From Natural Gas to Syngas: A Process Engineer’s Guide to Steam Methane Reforming

Almost every conversation about hydrogen today is really a conversation about syngas. Green, blue, grey, coal-derived, LNG-derived, biomethane-derived — the feedstocks and the politics differ enormously, but a very large share of the world’s industrial hydrogen still arrives through one common conceptual step: converting a hydrocarbon-rich gas into a controlled mixture of hydrogen and carbon oxides. This article is about how that step actually works, and about the engineering decisions a process engineer has to make before the syngas ever reaches the shift section.

1. The hydrogen revolution is actually a syngas revolution

Pick up any industry publication and you will find hydrogen sorted by colour. Green hydrogen from electrolysis. Blue hydrogen from natural gas with carbon capture. Grey hydrogen from natural gas without it. Alongside these sit coal-to-methanol, coal-to-ammonia, LNG-based hydrogen, and the growing interest in compressed biogas and biomethane as feedstocks.

These pathways look completely different at the front end. A coal gasifier and an electrolyser have essentially nothing in common. But look at what the downstream chemistry actually needs, and a pattern appears. Ammonia synthesis needs hydrogen and nitrogen in a particular ratio. Methanol synthesis needs hydrogen, carbon monoxide and carbon dioxide in a particular ratio. Fischer–Tropsch needs hydrogen and carbon monoxide in a particular ratio. Refinery hydrotreating needs hydrogen of a particular purity and pressure.

In every one of those cases except pure electrolysis, the plant is not really in the business of making hydrogen. It is in the business of making a synthesis gas of specified composition. Hydrogen is one component of that gas. The reformer — or the gasifier, or the autothermal reactor — is the unit that turns a carbon-and-hydrogen-bearing feed into that specified composition.

2. What this article covers — and what it does not

The scope is deliberately narrow, because the alternative is a survey that teaches nothing well.

Solid feeds are also out of scope. Coal and biomass gasification produce syngas by a different route with different contaminants and a different front end; the reforming chemistry discussed here begins with a gas that is already predominantly methane.

3. What is really entering the reformer?

Define the boundary first. This article assumes a gaseous feed in which methane is roughly 85–95 mol% or higher TYPICAL RANGE, with the balance made up of heavier hydrocarbons, carbon dioxide, nitrogen, hydrogen, sulfur species, water and traces of other contaminants.

That gas may have come from a transmission pipeline, from LNG regasification, from compressed biogas or upgraded biomethane, from associated gas at an oil field, from coal-bed methane, or from a refinery off-gas stream. Here is the point that beginners consistently miss:

What differs by source is the likelihood and variability of particular contaminants. Biomethane is more likely to carry siloxanes, oxygen and variable CO₂. Associated gas is more likely to carry heavy ends. Odorised pipeline gas carries deliberately added sulfur compounds. Design for the analysis; use the origin to decide how much margin and how much analytical vigilance you need.

The components, one at a time

4. Feed composition determines pretreatment

This is the single most useful table in the article. Read it as a chain, left to right: something is in the gas, it causes a specific kind of damage by a specific mechanism, so a specific unit operation is installed, and the reformer only tolerates a certain amount of what gets through.

5. The generic block flow

Before the detail, the shape of the thing. This is a generic, textbook-level arrangement — not any particular plant.

Two things are worth noticing immediately. First, the reformer sits at the junction of a chemical path and a thermal path, and the two are equally important. Second, the fuel side is roughly as complex as the process side — which is a hint about where the operating cost lives.

The same thing again, with the equipment in it

The block flow above is the shape. This is the same process drawn as a process flow diagram, with the individual items of equipment, the fuel and flue-gas circuit, and numbered streams you can follow through the table underneath.

What a reforming plant actually looks like

Block diagrams flatten a reforming plant into a single plane. On site the same equipment is stacked several floors high, insulated, wrapped in steelwork and largely hidden from the ground. These photographs are here so that the arrangement above has something physical attached to it, and so that the words radiant box, transfer line and convection section stop being abstractions.

Photographs are royalty-free stock images used for illustration only. They are representative industrial views — they do not depict any licensor’s design, any client plant, or any specific reformer described in this article.

6. Steam reforming versus autothermal reforming

There are two dominant routes from a methane-rich gas to syngas at industrial scale, and the difference between them comes down to a single question: where does the heat come from?

7. The downstream product defines the syngas

Here is the logic that beginners find counter-intuitive: the reformer is designed backwards, from the product.

8. The first engineering decision: pressure

Now we leave chemistry and enter engineering. Ask a student what pressure a reformer should run at and the answer is usually “low, because the reaction produces more moles.” That answer is thermodynamically correct and economically wrong, and understanding why is a genuine step up in engineering maturity.

