(97h) Corrosion Analysis of Non-Aqueous Solvent for Carbon Capture

RTI International is developing a novel non-aqueous solvent (NAS) based process for post-combustion CO₂ capture. The process has the potential of substantially reducing the thermal energy required for solvent regeneration compared to the NETL baseline cases.¹ RTI has already shown the reduction of thermal energy using NAS in a Bench-scale Gas Absorption System (BsGAS) at RTI with simulated flue gas, a pilot-scale unit at SINTEF’s Tiller facility in Trondheim, Norway with coal-derived flue gas, and testing at the National Carbon Capture Center (NCCC) in Wilsonville, Alabama with coal-derived flue gas. Further NAS testing was completed at Technology Centre Mongstad’s flexible amine-based capture plant (TCM DA) in Mongstad, Norway during March-September 2022 with 2752 hours of time on stream. This plant can treat up to 60,000 Sm³/hr flue gas supplied from a combined heat and power (CHP) plant and residual fluidized catalytic cracker (RFCC).

The capital cost of CO₂ capture plants is one of the largest drivers in the cost of CO₂ captured. As a result, efforts to reduce capital costs are one of the most effective ways to minimize the cost of CO₂ capture technologies. Most capture systems are currently built with stainless steel due to the corrosive nature of MEA and other aqueous solvents, but cheaper alternatives including carbon steel or plastic liners may be suitable for solvents like NAS. In addition to testing the performance of NAS under various operating conditions at TCM DA, corrosion due to the use of NAS was also monitored to identify suitable materials of construction (MOC) for future capture plants and to generate an accurate techno-economic analysis (TEA). Corrosion coupons made of a variety of materials (carbon steel, stainless steel, elastomer, resin) were inserted in the piping downstream of the absorber and reboiler and in the regenerator overhead condenser vapor stream. At the end of testing, the metal coupons were analyzed for corrosion rates, pitting, and stress cracking; elastomer and resin coupons were analyzed for changes in material hardness and mass. Solvent samples were also collected to determine dissolved metal, heat stable salt (HSS), water content, and conductance.

NAS (0.01 µm/yr) exhibited virtually no corrosion for CS 1018 under absorber conditions compared to MEA (4170 µm/yr)² due to the lower water content of the solvent. Reported literature values for aqueous solvent corrosion rates and NAS corrosion results for previous testing campaigns (NCCC and SINTEF) showed corrosion rate can be well predicted by a power law correlation with the conductivity of the solvent. The corrosion rate for stainless steel was about one order of magnitude lower than for the carbon steel. However, this correlation greatly overpredicted the NAS corrosion rates observed at TCM.

Solvent was routinely sampled at various points in the system during steady state operation to track system operation and detect any changes in solvent composition. Samples were analyzed for amine, water, and CO₂ content from both the absorber and regenerator side of the system. In addition, the concentrations of metals in the samples were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) for metal content. Chromium (Cr), iron (Fe), and nickel (Ni) are present in the TCM DA MOC and are not present in the flue gas in significant quantities, therefore the presence of these metal ions indicates corrosion in the process equipment. In a comparable TCM DA test campaign with MEA, concentrations of Fe, Cr, and Ni ions were an order of magnitude higher than concentrations measured in NAS.³ For instance, a maximum of ~17 ppm Fe was seen in the MEA campaign, whereas ~1.1 ppm Fe was observed in the NAS campaign.³

Fischer et al. (2017) suggested that a positive linear correlation exists between the rate of corrosion and the rate of formate ion production.⁴ Formate concentration was monitored throughout the TCM DA campaign but remained below detectable levels. Comparatively, studies of lean and rich 7 m MDEA (45.5 wt%) showed formate production of >400 ppm and >150 ppm respectively after 14 days.⁵ NAS’ ability to inhibit production of formate and other HSS may be a key contributor to reduced corrosivity.

The low corrosion rate of NAS indicates that CO₂ capture plants could use low-cost MOC, lowering capital costs significantly and reducing the cost of CO₂ capture. A TEA will also be included to show the reduction in cost of capture from the use of cheaper materials.

1. Schmitt, T.; Leptinsky, S.; Turner, M.; Zoelle, A.; White, C. W.; Hughes, S.; Homsy, S.; Woods, M.; Hoffman, H.; Shultz, T.; James III, R. E. Cost and Performance Baseline for Fossil Energy Plants Volume 1: Bituminous Coal and Natural Gas to Electricity; National Energy Technology Laboratory (NETL): United States, 2022.

2. Gunasekaran, P.; Veawab, A.; Aroonwilas, A., Corrosivity of Single and Blended Amines in CO2 Capture Process. Energy Procedia 2013, 37, 2094-2099.

3. Hjelmaas, S.; Storheim, E.; Flø, N. E.; Thorjussen, E. S.; Morken, A. K.; Faramarzi, L.; de Cazenove, T.; Hamborg, E. S., Results from MEA Amine Plant Corrosion Processes at the CO2 Technology Centre Mongstad. Energy Procedia 2017, 114, 1166-1178.

4. Fischer, K. B.; Daga, A.; Hatchell, D.; Rochelle, G. T., MEA and Piperazine Corrosion of Carbon Steel and Stainless Steel. Energy Procedia 2017, 114, 1751-1764.

5. Closmann, F.; Nguyen, T.; Rochelle, G. T., MDEA/Piperazine as a solvent for CO2 capture. Energy Procedia 2009, 1 (1), 1351-1357.