1. INTRODUCTION
Alkylation units are used in petroleum refineries to convert isobutane and low-molecular-weight alkenes into alkylate, a high-octane gasoline component. The process occurs in the presence of an acid, typically sulphuric or hydrofluoric acid, which acts as a catalyst. Hydrofluoric acid is normally the preferred catalyst as a typical hydrofluoric acid alkylation unit (HFAU) requires far less acid than a sulphuric acid unit to achieve the same volume of alkylate.
Despite significant technological advances, corrosion remains a major concern that affects the safety and reliability of an HFAU. Proper consideration needs to be given to the selection of materials in contact with hydrofluoric acid. Surprisingly, carbon steel is by far the most common material used, although it requires strict limits on composition and hardness, and its performance is highly dependent on control of the process parameters.
2. HYDROFLUORIC ACID ALKYLATION
A hydrofluoric acid alkylation unit, shown schematically in Figure 1 [1], is comprised of three main sections: reactor, fractionator, and defluorinator [2]. Its function is to convert feedstock containing isobutane and low-molecular-weight alkenes (primarily a mixture of propene and butene) into alkylate, a high-octane component used in the production of gasoline.
The feedstock is first treated in a coalescer to remove water, sulphur, and other contaminants. It is then passed to the combined reactor settler. From here, reaction products are fed to the main fractionator. This separates the alkylate and n-butane from other constituents through distillation. Unreacted isobutane is recovered and recycled back into the combined reactor settler.
Vapour from the fractionator is condensed and collected in the accumulator. A portion of the condensed fluid is returned to the combined reactor settler, a portion is used for reflux to the main fractionator, and the remainder is fed to the propane stripper. The vapour from the propane stripper is returned to the overhead condenser.
The propane and n-butane product streams are processed in the defluorination section to remove combined fluorides and any trace acid that may be present due to carryover.
2.1. Reactor Section
Isobutane and low-molecular-weight alkenes are injected into the combined reactor settler together with a hydrofluoric acid catalyst to produce alkylate. The process is highly exothermic, and the reactor vessel is water-cooled to maintain a reaction temperature of 32-38°C.
The hydrocarbon and acid phases separate rapidly in the combined reactor settler. The hydrocarbon rich phase is sent to the main fractionator. The denser acid phase is pumped or circulated under gravity to the acid regenerator from where it is returned to the combined reactor settler.
The water content of the circulating acid is critical. For optimum performance, it should have a water content in the region of 0.5-1.5 wt%. Below 0.5 wt%, alkylate production will be reduced. Above 1.5 wt%, carbon steel corrosion rates increase and the morphology of the protective iron fluoride scale changes. Particles of suspended iron fluoride scale become dislodged and cause fouling later in the process.
The acid strength is equally important. This is a measure of the purity of the acid and is normally maintained between 83% and 90%. From a corrosion standpoint, it is best to operate at the upper end of this range, preferably above 88%. Water, acid soluble oils and dissolved reactants reduce the strength of the acid. These impurities are removed via the acid regenerator.
2.2. Fractionator Section
The fractionator section typically consists of an iso-stripper, a propane stripper, and a hydrofluoric acid stripper. The iso-stripper is the main fractionator. It strips hydrofluoric acid overhead, separates unreacted isobutane, which is recycled to the combined reactor settler, and produces butane and alkylate product streams. Alkylate is drawn off from the bottom of the fractionator and n-butane drawn off through the side.
The stripped overhead vapor is a propane-enriched isobutane stream that contains hydrofluoric acid. It is condensed and collected in the accumulator, where it separates into a denser hydrofluoric acid phase and a lighter hydrocarbon phase. The acid phase is returned to the reactor section. The hydrocarbon phase is sent to the propane stripper.
In the propane stripper, a portion of the hydrocarbon phase containing most of the light hydrocarbons and hydrofluoric acid is driven overhead as vapour. The remaining portion of the hydrocarbon phase containing mainly propane and the heavier hydrocarbons, and which is essentially free of hydrofluoric acid, is recovered from the bottom of the stripper as liquid.
The vapour phase is passed to the overhead condenser, where it is condensed and collected in the accumulator. As discussed earlier, a portion of the condensed fluid is returned to the combined reactor settler, a portion is used for reflux to the main fractionator, and a portion is fed back to the propane stripper.
The liquid phase recovered from the bottom of the propane stripper is passed to the defluorinator section and treated as described below.
2.3. Defluorinator Section
Depending on the end use of the product, propane, n-butane and occasionally alkylate may need to be treated to remove any free acid and/or organic fluorides. The removal of organic fluorides is achieved by vaporising the fluid and contacting it with activated alumina in a defluorinator(s).