What thermodynamics says

The three reactions that matter:

Reactions (1) and (2) each convert two moles of reactant into four moles of product. By Le Chatelier’s principle, raising the total pressure pushes an equilibrium toward the side with fewer moles — here, backward, toward unreacted methane. Reaction (3) has no change in mole number and is essentially indifferent to pressure.

The consequence is unambiguous: at constant temperature and steam ratio, raising pressure increases methane slip. If equilibrium conversion were the only consideration, you would reform at the lowest pressure you could manage.

Two levers can be used to claw the conversion back. You can raise temperature, because both reforming reactions are endothermic and equilibrium therefore moves forward with temperature. Or you can raise the steam-to-carbon ratio, because steam is a reactant on the left-hand side. Both cost money, which brings us to the real question.

What economics says

A reformer does not exist in isolation. Look at where the gas came from and where it is going.

It came from a pipeline, quite possibly at a pressure well above what the reformer wants. It is going to a shift section, then to purification, then to a synthesis loop or a delivery header — and those downstream steps frequently want higher pressure than the reformer outlet. An ammonia synthesis loop, a methanol loop, a PSA-fed hydrogen header: all of them need pressure.

So if you reform at low pressure, you have to put the pressure back afterwards. And that is expensive in a particularly unpleasant way, because the syngas leaving the reformer has a very low molecular weight. Compression work per unit mass scales inversely with molecular weight, so compressing hydrogen-rich syngas is far more work than compressing the natural gas feed that produced it.

The compressor power relationship

For an ideal-gas polytropic compression the shaft power follows:

where Z is the compressibility factor, T1 the suction temperature, n the polytropic exponent, P2/P1 the pressure ratio, m the mass flow, M the molecular weight and eta the efficiency.

Read the assumptions before you use it. It assumes ideal-gas behaviour with a compressibility correction, a constant polytropic exponent across the stage, a single well-mixed gas composition, no phase change, no leakage and adiabatic operation. Real natural gas at high pressure deviates from all of these to some degree.

The important structural insight is in the exponent. Power does not scale with the pressure difference; it scales with the pressure ratio raised to a fractional power. Going from 10 to 20 bar and going from 20 to 40 bar both represent a ratio of 2 and cost broadly similar work per unit mass, even though the second step involves twice the pressure difference. Power also scales directly with suction temperature, which is why interstage cooling matters so much, and inversely with molecular weight, which is the syngas problem described above.

So what is the optimum?

The competing effects, laid out honestly:

Industrial fired reformers typically land somewhere in the range of about 15–40 bar at the outlet TYPICAL RANGE, with hydrogen plants often toward the middle and ammonia plants toward the upper end, precisely because that is where these curves cross for realistic equipment costs. It is a compromise, and it should be presented as one.

9. Joule–Thomson cooling and hydrate formation

This section exists because of a specific, common, and entirely avoidable mistake.

What actually happens across the valve

Throttling a real gas is an isenthalpic process, not an isothermal one. Because real gases have non-ideal intermolecular forces, expansion at constant enthalpy changes the temperature. The Joule–Thomson coefficient describes this, and for natural gas at normal pipeline conditions it is positive — meaning the gas gets colder as it expands.

A commonly quoted rule of thumb for natural gas is a cooling of roughly 0.4–0.6 °C for every bar of pressure drop TYPICAL RANGE, varying with composition, pressure level and temperature. Drop 40 bar and you can lose 20 °C or more.

Why cold gas is a problem

Three separate things can go wrong as the gas cools, and they can happen simultaneously.

Water condenses. Cooling lowers the water dew point margin. Free liquid water appears in the line.

Hydrocarbons condense. If the gas contains C₃+ material, cooling can cross the hydrocarbon dew point and drop liquid, which is bad news for absorbent beds and for any downstream compressor.

Hydrates form. This is the one that catches people out. Gas hydrates are ice-like crystalline clathrates in which water molecules cage light hydrocarbon molecules. They form at temperatures well above 0 °C when pressure is high — commonly at 15–20 °C at pipeline pressures with free water present TYPICAL RANGE. An operator who reasons “the gas is at 10 °C, water freezes at 0 °C, so we are fine” is about to plug a valve.

The variables that set the hydrate formation temperature are pressure (higher pressure means hydrates form at higher temperature), water content (no free water, no hydrate), and gas composition (heavier gases form hydrates more readily).

The engineering sequence

The essential move is the first one: heat before you throttle, not after. Heating downstream of the valve is too late — the hydrate has already formed at the point of lowest temperature, which is inside or immediately after the restriction. Preheating puts enough enthalpy into the gas that the post-throttle temperature lands safely above the hydrate curve with margin.

Where preheating is impractical, the alternatives are dehydration upstream to remove the water entirely, or injection of a thermodynamic inhibitor to depress the hydrate formation temperature. Both are legitimate; both cost more than a heater.

