Can you turn crude oil into sugar?

The question: can you take the long carbon chains in crude oil (say hexane, a C₆ chain) and convert them into something sugar-like (a C₆ ring like glucose)? Does it need a biological catalyst? Short answer: yes it's possible, life already does it, and the hard part is not the energy and not the ring — it's selective partial oxidation.

The reframe: sugar is half-burnt carbon

Rank carbon by oxidation state, from most-reduced to most-oxidised:

Sugar isn't above hydrocarbons in chemical energy — it sits below them, partway down the hill toward CO₂. Every carbon in glucose already carries an oxygen. That's why glucose yields far less energy per kg when burned: it's already half-oxidised.

So the conversion is an oxidation, and you don't have to push energy in to make it go.

The net reaction

On paper it balances cleanly as a partial oxidation, and it's exothermic:

C₆H₁₄  +  3.5 O₂   →   C₆H₁₂O₆  +  H₂O
hexane     oxygen        glucose     water

ΔH ≈ −1360 kJ/mol   (downhill)

Atom check: 6 C = 6 C; 14 H = 12 + 2 H; 7 O = 6 + 1 O. The ΔH comes from Hess's law on the two combustion enthalpies (hexane −4163, glucose −2805 kJ/mol).

hexane + O₂ → (some) glucose + (a lot of) CO₂ + H₂O + ATP + cell mass

The enzymatic pathway, step by step

The key insight: the cell does not reshape a hexane into a glucose ring. It dismantles the alkane into 2-carbon Lego bricks and builds a brand-new sugar from scratch.

1. C₆H₁₄          + O₂ + NADH  →  1-hexanol        [AlkB monooxygenase]
2. 1-hexanol      + NAD⁺       →  hexanal          [alcohol dehydrogenase]
3. hexanal        + NAD⁺ + H₂O →  hexanoic acid    [aldehyde dehydrogenase]
4. hexanoic acid  + ATP + CoA   →  hexanoyl-CoA     [acyl-CoA synthetase]
5. hexanoyl-CoA   — β-oxidation →  3 × acetyl-CoA
6. acetyl-CoA  — glyoxylate shunt →  C₄ acids → oxaloacetate → PEP
7. PEP  — gluconeogenesis →  fructose-6-P → glucose-6-P → glucose

Step 1 is the crux

The membrane enzyme alkane 1-monooxygenase (AlkB) inserts one oxygen atom from O₂ into the otherwise-inert terminal C–H bond; the other O becomes water, and the electrons come from NADH via a rubredoxin shuttle (AlkG):

R–CH₃  +  O₂  +  NADH  +  H⁺   →   R–CH₂OH  +  H₂O  +  NAD⁺

Mechanistically it's a diiron active site (coordinated by histidines) that activates the C–H bond, sitting at the end of a hydrophobic tunnel lined with leucines that grips the alkane and presents only the terminal carbon to the iron. That geometry is where the regioselectivity comes from — and it's exactly what brute-force chemistry can't reproduce, because uncontrolled oxidation runs away to CO₂. The enzyme does it at body temperature.

Why the ring isn't the hard part

You don't have to engineer the loop. An open-chain sugar (a polyhydroxy aldehyde) cyclises spontaneously in water: the –OH on C5 attacks the C1 aldehyde to form a hemiacetal ring, sitting in equilibrium between open and closed forms all on its own. Make the right linear molecule and the “loop that resembles sugar” forms for free.

Why step 6 (the glyoxylate shunt) is non-negotiable

You can't make net sugar from acetyl-CoA through the normal TCA cycle — it burns both carbons to CO₂. The glyoxylate bypass (isocitrate lyase + malate synthase) conserves the carbons into C₄ acids, which feed gluconeogenesis. The lovely confirmation: when Alcanivorax borkumensis grows on alkanes, proteomics shows exactly the glyoxylate-bypass and gluconeogenesis enzymes get upregulated — the cell visibly reconfigures to run carbon uphill back to sugar, and even secretes a glucose-lipid biosurfactant built from oil.

Does it have to be hexane?

No — and the reason is the most elegant part. Because step 5 chops everything to identical 2-carbon units, the original skeleton is destroyed; whatever you start with becomes the same acetyl-CoA pool. The six carbons in the final glucose are not “the same six” that were in your hexane. So chain length barely affects whether you get sugar — it only changes which enzyme does step 1:

What microorganisms you'd need

For hexane (C₆) specifically, the canonical workhorse is:

• Pseudomonas putida GPo1 (formerly P. oleovorans) — its AlkBGT pathway is literally the textbook route for C₆–C₁₂ n-alkanes. Default choice, and conveniently solvent-tolerant.

