Liquid Oxygen

Oxygen was not obtained in the liquid state by Faraday in his classical investigations on the liquefaction of gases, because the refrigerating agents used by him did not suffice for the attainment of the critical temperature of the gas, above which it is impossible to effect liquefaction, no matter how great the pressure.

The gas was first reduced to the liquid state by Cailletet in 1877, and almost simultaneously by Pictet. The former investigator, who effected the cooling merely by the sudden expansion of the gas from a pressure of 300 atmospheres, obtained only a mist of small globules liquid oxygen. Pictet, however, cooled the gas, already compressed to 320 atmospheres, to -140° C. in a bath of rapidly evaporating liquid carbon dioxide and was able to collect a small quantity of the liquid. Liquid oxygen was first produced in sufficient bulk for satisfactory examination by Wroblewski and Olszewski who made use of liquid ethylene, boiling rapidly under reduced pressure, as a refrigerant. The rapid evaporation of liquid ethylene in vacuo leads to a temperature of - 152° C, and Dewar utilised this in preparing liquid air and oxygen in large quantities.

The methods for the production of liquid air are divisible into two classes according to whether the cooling of the gases is due to the external or internal work performed by them.

The former method is based on the principle that the sudden, adiabatic expansion of gases against an external pressure causes external work to be done by them, accompanied by a proportional diminution in their own internal energy manifested by a reduction in temperature. Although this method was introduced by Cailletet in 1877 and was successfully applied by him to the liquefaction of oxygen, nitrogen, and air, it was not until 1905 that it was successfully applied on a commercial scale, namely, in the Claude Process.

Claude's apparatus for the production of loquid air.

The difficulty of lubrication appears to have been mainly responsible for the failure of previous attempts, and this was first overcome by the employment of petroleum ether which does not solidify, but merely becomes viscous at such low temperatures as -140° to -160° C. Later, however, it was found that leather retains its ordinary properties at these low temperatures, and in 1912 leather stampings were fitted to the working parts of the machinery to the entire exclusion of lubricants. Claude's apparatus is shown diagrammatically in fig.

Air, compressed to 40 atmospheres, passes along the inner tube T₁ of the usual concentric system to the branched tube B, where it is placed in connection with a "liquefier " whilst much of the gas passes on through the expansion machine. Cooled by its loss of energy during expansion, it proceeds to the tubes inside the liquefier, and finally passes along the outer of the concentric tubes, thus cooling the oncoming air which reaches the expansion machine at -100° C. So cold does the expanded gas become that the compressed air in the liquefier finally condenses and is tapped off periodically, whilst the gas, after exerting this cooling effect, flows from the tubes of the liquefier into the outer tube T₂ and reduces to approximately - 100° C. the temperature of the air reaching T₁.

The liquid air is usually collected and stored in Dewar vacuum flasks. These are double-walled glass vessels, the space between the walls being completely evacuated, so that the liquid in the flask is vacuum-jacketed. The heat conveyed by radiation across the vacuous space is only about one-sixth of that which would reach the liquid by conduction and convection if the space were filled with air; and this can be reduced to one-thirtieth by silvering the interior of the jacket. This latter procedure, however, is impracticable if for any purpose it is necessary to observe the contents-of the flask.

The second method of producing liquid air is based on the internal work performed by a gas upon expansion during passage from a high to a low pressure, the work being mainly that necessary to overcome the attraction between the gaseous molecules. This work is carried out at the expense of the sensible heat of the gas, and the effect is the greater the lower the temperature. It would not exist in the case of a perfect gas upon free expansion, namely, into a vacuum, and must be carefully distinguished from the cooling already considered as the result of adiabatic expansion against the external atmospheric pressure, and as utilised in the Claude Process. This thermal effect was first studied by Joule and Thomson, and is exhibited by oxygen and nitrogen and therefore by air, in the case of the last named, up to a temperature of 259° C. under normal pressure. The cooling, which is a small effect, amounting in the case of air at the ordinary temperature to only about 0.255° C. for a fall in pressure of one atmosphere, may be calculated from the expression



where p₁ is the initial high pressure, p₂the final low pressure, and T the initial absolute temperature, the cooling being expressed in degrees centigrade. By employing high pressures the cooling effect is proportionately enhanced. Thus, if a pressure difference of 100 atmospheres is employed, working at 0° C., the fall in temperature is 27.6 centigrade degrees.

The Linde liquid-air machine.

By allowing air to expand suddenly at ordinary temperatures a certain cooling is thus produced, and by applying this cooled gas to the reduction of the temperature of yet unexpanded gas, the latter after expansion will attain a still lower temperature. In this way it is possible to make the cooling effect cumulative so that at last the temperature of the air is reduced to the liquefying point. The Linde, Dewar, and Hampson lique- fiers are based on this principle.

