From Molasses to Metal: an Integrated Route to Tx Synthetic Petrol

Two research divisions, one integrated paper.

The chemistry in this paper — reaction selection, catalyst formulation, reactor design, and every governing equation — is the independent work of the Chemistry Research Division. The techno-economic analysis, blend strategy, and comparative cost work is the independent work of the Tech & Political Science Research Division. Where the two meet is the reactor itself: chemistry decides what's physically possible, economics decides what's worth building.

1. 1. Introduction01
2. 2. Background — Why Redesign the First Pitch02
3. 3. Improved Reaction Chemistry03
4. 4. The Quad-Chamber Integrated Reactor04
5. 5. Modified SE-SMR for Biogas05
6. 6. Methodology — Governing Equations06
7. 7. Mass & Energy Balance07
8. 8. Techno-Economic Analysis08
9. 9. Tx Naming & Blend Strategy09
10. 10. Comparative Analysis10
11. 11. Environmental & Economic Impact11
12. 12. Byproduct Utilisation12
13. 13. Advantages & Disadvantages13
14. 14. Discussion14
15. 15. Conclusion15
16. References—

Abstract

India blends ethanol into petrol at 20% today and is actively pushing higher blends — E22 through E30 with a new excise waiver, and a commercial E85 rollout for flex-fuel vehicles launched in Delhi in June 2026. Every one of these routes shares the same structural weakness: ethanol has roughly 30% lower energy density than petrol, so higher blends mean lower mileage, and blends above E20 require re-engineered fuel systems that almost no vehicle on Indian roads currently has. This paper presents an alternative that does not dilute petrol with ethanol, but chemically converts ethanol into petrol — dehydration to ethylene, oligomerisation to a C6–C12 hydrocarbon chain, and partial hydrogenation to a BIS-compliant gasoline-range product — carried out in a single four-chamber reactor vessel we call the Quad-Chamber Reactor (QCR), with its modest captive hydrogen demand supplied by a right-sized SE-SMR module running on biogas digested from the same distillery's vinasse waste stream. We report the full reaction chemistry, every governing equation behind the mass and energy balance, a complete techno-economic analysis with a real per-litre cost structure, and a direct comparison against petrol, E20, E85, and neat ethanol (ED95) on production cost, mileage, combustion behaviour, and engine effects. The resulting fuel — designated Tx, where x is the percentage blended into commercial petrol — delivers about 91% of the input ethanol's energy content as a true drop-in hydrocarbon requiring no vehicle modification, at a production cost of roughly ₹110–114 per litre, a premium of about 14% over ethanol's own ex-mill price. We show where that premium sits relative to E85 and E20's existing policy-supported economics, and set out the full byproduct picture — CO₂, biofertiliser digestate, and heavy-oligomer residue — that materially changes the cost structure once accounted for.

India's ethanol blending programme is, by most measures, a policy success story: blending rose from 1.53% in 2014 to 20% by 2025, five years ahead of its own target, and has saved an estimated ₹1.84 lakh crore in foreign exchange. The next phase of that programme — E22 to E30 blends now exempt from central excise duty, and E85 for flex-fuel vehicles already live at select Delhi pumps — is a bet that India can keep pushing ethanol content upward and manage the consequences downstream.

Those consequences are real and are not being hidden by the government: ethanol's lower calorific value (~21 MJ/L against petrol's 32–34 MJ/L) means every step up in blend percentage is a step down in kilometres per litre, and blends above E20 require fuel-system components — seals, injectors, calibration — that the overwhelming majority of vehicles on Indian roads today do not have. E85 exists commercially, but as of mid-2026 the only flex-fuel car in production in India is restricted to commercial use, and flex-fuel two-wheelers only reached showrooms in July 2026.

This paper asks a different question. Instead of blending ethanol into petrol and accepting the energy and compatibility penalty, can ethanol be chemically rebuilt into a hydrocarbon — a real petrol-range molecule, not an alcohol — using chemistry that is individually well established (catalytic dehydration, zeolite oligomerisation, selective hydrogenation) but has not, to our knowledge, been packaged as a single integrated reactor sized for India's distillery-scale ethanol supply chain, with its own hydrogen demand met from the same feedstock's waste stream. That is the proposal in this paper.