Supporting cast / alternatives:

• Pseudomonas aeruginosa strains (SJTD-1, NY3) — robust, multiple AlkB copies.
• Rhodococcus and Acinetobacter spp. — hardy, industrially friendly alkane oxidisers.
• Candida tropicalis / Yarrowia lipolytica — the yeasts from the BP single-cell-protein process; prefer longer C₁₀–C₁₈ paraffins but are the proven industrial organisms.
• Alcanivorax borkumensis — the marine specialist where the oil→sugar route is best documented; great model, fussy to culture.

A rough lab procedure

1. Strain: Pseudomonas putida GPo1 (or Rhodococcus sp. for robustness).
2. Medium: minimal salts (M9-type): N (NH₄⁺), P, S, Mg, trace metals. No sugar, no other carbon — hexane as the sole carbon source forces the cells to run the whole pathway.
3. Carbon feed: hexane as a second liquid phase, or via the vapour phase (a headspace reservoir), to dodge toxicity and evaporation.
4. Aeration: vigorous. Step 1 consumes O₂ directly, and the cell needs more O₂ to burn the rest for ATP. Strongly aerobic.
5. Grow at ~30 °C until biomass accumulates.
6. Harvest the sugar: the catch — the glucose is inside the cells (glycogen, cell-wall polysaccharide, metabolic pools), not secreted. To get free glucose you lyse the cells and acid- or enzyme-hydrolyse the polysaccharides, then purify. (The BP process skipped this and just dried the whole cells as protein-rich meal.)

If you genuinely wanted a secreted free-glucose stream, that's a synthetic-biology project: engineer a strain (knock out sugar consumption, add a glucose exporter, push gluconeogenic flux). Not an off-the-shelf bug.

This was real, at industrial scale

In the 1960s–70s, BP ran a “proteins-from-oil” process: Candida yeast grown on refinery n-paraffins with ammonia and salts, in big stirred fermenters. Started by Alfred Champagnat at the Lavéra refinery (pilot plant 1963), scaled to plants in several countries — it's the origin of the term single-cell protein. The biomass included all the carbohydrate the cells synthesised from oil. It was killed by economics (oil shocks) and consumer wariness, not by the chemistry failing.

The no-enzyme route: formose

Want sugar without biology? Then you don't go through C₆ chains at all — you go through C₁:

hydrocarbon → (oxidise) → CH₃OH → HCHO (formaldehyde) → [formose] → sugars

The formose reaction (Butlerov, 1861) is autocatalytic: under alkaline conditions formaldehyde slowly forms glycolaldehyde, which then “fixes” two more formaldehydes via aldol additions and keto-enol tautomerisations, fragments back by retro-aldol, and doubles — a self-amplifying loop climbing to C₃, C₄, C₅, C₆ sugars. Striking detail: ~3 ppm of glycolaldehyde is enough to ignite the autocatalysis.

The catch is it's a notoriously messy, racemic soup — dozens of sugars, no selectivity for glucose, with Cannizzaro side-reactions degrading the products. It's a leading hypothesis for prebiotic sugar formation, which is the whole lesson: sugars assemble spontaneously from one-carbon units, but only enzymes give you one clean, defined product with defined stereochemistry.

Bottom line

• Yes, it works — alkane → sugar is a real, documented metabolic capability, and ran at industrial scale.
• It's not an energy problem; it's downhill. It's a selectivity problem, solved by monooxygenase enzymes.
• The ring is free (spontaneous cyclisation).
• You won't get a clean glucose stream by just feeding hexane to a microbe — you get cells that contain sugar, and you extract it.
• Hexane is a poor feedstock (volatile, toxic); longer paraffins are easier. Chain length barely matters to the output because everything routes through acetyl-CoA.

Cell production — protein & food (the easy path)

Worth pausing on a confusion that runs through this whole topic: feeding oil to a microbe doesn't hand you a molecule, it hands you a cell — and a cell is mostly protein, with only a little sugar locked up inside it. That splits the goal into two genuinely different jobs: cell production (this section — protein and food, the path that actually worked industrially) and sugar production (next section — much harder, never done at scale). They share an identical first half and diverge entirely in the second.

The upstream is the same for both. Same inputs (reduced carbon + oxygen + nitrogen + minerals), same fermenter, same result: a tank of wet cells. What differs is which part of the cell you care about and how much you process it afterwards.

