How Does a Refinery Built to Process Sour Crude Differ From a Refinery to Process Sweet Crude?

You may think this is a stupid or irrelevant question to be asking when news is just breaking that Iran has launched some per-emptive strikes on US bases in the Persian Gulf and Jordan, but trust me, it is a very relevant question. (Note: I will deal with the military fallout from today’s new attacks in my post on Wednesday.)

The short answer to the question I posed above is that a sour-crude refinery is built around one dominant problem that a sweet-crude refinery is largely spared — sulfur — and almost every hardware difference cascades from that. But there are actually two variables doing the work here, and they’re usually conflated: sulfur content (sweet vs. sour) and density (light vs. heavy). They correlate in the real world — most sour crudes are also heavy (Arabian Heavy, Urals, Canadian dilbit, Venezuelan), most sweet crudes are lighter (WTI, Brent, Bonny Light, US shale) — so a “sour refinery” in practice is usually also a “heavy refinery,” and it’s worth keeping the two axes distinct because they drive different equipment.

Sour means high sulfur (roughly >0.5%, and Gulf grades run 1.5–4%+). Sulfur is a problem in two ways: it’s a contaminant you have to strip out to make spec-compliant fuel that you can sell, and it corrodes the plant itself. So a sour refinery is defined by four hydroprocessing and sulfur-handling systems that are not present in the same scale in a sweet refinery. I am talking about the following:

Hydrotreaters, and lots of hydrotreating capacity — units that combine the feed with hydrogen over a catalyst to convert sulfur into hydrogen sulfide (H₂S) so it can be pulled out. Every major cut (naphtha, kerosene, diesel, gasoil) needs its own treating.

A hydrogen source at scale. All that hydrotreating (and hydrocracking) consumes enormous volumes of hydrogen, so a sour/heavy refinery typically runs a dedicated steam methane reformer to manufacture hydrogen. A simple sweet refinery can often subsist on the hydrogen thrown off as a byproduct of its catalytic reformer.

Amine units + a Sulfur Recovery Unit (Claus process) + tail-gas treating. This is the back end that takes the H₂S you’ve stripped and converts it to elemental sulfur for sale. A sweet refinery may have a token SRU; a sour refinery’s sulfur block is a major, capital-intensive part of the plant. A big Gulf refinery can produce thousands of tons of sulfur a day as a side product.

Metallurgy. Because H₂S and organic acids attack carbon steel — via sulfidation and, where hydrogen is present, hydrogen-induced cracking and high-temperature hydrogen attack — sour plants are built with upgraded metallurgy: stainless and chrome-moly alloy cladding in the crude unit overhead, hydrotreater circuits, and hot sections. That alloy content is a big reason sour refineries cost more to build even before you count the extra units.

Are you with me so far? Let me explain why this asymmetry between sour crude refineries and sweet crude refineries matters strategically — The two plant types are not interchangeable, and the switching costs run mostly one direction.

A refinery configured and optimized for a specific sour/heavy diet (say, Urals, or Arabian Heavy, or Venezuelan/Canadian heavy) can’t freely swap to another crude without yield and economic penalties — its coker, hydrogen balance, and hydrotreater sizing were tuned to that feed. This is exactly why sanctioned-crude reroutings are sticky: Indian and Chinese refiners that took discounted Urals, or the Gulf/Chinese teapots’ appetite for heavy sour, reflect plants built for that assay, not spot opportunism.

A complex sour refinery can process sweet crude (it’s over-built for it, so it just runs its expensive units under-utilized — inefficient but feasible). A simple sweet refinery generally cannot process sour crude — it lacks the hydrotreating, hydrogen, sulfur recovery, and metallurgy to handle it at all. That one-way flexibility is why global sour/heavy conversion capacity is the real constraint, and why losing a specific complex refinery (or a specific heavy-crude supply it was built around) can’t be quickly back filled elsewhere.

So when you see the US or Iran strike a specific refinery, or a crude grade knocked off the market, the question that determines the real supply impact isn’t barrels-for-barrels — it’s which configuration was hit and whether the displaced crude has anywhere else to go that was actually built to digest it. A million barrels of Arabian Heavy stranded by a Hormuz disruption can’t be absorbed by simple sweet refineries sitting idle in some other region; it needs deep-conversion capacity specifically, and that capacity is concentrated, expensive, and slow to build.

The core asymmetry is that the crude coming out of the Gulf is overwhelmingly heavy and sour, and only four geographically concentrated set of refineries were ever built to digest it. These are:

Reliance’s Jamnagar (Gujarat, ~1.4 mbpd across two refineries) is the single most important node — it’s among the highest-complexity refineries on earth, purpose-built to run cheap heavy sour from anywhere and crack the resid into clean products.

China’s mega-refineries and Shandong teapots are the second great sink. The state-of-the-art complexes (Rongsheng/ZPC at Zhoushan, Hengli at Dalian) plus the independent teapot cluster in Shandong are configured for heavy sour and are the designated home for sanctioned barrels — Iranian Heavy, Venezuelan Merey, discounted Russian ESPO/Urals. This is the destination that makes Iran’s crude exports viable at all, and it’s Hormuz-and-Malacca exposed.

Northeast Asia — South Korea (SK, GS Caltex, S-Oil) and to a lesser extent Japan — runs high-complexity plants on a steady Gulf sour diet (historically heavy Saudi, Kuwaiti, and pre-sanctions Iranian). These are among the most Hormuz-dependent refiners in the world, with almost no domestic crude and limited ability to swap feed quickly.

The US Gulf Coast is the outlier that matters for the strategic picture — it’s enormous deep-conversion heavy-sour capacity (Valero, Motiva’s Port Arthur, ExxonMobil Beaumont, the coking belt), but it is not primarily Gulf-fed. It was rebuilt in the 2000s around Venezuelan and Mexican Maya, and now leans heavily on Canadian heavy via pipeline. So the USGC is heavy-sour hungry but largely insulated from a Hormuz event — which is exactly why it can’t easily backfill stranded Gulf barrels for Asia: it’s already running its coke drums full on Western Hemisphere heavy.

So the vulnerability isn’t measured in barrels off the water — it’s measured in coking and hydrocracking capacity that has lost its designed feed, and that capacity is concentrated in four places (Jamnagar, coastal China, Korea, and a USGC that’s already full on Canadian). That concentration is the thing worth watching: a sustained double-chokepoint event doesn’t ripple evenly across the refining world, it lands almost entirely on those Asian deep-conversion hubs and on the diesel market they anchor.

What I have outlined and discussed above just scratches the surface of the enormous economic stakes that are at risk if the Strait of Hormuz remains closed for an extended period of time. Besides the shortage of the sour crude and shortage of sulphur, there are also major deficits in urea, LNG and helium. It is conceivable that the economic damage from the US war on Iran will exceed any physical damage inflicted by missiles and bombs.

---

Here’s an interview I did with Joe Sansone three days ago. I just received the YouTube link:

I spoke with Glenn Diesen on Monday about the Odessa blockade by Russia and Urkaine’s attacks on Iran: