THE PRODUCTION AND TREATMENT OF VEGETABLE OILS T. W. CHALMERS E.Sn. A.M.I.Meoh. '/^^<«^^ •/'' i' ■ y / y • ^ -*/ IH^ P65c) I .r /L I (Lite B. M, ilm ICilirary ^ortlr Carnliita Stale CoIIeoe TP680 ^C4 't,'^ ^^;? \ ^4i S00589405 V Date Due \|Ci>ni May9'32 l:![%;3xZJft!«Mi' 19 Ian 3^ m ^lAii v9»i^ 25Feti^ Jj^-'tk! X^^yhoAJ <-4^, ■^.^■■Jh> IZkifft^W^j f iWIar'90;,' ir?far 5 Feb"* 2FMay34. _9JiajL 11^ ^4I& '^Ap'40F te- 22Mtt3 5 ''nV!y'40iV q 7Feb59 9Apr'36 OJA M)^ -^^'^^f^; ;3Mt42Ftf ^5^ay59t nNo'42K| ■ tAi'J :.S '4pt I J \?43- THE ENGINEER SERIES THE PRODUCTION AND TREATMENT OF VEGETABLE OILS ■THE ENGINEER SERIES THE PRODUCTION AND TREATMENT OF VEGETABLE OILS INCLUDING CHAPTERS ON THF, REFINING OF OILS, THE HYDROGENATION OF OILS, THE GENERATION OF HYDROGEN, SOAP MAKING, THE RECO\'ERY AND REFINING OF GLYCERINE, AND THE SPLITTING OF OILS *^ T. W^ CHALMERS, B.Sc, A.M.I.Mech.K. (On the Editorial Staff of "The Engineer") WITH NINE FOLDING PLATES AND 95 ILLUSTRATIONS IN THE TEXT LONDON CONSTABLE & COMPANY' LTD 10-12 ORANGE STREET LEICESTER SOUARE WC 2 1920 NEW YORK D. VAN NOSTRAND COMPANY EIGHT WARREN STREET First ritblished . lieprintcd . litis. I'.Hit. 1 '.ILM. Priiite'l ill dreat Ilrilaiii. PREFACE In this volume an attempt is made to deal with the production and treatment of vegetable oils primarily from the engineer's point of view, an aspect of the industry which hitherto has received in print scanty consideration as compared with the attention paid to the chemist's side of the matter. Everything connected with the recovery and treatment of vegetable oils has received a great stimulus from the conditions brought about by the war. The indu.stry is spreading or showing signsof spreading in many directions, and great as its importance in this country has been in the past, it is safe to prophesy that in the immediate future it will be greatci' still. It is hoped that this volume will do something towards assisting those interested or lilcely to become interested in the industry to understand the construction and working of the principal machines and plant which it depends upon. Sufficient information, it is believed, is given regarding the chemical and commercial aspects of the matter to make the book, although written from the engineer's standpoint, a more or less general treatise on the vegetable oil industry. The chapters which follow originally appeared as a series of articles * in The. Engineer. In planning this series it was at one time hoped to include in it sections devoted to certain aspects of the industrial employment of vegetable oils, notably on the employment of such oils in the manufacture of linoleum and margarine. While there is much of engineering interest in both these branches of industry it \\'as found that a consideraljle degree of secrecy was preserved regarding the machinery employed in the former, while the machinery for the latter came almost, if not quite, exclusively from abroad, notably from Holland. The.se reasons and the exigencies of space, compelled a considerable restriction in the account of the industrial employment of vegetable oils. On the other hand, much valuable infornu^tion was obtained regarding certain aspects of the production and treatment of vegetable oils, much of which infornuition, it is believed, has not hitherto been published, or has been published in an inaccurate, out-of-date, or incomplete form. In this connection special attention may perhaps be directed towards the sections dealing with the extraction of oils by means of chemical solvents, oil refuiing, oil hardening and the generation of hydrogen, the recovery and refining of glycerine, and the splitting of oils. Sincere thanks are due to all those who so courteously afforded their assistance in the preparation of the original series of articles, particularly to -Messrs. Manlove, AUiott & Co. : to Mr. H. J. Pooley, of Messrs. George Scott k Son ; and to Mr. Howard Lane. T. W. C, London, July, 1917. • "The Production anil Indu-strial ICniploynicnt of y^cgetiible Oils." Kifiliteen arlii'lcs. Pol)ni;iry 9tli to June 2'Jth, 1'JI7. 11356 CONTENTS CHAPTER I. Imtroductory a>ji) General ........... 1 CHAPTKU II. The rinxciPAi, Vf.grtable Oil? .......... CHAPTER III. Prepauatoky .Machinery for Copra an'd Linseed . . . ... . .12 (CHAPTER IV. Preparatory .Machinery ior Palm Fruit ano Palm Kiornels ..... 21 CHAPTER Y. Preparatory Machinery for Cotton Seer and Castor Seed ..... 29 CHAPTER VI Some Special Forms ok REDucriox Maciiinhry. ....... 39 CHAPTER VII. .Meal Kettles, Receiving Pans axd Moulding Machines. . . . ... 46 CHAPTER VI II. Oil Presses —Anglo-American Type .......... 56 CH.APTER IX. Oil Presses — Cage Type ............ 62 CHAPTER X. The General Arrangement of Oil Mills. ........ 73 CHAPTER XI. Extraction oi' Oils by Chemical Solvents ........ 83 CHAPTER XII. The Refining of Oils ............ 93 CHAPTER XIII. The Hydrogen ation or Hardening of t)iLs . . . . . . . .106 CHAPTER XIV. The Generation ov Hydrogen for Oil Hardening Purposes . . . . .113 ClhVPTER XV. The Manufacture of Soap . . . ■ • • • • .122 CHAPTER XVI. Glycerine Recovery and Refining and the Splitting of Oils .... 132 Index • I*'-' LIST OF ILLUSTRATIONS s. 4. 5. 6. 7. 8. i). 10. 11. 12. Ki. U. , 15. \ 16. 17. 18. 10. 20. 21. 22. 23. 24. 25.. 26. 27. 28. 2». 30. 31. 32. 33. 34. 3.'). 36. 37. 38. 38a. 3!). 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. Cocoa-nut Splitting Machine . M.\GNETIC SePAEATOK FOR COPEA, ETC. Preliminary Beeaking Machine Shredding and Crushing Rolls foe Copra, etc Shredding and Crushing Rolls Rolls foe further Reduction of Copra . Final Reduction Rolls .... Screening Machine for Linseed, etc. Rolls for Linseed, etc. .... Palm Fruit, Peric.\rp, Xuts, Shells and Kern Bunch of Palm Fruit . , . . Faikf.vx's Depericarping Machine Depericaeping Machine for Palm Fnun . Palm Xut Brushing Machine ... Palm Nut Cracking and Separating Machine Palm and Palm Kernel Oil Mill American and Egtpti.