The separation of ethane from natural gas mixtures affords an alternative source of feedstock for the production of ethylene by the reforming of ethane. Natural gas wells in Western Europe contain approximately 3% by volume of ethane (plus 97% methane, with trace quantitites of heavier hydrocarbons and rare gases); natural gas deposits containing up to 8% by volume of ethane have been found in other parts of the world. Separation or enrichment of small quantities of ethane from large volume flows of natural gas could be achieved by distillation or absorption but the inherent difficulties of such processes render them unattractive in this instance. Cryogenic distillation, for example, requires an excessively large capital outlay for compressors and heat exchange equipment while the chemical and physical similarity of the components to be separated considerably reduces the usual advantages of absorption.

Adsorption, on the other hand, offers potential possibilities in theory as a separation process for extracting low concentrations of ethane from natural gas, in view of the selective and adsorptive properties of some porous adsorbents.

Molecular sieves are already used in the gas processing industry for the removal of water, sulphur-containing compounds, carbon dioxide and oxygen from natural gas. However, it has not been at all certain that the equilibrium data of adsorption kinetics, which have hitherto been unquantified, would allow a practical separation of ethane from methane to be developed. According to the present invention, ethane is separated from methane by passing a gaseous mixture of the ethane and the methane (which mixture may include a carrier gas such as helium) at a temperature above 150K and at a pressure below 80 bar (preferably at least 1 bar, more preferably at least 3 bar) through a bed of an adsorber having an effective molecular sieve

size of from 4A (preferably from 4.5 A) to lOA until the methane concentration per unit volume of gas leaving the bed is at least as great as the methane concentration per unit volume of the mixture passing into the bed ( at the same pressure) , discontinuing the passage of the mixture (but optionally allowing any carri er gas to continue to flow) and collecting the gas which is released from the bed next after a complete cessation of methane release. This last-mentioned gas is relatively enriched in ethane, often containing only ethane .(plus any carri er gas ) . In practice, the rate of release of the ethane rises from zero to a peak, after which it falls more slowly. When the rate has fallen to a value where it ceases to be worth collecting the ethane, the separation process , as set forth above, may be repeated. By the time that the methane concentration of gas leaving - the bed has risen to be equal to the methane concentration of the mixture passing into the bed, methane is being desorbed from the bed, and when no further methane passes into the bed, this desorption proceeds quickly to completion (say 5 minutes in a typical case) , i . e. to the complete cessation of methane release. In such a -case, after a further 1 - 5 minutes , ethane starts to be desorbed, rising to a peak after 5 - 10 minutes and continuing at a reasonable rate for 30 - 40 minutes overall . As a gui de, the breakthrough time for ethane to be desorbed (if the bed is completely saturated) equals the duration of the adsorption part of the cycle plus l-J times the breakthrough time for methane.

The separation appears not to be grossly affected by the methane: ethane volume ratio , which may be (say) from 1: 1 to 3 5 1, nor by the presence of an inert carrier gas such as helium, which may be present as up to 9 % of the total volume.

The temperature of operation may be up to 400K but is preferably not more than 300K and more preferably below 270K. At lower temperatures , the above-mentioned complete cessation of gas release tends to lengthen , which is advantageous for ease of process control .

The adsorber is preferably a A molecular si eve . A 4A molecular sieve is less effective.

The invention will now be described by way of example. An adsorber bed was made up as follows . A 5A molecular si eve was supplied by Laporte Industri es Ltd as 1-2 mm spherical pellets (made from 1-5 ~ zeolite crystals ) containing 20% by weight of binder and calcined at 650 C. These pellets were then crushed ^' in a ball mill and the fraction between B.S . mesh 30 and

40 was collected. Macro-pore volume and average pore diameter, _, as determined by mercury pDrosimetry, were 0.322 cm 3.g -1 and

6θ3A respectively. The absence Of any hysteresis indicated an absence of pores containing narrow necks . This material was packed ( to a bulk density of 0.734 g .cm -3 ) into a stainless steel column 1.5m long and 5~~ in internal diameter immersed in a thermostatically controlled water bath, connected by means of copper pipework to a gas supply upstream and collection and chromatographic equipment downstream. The ends of the adsorption column were plugged with glass wool and any dead space remaining after the tube was packed with adsorbent was filled with 30 - 6θ mesh non-adsdrbent glass beads .

A gas mixture under a pressure of 3 -3 bar was passed through the bed at a rate of 2.5 cm /sec at 293K, and consisted (by volume) of 94.5% helium, 2.9% methane and 2.6% ethane. At 4 minutes from the start , the gas leaving the bed ( at the same temperature and pressure) started to show traces of methane, whi ch rose sharply

(within 1 further minute) to a concentration of 2.9%» where it levelled off . At 6 minutes from the start , desorption was accomplished by means of a step decrease to zero of adsorbate

(methane-plus-ethane) concentration, only helium carrier gas (at the former rate) being fed to the adsorbent bed from that moment until (as will be described) a second cycle of operations commenced at an increased total adsorbate pressure after all the ethane had been desorbed. The rate of desorption is slower than the rate of adsorption and the rate of desorption of methane is much more rapid than the rate of ethane desorption. Thus, the methane desorbed until Minute 8 at 2.9/94.5 of the rate of flow of the helium, thereafter falling to zero by Minute 11. Thereafter, only helium emerged from the bed until Minute 20, when ethane started to desorb. (Meanwhile, at about Minute 1 j the gas leaving the bed was switched to an ethane collecting reservoir, from which helium could be removed by known means). The rate of ethane desorption rose approximately linearly to a peak at Minute 27j when it was about 0.9/94.5 of the rate of flow of the helium (i.e. about a third of its flow rate into the bed at the start). Thereafter, the rate of ethane desorption fell, approximately exponentially, remaining reasonable at Minute 45 but becoming negligible after about Minute 55« Ethane collection was accordingly discontinued at Minute 50 and the whole process of the foregoing paragraph repeated, either with fresh gas for separation or if necessary with feedstock recycled from the ethane collecting reservoir.

The cycle time here is 1 hour, but depending on the size of the plant, the degree of separation required and other variables, cycle times of from 4θ minutes to 1 year are possible.

In another run with the same apparatus, using the same gas mixture, desorption was initiated not at Minute 6 but instead at Minute 10. By that time, methane was leaving the bed at 108% of the rate at which it was entering, and rose to 110% at Minute 11. By Minute 12 it had fallen to zero. As before, the gas was directed to ethane collection at Minute 15« Ethane started to

desorb at Minute 17 _. rising approximately linearly to a peak at Minute 25 , when it was about half of its original flow rate into the bed. Thereafter, the rate of ethane desorption fell , approximately exponentially, becoming very small around Minute 50. In trials where the gas feed consisted of 90% helium, 5% methane and 5% ethane, at 2 bar total pressure, methane desorption was completed just a few seconds before onset of ethane desorption, near Minute 12. Thus the separation there of ethane from methane according to the invention was still possible.