Wro
´
blewski began an initially fruitful partnership with a colleague
in the Chemistry Department, Karol Olszewski, and, by making
some modifications to Cailletet’s equipment, they were able to do
more than simply produce a fine mist of liquid droplets of oxygen:
in March 1883, they produced liquid oxygen quietly boiling away
by itself in a test tube! Two weeks later, they repeated the trick with
liquid nitrogen and Krako
´
w instantly became the world-leading
centre of low-temperature physics. Unfortunately, Wro
´
blewski
and Olszewski had a serious falling out and their professional
relationship broke up after a further six months. Thereafter they
worked independently in their own departments, despite working
on precisely the same project, that of attempting to liquefy
hydrogen. Toiling late one night in his laboratory in 1888,
Wro
´
blewski upset a kerosene lamp on his desk and was so badly
burned he died soon afterwards. Olszewski continued to work on
low-temperature problems, developing an improved Pictet-style
apparatus.
The principle of cooling by rapid expansion had been established
much earlier, in 1852, by James Prescott Joule, together with the
Belfast-born mathematical physicist William Thomson, later to be
known as Lord Kelvin. Their effect is known either as the Joule–
Thomson effect or, reflecting Thomson’s later elevation, as the
Joule–Kelvin effect . It works because, as a gas expands, the average
distance between molecules increases and this alters the effect of
the weak intermolecular attractive forces. It turns out that the
Joule–Thomson effect only leads to cooling if the gas is already at a
lowish temperature but, this complication notwithstanding, the
effect is hugely important for liquefying gases.
One method of getting gases to expand was by allowing high
pressure gas to squirt out of a fine nozzle or constriction into a
region of low pressure. This would cool the gas, allowing it to
liquefy, and any cold gas remaining could be recompressed and
forced around a circuit and back into the high pressure vessel.
In this way, a steady flow process could be produced and a gas
12
Superconductivity