-vn Cotton Seed Cotton Seed De-linter .... Cotton Seed Decorticating Machine Castor Seed — Pods, Beans and Kernels . Castor Seed Sheller .... Castor Seed Decorticator and Sepae.\tor Castoe Seed Decorticator and Separator Castor Seed Decorticator and Separator Horizontal Seed Rolls .... Edge Runner ...... Special Grinding Mill with Concave Pl.\tes Reducing Mill and Cake Breaker . Disintegrator ...... Disintegrator, showing Beater Ar.ms, Waved Meal Kettle ...... Double Kettle ...... Kettle and Moulding Machine Receiving Pan and Moulding Machines . Ordinary Form of Hydraulic Moulding Machine Special Form of Hydraulic Moulding Machine ! Se.mi-automatic Moulding Maciune Battery of Four Anglo-American Presses Small Anglo-American Press, etc. . Two Small Cage Presses .... Cage Press and Kettle in a .Mill Twin Cage Presses . . . Twin Cage Presses ..... Battery ok Four Cage Presses Revolving Cage Press .... Cross-section of a Mill with Anglo-.V.merican Presses Plan ok Mill with Cage and Anglo-American Presses High ani) Low-pressure Oil Press Pump . Oil Press Pump ...... Oil Mill Accumulators .... Facing LIST OF ILLUSTRATIONS .">2. ACCCMI'LATOK^ IX AX OlL MlLL ^>'^. acccmcxatok pcmp . 54. Relief Valve Detail? o5. Cake-paring Machixe. 56. Htukaflic Cake-pakixg Machlse 57. Bexzese Solvext Extractiox Plaxt 58. ."^COTT SOLVEXT EXTRACTOR WITH AGIT.ITIXG GEAII 59. .r Oil .... 62. Filter Press Plate .... 63. Two Forms of Filter Plates' . 64. Washing Machixe for Filter Cloths 65. Centrifugal Extractor for " Foots " 66. Cotton Seed Oil Refixert 67. Vacuum P-\x axd Condenser 68. Cocoa-nut On, Rehnert .... 69. Stearixe Presses for Demaegarixatixg Oil 70. Plax of Oil Htdrogexisixg Factory ...... Facing 71. Eaxe Autoclave Mr. Howard Laxe's Experimental OrL-HARDEXiXG Plaxt 73. Diagram of the Lane Htdrogex Retort Furnace 7-1. Three-wat Reversixg Cock for Htdrogex Retort Furnace Experimental Htdrogen Plant Soap Kettle Crutciiing Machine . 78. Twix Crutchixg Machixe? 79. Steam Driven" Crutcher 80. Soap Frame 81. Slab Cuttixg Machixe 82. Bar Ccttdcg Machixe 83. .Soap Drtixg Plant 84. Stamping Machine 85. Chipping Machine 86. Milling Machixe 87. Soap Squeezing Machine or Plodder 88. Single-effect Vacuum Evapc>rator for Concemratixg Crude Gltcerixe 89. Single-effect Vacuum Evaporator for Concextratixg Crude Gltcerfne 90. Removing Salt from a Vacuum Evaporator 91. .\uTOMATic Salt Discharging Device 92. Four Dourle-effect Vacuum Evaporators for Treatixg 500 Toxs of Soap Lte5 per Dat ........... 93. Double-effect Vacuum Evaporator for Concentrating Crude Gltcerin 94. Oil .4ND Fat Splitting Autoclaves ....... 95. Plant for Splitting Oii.s and Fats bt the TwrrcHEU. Process 142 96. (iLTCERINE RefIXTNG AXD CONCEXTRATIXG PlaXT 97. IXTERIOR OF A GlTCERIXE ReFIXERT . PAGE 78 94 95 95 96 97 98 loo 102 IU4 108 no 111 lit; ) iv 123 124 125 126 126 127 127 12s li» 12;' 12'.' ]:><> 134 135 136 136 138 139 140 143 145 146 LIST OF PLATES I'LATE I. 400-TOK AngloAmekkan Hydraulic Oil Press . . . Facing page 5C II. Cage Type Oil Press . . . . . . ..,,,. 62 III. Revolving Cage Type Oil Press . . . . • . ,, „ 69 IV. Copra Oil Mill with Cage Type Presses . . . ...,., 75 V. Benzene Solvent Extraction Plant for Vegetable Oils . . „ „ 87 VI. Oil-Hardening Autoclave and its Details . . ..... Ill VII. TiiF. Hardening of Oils^Tiie Lane Hydrogen Retort Flrnace . ,, ,, 115 CHAPTER I INTRODUCTORY AND GENERAL 'J'heke are three distinct systems or methods whereby oils in general can be classified into groups. These may be called the popular, the scientific, and the practical. The popular sy.stem divides them into animal, vegetable, and mineral oils according to the natural Idngdom from which they are derived. This scheme of classification has little scientific value, for it is more than doubtful if we are ever justified, scientifically, in speaking of a " mineral " oil. It is practically established that oil is always an organic and never an inorganic substance, and that so-called minei'al oil, whether obtained direct from a well, or recovered by distilling shale, is merely the transformed product of animal or vegetable organisms. Nevertheless " mineral " oil is so distinct in its properties from either vegetable or animal oil proper that in most practical circumstances its separate classification is highly desirable. In popular phraseology, it may be remarked, " mineral " oil means petroleum, the raw product, or one of the substances derived from it by treatment. It is not usual to speak of the coal tar oils as " mineral "' oils, although the well-established position of coal as a mineral would certainly justify us in doing so. The second scheme of classification is the chemical. Into this we do not projiose to enter here, for its interest is at present almost purely scientific. The third or practical system classifies oil primarily into two groups, namely, essential oils and fixed oils. An essential oil is one which can be volatilised without decomposing it. A fixed oil is one which cannot be so volatilised. The fixed oils are further subdivided into two groups, the mineral oils on the one hand and the fatty oils on the other. Essential oils, as we have said, are distinguished by the fact that they can be distilled without suffering an alteration in their chemical composition. They are obtained entirely from vegetable sources, commonly by distilling the leaves, flowers, fruit, or seeds of various plants. Certain barks and roots also yield essential oils, as do amber and other resinous exudations from trees. Distillation is not, however, universal, mechanical processes and extraction by means of .solvents being sometimes adopted. Essential oils ai-e u.sed for many minor purposes. Thus the oil distilled from pine needles finds employment in the manufacture of boot polishes, while cedar- wood oil, on account of its high