10. Compressor configuration

No mechanical design here — that is a specialist activity. What a process engineer needs is to understand how the feed conditions dictate the machine configuration.

The starting point is always the same two numbers: the suction pressure available at the battery limit, and the discharge pressure the reformer needs. Their ratio drives everything else.

Single stage or multistage?

The deciding factor is usually discharge temperature. Compression is close to adiabatic, so the gas heats up as it is compressed, and beyond a certain ratio the discharge temperature exceeds what the materials, the seals or the lubricant can tolerate. Splitting the compression into stages with cooling between them keeps every stage within limits and, as a bonus, reduces the total power because each stage starts from a lower suction temperature.

Interstage cooling brings its own requirement: cooling condenses whatever the gas was carrying, so every interstage cooler needs a knockout drum behind it. Liquid entering a compressor stage is a serious mechanical event.

The other configuration items follow from operability rather than thermodynamics. A recycle line from discharge back to suction lets the machine keep running when the plant turns down. For a centrifugal machine, an anti-surge system is essential — surge is a flow reversal that occurs when the machine is asked for too much head at too little flow, and it is destructive.

11. Inside the primary reformer

This is the heart of the article. A fired reformer is simultaneously a catalytic reactor, a fired heater and a high-temperature pressure vessel, and it has to succeed as all three at once.

The physical arrangement

The components, and what each one is really doing

Firing arrangements, flux profiles and catalyst

Three things decide how evenly heat arrives at the tubes and how well it is used: where the burners are, what flux profile that produces, and what the catalyst inside the tube can do with it.

12. Reforming reaction chemistry

The reactions that produce the product:

Both occur simultaneously on the same nickel catalyst inside the tube. The shift reaction is fast and stays close to equilibrium throughout, which means the CO/CO₂ split at the reformer outlet is essentially set by the outlet temperature and the steam content, not by anything you do separately. Detailed shift reactor design is outside this article’s scope; what matters here is that shift is already happening inside the reformer and shapes the gas you deliver.

The behaviour follows directly from the thermodynamics. Higher temperature increases conversion because the dominant reaction is endothermic. Higher pressure decreases conversion because moles increase. Higher steam-to-carbon increases conversion because steam is a reactant, and it also suppresses carbon. Longer residence time and higher catalyst activity both bring the outlet closer to equilibrium.

The reactions that destroy the plant

Three carbon-forming reactions compete with the useful chemistry, and understanding them is the difference between operating a reformer and merely running it.

Carbon does its damage in several ways. It physically blocks catalyst pores and active sites. It builds up in the bed and raises pressure drop, triggering the maldistribution feedback loop described earlier. Filamentous carbon growing from nickel crystallites can physically fracture catalyst pellets. And because carbon-covered catalyst cannot absorb heat by reaction, the tube wall above it overheats.

The defences, in order of importance: maintain adequate steam-to-carbon ratio, because steam gasifies deposited carbon as fast as it forms; remove or pre-reform heavy hydrocarbons before the radiant tubes; keep sulfur out, since a poisoned catalyst runs hot and hot catalyst cracks methane; fire evenly, because carbon forms first wherever the flux is highest; and never let steam-to-carbon fall during a transient, which is why low-steam trips exist on essentially every reformer ever built.

13. Kinetics versus equilibrium

Most textbooks teach reforming as an equilibrium problem. That is a useful first approximation and a dangerous stopping point.

Industry expresses the gap as the approach to equilibrium: the difference between the actual outlet temperature and the temperature at which the measured outlet composition would be exactly at equilibrium. A well-performing reformer with fresh catalyst might show an approach of a few degrees to perhaps 10–15 °C TYPICAL RANGE. As catalyst ages or is poisoned, the approach widens — and tracking that widening over months is one of the most valuable diagnostics a plant has.

There is a further subtlety that pure equilibrium thinking hides completely. Inside a reformer tube, the reaction is severely limited by heat transfer and by diffusion into the catalyst pellet. The effectiveness factor — the fraction of the catalyst volume actually participating — is very low, because the reaction is fast relative to the rate at which reactant can diffuse into the pellet interior. Most of the reaction happens in a thin shell near the pellet surface.

14. Why firing is required

Beginners sometimes assume the furnace is there to preheat the gas. It is not. It is there because the reaction itself consumes an enormous amount of energy, continuously, along the entire length of the tube.

The energy path is worth tracing explicitly. Burner combustion releases heat into the box. Radiation carries it to the tube outer surface. Conduction carries it through the tube wall. Convection and conduction carry it from the inner wall into the flowing gas and into the catalyst. Reaction absorbs it. Whatever the reaction does not absorb, the metal has to hold.

A useful order-of-magnitude picture: of the heat released by the burners, roughly half is absorbed by the process in the radiant section, a large further fraction is recovered in the convection section, and the remainder leaves as stack loss and casing loss TYPICAL, ORDER OF MAGNITUDE. The exact split varies with design, but the message is that the radiant box alone is not where the economics are decided.

