# sustainable aviation fuel

Topic radar · 3,759 documents · 2021-01-01–2026-10-16 · updated 2026-10-16

Canonical: https://slate.greyb.com/radar/sustainable-aviation-fuel/innovations

## In one minute

### Every route ends in the same upgrading step
The routes differ in how they build an intermediate. Fats and oils are deoxygenated and isomerised, wet waste is liquefied into a biocrude, plastic and tire oils are hydrocracked, biomass is gasified and rebuilt over a Fischer-Tropsch catalyst, sugars are joined by aldol condensation, and carbon dioxide is reduced to carbon monoxide or hydrogenated. They converge on a shared finishing step that trims the product to kerosene range. A palm-oil-mill-effluent process cleans its feedstock by adsorption, then hydrotreats and isomerises it [1]. Algal biocrude from continuous hydrothermal liquefaction has been hydrotreated to a jet cut at about 22.7% by weight. Corn-stover fast pyrolysis reaches a competitive selling price but overshoots the specification's final-boiling-point limit, so it blends at up to 10%. The shared back end is why the feedstocks compete for the same upgrading capacity.

1. [WO2025134933A1](/?paper=WO2025134933A1&utm_source=radar&utm_medium=dashboard)

### The premium is real, and the oil price moves it more than SAF does
SAF averaged about 2,872 a tonne in 2026, roughly 2.5 times conventional jet, adding an estimated 4.3 billion to airline fuel costs. The multiple is not stable. Argus reported the narrowest gap between SAF and jet since it began tracking prices, with SAF oversupply meeting fossil jet prices lifted by the Middle East conflict, and in March 2026 western European SAF traded near $2,500 a tonne with its premium falling for the same reason.


### Certification is where the commercial effort concentrates
The clearest commercial move in the field is a standards approval, and the licensing clusters around it. Methanol-to-jet cleared ASTM in July 2026. Honeywell UOP's eFining methanol-to-jet process was licensed to Verso Energy across seven planned sites in France, Finland and the United States, Topsoe markets its own methanol-to-jet route, and Axens completed a methanol-to-jet offer by buying Air Liquide's methanol-to-olefins portfolio. The effort is going into certified pathways and the licences around them, not into new chemistry.


### The certified routes all make the same narrow molecule
The certified pathways turn out iso- and n-alkanes and almost nothing else. Jet fuel also needs cyclic molecules, naphthenes and aromatics, to swell fuel-system seals, and the paraffinic routes make too few of them, which is why synthetic blendstock is capped at 50% in conventional jet fuel. Three lines of work are trying to make the missing classes. Hydrodeoxygenation of lignin oil yields jet-range naphthenes and aromatics. Ethylbenzene made from waste polystyrene restores seal swell and could lift the blend ceiling. A distilled-tall-oil route has produced all four hydrocarbon classes with aromatics above 80% and passed seal-swell testing. So the routes compete on cost and certification, and the product itself is not where they differ.


### Public money, not the passenger, carries much of the premium in Europe
EU ETS support for 2025 SAF uptake came to about 5.2 million allowances worth roughly €430 million, spread across 130 operators, on top of a €135 million incentive from zero-rating SAF on ETS routes: about €565 million against just over 530,000 tonnes of fuel claimed. The support is retrospective, paid for fuel already claimed, and it subsidises uptake rather than capacity. It is also the clearest case of the state acting as the buyer of the premium while the mandate is still being met from existing waste-lipid supply.


### Value is settling with the process owners, not the plant owners
Licensing revenue arrives at award, whether or not the plant ever runs at nameplate, while the developer carries the demand risk. KBR licensed its PureSAF technology for a plant on Jurong Island, for Kazakhstan's first SAF project, and for what is billed as Northern Europe's largest SAF and e-SAF plant, in Latvia. Axens licensed its HEFA technology for facilities in Africa and the Caribbean. Topsoe and Sasol returned to licensing after closing their production venture, signing a technology and engineering agreement for a $6.1 billion project in Uzbekistan. The scarce asset is a certified pathway, and the companies that own one are selling access rather than building.


### Methanol joins the biomass and power routes
Methanol can be made from biomass-derived syngas or from captured carbon dioxide and hydrogen, so one certified pathway now draws on both the collected-matter and the electricity economies. ASTM added methanol to the alcohol-to-jet annex of its synthetic-fuel specification in July 2026.


