Corrosive Environments

Aqueous Corrosion Background

Corrosion in aqueous solutions is an electrochemical process, involving ions (electrically charged atoms) and the transfer of electrical charges at the metallic surfaces. Anodes and cathodes occur locally on the surfaces; metal is removed at the anodes (in the form of positively charged ions, i.e. M > M⁺). Electrons flow within the metallic material from anodic sites to cathodic sites, and vice versa in the aqueous solution. The principle cathodic reaction during corrosion by acids is the reduction of positively charged hydrogen ions to regular hydrogen atoms (i.e. H⁺> H), and subsequently hydrogen molecules (gas). Hence the term reducing acid solution. An oxidizing acid solution is one that induces a cathodic reaction of higher potential; oxidizing acids tend to induce passivation (passivity), whereby protective films form on the metallic surfaces. These films can be multi-layered and can be oxides, hydroxides, or oxy-hydroxides.

The corrosion performance of even one metallic material is a very complex issue, given that there are many forms of corrosion, each dependent upon temperature, concentration, and the chemical purity of the solution. To simplify matters, therefore, this section deals with each form of corrosion in turn, with particular emphasis on the key industrial, inorganic chemicals, and upon the characteristics of each of the major alloy families (within the realm of corrosion- resistant, nickel- and cobalt-based alloys). The emphasis on inorganic chemicals is a reflection of their ionic nature, hence ability to induce an electrochemical (corrosive) process.

Uniform Corrosion in Hydrochloric Acid

Hydrochloric is a reducing acid. It pervades the chemical processing industries (CPI), both as a feedstock and by-product. It is extremely corrosive to most metals and alloys. As will be discussed, many nickel-based corrosion alloys (particularly those with high molybdenum contents) are able to withstand pure hydrochloric acid, within specific concentration and temperature ranges. Be aware, however, that the concentration and temperature dependencies can be strong with certain alloys, and that upset conditions in industry can result in significantly higher corrosion rates when these alloys are pushed close to their limits. Furthermore, some nickel alloys, notably those in the nickel-molybdenum family, are negatively affected by the presence of oxidizing impurities (which can occur in “real world” solutions of hydrochloric acid). Industrial field trials are therefore important, prior to use.

The alloys with the highest resistance to pure hydrochloric acid are those of the nickel-molybdenum family, whose molybdenum contents are close to 30 wt.%. Within this family, the wrought material with the highest level of corrosion resistance plus thermal stability is HASTELLOY^® B-3^® alloy. The corrosion rates of B-3^® alloy in pure, reagent-grade hydrochloric acid are shown in the figure below, as a function of concentration and temperature. Such charts (known as “Iso-Corrosion Diagrams”) will be used frequently in this manual, so some explanation is in order.

These diagrams were constructed mathematically from numerous laboratory data points, and each one defines, for a given alloy and solution, the “very safe”, “moderately safe”, and “unsafe” concentration/temperature regimes. These correspond to the corrosion rate ranges 0 to 0.1 mm/y, 0.1 to 0.5 mm/y, and over 0.5 mm/y. For those more familiar with the traditional American units (mils per year, or mpy), 0.1 mm/y is equivalent to 4 mpy, and 0.5 mm/y is equivalent to 20 mpy. It is noteworthy that, like all the materials in the nickel-molybdenum family, B-3^® alloy is able to withstand pure hydrochloric acid at all temperatures up to the boiling point curve, within the 0 to 20 wt.% concentration range. Tests of the type used to create these diagrams (involving unpressurized glass flask/condenser systems) are only accurate in hydrochloric acid up to a concentration of 20 wt.% (the azeotrope). At higher concentrations, hydrogen chloride gas can escape, resulting in concentration instability and the possibility of erroneous results.

Below is a graph illustrating the effects of oxidizing impurities (ferric ions and cupric ions) upon the performance of B-3^® alloy in hydrochloric acid (in this case boiling 2.5% HCl). For comparison, HASTELLOY^® HYBRID-BC1^® alloy, which contains 15 wt.% chromium in addition to a relatively high molybdenum level, is shown.

The alloys with the next highest resistance to pure hydrochloric acid are those of the nickel-chromium-molybdenum family, whose molybdenum contents range from about 13 to 22 wt.% (in some cases augmented by tungsten, which is half as effective as molybdenum on a wt.% basis). The most widely used, wrought material in this family is HASTELLOY^® C-276 alloy, which contains 16 wt.% of both chromium and molybdenum, plus 4 wt.% tungsten. Its iso-corrosion diagram for pure hydrochloric acid is shown below.

