Acetylene, the Principles of Its Generation and Use Part 12

HARMFULNESS OF IMPURITIES.--The reasons why the carbide impurities must be removed from acetylene before it is burned have now to be explained.

From the strictly chemical point of view there are three compounds of phosphorus, all termed phosph.o.r.etted hydrogen or phosphine: a gas, PH_3; a liquid, P_2H_4; and a solid, P_4H_2. The liquid is spontaneously inflammable in presence of air; that is to say, it catches fire of itself without the a.s.sistance of spark or flame immediately it comes in contact with atmospheric oxygen; being very volatile, it is easily carried as vapour by any permanent gas. The gaseous phosphine is not actually spontaneously inflammable at temperatures below 100 C.; but it oxidises so rapidly in air, even when somewhat diluted, that the temperature may quickly rise to the point of inflammation. In the earliest days of the acetylene industry, directly it was recognised that phosphine always accompanies crude acetylene from the generator, it was believed that unless the proportion were strictly limited by decomposing only a carbide practically free from phosphides, the crude acetylene might exhibit spontaneously inflammable properties. Lewes, indeed, has found that a sample of carbide containing 1 per cent of calcium phosphide gave (probably by local decomposition--the bulk of the phosphide suffering attack first) a spontaneously inflammable gas; but when examining specimens of commercial carbide the highest amount of phosphine he discovered in the acetylene was 2.3 per cent, and this gas was not capable of self-inflammation. According to Bullier, however, acetylene must contain 80 per cent of phosphine to render it spontaneously inflammable. Berdenich has reported a case of a parcel of carbide which yielded on the average 5.1 cubic foot of acetylene per lb., producing gas which contained only 0.398 gramme of phosphorus in the form of phosphine per cubic metre (or 0.028 per cent. of phosphine) and was spontaneously inflammable. But on examination the carbide in question was found to be very irregular in composition, and some lumps produced acetylene containing a very high proportion of phosphorus and silicon compounds. No doubt the spontaneous inflammability was due to the exceptional richness of these lumps in phosphorus. As manufactured at the present day, calcium carbide ordinarily never contains an amount of phosphide sufficient to render the gas dangerous on the score of spontaneous inflammability; but should inferior material ever be put on the markets, this danger might have to be guarded against by submitting the gas evolved from it to chemical a.n.a.lysis. Another risk has been suggested as attending the use of acetylene contaminated with phosphine (and to a minor degree with sulphuretted hydrogen), viz., that being highly toxic, as they undoubtedly are, the gas containing them might be extremely dangerous to breathe if it escaped from the service, or from a portable lamp, unconsumed. Antic.i.p.ating what will be said in a later paragraph, the worst kind of calcium carbide now manufactured will not yield a gas containing more than 0.1 per cent. by volume of sulphuretted hydrogen and 0.05 per cent. of phosphine. According to Haldane, air containing 0.07 per cent. of sulphuretted hydrogen produces fatal results on man if it is breathed for some hours, while an amount of 0.2 per cent. is fatal in 1- 1/2 minutes. Similar figures for phosphine cannot be given, because poisoning therewith is very rare or quite unknown: the cases of "phossy- jaw" in match factories being caused either by actual contact with yellow phosphorus or by inhalation of its vapour in the elemental state.

However, a.s.suming phosphine to be twice as toxic as sulphuretted hydrogen, its effect in crude acetylene of the above-mentioned composition will be equal to that of the sulphuretted hydrogen, so that in the present connexion the gas may be said to be equally toxic with a sample of air containing 0.2 per cent. of sulphuretted hydrogen, which kills in less than two minutes. But this refers only to crude acetylene undiluted with air; and being a hydrocarbon--being in fact neither oxygen nor common air--acetylene is irrespirable of itself though largely devoid of specific toxic action. Numerous investigations have been made of the amount of acetylene (apart from its impurities) which can be breathed in safety; but although these point to a probable recovery after a fairly long-continued respiration of an atmosphere charged with 30 per cent. of acetylene, the figure is not trustworthy, because toxicological experiments upon animals seldom agree with similar tests upon man. If crude acetylene were diluted with a sufficient proportion of air to remove its suffocating qualities, the percentage of specifically toxic ingredients would be reduced to a point where their action might be neglected; and short of such dilution the acetylene itself would in all probability determine pathological effects long before its impurities could set up symptoms of sulphur and phosphorus poisoning.

Ammonia is objectionable in acetylene because it corrodes bra.s.s fittings and pipes, and because it is partially converted (to what extent is uncertain) into nitrous and nitric acids as it pa.s.ses through the flame.

Sulphur is objectionable in acetylene because it is converted into sulphurous and sulphuric anhydrides, or their respective acids, as it pa.s.ses through the flame. Phosphorus is objectionable because in similar circ.u.mstances it produces phosphoric anhydride and phosphoric acid. Each of these acids is harmful in an occupied room because they injure the decorations, helping to rot book-bindings, [Footnote: It is only fair to state that the destruction of leather bindings is commonly due to traces of sulphuric acid remaining in the leather from the production employed in preparing it, and is but seldom caused directly by the products of combustion coming from gas or oil.] tarnishing "gold-leaf" ornaments, and spoiling the colours of dyed fabrics. Each is harmful to the human system, sulphuric and phosphoric anhydrides (SO_3, and P_4O_10) acting as specific irritants to the lungs of persons predisposed to affections of the bronchial organs. Phosphorus, however, has a further harmful action: sulphuric anhydride is an invisible gas, but phosphoric anhydride is a solid body, and is produced as an extremely fine, light, white voluminous dust which causes a haze, more or less opaque, in the apartment.

