Syngas CO2 capture on a steam methane reformer matches grey hydrogen at a carbon price of 54 USD/t, and flue-gas capture at 164 USD/t

CheAI Research Inc. | Research article | October 2026

Abstract

1. Introduction

In an SMR, the furnace burns the pressure swing adsorption (PSA) tail gas and trim natural gas, so CO₂ leaves both in the syngas and in the flue gas. The IEA Greenhouse Gas R&D Programme (IEAGHG) studied capture on a merchant-hydrogen plant [1], [2]. Syngas aMDEA captured 55.7 % of its CO₂. Flue-gas MEA captured 90.0 % at a higher energy and capital cost. Earlier work compared the capture points [3], and the debate over blue hydrogen turns on the capture rate and on upstream methane leakage [4], [5].

The IEAGHG plant has a pre-reformer and a steam cycle. The 2019 design studied here [6] has neither, and it fires cold combustion air. For such a plant, the capture point with the lowest cost per tonne of CO₂ avoided was not known, and the ranking could depend on how surplus heat is valued. We found that syngas capture had the lowest avoidance cost, and combined capture reached a higher capture fraction than flue-gas capture alone at a lower avoidance cost in all three thermal scenarios.

2. Methods

The 2019 design reformed 416.5 t/d of natural gas at 1000 °C, 20 bar and a steam-to-carbon ratio of 3.0. We modelled the reformer and shift reactors at equilibrium in DWSIM 9.0.5 [7]. Each capture unit was a separator with a fixed CO₂ removal and a literature reboiler duty, 98 % and 1.6 GJ/t for aMDEA and 90 % and 3.6 GJ/t for MEA.

The plant exports no steam, so we evaluated three thermal scenarios. In the primary heat-valued scenario, an auxiliary natural-gas boiler at 90 % efficiency raised the reboiler steam. In the free-heat scenario, rejected process heat raised it. In the heat-integrated scenario, flue gas preheated the furnace air and fuel to 400 °C. Well-to-gate emissions added upstream gas at 12.1 g CO₂e/MJ, re-based from GREET 2022 [8], and grid electricity.

The LCOH came from a 20-year discounted cash flow at a 10 % discount rate and 21 % tax, the terms of a published proton-exchange-membrane (PEM) electrolysis study [9]. Capital was scaled from IEAGHG [1], and natural gas cost 4.77 USD/MMBtu [10]. The carbon price at which a blue case matches grey is

c* = 10³ (LCOH_blue - LCOH_grey) / (e_grey - e_blue),

with LCOH in USD/kg at zero carbon price and e the direct emission in kg CO₂ per kg of hydrogen. All runs converged within 0.5 % closure of mass and carbon and 2 % closure of energy. An external audit reran selected cases and recomputed the results with an independent script.

3. Results

3.1 The 2019 report is internally inconsistent

At the stated conditions the rebuild makes 184.3 t/d of hydrogen, 18.2 % more than the report's 156 t/d. The report's flows imply 80.6-81.9 % methane conversion, against 98.3 % at its stated equilibrium, and we logged 11 inconsistencies. A reformer temperature fitted to 838.5 °C recovers the hydrogen output but not the shift volumes, and implies an implausible approach to equilibrium. We kept the stated basis, and the fitted case moves the grey LCOH from 1.988 to 2.085 USD/kg and the capture fractions by less than one point.

3.2 Syngas capture removes 57.6 % of the carbon, combined capture 95.1 %

The grey plant emits 10.22 kg CO₂ per kg of hydrogen, 13.7 % above the IEAGHG base case. Syngas capture lowers this to 4.30 kg/kg and removes 57.6 % of the natural-gas carbon, because the furnace still burns the tail-gas methane slip, CO and the trim gas. Flue-gas capture reaches 89.9 % and 1.28 kg/kg, and combined capture 95.1 % and 0.55 kg/kg (Figure 1a). Capture adds upstream gas and electricity, so the well-to-gate footprint falls less, from 12.48 kg CO₂e/kg to 6.71, 4.53 and 3.40.

3.3 The energy penalty depends on how surplus heat is valued

The grey plant rejects 102.6 MW of process heat, enough to raise all the reboiler steam. In the heat-valued scenario, B costs 4.33 GJ per tonne captured and A+B 1.98 GJ/t. Case A shows -0.002 GJ/t, because removing CO₂ from the syngas removes ballast that the furnace would otherwise heat. This saves 11.91 MJ of furnace gas per kg of hydrogen, while the aMDEA reboilers burn 10.39 MJ (Figure 1b). The near-zero penalty is a cancellation, not free capture. With 400 °C feed preheat the furnace burns less trim gas, and A captures 69.1 %.

3.4 Flue-gas capture sets the cost of deep capture

Grey hydrogen costs 1.988 USD/kg. Capture raises it to 2.307 (A), 3.453 (B) and 3.086 USD/kg (A+B) (Figure 2a). B needs the most hardware and steam, with a capture plant of 262.5 MUSD against 55.7 MUSD for A. A+B costs less than B, because the syngas unit removes most of the CO₂ at low cost. The break-even carbon prices are 53.75, 163.81 and 113.46 USD/t. Combined capture costs 2.11 times as much per tonne avoided as syngas capture. Flue-gas capture alone costs 3.05 times as much. The break-even prices depend little on the front-end cost, which grey and blue share. Across the sensitivities, A stays cheapest per tonne avoided and B most expensive (Figure 2b).