These diagrams were constructed mathematically from numerous laboratory data points, and each one defines, for a given alloy and solution, the “very safe”, “moderately safe”, and “unsafe” concentration/temperature regimes. These correspond to the corrosion rate ranges 0 to 0.1 mm/y, 0.1 to 0.5 mm/y, and over 0.5 mm/y. For those more familiar with the traditional American units (mils per year, or mpy), 0.1 mm/y is equivalent to 4 mpy, and 0.5 mm/y is equivalent to 20 mpy. It is noteworthy that, like all the materials in the nickel-molybdenum family, B-3^® alloy is able to withstand pure hydrochloric acid at all temperatures up to the boiling point curve, within the 0 to 20 wt.% concentration range. Tests of the type used to create these diagrams (involving unpressurized glass flask/condenser systems) are only accurate in hydrochloric acid up to a concentration of 20 wt.% (the azeotrope). At higher concentrations, hydrogen chloride gas can escape, resulting in concentration instability and the possibility of erroneous results.

Below is a graph illustrating the effects of oxidizing impurities (ferric ions and cupric ions) upon the performance of B-3^® alloy in hydrochloric acid (in this case boiling 2.5% HCl). For comparison, HASTELLOY^® HYBRID-BC1^® alloy, which contains 15 wt.% chromium in addition to a relatively high molybdenum level, is shown.

The alloys with the next highest resistance to pure hydrochloric acid are those of the nickel-chromium-molybdenum family, whose molybdenum contents range from about 13 to 22 wt.% (in some cases augmented by tungsten, which is half as effective as molybdenum on a wt.% basis). The most widely used, wrought material in this family is HASTELLOY^® C-276 alloy, which contains 16 wt.% of both chromium and molybdenum, plus 4 wt.% tungsten. Its iso-corrosion diagram for pure hydrochloric acid is shown below.

From the above diagram it is evident that the 16 wt.% molybdenum alloys exhibit strong temperature dependencies, especially at lower concentrations. The importance of molybdenum in resisting pure hydrochloric acid is illustrated in the next figure, which shows the corresponding iso-corrosion diagram for HASTELLOY^® HYBRID-BC1^® alloy, a material which contains 22 wt.% molybdenum (and 15 wt.% chromium, but no tungsten). Note the much broader “very safe” and “moderately safe” regimes, and generally higher temperature capabilities.

Two of the nickel-chromium alloys, namely 625 and HASTELLOY^® G-35^® alloy, contain sufficient molybdenum to provide good resistance to hydrochloric acid. The nominal molybdenum content of 625 alloy is 9 wt.%; that of G-35^® alloy is 8.1 wt.%. The chief differences between these two materials are the chromium contents (21.5 wt.% for 625 alloy versus 33.2 wt.% for G-35^® alloy) and the fact that G-35^® alloy contains little else, whereas 625 alloy has deliberate iron (2.5 wt.%) and niobium (3.6 wt.%, including any associated tantalum) additions. The corresponding iso-corrosion diagrams for 625 and G-35^® alloys are shown in the following two figures. The similarity of these two diagrams indicates that chromium content has little effect upon the performance of such alloys in pure hydrochloric acid.

To compare materials, it is customary to plot their 0.1 mm/y lines, i.e. the lines separating the “very safe” and “moderately safe” regimes. This does not work for all materials, in particular B-3^® alloy, which exhibits two 0.1 mm/y lines (one running horizontally at approximately 40°C), neither of which indicates that, unlike most other materials, B-3^® alloy is safe to use up to the boiling point curve in pure hydrochloric acid, at concentrations up to 20 wt.%. Nevertheless, a comparative 0.1 mm/y line plot such as that shown below does provide perspective on the hydrochloric acid resistance of two nickel alloys (625 and C-276 alloy) versus that of two commonly used, austenitic stainless steels.

Plots of 0.5 mm/y lines can also be used to compare materials. Below is a comparative 0.5 mm/y chart for several chromium- and molybdenum-bearing, nickel-based HASTELLOY^® alloys (including all of the versatile C-type materials), in reagent grade hydrochloric acid.

As already mentioned, copper is beneficial to the resistance of nickel alloys to reducing acids, of which hydrochloric acid is stereotypical. Hence it is relevant to present and discuss the iso-corrosion diagram for MONEL^® 400 alloy, the most commonly used nickel-copper alloy, the nominal copper content of which is 31.5 wt.%. This diagram is shown in the following figure and indicates that, while the nickel-copper alloys possess some resistance to pure hydrochloric acid (higher than that of the common austenitic stainless steels), their performance is well below that of 625 and G-35^® alloys (from the nickel-chromium group).

The only wrought, corrosion-resistant cobalt alloy that has been tested extensively enough in pure hydrochloric acid to enable the construction of an iso-corrosion diagram is ULTIMET^® alloy. This is presented below. Interestingly, despite its relatively low molybdenum and tungsten contents (5 and 2 wt.%, which are equivalent to 6 wt.% molybdenum), its performance in pure hydrochloric acid bears some resemblance to that of 625 and G-35^® alloys. The main difference is the narrower “moderately safe” regime at concentrations in the range 10 to 20 wt.%, indicating that it transitions from “safe” to “unsafe” over a smaller temperature range.