[Footnote: Lewes suggests that ammonia in the gas burnt may a.s.sist in the production of this haze, owing to the formation of solid ammonium salts in the state of line dust.] Immediately it comes in contact with atmospheric moisture phosphoric anhydride is converted into phosphoric acid, but this also occurs at first as a solid substance. The solidity and visibility of the phosphoric anhydride and acid are beneficial in preventing highly impure acetylene being unwittingly burnt in a room; but, on the other hand, being merely solids in suspension in the air, the combustion products of phosphorus are not so easily carried away from the room by the means provided for ventilation as are the products of the combustion of sulphur. Phosphoric anhydride is also partly deposited in the solid state at the burner orifices, perhaps actually corroding the steat.i.te jets, and always a.s.sisting in the deposition of carbon from any polymerised hydrocarbons in the acetylene; thus helping the carbon to block up or distort those orifices. Whenever the acetylene is to be burnt on the incandescent system under a mantle of the Welsbach or other type, phosphorus, and possibly sulphur, become additionally objectionable, and rigorous extraction is necessary. As is well known, the mantle is composed of the oxides of certain "rare earths" which owe their practical value to the fact that they are non-volatile at the temperature of the gas-flame. When a gas containing phosphorus is burnt beneath such a mantle, the phosphoric anhydride attacks those oxides, partially converting them into the respective phosphates, and these bodies are less refractory. A mantle exposed to the combustion products of crude acetylene soon becomes brittle and begins to fall to pieces, occasionally showing a yellowish colour when cold. The actual advantage of burning acetylene on the incandescent system is not yet thoroughly established-- in this country at all events; but it is clear that the process will not exhibit any economy (rather the reverse) unless the plant is provided with most capable chemical purifiers. Phosphorus, sulphur, and ammonia are not objectionable in crude acetylene because they confer upon the gas a nauseous odour. From a well-constructed installation no acetylene escapes unconsumed: the gas remains wholly within the pipes until it is burnt, and whatever odour it may have fails to reach the human nostrils.

A house properly piped for acetylene will be no more conspicuous by its odour than a house properly piped for coal-gas. On the contrary, the fact that the carbide impurities of acetylene, which, in the absolutely pure state, is a gas of somewhat faint, hardly disagreeable, odour, do confer upon that gas a persistent and unpleasant smell, is distinctly advantageous; for, owing to that odour, a leak in the pipes, an unclosed tap, or a fault in the generating plant is instantly brought to the consumer's attention. A gas wholly devoid of odour would be extremely dangerous in a house, and would have to be scented, as is done in the case of non-carburetted water-gas when it is required for domestic purposes.

AMOUNTS OF IMPURITIES AND SCOPE OF PURIFICATION.--Partly for the reason which has just been given, and partly on the ground of expense, a complete removal of the impurities from crude acetylene is not desirable.

All that need be done is to extract sufficient to deprive the gas of its injurious effects upon lungs, decorations, and burners. As it stands, however, such a statement is not sufficiently precise to be useful either to consumers of acetylene or to manufacturers of plant, and some more or less arbitrary standard must be set up in order to define the composition of "commercially pure" acetylene, as well as to gauge the efficiency of any process of purification. In all probability such limit may be reasonably taken at 0.1 milligramme of either sulphur or phosphorus (calculated as elementary bodies) per 1 litre of acetylene, _i.e._, 0.0-1.1 grain per cubic foot; a quant.i.ty which happens to correspond almost exactly with a percentage by weight of 0.01. Owing to the atomic weights of these substances, and the very small quant.i.ties being considered, the same limit hardly differs from that of 0.01 per cent. by weight of sulphuretted hydrogen or of phosphine--it being always recollected that the sulphur and phosphorus do not necessarily exist in the gas as simple hydrides. Keppeler, however, has suggested the higher figure of 0.15 milligramme of either sulphur or phosphorus per litre of acetylene (=0.066 grain per cubic foot) for the maximum amount of these impurities permissible in purified acetylene. He adopts this standard on the basis of the results of observations of the amounts of sulphur and phosphorus present in the gas issuing from a purifier charged with heratol at the moment when the last layer of the heratol is beginning to change colour. No limit has been given for the removal of the ammonia, partly because that impurity can more easily, and without concomitant disadvantage, be extracted entirely; and partly because it is usually removed in the washer and not in the true chemical purifier.

According to Lewes, the maximum amount of ammonia found in the acetylene coming from a dripping generator is 0.95 gramme per litre, while in carbide-to-water gas it is 0.16 gramme: 417 and 70.2 grains per cubic foot respectively. Rossel and Landriset have found 4 milligrammes (1.756 grains [Footnote: Milligrammes per litre; grains per cubic foot. It is convenient to remember that since 1 cubic foot of water weighs 62.321 x 16 - 997.14 avoirdupois ounces, grammes per litre are approximately equal to oz. per cubic foot; and grammes per cubic metre to oz. per 1000 cubic feet.]) to be the maximum in water-to-carbide gas, and none to occur in carbide-to-water acetylene. Rossel and Landriset return the minimum proportion of sulphur, calculated as H_2S, found in the gaseous state in acetylene when the carbide has not been completely flooded with water at 1.18 milligrammes per litre, or 0.52 grain per cubic foot; and the corresponding maxima at 1.9 milligrammes, or 0.84 grain. In carbide-to- water gas, the similar maxima are 0.23 milligramme or 0.1 grain. As already stated, the highest proportion of phosphine yet found in acetylene is 2.3 per cent. (Lewes), which is equal to 32.2 milligrammes of PH_3 per litre or 14.13 grains per cubic foot (Polis); but this sample dated from 1897. Eitner and Keppeler record the minimum proportion of phosphorus, calculated as PH_3, found in crude acetylene, as 0.45 milligramme per litre, and the maximum as 0.89 milligramme per litre; in English terms these figures are 0.2 and 0.4 grain per cubic foot. On an average, however, British and Continental carbide of the present day may be said to give a gas containing 0.61 milligramme of phosphorus calculated as PH_3 per litre and 0.75 milligramme of sulphur calculated as H_2S. In other units these figures are equal to 0.27 grain of PH_3 and 0.33 grain of H_2S per 1 cubic foot, or to 0.041 per cent. by volume of PH_3 and 0.052 per cent. of H_2S. Yields of phosphorus and sulphur much higher than these will be found in the journals and books, but such a.n.a.lytical data were usually obtained in the years 1896-99, before the manufacture of calcium carbide had reached its present degree of systematic control. A commercial specimen of carbide was seen by one of the authors as late as 1900 which gave an acetylene containing 1.12 milligramme of elementary sulphur per litre, i.e., 0.096 per cent, by volume, or 0.102 per cent, by volume of H_2S; but the phosphorus showed the low figure of 0.36 milligramme per litre (0.031 per cent, of P or 0.034 per cent, of PH_3 by volume).

The British Acetylene a.s.sociation's regulations relating to carbide of calcium (_vide_ Chap. XIV.) contain a clause to the effect that "carbide which, when properly decomposed, yields acetylene containing from all phosphorus compounds therein more than 0.05 per cent, by volume of phosph.o.r.etted hydrogen, may be refused by the buyer." This limit is equivalent to 0.74 milligramme of phosphorus calculated as PH_3 per litre. A lat.i.tude of 0.01 per cent, is, however, allowed for the a.n.a.lysis, so that the ultimate limit on which carbide could be rejected is: 0.06 volume per cent. of PH_3, or 0.89 milligramme of phosphorus per litre.