refractive index, is in demand for microscopic purposes. (Jeneially speaking, however, these oils may be said to be in chief use as perfumes, as flavourings, and as medicines, typical examples of the three classes being lavender oil, peppcnnint oil, and eucalyptus oil. Fixed oils, as already remarked, cannot be distilled without suffering chemical decomposition, and are divided into two gi'oups, the muieral and the fatty. The mineral group comprises petroleum and its derivatives, and is identical with the mmeral group of the popular classification. The fatty oil group, on the other hand, is not identical with the animal oil group of that classification, although it is generally true that all animal oils are fatty oils. Many fatty oils, we might even saj' the most important fatty oils, are of vegetable origin. Certain of the heavier mineral oil deriva- tives may look like fatty oils, and are used, as are some fatty oils, for lubricating and 2 THE PRODUCTION AND TREATMENT OF ^T:GETABLE OILS burning purposes. Nevertheless the two classes are radically different in their cliemical composition. The broad and most important and practical difference between them lies in the fact that fattj- oils can be converted into soaps by acting upon them witli caustic alkalis and otiier inorganic substances, whereas mineral oils cannot be saponified. Fatty oils are thus either of vegetable or animal origin. Neither " vegetable " nor " animal," as here used, is. of course, to be interpreted in a restricted sense. Speaking botanically, very few oils are obtained from vegetables, the only one, in fact, which conies readily to mind being that derived from radishes, the seeds of which yield an oil of the rape cr colza type. The majority of and the most important vege- table oils are extracted from the seeds of plants, for example, linseed, cotton seed, rape seed, hemp seed, poppy seed, and smaflower seed. Important vegetable oils are also obtained from the nuts of certain trees, such as the wabiut. cocoa-nut. and hazel nut. Other trees yield oils from their fruit, either from the fruit itself . as in the case of olive oil. or from the fruit kernels, as in the case of cherry, apricot, plum, peach and palm kernel oils. Speakmg zoologically, the term " animal oil " is more or less justified. Thus the sheep, the horse, the ox, the whale, the seal, and the porpoise are all animals and jaeld important oils. It may perhaps be remarked parenthetically that the hoof of the ox is the soui'ce of the well-knowai neats foot oil, and that from the jaw of the porpoise there is obtained a veiy valuable oil used for lubricating watch.es and other delicate machinery. It must be admitted, however, that some important "' animal "' oils are recovered from fish, notabh' from the livers of fish. Such oils find employment chiefly in currying leather, to some extent in soap making, and to a small degi-ee in medicine. The birds also su])ply " animal " oils or fats. Thus the egg of the common hen yields an oil used in leather dressmg. while fats are obtained from tl:e blackcock, duck, and goose. Even the reptile kingdom is drawai upon, foi' tl:e rattlesnake gives a fat which is used in pharmacj'. So far as we laiow the insect kuigdom is not used as a source of " animal " oil. This is so presumabty because of the difficulty of collecting insects in quantities sufficient for treatment, and because of the difficulty of treating them. It is certainly not due to the absence of oil in their composition. Thus, cochineal extracted with benzene can be made to yield an oil. Cochineal is, of course, the dried bodies of an insect which lives on a species of cactus cultivated, for the sake of the insect, in Mexico and Central Ameiica. Vegetable and animal oils are frequently closely similar in composition and behaviour. Until recently it was in general impo.ssible to determine to which class a given oil belonged solely by chemical examination. Means, however, are now available for discriminatmg chemically between the two classes. Thus, animal oils contain a certain alcohol laiown to the chemist as cholesterol. This body can be isolated from the oil recovered from sheep's wool, and is well kno\\Ti to the general public under the name of lanolm. A similar alcohol is contained in vegetable oils. This body is known as phytosterol. It appears to have the same molecular formula as cholesterol, but mider certain ciicumstances it behaves different]}'. Thus under the microscope the crystals of the two bodies are foimd to be of different shape. This and the fact th.at their acetates melt at different temperatures form a basis for di.stinguishing chemically between a vegetable and an animal oil. Apart from the existence of these two bodies in animal and vegetable oils respec- tively, the chemical composition of such oils cannot be regarded at present as being complete!}^ understood. This composition, of course, varies, and varies greatly, from oil to oil. Further, although to a lesser extent, it varies in any one oil accordmg to the soil and climate in which the plant from which it was derived was grown. INTRODUCTORY AND GENERAL 3 or according to the age, food, even the personal habits, of the animal from which it was obtamed. There need therefore be little wonder if the classification of oils on a chemical basis is, as we have stated it to be, a matter at present almost solely of scientific interest. Even in the matter of definmg what an oil in general is the chemist cannot do more than adopt the popular description of an oil as a substance, usually liquid at ordinary temperatures, which is insoluble in water, combustible, and more or less viscous. This deflniticn succeeds in that it excludes oil of vitriol — the popular term for concentrated sulphuric acid — from among the oils. It fails only in so far as it does not distinguish between an oil and a fat. There is no chemical distinction between