Next-gen carbon capture: Life cycle metrics of solar-integrated direct air capture technology

[Geoengineering News]

https://www.sciencedirect.com/science/article/pii/S2590123026017627

Authors: Mohsen Nasrollahi, Farzin Hosseinifard, Mohsen Salimi, Majid Amidpour

05 May 2026

https://doi.org/10.1016/j.rineng.2026.110729

Highlights

•PV shows marginally lower total environmental impacts, ∼0.2–0.3% than CSP options.

•CSP-PTC exhibits minimally higher exergy demand at 1.262 MWh per ton of CO₂.

•KOH production dominates exergy demand, contributing 1.097 MWh per ton of CO₂.

•Weighted ReCiPe scores indicate higher total damages for CSP-PTC than PV.

•Human health impacts account for over 90% of total ReCiPe endpoint damages.

Abstract

This study presents a cradle-to-gate life cycle assessment (LCA) of a Direct Air Capture (DAC) system powered by three renewable energy configurations: Photovoltaic panels (PV), parabolic trough collectors (CSP-PTC), and solar tower technology (CSP-T). The DAC process was modeled under steady-state conditions using Aspen Plus V11, while the solar energy systems were simulated in Thermoflex and PV-syst to ensure consistent thermodynamic evaluation. Environmental impacts were assessed in SimaPro using the CExD and ReCiPe 2016 methods for a functional unit of 1 ton of captured CO₂. Results indicate that CSP-PTC exhibits marginally higher exergy demand (1.262 MWh/ton CO₂), followed by CSP-T (1.246 MWh/ton CO₂) and PV (1.244 MWh/ton CO₂). Potassium hydroxide (KOH) is identified as the dominant contributor to exergy demand across all scenarios. ReCiPe midpoint and endpoint results show that PV yields slightly lower environmental impact, with total weighted damages of 70.515 Pt, compared with 70.606 Pt for CSP-T and 70.804 Pt for CSP-PTC. More than 90% of total damage is associated with human health, largely due to the ReCiPe weighting structure. Overall, while PV-powered DAC shows marginally lower environmental impacts across all assessed categories, the three renewable configurations perform comparably given the small magnitude of differences and uncertainties inherent in LCA modeling.

Source: ScienceDirect