The existence in appreciable quant.i.ty of combined silicon as a normal impurity in acetylene seems still open to doubt. Calcium carbide frequently contains notable quant.i.ties of iron and other silicides; but although these bodies are decomposed by acids, yielding hydrogen silicide, or siliciuretted hydrogen, they are not attacked by plain water. Nevertheless Wolff and Gerard have found hydrogen silicide in crude acetylene, and Lewes looks upon it as a common impurity in small amounts. When it occurs, it is probably derived, as Vigouroux has suggested, from "alloys" of silicon with calcium, magnesium, and aluminium in the carbide. The metallic const.i.tuents of these substances would naturally be attacked by water, evolving hydrogen; and the hydrogen, in its nascent state, would probably unite with the liberated silicon to form hydrogen silicide. Many authorities, including Keppeler, have virtually denied that silicon compounds exist in crude acetylene, while the proportion 0.01 per cent. has been given by other writers as the maximum. Caro, however, has stated that the crude gas almost invariably contains silicon, sometimes in very small quant.i.ties, but often up to the limit of 0.8 per cent.; the failure of previous investigators to discover it being due to faulty a.n.a.lytical methods. Caro has seen one specimen of (bad) carbide which gave a spontaneously inflammable gas although it contained only traces of phosphine; its inflammability being caused by 2.1 per cent. of hydrogen silicide.

Practically speaking, all the foregoing remarks made about phosphine apply equally to hydrogen silicide: it burns to solid silicon oxide (silica) at the burners, is insoluble in water, and is spontaneously inflammable when alone or only slightly diluted, but never occurs in good carbide in sufficient proportion to render the acetylene itself inflammable. According to Caro the silicon may be present both as hydrogen silicide and as silicon "compounds." A high temperature in the generator will favour the production of the latter; an apparatus in which the gas is washed well in lime-water will remove the bulk of the former.

Fraenkel has found that magnesium silicide is not decomposed by water or an alkaline solution, but that dilute hydrochloric acid acts upon it and spontaneously inflammable hydrogen silicide results. If it may be a.s.sumed that the other silicides in commercial calcium carbide also behave in this manner it is plain that hydrogen silicide cannot occur in crude acetylene unless the gas is supposed to be hurried out of the generator before the alkaline water therein has had time to decompose any traces of the hydrogen silicide which is produced in the favouring conditions of high temperature sometimes prevailing. Mauricheau-Beaupre has failed to find silica in the products of combustion of acetylene from carbide of varying degrees of purity. He found, however, that a mixture of strong nitric and hydrochloric acids (_aqua regia_), if contaminated with traces of phosphoric acid, dissolved silica from the gla.s.s of laboratory vessels. Consequently, since phosphoric acid results from the phosphine in crude acetylene when the gas is pa.s.sed through aqua regia, silica may be found on subsequently evaporating the latter. But this, silica, he found, was derived from the gla.s.s and not through the oxidation of silicon compounds in the acetylene. It is possible that some of the earlier observers of the occurrence of silicon compounds in crude acetylene may have been misled by the solution of silica from the gla.s.s vessels used in their investigations. The improbability of recognisable quant.i.ties of silicon compounds occurring in acetylene in any ordinary conditions of generation is demonstrated by a recent study by Fraenkel of the composition of the deposit produced on reflectors exposed to the products of combustion of a sample of acetylene which afforded a haze when burnt. The deposit contained 51.07 per cent. of phosphoric acid, but no silica. The gas itself contained from 0.0672 to 0.0837 per cent. by volume of phosphine.

PURIFYING MATERIALS.--When acetylene first began to be used as a domestic illuminant, most generator builders denied that there was any need for the removal of these carbide impurities from the gas, some going so far as to a.s.sert that their apparatus yielded so much purer an acetylene than other plant, where purification might be desirable, that an addition of a special purifier was wholly unnecessary. Later on the more responsible members of the trade took another view, but they attacked the problem of purification in a perfectly empirical way, either employing some purely mechanical scrubber filled with some moist or dry porous medium, or perhaps with c.o.ke or the like wetted with dilute acid, or they simply borrowed the processes adopted in the purification of coal-gas. At first sight it might appear that the more simple methods of treating coal-gas should be suitable for acetylene; since the former contains two of the impurities--sulphuretted hydrogen and ammonia--characteristic of crude acetylene. After removing the ammonia by washing with water, therefore, it was proposed to extract the sulphur by pa.s.sing the acetylene through that variety of ferric hydroxide (hydrated oxide of iron) which is so serviceable in the case of coal-gas. The idea, however, was quite unsound: first, because it altogether ignores the phosphorus, which is the most objectionable impurity in acetylene, but is not present in coal- gas; secondly, because ferric hydroxide is used on gasworks to extract in a marketable form the sulphur which occurs as sulphuretted hydrogen, and true sulphuretted hydrogen need not exist in well-generated and well- washed acetylene to any appreciable extent; thirdly, because ferric hydroxide is not employed by gasmakers to remove sulphur compounds (this is done with lime), being quite incapable of extracting them, or the a.n.a.logous sulphur compounds of crude acetylene.

About the same time three other processes based on somewhat better chemical knowledge were put forward. Pictet proposed leading the gas through a strong solution of calcium chloride and then through strong sulphuric acid, both maintained at a temperature of -20 to -40 C., finally washing the gas in a solution of some lead salt. Proof that such treatment would remove phosphorus to a sufficient degree is not altogether satisfactory; but apart from this the necessity of maintaining such low temperatures, far below that of the coldest winter's night, renders the idea wholly inadmissible for all domestic installations.

Willgerodt suggested removing sulphuretted hydrogen by means of pota.s.sium hydroxide (caustic potash), then absorbing the phosphine in bromine water. For many reasons this process is only practicable in the laboratory. Berge and Reychler proposed extracting both sulphuretted hydrogen and phosphine in an acid solution of mercuric chloride (corrosive sublimate). The poisonousness of this latter salt, apart from all other objections, rules such a method out.