A diagrammatic representation of the Linde machine is given in fig. Air compressed to 200 atmospheres passes through the steel bottle B where it deposits its moisture, and thence proceeds to the worm surrounded by a refrigerating medium. Here the temperature is reduced to -50° C, and the last traces of water-vapour are removed. The gas passes thence down the innermost of the concentric copper tubes T, by way of which it reaches the needle-valve V, where it expands to a pressure of 40 atmospheres. This limited expansion yields the major portion of the Joule-Thomson effect and at the same time reduces the subsequent necessary work of compression. The cooled expanded gas returns through the second concentric tube to the compressor, cooling the oncoming air as it passes. As this process is continued, the air reaching V steadily falls in temperature until at last it begins to condense to the liquid state, when the liquid is allowed periodically to pass through the valve Y where, on account of the further decrease of pressure to one atmosphere, the liquid evaporates vigorously until its temperature falls to its normal boiling-point for this pressure; the cold gas from the evaporation passes away through the outermost concentric tube and so assists in cooling the compressed air, whilst the liquid air collects in the receiver R and can be drawn off as required by the tap C. The apparatus is enclosed in a packing of non-conducting material such as wool and is supported externally by a wooden or metallic case. In the earlier forms of this type of liqucfier the process was somewhat simpler because the pressure was allowed to fall directly to the ordinary external atmospheric pressure by one expansion only. Machines of the more modern type have been constructed to yield over 50 litres of liquid air per hour.

On account of the great importance of oxygen and the increasing importance of nitrogen for industrial and other purposes, the liquid mixture of these elements provides a promising field for a successful process for the production of the gases on a large scale.

The vapour of boiling liquid air is richer in nitrogen than the liquid, hence careful fractional distillation or evaporation should finally yield the oxygen in a pure condition because the boiling-point rises steadily as the percentage of oxygen increases. Bearing in mind the proximity to the absolute zero, it will be easily recognised that the relative difference between the boiling- points of the two constituents, namely oxygen -182.9° C. and nitrogen -195.67° C., is very considerable and that the main difficulties are likely to be of a mechanical type.

The Linde liquid-air machine.

Several forms of apparatus have been proposed. One of the earlier forms suggested by Linde consisted of a modification of the apparatus represented in fig.; this was supplied with only one valve which allowed immediate expansion to atmospheric pressure, the liquid air produced by the cooling being collected in a suitable receiver. The compressed gas, before reaching the valve, was made to circulate through a copper coil actually inside the receiver so as to be covered by the liquid air already formed. The relative warmth of this gas caused an evaporation of the more volatile nitrogen, the liquid lost by evaporation being replaced by fresh liquid air produced by the expansion of the cooled gas. Proceeding in this way, the receiver soon contains fairly pure liquid oxygen which can be drawn off as necessary and transported, in the form of compressed gas, in steel cylinders. As the gaseous nitrogen which passes away from the apparatus is formed by the evaporation of a liquid containing at least 21 per cent, of oxygen, the nitrogen is not pure but must contain at least 7 per cent, of oxygen. A recent form of the Linde oxygen plant is shown in figs.

The Linde oxygen plant.

Prior to admission to the plant, the air is compressed to 135 atmospheres (2000 lb. per square inch), and cooled to -20° C. in an ordinary refrigerating apparatus. This serves to freeze out atmospheric moisture. Carbon dioxide is removed by passage through a slaked lime purifier. Thus treated, the air is admitted to the Linde plant at the mouth of the regenerator spiral AA' through three small pipes a, one of which is surrounded by a wider concentric pipe b, as indicated in fig. These small pipes continue, inside AA', to encircle the rectifying column D and merge into the smaller spiral surrounded by liquid oxygen in B. The air on its passage becomes increasingly cooler, and escapes by way of the throttle-valve C to the top of the rectifying column D, a fall in temperature occurring at C owing to the Joule-Thomson effect. Ultimately a liquid rich in oxygen collects in B, whilst gas, rich in nitrogen and containing only about 7 per cent, of oxygen, escapes at E and leaves the apparatus through the regenerator spiral AA', cooling in its passage, by conduction, the incoming air in a. The oxygen at F leaves through the tube b passing up inside A'A.

When the apparatus has been at work a sufficient time to become steady, the liquid in B is continuously evaporated by the warmer air passing through the spiral, and the vapours escaping from B are rich in nitrogen, whilst the liquid remaining is rich in oxygen. The rectifying tower, with its baffle plates, reduces the amount of oxygen in the vapours escaping at E to about 7 per cent., for the ascending gases are constantly meeting liquids whose temperatures further up the rectifying column are increasingly lower. The oxygen thus condenses and joins the descending liquid stream. On the other hand, the nitrogen in that stream meets increasingly warmer gases as it falls, and having a lower boiling-point than the oxygen, it evaporates away and escapes at E. Liquid oxygen of 98 to 99 per cent; purity thus collects in B and is finally drawn off at H through b.