a fatty oil and a fat. It is purely a matter of temperature, for a fatty oil when frozen becomes a fat, and a fat %\hen melted becomes a fatty oil. A particular instance of how climate affects the nomenclature is to be found in the case of cocoa-nut oil. In Lidia this substance is liquid, and is therefore to be regarded as an oil. In this country it is usually solid and should properly be spoken of as a fat. It has been proposed to adopt 20' C. as the standard temperature at wliich to judge fats and oils. The hardest fats, it may be added, melt at about 50° C, while some oils are still liquid at and below the freezing point of water. As the title of this volume indicates, we propose to describe and discuss the production and treatment of vegetable oils. It is not our mtention to deal with either essential, mmeral or animal oils. The whole field is too vast to be treated conve- niently in one volume, and, moreover, is not sufificiently well connected to have a common interest. Li fastenmg our attention upon vegetable fatty oils considerations other than these have also weighed with us. Among such considerations may be mentioned the fact that these oils, particularly those suitable for edible purposes, are now attracting attention in this country to an extent hardly contemplated before the war. As is well laio\\ii, Germany had in recent years become a very formidable rival to this coimtry in its command over the^vegetable oil industries, and had, as in Nigeria, for example, secured virtual monopolies over certain of the raw materials. These sources of supply have now been largely set free and, let us hope, will never again pass into our enemies' hands. Then agaui the war has led, as most of us laiow from experience, to an enormous increase in the demand for margarine, a very impor- tant outlet for certam varieties of fatty vegetable oils. Although much of this substi tute for butter still comes from Holland,' efforts on a gratifying scale are being made to meet the demand witii margarine produced in this country. Aparti from the margarine industry, the fatty vegetable oils constitute the raw product or one of the raw products of several important industries. Thus they are used hi the manufacture of paints and vaniishes, soap and candles, linoleum and oilcloth. In these industries the engineer plays a veiy considerable part, so that both in the production and in the industrial employment of vegetable oils much of great engmeering interest is to be found. It is from the engmeerhig standpomt, and in particular from the British engineer's standpomt, that we propose chiefly to regard the matter. In the next chapter we shall discuss the sources from which the better known vegetable oils are obtained, and the principal uses to which they are put. For the present it will be u.seful to describe without going into details the general methods adopted in their production. There are two broad methods of extracting fatty oils from vegetable products — one, that employmg pressure, being purely a mechanical process, and the other, that extracting the oil by means of solvents, being more or less a chemical process. Under the first method the seed, if small, is simply crushed in a hydraulic press. The oil 4 THE PRODUCriOX AND TREATJIENT OF VEGETABLE OILS forced out of the seed is caught and drained off. If the seed is large or if the raw material is copra or some such stuff, it is first giound up in special machmes before being crushed. The seed or seed " meal " is sometmies heated during the process of crushing. The oil then produced is known as '" hot pressed '" or " hot drawn " oil. Such oil, however, is apt to be unduly discoloured by reason of its having dissolved from the seed dm-ing expression an excessive amoiuit of colouring matter. For certain purposes, therefore, notably for edible purposes, " cold drawii "" oil is preferred. The " cold drawing "" process usuallj' leaves quite a considerable quantity of oil remaining behmd in the seed. Consequently it is a common practice to break up the cake left in the press after cold drawmg, heat it and extract a " second expression oil "' bj' the hot process. The cake left may be once again broken up, heated, and expressed a tliird time, but even so it is scarcely possible to extract more than 90 to 95 per cent, of the total oil in the seed by the crushing process alone. The second process extracts the oil from the seed or seed meal by means of chemical solvents, the seed being treated either hot or co!d^ The three cliief solvents in use are ^ benzene, carbon disulphide, and carbon tetrachloride. The process in outline consists of aUowi:ig the solvent to percolate through the seed or meal in a closed vessel, draining off the solvent and dissolved oil, transferrmg it to a heated still and there driving off the volatile solvent so as to leave the oil behind. The solvent is condensed and re-used. So far as the percentage of oil recovered from the seed is concerned, this process is distinctly superior to the pressure process, for under it as much as 99 per cent, oi the oil can readily be extracted from the raw material. A further advantage of the process undoubtedly lies in the simplicitj- and cheapness of the plant required as compared with that used under the pressure method. The relative advantages of these two processes form a subject of much discussion. As the reader is doubtlessly aware, the residue left after the oil has been extracted from linseed, cotton seed, copra, and certain other oil-bearing substances, is in great demand as a cattle food. While it is admitted generally that the solvent extraction process recovers the oil more thoroughly from the seed, etc., than does the pressure process, it is frequently urged that its verj' efficiencj' in this respect deprives the residue of much, if not quite the whole of its value as a feeding stuff. The 5 to 10 per cent, of oil remaming