BLEACHING POWDER.--The next idea, first patented by Smith of Aberdeen, but fully elaborated by Lunge and Cedercreutz, was to employ bleaching- powder [Footnote: Bleaching-powder is very usually called chloride of lime; but owing to the confusion which is constantly arising in the minds of persons imperfectly acquainted with chemistry between chloride of lime and chloride of calcium--two perfectly distinct bodies--the less ambiguous expression "bleaching-powder" will be adopted here.] either in the solid state or as a liquid extract. The essential const.i.tuent of bleaching-powder from the present aspect is calcium hypochlorite, which readily oxidises sulphuretted hydrogen, and more particularly phosphine, converting them into sulphuric and phosphoric acids, while the acetylene is practically unattacked. In simple purifying action the material proved satisfactory; but since high-grade commercial bleaching-powder contains some free chlorine, or some is set free from it in the purifier under the influence of the pa.s.sing gas, the issuing acetylene was found to contain chlorine, free or combined; and this, burning eventually to hydrochloric acid, is hardly less harmful than the original sulphur compounds.

Moreover, a mixture of acetylene, chlorine, and air is liable to catch fire of itself when exposed to bright sunlight; and therefore the use of a bleaching-powder purifier, or rather the recharging thereof, was not unattended by danger in the early days. To overcome these defects, the very natural process was adopted of diluting the bleaching-powder, such diluent also serving to increase the porosity of the material. A very unsuitable substance, however, was selected for the purpose, viz., sawdust, which is hygroscopic organic, and combustible. Owing to the exothermic chemical action between the impurities of the acetylene and the bleaching-powder, the purifying ma.s.s became heated; and thus not only were the phenomena found in a bad generator repeated in the purifying vessel, but in presence of air and light (as in emptying the purifier), the reaction proceeded so rapidly that the heat caused inflammation of the sawdust and the gas, at least on one occasion an actual fire taking place which created much alarm and did some little damage. For a time, naturally, bleaching-powder was regarded as too dangerous a material to be used for the purification of crude acetylene; but it was soon discovered that danger could be avoided by employing the substance in a proper way.

HERATOL, FRANKOLINE, ACAGINE AND PURATYLENE.--Setting aside as unworthy of attention certain compositions offered as acetylene purifying materials whose const.i.tution has not been divulged or whose action has not been certified by respectable authority, there are now three princ.i.p.al chemical reagents in regular use. Those are chromic acid, cuprous chloride (sub- or proto-chloride of copper), and bleaching- powder. Chromic acid is employed in the form of a solution acidified with acetic or hydrochloric acid, which, in order to obtain the advantages (_see_ below) attendant upon the use of a solid purifying material, is absorbed in that highly porous and inert description of silica known as infusorial earth or "kieselguhr." This substance was first recommended by Ullmann, and is termed commercially "heratol" As sold it contains somewhere about 136 grammes of chromic acid per kilo. Cuprous chloride is used as a solution in strong hydrochloric acid mixed with ferric chloride, and similarly absorbed in kieselguhr. From the name of its proposer, this composition is called "frankoline." It will be shown in Chapter VI. that the use of metallic copper in the construction of acetylene apparatus is not permissible or judicious, because the gas is liable to form therewith an explosive compound known as copper acetylide; it might seem, therefore, that the employment of a copper salt for purification courts accident. The objection is not sound, because the acetylide is not likely to be produced except in the presence of ammonia; and since frankoline is a highly acid product, the ammonia is converted into its chloride before any copper acetylide can be produced. As a special acetylene purifier, bleaching-powder exists in at least two chief modifications. In one, known as "acagine," it is mixed with 15 per cent.

of lead chromate, and sometimes with about the same quant.i.ty of barium sulphate; the function of the latter being simply that of a diluent, while to the lead chromate is ascribed by its inventor (Wolff) the power of retaining any chlorine that may be set free from the bleaching-powder by the reduction of the chromic acid. The utility of the lead chromate in this direction has always appeared doubtful; and recently Keppeler has argued that it can have no effect upon the chlorine, inasmuch as in the spent purifying material the lead chromate may be found in its original condition unchanged. The second modification of bleaching-powder is designated "puratylene," and contains calcium chloride and quick or slaked lime. It is prepared by evaporating to dryness under diminished pressure solutions of its three ingredients, whereby the finished material is given a particularly porous nature.

It will be observed that both heratol and frankoline are powerfully acid, whence it follows they are capable of extracting any ammonia that may enter the purifier; but for the same reason they are liable to act corrosively upon any metallic vessel in which they are placed, and they therefore require to be held in earthenware or enamelled receivers. But since they are not liquid, the casing of the purifier can be safely constructed of steel or cast iron. Puratylene also removes ammonia by virtue of the calcium chloride in it. Acagine would probably pa.s.s the ammonia; but this is no real objection, as the latter can be extracted by a preliminary washing in water. Heratol changes, somewhat obscurely, in colour as it becomes spent, its original orange tint, due to the chromic acid, altering to a dirty green, characteristic of the reduced salts of chromium oxide. Frankoline has been a.s.serted to be capable of regeneration or revivification, _i.e._, that when spent it may be rendered fit for further service by being exposed to the air for a time, as is done with gas oxide; this, however, may be true to some extent with the essential const.i.tuents of frankoline, but the process is not available with the commercial solid product. Of all these materials, heratol is the most complete purifier of acetylene, removing phosphorus and sulphur most rapidly and thoroughly, and not appreciably diminishing in speed or efficiency until its chromic acid is practically quite used up. On the other hand, heratol does act upon pure acetylene to some extent; so that purifiers containing it should be small in size and frequently recharged. In one of his experiments Keppeler found that 13 per cent. of the chromic acid in heratol was wasted by reacting with acetylene. As this waste of chromic acid involves also a corresponding loss of gas, small purifiers are preferable, because at any moment they only contain a small quant.i.ty of material capable of attacking the acetylene itself. Frankoline is very efficacious as regards the phosphorus, but it does not wholly extract the sulphur, leaving, according to Keppeler, from 0.13 to 0.20 gramme of the latter in every cubic metre of the gas. It does not attack acetylene itself; and if, owing to its free hydrochloric acid, it adds any acid vapours to the purified gas, these vapours may be easily removed by a subsequent pa.s.sage through a vessel containing lime or a carbide drier. Both being essentially bleaching-powder, acagine and puratylene are alike in removing phosphorus to a satisfactory degree; but they leave some sulphur behind. Acagine evidently attacks acetylene to a slight extent, as Keppeler has found 0.2 gramme of chlorine per cubic metre in the issuing gas.