### A corporate SAF claim rests on a registry rulebook, not on fuel in the tank
Book-and-claim decouples the environmental attribute from the molecule: a SAF certificate equals one tonne of neat certified SAF and carries the right to claim scope 1 and scope 3 reductions once retired in a registry. The channel is growing — IATA, Airbus and Accenture put corporate purchases at 1.0 to 1.9 million tonnes by 2030, from about 323,000 tonnes in 2024.


## Overview


The chemistry of making sustainable aviation fuel is settled. Every route in the field, whether it starts from fats and oils, wet waste, plastic, biomass, sugars or captured carbon dioxide, converges on a shared finishing step and turns out the same narrow set of paraffinic molecules. That uniformity is why the certified pathways are capped at a 50% blend in conventional jet fuel: they make too few of the cyclic molecules the aircraft's fuel-system seals need. Three research lines are trying to make the missing classes, and none is certified. The routes therefore compete on cost and on certification, and the product itself is the same across them.


## What's new


The only capacity here with money and a date is a modification to something already running. Topsoe will license kerosene hydrotreating so BPCL's 4,000 metric ton per day Mumbai hydrotreater can take 5% used cooking oil alongside conventional feed, decision taken, commissioning targeted for end-2026. Phillips 66's Wood River is the same pattern in a hydrocracker, conventional jet blended into SAF and counted through a United, Microsoft and State of Illinois chain. Everything larger is a study, a target or a bench: Pemex's 5,000 barrel per day Pajaritos feasibility reports in early 2027, GeorgiaSAF's 20,000 barrels a day is described by the company as a development target, and the on-spec Fischer-Tropsch result is 2.5 barrels a day.


## Research

### A standards approval is what makes a chemistry count
ASTM D7566 governs whether a synthetic fuel can be blended into jet fuel, and it is the real gate to market. Co-processing is governed separately and the limits differ by jurisdiction. The UK raised its co-processing cap to 30% in May 2025, while the international D1655 route still caps renewable esters and fatty acids at 5%. A chemistry matters commercially once a standards body accepts it.


### The engine sees much the same fuel
SAF behaves much like Jet A in bulk physical terms and differs from it chemically. Cavitation behaviour and near-nozzle spray barely moved across diesel, Jet A, two SAF samples and dodecane, with the most viscous test fluid the exception. In a three-sector combustor with commercial airblast injectors, SAF droplets came out slightly smaller and soot volume fraction fell two to three times. What changes is molecular: how little aromatic and sulfur content the fuel carries, how its light and heavy ends separate during evaporation, and how that shifts flame position and dynamic response.


### Soot falls and nitrogen oxides do not move
The emissions pattern follows that split. ECLIF3 found no measurable fuel effect on the nitrogen oxide emission index, and the VOLCAN flights found no significant difference across Jet A-1, SAF and two blends. Soot responds to composition. Nitrogen oxides respond to temperature and residence time, and swapping the fuel does not change those enough to register.


### The soot cut is largest where the engine is least efficient
At ground idle the combustor is cool and poorly mixed, so fuel chemistry dominates. At high power it is hot enough that the fuel matters less. A commercial turbofan cut non-volatile particulate mass by about 70% and number by 60% at idle on a 32% blend, with the relative reduction falling as thrust rose. A turboshaft cut particle number by up to 81% on 100% SAF and shifted the size mode smaller. An in-flight A350 cut particle number by about 41% at low cruise, 29% at typical cruise and 22% at high cruise. A business jet on a 30% blend cut mass by about 35% and number by about 20% at idle, again diminishing with thrust.


### Contrail ice stops following soot once soot is scarce
The early result was proportional: less aromatic fuel, less soot, fewer ice crystals. The newer flight data says proportionality holds only while soot is abundant. In ECLIF3, ice particle number fell 56% while soot fell 35%. On lean-burn LEAP-1A engines, soot dropped by orders of magnitude and ice numbers stayed high, tracking total aerosol. A 2026 measurement behind a lean-burn A321neo found ice crystal numbers exceeding soot particle emissions by three orders of magnitude. The microphysics accounts for it. Ice activation is governed by the size of the primary soot particles, and smaller primaries delay ice onset and let volatile particles contribute. Real aviation soot does not become a better ice nucleus after contrail processing, and in the soot-poor regime the apparent ice emission index can exceed the soot emission index by up to two orders of magnitude. Sulfur and lubrication-oil species and entrained ambient particles take over the ice-forming burden.