The efficiency of the apparatus depends upon the temperature gradient between D and F, and this is controlled by the throttle-valve C. In practice it is found that a pressure of 50 to 60 atmospheres is sufficient, when the plant is in steady running, the temperature of the entering liquid air at D being -192° C., and at F -181.5° C.

The Claude separator

In Claude's process compressed air, cooled by passage through a coil surrounded by the cold gases issuing from other parts of the apparatus, enters the lower portion of the apparatus (fig.) at A where it reaches the inner part of the tubular vessel B of annular cross-section; this vessel is surrounded by liquid oxygen. During its ascent through B, the air becomes partially condensed to a liquid which, will contain up to 47 per cent, of oxygen. If the pressure of the incoming gas is correctly adjusted, the residual gas will consist of almost pure nitrogen, which will pass over into the external tubular space C, where it becomes entirely liquefied. The liquids condensed in D and E are therefore greatly enriched in oxygen and nitrogen respectively before admission to the "still" proper. The liquid collecting in D is caused by its pressure to rise through a regulating-valve into the fractionating column at F, and, overflowing downwards, meets the ascending gases from the liquid oxygen in H. On account of the contact between these two currents, the descending liquid grows steadily richer in oxygen until it reaches the vessel H, which is in connection with the tubes in B, as liquid oxygen. The gases rising up the column beyond F become submitted to further " scrubbing " by the liquid nitrogen reaching G from E, the effect of this being to condense any oxygen still remaining in the gas so that it returns to scrub the ascending gases in the lower portion of the column, whilst the gas issuing at the top is reduced to pure nitrogen. Gaseous oxygen can be drawn from I above the condensed liquid. Thus almost pure oxygen and nitrogen are simultaneously produced.

Commercial liquid oxygen may contain argon. Morey found the composition of liquid oxygen to be as follows:

Oxygen	96.9 per cent.
Argon	2.8 per cent.
Nitrogen	0.3 per cent.

Liquid oxygen is a transparent liquid, possessed of a bluish tinge. The boiling-point of liquid oxygen varies with the pressure, as indicated in the following table

Variation of the boiling-point of oxygen with the pressure

Pressure, mm.	Absolute Boiling-point.
	Hydrogen Scale	Helium Scale.
800	90.60	90.70
760	90.10	90.20
700	89.33	89.43
600	87.91	88.01
500	86.29	86.39
400	84.39	84.49
300	82.09	82.19
200	79.07	79.17

The vapour pressure rises from 9.096 atm. At -154.91° C. to 49.640 atm. At -118.70° C.

The vapour pressure of oxygen at any temperature between 57° and 90° abs. may be calculated from the expression

log p = -419.31/T + 5.2365 - 0.0648T

where the pressure p is expressed in atmospheres, T being the absolute temperature.

For p = 1 atm., the value for T becomes 90.13°, which agrees very satisfactorily with the boiling-point under normal pressure as given in the preceding table.

Densities of liquid oxygen at various temperatures

Temperature, ° C.	Density.
-183.6	1.1321
-183.3	1.1310
-182.5	1.1181
-195.5	1.1700
-210.5	1.2386
-193.93	1.203
-198.30	1.223

When the values obtained by Dewar for the densities are plotted against the absolute temperatures, they are seen to lie very closely to a straight line, so that the densities at intermediate temperatures can readily be calculated. The expression is

Density = 1.5154 – 0.004420T

where T is the absolute temperature.

Using the data given by Ramsay and Drugman, the specific volume of oxygen at -183° C. is 0.8838, and the molecular volume 28.28.

When exposed to the air, liquid oxygen absorbs appreciable quantities of nitrogen.

Liquid oxygen is more compressible than water, its coefficient of compressibility being 0.00195 between 10 and 20 atmospheres.

The observed surface tension of the liquid is 13.074 dynes per cm. - a value in fair agreement with that expected for a liquid of the same molecular weight as gaseous oxygen, although the possibility of slight association is not excluded. From other data Inglis and Coates conclude that the degree of association of liquid oxygen at about -195° C. is 109.

Densities of liquid oxygen

Temperature, ° Abs	Temperature, ° C	Density.
68.0	-205	1.2489
70.0	-203	1.2393
74.0	-199	1.2200
78.0	-195	1.2008
80.0	-193	1.1911
82.0	-191	1.1815
86.0	-187	1.1623
89.0	-184	1.1479

The specific heat of liquid oxygen between -200° and -183° C. is 0.347, and the heat of evaporation is 51.3 calories per gram at 763 mm. pressure, its molecular heat of vaporisation being 1599 calories according to another computation. Its coefficient of expansion with rise of temperature is 0.00157 at -252.6° C. The refractive index for sodium light, is 1.2236, and the spectrum absorption similar to that of gaseous oxygen.

Liquid oxygen is a non-conductor of electricity; but it is strongly attracted by a magnet. It readily absorbs nitrogen from the atmosphere, and can be mixed with liquid fluorine without suffering chemical change. The magnetic rotatory power and dispersion have been determined.