Uniform Corrosion in Sulfuric Acid

Sulfuric acid is also a very important and very corrosive industrial chemical. It is used in the manufacture of fertilizers, detergents, plastics, synthetic fibers, and pigments. It is also used as a catalyst in the petroleum industry. As with hydrochloric acid, the concentration and temperature dependencies can be strong for certain nickel- and cobalt-based materials. Also, “real world” conditions (in terms of chemical purity and flow) are unlike those in the laboratory used to generate the data in the following diagrams, so field trials are recommended. In pure sulfuric acid, again molybdenum is highly beneficial; copper is also advantageous.

Although the high-molybdenum alloys perform well across the whole concentration range in pure sulfuric acid, this is despite the fact that the nature of the cathodic reaction changes for many metallic materials at a concentration of approximately 60 to 70 wt.% (Sridhar, 1987). At lower concentrations, the cathodic reaction is believed to be the reduction of positively charged hydrogen ions (and the evolution of hydrogen gas), whereas mixed cathodic reactions appear to be in play at high concentrations. This affects the behavior of metallic materials such as the zirconium alloys and nickel-copper alloys.

As with pure hydrochloric acid, the nickel alloys with the highest resistance to pure sulfuric acid are those of the nickel- molybdenum family, as indicated in the following iso-corrosion diagram for HASTELLOY^® B-3^® alloy. Note that this diagram covers a large concentration range (up to 90 wt.%), and that the “very safe” regime is extremely large for B-3^® alloy. As with hydrochloric acid, however, the presence of oxidizing species in sulfuric acid negatively affects the nickel- molybdenum alloys.

With their moderately high molybdenum contents (in the 13 to 22 wt.% range), the nickel-chromium-molybdenum alloys also possess high resistance to pure sulfuric acid. Moreover, they are somewhat protected from the oxidizing species that can occur in industrial solutions by chromium. Indeed, the higher the chromium content of the Ni-Cr-Mo alloys, the greater is the level of protection.

To illustrate the performance of the nickel-chromium-molybdenum alloys in pure sulfuric acid, the iso-corrosion diagrams for HASTELLOY^® C-276 alloy and HASTELLOY^® HYBRID-BC1^® are shown below. Note that the temperature capability of both these nickel-chromium-molybdenum alloys does not vary much over the whole concentration range, in pure sulfuric acid. However, the higher molybdenum content of HYBRID-BC1^® alloy (22 wt%) enables the material to be used at considerably higher temperatures.

The iso-corrosion diagrams for 625 and G-35^® alloys (from the nickel-chromium family of materials) in pure sulfuric acid are shown below.

From these diagrams, it is evident that the characteristics of these two, molybdenum-bearing, nickel chromium alloys are almost identical in pure sulfuric acid and that temperature is critical to their performance. Indeed, at many concentrations, 625 and G-35^® alloys exhibit no “moderately safe” regime. For perspective, their temperature capabilities in pure sulfuric acid are much higher than those of common, austenitic stainless steels, as indicated by the comparative 0.1 mm/y line plot shown below, which includes 625 alloy.

A comparative 0.5 mm/y chart for several chromium- and molybdenum-bearing, nickel-based HASTELLOY^® alloys (including all of the versatile C-type materials), in reagent grade sulfuric acid is presented below.

As already mentioned, the performance of the nickel-copper alloys in pure sulfuric acid is affected by the change in cathodic reaction at concentrations in the range 60 to 70 wt.%. This is illustrated in the following iso-corrosion diagram for MONEL^® 400 alloy.

The iso-corrosion diagram for ULTIMET^® alloy in pure sulfuric acid is shown below. Comparison of this diagram with those for 625 and G-35^® alloys reveals many similarities, in particular the absence of a “moderately safe” regime at certain concentrations (i.e. there is a strong temperature dependency).

Uniform Corrosion in Highly Concentrated,
Industrial Sulfuric Acid

A major source of sulfuric acid is the mining and metal extraction industry, which produces highly-concentrated solutions (typically 92-99 wt.%) from smelter off-gases. These solutions are described as “super-oxidizing” and induce cathodic reactions of very high potential, beyond the range capable of supporting chromium-rich passive films. Nevertheless, nickel-chromium-molybdenum alloys can be used in such solutions, up to about 95°C (Sridhar, 1987). For higher temperatures, materials which form alternate and sustainable protective films at these potentials are required.