What the microbe actually makes

A harvested cell, by dry weight, is roughly:

Protein dominates — that's just what cells are (enzymes, ribosomes, membranes). The “sugar” is the thin carbohydrate slice: storage glycogen plus the structural glucans/mannans of the cell wall. There is no pool of free glucose sitting in the broth waiting to be skimmed off.

One fermentation, three destinations

Protein is the easy one — and it's why every real industrial project (BP, ICI, the USSR) made protein, never sugar. To “make protein” you barely process at all: separate the cells, wash, dry. The dried biomass is ~60% protein. That powder is “single-cell protein.”

Food is protein-plus-finishing. For animal feed, the dried meal is essentially done. For human food you must do more, because cells carry ~10% nucleic acid (RNA) — eat too much and the uric acid gives you gout. So you add a heat step that destroys RNA (sacrificing ~30% of the biomass), then texturise the result into something fibre-like. That extra processing is exactly the line between Pruteen/BVK (animal feed, never RNA-reduced) and Quorn (human food, RNA-reduced and texturised).

Efficiency & inputs — the oxygen problem

“Efficiency” means three different things here, and SCP scores very differently on each:

• Mass yield — about 1 kg of cells per kg of n-paraffin (alkanes are reduced and carbon-rich), but only ~0.4–0.5 kg per kg of methanol or sugar.
• Carbon conversion — only about half (~50–60%) of the feed carbon ends up in cells; the rest is burned to CO₂ to power the build.
• Vs. livestock — the selling point. Microbes double in hours (harvested daily or weekly, vs. an annual crop or a multi-year steer) and put ~60% of their dry mass into protein. On protein per unit feed per unit time they beat animals by orders of magnitude.

But the process efficiency is dragged down by everything that isn't the oil. Roughly, per tonne of dry SCP:

Oxygen is the dominant non-carbon input — ~1.5–2.5 t per tonne of product, second only to the feed itself. And it carries a twin cost: you must compress and sparge that much air through a broth the oily substrate barely dissolves in, and you must remove the heat it generates — about 14 MJ for every kg of O₂ consumed (~30 GJ per tonne of SCP on the alkane route). The fermenter is as much heat-exchanger as bioreactor. Note the trade-off in the table: the alkane route needs only half the carbon feed by mass, but ~40% more oxygen and cooling than methanol — one reason ICI chose clean, water-soluble methanol.

This is the deeper reason the economics were so fragile: the real running cost was oxygen + electricity + cooling + drying, not the petroleum. The process had little slack — so when the 1973 oil shock lifted feedstock prices, there was nothing left to absorb it.

Figures per tonne of dry cells, anchored to an integrated methanol-SCP design (Phillips Petroleum, US Patent 4,145,445: ~1,000 t/day methanol · 790 t O₂ · 100 t NH₃ → 500 t SCP + 482 t CO₂) plus n-hexadecane oxygen-demand data (~2.2 kg O₂/kg cells); heat from the oxycaloric value ~14.4 MJ/kg O₂. The alkane-route CO₂ (~1–1.4 t) and ammonia are approximate, and that patent describes a larger plant than ICI's actual ~50,000–75,000 t/yr Billingham unit.

Sugar production — the hard path

Now the other job. Getting free sugar out is a different problem — and it's why, unlike protein, nobody ever ran oil→sugar industrially. There are only two ways to attempt it, and both are awkward:

The cellular route is the hard one — and nobody did it at scale. The glucose is bound up as glycogen and cell-wall polymer inside the cells, so you'd have to break the cells open, hydrolyse those polymers back to glucose, then purify it out of a slurry of protein and debris. Worse, it's backwards from the microbe's point of view: the cell wants to spend sugar building itself, not hoard free glucose for you. A clean secreted-glucose stream would mean engineering a strain against its own metabolism — a synthetic-biology project, not an off-the-shelf bug.

The abiotic route skips cells entirely. Oxidise the hydrocarbon down to one-carbon units (methanol → formaldehyde) and run the formose reaction from Part 1. It needs no enzymes, but it's unselective: you get a racemic mess of many sugars, not clean D-glucose. So neither path gives you the tidy “sugar out of oil” you might picture — which is exactly why the world built protein factories, not sugar ones.

Part 2 — The Oil-to-Protein Story

The chemistry above isn't hypothetical: from the late 1950s to the 1980s, three industrial programs grew microbes on petroleum to make protein-rich animal feed — single-cell protein (SCP). They were a technical success and a commercial failure, and the story is a neat case study in how a real, working bioprocess dies anyway. (This section was cross-checked across a deep web-research pass and two independent AI fact-checkers, Codex and Gemini; corrections from that triangulation are baked in, and the genuinely contested points are flagged below.)

