Apr-2013

Solubility of hydrocarbons and light ends in amines

Modelling procedures predict with considerable certainty what the expected hydrocarbon content of an amine solution in a gas processing plant ought to be

NATHAN HATCHER, CLAYTON JONES and RALPH WEILAND
Optimized Gas Treating

Article Summary

In processing and treating both wellhead and refinery gases for H2S and CO2 removal, an unfortunate side effect is the loss of hydrocarbons and other light ends gases that are also soluble to varying degrees. Hydrocarbon solubility is never good news, unless the solvent is intentionally being used for dew point control. Any hydrocarbons that dissolve into the solvent in the absorber must reappear somewhere else and be discharged from the process. Gas removed from the rich solvent via a flash tank immediately downstream from the absorber (so-called “flash gas make”) is mostly hydrocarbon but with high levels of CO2 and H2S. It may need to be treated further in a low-pressure absorber to remove H2S before it can be used as fuel gas, for example, so its flow rate and composition might need to be known with reasonable accuracy. Incidentally, gas dissolved in the amine solvent is not the only way hydrocarbons can enter the flash tank. Carry-under of gas entrained (but not dissolved) in the solvent can be a much bigger source than dissolved gas if care is not taken to prevent it.

Once the rich solvent enters the regenerator, its hydrocarbon content is almost completely stripped and it enters the sulphur plant along with the acid gas. Low molecular weight hydrocarbons take up air demand and hydraulic volume in the sulphur plant, lowering sulphur recovery capacity and efficiency. BTEX aromatics are real troublemakers in sulphur plants and can lead to catalyst deactivation in units with acid gas feeds that are lean in H2S. With richer feeds, and especially oxygen-enriched units, temperature moderation problems can result. In some cases, heavy hydrocarbons in the right boiling range can be steam stripped and trapped in the overhead circuit of the regenerator, leading to foaming.

When conditions are right, hydrocarbons can form a second liquid phase. Foaming then becomes almost inevitable, and it is usually alleviated by the spewing of foamy solvent overhead. Although not a solubility issue, foamed solvent may also carry large amounts of hydrocarbons from the absorber bottom and into the flash tank, where they contribute in a large way to flash gas make rates far in excess of what would be expected from solubility alone.

To estimate solubility in amine treating solutions, a good place to start is with the solubility in water, because water makes up at least 85 mol% of most treating solutions. However, amines are organic molecules and have a much higher affinity for hydrocarbons than water does. Thus, hydrocarbon solubility in amine-treating solvents should be expected to be higher than in water, and the solubility should depend not only on the particular amine but also on its strength. One might say that the amine “salts” in hydrocarbons. Dissolved acid gases, on the other hand, cause what is termed “salting-out” of hydrocarbons (and fixed gases). When acid gases dissolve in (and react with) the solvent, they produce copious amounts of ionic components such as protonated amine, sulphide and bisulphide ions, carbonate and bicarbonate and, in the case of primary and secondary amines, amine carbamates. These are all ionic and result in a high total ionic strength. This causes dissolved hydrocarbon and fixed gas concentrations to be much lower than would be expected simply on the basis of solubilities in unloaded amines. Unfortunately, unloaded amines are usually what have been used to assess hydrocarbon and fixed gas solubilities in amine-treating solutions. Assessments based solely on unloaded solvent data can be erroneously high, by 50% at high loadings.

This article aims to place hydrocarbon and fixed gas solubilities in amine-treating solutions on a solid, generalised footing. The intended result for practitioners is the ability to predict with considerable certainty just what the expected hydrocarbon content of any amine solution in a properly operating plant ought to be. This is particularly important for rich solvents under absorber bottom conditions, because the rich amine is the normal, unavoidable source of hydrocarbons in flash gas make and Claus plant feed streams. And it is precisely here that ion-induced salting-out is at its strongest. The importance of being able to make this prediction accurately is that it provides a quantitative assessment of the situation expected under normal circumstances in a well-managed facility. If the flash gas make rate is higher, there is probably carry-under from the absorber either as entrained gas or as an emulsified second hydrocarbon phase. It points troubleshooting in the right direction, and it also allows a downstream flash gas scrubber to be properly sized.

Modelling and theory
The basis for determining the solubility of any sparingly soluble gas, i, uses temperature-dependent Henry’s Law data for pure water, Hio and a Setschenow (1892) salting (in and out) coefficient Si to compute the Henry’s Law constant in the actual solution:

Hi = Hio
        Si

There is extensive literature on the Setschenow approach.4 The original work of Setschenow provides a method for calculating phase equilibria for systems for which conventional activity coefficient models are unsuited. The trick, however, is in determining the Setschenow coefficient for the gas of interest. This coefficient consists of two terms. There is a contribution from the molecular amine (a salting-in contribution given by kCa) and the various ions (salting-out contributions):

In(Si) = kCa - 2.302585 h I

where I is the ionic strength. The system-dependent parameter h has anion, cation and gas contributions:

h = h_ + h+ + hg

(Ionic strength is defined as 
I = ½∑m3z23, where m3 is the molality of the salt ion and z3 is its charge number.)
The value of the salting parameter, k, depends on the specific hydrocarbon amine pair and is also temperature dependent. The salting-out part depends on the specific amine, with contributions from protonated amine, carbamate, bicarbonate, carbonate and, in principle at least, on bisulphide and sulphide ions, although no measurements appear ever to have been reported of the interactions of HS- and S= with any soluble gas (a worthy subject for research).