Although some of these materials attack acetylene slightly, and some leave sulphur in the purified gas, they may be all considered reasonably efficient from the practical point of view; for the loss of true acetylene is too small to be noticeable, and the quant.i.ty of sulphur not extracted too trifling to be harmful or inconvenient. They may be valued, accordingly, mainly by their price, proper allowance being made for the quant.i.ty of gas purified per unit weight of substance taken. This quant.i.ty of gas must naturally vary with the proportion of phosphorus and sulphur in the crude acetylene; but on an average the composition of unpurified gas is what has already been given above, and so the figures obtained by Keppeler in his investigation of the subject may be accepted.

In the annexed table these are given in two forms: (1) the number of litres of gas purified by 1 kilogramme of the substance, (2) the number of cubic feet purified per lb. It should be noted that the volumes of gas refer to a laboratory degree of purification; in practice they may all be increased by 10 or possibly 20 per cent.

_________________________________________________ | | | | | | Litres | Cubic Feet | | | per Kilogramme. | per Lb. | |______________|___________________|______________| | | | | | Heratol | 5,000 | 80 | | Frankoline | 9,000 | 144 | | Puratylene | 10,000 | 160 | | Acagine | 13,000 | 208 | |______________|___________________|______________|

Another method of using dry bleaching-powder has been proposed by Pfeiffer. He suggests incorporating it with a solution of some lead salt, so that the latter may increase the capacity of the calcium hypochlorite to remove sulphur. a.n.a.lytical details as to the efficiency of this process have not been given. During 1901 and 1902 Bullier and Maquenne patented a substance made by mixing bleaching-powder with sodium sulphate, whereby a double decomposition occurs, sodium hypochlorite, which is equally efficient with calcium hypochlorite as a purifying material, being produced together with calcium sulphate, which, being identical with plaster of Paris, sets into a solid ma.s.s with the excess of water present, and is claimed to render the whole more porous. This process seemed open to objection, because Blagden had shown that a solution of sodium hypochlorite was not a suitable purifying reagent in practice, since it was much more liable to add chlorine to the gas than calcium hypochlorite. The question how a solidified modification of sodium hypochlorite would behave in this respect has been investigated by Keppeler, who found that the Bullier and Maquenne material imparted more chlorine to the gas which had traversed it than other hypochlorite purifying agents, and that the partly foul material was liable to cause violent explosions. About the same time Rossel and Landriset pointed out that purification might be easily effected in all generators of the carbide-to-water pattern by adding to the water of the generator itself a quant.i.ty of bleaching-powder equivalent to 5 to 20 grammes for every 1 kilogramme of carbide decomposed, claiming that owing to the large amount of liquid present, which is usually some 4 litres per kilogramme of carbide (0.4 gallon per lb.), no nitrogen chloride could be produced, and that owing to the dissolved lime in the generator, chlorine could not be added to the gas. The process is characterised by extreme simplicity, no separate purifier being needed, but it has been found that an introduction of bleaching-powder in the solid condition is liable to cause an explosive combination of acetylene and chlorine, while the use of a solution is attended by certain disadvantages. Granjon has proposed impregnating a suitable variety of wood charcoal with chlorine, with or without an addition of bleaching-powder; then grinding the product to powder, and converting it into a solid porous ma.s.s by the aid of cement.

The material is claimed to last longer than ordinary hypochlorite mixtures, and not to add chlorine to the acetylene.

SUBSIDIARY PURIFYING MATERIALS.--Among minor reagents suggested as purifying substances for acetylene may be mentioned pota.s.sium permanganate, barium peroxide, pota.s.sium bichromate, sodium plumbate and a.r.s.enious oxide. According to Benz the first two do not remove the sulphuretted hydrogen completely, and oxidise the acetylene to some extent; while pota.s.sium bichromate leaves some sulphur and phosphorus behind in the gas. Sodium plumbate has been suggested by Morel, but it is a question whether its action on the impurities would not be too violent and whether it would be free from action on the acetylene itself. The use of a.r.s.enious oxide dissolved in a strong acid, and the solution absorbed in pumice or kieselguhr has been protected by G. F. Jaubert. The phosphine is said to combine with the a.r.s.enic to form an insoluble brownish compound. In 1902 Javal patented a mixture of 1 part of pota.s.sium permanganate, 5 of "sulphuric acid," and 1 of water absorbed in 4 parts of infusorial earth. The acid constantly neutralised by the ammonia of the crude gas is as constantly replaced by fresh acid formed by the oxidation of the sulphuretted hydrogen; and this free acid, acting upon the permanganate, liberates manganese peroxide, which is claimed to destroy the phosphorus and sulphur compounds present in the crude acetylene.

ePUReNE.--A purifying material to which the name of epurene has been given has been described, by Mauricheau-Beaupre, as consisting of a mixture of ferric chloride and ferric oxide in the proportion of 2 molecules, or 650 parts, of the former with one molecule, or 160 parts, of the latter, together with a suitable quant.i.ty of infusorial earth. In the course of preparation, however, 0.1 to 0.2 per cent. of mercuric chloride is introduced into the material. This mercuric chloride is said to form an additive compound with the phosphine of the crude acetylene, which compound is decomposed by the ferric chloride, and the mercuric chloride recovered. The latter therefore is supposed to act only as a carrier of the phosphine to the ferric chloride and oxide, by which it is oxidised according to the equation:

8Fe_2Cl_6 + 4Fe_2O_3 + 3PH_3 = 12Fe_2Cl_4 + 3H_3PO_4.

Thus the ultimate products are phosphoric acid and ferrous chloride, which on exposure to air is oxidised to ferric chloride and oxide. It is said that this revivification of the fouled or spent epurene takes place in from 20 to 48 hours when it is spread in the open in thin layers, or it may be partially or wholly revivified _in situ_ by adding a small proportion of air to the crude acetylene as it enters the purifier. The addition of 1 to 2 per cent. of air, according to Mauricheau-Beaupre, suffices to double the purifying capacity of one charge of the material, while a larger proportion would achieve its continuous revivification.

epurene is said to purify 10,000 to 11,000 litres of crude acetylene per kilogramme, or, say, 160 to 176 cubic feet per pound, when the acetylene contains on the average 0.05 per cent, by volume of phosphine.

