# From crude oil and natural gas to the C1-C4 building blocks: Steam cracking: the engine of the petrochemical industry

Source page: https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p2/

[Understanding Chemicals](https://www.materialinsights.org/learn/chemicals/) / [Module 5: From crude oil and natural gas to the C1-C4 building blocks](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/) Part 2 of 5

## Steam cracking: the engine of the petrochemical industry

Steam cracking makes almost all the world's ethylene and much of its propylene and butadiene. It uses no catalyst, only heat, speed and a very cold separation train. This part follows the gas through the furnace, the quench and the cold end. It then shows how the choice of feed, ethane or naphtha, changes everything that comes out and what each tonne of ethylene really costs.

### 2.1 The hot end: a few hundredths of a second at 800 C

A steam cracker breaks saturated hydrocarbons into smaller, unsaturated ones. Wittcoff describes the furnace. The feed is vaporised if it is a liquid, mixed with steam, and sent through tubes about an inch across that run through a fired furnace. Conditions are a high temperature (650-900 C) and a low hydrocarbon pressure. The residence time is very short, 30-100 milliseconds. Shreve (1984) gives about 925 C for the same step.

Each choice has a reason. Breaking carbon-carbon bonds needs heat, so the reaction is strongly endothermic. Low pressure favours the side of the reaction with more molecules, because one large molecule becomes several small ones. Running below atmospheric pressure would risk air leaking in and forming explosive mixtures. So steam is added as an inert diluent to lower the hydrocarbon's partial pressure instead. The steam also reduces coke on the tube walls. Time is kept short because coke and hydrogen are the thermodynamically favoured end products. Left longer, the valuable olefins would break down further.

The chemistry is a chain reaction of free radicals, which are fragments with an unpaired electron. With ethane, a C-C bond splits into two methyl radicals. These pull hydrogen atoms off other ethane molecules. The resulting ethyl radicals shed a hydrogen atom to give ethylene, and the chain continues. Propane can also lose a hydrogen to form an isopropyl radical, which can only become propylene. That is why propane gives far more propylene than ethane. Long naphtha chains break readily, and the favoured step (beta-scission) keeps peeling off ethylene units. Gas leaves the furnace at about 800 C.

Process flowchart Naphtha steam cracker: furnace to finished products Inputs **Naphtha**about 3.2 t per t ethylene **Dilution steam**mixed with feed before the furnace **Furnace fuel**largely the cracker's own methane and hydrogen Process steps 1 **Cracking furnace**

1-inch tubes, 650-900 C, 30-100 ms, low hydrocarbon partial pressure; radical chain reactions
 2 **Transfer-line quench exchanger**

cools gas from about 800 C to about 400 C and raises steam for the compressor turbines
 3 **Oil quench tower**

spray of recycled fuel oil cools to a little above 100 C; fuel oil drawn off
 4 **Water quench and decanter**

cooling to ambient condenses pyrolysis gasoline and the dilution steam
 5 **Acid gas removal and drying**

ethanolamine scrub removes CO2 and H2S; molecular sieves remove water so ice cannot form
 6 **Compression and chilling**

centrifugal compressors to about 40 bar; refrigeration cascade condenses the gas
 7 **Demethaniser**

overhead condenser at about -95 C using liquid ethylene; H2 and CH4 off the top
 8 **Deethaniser, acetylene hydrogenation, C2 splitter**

acetylene selectively hydrogenated; ethylene overhead, ethane recycled to furnace
 9 **Depropaniser, C3 splitter, debutaniser**

propylene (propyne removed); crude C4 stream; C5+ joins pyrolysis gasoline
 Outputs **Ethylene**1 t **Propylene**about 0.5 t **Butadiene**about 0.15 t **Other C4s**about 0.25 t **Pyrolysis gasoline**about 0.75 t (incl. 0.15 t benzene) **Fuel oil**about 0.1 t **Hydrogen and methane**about 0.45 t, mostly burned as furnace fuel

Quantities per tonne of ethylene from Wittcoff's Table 2.7 (naphtha column, per 100 lb ethylene scaled to 1 t).
 Source: Wittcoff (2004), Ch 2 section 2.2.1 and Table 2.7, pdf pp. 93-96  Worked example

#### Why yields jumped

Wittcoff reports that over the quarter-century before 2004 the ethylene yield from fairly severe naphtha cracking roughly doubled, from about 16% to 31% by weight. The gain came from hotter, shorter cracking. Ethylene is the premium product, so a few points of yield are worth a great deal on a plant making hundreds of thousands of tonnes a year. The same feed simply yields more of the most valuable product.