3. MATERIAL SELECTION
NACE International Publication 5A171 provides an excellent guide when selecting suitable materials for use in the construction of a hydrofluoric acid alkylation unit. Figure 2 is taken from NACE International Publication 5A171 and indicates the ranges in which various alloys can be used. It has been developed based on data taken from laboratory results and experience in low-flow, low-oxygen commercial aqueous hydrofluoric acid and anhydrous hydrogen fluoride.
The choice of carbon steel is dictated primarily by the predicted corrosion rate. Figure 3 [3] shows the corrosion rate of carbon steel as a function of the hydrofluoric acid concentration at ambient temperature. Carbon steel relies on the formation of a film of iron fluoride for its corrosion performance. Consequently, fluid velocity is important in minimising the risk of erosion-corrosion, which will damage the protective fluoride film. For this reason, the velocity of liquid anhydrous hydrogen fluoride in carbon steel piping is generally limited to 1.8 m/s.
A further consideration is the susceptibility of carbon steel hydrogen stress cracking (HSC), hydrogen embrittlement (HE), and/or hydrogen induced cracking (HIC). This is covered later in the section on hydrogen damage.
Whilst carbon steel has been used successfully in aqueous hydrofluoric acid service at concentrations as low 70%, as discussed previously, it is normal to maintain a concentration of between 83% and 90% and to limit the water content to a maximum of 2.5%.
In anhydrous hydrogen fluoride, the corrosion rate is closely linked to and increases with the velocity of the liquid or gas as well as the temperature. Carbon steel has been found to exhibit satisfactory corrosion resistance at temperatures up to 66°C and can withstand temperatures of up to 82°C, but the corrosion rates will be significantly higher, combined with a thicker accumulation of iron fluoride. The corrosion rate in anhydrous hydrogen fluoride storage vessels at ambient temperature is generally very low (in the region of 0.025 mm/y).
As noted above, it is important to limit the velocity of liquid anhydrous hydrogen fluoride in carbon to avoid damaging the protective fluoride film. NACE International Publication 5A171 [3] suggests a maximum velocity of 1.8 m/s at ambient temperature, while Eurofluor [4] recommends a lower limit of 1.5 m/s. The protective film is more easily damaged at higher temperatures, so at 66°C, the velocity should be limited to less than 0.6 m/s.
As indicated in Figure 3 [3], the corrosion rate of carbon steels in aqueous hydrofluoric acid increases with dilution of the hydrofluoric acid solution [4].
- At concentrations between 85% and 100%, the corrosion rate is low and remains roughly similar.
- At concentrations between 70% and 85%, the corrosion rate increases but is still low enough to allow carbon steel to be used at room temperature.
- At concentrations below 70%, the corrosion rate increases noticeably, and the use of carbon steel is not recommended.
Even when the corrosion rate is low (i.e. at acid concentrations above 85% or in anhydrous hydrogen fluoride), atomic hydrogen will be generated and as already mentioned, hydrogen damage can occur. As discussed later, this may require additional restrictions to be placed on the chemical composition and processing of the steel.
4. CHEMICAL COMPOSITION
Chemical composition is important for two reasons, its influence on the general corrosion of the steel, and its effect on the resistance of the steel to hydrogen damage. The importance of chemical composition in determining the resistance of a carbon steel to hydrogen damage is discussed in the next section in relation to the different types of damage that can occur.
The chemical composition of carbon steel is known to have a major influence on the variability of non-uniform corrosion in hydrofluoric acid service. The corrosion allowance will normally be established based on the assumption that corrosion will be uniform. Non-uniform corrosion can be extremely damaging as corrosion is restricted to a localised area with the result that metal is lost from a fixed location, rather than being spread uniformly across the surface.
In a study [5, 6], it was observed that the carbon content in combination with certain residual elements (nickel, copper, and chromium) could increase non-uniform corrosion by up to five times in comparison to moderate measured corrosion rates. In the same study [5, 6], it was concluded that non-uniform corrosion would be minimised by having a carbon content greater than 0.18% and a combined copper and nickel content of less than 0.15%. Non-uniform corrosion was maximised when the carbon content was less than 0.15% and the combined copper, nickel, and chromium content was greater than 0.30%.
5. MATERIAL REQUIREMENTS
The observations regarding the optimum chemical composition of carbon steel have been captured in ASTM A106 [7], Supplementary Requirement S9, “Requirements for Carbon Steel Pipe for Hydrofluoric Acid Alkylation Service”. This specifies that:
- Pipe is to be supplied in the normalised heat-treated condition.