4. Discussion

Syngas capture is cheapest because its CO₂ is concentrated and pressurised. It reaches only the carbon that passes through the shift, and the furnace heat balance sets how much bypasses it. Case A removes 57.6 %, 61.0 % or 69.1 % in the three scenarios. Among the configurations tested, about 90-95 % capture required flue-gas treatment, and syngas capture in front of it lowered the break-even price from 164 to 113 USD/t. Our capture fractions are within 2 points of the IEAGHG cases [1], but only for comparable furnace balances, since furnace integration moves the syngas fraction by 11.5 points.

The published PEM reference design costs 11.34 USD/kg [9], 3.3 to 5.7 times the cases here. The reference is one screening design from the same organisation and delivers hydrogen at 30 bar against our 20 bar. The comparison applies to that design only. On the US-average grid it emits 22.9 kg CO₂e per kg at 80 bar [11], more than grey SMR. Dedicated low-carbon electricity can bring it below 1 kg CO₂e per kg [11].

5. Limitations

This is a screening model of one design, with equilibrium reactors, a PSA of fixed recovery and purity, and pure methane feed. Capture units are separators with literature reboiler duties. No amine column, solvent chemistry, co-absorption or hydraulics is solved, so CO₂ purity is 1.0 by construction. The furnace model omits the air and fuel preheat that the 2019 report mentions but does not size. This raises the grey footprint above IEAGHG and makes the ballast saving specific to this flowsheet. The heat-integrated scenario has no steam export, pre-reformer, exchanger design or preheater capital, and none of the three scenarios brackets a real plant.

Several inputs are weakly sourced. The aMDEA duty of 1.6 GJ/t is the least well-sourced input; at 3.2 GJ/t the penalty of A is +1.78 GJ/t. The 90 % boiler efficiency is assumed. The boiler raises 87.93 MW of steam in case B but carries no capital charge, which understates the cost of flue-gas capture. The upstream methane factor is an approximate re-base of GREET 2022, not the value behind the PEM footprint, and leakage above the 2.3 % tested [12] would raise the blue footprints.

On cost, the absolute LCOH is for a new, fully capitalised front end, not a retrofit, although the front end largely cancels in the break-even prices. Capital accuracy is roughly -30/+50 %. The battery-limits split is assumed, the 2014 cost index comes from a secondary source, and the 2019-method index was recalled from a textbook and not checked. Monetary years are also mixed: a 2 % gas-price deflation would lower the grey LCOH from 1.988 to 1.970 USD/kg.

Finally, hydrogen firing in the furnace and low-carbon electricity for the PEM plant were not modelled. The PEM cost was reproduced only within 6.4 % and was held fixed in all sensitivities, and the audit did not derive the plant thermodynamics independently.

6. Conclusions

With reboiler heat charged as natural gas, syngas capture on this SMR removes 57.6 % of the natural-gas carbon at a break-even carbon price of 54 USD/t. Flue-gas capture removes 89.9 % at 164 USD/t, and both together 95.1 % at 113 USD/t. Among the configurations tested, about 90-95 % capture requires flue-gas treatment, and combined capture has a lower avoidance cost than flue-gas capture alone. The energy penalty depends on how surplus heat is valued, so a capture decision for an existing plant needs that plant's heat balance.

References

1. IEAGHG. Techno-Economic Evaluation of SMR Based Standalone (Merchant) Hydrogen Plant with CCS. Technical Report 2017/02, IEA Greenhouse Gas R&D Programme, Cheltenham, 2017.
2. Collodi G, Azzaro G, Ferrari N, Santos S. Techno-economic evaluation of deploying CCS in SMR based merchant H2 production with NG as feedstock and fuel. Energy Procedia 114, 2690-2712, 2017. DOI 10.1016/j.egypro.2017.03.1533
3. Soltani R, Rosen MA, Dincer I. Assessment of CO2 capture options from various points in steam methane reforming for hydrogen production. International Journal of Hydrogen Energy 39(35), 20266-20275, 2014. DOI 10.1016/j.ijhydene.2014.09.161
4. Howarth RW, Jacobson MZ. How green is blue hydrogen? Energy Science & Engineering 9(10), 1676-1687, 2021. DOI 10.1002/ese3.956
5. Romano MC, Antonini C, Bardow A, et al. Comment on "How green is blue hydrogen?". Energy Science & Engineering 10(7), 1944-1954, 2022. DOI 10.1002/ese3.1126
6. Ho T. Advanced Ammonia Plant Design. Design report (team report, unpublished), University of California, Berkeley, 2019.
7. Medeiros D. DWSIM, open-source chemical process simulator, version 9.0.5. 2025. https://dwsim.org
8. Elgowainy A, Frank ED, Vyawahare P, et al. Hydrogen Life-Cycle Analysis in Support of Clean Hydrogen Production. Argonne National Laboratory, Lemont, IL, 2022. DOI 10.2172/1892005
9. CheAI Research. Green hydrogen PEM plant, techno-economic analysis. 2024. cheairesearch.com/reports/green-hydrogen-tea.html
10. U.S. Energy Information Administration. Short-Term Energy Outlook, October 2026, Table 2. https://www.eia.gov/outlooks/steo/archives/oct26.pdf
11. CheAI Research. PEM hydrogen well-to-gate carbon footprint. 2026. cheairesearch.com/reports/pem-hydrogen-lca.html
12. Alvarez RA, Zavala-Araiza D, Lyon DR, et al. Assessment of methane emissions from the U.S. oil and gas supply chain. Science 361(6398), 186-188, 2018. DOI 10.1126/science.aar7204

Related studies

• Green hydrogen PEM plant, techno-economic analysis, the cost reference used here
• PEM hydrogen well-to-gate carbon footprint, the emissions reference used here
• Cost-optimal condenser temperature of an ammonia synthesis loop, the synthesis section of the same 2019 design