Hydrogen (H₂)
Hydrogen burns to water. No carbon comes out of the pipe, so the combustion emission factor is genuinely, defensibly zero — and a reporter who stops there books a fuel switch as an instant decarbonisation win. The trouble is that the zero is real and the win may not be.
Hydrogen is an energy carrier, not an energy source. Nothing about it is low-carbon in itself: every emission attributable to a kilogram of hydrogen was released before it reached you, and depending on how it was made that figure ranges across more than a factor of eleven.
Hydrogen (H₂) is an energy carrier that releases no CO₂ when burnt or used in a fuel cell — all of its climate impact sits upstream, in production. Grey hydrogen from steam methane reforming carries 12.516 kg CO₂e per kg well-to-tank; green, from renewable electrolysis, 1.092.
Definition — An Energy Carrier, Not an Energy Source
Hydrogen is the lightest element and, in its molecular form H₂, a combustible gas that releases energy by oxidising to water. Because the molecule contains no carbon, complete oxidation produces no carbon dioxide — only water vapour, plus a small quantity of nitrogen oxides where it is burned in air at high temperature rather than used in a fuel cell.
That single chemical fact drives everything else on this page. Unlike natural gas, diesel or coal, hydrogen has no intrinsic carbon content to release, so it cannot be characterised by a combustion emission factor in the usual way. It is not found in commercially useful quantities in free form either: it must be manufactured, from methane, from coal, or from water using electricity. Hydrogen is therefore a carrier — it moves energy that came from somewhere else, and it inherits the carbon intensity of wherever that was.
Definition at a glance
| What it is | Combustible gas (H₂) used as a fuel and industrial feedstock; an energy carrier, not a primary source |
|---|---|
| Combustion emissions | No CO₂ — factor is 0 kg CO2e per kg |
| Where emissions sit | Entirely upstream, in production — see well-to-tank |
| Range by pathway | 1.092 to 12.516 kg CO₂e per kg — an eleven-fold spread |
| Energy content | 119.953 GJ/tonne net, 141.781 gross — the widest gap of any fuel |
| Under CBAM | A covered good (CN 2804 10 00) with country-specific default values |
The Colour Taxonomy — Where the Emissions Actually Live
Hydrogen is conventionally labelled by the production route that made it. The colours are industry shorthand rather than a formal standard — no regulator certifies “green” on the basis of the word alone — but they map cleanly onto very different carbon intensities, and they are the vocabulary every supplier contract uses.
| Label | Production route | Where the carbon comes from |
|---|---|---|
| Grey | Steam methane reforming (SMR) of natural gas | The methane feedstock itself, released as CO₂ during reforming |
| Brown / black | Coal gasification | The coal feedstock — the most carbon-intensive route |
| Blue | SMR with carbon capture and storage | Whatever the capture rate misses, plus upstream methane leakage |
| Green | Electrolysis of water using renewable electricity | The grid or generation behind the electrolyser |
| Pink / purple | Electrolysis using nuclear electricity | The nuclear lifecycle |
| Turquoise | Methane pyrolysis, yielding solid carbon | The heat input; carbon leaves as a solid rather than CO₂ |
The gap between the ends of that table is the entire point. On the Dutch national factor set — one of the few published sources that gives hydrogen separate grey and green well-to-tank values rather than a single blended figure — grey hydrogen carries 12.516 kg CO2e per kg [GreenCalculus fuels.nld.hydrogen_grey.wtt.kg · NL 2026 · v2026.284] and green carries 1.092 [GreenCalculus fuels.nld.hydrogen_green.wtt.kg · NL 2026]. Both have a combustion factor of zero. Both are “hydrogen” on the invoice.
Grey and green sit at opposite ends and nobody disputes which is which. Blue hydrogen is where the accounting argument happens, because its footprint depends on two numbers a supplier controls and rarely publishes: the capture rate of the CCS plant, and the upstream methane leakage of the gas feeding it. A high capture rate on leaky gas can still produce a poor result, because methane’s global warming potential does the damage long before the reformer does. Treat an unqualified “blue” claim as an input you still need to source, not a figure you can use.
Energy Basis — the Widest Gross–Net Gap of Any Fuel
Hydrogen has the largest divergence between its two calorific values of any common fuel. Its gross calorific value is 141.781 GJ per tonne [GreenCalculus fuel_properties.hydrogen.gross_cv_gj_tonne · NIST CODATA]; its net calorific value is 119.953 [GreenCalculus fuel_properties.hydrogen.net_cv_gj_tonne · NIST CODATA]. The gross figure is 18.2% higher than the net one — against roughly 10% for natural gas and around 5% for diesel.