in the cake left after crushing in a press is not. it is claimed, a source of loss, for without it the residue could at best command a market only as manure. On the other hand, it is stated that the oil left in the cake is only a heat- forming substance, and that the husks, etc., of the seed foj"m the real food value of the cake. Further, oil press cake, it is argued, cannot be fed imdiluted to cattle, but has to be mixed with bran and other substances, a fact which would seem to imply that oil cake is a richer food than it need be. The residue left by the solvent extraction process retains all the hu.sks, etc., wliile its richness in oil is not such as to prevent its being fed directly to cattle. Whatever maj- be the true way of looking at this matter we have next to note that the advocates of the pressure system urge a further objection to the solvent extraction process. This is to be fomad in the alleged difficulty o ■ impos.sibihty of getting rid of the last traces of the solvent used eitl.er from the oil or the residue. The point is of importance, for the solvents commonly used are either poisonous or have a nauseous taste. If the allegation were well fomided, therefore, solvent extracted oils could not be readily used for edible purposes, and would have to find an outlet solely in industrial appUcations, such as soap making, while the residue would probably be refused as food by cattle and would have to be used as manure. Whatever may at one time have been the case, and may still be where old-fasliioneil INTRODUCTORY AND GENERAL 5 Germ an -made solvent extraction plant is in use, it seems certain that recent progress has overcome these objections to the process. We are credibly informed that horses and cattle will eat extracted meal with avidity. AVe have examined oil extracted with benzene, and neither to the taste nor smell did it reveal any trace of the solvent, although benzene is said to be the most difficult of all the solvents to elimmate. It is advisable, we think, to discard the idea that the two processes ai'e essentially rivals. It is certainly undoubted that they can be very profitably worked side by side in the same mill, for the solvent process can be made to supplement the pressure process frequently with great advantage. Thus certain seeds can profitably be crushed to recover a high-class edible or other oil, and thereafter treated with .solvents to recover the remaining oil. It is to be noticed tl at in discussing the relative advantages of the two processes it is not wise always to confine our argument to the general case. Oui conclusions must be modified not only by local conditions as to the outlet for the oil and seed residue — that is, the press cake or extracted meal — but also by the particular oil-bearing seed which is to be treated. Thus the residue of certain seeds, rape seed, for example, has little or no value as a foodstuff however it is obtamed. It seems, therefore, only reasonable in such ca.ses to adopt that process which recovers most oil from the seed, and which, moreover, leaves the residue in a form which is directly suitable for manurial purposes. On the other hand, the solvent extraction process should be studied cautiously if castor seeds are in question. Castor oil is in several respects an exceptional oil and appears to suffer some chemical change by the action of solvents. In conclusion, it may be remarked that any objection to solvent extracted meal as a foodstuff on the ground that it is deficient in oil can be overcome by mixing it with the desired proportion of oil and moulding it into cakes. Again, it can be mixed with gromid-up press cake and the whole remoulded. Both practices are followed on the Continent. We may add that so far as we can discover there is no ground for the assertion made in an authoritative work that extracted meal cannot be sold in this country as a cattle food. CHAPTER II THE PRINCIPAL VEGETABLE OILS It is not quite easy to compile a list which is likely to meet with imiversal approval as representing the principal vegetable oils. In presenting the following brief summary of the sources, characteristics, and chief uses of certain oils, we will therefore not insist upon its being a complete li.st of the principal vegetable oils. It may be taken, however, that all the oils mentioned are of first-class or of considerable industrial importance. Linseed Oil. Few, we think, will quarrel ■nith our selection or this oil for notice before all others. It is midoubtedly one of the most important, if not the most important, oil known to man. The flax plant is a herb consisting of a single stem 20 to 40 inches high, and is widely cultivated in many temperate climates, notablj^ in Ireland, Belgium. Holland, France, Rus.sia, India, Canada, the United States, and the Argentine. Its stem consists of a core, an outer covering, and an intermediate layer of tissue kno\Mi as the " bast." The bast, suitably separated from the other parts of the stem, fonns, of course, the basis of the linen industry. The fruit of the plant yields a seed, " linseed," which forms the basis of the linseed oil industry. The plant is cultivated in two distmct forms, one more richly flowered than the other, and therefore gi-o^\ii tor the sake of its seeds. This varietj' is chiefly cultivated in Russia, India. Canada, the United States, and the Argentine. The Russian, and particularly that coming from the Baltic districts, is perhaps the mo.st highly esteemed source of linseed oil. The seed contains from 38 to 40 per cent, of oil. The oil is recovered from the seed very commonly by hot pressing. The hot press cake retains about 10 per cent, of oil and forms an extremely valuable and wholesome cattle food. Occasionally, the seeds are pressed cold for the recovery of an edible oil. The hot pressed oil is of v,ic\e application in the arts. It is used exten- sively in the manufacture of soft