For employment in all acetylene installations smaller than those which serve complete villages, a solid purifying material is preferable to a liquid one. This is partly due to the extreme difficulty of subdividing a stream of gas so that it shall pa.s.s through a single ma.s.s of liquid in small enough bubbles for the impurities to be removed by the time the gas arrives at the surface. This time cannot be prolonged without increasing the depth of liquid in the vessel, and the greater the depth of liquid, the more pressure is consumed in forcing the gas through it. Perfect purification by means of fluid reagents unattended by too great a consumption of pressure is only to be effected by a mechanical scrubber such as is used on coal-gas works, wherein, by the agency of external power, the gas comes in contact with large numbers of solid surfaces kept constantly wetted; or by the adoption of a tall tower filled with porous matter or hollow b.a.l.l.s over which a continuous or intermittent stream of the liquid purifying reagent is made to trickle, and neither of these devices is exactly suited to the requirements of a domestic acetylene installation. When a solid material having a proper degree of porosity or aggregation is selected, the stream of gas pa.s.sing through it is broken up most thoroughly, and by employing several separate layers of such material, every portion of the gas is exposed equally to the action of the chemical reagent by the time the gas emerges from the vessel. The amount of pressure so consumed is less than that in a liquid purifier where much fluid is present; but, on the other hand, the loss of pressure is absolutely constant at all times in a liquid purifier, provided the head of liquid is maintained at the same point. A badly chosen solid purifying agent may exhibit excessive pressure absorption as it becomes partly spent. A solid purifier, moreover, has the advantage that it may simultaneously act as a drier for the gas; a liquid purifier, in which the fluid is mainly water, obviously cannot behave in a similar fashion For thorough purification it is necessary that the gas shall actually stream through the solid material; a mere pa.s.sage over its surface is neither efficient nor economical of material.

DISPOSITION OF PURIFYING MATERIAL.--Although much has been written, and some exaggerated claims made, about the maximum, volume of acetylene a certain variety of purifying material will treat, little has been said about the method in which such a material should be employed to obtain the best results. If 1 lb. of a certain substance will purify 200 cubic feet of normal crude acetylene, that weight is sufficient to treat the gas evolved from 40 lb. of carbide; but it will only do so provided it is so disposed in the purifier that the gas does not pa.s.s through it at too high a speed, and that it is capable of complete exhaustion. In the coal- gas industry it is usually a.s.sumed that four layers of purifying material, each having a superficial area of 1 square foot, are the minimum necessary for the treatment of 100 cubic feet of gas per hour, irrespective of the nature of the purifying material and of the impurity it is intended to extract. If there is any sound basis for this generalization, it should apply equally to the purification of acetylene, because there is no particular reason to imagine that the removal of phosphine by a proper substance should occur at an appreciably different speed from the removal of carbon dioxide, sulphuretted hydrogen, and carbon bisulphide by lime, ferric oxide, and sulphided lime respectively, Using the coal gas figures, then, for every 10 cubic feet of acetylene generated per hour, a superficial area of (4 x 144 / 10) 57.6 square inches of purifying material is required. In the course of Keppeler's research upon different purifying materials it is shown that 400 grammes of heratol, 360 grammes of frankoline, 250 grammes of acagine, and 230 grammes of puratylene each occupy a s.p.a.ce of 500 cubic centimetres when loosely loaded into a purifying vessel, and from these data, the following table has been calculated:

__________________________________________________________ | | | | | | | Weight | Weight | Cubic Inches | | | per Gallon | per Cubic Foot | Occupied | | | in Lbs. | in Lbs. | per Lb. | |_____________|____________|________________|______________| | | | | | | Water | 10.0 | 62.321 | 27.73 | | Heratol | 8.0 | 49.86 | 31.63 | | Frankoline | 7.2 | 41.87 | 38.21 | | Acagine | 6.0 | 31.16 | 55.16 | | Puratylene | 4.6 | 28.67 | 60.28 | |_____________|____________|________________|______________|

As regards the minimum weight of material required, data have been given by Pfleger for use with puratylene. He states that 1 Kilogramme of that substance should be present for every 100 litres of crude acetylene evolved per hour, 4 kilogrammes being the smallest quant.i.ty put into the purifier. In English units these figures are 1 lb. per 1.5 cubic feet per hour, with 9 lb. as a minimum, which is competent to treat 1.1 cubic feet of gas per hour. Thus it appears that for the purification of the gas coming from any generator evolving up to 14 cubic feet of acetylene per hour a weight of 9 lb of puratylene must be charged into the purifier, which will occupy (60.28 / 9) 542 cubic inches of s.p.a.ce; and it must be so spread out as to present a total superficial area of (4 x 144 x 14 / 100) 80.6 square inches to the pa.s.sing gas. It follows, therefore, that the material should be piled to a depth of (542 / 80.6) 6.7 inches on a support having an area of 80.6 square inches; but inasmuch as such a depth is somewhat large for a small vessel, and as several layers are better than one, it would be preferable to spread out these 540 cubic inches of substance on several supports in such a fashion that a total surface of 80.6 square inches or upwards should be exhibited. These figures may obviously be manipulated in a variety of ways for the design of a purifying vessel; but, to give an example, if the ordinary cylindrical shape be adopted with four circular grids, each having a clear diameter of 8 inches (_i.e._, an area of 50.3 square inches), and if the material is loaded to a depth of 3 inches on each, there would be a total volume of (50.3 x 3 x 4) = 604 cubic inches of puratylene in the vessel, and it would present a total area of (50.3 x 4) = 201 square inches to the acetylene. At Keppeler's estimation such an amount of puratylene should weigh roughly 10 lb., and should suffice for the purification of the gas obtained from 320 lb. of ordinary carbide; while, applying the coal-gas rule, the total area of 201 square inches should render such a vessel equal to the purification of acetylene pa.s.sing through it at a speed not exceeding (201 / 5.76) = 35 cubic feet per hour. Remembering that it is minimum area in square inches of purifying material that must govern the speed at which acetylene may be pa.s.sed through a purifier, irrespective probably of the composition of the material; while it is the weight of material which governs the ultimate capacity of the vessel in terms of cubic feet of acetylene or pounds of carbide capable of purification, these data, coupled with Keppeler's efficiency table, afford means for calculating the dimensions of the purifying vessel to be affixed to an installation of any desired number of burners. There is but little to say about the design of the vessel from the mechanical aspect. A circular horizontal section is more likely to make for thorough exhaustion of the material. The grids should be capable of being lifted out for cleaning. The lid may be made tight either by a clamp and rubber or leather washer, or by a liquid seal. If the purifying material is not hygroscopic, water, calcium chloride solution, or dilute glycerin may be used for sealing purposes; but if the material, or any part of it, does absorb water, the liquid in the seal should be some non-aqueous fluid like lubricating oil. Clamped lids are more suitable for small purifiers, sealed lids for large vessels. Care must be taken that condensation products cannot collect in the purifying vessel. If a separate drying material is employed in the same purifier the s.p.a.ce it takes must be considered separately from that needed by the active chemical reagent. When emptying a foul purifier it should be recollected that the material may be corrosive, and being saturated with acetylene is likely to catch fire in presence of a light.

