# From building blocks to plastics, fibres and detergents: How polymers are made

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

[Understanding Chemicals](https://www.materialinsights.org/learn/chemicals/) / [Module 6: From building blocks to plastics, fibres and detergents](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/) Part 2 of 5

## How polymers are made

A polymer is a long chain built from many small molecules called monomers. How the links form, what catalyst guides them, and how the reactor gets rid of the heat decide what chain comes out: its length, its branches and how regularly it is built. Those three features then decide whether the product is film, bottle, fibre or rubber.

### 2.1 Chains grow either one monomer at a time or by joining fragments

There are two ways to build a polymer. In chain growth, the older name for which is addition polymerisation, an initiator turns one monomer into a reactive species, usually a free radical (a fragment with an unpaired electron). That species adds monomers one after another in a fraction of a second until something stops it. The three stages are initiation, propagation and termination.

Chain length depends on the race between propagation and termination. If many radicals are started at once, they meet and kill each other early, giving short chains; long chains need slow initiation and fast propagation, thousands of times faster than termination. A growing radical can also steal a hydrogen from the middle of a finished chain, which then grows a branch. In high-pressure polyethylene, the radical end bends back into a ring-like loop and moves its radical to a carbon a few atoms down the chain; growth restarts there, leaving a four-carbon branch, which is why that polymer is full of short butyl branches.

In step growth, or condensation, molecules with reactive groups at each end join repeatedly, often giving off water. A glycol and a dibasic acid make a polyester; a diamine and a diacid make a nylon. The monomer disappears early into short pieces called oligomers, which then join each other, so chain length keeps rising the longer the reaction runs.

Functionality, the number of links each monomer can form, sets the overall shape. Two-ended monomers give linear chains that melt and can be reshaped: thermoplastics. Monomers with three or more links give cross-linked networks, thermosets, that never melt. Strength rises quickly between about 50 and 500 repeat units and changes little beyond that.

Data Chain growth and step growth compared

| Feature | Chain growth (addition) | Step growth (condensation) |
|---|---|---|
| Typical monomers | Ethylene, propylene, vinyl chloride, styrene, butadiene | Terephthalic acid and glycol, adipic acid and diamine, caprolactam |
| What drives it | Radical, ion or metal catalyst site adds monomer after monomer | End groups react with each other, often releasing water |
| Intermediates | Short-lived, cannot be isolated | Oligomers that can be isolated |
| Effect of time | Raises yield; chain length set early | Raises chain length |
| Main control | Initiator rate, temperature, chain-transfer agents | Water removal, chain stoppers, purity of 1:1 balance |
| Products in this module | PE, PP, PVC, PS, SBR, acrylic | PET, nylon 66, nylon 6 |

 Source: Wittcoff (2004), Ch 15 How polymers are made, pdf pp. 490-499  Common misreading

#### Longer reaction, longer chains?

Only for step growth. In chain growth each chain is finished within a fraction of a second, so leaving the reactor running longer raises the yield, not the chain length. Length is set by the balance of initiation, propagation and termination, and by chain-transfer agents added on purpose, such as hydrogen in Phillips polyethylene. Step growth is the opposite. Its chains keep lengthening with time, so nylon makers add a little acetic acid to cap the chain ends and stop the melt becoming too viscous to spin.

- Wittcoff (2004), Ch 15 How polymers are made · pdf pp. 486, 490-500, 538
- Kent (2012), Ch 12 Manufactured textile fibers · pdf p. 447

### 2.2 Catalysts decide what shape the chain takes

Until the 1950s ethylene could only be polymerised by free radicals at around 2,000 bar and 200 C, giving branched low-density polyethylene. Then, in 1954, Karl Ziegler announced that titanium tetrachloride combined with an aluminium alkyl made linear, crystalline polyethylene at about atmospheric pressure and 60 C. A titanium atom with one empty position grabs a monomer and inserts it into the growing chain, then the vacancy reappears for the next one.

