How DEA affects CO2 removal by 
Hot Pot

In ammonia production, CO2 removal from synthesis gas is an important processing operation.

Ralph H Weiland, G Simon A Weiland and Mathew Bailey
Optimized Gas Treating, Inc.

Viewed : 2463

Article Summary

There are a number of technologies for taking raw syngas from 17% CO2 to a few 100s of ppmv. This article focuses on CO2 removal by a two-stage Benfield process, probably the most commonly-seen carbonate-based technology.

Carbonate processes are characterized by absorbers and regenerators both running very hot (typically ranging from 100-130°C) so that heat integration in the form of large lean-rich cross exchangers is unnecessary. Regeneration is forced by a pressure swing from a high absorber pressure to a low regenerator pressure, aided by steam stripping, as opposed to the temperature swing and predominantly steam stripping on which amine absorption systems rely. Nevertheless, removing CO2 is still energy intensive, and a variety of conservation schemes are used.

High temperatures in CO2 service make for a very corrosive environment. To avoid the use of corrosion inhibitors, vessels are sometimes stainless clad with stainless process piping. Characteristically, the absorber and regenerator tend to be very tall (50-60m overall) holding five or six beds of random packing, each between 5 and 8m deep.

Process Chemistry
Aqueous potassium carbonate and bicarbonate are in the form exclusively of K+, HCO3- andCO3= ions. When CO2 dissolves into water it forms carbonate and bicarbonate ions, and very little is present as molecular CO2. Hot potassium carbonate solutions are an ionic soup and the notion that these ions are associated with each other in the form, for example, of K2CO3 is fictitious. Carbon dioxide hydrolyses in solution with the hydroxide ion from dissociated water:

H2O⇌╰H+ +〖OH-                                                            (1)

CO2+〖OH-⇌〖⇌╰ HCO〗3-                                           (2)

The hydrogen ion that remains after hydrolysis immediately and instantaneously reacts with carbonate to form bicarbonate:

H+ +〖CO3=⇌╰〖HCO3-                                                  (3)

Potassium is merely a spectator ion. It takes no part in any reactions and, beyond affecting the ionic strength of the solution and its non-ideality, potassium itself has no effect on the solubility of CO2 in Hot Pot solutions. The vapour-liquid equilibrium associated with the solubility of CO2 in Hot Pot is modelled in the ProTreat® simulator on the basis of a concentrated solution of electrolytes.
The rate of the CO2 hydrolysis reaction (Reaction 2) is fairly slow because the OH- ion concentration is low, and CO2 is a sparingly soluble gas. This leads to quite tall absorption and regeneration towers without a promoter. DEA is a secondary amine and reacts readily with CO2, so its addition to carbonate solutions tends to speed up the absorption process. DEA reacts with CO2 according to the mechanism:

CO2 + R2NH⇌╰ R2NH+ COO-                                          (4)

R2 NH+ COO- + R2NH⇌╰ R_2 NH2+ + R2NCOO-           (5)

Reaction (4) occurs at finite rate while Reaction (5) involves only a proton transfer and so is instantaneous. Apart from the molecular species CO2, DEA, and of course water, the solvent again is an electrolyte soup and when combined with Hot Pot, the correct way to determine CO2 solubility is with an electrolyte model. This is the way ProTreat simulation does phase equilibrium calculations.

The amine of choice for promoting Hot Pot is DEA. As a secondary amine, DEA binds less strongly to CO2 so carbamate decomposition in the regeneration step requires less energy. MEA reacts faster with CO2, which for the same molar concentration would enhance the absorption rate; however, the cost is a higher regeneration energy requirement compared to DEA, and MEA’s absorption rate advantage can be easily achieved using DEA with a small amount of additional packing. As will be seen, a small amount of DEA also lowers the CO2 equilibrium backpressure over the treating solution.

Fractional Conversion
Fractional Conversion, Fc, is the extent to which a carbonate solvent is saturated with CO2:

If the solvent is promoted with DEA, then Fractional Conversion is:

The subscript ‘o’ signifies the concentration of the component in the completely CO2-free state, i.e., the fresh solvent before it has been exposed to carbon dioxide. DEACOOH is equivalent to R2NCOO- in Equation (5). These definitions are the exact equivalents of the term ‘loading’ as used with amines in the natural gas industry.

The ProTreat® simulator was used to develop solubility curves for CO2 in Hot Pot with and without DEA. The cases considered were 30 wt% K2CO3 and 30 wt% K2CO3 + 2.5 wt% DEA because the latter corresponds to the solvent formulation in the case study to be considered later. Figure 1 shows the extent to which 2.5% DEA reduces the CO2 backpressure at absorber lean-end conditions. Fractional Conversions between 0.1 and 0.25 have CO2 levels in the gas between 100 and 3,000 ppmv. There, 2.5 wt% DEA reduces equilibrium CO2 pressures between 10 and 40% (i.e., the ppmv ratio is 0.6-0.9).

Case Study – 1,000 MTPD Ammonia Plant
The case study is based on a 1,000 MTPD ammonia plant. The CO2 section uses the two-stage DEA-promoted Hot Pot system shown in Figure 2. The simplified drawing omits several energy conservation measures but retains the features essential to the discussion. Table 1 shows the parameters pertinent to the raw gas (Stream 3).

Both towers contained more than one type and size random packing in multiple beds. The absorber had two water wash trays at the top, and a total of 32m of packing distributed roughly equally between the lean (2600mm diameter) and semi-lean (4250mm diameter) tower sections. The regenerator was 5000mm diameter above the semi-lean draw point holding 25.6m of packing and 3050mm diameter in the lean section with 18m of packing.

This plant actually operates with 30 wt% potassium carbonate and 2.5 wt% DEA and the ProTreat® simulator predicted performance parameters very close to measured data without the need for any adjustment or manipulation of any parameters to achieve agreement between simulation and measurement. In other words, the simulation is fully predictive without adjustable parameters. The same unit was simulated without DEA, all other parameters being identical between the two cases. The effect of DEA on overall performance of both absorber and regenerator is summarized in Table 2. Using 2.5 wt% DEA provides a very satisfactory synthesis gas.

Add your rating:

Current Rating: 3

Your rate:

  • Responsive image Shut-off valves
  • Responsive image Tri block, double block and bleed valve from ZWICK
  • Responsive image Process catalysts
  • Responsive image Monthly technical bulletin
  • Responsive image PROVALVE® Fixed Valve
  • Responsive image ITW technologies online cleaning
  • Responsive image Trap more foulants with ActiPhase®
  • Responsive image Level and density in resid hydrocracking processes
  • Responsive image OpX® Digital
  • Responsive image Digital Refinery