Purifiers charged with heratol are stated, however, to admit of a more rapid flow of the gas through them than that stated above for puratylene.

The ordinary allowance is 1 lb. of heratol for every cubic foot per hour of acetylene pa.s.sing, with a minimum charge of 7 lb. of the material. As the quant.i.ty of material in the purifier is increased, however, the flow of gas per hour may be proportionately increased, _e.g._, a purifier charged with 132 lb. of heratol should purify 144 cubic feet of acetylene per hour.

In the systematic purification of acetylene, the practical question arises as to how the attendant is to tell when his purifiers approach exhaustion and need recharging; for if it is undesirable to pa.s.s crude gas into the service, it is equally undesirable to waste so comparatively expensive a material as a purifying reagent. In Chapter XIV. it will be shown that there are chemical methods of testing for the presence, or determining the proportion, of phosphorus and sulphur in acetylene; but these are not suitable for employment by the ordinary gas-maker. Heil has stated that the purity of the gas may be judged by an inspection of its atmospheric flame as given by a Bunsen burner. Pure acetylene gives a perfectly transparent moderately dark blue flame, which has an inner cone of a pale yellowish green colour; while the impure gas yields a longer flame of an opaque orange-red tint with a bluish red inner zone. It should be noted, however, that particles of lime dust in the gas may cause the atmospheric flame to be reddish or yellowish (by presence of calcium or sodium) quite apart from ordinary impurities; and for various other reasons this appearance of the non-luminous flame is scarcely to be relied upon. The simplest means of ascertaining definitely whether a purifier is sufficiently active consists in the use of the test-papers prepared by E. Merck of Darmstadt according to G. Keppeler's prescription. These papers, cut to a convenient size, are put up in small books from which they may be torn one at a time. In order to test whether gas is sufficiently purified, one of the papers is moistened with hydrochloric acid of 10 per cent. strength, and the gas issuing from a pet-c.o.c.k or burner orifice is allowed to impinge on the moistened part.

The original black or dark grey colour of the paper is changed to white if the gas contains a notable amount of impurity, but remains unchanged if the gas is adequately purified. The paper consists of a specially prepared black porous paper which has been dipped in a solution of mercuric chloride (corrosive sublimate) and dried. Moistening the paper with hydrochloric acid provides in a convenient form for application Berge's solution for the detection of phosphine (_vide_ Chapter XIV.). The Keppeler test-papers turn white when the gas contains either ammonia, phosphine, siliciuretted hydrogen, sulphuretted hydrogen or organic sulphur compounds, but with carbon disulphide the change is slow.

Thus the paper serves as a test for all the impurities likely to occur in acetylene. The sensitiveness of the test is such that gas containing about 0.15 milligramme of sulphur, and the same amount of phosphorus, per litre (= 0.0655 grain per cubic foot) imparts in five minutes a distinct white mark to the moistened part of the paper, while gas containing 0.05 milligramme of sulphur per litre (= 0.022 grain per cubic foot) gives in two minutes a dull white mark visible only by careful inspection. If, therefore, a distinct white mark appears on moistened Keppeler paper when it is exposed for five minutes to a jet of acetylene, the latter is inadequately purified. If the gas has pa.s.sed through a purifier, this test indicates that the material is not efficient, and that the purifier needs recharging. The moistening of the Keppeler paper with hydrochloric acid before use is essential, because if not acidified the paper is marked by acetylene itself. The books of Keppeler papers are put up in a case which also contains a bottle of acid for moistening them as required and are obtainable wholesale of E. Merek, 16 Jewry Street, London, E.C., and retail of the usual dealers in chemicals. If Keppeler's test-papers are not available, the purifier should be recharged as a matter of routine as soon as a given quant.i.ty of carbide--proportioned to the purifying capacity of the charge of purifying material--has been used since the last recharging. Thus the purifier may conveniently contain enough material to purify the gas evolved from two drums of carbide, in which case it would need recharging when every second drum of carbide is opened.

REGULATIONS AS TO PURIFICATION.--The British Acetylene a.s.sociation has issued the following set of regulations as to purifying material and purifiers for acetylene:

Efficient purifying material and purifiers shall comply with the following requirements:

(1) The purifying material shall remove phosphorus and sulphur compounds to a commercially satisfactory degree; _i.e._, not to a greater degree than will allow easy detection of escaping gas through its odour.

(2) The purifying material shall not yield any products capable of corroding the gas-mains or fittings.

(3) The purifying material shall, if possible, be efficient as a drying agent, but the a.s.sociation does not consider this an absolute necessity.

(4) The purifying material shall not, under working conditions, be capable of forming explosive compounds or mixtures. It is understood, naturally, that this condition does not apply to the unavoidable mixture of acetylene and air formed when recharging the purifier.

(5) The apparatus containing the purifying material shall be simple in construction, and capable of being recharged by an inexperienced person without trouble. It shall be so designed as to bring the gas into proper contact with the material.

(6) The containers in purifiers shall be made of such materials as are not dangerously affected by the respective purifying materials used.

(7) No purifier shall be sold without a card of instructions suitable or hanging up in some convenient place. Such instructions shall be of the most detailed nature, and shall not presuppose any expert knowledge whatever on the part of the operator.

Reference also to the abstracts of the official regulations as to acetylene installations in foreign countries given in Chapter IV. will show that they contain brief rules as to purifiers.