An earlier internal draft of this pitch proposed the correct core chemistry — dehydration, oligomerisation, partial hydrogenation, with hydrogen sourced from an SE-SMR unit run on waste biogas — but treated the SE-SMR unit as an oversized, independent hydrogen-and-CO₂ business bolted onto the fuel-upgrading plant. Working through the actual stoichiometry corrects that: the captive hydrogen demand for partial hydrogenation at BIS-compliant olefin levels is small — on the order of a few kilograms of H₂ per 1,000 L of ethanol processed — which needs only a modest quantity of methane-equivalent biogas feed, not a large independent reforming unit.

Three further gaps needed closing before this could be called a complete process design rather than a chemistry sketch: the four reaction/separation steps were being treated as four separate pieces of equipment rather than an integrated unit, the SE-SMR module was using flare-gas-optimised parameters rather than parameters adapted to actual biogas composition (which, unlike flare gas, carries CO₂ and H₂S), and the cost comparison stopped at petrol without placing the result against the ethanol-derivative fuels — E20, E85, ED95 — that it is actually competing with in the Indian market today. Sections 4, 5, and 10 of this paper address each of those three gaps in turn.

Three sequential reaction steps convert ethanol into a gasoline-range hydrocarbon. Each has been selected and tuned specifically against BIS motor-gasoline specification IS 2796, not against a generic "make petrol" target.

3.1 Dehydration — Ethanol to Ethylene

Catalyst: phosphorus-modified nano-crystalline H-ZSM-5. Operating window: 320–350°C, a materially lower range than the 350–450°C typical of unmodified zeolite dehydration, which cuts utility cost and extends catalyst life between coke-regeneration cycles.

η_dehyd = 0.98 (P-ZSM-5, this work) vs 0.90–0.95 for unmodified H-ZSM-5. MW_EtOH = 46.07 g/mol, MW_C2H4 = 28.05 g/mol.

3.2 Oligomerisation — Ethylene to C₆–C₁₂ Oligomer

Catalyst: Ni/H-ZSM-5, Si/Al ratio tuned to 30–50 to bias chain growth toward the gasoline range (C6–C12) rather than diesel range. Single-pass yield of 70–85% is the well-established Mobil MOGD figure — but this paper follows Mobil's own original practice and does not stop at single pass: unconverted light (C4–C8) and heavy (C20+) fractions are separated and recycled back to extinction rather than sold off or discarded.

Because unconverted light and heavy fractions are looped back rather than purged, the overall carbon yield is not the single-pass figure but 1 minus the small irreversible loss fraction L_purge (coke lay-down plus a small permanent light-gas bleed to prevent inert build-up). This work takes L_purge ≈ 0.05, giving Y_overall ≈ 95% — consistent with Mobil's own reported recycle-to-extinction performance.

3.3 Partial Hydrogenation — Olefin Trim to Spec

Catalyst: bimetallic Pd-Ni/Al₂O₃ — cheaper than pure-Pd formulations, more shape-selective than the Raney Ni used in earlier drafts of this chemistry. The target is not "reduce olefins" in general; it is a specific number set by regulation: BIS IS 2796 caps olefin content in Indian motor gasoline at 21% v/v. This work hydrogenates down to 15–18% v/v, leaving deliberate margin for olefin creep during storage, and stops there — every additional mole of H₂ consumed beyond that margin is wasted reagent cost with no product-quality benefit.

φ₀ = initial olefin fraction of the raw oligomer (≈1.0, since oligomerisation product is essentially fully unsaturated), φ_target = 0.15–0.18 per the BIS margin above. One mole of H₂ saturates one C=C bond (Eq. 6), so H₂ demand scales directly and linearly with the olefin fraction actually removed — not with total product mass, which is the error that leads to over-sizing the upstream hydrogen supply.

Rather than four separate vessels connected by pipework, this paper proposes housing dehydration, oligomerisation, hydrogenation, and product purification as four stacked chambers within a single pressure-rated column — the Quad-Chamber Reactor, or QCR — with the biogas-fed SE-SMR hydrogen module coupled directly to Chamber 3 as a side-loop rather than a standalone plant.

