Inside the reformer’s side hustle: hydrogen supply, butane blending, and sulfur credits

Catalytic (naphtha) reforming doesn’t just make high‑octane reformate for gasoline. It throws off a hydrogen‑rich gas — typically ~70–85% H₂ with the balance mainly C1–C4 paraffins — that refineries increasingly monetize and control, from hydrogen networks to LPG pools (en.citizendium.org).

In practice, that off‑gas is amine‑scrubbed to remove hydrogen sulfide (H₂S), hydrogen is separated, and the residual light hydrocarbons (C₃–C₄) are recovered as fuel gas/LPG or blended. The economic and regulatory trade‑offs are real: volatility limits often push C₃–C₄ to fuel gas or LPG, but high gasoline prices and permissive seasonal Reid vapor pressure (RVP, a volatility measure) can make blending lucrative (oms-elearning-academy.com) (oms-elearning-academy.com).

Meanwhile, the H₂S that comes with sour feeds doesn’t leave the site: it is conventionally converted to saleable sulfur in a Claus sulfur recovery unit (SRU), with tail‑gas incineration to meet stringent stack limits. The sequence is amine absorber to SRU, and done well, >99% sulfur recovery is standard practice (sulfurrecovery.com) (thepetrosolutions.com).

Hydrogen recovery and refinery value

Reformers generate very large volumes of hydrogen — often supplying the bulk of a refinery’s hydrogen demand (for hydrodesulfurization, hydrocracking, and more), routinely “dozens of MMSCFD” in large refineries. Any excess H₂ can be sold or used for new processes (e.g., hydrogenation). Recovering this H₂ has huge value: replacing it via steam methane reforming would cost on the order of $1–3/kg H₂ (sciencedirect.com).

Yield can also be engineered. In one refinery study, upgrading the feed composition — specifically shifting C₆’s to isomerization — boosted net H₂ yield by 28–48% over conventional reforming (digitalrefining.com).

Separation tech matters. Pressure‑swing adsorption (PSA) and membranes typically recover hydrogen from the treated off‑gas; one membrane unit produced ≈98% pure H₂ at 250 psig with an initial ~36% recovery, and a properly scaled membrane train could reach ≈93% H₂ recovery (sciencedirect.com). With reformer off‑gas commonly ~80% H₂, these routes are central to cutting purchased hydrogen (sciencedirect.com).

Utilization of light hydrocarbons

The remaining off‑gas components are small alkanes and olefins — methane, ethane, propane, iso-/n‑butane. After hydrogen removal and dehydration, many refineries route this C₁–C₄ mix to the fuel gas system or to LPG storage.

Other outlets exist. In integrated sites, C₃ can be fed to propane dehydrogenation (PDH) or a steam cracker, and C₄ can be used for butadiene or alkylation feed. In practice, most Indonesian refineries lack dedicated PDH/cracking, so direct LPG sales are simpler. Indonesia’s LPG (mixed propane/butane) is heavily consumed (cooking fuel etc.), so large volumes of reformer C₃–C₄ typically enter the LPG pool.

Economically, refineries often keep C₃–C₄ out of the gasoline pool because blending introduces RVP constraints. But when gasoline prices rise relative to LPG, and when seasonal RVP specs allow, blending can be lucrative (oms-elearning-academy.com) (oms-elearning-academy.com).

Gasoline blending: octane vs. vapor pressure

The C₃/C₄ by‑products carry high octane — propane research octane number (RON) ≈ 105 and n‑butane RON ≈ 92 — making them powerful blendstocks (wikipedia77.en-us.nina.az) (wikipedia77.en-us.nina.az). The catch is volatility: they raise RVP sharply, especially propane.

That’s why blending is selective and seasonal. In temperate markets, summer RVP limits are ~7–9 psi (≈50–62 kPa); winter blends can be ~9–11 psi, opening room for butane (oms-elearning-academy.com). Analysts cite an average price spread of ≈US$2/gallon between winter gasoline (~$3.00/gal) and LPG‑butane (~$0.94/gal), enabling 10–12% butane by volume in winter gasoline to meet RVP without sacrificing octane — worth up to +$6 million margin on a 100,000‑barrel batch (oms-elearning-academy.com) (oms-elearning-academy.com).

