Pinched performance: Part 1

Absorbers become lean-end pinched when the solvent’s capacity for acid gas greatly exceeds the acid gas to be removed. Carbon dioxide and hydrogen sulphide removal by the fast-reacting amines MEA and DGA® are typical examples.

Ralph Weiland, Nate Hatcher and Scott Alvis
Optimized Gas Treating, Inc.

Viewed : 4627

Article Summary

Introduction: Absorbers used in amine-based gas treating almost invariably show a maximum temperature somewhere along the height of the column, or occasionally at the very top or bottom. However, their operation usually falls squarely into one of three modes: (1) lean-end pinched, (2) rich-end pinched, and (3) bulge pinched. Understanding pinches and their cause can be very useful in diagnosing column malperformance, unit bottlenecks, and in unit design and optimisation.

An operational pinch of a component occurs in a column when the actual partial pressure of the component in the vapour leaving a tray, for example, approaches closely to the equilibrium vapour pressure of that component on the same tray. When this happens, the driving force for absorption (or stripping) of the component essentially vanishes and its absorption ceases.

The concept of a pinch may be more familiar in the context of distillation where, for example, a pinch may occur as the result of an operating line lying too close to an equilibrium line when stepping off ideal stages on a McCabe-Thiele diagram. In such cases, it may be impossible to step past the pinch point on the diagram. However, pinches are not limited to distillation, McCabe Thiele diagrams, and ideal stage calculations. In fact, before the advent of selective treating, at which time it became possible to leave a substantial amount of carbon dioxide in the treated gas, most gas treating was lean-end pinched, meaning the lean end (upper end) of the tower was mass transfer pinched. Traditionally, MEA was used to remove all the carbon dioxide and hydrogen sulphide from the gas. Nearly all contactors had 20 trays and the treated gas contained concentrations of the acid gases that were essentially in equilibrium with the lean solvent entering the top of the absorber. Indeed, the gas and liquid in a substantial region of the upper part of the column were frequently in equilibrium, because the tower contained many more trays than needed to achieve the treating desired. However, even if its existence is revealed, the true nature of the pinch cannot be described by the ideal stage approach because the absorption rates of the acid gases are mutually interdependent.

The concept of a bulge pinch is relatively new, and was not recognised until it was revealed by the ProTreat® mass transfer rate-based simulator. Further discussion of bulge pinches is deferred to Part 2 of this two-part series. The rest of this article consists of a series of case studies that demonstrate the nature of rich- and lean-end mass transfer pinches. These simulation studies were all done on a mass transfer rate basis.

Lean-End Pinches: Two Case Studies
Classic Gas Treating with MEA

Before selectivity became an important consideration, gases were treated with MEA, and later with DEA and DGA, to remove the acid gases completely. Figure 1 shows the gas and liquid temperature profiles across a 20-tray absorber treating 2 x 106 SCMD of 1% H2S and 2% CO2 in predominantly methane at 66 barg. Solvent flow is 1325 L/min of 20 wt% MEA.

Across the upper 10 trays the phase temperatures change by less than 1°C, and no change is discernable across the top five trays. This has the outward appearance of a lean end pinch because temperature change is often a good first indication of the absolute extent of acid gas absorption. However, true pinch conditions can be certified only by examining composition profiles. The actual concentrations of H2S and CO2 (ppmv units) over each tray are compared with the equilibrium values in Figures 2(a) and (b). Note the logarithmic concentration scales.

(a) H2S Distribution across Absorber   

(b) CO2 Distribution across Absorber

Across the upper six trays there is no discernable departure of the hydrogen sulphide concentration from the purely equilibrium value, both of which remain constant (Figure 2a). Clearly, hydrogen sulphide absorption is truly lean-end pinched — the treated gas is in nearly perfect equilibrium with the lean solvent. The pinch criterion that the actual concentration is at the equilibrium value is met and hydrogen sulphide absorption is pinched across the upper five or six trays. Carbon dioxide absorption is nearly (but not quite) pinched only on the very top tray. Using a few more trays would generate a more clearly defined pinch for CO2 absorption.

This case is an example of the fact that a lean-end pinch does not imply both acid gases have their removal limited by a close approach to equilibrium at the top, or lean end, of the column. As seen here, hydrogen sulphide removal is finished by the time gas reaches the sixth tray from the top. Carbon dioxide continues to be absorbed across the whole column.

Deep CO2 Removal Using Piperazine-Promoted MDEA
We will investigate the use of piperazine promoted MDEA for deep CO2 removal in LNG production and other applications in detail in Part 2 of this series. However, the following case study shows the fallacy of relying solely on temperature profiles to declare a lean-end pinch.

Generic MDEA is generally unsuitable for reducing carbon dioxide to even a few tenths of a mole percent, and most attempts to use it for this purpose have failed. Indeed, we have reported a case in which a design could not be made to work even using 100 trays because the column became bulge pinched1. MDEA alone is certainly incapable of getting to the 50 ppmv level required for natural gas liquefaction. It lacks reactivity with CO2 so absorption rates are just too slow to achieve low carbon dioxide levels in a reasonable number of trays. Although other reactive amines such as MEA, MMEA, and DGA can be made to work for CO2 removal in LNG applications, the most commonly used chemical system is MDEA promoted with piperazine. At 20°C, piperazine has a second-order rate constant for reaction with CO2 about an order of magnitude higher than MMEA, the next most reactive amine.

The case concerns removing 3 mole% carbon dioxide from a gas at 65 barg using a solvent with 45 wt% MDEA and 5 wt% piperazine loaded to 0.001 moles CO2 per mole of total amine (MDEA + piperazine). Figure 3 shows the gas and liquid temperature profiles. These temperature profiles look remarkably similar to those in Figure 1 for MEA. On the basis of the uniform temperature over the upper three or four trays, one might be tempted to say that this absorber is also lean-end pinched. But Figure 4 shows this is decidedly not the case.

Although the equilibrium CO2 concentration in the gas has leveled off to a constant value of about 0.01 ppmv by the fifth tray from the top of the absorber, the actual carbon dioxide concentration in the gas continues to fall exponentially across the remainder of the column and, no doubt, would continue to do so even with the addition of another 10 trays. An absorber is lean-end pinched if the driving force for absorption shrinks to nearly zero towards the upper (lean) end of the column. This absorber shows nothing of the sort and, despite the nearly constant temperature across the upper trays, it is not lean-end pinched. In fact, it might be closer to bulge pinched, because the driving force close to the temperature bulge is minimal. Fixed temperature across the top trays is deceptive. Further discussion is deferred to Part 2.

Add your rating:

Current Rating: 4

Your rate:

  • Responsive image FCC online diagnostic services for FCCU
  • Responsive image Digital Refinery
  • Responsive image Sabin Metal profitable hydrocracking
  • Responsive image PROVALVE® Fixed Valve
  • Responsive image Galexia™ hydroprocessing platform
  • Responsive image Hardware for your operations
  • Responsive image Downstream industry process diagnostics
  • Responsive image Reactor Internals Knowledge Hub
  • Responsive image Refining GT-BTX PluS
  • Responsive image Designing deepcut vacuum units that work