### The hardware answers are nozzle staging, fuel heating and composition sensing
The recurring invention splits a combustor's nozzles into two subsets, feeds one at a higher rate, and pairs the arrangement with a measured emissions ratio comparing SAF against fossil kerosene. Non-volatile particulate matter is the currency those claims are written in. A second family treats SAF's lower viscosity as the problem and heats the fuel to at least 135°C at cruise to bring viscosity at combustor entry to 0.58 mm²/s or below. A third and larger family reads composition from the tank and adapts engine parameters to it. The heat that fixes SAF viscosity is bounded by the wear scar diameter the pumps and valves need, so the atomization gain available from heat exchange has a ceiling set by hardware designed for kerosene.


### The combustion modelling is being claimed in Chinese filings
A method that builds a validated combustion mechanism in hours, or predicts combustion characteristics from molecular descriptors without experiments, is the screening layer between a producer's candidate molecule and an expensive test campaign. The Western primes appear in this area as paper co-authors and campaign partners, and their patent filings are absent. That is a filing pattern, and a company can hold combustion capability without patenting it.


### Feedstock engineering sits at the front of the chain
The conversion step is largely shared, so the differences sit in how the feedstock is grown, recovered or fermented. Winter oilseeds grow in a rotation window that already exists, and life-cycle work puts hydroprocessed jet from camelina, carinata and pennycress at 50 to 66 percent below petroleum jet. POME's value comes from being counted as a residue, and the recovery step is a purification problem. Algal fuel has been demonstrated end to end, including the first reported jet fuel from autotrophically grown marine algal biomass, but the cost models keep it far above the market without policy. Precision fermentation is where the technical progress is fastest, with engineered microbes making defined precursors and machine learning lifting titres several-fold.


### The co-product carries the plant
For winter oilseeds, the protein meal can carry much of the economics, and a national model found meal credits could cut costs by up to 70 percent. For POME, the co-product is the methane in the effluent, and whether capturing it for mill energy and claiming avoided methane for the fuel are additive is unresolved. For algae, the co-product is wastewater treatment, which changes the customer from a refiner to a utility. In all three, the fuel makes the marginal economics work and the co-product makes the plant bankable.


### The cheapest capacity reuses existing plants
Cost is set before the conversion step. Repurposing a petroleum refinery cuts baseline capital cost by 36%, and repurposing pulp and paper facilities reaches $1.16/L, about 19% below a matching greenfield plant. Where a study reports below $2/L, a favourable feedstock and a co-product credit are doing the work. Carbon is set by feedstock and land, and the conversion step is a minority of the total: renewable hydrogen for upgrading cuts emissions by 92 to 96.6%.


### The accounting layer is being built as infrastructure
The filings in the assessment part of the field claim the method: how to optimise a low-carbon supply chain, how to build a distributed carbon footprint database, how to attribute the spread in published life-cycle results, and how to couple SAF's environmental effect with tourism. A certification workflow publishes an electronic certificate with its number, scope and validity period, and a tokenisation platform trades low-carbon fuel batches and contracts. All the assessment filings come from Chinese civil aviation institutes and universities. The groups publishing the most cost and carbon accounting file nothing on how the accounting is done.


### Three lines of work sit in filings but not in public coverage
Wet-waste liquefaction has dense academic work and a thin commercial layer. One company holds a real patent family on sewage-sludge liquefaction with upgrading to jet fuel, and the rest are single filings. Bioethylene valorisation is in the filings and absent from the market record: Chevron Phillips describes oligomerising bioethylene and upgrading the decene byproduct to a C16 paraffin, while the public ethanol-to-jet coverage leaves the olefin and decene steps out. The reverse water-gas shift and co-electrolysis chemistry is essentially unclaimed in patents; the filings that touch it claim electrolyser operation and power accounting, and leave the catalyst and the cell unclaimed.


