plant. Using these tubes at the experimental site (designated Lu398) in the
main Ludwigshafen compound, the BASF first began to synthesis liquid
ammonia as early as 18 May 1910. By 19 July, enough ammonia had been
produced to fill a 5 kg container,
5 but there were clear problems developing
with the reactor tubes. These tubes performed well within a circulation system for about 80 h, but then they failed catastrophically and burst, Bosch
later observing that, ‘Had these tubes been charged with osmium, the world’s
supply of this rare metal would have disappeared!’
1
Examination of the burst tubes showed that they were swollen. Some
change in the material of the inner wall had resulted in a loss of elasticity and
this change had progressed outward until the undamaged part was so thin
that it yielded to the internal pressure. A chemical attack was suspected and
nitrogen was at first thought to be the culprit because the literature showed
that iron nitride, a silvery brittle compound, was apt to form when hot iron
and ammonia come into contact. Chemical analysis of the changed material
however showed no sign of nitrogen. Now Bosch’s provenance was in metallurgy, and he applied a technique little known to chemical engineering at
the time viz, metallurgical ‘etching’. Consequently, the following picture
slowly emerged.
Because of its high mechanical strength, carbon steel was the obvious
candidate to construct the reactors. This steel has a structure in which carbonaceous perlite is dispersed in a matrix of pure iron, but in the changed
parts of the reactor tubes, there was no sign of the perlite, and the resulting
structure had been destroyed by cracking. Bosch realised that the perlite had
disappeared because of the ‘de-carburisation’ of the steel, but this in turn
should have resulted in ‘soft iron’ which should not have cracked. On the
contrary, the changed material was hard and brittle, rather like ‘cast’ iron.
Desperate laboratory investigations soon revealed the true problem. Diffusion
of hydrogen
6 into the steel caused de-carburisation via the formation of
methane gas leaving a brittle alloy of iron hydride. The gas, trapped under
extreme high pressure within an embrittled metal, added to the mechanical
stress on the material, loosening the structure and contributing to a complete
wall failure. Bosch concluded that all carbon steels would suffer this ‘hydrogen
attack’ and fail in a matter of hours or days. He also concluded that even in
the absence of ‘hydrogen attack’, external heating of the steel jackets of these
converters put further stress on their walls because the temperature and
pressure gradients involved led to buckling and cracking, causing heavy
explosions and costly repairs. The original Haber-Le Rossignol experiments of
course, had only been run for a few hours and so these problems had never
arisen—although to be fair, Haber and Robert had also converted to internal
216
D. Sheppard
main Ludwigshafen compound, the BASF first began to synthesis liquid
ammonia as early as 18 May 1910. By 19 July, enough ammonia had been
produced to fill a 5 kg container,
5 but there were clear problems developing
with the reactor tubes. These tubes performed well within a circulation system for about 80 h, but then they failed catastrophically and burst, Bosch
later observing that, ‘Had these tubes been charged with osmium, the world’s
supply of this rare metal would have disappeared!’
1
Examination of the burst tubes showed that they were swollen. Some
change in the material of the inner wall had resulted in a loss of elasticity and
this change had progressed outward until the undamaged part was so thin
that it yielded to the internal pressure. A chemical attack was suspected and
nitrogen was at first thought to be the culprit because the literature showed
that iron nitride, a silvery brittle compound, was apt to form when hot iron
and ammonia come into contact. Chemical analysis of the changed material
however showed no sign of nitrogen. Now Bosch’s provenance was in metallurgy, and he applied a technique little known to chemical engineering at
the time viz, metallurgical ‘etching’. Consequently, the following picture
slowly emerged.
Because of its high mechanical strength, carbon steel was the obvious
candidate to construct the reactors. This steel has a structure in which carbonaceous perlite is dispersed in a matrix of pure iron, but in the changed
parts of the reactor tubes, there was no sign of the perlite, and the resulting
structure had been destroyed by cracking. Bosch realised that the perlite had
disappeared because of the ‘de-carburisation’ of the steel, but this in turn
should have resulted in ‘soft iron’ which should not have cracked. On the
contrary, the changed material was hard and brittle, rather like ‘cast’ iron.
Desperate laboratory investigations soon revealed the true problem. Diffusion
of hydrogen
6 into the steel caused de-carburisation via the formation of
methane gas leaving a brittle alloy of iron hydride. The gas, trapped under
extreme high pressure within an embrittled metal, added to the mechanical
stress on the material, loosening the structure and contributing to a complete
wall failure. Bosch concluded that all carbon steels would suffer this ‘hydrogen
attack’ and fail in a matter of hours or days. He also concluded that even in
the absence of ‘hydrogen attack’, external heating of the steel jackets of these
converters put further stress on their walls because the temperature and
pressure gradients involved led to buckling and cracking, causing heavy
explosions and costly repairs. The original Haber-Le Rossignol experiments of
course, had only been run for a few hours and so these problems had never
arisen—although to be fair, Haber and Robert had also converted to internal
216
D. Sheppard