soaps. Its high specific gravity and its fine drying qualities make it of first importance in the manufacture of paints and vaniishes. The chemical changes which occur when linseed oil " dries " are not clear, but it is certain that the main feature is the oxidation of the oil. The oxygen is taken up rapidty and transforms the oil into a flexible solid mass, lcno\\ii as .solidified or oxidi.sed lin.seed oil or as " linoxyn." This substance is manufactured on a large scale, for it is the principal raw material of the linoleum and oilcloth industry. Linseed oil in the natural state dries to an elastic .skin in about three days. If, however, it is prepared by heating it with various salts of lead or manganese, it Avill dry within six or eight hours. So treated, it is known as " boiled " oil and is much used by painters and artists. Heated with sulphur, linseed oil is used medicinally. Acted upon otherwise by sulphur or by chloride of sulphur the oil solidifies into a vulcanised oil, which is used as a substitute for rubber. This material is chiefly used as an addition to genuine raw rubber. That the one plant should give us so many and .such different materials is a remarkable illustration of Nature's economy. The picture would be complete were THE PRINCIPAL VEGETABLE OILS 7 linseed oil available for burning and lubricating pui-poses. Its quick drying properties, however, render it unsuitable for these uses, particularly for the latter. Cotton Seed Oil. Several oils have claims to be ranked next in importance to linseed oil. One of these is cotton seed oil, as obtained from the seeds of the cotton plant. There are several varieties of cotton plant, such as the Upland and Sea Island varieties, as grown in the United States, and the Egyptian, Indian, Brazilian, and Peruvian. The fruit or " boll " of the plant, when ripe, bur.sts opeii and exposes the cotton in the form of a fluffy mass. In this condition it is gathered and is Itnown as seed cotton,, for it consists of about one-third by weight of fibre and two-thirds of seed. The fibre is attached to the seeds and has to be separated therefrom in a cotton gm. The degree of success attending the ginning is an important consideration in the subsequent recovery of the oil from the seeds. The Egyptian plant yields a black seed, from which the fibre is easily removed completely in the gin. The American and Indian seeds, on the other hand, are white and leave the gin with a considerable amount of the fibre still adhering to their husks. Such seeds are sometimes reginned in the mill before being further treated. Practice differs, however. In some cases the seeds are crushed as received, the fibres and husks being allowed to pass into the cake. In other cases, as, for example, very frequently in the case of American " Upland " seed, the fibre is removed, the husks are taken off in a decorticating machine, and the kernels, or " meats " as they are called, are alone crushed. The seeds on the average consist of 60 per cent, keiuiel and 40 per cent. husk. The amoimt of oil which they contam varies from 18 to 24 per cent., according to the country and plant producing them. Cotton seed kernels, the real source of the oil, contain a strong deep-brown colouring matter. Owing to the difficulty of refining the crude oil, the seed not required for planting was, mitil a little over sixty years ago, thrown away, although as long ago as 1783 the Royal Society of Arts endeavoured to encourage the production of cotton seed oil and cotton seed cake. In America, up to 1860, the disposal of the seed fi'om the ginning plant was, in fact, a problem of no little concern to the proprietor. He was heavily penalised if he allowed the seed to accumulate near the gin, and was strictly forbidden to get rid of it by throwing it into any river or stream. The waste seed was to a small extent used as a maniu'e, but the bulk of it had to be bunied. Later on it was discovered that the residue left after millmg the seed for its oil retained all the fertilismg properties. Cotton seed meal, that is, the press cake ground up, is still largely used as a manure for sugar cane, cotton, com, tobacco, and so on. It is now, howev^er, realised that the most economical mamier of using it is to feed it to cattle and to use the resulting manure, which retains 80 to 90 per cent, of the original fertilising value, on the ground. Cotton .seed oil is a so-called .semi-drying oil. It absorbs oxygen slowly under ordinary conditions, but by blowing air through it at about 100" C. the absorption can be increased. Blown cotton seed oil and other semi-drying oils similarly treated become thickened and appear in den.sity and viscosity like castor oil. They are produced on a considerable scale and when dissolved in light mineral oils are used as lubricants. Refined cotton seed oil is in extensive use for edible purposes. It appears on the table as salad oil, it is used by the sardme tinning industry, and under the name of butter oil it fonns one of the chief raw materials of the margarine manufacturer and of the manufacturer of lard substitute, or compound lard as it is called. Apart, from the very great use of cotton seed oil for edible purposes its chief industrial emplojTuent is in the soap-making industry. It is frequently "used in this connection by itself. As an 8 THE PRODUCTION AND TREATMENT OF \'EGETABLE OILS ingredient of toilet soap it is commonly mixed \rith tallow or cocoa-nut oil. It is also widely used in the manufacture of soap powder. Olive Oil. Olive oil is in several respects chemically and industriaUy closely similar to cotton seed oil. The latter being cheaper :s frequently substituted for it. notably for edible purposes. The reputation of oUve oil as an edible oil is. however, too great for it ever to be supplanted completely by any other. The olive tree is chiefly cultivated in the countries bordering the