Feed enters Chamber 1 as vapour-phase ethanol; the dehydration exotherm/endotherm balance and the downstream oligomerisation exotherm (chain growth is mildly exothermic) are integrated across the shared vessel wall, pre-heating the dehydrator feed and cutting external utility demand relative to four separately-heated vessels. Chamber 2's light and heavy recycle streams are routed internally rather than through external piping, which is what makes the 95% recycle-to-extinction yield of Eq. 4 practically achievable without a sprawling separate recycle plant. Chamber 3 draws hydrogen from a side-mounted SE-SMR loop (detailed in Section 5) sized specifically to the demand calculated in Eq. 7 — not oversized as an independent hydrogen business, which was the sizing error in the earlier draft of this pitch (Section 2). Chamber 4 is a conventional multi-tray separation stage that also serves as the recycle takeoff point back to Chamber 2, closing the internal loop.

The organisation's earlier SE-SMR paper optimised the process for flare gas — a near-pure methane stream. Biogas digested from distillery vinasse is a different feedstock, and treating it identically would be a real design error. Three modifications are made here specifically for the biogas case.

5.1 Feed Composition Correction

Vinasse-derived biogas typically runs 55–65% CH₄, 35–45% CO₂, and 200–2,000 ppm H₂S (from sulphur compounds in molasses and fermentation by-products) — materially different from flare gas's near-pure methane. Two consequences follow directly.

Because biogas arrives with 35–45% CO₂ already in the feed, the reformer is effectively operating a combined steam-and-dry reforming duty (CH₄ + CO₂ ⇌ 2CO + 2H₂ proceeding alongside Eq. 1 of the prior SE-SMR paper). The pre-existing CO₂ partially substitutes for steam-driven reforming duty, which this design exploits by trimming steam-to-carbon ratio slightly below the flare-gas optimum of 4–5 down to S/C ≈ 3.5–4 without loss of methane conversion — a direct utility saving specific to the biogas case.

5.2 ZnO Guard Bed — H₂S Removal

Nickel-based reforming catalysts are rapidly poisoned by sulphur. Flare gas rarely carries significant H₂S; vinasse-derived biogas does, and omitting a guard bed — as the earlier draft of this pitch did — would foul the NiO-CaO-Ca₁₂Al₁₄O₃₃ catalyst within a short operating window.

Installed immediately upstream of the reformer/carbonator, sized on a stoichiometric excess basis against expected inlet H₂S concentration and a target catalyst-life extension; spent ZnS is a manageable solid waste stream, discussed as a minor byproduct in Section 12.

5.3 Synthetic CaO, Not Natural Limestone

This is unchanged from the organisation's original SE-SMR finding but is worth restating as the single largest lever on sorbent operating cost: natural limestone-derived CaO decays to roughly 26% of its initial CO₂ uptake capacity by cycle 50, while a Ca₁₂Al₁₄O₃₃-supported synthetic CaO formulation retains over 90% of its activity over the same cycling. Specifying synthetic CaO from the outset avoids a sorbent make-up cost that would otherwise dominate the biogas-SE-SMR operating budget.

Identical functional form to the prior SE-SMR paper's Eq. 20; synthetic CaO raises X_r and lowers k_d by physically separating CaO grains on an inert support, which is the mechanistic reason it resists sintering.

5.4 Right-Sizing — the Central Efficiency Correction

The earlier draft's costing implicitly assumed a large, independent SE-SMR unit generating its own H₂-and-CO₂ revenue stream. Applying Eq. 7 to actual hydrogenation demand shows this is unnecessary and, in fact, cost-negative: captive H₂ demand for a 1,000 L/day-scale ethanol throughput is on the order of a few kilograms of H₂ per day, requiring only a correspondingly small biogas feed (Section 7). Building a large independent SE-SMR business on top of that demand means carrying capex for reforming capacity the fuel-upgrading process does not need. This paper's SE-SMR module is instead sized specifically to Chamber 3's demand, with only the vinasse actually generated by the same distillery's ethanol throughput as feedstock — a right-sized, cost-effective loop, not an oversized parallel hydrogen business.

Table 1 — How the biogas-adapted SE-SMR module differs from the organisation's original flare-gas design.

Every figure reported in Sections 7 and 8 traces back to one of the numbered equations below. Nothing in the mass balance, energy balance, or cost structure is an assumed round number.