Giulio Natta tried the catalyst on propylene. Free radicals had turned propylene into a useless, low-molecular-weight tangle, but the Ziegler catalyst gave long chains in which every methyl group pointed the same way. Such isotactic polymer packs into crystals and is stiff and high-melting; random, atactic polymer is a soft gum. Phillips developed a rival, chromium oxide on silica, which gives very long chains that are tamed with hydrogen as a chain stopper and about 1% butene; it works at around 100 C and 40 bar. By 2004 Ziegler-type catalysts made about 60% of HDPE in the United States and Western Europe.

Catalyst productivity has changed the economics of whole plants. The first polypropylene catalysts gave 1,000-2,000 lb of polymer per lb of catalyst. Modern ones give up to 75,000 lb, with 97% isotactic content, so neither the catalyst residue nor the atactic fraction has to be removed.

Metallocenes, commercialised by Exxon in 1991 and Dow in 1993, are single-site catalysts: every active site is identical, unlike the many different sites on a Ziegler-Natta solid. They give a very narrow spread of chain lengths and spread comonomer evenly, which means clearer, tougher films with fewer of the short chains that add taste and odour. The trade-off is poorer melt processing and an expensive methylaluminoxane co-catalyst, so in 2000 metallocene polypropylene had under 1% of its market.

Data Catalyst families for polyolefins

| Family | Typical make-up | Conditions | What it gives | Limits |
|---|---|---|---|---|
| Free radical (oxygen, peroxide) | Initiator only | 1,300-2,600 bar, 200-300 C | Branched LDPE | Very high pressure; cannot make useful PP |
| Ziegler-Natta | TiCl4 or TiCl3 with aluminium alkyl, often on MgCl2 | 40-80 C (PP); low pressure | Linear HDPE, LLDPE, isotactic PP | Many different sites; broad chain-length spread |
| Phillips metal oxide | Chromium oxide on silica or alumina | About 100 C, 40 bar, hydrogen to limit chain length | HDPE | Cannot make polypropylene |
| Metallocene, single site | Zirconocene with methylaluminoxane | Similar to Ziegler-Natta | Narrow distribution, even comonomer, syndiotactic PP | Costly co-catalyst; harder to process |

 Source: Wittcoff (2004), Ch 3, Ch 4, Ch 15, pdf pp. 133, 199-200, 515-526  Case

#### The plant a catalyst removed

Kent describes the first polypropylene plants: catalyst slurried in hexane, the isotactic polymer settling out while the atactic part stayed dissolved. The catalyst then had to be killed and dissolved with hydrochloric acid and alcohol, the polymer separated by centrifuge or filter, and the atactic fraction recovered by evaporating solvent. Wittcoff's comparison of old and new flowsheets shows modern high-yield, high-isotactic catalysts deleting the de-ashing and atactic-removal sections altogether. Fewer vessels mean lower capital, less solvent and less energy per tonne.

- Wittcoff (2004), Ch 3 Ethylene polymers · pdf pp. 131-133
- Wittcoff (2004), Ch 4 Propylene polymers · pdf pp. 199-200
- Wittcoff (2004), Ch 15 Metal complex and metallocene catalysts · pdf pp. 515-526
- Kent (2012), Ch 12 Polypropylene fibre synthesis · pdf p. 462

### 2.3 Reactor design is mostly about getting the heat out

Polymerisation releases a lot of heat, and as chains grow the mix turns viscous, so heat cannot easily reach a cooling surface. Each industrial process is largely an answer to that problem. Wittcoff lists four basic procedures: bulk (monomer and initiator alone), solution (a solvent soaks up heat and thins the mix), suspension (monomer droplets stirred in water) and emulsion (monomer held in soap micelles in water).

High-pressure LDPE is a bulk process run in a tubular reactor or an autoclave at 1,300-2,600 bar and 200-300 C. Only about 30-35% of the ethylene reacts on each pass; going further would add too many branches. The rest is separated at high pressure and recycled, so the plant is built around high-pressure compression and recycle.

The gas-phase fluidised bed, first Union Carbide's Unipol process, has been licensed extensively for HDPE and LLDPE, and Wittcoff calls it the most important process in use. Small polymer particles are kept floating by a rising stream of ethylene and comonomer at about 85-105 C and 20 bar. Catalyst is sprayed in, polymer grows on the particles, and the circulating gas carries away the heat. Only 2-3% of the gas reacts per pass, so it is cooled and recompressed. Particles grow over three to five hours, then leave through a gas lock as dry powder.