- The carbon equivalent (CEV), based on the heat analysis, and calculated using the following equation, shall not exceed 0.43% when the specified wall thickness is equal to or less than 25.4 mm, or 0.45% when the specified wall thickness is greater than 25.4 mm.
- Based on heat analysis in mass percent, the vanadium content shall not exceed 0.02%, the niobium content shall not exceed 0.02%, and the sum of the vanadium and niobium contents shall not exceed 0.03%.
- Based on heat analysis in mass percent, the sum of the nickel and copper contents shall not exceed 0.15%.
- Based on heat analysis in mass percent, the carbon content shall not be less than 0.18%.
It is noted that the above requirements relating to the niobium and vanadium contents are less restrictive than those recommended below to ensure an effective post weld heat treatment. Further, no additional restrictions are placed on the sulphur content or steelmaking processes.
6. HYDROGEN DAMAGE
Atomic hydrogen is evolved when corrosion occurs. Atomic hydrogen will normally combine to form molecular hydrogen, which from a materials viewpoint is not a major issue as it will not diffuse into the steel. However, the presence of sulphides and arsenic, which are common impurities in the process stream, poison this reaction such that atomic hydrogen will diffuse into the steel. There are three types of damage that can occur when evolved hydrogen is absorbed by the steel [8].
6.1. Hydrogen Assisted Stress Corrosion Cracking
This occurs in the hardened heat affected zone adjacent to welds. Hardening occurs during welding, when the steel cools rapidly, and is a function of the carbon equivalent and cooling rate. A higher carbon equivalent and/or a rapid cooling rate increases the tendency for hardening to occur. It is good practice to limit the carbon equivalent to 0.43%. The cooling rate can be minimised by applying preheat and by using high heat input welding processes.
It is recommended that post weld heat treatment is carried out. This will reduce the risk of cracking by lowering the hardness and relieve a lot of the residual stress, thereby reducing, but not eliminating, the potential for cracking. For post weld heat treatment to be effective, ideally the combined vanadium and niobium content should not exceed 0.1%.
6.2. Hydrogen Induced Cracking
Hydrogen induced cracking is caused when atomic hydrogen that diffuses through the steel collects at microvoids or inclusions within the structure. The atomic hydrogen atoms combine to form molecular hydrogen, which becomes trapped. As the molecular hydrogen accumulates, it produces a high pressure within the cavity. Internal cracks form and propagate due to the increasing pressure [9].
Hydrogen induced cracking can take various forms. Stress-orientated hydrogen-induced cracking occurs under the influence of an applied load and takes the form of a vertical array of alternate planar hydrogen induced cracks linked by stress cracks. The array is orientated normal to the direction of the maximum applied load. With classic stepwise cracking, an applied load is not required, individual cracks link up due to tearing caused by the intense stress fields generated at the ends of the cracks.
The cracking mechanism in stress-orientated hydrogen-induced cracking is unclear. One hypothesis is that it occurs in steels strengthened by microalloying. If this is the case, it would be more likely to occur in higher strength steels, i.e. API 5L X65 and above. For this reason, it is normal practice to limit the yield strength of steels in hydrofluoric acid service to 415 MPa.
There are many microstructural features that can act as initiation sites for cracks. Elongated manganese sulphide (MnS) inclusions are probably the most common cause of stepwise cracking, but there are other inclusions that can have incoherent interfaces which may include microvoids in the structure, principally silicon oxide (SiO2) and aluminium oxide (Ai2O3).
The risk of hydrogen induced cracking can be significantly reduced using clean steels. An ultra-low sulphur content (typically less than 0.003%), vacuum degassing, and calcium treatment (inclusion shape control) will reduce the number and cross-sectional area of detrimental inclusions [10].
A low sulphur content on its own will not ensure that a steel is immune from hydrogen induced cracking. The oxygen content is also important [11]. Hence, it is good practice to specify both vacuum degassing and a fully killed steel. It should be noted that in the case of forged components such as flanges, hydrogen induced cracking is not a major concern, so it is not necessary to impose such a restrictive limit on the sulphur content.
6.3. Blistering
Hydrogen blistering is characterised by planar cavities that appear as bulges in the steel. Like hydrogen induced cracks, blisters form when atomic hydrogen, diffusing through the steel, accumulates in voids. This form of hydrogen damage typically occurs in low-strength metals and decreases the mechanical strength of the component, which can lead to its premature failure even under moderate loads.