Such materials include high-silicon nickel-based alloys and high-silicon stainless steels, both of which form protective silica films in solutions of this type. An example is HASTELLOY^® D-205^® alloy. Due to its weld mechanical properties, this wrought alloy is only available in thin sheet form for gasketed plate heat exchangers.

Uniform Corrosion in Hydrobromic Acid

Hydrobromic is another strong, reducing acid to which the molybdenum-bearing nickel alloys are resistant, within certain concentration and temperature ranges. It is not widely encountered within the chemical process industries, but is important in the production of inorganic bromides and of brominated organic compounds. The characteristics of the corrosion-resistant nickel alloys in pure hydrobromic acid are closely allied to those of the same materials in hydrochloric acid.

The iso-corrosion diagram for HASTELLOY^® B-3^® alloy in pure hydrobromic acid is shown below. Note that the azeotrope for hydrobromic acid is 40 wt.%, hence the extended concentration range. Comparing this diagram with the corresponding chart for hydrochloric acid, it is apparent that they are quite similar, in that they feature a large “moderately safe” regime, and no “unsafe” zone.

The iso-corrosion diagrams for C-276 and HYBRID-BC1^® alloys in pure hydrobromic acid are shown in the following figures. Assuming the azeotropes for hydrobromic and hydrochloric acids to be equivalent, the diagrams are quite similar to those for the same alloys in hydrochloric acid, although C-276 and HYBRID-BC1^® appear to be more resistant to hydrobromic acid. Particularly notable are the high position of the 0.5 mm/y line in the HYBRID-BC1^® diagram, indicating a “moderately safe” zone to temperatures in excess of 100°C, and the increase in temperature capability of C-276 alloy in the concentration range 15 to 40 wt.%.

The molybdenum-bearing nickel-chromium materials, such as 625 and G-35^® alloys also exhibit good resistance to pure hydrobromic acid, as indicated by the following iso-corrosion diagrams.

Comparing the diagrams for these two alloys in hydrochloric and hydrobromic acids, it is evident that hydrobromic is less corrosive. Also, comparing the hydrobromic acid diagrams for 625 and G-35^® alloys, it appears that the latter is considerably more resistant to this acid at concentrations up to about 25 wt.%. This infers that the high chromium content of G-35^® alloy might be beneficial.

Uniform Corrosion in Hydrofluoric Acid

Aqueous solutions of hydrofluoric acid (HF) are among the most difficult to deal with. Not only can such solutions attack glass, but also alloys of elements such as titanium and zirconium, which are normally protected by oxide films, are rendered useless by hydrofluoric acid. In fact, select nickel alloys are among the few metallic materials resistant to hot hydrofluoric acid solutions, and even they have restricted temperature and concentration capabilities. Moreover, hydrofluoric acid can induce other forms of corrosion, such as internal attack and stress corrosion cracking (both of which will be discussed in later sections of this manual).

Hydrofluoric acid is also unusual (for a reducing acid) in that it can induce the formation of pseudo-passive films on nickel-based alloys. Indeed, the outstanding performance of MONEL^® 400 alloy (from the nickel-copper family of materials) has been attributed to the formation of protective fluoride films on exposed surfaces.

As to industrial uses of hydrofluoric acid, it is found in the etching and cleaning of metals and ceramics, acid treating of oil and gas wells, and extractive metallurgy. Although beyond the scope of this manual, anhydrous hydrogen fluoride is used in the manufacture of the industrially important fluoro-chemicals (Jennings 2006). While all acids pose a safety issue, hydrofluoric acid is by far the most dangerous. No skin should be exposed when handling solutions of the acid, and fumes should be avoided.

Since glass is attacked by hydrofluoric acid, laboratory corrosion tests are carried out at Haynes International in TEFLON^® flasks, with TEFLON^® condensing systems. Due to the dangers of HF, tests (of 96 h duration) are normally carried out without interruption (tests in other acids involve four 24 h test periods, with interruptions for cleaning and weighing of samples, and replenishment of solutions, if needed).

Within the realm of corrosion-resistant nickel- and cobalt-based alloys, the nickel-copper alloys are the most commonly utilized for industrial applications involving hot, pure, aqueous solutions of hydrofluoric acid. However, they are negatively affected by the presence of oxygen. In such circumstances, the nickel-chromium-molybdenum alloys are used, although they are restricted to lower operating temperatures.

An iso-corrosion diagram for MONEL^® 400 alloy in aqueous solutions of hydrofluoric acid is shown in Crum et al, 1999. Essentially, it indicates that the alloy generally exhibits rates of less than 0.5 mm/y at all concentrations up to 100 wt.%, and at temperatures up to boiling. Jennings 2006 states that commercial aqueous hydrofluoric acid is available at concentrations of 49 and 70%; this is consistent with the experimental procedure of Crum et al, 1999, which states that their aqueous hydrofluoric acid tests involved solutions prepared from a concentration of 49 wt.%. It does not explain how the higher concentrations were attained, nor does it give the duration of the tests, which experience at Haynes International suggests is important. Tests of 24 h duration gave variable results, suggesting an incubation period, during which the pseudo-passive fluoride films are becoming established. Tests of 240 h (without interruption) resulted in concentration changes, due to the escape of hydrogen fluoride gas at the thermometer seals in the TEFLON^® system.