- Wittcoff (2004), Ch 2 sections 2.2.1-2.2.1.3 · pdf pp. 93-96, 102-103
- Shreve (1984), Ch 37 Petroleum Processing · pdf pp. 743-744

### 2.2 The cold end: separating a cracked gas costs more than cracking it

What leaves the furnace is a hot, mixed gas: hydrogen, methane, ethylene, propylene, butadiene, a little acetylene and a liquid tail of aromatics. Turning it into polymer-grade products is, in Wittcoff's words, a classic chemical engineering problem. The gas must be cooled by hundreds of degrees with as little wasted heat as possible. Then it is chilled far below zero for distillation.

The flow figure above shows the sequence. The first quench exchanger cools the gas from about 800 C to about 400 C and raises high-pressure steam. That steam drives the turbines of the compressors further down the line, so the furnace heat is recycled. A spray tower using fuel oil then cools the gas to just above 100 C. Further cooling condenses pyrolysis gasoline and the dilution steam, which separate in a decanter like oil and water. Next, an amine scrub (ethanolamines) removes carbon dioxide and hydrogen sulphide. Molecular sieves dry the gas, because ice would scour the cold equipment.

The dry gas is compressed to about 40 bar and condensed in a cascade of refrigeration units. Engineers often call this chilling section the cold box. The first column, the demethaniser, runs with a condenser at about -95 C cooled by liquid ethylene. Hydrogen and methane leave overhead. Later columns split C2 from C3 and C4. Acetylene is hydrogenated at low temperature, which attacks the triple bond but not the double bond. Left in, acetylene would put cross-links into polyethylene. Unreacted ethane is recycled to the furnace.

Wittcoff argues that cracking only became economic once centrifugal compressors could run these refrigeration duties cheaply. He calls them mechanical engineering's greatest contribution to the chemical industry.

Data What a high-severity naphtha crack produces before separation

| Component | wt% of cracked product |
|---|---|
| Hydrogen | 1.2 |
| Methane | 15.2 |
| Acetylene | 1.3 |
| Ethylene | 31.8 |
| Ethane | 1.2 |
| Propadiene | 1.2 |
| Propylene | 11.6 |
| Propane | 0.3 |
| Butadiene | 4.7 |
| Butylenes | 2.2 |
| C5+ liquids | 27.7 |

A 1970s high-severity, short-time crack quoted by Shreve. Every component on this list needs its own column or reactor.
 Source: Shreve (1984), Ch 37, pdf p. 745  Common misreading

#### Steam cracking is not a refinery reaction

Wittcoff notes that, strictly, steam cracking is not refining. Crackers are usually built close to refineries because they draw feed from them. For gasoline, thermal cracking was replaced long ago by catalytic cracking, which gives branched molecules with better octane. Only about 9% of US cracking capacity is steam cracking. Yet that 9% is the base of the petrochemical industry.

- Wittcoff (2004), Ch 2 section 2.2.1 · pdf pp. 93, 95, 123
- Shreve (1984), Ch 37 Petroleum Processing · pdf pp. 743-745

### 2.3 The feedstock decides the co-product slate, and the slate decides the economics

The same furnace gives very different products depending on what goes in. Wittcoff's table below shows that ethylene yield falls and co-products rise as the feed gets heavier. Ethane gives 80% ethylene. Propane gives 42%, naphtha 31% and vacuum gas oil 23%. Each tonne of ethylene needs about 1.2 t of ethane but 3.2 t of naphtha. The extra 2 t from naphtha comes out as propylene, C4s, pyrolysis gasoline and fuel. Brennan (2020) gives the same picture with slightly different numbers. In his figures, 125 t of ethane makes 100 t of ethylene and about 3 t of propylene. It takes 330 t of naphtha to make 100 t of ethylene plus about 52 t of propylene, 14 t of butadiene and 22 t of benzene.

The IEA (2018) puts it in terms of all high-value chemicals. An ethane cracker yields about 50% more HVCs per tonne of feed than a naphtha cracker, but almost all of it is ethylene. Naphtha gives a balanced mix of ethylene, propylene and aromatics. The heavier the feed, the more the product resembles the feed's structure. That is the trade-off between yield and product diversity.

Severity, meaning how hot and how fast the crack is run, is a second lever. Wittcoff's Table 4.1 shows the propylene-to-ethylene weight ratio for light naphtha. It rises from 0.43 at high severity to 0.57 at low severity. Gas oil at moderate severity gives 0.63. Ethane gives only 0.019, so little that its propylene is often not recovered.