Mediterranean. Recently attempts, not always ^^-ith success, have been made to grow it in India, California. South Africa, and Australia. The fruit of the oUve consists of rind, flesh, stone, and seed kernel. All parts contaui oil. The fleshy part, forming 80 per cent, of the whole, contains from 40 to fiO per cent, of oil and yields the best oil for edible purposes. To produce tliis oil the fruit is gathered before it is quite ripe and is peeled and stoned. The flesh is then pressed by itself. The kernels are crushed separately and yield an inferior ' olive kernel oil." The pulp left after the pressing of the flesh may contain as much as 20 per cent, of oil. It is gromid up with hot water and allowed to stand nntil the broken up cellular tissue rises to the surface. This is again pressed for a second quahty oil. The residue is finally extracted with solvents, commonly carbon disulphide. Such extracted oil acquires a deep-green colour from the chlorophyll in the fruit, and is principally used for soap-maldng. In some mills the original fruit is not stoned before being pressed for the first time, but is crushed as a whole. The oil yielded is of a less perfect quality than that obtained by the other process, for it contains the poorer oil derived from the kerne's. The oil derived from the fir.st pressing of the fruit is almost invariably used for edible purposes. A second or third pressing is commonly adopted. The oil so obtained is used for soap-making and for lubricating and burning purposes, for olive oil is a non-drj-ing oil. The press cake is sometimes used locally as a cattle food. The value of the oil. however, makes it pay to carry the recoven,- to the greatest pos.«ib!e extent. Hence the last drop of oil is usually recovered by the chemical solvent process and the residue sold as manure. Castor Oil. The castor tree or shrub — it is found in both forms — grows in all tropical and sub-tropical countries. The seeds are enclosed withm a rough outer shell or pod. and themselves consist of a husk containing a white soft kernel. The kernel forms 80 per cent, of the seed and j-ields from 46 to 53 per cent, of its weight in oil. The husks are oil-less. The bulk of the seed u.«ed in commerce comes from the East Indies. Ca«;tor oU is of the non-drying class and is of great value as a lubricant. It is extensively used in the soap industr\-. Treated with concentrated sulphuric acid it yields a fatty substance knowii as Turkey Red oil. which is used in preparing cotton fibre for dyeing in the Turkey Red industry-. Its medicinal use depends upon the fact that it contains an alkaloid. This alkaloid in excess is poisonous, and as it is retained in considerable quantity in the residue left after crushing the seeds, castor oil cake is unfit for a cattle food. Consequently it is extracted with solvents to recover a quality of oil suitable for soap-making and other technical pirrposes. The ultimate residue is used as a manure. Castor seeds are commonly pressed cold to obtain medicinal oil and then pressed a second or third time in a hot condition to obtain technical quality oils. The seeds THE PRINCIPAL VEGETABLE OILS 9 are frequently crushed with their husks on, but sometimes they are previously decorticated and their kernels alone placed in the press. Palm, Palm Kernel, and Cocoa-nut Oils. The fruit of the African oil palm consists of a fleshy outer layer or pericarp surrounding a hard woody shell within which is the seed kernel. Roughly, the shell forms 50 per cent, of the whole, the fleshy pericarp 35 per cent., and the kernel 15 per cent. Of the pericarp 50 per cent, or so is oil, while the kernel yields about 45 per cent. In the case of the olive the oils recovered from the fleshy part and from the kernels are practically the same. In the case of the palm tree fruit they are quite different. Palm oil, the product derived from the pericarp, is used principally in the making of soap and candles. The pericarp, owing to its nature, has to be worked up as soon as the fruit is pulled. Consequently, the factory is erected near the plantation. The kernels, separated from the pericarp, are shipped to the United Kingdom, Ham- burg, etc., and are treated by cru.shing and extraction with solvents for the recovery of palm kernel oil. This oil in a fresh condition is largely used in the manufacture of margarine, and to a considerable extent, when suitably treated, in the manufacture of chocolate. The poorer qualities and the extracted oil are suitable for soap, candle and paint making. Palm kernel oil cake is somewhat deficient in nitrogen, so that its value as a cattle food is less than that of some other qualities of cake. This deficiency also renders the residue from the extraction process of low value as a manure. Cocoa-nuts are obtained from a tree of the palm family, not, of course, from the cocoa tree. The fleshy layer inside the nut, when dried either in the sun or by artificial heat, is kno^vn as " copra." The undried flesh contains about half its weight of water, so that by drying it, an operation carried on at the place where the nuts are gathered, a considerable saving of freight is affected. The copra, shipped to the oil mills, is shredded and crushed hot. It yields roimd about 64 per cent, of its weight in oil, but this figure is subject to variation according to the precise method adopted for drying the copra by the gatherers. Cocoa-nut oil is very closely similar to palm kernel oil and is used for much the same purposes, that is to say, chiefly in the making of margarine and of soap. These three oils, palm, palm kernel, and cocoa-nut oils, are all of the non-drying type, and at ordinary European temperatures are to be regarded as vegetable fats rather than as oils. It may be noted here that although cocoa-nuts do not grow on cocoa trees, still cocoa-nut oil — and also pahn kernel oil — is of great value to the chocolate manufacturer. The