How a classic specialty chemical market works: the synthetic dye industry: How firms competed: laboratories, chemists, salesmen and cartels

4.1 Making an azo dye: a few reactions, millions of combinations

Kent describes the toolkit that turns coal-tar aromatics (benzene, naphthalene and anthracene) into intermediates. Nitration adds a nitro group using "mixed acid", nitric plus sulphuric; toluene, for example, is nitrated near 20 C. Reduction turns the nitro group into an amino group; nitrobenzene to aniline is done commercially by high-temperature hydrogenation, or with iron and acid. Sulphonation adds sulphonic acid groups that make a dye soluble in water and give it a grip on fibres, using sulphuric acid or, for a second group, oleum. Alkali fusion swaps a sulphonic group for a hydroxyl group, giving naphthols. Most of these steps run in water, in batches.

Azo dyes are then made in two steps. Diazotisation: an aromatic amine such as aniline reacts with nitrous acid, generated in the vessel from hydrochloric acid and sodium nitrite, at 0 to 10 C because the product, a diazonium salt, decomposes when warm. The theory needs two moles of acid per mole of amine; plants use 2.2 to 2.5. Coupling: the diazonium salt is added to a coupler. Phenols and naphthols couple at pH 8 to 9, amines at pH 5 to 6. The product has an N=N azo link joining two aromatic rings, which gives colour.

The power of the method is combinatorial. Any of hundreds of amines can be paired with any of hundreds of couplers, once, twice or more, and a diamine can be coupled at both ends. Peter Griess found the reaction in 1858, but its full use came only after chemists understood the benzene ring in the 1870s. By 1912, 462 of 921 dyes on sale were azo dyes. Caro called the search an "endless combinatorial game"; Bayer's Carl Duisberg said there was no trace of a flash of genius in it.

• Kent (2012), Ch 13 Dye Application, Manufacture of Dye Intermediates and Dyes · pdf pp. 516-530
• Pollak (2011) · pdf p. 153
• Murmann (2003) · pdf pp. 275-276, 167

4.2 The research laboratory: a factory for inventions

Before about 1877 no firm had a research strategy. New dyes came from university chemists, dyers and tinkerers, and firms scrambled to buy or copy them. The German patent law of 1877 and the new structural chemistry changed that. Between 1877 (BASF) and 1886 (Bayer), the seven largest German dye firms set up laboratories devoted only to research.

Bayer came late. In 1883 Carl Rumpf could not persuade his partners to pay chemists just to invent, so he hired three himself and sent each to a university laboratory for six months. One was Carl Duisberg, whose first lab was a room behind the plant director's bathroom. In 1885 Bayer paid no dividend; Duisberg's patented Benzopurpurin 4B then restored its profits and convinced the board. In 1889 Duisberg built a central laboratory costing half a million marks, with twelve chemists sharing open rooms to encourage talk.

The lab became a funnel. In 1906 Bayer's chemists made 2,656 new compounds; 60 passed first screening and went to larger trials; 36 dyes reached the market. Hoechst once tested 3,500 colours to launch 18. After a new blue faded badly in sunlight, Bayer set up a testing department in 1887 so that no dye reached customers untested, and dyeing and printing labs where chemists worked beside trained colourists.

Patents were run like a weapon. Bayer created a patent department in 1896 and held 8,000 patents by 1913. "Patent laboratories" tried to find other routes to rivals' dyes. Firms filed in every country and in every corner of chemistry that might yield dyes, which Murmann calls saturation patenting, so that any newcomer would need a cross-licence. The cost was spread over a growing output: research cost per unit falls as sales rise, which favoured whoever grew fastest.

• Murmann (2003) · pdf pp. 50-51, 115-116, 155-158, 161, 164, 174-182

4.3 Selling application service around the world

German success is often put down to helping dyers use the new dyes on their own fabrics, and Murmann agrees, with a caveat: Perkin did the same. The difference was organisation and scale.

Bayer had the dyer-customer link built in from the start, as a partnership between a natural-dye dealer and a working dyer. Its partners visited customers in person. From the late 1870s it began replacing commission agents, who sold several makers' dyes and kept market news to themselves, with its own employees selling only Bayer dyes. Steady orders were essential to keep shared production lines full. In 1892 it sent out six travelling technicians to help dyers apply its dyes; by 1894 there were sixteen, hired locally wherever possible because locals understood local mills. Around 1900 it began year-long free courses at its works for the sons of mill owners, on condition they attended no rival's course for a year. Trainees learned to dye with Bayer products and told Bayer which problems its chemists should solve next.

Bayer translated its catalogues and printed labels to suit each market. Customers served each year rose from 10,000 in 1890 to 25,000 in 1902. By 1913 Bayer had 44 sales subsidiaries and 123 agents, from Manchester to Bombay, Shanghai and Kobe, each also feeding back intelligence about local tastes. BASF and Hoechst built similar machines.

It was not all clean. Bribing the foreman who chose a mill's dyes was common, and Bayer's kickbacks in the United States rose from 10% of receipts ($250,000) in 1902 to 18% ($700,000) in 1912. Britain's Levinstein and BS&S were too small to afford application departments on this scale.

• Murmann (2003) · pdf pp. 50, 134, 136, 140, 152-153, 171-173

4.4 Integration, scope and the taming of competition

Every firm had to decide how far back up the chain to make its own inputs. Founders' backgrounds shaped the choice. BASF's founders came from coal tar and built an integrated works making acids, soda, aniline and dyes. Bayer and Jager came from dye trading and at first made only dyes, buying aniline. SM&N became so good at nitrobenzene that even Perkin bought from it. After the 1880s independent aniline makers largely vanished, through integration, mergers and an aniline cartel in 1883. Bayer stopped making fuchsine in 1889 because it could not buy aniline and toluidine cheaply enough.

The winners integrated around a large, planned site. Duisberg's 1895 plan for Leverkusen centralised intermediates in one plant, kept each dye class in its own unit, standardised every pipe and valve, built an experimental plant holding small copies of all the equipment. Reactors grew from 30 litres in 1868 to 2,000 litres by 1905, and azo vessels to 20,000 litres by 1907, but this was still batch production of many products. A 1908 study for Schoellkopf found that an aniline plant of 3 million pounds a year cost $70,000 to build in Germany against $104,000 in the United States, with wages of $61,493 against $116,236 and materials of $317,000 against $443,000.

Competition was then tamed. Ten alizarin survivors formed a price-fixing convention in 1881; it broke down in 1885 and a new one ran from 1901. In 1904 Duisberg proposed merging all German dye firms while profits were high. They settled for two groups: Bayer, BASF and AGFA, and Hoechst, Cassella and Kalle. Even Schoellkopf agreed in 1910 to sell direct blacks at the cartel price.

• Murmann (2003) · pdf pp. 130-132, 165-171, 184-185, 243

See it in the table

Check your understanding

Why is diazotisation run at 0 to 10 C?

In 1906 Bayer's labs made 2,656 new compounds. How many reached the market that year?

What did Bayer gain from its year-long 'volunteer' courses for mill owners' sons?