Blistering is due to the combined effect of atomic hydrogen and dirty steel. Semi-killed steels are particularly susceptible to blistering in both hydrofluoric acid [12] and anhydrous hydrogen fluoride [13]. It was previously believed that silicon fully killed steels performed better than aluminium fully killed steels, but this is no longer the case. It is now accepted that aluminium fully killed steels, with low sulphur and oxide levels are equally resistant to blistering and, when produced by fine grain practice, have superior low temperature toughness [10].
7. OPERATING PRACTICE
Corrosion and fouling in hydrofluoric acid alkylation units are closely linked to feed quality and operating conditions [10]. Control of the feed quality is essential in keeping contaminants out of the hydrofluoric acid alkylation unit.
- Water is a major contaminant and a key factor in ensuring the successful operation of the unit. The presence of water content promotes corrosion by reducing the acid strength and encouraging the formation of corrosive acid soluble oils.
- Sulphur, in the form of hydrogen sulphide, mercaptan, and small quantities of carbonyl sulphide, is another contaminant that can affect corrosion by reacting to produce light acid soluble oils. These are not only corrosive, but their formation reduces the acid strength further promoting corrosion.
- Oxygen is another potential contaminant and can enter the unit via several routes. If it does, it dissolves in and tends to stay with the hydrofluoric acid. If present, it will lead to accelerated corrosion of carbon steel as well as certain corrosion resistant copper-based alloys.
- The presence of nitrogen compounds such as amines, aceto-nitriles, and acrylo-nitriles also promote corrosion by reducing the acid strength and must be removed through acid regeneration.
Other contaminants that can lower acid strength may be present, but which are not normally monitored unless there is an issue. These can include arsenic, which as discussed earlier, can cause hydrogen damage if corrosion occurs.
REFERENCES
- J.D. Dobis, J. E. Cantwell, and M. Prager, “Damage Mechanisms Affecting Fixed Equipment in the Refining Industry”, WRC Bulletin 490, June 2004, Welding Research Council, New York.
- Jonathan D. Dobis, Dana G. Williams, and David L. Bryan, Jr, “The Effect of Operating Conditions on Corrosion in HF Alkylation Units - Part 1”. Inspectioneering Journal, May/June 2004 (Spring, TX: Inspectioneering, LLC).
- NACE International Publication 5A171 (2007 Edition), “Materials for Storing and Handling Commercial Grades of Aqueous Hydrofluoric Acid and Anhydrous Hydrogen Fluoride”. (Houston, TX: NACE International).
- Eurofluor-GD04-STS90 (Version 2017.11.02), “Recommendation on Materials of Construction for Anhydrous Hydrogen Fluoride (AHF) and Hydrofluoric Acid Solutions (HF)”, (EUROFLUOR, the European Technical Committee for Fluorine, Brussels, Belgium).
- H.H. Hashim, W.L. Valerioti, “Effect of Residual Copper, Nickel, and Chromium on the Corrosion Resistance of Carbon Steel in Hydrofluoric Acid Alkylation Service”, Paper 623, Corrosion/1993, (Houston, TX, NACE International, 1993).
- A.C. Gysbers, H.H. Hashim, D.R. Clarida, G. Chirinos, J. Marsh, J. Palmer, “Specification for Carbon Steel Materials for Hydrofluoric Acid Alkylation”, Corrosion/2003, paper no. 03651, (Houston, TX, NACE International, 2003).
- ASTM A106, “Standard Specification for Seamless Carbon Steel Pipe for High-Temperature Service”, 2024, (ASTM International, West Conshohocken, PA)
- “Materials of Construction Guideline for Anhydrous Hydrogen Fluoride”, revised December 27, 2004, (The Hydrogen Fluoride Industry Practices Institute, Washington, DC).
- Garofalo, F., Chou, Y.T., Ambegaokar, V., 1960, “Effect of hydrogen on stability of microcracks in iron and steel”, Acta Metall. 8, 504–512.
- H.S. Jennings, “Materials for Hydrofluoric Acid Service in the New Millenium”, Corrosion/2001, paper no. 01345, (Houston, TX, NACE International, 2001).
- C.C. Seastrom, “Minimising Hydrogen Damage in Carbon Steel Vessels Exposed to Anhydrous Hydrogen Fluoride”, 1990 Mechanical Working and Steel Processing Proceedings, p. 507-514,Iron and Steel Society.
- M.E. Holmberg, F.A. Prange, Industrial and Engineering Chemistry, Volume 37, № 11, November 1945, pp. 1030-1033.
- R.L. Schuyler, “Hydrogen Blistering of Steel in Anhydrous Hydrofluoric Acid”, Materials Performance 18, 8 (1979): p. 9.