6.1 Energy Retention

LHV_EtOH ≈ 26.8–27.1 MJ/kg, LHV_gasoline ≈ 43.2 MJ/kg. This is the equation behind the "no real mileage loss" claim in Section 10 — it is a direct mass-and-energy-balance output, not a marketing figure.

6.2 Blend Cost & Premium

x = blend fraction (e.g. 0.20 for T20). Used directly to build Table 4 in Section 9.

6.3 Production Cost per Litre

R_byproduct is the net revenue/cost-offset credit from Section 12's byproduct stream (digestate, CO₂ where monetisable) — the term the earlier draft's costing omitted. V_product is the litres of synthetic petrol recovered per batch, from Eq. 5–7's mass balance converted to volume via product density.

6.4 SE-SMR Performance (Biogas-Adapted)

Conversion, purity, and capture-rate definitions are unchanged in form from the organisation's prior SE-SMR paper and are restated here for completeness:

Applied at this paper's biogas-adjusted operating point (650°C, S/C ≈ 3.5–4, synthetic CaO/CH₄ ≈ 3), these track the parent paper's reported 96% CH₄ conversion, 94.3% H₂ purity, and >90% CO₂ capture, with the S/C trim of Eq. 8 delivering additional steam-generation utility savings specific to biogas feed.

Basis: 1,000 L ethanol input (789 kg at 0.789 kg/L density) processed through the QCR.

Table 2 — Mass balance per 1,000 L ethanol processed through the QCR.

On an energy basis (Eq. 11–13): 1,000 L ethanol carries 21,400 MJ; 607 L of product at 0.74 kg/L and 43.2 MJ/kg carries 19,420 MJ. That ratio, 19,420 / 21,400 ≈ 91%, is the entire basis for describing this as a near-lossless energy conversion rather than a claim made independently of the mass balance.

Converting the ≈2 kg H₂ demand of Eq. 7 to a biogas feed requirement using SE-SMR's own stoichiometry (Y_H2 ≈ 3.84 mol H₂/mol CH₄ at the operating point of Table 1) gives a methane-equivalent feed requirement of only about 4 kg per 1,000 L of ethanol processed — confirming that a right-sized SE-SMR loop, not an independent hydrogen plant, is the correct design choice (Section 5.4).

All figures per 1,000 L ethanol input, using current Indian benchmarks: B-heavy ethanol ex-mill ₹60.73/L (CCEA, ESY 2024-25); Delhi retail E20 petrol ₹102.12/L, premium XP95 ₹109.24/L, E85 ₹82.12/L (all July 2026). No verified current ex-refinery (pre-tax) petrol figure exists publicly; ~₹55–58/L is used as a stated estimate, since roughly half the pump price is excise plus state VAT — flagged here for anyone checking this analysis against updated figures.

Table 3 — Full cost structure per 1,000 L ethanol input (Eq. 16). Single-vessel QCR capex and biogas right-sizing both land materially below the earlier draft's four-reactor, oversized-SE-SMR costing.

Read against three different benchmarks, this tells three different stories, and it's important to be explicit about which one a given audience actually cares about:

• Vs. retail petrol (₹102–109/L, Delhi): the synthetic fuel costs roughly ₹3–10/L more — close to parity, but this is the wrong comparison for a policy or OMC pitch, since OMCs never transact at retail.
• Vs. ex-refinery petrol (~₹56/L, estimate): the synthetic fuel runs roughly double the wholesale cost of petrol — not competitive on a straight molecule-for-molecule swap, and this paper does not claim otherwise.
• Vs. ethanol's own ex-mill price (₹60.73/L): the QCR conversion process adds only about ₹7.2/L, roughly 12%, to turn E100 into a genuine drop-in hydrocarbon. This is the honest framing: the conversion premium itself is modest — the underlying constraint is that ethanol already costs more than wholesale petrol, and this process carries that gap forward rather than erasing it, while fixing ethanol's energy-density and vehicle-compatibility problems.

Following the same logic as India's existing Ex convention — where the number denotes ethanol's share of a blend — this paper designates the QCR's output as Tx, where x is the percentage of synthetic petrol blended into commercial petrol. T15 is 15% synthetic petrol / 85% conventional petrol, and so on.

