goerz, NACE Annual Corrosion Conference, Paper No 26, 1996 N N Bich, “Fundamental of Wet Sour Gas Corrosion”, Presentation/Private communication, 1999 F M Pots, R C John, I J Rippon, M J J Simon Thomas, S D Kapusta, M M Girgis and T Whitman, NACE Annual Corrosion Conference, Paper 02235, 2002 S N Smith and R.S. Pakalapati, NACE Annual Corrosion Conference, Paper 04744, 2004 J L Crolet, private communications, 2005 A K Dunlop, H L Hassell and P R Rhodes, NACE Annual Corrosion Conference, Paper No 46, 1983 A K Agrawal, C Durr and G H Koch, NACE Annual Corrosion Conference, Paper 04383, 2004 S N Smith and J L Pacheco, NACE Annual Corrosion Conference, Paper 02241, 2002 Z Foroulis, Werkstoffe and Korrosion, pp.463-470,1980 P R Russ and C Rainsford, SPE Asia Pacific Oil and Gas Conference and Exhibition, Paper No 88570, October 2004 L M Smith and C de Waard, NACE Annual Corrosion Conference, Paper 05648, 2005 B Brown, S R Parakala and S Nesic, NACE Annual Corrosion Conference, Paper 04736, 2004 D Paisley, BP Internal Report, 1993 C de Waard and U Lotz, NACE Annual Corrosion Conference, Paper No. 69, 1993 B Brown and S Nesic, NACE Annual Corrosion Conference, Paper 05625, 2005 I H Omar, Y M Gunaltun, J Kvarekval and A Dugstad, NACE Annual Corrosion Conference, Paper 05300, 2005 J A Dougherty, NACE Annual Corrosion Conference, Paper 04376, 2004 M W Joosten, J Kolts, J W Hembree and M Achour, NACE Annual Corrosion Conference, Paper 02294, 2004 M W Joosten, G D Harris, R L Hudgins, D A Daniels, and K M Cloke, NACE Annual Corrosion Conference, Paper 05114, Corrosion 2005 W M Hedges and L McVeigh, NACE Annual Corrosion Conference, Paper No 21, 1999 J J Perdomo, J L Morales, A Viloria and A J Lusinchi, Materials Performance, pp 54-58, March 2002 M B Kermani, J C Gonzales, G L Turconi, T Perez and C Morales, NACE Annual Corrosion Conference, Paper 04111, 2004 L Pigliacampo, J C Gonzales, G L Turconi, T Peres, C Morales and M B Kermani, NACE Annual Corrosion Conference, Paper 06133, 2006 L M Smith and C de Waard, Industrial Corrosion, p 14-18, 2004. S Srinivasan and R D Kane, NACE Annual Corrosion Conference, Paper No. 11, 1996 J L Crolet, S Olsen, W Wilhelmsen, NACE Annual Corrosion Conference, Paper 127, 1995 R Nyborg, NACE Annual Corrosion Conference, Paper 02233, 2002 M B Kermani, J C Gonzales, G L Turconi, T Perez and C Morales, NACE Annual Corrosion Conference, Paper 05111, 2005 J L Crolet and M R Bonis, SPE Production Engineering, pp. 449-453, Nov, 1991 A Ikeda, M Ueda and S Mukai, Advances in CO2 Corrosion Volume 2, pp 1-22, NACE International, 1985 K. Videm and J. Kvarekval, NACE Annual Corrosion Conference, Paper No. 94012, 1994 J Kvarekval, EUROCORROSION 97, Trondheim, Norway. A Valdes, R Case, M. Ramirez and A. Ruiz, NACE Annual Corrosion Conference, Paper No. 22, 1998 J L Crolet, in “Modelling Aqueous Corrosion From Individual Pits to System Management”, ed. K R Trethewey and P R Roberge, NATO ASI Series, Series E: Applied Sciences, Vol 266, pp 1-28, 1994.
13
Table 1 The analysis of the synthetic produced waters used in the tests Component
Formation Water (ppm)
Chloride SO4-Bicarbonate Sodium Potassium Calcium Magnesium Acetate pH
52,000 10 500 29,500 380 3200 500 50 5.5-5.8
Chloride Containing Fluid (ppm) 42500
27540
3.8-4.0
Table 2 H2S-CO2 Corrosion Dominance and Prediction Guides (A Rule of Thumb)
CO2/H2S Ratio
< 20
Sub category
-
Operating Parameters
Dominating Corrosion Process
Primary Corrosion Product
Low: Subject to the formation of a protective FeS Mixed - the highest localised corrosion rate does not exceed predicted sweet corrosion rate.
Known
H2 S
FeS
Known
Mixed H2S/CO2
FeS and FeCO3
Fully known
Mixed H2S/CO2
FeS and FeCO3
Mixed – Corrosion rate determined by the nature of FeS
Known
CO2
FeCO3
CO2 driven
1000
Fully known
CO2
FeCO3
CO2 driven
10000
Fully known
CO2
FeCO3
CO2 driven
>10000
Fully known
CO2
FeCO3
CO2 driven
20 to 500
100
> 500
14
Corrosion Damage Risk Factor (A Rule of Thumb)
Possible Pattern of Corrosion Damage
-
CO2 Corrosion Model CO2 Corrosion Model with possible reduction of 4 CO2 Corrosion Model with possible reduction of 4 CO2 Corrosion Model with possible reduction of 3 CO2 Corrosion Model with possible reduction of 3 CO2 Corrosion Model
Figure 1.
Comparative corrosiveness of three common gases in water solutions (25oC, 5-7 day exposure, 2-5 g/litre NaCl, HCO3 alkalinity < 50 mg/l – computed from several data sources) [after Ref 11].
Figure 2. CO2-H2S corrosion domains [after Ref 14].
15
Figure 3. Parameters affecting CO2-H2S corrosion.
Figure 4.
Corrosion rate of X65 linepipe steel under sweet and mildly sour conditions in chloride containing conditions at pH 3.8-4.0 [24].
16
Figure 5.
Corrosion rate of X65 linepipe steel under sweet and mildly sour conditions in simulated formation water at pH 5.5-5.8 [24].
17
18
Figure 6.
Materials optimisation strategy flow chart [after Ref 39].