DRYING.--It has been stated in Chapter III. that the proper position for the chemical purifiers of an acetylene plant is after the holder; and they therefore form the last items in the installation unless a "station"

governor and meter are fitted. It is therefore possible to use them also to remove the moisture in the gas, if a material hygroscopic in nature is employed to charge them. This should be true more particularly with puratylene, which contains a notable proportion of the very hygroscopic body calcium chloride. If a separate drier is desirable, there are two methods of charging it. It may be filled either with some hygroscopic substance such as porous calcium chloride or quicklime in very coa.r.s.e powder, which retains the water by combining with it; or the gas may be led through a vessel loaded with calcium carbide, which will manifestly hold all the moisture, replacing it by an equivalent quant.i.ty of (unpurified) acetylene. The objection is sometimes urged against this latter method, that it restores to the gas the nauseous odour and the otherwise harmful impurities it had more or less completely lost in the purifiers; but as regards the first point, a nauseous odour is not, as has previously been shown, objectionable in itself, and as regards the second, the amount of impurities added by a carbide drier, being strictly limited by the proportion of moisture in the damp gas, is too small to be noticeable at the burners or elsewhere. As is the case with purification, absolute removal of moisture is not called for; all that is needed is to extract so much that the gas shall never reach its saturation-point in the inaccessible parts of the service during the coldest winter's night.

Any accessible length of main specially exposed to cold may be safeguarded by itself; being given a steady fall to a certain point (preferably in a frost-free situation), and there provided with a collecting-box from which the deposited liquid can be removed periodically with a pump or otherwise.

FILTRATION.--The gas issuing from the purifier or drier is very liable to hold in suspension fine dust derived from the purifying or drying material used. It is essential that thin dust should be abstracted before the gas reaches the burners, otherwise it will choke the orifices and prevent them functioning properly. Consequently the gas should pa.s.s through a sufficient layer of filtering material after it has traversed the purifying material (and drier if one is used). This filtering material may be put either as a final layer in the purifier (or drier), or in a separate vessel known as a filter. Among filtering materials in common use may be named cotton-wool, fine canvas or gauze, felt and asbestos-wool. The gas must be fairly well dried before it enters the filter, otherwise the latter will become choked with deposited moisture, and obstruct the pa.s.sage of the gas.

Having now described the various items which go to form a well-designed acetylene installation, it may be useful to recapitulate briefly, with the object of showing the order in which they should be placed. From the generator the gas pa.s.ses into a condenser to cool it and to remove any tarry products and large quant.i.ties of water. Next it enters a washing apparatus filled with water to extract water-soluble impurities. If the generator is of the carbide-to-water pattern, the condenser may be omitted, and the washer is only required to retain any lime froth and to act as a water-seal or non-return valve. If the generator does not wash the gas, the washer must be large enough to act efficiently as such, and between it and the condenser should be put a mechanical filter to extract any dust. From the washer the acetylene travels to the holder. From the holder it pa.s.ses through one or two purifiers, and from there travels to the drier and filter. If the holder does not throw a constant pressure, or if the purifier and drier are liable to cause irregularities, a governor or pressure regulator must be added after the drier. The acetylene is then ready to enter the service; but a station meter (the last item in the plant) is useful as giving a means of detecting any leak in the delivery-pipes and in checking the make of gas from the amount of carbide consumed. If the gas is required for the supply of a district, a station meter becomes quite necessary, because the public lamps will be fed with gas at a contract rate, and without the meter there would be no control over the volume of acetylene they consume. Where the gas finally leaves the generating-house, or where it enters the residence, a full-way stopc.o.c.k should be put on the main.

GENERATOR RESIDUES.--According to the type of generator employed the waste product removed therefrom may vary from a dry or moist powder to a thin cream or milk of lime. Any waste product which is quite liquid in its consistency must be completely decomposed and free from particles of calcium carbide of sensible magnitude; in the case of more solid residues, the less fluid they are the greater is the improbability (or the less is the evidence) that the carbide has been wholly spent within the apparatus. Imperfect decomposition of the carbide inside the generator not only means an obvious loss of economy, but its presence among the residues makes a careful handling of them essential to avoid accident owing to a subsequent liberation of acetylene in some unsuitable, and perhaps closed, situation. A residue which is not conspicuously saturated with water must be taken out of the generator- house into the open air and there flooded with water, being left in some uncovered receptacle for a sufficient time to ensure all the acetylene being given off. A residue which is liquid enough to flow should be run directly from the draw-off c.o.c.k of the generator through a closed pipe to the outside; where, if it does not discharge into an open conduit, the waste-pipe must be trapped, and a ventilating shaft provided so that no gas can blow back into the generator-house.

DISPOSAL OF RESIDUES.--These residues have now to be disposed of. In some circ.u.mstances they can be put to a useful purpose, as will be explained in Chapter XII.; otherwise, and always perhaps on the small scale-- certainly always if the generator overheats the gas and yields tar among the spent lime--they must be thrown into a convenient place. It should be remembered that although methods of precipitating sewage by adding lime, or lime water, to it have frequently been used, they have not proved satisfactory, partly because the sludge so obtained is peculiarly objectionable in odour, and partly because an excess of lime yields an effluent containing dissolved lime, which among other disadvantages is harmful to fish. The plan of running the liquid residues of acetylene manufacture into any local sewerage system which may be found in the neighbourhood of the consumer's premises, therefore, is very convenient to the consumer; but is liable to produce complaints if the sewage is afterwards treated chemically, or if its effluent is pa.s.sed untreated into a highly preserved river; and the same remark applies in a lesser degree if the residues are run into a private cesspool the liquid contents of which automatically flow away into a stream. If, however, the cesspool empties itself of liquid matter by filtration or percolation through earth, there can be no objection to using it to hold the lime sludge, except in so far as it will require more frequent emptying. On the whole, perhaps the best method of disposing of these residues is to run them into some open pit, allowing the liquid to disappear by evaporation and percolation, finally burying the solid in some spot where it will be out of the way. When a large carbide-to-water generator is worked systematically so as to avoid more loss of acetylene by solution in the excess of liquid than is absolutely necessary, the liquid residues coming from it will be collected in some ventilated closed tank where they can settle quietly. The clear lime-water will then be pumped back into the generator for further use, and the almost solid sludge will be ready to be carried to the pit where it is to be buried. Special care must be taken in disposing of the residues from a generator in which oil is used to control evolution of gas. Such oil floats on the aqueous liquid; and a very few drops spread for an incredible distance as an exceedingly thin film, causing those brilliant rainbow-like colours which are sometimes imagined to be a sign of decomposing organic matter. The liquid portions of these residues must be led through a pit fitted with a depending part.i.tion projecting below the level at which the water is constantly maintained; all the oil then collects on the first side of the part.i.tion, only water pa.s.sing underneath, and the oil may be withdrawn and thrown away at intervals.

CHAPTER VI