The three programs

“Oil-to-protein” was really three parallel efforts with importantly different chemistry — and one of them (Pruteen) isn't a hydrocarbon product at all:

A distinction worth keeping straight: ICI's Pruteen is the odd one out — methanol-fed bacteria, not petroleum yeast. The true “oil-to-protein” line (growing on petroleum paraffins) is BP's Toprina and the Soviet BVK.

BP — “Proteins-from-Oil” (Toprina)

Initiated by Alfred Champagnat at BP's Lavéra refinery in France (research from the late 1950s); pilot plant at Lavéra in March 1963, second pilot at Grangemouth, Scotland. Champagnat's “Protein from Petroleum” (Scientific American, 1965) was the field's founding manifesto, and he won the 1976 UNESCO Science Prize. Two product variants:

• Toprina L — C. tropicalis on gas-oil at Lavéra (the “dirty” route: yeast eats the paraffins out of gas-oil, leaving residual oil to separate).
• Toprina G — Y. lipolytica (= C. lipolytica) on purified n-paraffins at Grangemouth (the “clean” route; the smaller plant at ~4,000 t/yr vs Lavéra's ~16–17k).

That purified-vs-gas-oil split is the whole engineering trade-off in miniature: purifying the feed costs money but avoids a nasty downstream separation and reduces carcinogen carry-over.

ICI — Pruteen, the engineering monument

ICI grew the bacterium Methylophilus methylotrophus on methanol (~2 t methanol → 1 t protein). The Billingham plant: 50,000 t/yr, sanctioned 1976 (~£40M initial; total ICI spend eventually >£100M), contractor John Brown, built around a single airlift fermenter ~60 m tall, claimed to be the world's largest (working volume ~1,500 m³; the ~600-tonne steel vessel was fabricated in Dunkirk and shipped to Billingham in one piece). It was a technical triumph and a commercial flop, discontinued in the late 1980s. ICI reused the fermentation know-how (not the organism) for Quorn — which is the fungus Fusarium venenatum, the one lasting success of the SCP era.

The Soviet BVK giant

By far the largest program (product BVK, belkovo-vitaminny kontsentrat). Principal organism a high-yield Candida guilliermondii on 99%-pure n-paraffins, ~65% protein, claimed yield ~1 g cells per g paraffin (best case — >1 g/g once O₂ and minerals are counted is not physical). Commercial plants at Kstovo (1973) and Kirishi (1974); eight plants by 1989 under the Ministry of Microbiological Industry.

Why it collapsed

(a) Economics — the primary killer. The 1970s oil shocks drove up petroleum/methanol feedstock prices (feedstock was roughly half the production cost) exactly while soybean meal and fishmeal stayed cheap. Pruteen sold ~$600/ton vs. soy meal ~$125–200/ton (1984). SCP was a technical success that simply couldn't match the price of conventional protein.

(b) Toxicology — real concerns, but overstated in the popular telling. n-paraffins can carry benzo[a]pyrene (a carcinogenic PAH) and hydrocarbon-grown cells incorporate odd-chain fatty acids and residual alkanes that accumulate in fed-animal tissue. But the benzo[a]pyrene objection applies to the gas-oil route — purified (>99%) n-paraffins largely remove the PAH carry-over. The Italian Istituto Superiore di Sanità studied Toprina in pigs and cattle (results c. 1975, published in the Annali ISS, 1979): petroleum n-paraffins did accumulate in the animals' fat (~50–70 ppm) and heart — but not liver or brain — and odd-chain fatty acids (e.g. heptadecanoic) built up dose-dependently in tissue, milk, and offspring. Crucially, the studies showed no carcinogenicity; benzo[a]pyrene was a concern (a known petroleum contaminant), not a demonstrated finding. Italy's Consiglio Superiore della Sanità nonetheless ruled the product “not yet advisable” in 1978 on residue/contaminant grounds. Modern consensus: purified-substrate SCP is likely safe; the collapse was mostly economics + regulation + public acceptance, not proven carcinogenicity.