A more important graph generated by Crum et al is reproduced in the following figure. It compares the positions of the 0.5 (or more precisely 0.51) mm/y lines of many nickel alloys on the iso-corrosion diagram for hydrofluoric acid. In particular, it indicates that the nickel-chromium-molybdenum alloys fall within a fairly narrow performance band, along with pure nickel, and well above that for the nickel-chromium, nickel-chromium-iron, and nickel-iron-chromium materials.

Much hydrofluoric acid testing was performed at Haynes International during the period 1995 to 2010, resulting in several technical papers, notably Rebak et al 2001 and Crook et al, 2007. The former paper focuses on the behavior of HASTELLOY^® C-2000^® alloy versus MONEL^® 400 alloy, both immersed in, and in the vapor phase above, solutions of hydrofluoric acid. The latter paper describes the internal attack that can occur in the nickel-chromium-molybdenum alloys (in particular C-22^®, C-276, and C-2000^® alloys).

Only one iso-corrosion diagram involving hydrofluoric acid is currently in use at Haynes International, and that is for C-2000^® alloy (shown below). Note that it only covers concentrations up to 30 wt.%, and that there is considerable difference between the slope of the 0.5 mm/y line and the slope for the nickel-chromium-molybdenum materials in in the Crum comparison, suggesting a greater temperature dependency.

One of the problems of using iso-corrosion diagrams for hydrofluoric acid is that they do not indicate whether or not internal attack is occurring. To remedy this, five alloys (from the Ni-Cr-Mo and Ni-Cr families) were subjected to extensive tests in hydrofluoric acid, then studied for internal attack. The results were used to create simplified corrosion rate charts, with indicators (black crosses) as to the temperatures and concentrations which can induce internal attack over 96 hours. These charts follow the iso-corrosion diagram for C-2000^® alloy.

To indicate the “very safe”, “moderately safe”, and “unsafe” regimes for the five alloys, the enlarged squares (centered on the test concentrations and temperatures) have been colored green, blue and red, respectively. From these charts, it is evident that:

1. Of the three nickel-chromium-molybdenum materials tested, C-2000^® alloy is the least susceptible to internal attack; this might be because C-2000^® alloy contains a small (1.6 wt.%) addition of copper.
2. The nickel-chromium materials are much less resistant to hydrofluoric acid; any internal attack is probably overwhelmed by the high rates of general attack in the “unsafe” zones.

Uniform Corrosion in Nitric Acid

Unlike the acids that have been discussed so far, nitric acid induces a high-potential, oxidizing, cathodic reaction with metallic materials. Under such circumstances, protective (passive) films form readily on materials containing sufficient chromium, such as the stainless steels, chromium-bearing nickel alloys, and chromium-bearing cobalt alloys.

For industrial applications involving only pure nitric acid, the stainless steels generally possess sufficient corrosion resistance. The chromium-bearing nickel alloys are only required if other chemicals, notably the halides and/or halogen acids, are present, or if the equipment has multiple uses involving different chemicals.

The performance of nickel-based alloys in pure nitric acid is very much a function of chromium content. Thus, HASTELLOY^® G-35^® alloy, which has a very high chromium content (33.2 wt.%), exhibits outstanding resistance, as shown below (the corresponding iso-corrosion diagram). In fact, a rate of less than 0.1 mm/y is expected at all temperatures below the boiling point curve, in the concentration range 0 to 70 wt.% (the highest concentration for which reagent grade, laboratory acid is available).

Two iso-corrosion diagrams showing the performance of nickel-chromium-molybdenum alloys in pure nitric acid are presented below. These relate to C-276 alloy (16 wt.% chromium) and C-2000^® alloy (23 wt.% chromium). Note how the 0.1 mm/y and 0.5 mm/y lines have been pushed to higher concentrations and temperatures by the higher chromium content of C-2000^® alloy.

B-3^® alloy, the chromium content of which is only 1.5 wt.%, corrodes rapidly in nitric acid. On the other hand, HYBRID-BC1^® alloy, with a chromium content of 15 wt.%, can withstand even 70% nitric acid (i.e. exhibit corrosion rates less than 0.5 mm/y) up to a temperature of approximately 50°C.