Slurry processes disperse catalyst and growing polymer in a light hydrocarbon such as hexane. For polypropylene the slurry can be in liquid propylene itself, pumped round a loop of large-diameter pipe; such bulk plants were over half of all PP plants in 2000, and gas-phase plants rose from about 15% in 1990 to almost 30%. Solution processes dissolve the polymer; they fell out of use for polypropylene because of solvent losses and the difficulty of removing the last traces.

Process flowchart Gas-phase fluidised-bed polyethylene (Unipol type) Inputs **Polymer-grade ethylene**about 1 t per t of polymer **Comonomer (1-butene or 1-hexene)**about 1% for HDPE; 6-8% for LLDPE **Ziegler, chromium or metallocene catalyst**sprayed in continuously **Hydrogen**small; controls chain length Process steps 1 **Feed purification**

Remove water, oxygen and other catalyst poisons
 2 **Fluidised-bed reactor**

85-105 C, about 20 bar; gas keeps polymer particles floating and removes heat
 3 **Cycle gas loop**

Only 2-3% of gas reacts per pass; unreacted gas is compressed, cooled and returned
 4 **Product discharge**

Granules grow over 3-5 h to 0.5-1 mm, leave through a gas lock and are purged with nitrogen
 5 **Compounding extruder**

Melt with stabilisers and additives, cut into pellets; energy-intensive
 Outputs **HDPE or LLDPE pellets**1 t **Purge gas**small; recovered or flared

Wittcoff gives 85-105 C, 20 bar and 500-micron particles; Shreve gives 100 C, 0.7-2 MPa and about 1,000 microns.
 Source: Wittcoff (2004), Ch 3; Shreve (1984), Ch 34, pdf pp. 133-134; 674-675  In the data

#### Why the low-pressure route won

LDPE needs 1,300 bar and more. HDPE and LLDPE can be made as low as 60 C and 1 bar, though commercial plants run at 130-270 C and 10-160 bar with conversion per pass approaching 100% in some designs. LLDPE copies LDPE's softness by adding 6-8% butene to a low-pressure copolymer. Its energy advantage is partly eaten by the cost of the comonomer, but it ran on HDPE equipment and made stronger film, so it grew largely at the expense of LDPE.

- Wittcoff (2004), Ch 3 Ethylene polymers · pdf pp. 132-134
- Wittcoff (2004), Ch 4 Propylene polymers · pdf p. 200
- Wittcoff (2004), Ch 15 Polymerisation procedures · pdf pp. 504-506
- Shreve (1984), Ch 34 Plastic industries · pdf pp. 674-676

### 2.4 Grades are set by chain length, branching and regularity

Producers sell dozens of grades of each polymer, and three molecular features do most of the tuning.

The first is chain length. Because chains in a sample vary, it is quoted as averages: the number average and the weight average, which leans towards the heavier chains. Their ratio measures the spread. A broad spread helps processing, since short chains act as a built-in lubricant in the melt; a narrow one, as from metallocenes, gives more uniform properties but harder processing.

The second is branching. Branches stop chains packing side by side, so they lower crystallinity, density, melting point and stiffness. Wittcoff gives HDPE 85-95% crystallinity and LDPE about 55%, with densities ranging from 0.965 down to 0.915 g/cm3. LLDPE reaches LDPE's density with short, regular branches from 6-8% butene, hexene or octene, against about 1% in HDPE; regular branches let the crystalline parts pack better, so its film is stronger and can be thinner.

The third is regularity of structure, such as tacticity in polypropylene or the cis or trans form of polybutadiene. Regular chains crystallise; irregular ones stay amorphous or rubbery.

Two temperatures follow from these features. The glass transition temperature, Tg, is where amorphous regions turn from glassy to rubbery; the crystalline melting point, Tm, is where crystals melt. A plastic used above its Tg but below its Tm is tough; below Tg it is brittle. Additives then finish the grade: Kent notes that HDPE for detergent bottles is made tougher against stress cracking by raising molecular weight or adding 1-2% butene, and that flexible PVC carries 20-80 parts of plasticiser per hundred of resin.

