potential criticism was anticipated and addressed. The two men were now
‘playing hard-ball’ and in a substantial (16 page) paper, Nernst was questioned on almost every page.
They began by announcing that in view of their new results, ‘Herr Nernst
was not entitled to his previous objection’. Nernst’s values for the equilibrium,
they claimed, lay so far to the nitrogen:hydrogen side that they could not be
justified. Much of this they felt could be attributed to Nernst’s experimental
arrangement which he had already admitted was ‘simple’.
11 Their arrangement however, was ‘substantially’ different. Haber and Robert continued by
criticising Nernst’s use of an internal heating element (‘Innenheitzung’) in his
oven. This, they argued, produced a non-uniform temperature profile; in
their experiments the pressurised ‘ammonia bomb’ was uniformly heated,
12
externally, by immersion in an oil bath, and even with determinations being
made from both sides of the equilibrium at 30 atm, they obtained the same
results as at 1 atm. The presence of impurities too, they felt, dramatically
influenced the results. Nernst had revealed that early inconsistencies were
observed in his experiments because of the presence of small amounts of
oxygen in the ammonia which formed water. Haber and Robert found that
with an iron catalyst, any oxygen was removed as the oxide. The catalytic
effect however was not impaired
13 and so for iron this impurity was not a
problem. For other metals however oxygen had to be removed. But Robert
also found that in the commercial gases used by both himself and Nernst,
residual carbon dioxide
14 was equally problematic because it was an unexpected partial pressure and unaccounted for in he calculations. Finally, Robert
simulated Nernst’s experiments by using platinum as catalyst, but when this
metal was finely distributed on an asbestos bed he found that ‘platinsilicid’
formed on it’s surface. As a result the catalytic effect was then ‘very slight’.
None of this had been addressed by Nernst, so his figures were always going
to lie on the side of reduced yields of ammonia at equilibrium.
Once the shortcomings of Nernst’s experimental design had been
addressed the two men had to establish the integrity of their own investigation. The apparatus
15 and quantitative determinations they described were
essentially the same as that used in their Hamburg paper with of course the
addition of the pressurised reaction chamber, a clear quartz tube about 60 cm
long with walls 3 mm thick, which unlike Nernst’s metal tube, was less
permeable to diffusion. Gases from commercial high pressure cylinders
16 were
brought to 30 atm by means of a combined pressure reducing valve/mercury
manometer
17 from the firm J. Amsler-Laff. The Mechanical Institute of the
University was well experienced in the use of this technology and they also
provided a Schaffer-Budenberg manometer by means of which the
132
D. Sheppard
‘playing hard-ball’ and in a substantial (16 page) paper, Nernst was questioned on almost every page.
They began by announcing that in view of their new results, ‘Herr Nernst
was not entitled to his previous objection’. Nernst’s values for the equilibrium,
they claimed, lay so far to the nitrogen:hydrogen side that they could not be
justified. Much of this they felt could be attributed to Nernst’s experimental
arrangement which he had already admitted was ‘simple’.
11 Their arrangement however, was ‘substantially’ different. Haber and Robert continued by
criticising Nernst’s use of an internal heating element (‘Innenheitzung’) in his
oven. This, they argued, produced a non-uniform temperature profile; in
their experiments the pressurised ‘ammonia bomb’ was uniformly heated,
12
externally, by immersion in an oil bath, and even with determinations being
made from both sides of the equilibrium at 30 atm, they obtained the same
results as at 1 atm. The presence of impurities too, they felt, dramatically
influenced the results. Nernst had revealed that early inconsistencies were
observed in his experiments because of the presence of small amounts of
oxygen in the ammonia which formed water. Haber and Robert found that
with an iron catalyst, any oxygen was removed as the oxide. The catalytic
effect however was not impaired
13 and so for iron this impurity was not a
problem. For other metals however oxygen had to be removed. But Robert
also found that in the commercial gases used by both himself and Nernst,
residual carbon dioxide
14 was equally problematic because it was an unexpected partial pressure and unaccounted for in he calculations. Finally, Robert
simulated Nernst’s experiments by using platinum as catalyst, but when this
metal was finely distributed on an asbestos bed he found that ‘platinsilicid’
formed on it’s surface. As a result the catalytic effect was then ‘very slight’.
None of this had been addressed by Nernst, so his figures were always going
to lie on the side of reduced yields of ammonia at equilibrium.
Once the shortcomings of Nernst’s experimental design had been
addressed the two men had to establish the integrity of their own investigation. The apparatus
15 and quantitative determinations they described were
essentially the same as that used in their Hamburg paper with of course the
addition of the pressurised reaction chamber, a clear quartz tube about 60 cm
long with walls 3 mm thick, which unlike Nernst’s metal tube, was less
permeable to diffusion. Gases from commercial high pressure cylinders
16 were
brought to 30 atm by means of a combined pressure reducing valve/mercury
manometer
17 from the firm J. Amsler-Laff. The Mechanical Institute of the
University was well experienced in the use of this technology and they also
provided a Schaffer-Budenberg manometer by means of which the
132
D. Sheppard