How does an oil refinery work, and how is it changing?

By the Vrulum team

Introduction

A petroleum refinery is the plant where crude oil is turned into products people can use. Crude oil is a mixture, and a refinery's job is to separate it and reshape the parts. The EBSCO refining overview lists gasoline (petrol), kerosene, and diesel and heating oils among the products that go on the market (EBSCO refining overview). An EBSCO research starter on cracking describes short-chain products as more commercially valuable than longer-chain ones.

The sources below trace how the process developed and what it has done to some emissions.

How it works

The first stage is distillation. EBSCO's refining overview defines distillation as "the separation of components based on boiling point". The crude is split into boiling ranges, called cuts. Material that does not distil is called residuum. That heavy residue can be treated to separate lubricating oils and waxes (refining overview).

The second stage changes the molecules. Cracking "breaks down long-chain hydrocarbon molecules" into shorter ones, as the cracking page explains. Catalytic reforming uses reforming catalysts, which usually include a metal such as nickel or platinum. Hydrotreating uses hydrogen to treat fuels, and hydrodesulfurization, which removes sulphur, is one of the processes it covers (refining overview).

Petrol quality is judged by octane number, which describes how well a fuel resists knocking in an engine. The refining overview says most regular grade gasolines have octane numbers of 87, and gives iso-octane a value of 100.

What has been demonstrated

These are historical and process facts that the sources report directly.

Samuel Kier sent a sample of oil to Professor James Curtis Booth, who suggested distilling it for lighting. Kier then built a cast-iron, one-barrel still and began selling kerosene (American Chemical Society release, via EurekAlert). Edwin Drake's well at Titusville struck oil on 27 August 1859 at a depth of 69 and a half feet (Drake Well Museum).

Cracking was developed in stages. The EBSCO Burton article dates the first battery of twelve stills used in the thermal cracking of petroleum products to January 1913. The Burton process was a batch process, but by the 1920s continuous thermal cracking processes were developed (EBSCO on Burton). Eugene Houdry had developed a method known as cracking by 1925, and by 1936 his first plant at Paulsboro, New Jersey, was producing over 2,000 barrels per day of high-octane gasoline (EBSCO on Houdry).

Catalytic cracking can give gasolines with octane numbers above 90, and the same source says it raises the gasoline yield of a refinery to about 45 per cent. The source contrasts this with straight-run gasoline yields of not more than 20 per cent from very good quality crude and possibly less than 10 per cent from poorer crudes, but it does not state the crude type or date behind the 45 per cent figure (refining overview).

On emissions, the EPA enforcement results page, last updated on 20 October 2026, reports a significant and steady decline in sulphur dioxide and nitrogen oxide emissions from the refining sector since the first settlement in 2000. It gives a figure of approximately 75 per cent less than pre-Refinery Initiative levels. The page text does not state the baseline year for that figure.

What is only reported

Some claims come from a single secondary source and have not been checked against primary data. An EBSCO article on cracking states that Burton's thermal cracking "doubled gasoline production at American refineries". The page ties the statement to Burton's 1913 patent but gives no data behind it, so it is treated here as unverified. The EBSCO Burton article does not itself state that gasoline production doubled.

The sources used for this article contain no company claims about lower-emission refining, so none are reported here. Any such claim would need its baseline, method and measurements stated before it could be weighed.

Limits and open problems

The emissions evidence here is narrow. The EPA page decline statement covers sulphur dioxide and nitrogen oxides, and the page does not mention greenhouse gases. The sources reviewed give no verified greenhouse gas figures for refining, so this article does not assess its climate impact.

No source reviewed here shows a refinery running on non-fossil feedstock at scale. The sources also do not quantify the cost of changing plants, the pace at which the industry could adapt, or how fuel demand may shift. Any answer to those questions would need its own evidence.

What to watch next

Watch for updates to the EPA emissions page, which carries a date but describes its baseline only as pre-Refinery Initiative levels. Watch for company claims about lower-emission refining that give baselines and measured results. Watch for any refinery that reports measured output from waste or bio-based feedstock at a commercial scale.