May-2014

Predicting corrosion rates in amine and sour water systems

A chemistry-based predictive model predicts corrosion rates in specific processing conditions. Corrosion is a ubiquitous problem in gas treating in the petroleum and natural gas industries, in syngas plants, in processing unconventional gases such as shale and coal seam gas, and in numerous other treating applications.

NATHAN A HATCHER, CLAYTON E JONES, G SIMON WEILAND and RALPH H WEILAND
Optimized Gas Treating, Inc.

Article Summary

The primary impurities removed in the treating process are the acid gases carbon dioxide and hydrogen sulphide. The corrosion of equipment and piping is an inevitable consequence of removing these very gases with amines, and of handling sour water. There are other corrosive impurities that either enter in small amounts with the gas, such as HCN and oxygen, or that are produced in the amine system itself, mostly heat stable salts (HSS) derived from HCN. Corrosion rates are affected by the nature of the corrosive agent, temperature, fluid velocity, the presence of solids, and the metallurgy involved.1 To prevent equipment failures, mitigate risk and select optimal materials, one must be able to predict corrosion rates pertinent to the particular processing conditions. This article describes the underpinnings of a chemistry-based predictive corrosion model built on both public and much proprietary corrosion rate data. The model includes dependence on ionic solution composition (speciation), fluid velocity, temperature, HSSs, and metallurgy.

Corrosion in alkaline systems
Although the concepts presented here apply equally to pH-neutral and acidic systems, these systems are not addressed because the amount of corrosion data available for modelling is not as extensive. The corrosive action of H2S is inherently different from that of CO2 in that H2S can and does form a relatively robust, protective iron sulphide layer on the metal surface. On the other hand, iron carbonate forms a more fragile layer, so it offers much less protection. There are several tenets embedded in the model:
• The corrosive agents are acids
• In and of itself, the amine (or ammonia) is not corrosive
• The iron sulphide film can protect against further corrosion
• Iron carbonate also offers protection but to a lesser degree
• High fluid velocities physically increase corrosion rates
• Higher temperature increases corrosion rates
• Heat stable salts chemically exacerbate corrosion.

Although technically incorrect, the industry continues to bandy about such terms as ‘amine corrosion’ and ‘alkaline stress corrosion cracking’ to describe corrosion that, at the root level of chemistry, is really caused by dissolved acid gases in various forms. For example, nearly 60 years ago, Polderman2 reported that 20 wt% MEA without acid gas was actually less corrosive to steel than pure water. As far as the corrosive agents themselves are concerned, the important parameter is the chemical activity of the dissolved acid gas species responsible for corrosion. The activity (vs concentration) changes with the amine type, amine concentration, acid gas loadings, the concentrations and identities of HSSs, and temperature. This may make some amine systems appear to be more susceptible to corrosion than others; however, the essential point is that it is the activity of the corrosive species that is of direct importance, not the type of amine per se.

The chemical species of interest are: bisulphide ion (HS–), free physically dissolved H2S, bicarbonate ion (HCO3–), and free physically dissolved CO2, all of which are oxidising agents. These species are called protonic acids because they can give up a hydrogen ion.3 Sulphide (S=) and carbonate (CO3=) ions are also present; however, they themselves are final reaction products and are unable to provide the hydrogen ion necessary for the oxidation of iron. Molecular hydrogen sulphide and carbon dioxide react with iron only in the presence of water. The final distribution of molecular and ionic species is found by solving the equations of chemical reaction equilibria, atom balances, and a charge balance. The resulting set of species concentrations is termed the solution’s speciation.

In their simplest stoichiometric forms, the basic corrosion reactions of dissolved H2S species with iron are:
 
H2S(aq) + Fe(s) → FeS(s) + H2(g)

2HS-(aq) + Fe(s) → FeS(s) + H2(g) + S=(aq)
 
For CO2, the relevant reactions are:
 
CO2(aq) + Fe(s) + H2O → FeCO3(s) + H2(g)

2HCO3-(aq) + 2Fe(s) → 2FeCO3(s) + H2(g)

The oxidation reaction with hydrogen sulphide is faster than the reaction with bisulphide; however, the alkalinity of the amine (and ammonia) solutions means that the dissolved H2S is predominantly in the bisulphide form, with very little remaining as free molecular hydrogen sulphide. This is also true of dissolved carbon dioxide. The concentrations of free H2S and CO2 are pH dependent and pH is a function of amine strength, total dissolved acid gas, temperature, and to a lesser extent HSS concentrations. However, heat stable salt species and their concentrations do affect the speciation of the solution, especially in lean solvents.

As discussed by Cummings et al1, the sequence of physico-chemical steps in the process of oxidising iron consists of transporting the acid from the bulk solution to the metal surface, adsorption of the acid onto the surface, reaction with iron, and transport of reaction products back into the bulk solution. The steps are similar to what occurs in heterogeneous catalysis. The reaction of H2S with the iron component of various iron-based metallurgies forms solid iron sulphide and hydrogen gas, and as the reaction proceeds, the surface of the iron is changed to a mosaic of iron and sulphide ions. The surface expands by addition of sulphide, and the liberation of hydrogen gas exacerbates the expansion. The surface layer is somewhat porous, and it adheres to the surface of the free metal. The iron-H2S and iron-HS– reactions form reaction products in completely different phases, as do the reactions with carbon dioxide and bicarbonate. Thus, because reaction products are continually removed from the reacting solution, it follows from Le Chatelier’s principle that there is a strong thermodynamic driving force powering continued corrosion.
What limits the corrosion reactions is primarily the amount of bare, unreacted iron that the passivating film leaves available at the metal surface. A secondary factor is the concentrations of dissolved reactant gases, H2S and CO2, which are other parameters. Thus, H2S, HS–, CO2 and HCO3– all react with unprotected iron. To control the concentration of dissolved gas in their various forms, most practitioners adopt a rich amine acid gas loading, upper limit of <0.4 to 0.5 moles of total acid gas per mole of molecular amine.