15. The heat balance is an economic story

Presented as a calculation, the heat balance is dull. Presented properly, it is the plant’s profit and loss account.

In a well-integrated hydrogen or ammonia front end, feedstock and fuel together dominate the cash cost of production — commonly quoted as the large majority of operating cost for natural-gas-based hydrogen TYPICAL. That means the heat balance is not a checking calculation performed at the end of the design. It is the design.

What the recovered heat is doing

What poor heat recovery actually costs

Let the heat recovery degrade — fouled convection coils, air preheater bypassed, high excess air, a leaking damper — and the consequences compound. Fuel consumption rises for the same production. Stack temperature rises, taking useful energy out of the plant. Steam generation falls, so the shortfall must be made up by a fired boiler somewhere else. Specific energy consumption per tonne of product rises. And because the CO₂ emission of the plant is essentially proportional to the carbon burned, the carbon footprint rises in direct proportion — which in a carbon-priced market is now a cash cost, not a reporting line.

16. What the operator is actually trying to control

Everything above is design. This section is about the shift. An operator is holding several competing objectives in tension, all day, with instruments that only partially show what is happening inside the tubes.

The six trade-offs, and what happens when you get them wrong

17. How configuration changes with the application

Reading across a row is the exercise. The same unit operation is being tuned toward four different targets, and the design decisions that look arbitrary in isolation are entirely determined once you know which column you are in.

18. Worked numerical examples

Example 1 — Steam-to-carbon ratio for a feed containing heavy ends

Assumptions: feed of 100 kmol/h containing 90% CH₄, 5% C₂H₆, 2% C₃H₈, 2% CO₂, 1% N₂. Target S/C = 3.0 on total hydrocarbon carbon.

Lesson: 7 mol% of C₂/C₃ raised the steam demand by 6% at identical molar feed. Put the other way round: if the steam flow controller is holding the pure-methane setpoint and the feed turns rich, the actual S/C has silently fallen to about 2.83 — and the operator sees nothing on the panel unless the ratio is computed from a live analysis.

Example 2 — Joule–Thomson temperature drop and hydrate margin

Assumptions: supply at 55 barg and 25 °C, let down to 30 barg. J-T coefficient taken as 0.5 °C/bar. Hydrate formation temperature at 30 barg with free water present assumed to be 15 °C.

Now add preheat upstream of the valve to 45 °C:

Lesson: the fix is upstream heating, and the required preheat comes straight out of the arithmetic. Check it at the lowest credible supply temperature and the highest credible pressure drop, not at the design point — that is where the margin disappears.

Example 3 — Multistage compression and the pressure-ratio rule

Assumptions: ideal gas, k = 1.31, suction 5 bara at 40 °C, discharge 40 bara.

Lesson: splitting the ratio and cooling between stages dropped the discharge temperature from 219 °C to 122 °C, bringing it inside normal material and seal limits — and it reduces total power as well, because the second stage starts cold. Equal ratio per stage is the classic first estimate for minimum total work.

Example 4 — Reformer heat duty, first-pass estimate

Assumptions: 100 kmol/h methane fed, 92% conversion, heat of reaction for CH₄ + H₂O → CO + 3H₂ taken as 206 kJ/mol at 298 K. Sensible heating of feed and steam from 500 to 850 °C estimated at 35% of the reaction duty for this illustration.

Lesson: the reaction is only part of the duty, and roughly half of what you fire never reaches the process in the radiant box. Whether that half is recovered or thrown away is the difference between an efficient plant and an expensive one.

Example 5 — The economic cost of one degree of efficiency

Assumptions: a unit firing 50 MW of fuel, 8000 operating hours per year, fuel valued at USD 8 per GJ, natural gas emission factor 56 kg CO₂ per GJ.

Lesson: one percentage point. That is the value of keeping the convection coils clean, the excess oxygen trimmed and the air preheater in service. Multiply by the number of percentage points a neglected unit typically loses and the case for condition monitoring writes itself.

Try it yourself — steam-to-carbon calculator

19. Oversimplifications to avoid

A short list of statements that sound reasonable, appear in introductory material, and will mislead you.

20. Self-check

Where to read further

This article deliberately cites categories of source rather than proprietary documentation. For deeper study, the most reliable open material is found in: standard chemical engineering texts on hydrogen and synthesis gas manufacture; peer-reviewed literature on nickel reforming catalysts, carbon formation mechanisms and reforming kinetics; published technical papers from ammonia, methanol and hydrogen safety and technical symposia; catalyst manufacturers’ publicly issued technical bulletins; national laboratory and government technical reports on hydrogen production pathways and life-cycle emissions; and international standards covering fired heater design, gas dehydration and hydrate prevention.

The one-paragraph summary