Data Typical glass transition and melting temperatures

| Polymer | Tg (C) | Tm (C) | What it means in use |
|---|---|---|---|
| cis-Polybutadiene | -101 | 4 | Rubbery at all normal temperatures: a tyre rubber |
| Linear polyethylene | -70 to -20 | 132 | Tough at room temperature; melts near 132 C |
| Polypropylene | -16 | 170 | Higher service temperature than PE; brittle in the cold unless copolymerised |
| Nylon 66 | 47 | 235 | Strong fibre and engineering plastic |
| PVC | 70 | 140 | Rigid unless plasticised |
| Polystyrene | 94 | 227 | Glassy and brittle at room temperature |
| PET | about 80 | not tabulated; melt handled at 260-300 C | Must be drawn above 80 C to orient |

PET is not in Wittcoff's table: Tg from Kent, melt temperature from Shreve's polyester flowchart.
 Source: Wittcoff (2004), Ch 15, Table 15.7; Kent (2012), Ch 12; Shreve (1984), Ch 35, pdf p. 543; pdf p. 452; pdf p. 692  Worked example

#### Reading three polyethylene grades

Kent's typical values show how much branching changes. HDPE: density 0.96, melting point 130 C, modulus 1,400 MPa, tensile strength 26 MPa. LLDPE: 0.93, 124 C, 450 MPa, 14 MPa. LDPE: 0.92, 108 C, 180 MPa, 12 MPa. A density difference of only 0.04 g/cm3 comes with an eightfold drop in stiffness. That is why HDPE makes milk bottles and pipe, where stiffness lets walls be thinner, while LDPE and LLDPE make film and bags.

- Wittcoff (2004), Ch 3 Ethylene polymers, Table 3.2 · pdf pp. 132-134
- Wittcoff (2004), Ch 15 Molecular weight; Polymer properties · pdf pp. 503-504, 538-547
- Kent (2012), Ch 15 Synthetic resins and plastics, Tables 15.2, 15.5 · pdf pp. 596, 598

### See it in the table
[H₂O Water](https://www.materialinsights.org/commodities/water/)[Al Aluminium](https://www.materialinsights.org/commodities/aluminum/)[Cr Chromium](https://www.materialinsights.org/commodities/chromium/)[Ti Titanium](https://www.materialinsights.org/commodities/titanium/)[HCl Hydrochloric acid](https://www.materialinsights.org/chemicals/hydrochloric_acid/)[H2 Hydrogen](https://www.materialinsights.org/chemicals/hydrogen/)[C2H4 Ethylene](https://www.materialinsights.org/chemicals/ethylene/)[C3H6 Propylene](https://www.materialinsights.org/chemicals/propylene/)[AcOH Acetic acid](https://www.materialinsights.org/chemicals/acetic_acid/)[PP Polypropylene](https://www.materialinsights.org/chemicals/polypropylene/)[PVC PVC](https://www.materialinsights.org/chemicals/pvc/)

### Check your understanding

In free-radical polymerisation, what happens if the initiation rate is set very high?
 Many radicals at once means frequent collisions and early termination, so molecular weight falls. Long chains need slow initiation and fast propagation.

Why does the gas-phase polyethylene reactor convert only 2-3% of the ethylene per pass?
 The circulating gas is both the fluidising medium and the coolant. A big flow with small conversion per pass keeps the bed at a steady 85-105 C.

LLDPE and LDPE can have the same density. What gives LLDPE its higher film strength?
 Comonomers such as butene give identical short branches. LDPE's branches are irregular, with branches on branches, so its crystals are less well ordered.

[← Part 1: Building blocks and their family trees](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/p1/)[Mark read · Part 3: The big bulk plastics →](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/p3/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/index.md)
- [Building blocks and their family trees](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/p1/index.md)
- [The big bulk plastics](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/p3/index.md)
- [Synthetic fibres](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/p4/index.md)
- [Detergents: from benzene, paraffins and ethylene oxide](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/p5/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/chemicals/building-blocks-to-polymers/review/index.md)