Novel High Capacity Oligomers for Low Cost CO2 Capture Page: 4 of 213

                                
Executive Summary

The novel concept of using a molecule possessing both physi-sorbing and chemi-sorbing properties
for post-combustion CO2 capture was explored. A variety of candidate materials with physi-sorbing
backbones and chemically reactive peripheral groups were considered with the final selection
primarily focusing on aminosilicones. A small effort was also devoted to derivatized plant oils.
None of the plant oil derivatives were effective in absorbing CO2 in laboratory experiments. Model
reactions suggest that both intra-molecular H-bonding between adjacent hydroxyl and amino groups
and potential micelle formation suppressed reactions with CO2. This route was abandoned in favor of
the aminosilicones.
A variety of aminosilicones with differing architectures and type and placement of amine groups were
examined both experimentally and computationally for physical properties as well as CO2 capture
efficacy. Modeling indicated that the heat reaction of sterically hindered amines with CO2 was lower
than for unhindered amines and that less basic amines also decreased the heat of reaction. This
provided options to tune the heat of reaction for optimal plant performance. Physical property
predictions were also made for viscosity, vapor pressure, density and CO2 solubility. Limited
experimental data confirmed the accuracy of the density and solubility models as well as trend
predictability in heats of reaction. However, viscosity predictions were not accurately modeled and
vapor pressure data was unavailable.
Synthetically, GEN 1 aminosilicone solvents with linear, branched, cyclic and star architectures were
made that possessed mono and di-amine groups while other solvent candidates had varying degrees
of steric hindrance. Oligomers as well as discrete small molecules were prepared and evaluated. These
included solvents with covalently bound polyether units. CO2-capture experiments were performed
using both high throughput screening (HTS) techniques as well as small-scale laboratory testing. Mass
transfer issues prevented the HTS methodology from being as useful as anticipated. However, efficient
mixing on the lab-scale provided reliable data. These experiments also showed that pure solvents did
not maintain their desired liquid state on exposure to CO2. To circumvent this problem, a co-solvent
was added. The optimal co-solvent was triethylene glycol (TEG). This material prevented solids
formation upon generation of the carbamate salts.
Lab-scale experiments indicated that the aminosilicone designated as GAP-0 provided a CO2-capture
capacity in a 50/50 mixture with TEG commensurate with 30% MEA. Lab-scale absorption and
desorption experiments showed complete reversibility of the reaction with CO2 over several cycles and
isotherm data generated indicated a dynamic range (equivalent to net CO2 loading) of CO2-capture of
5-6%. This solvent composition was scaled-up and used for further studies.
A continuous absorption system as well as a bench-scale, continuous absorption/desorption system
was designed and assembled to validate the lab results seen. Under comparable conditions, the
continuous absorber system showed a CO2 absorption efficiency of >99% for both 50% GAP-0 and
30% MEA. Mass transfer coefficients for that system were of the same order of magnitude. The
continuous absorption/desorption system functioned well and provided some initial information on the
GAP-0/TEG mixture. However, solid formation during continuous operation was a serious issue. This
problem was circumvented by the designed the GEN 2 solvent designated as GAP-1. A 60/40 blend of
GAP-1/TEG allowed continuous operation of the absorber/desorber unit with no solid precipitation.
Absorption of CO2 occurred as expected in this system but desorption under pressure was less than
expected with a dynamic range of 2.6%.
Corrosion studies comparing 30% MEA and 50% GAP-0 were also conducted and showed that the
aminosilicone was equivalent or superior to the organic amine over 3000 hours. Thermal aging at 100
and 120 oC of both GAP-0 and GAP-1 showed little degradation of the materials over 80 days.

4