The reason is the same latent-heat mechanism that separates net from gross for every fuel, amplified because hydrogen’s only combustion product is water. Every atom of the fuel ends up in the vapour whose condensation heat the gross basis counts and the net basis does not, so the gap is as wide as the chemistry permits.
Per kilogram, those figures are 39.38 kWh gross and 33.32 kWh net — dividing the megajoule-per-kilogram values above by 3.6. Fuel-cell specifications and vehicle range figures almost always quote the net number; energy-content and billing data more often quote gross. An 18% error is available to anyone who takes one and pairs it with the other, and because hydrogen is usually metered in kilograms rather than energy units, the basis is rarely stated on the document you are reading.
Quick converters: move a mass of hydrogen onto an energy basis with kg of hydrogen to kWh, and see the energy-unit table for the constants behind it.
Scope Routing — Why a Zero Factor Is Not a Zero Footprint
Hydrogen is the clearest case in GHG accounting of a fuel whose scope placement decides whether its emissions appear at all. The combustion is genuinely zero-carbon. Where the production emissions land depends entirely on who did the producing.
| Situation | Where it goes |
|---|---|
| You burn purchased hydrogen in a boiler or vehicle | Scope 1 — but the CO₂ line is zero |
| The upstream production of that purchased hydrogen | Scope 3 Category 3 (well-to-tank), or Category 1 if bought as a good |
| You produce hydrogen on site by steam methane reforming | Scope 1 — the reformer’s CO₂ is yours |
| You produce it on site by electrolysis, using grid power | Scope 2 — the electricity carries the footprint |
| You produce it on site using a contracted renewable supply | Scope 2 market-based, per your energy attribute certificates |
The practical consequence is that an inventory reporting only Scope 1 and Scope 2 can show a fuel switch to grey hydrogen as a reduction to zero, while the emissions it displaced simply moved into a Scope 3 category that entity may not report. Nothing about that is a rule violation. It is why the well-to-tank line matters more for hydrogen than for any other fuel, and why a hydrogen claim without a stated production pathway carries no information.
Hydrogen is also an indirect greenhouse gas. Leaked H₂ does not warm the atmosphere directly, but it consumes the hydroxyl radicals that would otherwise break down atmospheric methane, extending methane’s lifetime, and it contributes to stratospheric water vapour and tropospheric ozone. This effect is established in the literature and actively under review, but it is not yet carried in DEFRA, GHG Protocol or IPCC corporate reporting factor sets — so it appears in no number on this page and in no hydrogen inventory line you are likely to file. Treat leakage as a real physical concern that current reporting conventions do not yet price.
Hydrogen Under CBAM
Hydrogen is one of the goods covered by the EU Carbon Border Adjustment Mechanism, under commodity code CN 2804 10 00. Importers into the EU must report embedded emissions, and where actual verified data is unavailable the published country default applies. The compressed-hydrogen default is 26.64 kg CO2e per kg [GreenCalculus materials.cbam.hydrogen.compressed · CBAM 2026] of good.
Those defaults are deliberately conservative — they are set high enough that using actual data is normally to the importer’s advantage. For hydrogen specifically the defaults sit near the grey end of the range, which means an importer of genuinely green hydrogen who fails to supply verified production data will be charged as though it were made by reforming methane. See the CBAM methodology for how embedded emissions are calculated, and the UK CBAM calculator for the parallel UK regime.
Production Routes and Their Feedstock Carbon
Beneath the colour labels sit measurable feedstock carbon contents. The IPCC 2019 Refinement publishes these per unit of feedstock energy, which is the form national inventory compilers use and the most direct way to compare routes without assuming a plant’s efficiency:
| Production route | Feedstock carbon | Unit |
|---|---|---|
| Steam reforming — natural gas | 0.0153 | tonne C per GJ feedstock |
| Steam reforming — LPG | 0.0172 | tonne C per GJ feedstock |
| Steam reforming — naphtha | 0.02 | tonne C per GJ feedstock |
| Gasification — coal / coking coal | 0.0258 | tonne C per GJ feedstock |
| Gasification — wood waste | 0.0305 | tonne C per GJ feedstock |
| General default | 0.0183 | tonne C per GJ feedstock |
Coal gasification carries roughly 1.7 times the feedstock carbon of natural-gas reforming per unit of energy in, which is why the brown route sits below grey in every ranking. Note that the wood-waste figure is not a licence to treat that route as carbon-neutral: whether those emissions are reported as biogenic depends on the sourcing rules that apply to the feedstock, not on the gasifier.
