May-2018

Accuracy of measured lean amine 
H2S loading

When solvent regeneration is done robustly the most accurate assessment of lean loadings can be by simulation rather than by measurement.

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

Article Summary

The regenerator in an amine unit for acid gas removal is a huge consumer of energy, accounting for roughly 80% of the unit’s opex. Yet the regenerator is probably the most neglected and imperfectly understood piece of equipment in the plant. The purpose of the regenerator of course is to strip acid gases from the solvent to prepare it for reuse in the absorber where gas is actually processed to meet some treating specification. Treated gas purity is at least partially, and often completely, determined by the loading of acid gas in the lean solvent produced by the regenerator. Absorbers in this condition are called lean-end pinched, a normal state for tail gas treating units (TGTU) for example.

The absorber and regenerator operate together in a closed circuit and are obviously mutually dependent, with solvent lean loading being a parameter critically important to the performance of the unit as a whole. Wrong values of solvent lean loading can have enormous consequences in design and troubleshooting. For example, if the lean H2S loading is wrongly measured to be a lot lower than it really is, the regenerator designed to produce a low measured loading will be specified with the steam flow rate to the reboiler considerably higher than necessary. This translates into a higher anticipated opex to produce the falsely low loading value. Consequently, the reboiler will end up being oversized too, and the needlessly high capex and opex may collude into a failed revamp bid.

There are a number of methods (mostly wet chemical as well as gas chromatography) to measure lean acid gas loading, but all of them rely on amine solvent samples collected from somewhere within the gas treating unit. How lean amine samples are drawn, how they are transported, and how they are treated during analysis can have a significant effect on the reliability of the analytical results. The issue is one of assuring sample integrity. Should one be sceptical? The question to be addressed then is how reliable the loading values as reported by the laboratory really are and how reliability can be improved by proper and careful sample handling.

Chemistry of H2S oxidation
H2S is produced naturally by the anaerobic breakdown of organic matter, but it is also slowly oxidised in nature to a variety of product species. Air oxidation of H2S in fresh and seawater is an area of great interest in water treatment and environmental science, as is the oxidation of sulphide and bisulphide.

At concentrations of H2S in water below 2 mg/l, aeration is considered to be an ideal process for completely oxidising it to sulphate. This level of H2S is roughly equivalent to a lean loading of 0.000015 in 45 wt% MDEA. The message is that at this loading, H2S is likely to disappear altogether from a sample exposed to air – the laboratory will probably report zero. At higher loadings it will still tend to disappear in an aerated sample; just perhaps not completely.

Luther et al¹ have shown that the kinetics of H2S oxidation in seawater follows the kinetic expression:

Vazquez et al² have shown that the addition of trace metals, in particular Fe(II), to the reaction vessel they used in their study increased the rate of oxidation such that the half-life of HS– is on the order of minutes in alkaline solutions at room temperature.

The form of H2S that oxidises is bisulphide ion, not molecular H2S. But it should also be remembered that when H2S dissolves in water it immediately dissociates into H+ and HS– ions. The dissociation reaction is instantaneous so as soon as an HS– ion is consumed it is immediately replaced, and oxidation proceeds at constant rate until either oxygen or H2S is completely consumed.

Iron, particularly in the form Fe(II), is a catalyst for HS– oxidation, and it is a very unusual gas treating solvent that is free from iron, and indeed other ionic metal contaminants. The scene is set for H2S oxidation in amine solvents. Amine solvent samples are usually collected into air filled sample bottles, they are shipped to the laboratory in air-filled containers, the samples are then given a good shaking in their air-filled bottles, they are pipetted for analysis without taking any care to exclude oxygen, and then they are analysed in the presence of air. Consequently, the measured solvent loading invariably will be wrong, at least to some degree, and the leaner the original sample, the more erroneous the analysis in relative terms.

The presence of the heat stable salts (HSS) sulphate and thiosulphate in an amine sample is often indicative of HS– oxidation already having taken place, either by oxygen ingress into the treating system, or quite possibly by oxygen introduced into the lean amine sample itself. Oxidation is invariably worse in rich amine samples where a greater abundance of H2S reactant is present; however, the relative errors are lower, maybe to the point of being unnoticeable.

Incidentally, there is an old wives’ tale to the effect that purging regenerator reflux also purges HSS. This tale probably originates from reflux water samples sometimes containing small amounts of sulphate and thiosulphate. However, HSS are ionic species and, therefore, they are completely non-volatile; after all, they are salts just like their name says they are, so the question should be how they got into reflux water in the first place. Perhaps the answer is that they were formed in the reflux water via oxidation of the bisulphide present there (by oxygen dissolved in the rich amine feed to the regenerator), or perhaps they were formed simply by oxidation of H2S by air introduced in the reflux water sampling and analysis procedure. The only other source is liquid entrained with the vapour leaving the feed tray of the regenerator. Purging reflux water removes ammonia and trapped H2S quite effectively. However, any sulphate or thiosulphate it removes via blowdown was either generated in situ in the actual reflux water or samples of it, or it arrived via entrainment. Purging via reflux blowdown is certainly not an effective way to remove HSS from the system. Instead, one should either look for the reason HSS are being generated in the overhead system, or deal with excessive entrainment by proper tray design and demister selection.

The fact that amine treating is carried out in carbon steel vessels makes it a near certainty that every amine sample from an industrial facility will contain a measurable level of dissolved iron in the form Fe(II), a very effective catalyst for HS– oxidation. It is also a fact that there are multiple opportunities for oxygen to find its way into an amine sample. Even taking the sample into an air-filled sample bottle is sufficient to compromise the sample.

A solid mass transfer rate based regenerator model is perfectly capable of producing an accurate rendition of the lean loading to be expected in a well run system. One of the most frequently overlooked and violated physical requirements of a sample analysis is that the measured ionic composition meets a charge balance; that is, the analysis leaves the solution charge neutral. There is a very useful spreadsheet tool available at www.protreat.com/downloads that can be used to ensure a given chemical analysis meets the charge neutrality requirement. It also suggests how the charge balance can be enforced by appropriately adjusting the measured H2S loading of the solvent. Because ProTreat calculations always enforce charge neutrality, the simulator’s calculated value of lean H2S loading is likely to be more accurate than all but the most carefully done measurement.