Uniform Corrosion in Phosphoric Acid

There are two types of phosphoric acid, a pure, “food” grade made from elemental phosphorus, and an impure, “fertilizer” grade, which is made by reacting phosphate rock with sulfuric acid. Pure, “food” grade phosphoric acid is not nearly as corrosive as the other reducing acids, such as hydrochloric. Indeed, many alloys from the nickel-chromium, nickel-molybdenum, and nickel-chromium-molybdenum systems exhibit corrosion rates of less than 0.1 mm/y over large temperature and concentration ranges, as indicated in the iso-corrosion diagrams shown below for G-35^® alloy, B-3^® alloy, and C-276 alloy.

With regard to “fertilizer” grade phosphoric acid, important steps in the manufacture of fertilizers are the production and concentration of phosphoric acid. This acid is made by reacting phosphate rock with sulfuric acid, hence the alternate name “wet process” phosphoric acid. Not only does this fertilizer-grade/wet process acid contain traces of sulfuric acid, but also it contains impurities from the phosphate rock, which serve to increase its corrosivity.

The primary constituents of the phosphate rock are tricalcium phosphate (from which phosphoric acid and calcium sulfate are created), calcium fluoride (which generates hydrogen fluoride and calcium sulfate), and calcium carbonate (which yields carbon dioxide, calcium sulfate, and water). The concentration of the phosphoric acid (or rather the P2O5 content) is determined by the amount of rinse water needed to separate it from the calcium sulfate, and is normally in the range 30 to 32 wt.% (Crook and Caruso, 2004).

Impurities include fluoride ions, from the hydrogen fluoride (although these generally form complexes with metallic ions), chloride ions, silica, aluminum, iron (which serves to increase the oxidizing potential of the acid), calcium, and sodium.

A typical concentration sequence (to enable transportation and effective use) is:

Step 1: 30-32% to 37-39%

Step 2: 37-39% to 45-48%

Step 3: 45-48% to 52-54%

Step 4: 52-54% to 69-72%

The first three steps typically involve temperatures of 90°C or higher, and the fourth step typically involves temperatures up to 150°C.

Geographically, producers of fertilizer-grade/wet process phosphoric acid are located close to the largest accessible reserves of phosphate rock. In the United States, activity is centered in Florida, although there are also plants in the North-West. Across the Atlantic, most of the action is centered around North African and Middle Eastern deposits. Differences in impurity levels at the various locations result in differences in corrosivity.

The graph below compares the results for several materials in fertilizer-grade/wet process phosphoric acid (in the concentration range 36 to 54 wt.% P2O5). Acid 1 and acid 2 were from different locations in Florida. The tests were performed in autoclaves at 121°C, for a duration of 96 hours (uninterrupted). The choice of this temperature was based on previous tests of the high-chromium stainless steels (Alloys 28 and 31), for which 121°C was the upper limit. It should be noted that, for the nickel-based materials (i.e. C-2000^® alloy with 23 wt.% chromium, G-30^® alloy with 30 wt.% chromium, and G-35^® alloy with 33.2 wt.% chromium), there is a strong correlation between chromium content and corrosion resistance, due to the presence of oxidizing species.

Uniform Corrosion in Sodium and Potassium Hydroxides

Sodium hydroxide (NaOH, caustic soda) and potassium hydroxide (KOH, caustic potash) are widely used chemicals. Applications include the manufacture of soap, paper, and aluminum. They are also used to neutralize acids, especially in the petrochemical industries. Molten NaOH is used to descale stainless steels and other alloys, in the metals industry. The materials most commonly used to resist NaOH and KOH are the commercially pure, wrought nickel products, Ni 200 and Ni 201 (Dillon, 1994, Klarstrom, 1987, Friend, 1980). The latter is a low carbon grade for use at temperatures above 315°C, where graphitization is a possibility with the higher carbon material (Ni200). Where greater strength is required, Ni301, which can be age-hardened, is used. The iso-corrosion diagram for Nickel 200 in sodium hydroxide (Ref. Crum and Schumaker, 2006) is shown below.

In many heat transfer systems, the material of choice must resist NaOH (or KOH) on one side and chlorinated or contaminated water on the other (Rebak, 2005). Worse still, some situations require that the material be exposed sequentially to both acids and alkalis. In such cases, nickel alloys containing chromium and/or molybdenum are necessary.

As to the resistance of these chromium and molybdenum bearing nickel alloys to NaOH and KOH, several are susceptible to “caustic dealloying” at high concentrations and temperatures (Chambers, 2002, Crook and Meck, 2005). This phenomenon is characterized by the selective removal of elements other than nickel from near-surface regions of the microstructure. In tests involving 50 wt.% sodium hydroxide at temperatures up to 107°C, it was determined that B-3^® and C-2000^® alloys are susceptible to dealloying at 66°C and 79°C (and above), respectively, whereas G-35^® alloy does not exhibit dealloying, at least up to the maximum test temperature, and over the 72 h test duration. This infers that molybdenum is a bad actor regarding caustic dealloying, and/or that a higher chromium content is beneficial.

