Open tool for CO₂ capture in cement and steel

The problem

Cement and steel emit CO₂ that clean electricity alone cannot remove: part of it comes from the process chemistry itself. Capturing it at the stack is one of the few options, but the mature technology (amines) regenerates the solvent with a lot of heat.

My thesis explores another route: absorb the CO₂ with sodium hydroxide and regenerate that hydroxide in an electrochemical cell, without heat. The question: how much does it cost to capture a tonne this way, and what drives that cost?

What I built

• An open Python model that takes a plant’s flue gas and computes, step by step, the absorption column (mass transfer, chemical reaction, packing hydraulics), the cell that regenerates the solvent, and the levelized cost of capture (LCOC).
• A physical prototype: an acrylic column with 3D-printed triply periodic minimal surface (TPMS) packing, and a porous-solid-electrolyte cell.
• A web tool where anyone can change the case and see, for every result, the equation behind it, the values used and the paper it comes from.

What I found

For a Colombian cement plant with 150,000 Nm³/h of flue gas at 25 % CO₂, capturing 90 %:

• The column is the cheap part. Two columns of 6.7 m diameter with a 4.5 m bed of Pall rings cost less than 1 USD per tonne captured.
• The cell decides everything. With electrolyser costs from the DOE H2A model and grid power at 0.12 USD/kWh, capture costs about 450 USD/t: more than half is electricity and a third is the cell investment.
• Electricity price rules. At Colombia’s 2024 solar auction price (0.018 USD/kWh), the same design drops to about 244 USD/t.
• On the grid, net capture shrinks. The electricity the cell uses emits about 42 % of what is captured on the Colombian grid; with solar power that penalty almost disappears.

The practical conclusion: this technology needs cheap renewable electricity and cheaper cells per square metre. The column is not the bottleneck.

How it is built

The model uses published correlations and cites them: Billet & Schultes for the packing, Pohorecki & Moniuk for the kinetics, Weisenberger & Schumpe for solubility, Zhang et al. for the cell and its costs. Every equation has a card with its meaning, units, validity range and whether it comes from theory, literature or an own closure. The code has automated tests that compare it against published data.

It is a pre-feasibility tool (AACE class 5): for comparing options and orders of magnitude, not for designing a plant.

Mechanical Engineering thesis, Universidad de los Andes (2026). Advisors: A. González Mancera and R. Sierra Ramírez. Full text in the Uniandes repository (in Spanish). Sustainability recognition at the Mechanical Engineering Innovation Fair.