(c) Regulation & public resistance. Italy's Sarroch / Italproteine plant (a ~100,000 t/yr BP–Italian JV) was blocked by health authorities and never reached commercial production — it managed only a few hundred tonnes of trial output (Dec 1976–Apr 1977) before being moved to liquidation in 1978, killed by the “polli al petrolio” (“petroleum chickens”) press scare. Japan's industry abandoned petro-protein amid 1972–73 opposition. BP wound down its SCP plants across the mid-to-late 1970s — museum historian Fortier dates Lavéra to 1975, Grangemouth and Sarroch to 1978 (these specific years trace to that one source; academic reviews place the European wind-down around 1977, after the 1973 oil shock). The Soviet plants were hit later by a different problem — the 1987 Kirishi disaster, which reportedly involved poorly-filtered airborne protein dust causing mass allergic asthma in neighbours and triggering some of the USSR's first large environmental protests.

What's verified vs. contested

This section was triangulated across the deep-research pass + Codex + Gemini. To be honest about confidence:

• Solid (multi-source / primary-confirmed): the people, organisms, sites, the Toprina G/L split, plant capacities (16k / 4k / 100k t/yr), and Pruteen's core engineering; Sarroch never reached commercial production; Soviet 1.8M = total microbial-protein capacity vs ~1M actual output; the carcinogen story is overstated for purified feed. The ISS findings are now confirmed against the primary 1979 paper — n-paraffins ~50–70 ppm in fat/heart (none in liver/brain), dose-dependent odd-chain fatty-acid build-up, and no carcinogenicity (this also corrects an earlier “~1,000 ppm” figure, which was ~20× too high).
• Reported but not independently locked down: the exact closure dates (esp. Lavéra 1975, single-source; reviews say ~1977), the few-hundred-tonne Sarroch trial figure, the Kirishi airborne-protein-dust health mechanism (one 1990 advocacy source + general occupational-dust toxicology — and there is no formally named “Kirishi syndrome”), and whether Japan's exit was a formal ban or an industry withdrawal under pressure. The Kirishi plant, its 1987 trigger, the protests, and the 1989 closures are well-attested; only the specific health-causation detail is thin.
• Refuted in verification: the tidy “Soviet plants closed because of alkane toxicity + green pressure” story (the closures were real but multi-causal), and specific Pruteen 1979–1985 production-period dates.

For the full source-by-source breakdown, see the annotated reading guide.

Sources

• Alkane hydroxylase structural diversity & n-alkane biodegradation mechanisms — Frontiers in Microbiology, 2013
• Specificity at the End of the Tunnel: AlkB substrate-length discrimination — PMC
• Electrochemical hydroxylation of C3–C12 n-alkanes by recombinant AlkB — PMC
• Proteomic insights into metabolic adaptations in Alcanivorax borkumensis (glyoxylate + gluconeogenesis) — J. Bacteriol., 2006
• The Messy Alkaline Formose Reaction and Its Link to Metabolism — PMC
• Exploring the Core Formose Cycle: Catalysis and Competition — PMC
• Single-cell protein / BP “proteins-from-oil” process — Wikipedia

Part 2 (history) — selected; full source-by-source breakdown in the reading guide:

• USSR: Single-Cell Protein Industry (SI 77-10014K) — declassified CIA assessment, 1977 (Soviet capacities, plants, quality standards)
• US Office of Technology Assessment — 1984 (Soviet actual output ~1M t/yr)
• Fortier, “petroleum steak tartare…” — Ingenium, 2024 (plants & closure dates: Lavéra 1975, Sarroch & Grangemouth 1978)
• Garattini, Paglialunga & Scrimshaw (eds.), SCP: Safety for Animal and Human Feeding — Pergamon, 1979 (the 1977 Milan toxicology symposium)
• Champagnat, “Protein from Petroleum” — Scientific American, 1965 (the original manifesto)
• Rogers, “Petroprotein” — Yale, 2025 (the dedicated scholarly history)
• Alimenti, Boniforti, Silano et al., “Indagini nutrizionali sulla Toprina” — Annali ISS 15:649–690, 1979 (the primary Italian toxicology paper; n-paraffin & odd-chain-fatty-acid tissue data)
• Bizzi & Garattini, “Accumulation and metabolism of uneven fatty acids in SCP” — Toxicology Letters 5, 1980
• Groenewald et al., Yarrowia lipolytica safety assessment — Crit. Rev. Microbiol., 2013 (Toprina plants & safety)
• Ritala et al., “Single Cell Protein — state-of-the-art” — Front. Microbiol., 2017
• Ekenvall et al., “Single cell protein as an occupational hazard” — Br. J. Ind. Med., 1983 (SCP-dust respiratory reactions)

Research note assembled 2026-06-08 from a chat exploring whether crude-oil polycarbons can be converted to sugar. Web-sourced claims are cited above; combustion enthalpies and oxidation-state arithmetic are standard textbook values.