Pitting and Crevice Corrosion

Pitting and crevice corrosion are forms of attack associated with the presence of chlorides (or other halides) in aqueous solutions. Pitting is typically initiated by the localized breakdown of passive films, or by the occurrence of localized electrochemical cells (normally the anodic and cathodic sites move around on a metallic surface, but metallurgical inhomogeneity can result in static conditions). As its name suggests, crevice corrosion occurs in crevices or narrow gaps between structural components. Both forms of attack are associated with the localized build-up of positive charge, and the attraction of negatively charged chloride (or other halide) ions to the pit or gap, followed by the formation of the corresponding acid (hydrochloric acid, in the case of chloride ions). This acid accelerates the attack, and the process becomes auto-catalytic.

It was once stated that the commercial success of the nickel-chromium-molybdenum alloys has been largely due to their resistance to pitting and crevice corrosion (i.e. chloride-induced localized attack). Indeed, this is a major attribute of this family of materials, another being their resistance to chloride-induced stress corrosion cracking, which will be discussed later.

The pitting resistance of a material is hard to assess. Short-term tests can result in miniscule pits or large pits, both of which are treated equally. Corrosion rates are generally misleading, since there is great random error associated with the time of pit initiation and progression; in a 24 h test of identical samples, for example, the first pit on each sample can start either early or late in the test, giving widely different corrosion rates for the two samples.

The pitting and crevice corrosion tests preferred by the author are those described in ASTM Standard G48: Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by Use of Ferric Chloride Solution. Six test methods are described in this standard, two of which (Methods C and D) are related to nickel-based and chromium-bearing alloys. Method C enables determination of the critical pitting temperature (CPT) of a material, i.e. the lowest temperature at which pitting is observed in a solution of 6 wt.% ferric chloride + 1 wt.% hydrochloric acid, over a 72 h period. Method D requires a crevice assembly to be attached to the sample, to enable determination of the corresponding critical crevice temperature (CCT). In the inter-laboratory test program associated with ASTM Standard G48, a maximum temperature of 85°C was used. The standard does allow testing at higher temperatures, but does not address the equipment (i.e. autoclaves) required for tests above the boiling point.

Generally speaking, the ranking of alloys in Method C correlates well with the ranking in Method D, although crevice corrosion tends to be encountered at much lower temperatures than pitting. With this in mind, along with industry concerns about high temperature autoclave testing, the focus at Haynes International has been on Method D, and the associated CCT’s of the nickel-based alloys (relative to stainless steels).

The CCT’s of several nickel-chromium and nickel-chromium-molybdenum alloys are given in the following bar chart, along with those of 316L (CCT = 0°C) and 254SMO^® alloy. They suggest that both molybdenum and chromium are important to chloride-induced pitting and crevice corrosion resistance, with the former being more influential (which is not surprising, since hydrochloric acid is the aggressive chemical of concern in chloride solutions).

With regard to the pitting and crevice corrosion resistance of the cobalt-based alloys, only ULTIMET^® alloy and HAYNES^® 25 alloy have been seriously compared with the nickel-based CRA’s. This is partly due to the fact that the tungsten- bearing, high-carbon cobalt alloys are not nearly as corrosion-resistant as the nickel-based CRA’s, since portions of their chromium and tungsten contents contribute to carbide formation. Furthermore, from the standpoint of chloride- induced localized attack, their carbide-laden microstructures are inhomogeneous and therefore likely to cause localized electrochemical conditions. ULTIMET^® alloy and HAYNES^® 25 alloy, on the other hand, are homogenous wrought products, the former designed specifically to excel in aqueous solutions of corrosive chemicals.

Tests to determine the pitting resistance of ULTIMET^® alloy were carried out in the early 1990’s, before the ASTM G48 test became favored at Haynes International. At that time, the critical pitting temperature in Green Death (a solution designed to simulate condensates in flue gas desulfurization systems) was the definitive measure of a material’s resistance to chloride-induced pitting. Green Death is similar to the ASTM G28B quality control solution and comprises 11.5% H2SO4 + 1.2% HCl + 1% FeCl3 + 1% CuCl2 (where all percentages are by weight). The duration of the Green Death test is 24 h.

Surprisingly, ULTIMET^® alloy was equal to the nickel-chromium-molybdenum alloy with the highest critical pitting temperature (i.e. C-22^® alloy), as shown in the table below, despite the much lower combined molybdenum plus tungsten level of ULTIMET^® alloy. It perhaps indicates that these two elements are more effective in a cobalt, rather than nickel, base. Alternatively, nitrogen (which is more soluble in cobalt alloys than nickel alloys, and was deliberately added to ULTIMET^® alloy to enhance its resistance to chloride-induced pitting) could be responsible.

