Benchmarking solvents for carbon capture

Existing technological approaches to carbon capture (CC) from conventional power plant flue gases are almost all based on solvents.

Ralph H Weiland, Nathan A Hatcher and Jaime L Nava
Optimized Gas Treating

Viewed : 2995

Article Summary

High-strength MEA has received the most attention and is the solvent against which other technologies are usually benchmarked. Other solvents include 2amino-2-methyl-1-propanol (AMP), piperazine, cold ammonia, a variety of caustic-neutralised amino acids, amine-promoted potassium carbonate, and physical solvents including ionic liquids, a new and interesting class of fluids. Our focus in this paper is with AMP, sodium glycinate, and the MEA benchmark.

CC from atmospheric-pressure flue gas presents a unique set of difficulties usually not experienced in more-conventional gas treating. On the absorption side, problems arise from the gas being at low pressure, and its unavoidable oxygen content. Even a small power plant generates an enormous volume of flue gas at atmospheric pressure which needs sizable power to drive the gas through the contactor. The driving force for CO2 absorption is its partial pressure in an atmospheric-pressure gas. Amine vapourisation losses and oxidative degradation are also problems. The over-riding drawback however is prohibitive regeneration energy requirements.

Fast-reacting carbamate-formers like MEA, although very effective at reacting with and removing CO2, are energy-intensive to regenerate. However, the technology of using MEA in gas treating is rather better established than other solvents, and it remains high on the list of interesting CC solvents.

The higher loading potential of noncarbamate-forming amines (one amine molecule for each molecule of CO2 absorbed) has made the moderately hindered amine AMP more interesting. AMP is a primary amine, but the secondary methyl group shields the amino group to a significant extent, and carbamate formation is made more difficult. Because the reaction product is carbonate rather than carbamate, regeneration energy ought to be lower than for MEA. Steric hindering also means that each CO2 molecule theoretically uses only one AMP molecule, potentially doubling the capacity of the solvent.

Academic laboratories began to characterise AMP almost as soon as Exxon’s first Flexsorb patents were issued more than 20 years ago. In the literature today there are enough phase equilibrium, kinetics, and physical property data of good quality and reliability to allow AMP to be process simulated with fair accuracy. AMP was added to the ProTreat amine simulator’s solvent offerings in mid-2009.

Salts of amino acids have been used since 1935 for acid gas removal from refinery, coke oven, and natural gas, among others, mostly in Europe and especially in Germany. These are the so-called Alkazids (potassium salts of N,N-dimethylglycine and Nmethylalanine) developed by BASF. Within the last five years, interest has developed in the sodium and potassium salts of glycine, the simplest amino acid, for CO2 capture.

Sufficient kinetic and equilibrium data have now been published to permit the detailed simulation of a CO2 capture plant using this solvent.

Sodium glycinate (NaGly) has recently been implemented within ProTreat. Unlike AMP, NaGly is a carbamate former because the amino group in neutralised glycine is perfectly capable of reacting with CO2, just like any other primary amine. It has the advantages of zero volatility (NaGly is a salt), and greater resistance to oxidation (it is already oxidised to an acid) and thermal degradation (NaGly is a small, stable molecule).

MEA: benchmarking standard
the basis for developing this benchmark and making subsequent comparisons is 30 wt% MEA to remove 3000 tonnes/day (3,300 short tons/day) CO2 produced by burning coal in a roughly 300 MW power station. By world standards, this is a small power station, but of sufficient scale to be indicative of solvent performance.

We are interested in simply comparing the regeneration energy required by AMP and NaGly relative to MEA, not in developing intricate processing schemes to minimise energy consumption per se, nor in minimising water and amine losses from the plant. So as long as comparisons between solvents are made on the same basis, small efficiency gains made by tweaking the flowsheet will have minimal effect. The gas volumes in CO2 capture are extremely large and providing energy to overcome absorber pressure drop can be significant. Therefore, the PFD included a booster blower but it omitted any peripheral equipment designed to prevent solvent vapourisation losses with the treated flue gas or do any form of heat integration beyond a conventional lean-rich cross-exchanger (Figure 1).   

Because only a relatively low recovery of CO2 is ever required, the most economical way to operate the absorber is with it rich-end pinched. In other words, the loaded solvent should be as close as practicable to being in equilibrium with the entering flue gas—in this case 0.5 mol/mol. So, although the temperature and flow rate may vary, the regenerator is always presented with 30 wt% MEA loaded to 0.5 mol/mol. The discussion mostly focuses on 85% CO2 recovery and the rich-amine feed temperature to the 20-tray regenerator was set at 96°C (205°F) to avoid a temperature cross in the lean-rich exchanger from a regenerator operated at relatively low pressure (typically 1.5–2.5 bar, 10–20 psig).

Mellapak-Plus M252.Y packing was taken as the base case for comparison with AMP. This structured packing has a reasonable specific area of around 250 m-1 (76 ft-1) with special treatment at the top and bottom of each block to reduce pressure drop. It was found that 85% recovery could be safely achieved over a range of solvent rates from 1400 to 2000 m3/s using 115 MW reboiler energy (110 MW barely made 85% recovery at a single solvent flow and 105 MW could not reach 85% recovery at all). 

As circulation rate was lowered in the simulations, the rich amine continued to leave the scrubber at essentially a constant loading of 0.5 mol/mol, and the treating capacity lost through reduced circulation was very closely made up by improved regeneration (at constant reboiler duty). At a low enough circulation rate, of course, the stripability limit of the solvent (about 0.1 mol/mol for MEA) is reached and very little further reduction can be realised; thus, the recovery curve dropped markedly below a circulation rate of about 1,400 m3/hr (6,160 gpm). Conventional gas treating experience would suggest that if the circulation rate can be dropped by 30% while still treating satisfactorily, then the reboiler duty ought to be able to be decreased in the same way, since obviously the solvent has excess capacity that’s not being used. However, in the case of CO2 capture, the absorber is always operated severely rich-end pinched, so this thinking simply doesn’t apply.

Add your rating:

Current Rating: 1

Your rate:

  • Responsive image Downstream industry process diagnostics
  • Responsive image Reinventing reliability
  • Responsive image Galexia™ hydroprocessing platform
  • Responsive image Multi-element and total S process analyzers
  • Responsive image Get the right answer faster
  • Responsive image Advanced sulphur analysis in hydrocarbons
  • Responsive image Axens on Linkedin
  • Responsive image Worldwide refinery processing review
  • Responsive image Becht fired heaters services
  • Responsive image Hydrotreating and hydrocracking catalysts