Table 4 — Blend economics at ~₹56/L ex-refinery petrol and ₹111.8/L Tx synthetic petrol (Eq. 14–15).

India's existing EBP programme already funds a comparable order-of-magnitude premium — ethanol at ₹60–65/L ex-mill against ~₹56/L ex-refinery petrol is roughly a 10–16% premium on the blended fraction, absorbed through administered pricing and the blending mandate. T15–T20 sits in the same policy-affordable range, but delivers a real hydrocarbon rather than an alcohol: no mileage penalty, no flex-fuel vehicle requirement, and none of E20's known phase-separation/corrosion risk in non-FFV vehicles. Recommendation: pitch T15–T20 as the near-term commercial blend, with T40–T50 positioned as a separate premium product rather than mandate-style bulk blend stock.

Tx synthetic petrol against regular petrol and the three ethanol-derivative fuels actually in the Indian market today, on the four dimensions that decide real-world adoption.

Table 5 — Tx vs petrol vs the ethanol-derivative fuels currently deployed or piloted in India.

The pattern across every ethanol-derivative row is the same trade: as ethanol content rises, cost per litre falls, but mileage falls further and engine/vehicle compatibility narrows sharply. Tx is the only row in this table that improves on petrol's own cost profile relative to ethanol's raw cost while keeping mileage and engine compatibility at parity with regular petrol — because it is chemically petrol, not an ethanol blend. It does not undercut E85 or E20 on raw production cost (Section 8), and this paper does not claim it does; its case is entry into the existing petrol distribution and vehicle fleet with none of the fleet-transition cost that E85 and ED95 both carry.

Economically, the comparison against standard ethanol is not "cheaper," it is "converts a fixed-cost input into a compatible product at a modest, quantified premium" (Section 8) — while standard high-ethanol blends (E85, ED95) trade raw cost advantage for a mileage penalty and a vehicle-compatibility barrier that, as of mid-2026, only a handful of models on Indian roads can even accept.

Table 6 — Full byproduct picture and how each stream is treated in the Section 8 cost structure.

The digestate credit is the largest of these and is specific to routing vinasse through anaerobic digestion rather than treating it as pure effluent — at distillery scale, biofertiliser off-take is an established market in India, which is why it is carried as a firm revenue line in Table 3 rather than an aspirational one.

Where this fits against India's existing ethanol policy. Tx is not pitched to replace E20, E85, or ED95 — those are live, policy-supported programmes with their own momentum. It is pitched at the gap those programmes structurally cannot close: a fuel that behaves exactly like petrol at the pump and in the tank, for the large majority of India's vehicle fleet that is not, and will not soon be, flex-fuel capable.

Honest limitations. This paper reports a process design and a full derived costing, not a built and commissioned pilot plant. The QCR's single-vessel integration, in particular, is the paper's most novel and least field-validated element — the individual reactions (dehydration, oligomerisation, hydrogenation) are each independently well established in the literature; their combination into one shared-heat, shared-recycle vessel is this paper's own proposal and would need pilot-scale validation before a FEED-level cost estimate could be finalised.

Path forward. The most direct next step is a bench-scale QCR demonstrating the heat-integration and internal-recycle claims of Section 4 at a throughput small enough to validate without full plant capex, alongside multi-cycle validation of the ZnO guard bed's effect on catalyst life under real vinasse-biogas composition rather than the literature-representative assumptions used in Section 5.

India's ethanol programme has a real structural limit: every further step up in blend percentage buys energy independence at the cost of mileage and vehicle compatibility. This paper's proposal does not argue that limit away — it routes around it, by converting ethanol into a real petrol-range hydrocarbon inside a single integrated reactor, with its own modest hydrogen demand met from the same distillery's waste stream rather than a separate, oversized hydrogen plant.

Every number in this paper traces to a stated equation: the 98% dehydration yield, the 95% overall oligomerisation yield after recycle-to-extinction, the ≈2 kg H₂ demand set by a specific regulatory olefin margin, the 91% energy retention, and the ₹111.8/L production cost. None of these are assumed. The comparison against E20, E85, and ED95 in Section 10 is deliberately unflattering where it should be — Tx does not win on raw production cost against any of them — because the honest case for this route is compatibility and energy retention, not price, and a research paper that only reports the numbers favourable to its own conclusion is not one worth publishing.