## Market

### Regulators now treat SAF marketing as a claim to substantiate
A Danish court fined KLM 3 million kroner, about €401,000, in July 2026 over a 2023 campaign that said it was using "a larger share of sustainable aviation fuel for all tickets" when SAF was about 1% of its fuel. The UK advertising regulator ruled Virgin Atlantic's unqualified "100% sustainable aviation fuel" claim about Flight100 misleading, because the flight was not commercial and SAF's limitations were material to the audience. Eurowings was ruled against over a checkout payment presented as offsetting a specific flight's emissions.


### The 100% SAF claims run ahead of what is certified
Rolls-Royce completed a ground and flight test programme across its in-production civil engines on 100% SAF in 2023 and said the tests confirmed no effect on engine performance. Boeing has committed that its commercial airplanes will be 100% SAF compatible by 2030.


### Suppliers are marketing the fuel system as the SAF-compatible part
Seal and elastomer makers are positioning on SAF compatibility. One markets seals for higher-percentage and eventual 100% operation, and Parker's VX065 fluorocarbon is sold on SAF compatibility and marketed as the only compound on the −65°F fluorocarbon qualified products list. The claims rest on bench and field experience rather than a service record on neat SAF, because no aircraft flies on it commercially. That is the same gap as the engine claims, one layer down.


## Strategy

### The committed capital is going into existing assets, not new synthetic plants
Where money is being committed, it is incremental and asset-based. Montana Renewables cut the remaining capital for its expansion to 137 million from the 1.2 billion of the original plan, by repurposing a hydrotreater, hydrogen plant and naphtha splitter from the adjacent refinery. Where a project needs a greenfield plant and a new carbon-dioxide supply, it stalls: Gevo abandoned its AtJ-60 plant after failing to secure CO2 transport and storage. No European e-SAF plant above 10,000 tonnes a year is operating or past final investment decision. The announcements are there; the commitments are not.


### The near-term effort is in the waste economy
Almost all announced and financed capacity is lipid-based, and the feedstock is a globally traded waste stream, so the announced plants compete for the same barrels. As used cooking oil is bid up, the routes that use no lipids gain a relative feedstock advantage they did not have when lipids were cheap, but they carry higher capital cost and depend on cheap hydrogen or renewable electricity. The risk runs the other way too: Shell abandoned its Rotterdam biofuels plant, 820,000 tonnes a year, on competitiveness grounds.


### The mandate is being held at the headline and deferred at the sub-target
The European Commission reaffirmed in March 2026 that the ReFuelEU mandates would not be weakened, but the 2030 e-SAF sub-mandate stays on the books without enforcement. That is where the effort goes: bio-SAF now, synthetic later. The pathways the later targets depend on lose the one thing that would make them investable today, a compliance deadline that bites.


### Feedstock rules are redirecting effort geographically
The 45Z clean fuel production credit now requires that fuel produced after 2025 be derived exclusively from feedstocks grown or produced in the United States, Mexico or Canada. That excludes palm-oil-mill effluent, which comes from Indonesia and Malaysia, and it favours the domestic winter oilseeds. The effort follows the credit: Bunge certified the first soybeans under a low-land-use-change protocol, Petrobras sold the first CORSIA-certified soy-based SAF, Nufarm and bp expanded their carinata partnership, and the three winter oilseed crops now have their own trade association. The measurement layer, soil carbon and traceability, becomes the decisive capability.


### Offtake is moving to non-airline buyers
The counterparties are increasingly logistics groups, distributors and corporate buyers rather than airlines alone. DHL Express signed a ten-year offtake for 25,000 tonnes a year from SAF One's Bahrain plant, a Madrid deal with Moeve, and deepened a multi-year agreement with IAG Cargo. Cargo and book-and-claim turn SAF into a premium logistics product, a demand channel that does not depend on the mandate the way an airline's purchase does.


### The mitigation effort is moving off the fuel molecule
ICAO names contrail avoidance as the most feasible short-term lever for aviation's non-CO2 impact, citing a 73% reduction in warming for a 0.1% fleetwide fuel-burn increase. As soot-poor engines weaken the composition lever, the ice-forming burden shifts to sulfur and lubrication-oil species and to particles entrained from ambient air, which fuel composition does not reach. The competitive ground moves from the combustor to the fuel system: pumps, seals, valves and the property sensing that has to work.


## Areas

- Conversion Technologies · 1,345 documents
- Policy & Commercial Deployment · 1,091 documents
- Combustion & Emissions · 478 documents
- Feedstock Supply · 351 documents
- Cross-Cutting Assessment · 449 documents