Indonesia’s tropical climate raises vapor‑lock risk, and its fuel regulations have historically not restricted RVP stringently; however, care is still needed. Indonesia also raised its minimum gasoline octane from RON88/89 to RON90 in 2023, aligning with the high‑octane nature of LPG (qcintel.com). Even with high RON, the vapor‑pressure penalty generally caps blending (often ~10% for winter grade, and 0% butane in summer gasoline in some regimes) (oms-elearning-academy.com). Formulators typically balance octane gain by adding non‑volatile high‑octane components (e.g., aromatics or ethers) and using butane only seasonally.

Alternatively, the same light hydrocarbons can be directed to the LPG pool. Given local demand and subsidies, LPG often fetches a controlled price, so refiners compare the incremental value of butane in gasoline versus LPG and monitor fuel prices/specs closely. Any blending scheme must meet vehicle VOC/NOx emission standards and fuel stability requirements.

Amine sweetening and sulfur recovery

Any H₂S produced must be removed before the reformer to protect catalysts. The conventional route is an amine absorber (acid‑gas removal), often using an amine solvent such as CO₂/H₂S removal amine solvent, followed by a sulfur recovery unit (Claus process). Sequence: amine sweetening yields an H₂S‑rich acid gas; then the SRU converts it to elemental sulfur.

In the Claus SRU, ~30–40% of H₂S is combusted with air to SO₂ in the reactor furnace; that portion provides the heat for downstream conversion. The remaining H₂S plus generated SO₂ pass over catalyst beds at 150–300°C, yielding elemental sulfur and water — 2H₂S + SO₂ → 3S + 2H₂O; overall, 2H₂S + O₂ → 2S + 2H₂O. Molten sulfur is condensed after each reactor (sulfurrecovery.com).

A two‑stage Claus unit typically recovers ~95–98% of the inlet sulfur — standard design yields ≈96+% sulfur recovery. Adding a third catalytic stage or a tail‑gas cleanup (TGCU) unit such as Shell SCOT pushes recovery past 99%; well‑operated SRUs achieve >99% sulfur recovery. SCOT units convert remaining SO₂/S back to H₂S via hydrogenation, which is re‑absorbed, pushing net recovery toward 99.9% (sulfurrecovery.com) (thepetrosolutions.com).

The output sulfur is commercial‑grade molten sulfur (often ≥99% purity). Residual tail gas — trace H₂S/SO₂ — is sent to a thermal incinerator. Modern Claus/SCOT units burn tail gas to ~750–800°C, converting oxygenates to SO₂ and emitting SO₂ at low concentrations (typically a few hundred mg/Nm³ or less). With careful temperature control, residual H₂S can be ~0.1 ppm and SO₂ on the order of 250–300 ppm, meeting stringent norms (Indonesian regulations would require similarly low SOx/H₂S stack limits) (thepetrosolutions.com).

In Indonesia, gasoline sulfur specs currently allow up to 500 ppm S, and Euro‑IV (50 ppm) rules are phased in by 2025 (transportpolicy.net). One Indonesian refinery reported its Claus SRU processing ~73% of outlet H₂S vapor and yielding stack SO₂ ~14.7 mg/m³ — well below typical limits (wjarr.com). In short, the by‑product H₂S is converted almost entirely to saleable sulfur via the Claus process, with any remaining trace gases incinerated to comply with environmental standards (sulfurrecovery.com) (thepetrosolutions.com).

Decision parameters and documented figures

Key data points underpin choices on hydrogen recovery, LPG routing, gasoline blending, and SRU design: hydrogen yields can often be boosted 28–48% by feed optimization (digitalrefining.com); reformer off‑gas is ~80% H₂ and membranes have produced ≈98% pure H₂ (250 psig) with initial ~36% recovery and potential ≈93% recovery (sciencedirect.com); LPG blending (≈10–12% C₄) can add several $M per fuel batch when RVP allows (oms-elearning-academy.com); and Claus SRUs typically recover 96–99.9% of sulfur with TGCU (SCOT) (sulfurrecovery.com) (thepetrosolutions.com). These figures inform decisions on whether to redirect by‑products to gasoline versus LPG markets and how to size/operate the SRU.