Testing in Green Death revealed that HAYNES^® 25 alloy also possesses high resistance to chloride-induced pitting. Its critical pitting temperature in this solution is close to that of C-276 alloy, despite that, in atomic terms, its combined molybdenum plus tungsten content is relatively low.

Note that HAYNES^® 6B alloy, the wrought version of the most popular high-carbon cobalt-based alloy, has a Green Death critical pitting temperature of 45°C, which is significantly higher than that of 316L stainless steel.

Critical Pitting Temperatures in Green Death Solution

Perhaps the most common chloride solution encountered by the wrought, corrosion-resistant nickel alloys is sea water. Sea water is encountered by marine vessels, oil rigs, and coastal structures and facilities (which typically use sea water as a coolant). As a chloride, it can induce pitting, crevice attack, and stress corrosion cracking of metallic materials, as well as uniform attack. Furthermore, marine equipment can become encrusted, leading to a form of crevice attack known as “under-deposit” corrosion; biofouling is also an issue.

Fortunately, the nickel alloys possess good sea water resistance. In particular, those with high copper contents, such as alloy 400, resist biofouling (copper being a poison to microbes). For stagnant or low velocity conditions, chromium- and molybdenum-bearing nickel alloys are favored, due to their higher resistance to pitting and crevice attack.

Some crevice corrosion data for sea water, generated as part of a U.S. Navy study at the LaQue Laboratories in Wrightsville Beach, North Carolina, are presented in the following table created by Aylor et al, 1999. Crevice tests were performed in both still (quiescent) and flowing sea water, at 29°C, plus or minus 3°C. Two samples of each alloy were tested in still water, for 180 days, and two samples of each alloy were tested in flowing water, for 180 days. Each sample contained two possible crevice sites. In quiescent sea water, the results mirror those generated in acidified ferric chloride, with C-22^® and C-2000^® alloys as the most resistant. In flowing sea water, crevice attack of the stainless steels was shallower, and none of the Ni-Cr-Mo alloys exhibited crevice corrosion.

Stress Corrosion Cracking

One of the chief advantages of the wrought, corrosion-resistant, nickel-based alloys is their high resistance to chloride- induced stress corrosion cracking (SCC), a form of corrosion to which the stainless steels are particularly prone. In fact, evidence points to the fact that the higher the iron content of the nickel alloys, the less is their resistance to this extremely harmful form of attack.

Chloride-induced SCC falls within the broader category of environmentally assisted cracking (EAC), which also includes hydrogen embrittlement, sulfide stress cracking, and liquid metal embrittlement (Rebak, 2005). All of these phenomena describe circumstances under which ductile materials exhibit embrittlement when subjected to tensile stresses in specific corrosive environments.

While chlorides, hence chloride-induced SCC, are the main concern of this manual, other halides (bromides, fluorides, and iodides) can be equally damaging. Also, hydrofluoric acid is known to cause stress corrosion cracking, especially in the presence of dissolved oxygen. Resistance to hydrogen embrittlement and sulfide stress cracking are important attributes for down-hole applications in the oil and gas industry; fortunately, the nickel-chromium and nickel- chromium-molybdenum alloys are very resistant to these forms of environmentally assisted cracking, and are consequently used in some of the most severely corrosive wells.

A test commonly used to assess the resistance of metallic materials to chloride-induced SCC is that described in ASTM Standard G36, which involves a boiling solution of 45 wt.% magnesium chloride. The following table gives the times required to cause cracking of U-bend samples (made from 3.2 mm sheet) of various alloys in this solution. For the low- iron nickel-based CRA’s, which did not exhibit cracking, the test was stopped after 1,008 h (six weeks).

EAC and chloride-induced SCC are complicated subjects. In the case of nickel-based alloys and stainless steels, and aqueous solutions, it is believed that local electrochemical conditions occur at crack tips, to ensure rapid crack propagation. Microstructure also plays a part, cold-worked materials being much more prone to cracking than solution annealed materials.

With regard to the propagation of cracks during chloride-induced SCC, a branching transgranular pattern is common with the austenitic stainless steels; however, other propagation paths are possible. Hydrofluoric acid, for example, can induce rather peculiar crack patterns in stressed nickel-based CRA’s, as illustrated in the following photomicrographs, which show a) extremely fine/barely resolvable cracking in the outer surface of a C-2000^® U-bend sample exposed to 20% hydrofluoric acid at 79°C for 240 h, an b) cracking of a more normal nature in a C-22^® U-bend sample exposed to the same conditions.

Cracking in C-2000^® U-Bend Sample Exposed to 20% HF at 79°C

(Crook et. al. 2007)

Cracking in C-22^® U-Bend Sample Exposed to 20% HF at 79°C

(Crook et al, 2007)