This is why feedstock choice is an economic decision about co-products as much as about feed cost. A naphtha cracker is a multi-product plant. Its profit depends on selling propylene, butadiene and aromatics, not just ethylene.

Data Steam cracker yields by feedstock (lb per 100 lb ethylene)

|  | Ethane | Propane | Butane | Naphtha | Atm. gas oil | Vacuum gas oil |
|---|---|---|---|---|---|---|
| Feed required | 120 | 240 | 250 | 320 | 380 | 430 |
| Ethylene yield,% of feed | 80 | 42 | 38 | 31 | 26 | 23 |
| Propylene | 3 | 50 | 40 | 50 | 55 | 60 |
| Butadiene | 2 | 5 | 9 | 15 | 17 | 17 |
| Other C4 olefins | 1 | 3 | 17 | 25 | 18 | 18 |
| Pyrolysis gasoline | 2 | 15 | 18 | 75 | 70 | 65 |
| of which benzene | 1 | 5 | 6 | 15 | 23 | 24 |
| of which toluene | neg. | neg. | 3 | 10 | 12 | 12 |
| Fuel oil | neg. | 2 | 4 | 10 | 70 | 125 |
| Other (mainly H2 and CH4) | 17 | 65 | 62 | 45 | 50 | 45 |

Read each column as the co-products made alongside 100 lb (or 100 t) of ethylene. Figures are circa 2000-2004.
 Source: Wittcoff (2004), Ch 2, Table 2.7, pdf p. 96  Worked example

#### Worked example: net feed cost per tonne of ethylene, Europe 2002

Using Wittcoff's 2002 European prices (contract, or spot where no contract is given): naphtha costs 3.2 t x $223 = $714 per t ethylene. Credits: propylene 0.5 x $419 = $210; butadiene 0.15 x $407 = $61; pyrolysis gasoline 0.75 x $241 = $181; fuel oil 0.1 x $131 = $13; other C4s 0.25 t valued at butane's $207 = $52. Net feed cost is about $197. For propane: 2.4 x $240 = $576, less about $275 of credits, gives about $301. So naphtha was cheaper that year, before fuel gas credits and the naphtha plant's higher capital.

- Wittcoff (2004), Ch 2 section 2.2.1.1 and Ch 4 Table 4.1 · pdf pp. 96, 193-194
- Brennan (2020), Ch 2 Case C Ethylene · pdf p. 55
- IEA (2018), Ch 2 · pdf pp. 31-32

### See it in the table
[H₂O Water](https://www.materialinsights.org/commodities/water/)[H2 Hydrogen](https://www.materialinsights.org/chemicals/hydrogen/)[CO2 Carbon dioxide](https://www.materialinsights.org/chemicals/carbon_dioxide/)[C2H4 Ethylene](https://www.materialinsights.org/chemicals/ethylene/)[C3H6 Propylene](https://www.materialinsights.org/chemicals/propylene/)[BD Butadiene](https://www.materialinsights.org/chemicals/butadiene/)[Bz Benzene](https://www.materialinsights.org/chemicals/benzene/)

### Check your understanding

Why is steam added to the feed of a steam cracker?
 Low pressure favours cracking. Running below atmospheric pressure would risk air leaking in and forming an explosive mixture, so steam acts as an inert diluent. Wittcoff notes it also reduces coke formation.

Per tonne of ethylene, a naphtha cracker needs about 3.2 t of feed and an ethane cracker about 1.2 t. What happens to the extra naphtha?
 Wittcoff's Table 2.7 shows the naphtha balance per 100 lb ethylene: 50 propylene, 15 butadiene, 25 other C4s, 75 pyrolysis gasoline, 10 fuel oil and 45 hydrogen plus methane. Only part of this is burned.

A naphtha cracker operator wants more propylene relative to ethylene. What should they do?
 Wittcoff's Table 4.1 gives the light-naphtha propylene/ethylene ratio as 0.43 at high severity and 0.57 at low severity. Ethane gives almost no propylene.

[← Part 1: The raw materials: crude oil, natural gas and the refinery](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p1/)[Mark read · Part 3: Propylene, the C4 stream and the aromatics →](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p3/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/index.md)
- [The raw materials: crude oil, natural gas and the refinery](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p1/index.md)
- [Propylene, the C4 stream and the aromatics](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p3/index.md)
- [C1 chemistry: synthesis gas, methanol, ammonia and coal](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p4/index.md)
- [The economics: feedstock, scale, integration and carbon](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/p5/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/chemicals/oil-gas-to-building-blocks/review/index.md)