A Note on Factor Availability
There is no UK DEFRA hydrogen emission factor equivalent to the ones published for natural gas, diesel or LPG. This is not an oversight in our data: because combustion emits no CO₂, a conventional combustion factor for hydrogen would simply be zero, and the number a UK reporter actually needs — the well-to-tank intensity of the specific hydrogen they bought — depends on a production pathway DEFRA cannot know.
The practical consequence is that hydrogen is one of the few fuels where a supplier-specific figure is not merely preferable but necessary. The national factor sets that do publish split values, such as the Dutch set used above, are useful for scale and comparison. They are not a substitute for asking your supplier which colour they sold you, and on what evidence.
Common Confusions
- Reading the zero combustion factor as a zero footprint. It is a correct number describing one stage. Hydrogen’s emissions are real and sit upstream — report the well-to-tank line or you have not reported the fuel.
- Treating “hydrogen” as one fuel with one factor. Grey and green differ by more than a factor of eleven on the same Dutch factor set. A supplier invoice that says only “hydrogen” is missing the number that matters.
- Accepting “blue” as a figure. Blue hydrogen’s footprint is a function of the CCS capture rate and the upstream methane leakage of the feed gas. Without both, the label is a description of equipment, not a carbon intensity.
- Mixing the gross and net energy bases. Hydrogen’s two calorific values differ by 18.2% — the widest gap of any fuel. Fuel-cell data is usually net; energy-content data is often gross.
- Assuming the CBAM default is a penalty-free fallback. The defaults sit near the grey end, so importing genuinely green hydrogen without verified production data is charged as though it were reformed from methane.
- Calling electrolytic hydrogen green because the electrolyser is new. The colour describes the electricity, not the equipment. An electrolyser on an unabated grid produces hydrogen whose footprint is that grid’s carbon intensity, divided by efficiency.
Related Terms, Standards, and Tools
Frequently Asked Questions
No. Hydrogen contains no carbon, so complete combustion produces only water vapour — the combustion emission factor is 0 kg CO2e per kg [GreenCalculus mobile_combustion.nld.hydrogen_grey.per_kg · NL 2026]. Burning it in air at high temperature does produce nitrogen oxides, which are an air-quality pollutant rather than a CO₂e line, and fuel cells avoid even that. But a zero combustion factor is not a zero footprint: every emission attributable to hydrogen was released upstream, when it was made.
There is no single one, and that is the important answer. Combustion is zero for all hydrogen. The well-to-tank figure depends entirely on production: on the Dutch national factor set, grey hydrogen from steam methane reforming is 12.516 kg CO2e per kg [GreenCalculus fuels.nld.hydrogen_grey.wtt.kg] while green hydrogen from renewable electrolysis is 1.092 [GreenCalculus fuels.nld.hydrogen_green.wtt.kg]. Any hydrogen figure quoted without a production pathway is incomplete.
All three, depending on what you did. Burning purchased hydrogen is Scope 1, with a zero CO₂ line. The production of that purchased hydrogen is Scope 3 — Category 3 as a well-to-tank fuel item, or Category 1 if you bought it as a good. Making it yourself by reforming puts the reformer’s CO₂ in your Scope 1; making it by electrolysis puts the electricity in your Scope 2.
Because a conventional combustion factor would be zero and would tell a reporter nothing useful. The figure that matters is the well-to-tank intensity of the specific hydrogen purchased, which depends on a production pathway a national factor publisher cannot know. Hydrogen is therefore one of the few fuels where a supplier-specific figure is not just preferable but necessary — national sets that publish split grey and green values are useful for scale, not as a substitute for asking.
119.953 GJ per tonne net [GreenCalculus fuel_properties.hydrogen.net_cv_gj_tonne] and 141.781 gross [GreenCalculus fuel_properties.hydrogen.gross_cv_gj_tonne] — which per kilogram is 33.32 kWh net and 39.38 kWh gross. That 18.2% spread is the widest of any fuel, because hydrogen’s only combustion product is the water whose condensation heat separates the two bases. Fuel-cell and vehicle-range figures use the net number; energy-content data often uses gross. Use the converter and state which basis you took.
It depends on two numbers the label does not carry: the capture rate of the CCS plant, and the upstream methane leakage of the gas feeding the reformer. A high capture rate on leaky gas can still give a poor result, because methane’s global warming potential means leakage does its damage before the reformer is involved. Treat an unqualified blue claim as an input still to be sourced rather than a carbon intensity you can use.
Need someone who does this? 6 carbon accounting & inventory providers in our directory →