cocoa bean when roasted and ground contains about 50 per cent, of fat, or " cocoa butter " as it is called. This fat lenders the cocoa powder difficult of mixture with boiling water and indige.'ftible. It i.s besides a valuable .substance in itself, being used in medicine and soap-making. Hence it is frequently removed to the extent of about half its original amount by submitting the ground cocoa powder to hydraulic pressure. In working up the cocoa powder into chocolate of the best quality a portion of the abstracted cocoa butter is returned to it. In other chocolates the valuable cocoa butter is omitted, and cocoa-nut oil, suitably treated, or palm kernel oil is used instead. Soya Bean Oil. The soya bean flourishes in Manchuria, China, and Japan. Li Manchuria the cultivation of the plant is .stated to have been the main agricultural industry for centuries, while the production of soya bean oil and .soya bean cake formed the most important manufactures of the country. The bean cakes have for long formed one 10 THE PRODUOTIOX AXD TREATMEXT OF \T:GETABLE OILS of the chief articles of food for the inhabitants. Nevertheless, the bean and the oil it yields were almost unlciio\\ii in Eiuope until the Russo-Japanese Wav. Since then the production and use of soj a bean oil and soya bean cake have developed phenomenally. The oil in Europe now rivals that obtained from the cotton seed, while the cake, on the Continent at least, is contesting the position as a food for milch cows held by linseed and cotton seed cake. The oil belongs to the semi-diying class, and is used for edible purposes, as an illuminant, in soap-making, and in the manufacture of linoleum. The bean contains about 18 per cent, of oil, and in the press yields from 10 to 13 per cent. Rape or Colza Oil. The rape plant is gi'OMii extensively in many European countries, notably in Russia. It is cultivated in British India to an extent which renders the annual crop second onh' in importance to the linseed crop. The bullv of the Indian seed is shipped to England, but C4ermany used to have a preponderating hold on other sources of supply. Rape oil belongs to the semi-drjang class, and is principally used for burning purposes and as a lubricant. In the latter case the oil is frequently '" blowTi," as mentioned above under cotton seed oil. To a small extent rape oil when obtained by " cold drawing " is used for edible purposes, notably by bakers in the production of bread. It is commonly employed as a quenching medium for steel plates, etc., and on the Continent is used occasionally in the manufacture of soft soap. The seed contains anything from 33 to 43 per cent, of oil. It is frequently extracted by means of solvents. The oil apparently contains a poisonous element. Consequently rape seed cake is not greatly valued as a cattle food. It may, in fact, be said that the bulk of the residue left after either crushing or extraction M'ith solvents is used as a manure. Mustard Oil. Tliis oil is obtained from the black, white, or wild mustard plant, and is used in soap-maliing and as a substitute for or adulterant in rape oil, to which it is closely similar. The cake left after crushing is, however, a more important product than the oil. When ground this cake gives the mustard of the domestic table. ScjNFLowER Oil. The sunflower is cultivated for the sake of its seeds on an immense scale in Russia, Italy, India, and China. The seeds, raw or roasted, are used in Russia as an article of diet. The oil recovered from them by crushing is, when refined, considered by some to equal olive oil for edible purposes. Its chief use, however, is in soap and candle-making. The seeds contain from 20 to 23 per cent, of oil. For cattle-feeding purposes the cake is not only veiy palatable, but being rich in nitrogenous matter is of great food value. Sunflower oil belongs to the drs'ing class. The sunflower is verj- readily cultivated, and produces a high yield of seeds. It is believed that the Central Empires, cut off as they are at present from many important sources of oils and fats, are cultivating the sunflower on an exten.sive scale in an attempt to reduce the deficiency. They are probably growing flax — for linseed oil — also on a considerable scale ; but flax, it is to be noted, rapidly exhausts the soil and is therefore in all hkehhood bemg cultivated to an extent only slightly greater than in peace time. It maj' perhaps be added that the rumours recently in circulation as to Germany's shortage of glycerine and the horrible means she is adopting to make it good cannot be accepted as true by those qualified to judge. In the first place Germany uses little or no glycerme in the production of her explosives, differing in tlus respect from this country, which, of THE PRINCIPAL VEGETABLE OILS 11 course, dt-pends extensively upon nitro-glycerine. In the second place the yield of glycerine from the source suggested would be altogether too insignificant to justify the cost, trouble, and difficulty of recovering it.* Poppy Seed Oil. The seed of the poppy contains from 45 to 50 per cent, of an oil which, when '■ cold dra^^^l," is almost colourless, has little odour, and possesses a pleasant taste. It is in extensive use on the table as a salad oil, and is highly valued by arti.sts and artists' colourmen. The seeds are usually expressed twice, the second pressing being carried out hot and yielding an inferior oil which is extensively employed in making paints and soft soaps. The oil belongs to the drying class. Poppy seed cake is rich in nitrogen and is highly valued as a cattle food. * Since these remarks were written, the rumours referred to have become a popular article of belief, iu this country at least, and have received what would appear to be semi-official confirmation. CHAPTER III PREPARATORY .ALICHINERY FOR COPRA AND LINSEED From the engineering point of view the machinery and plant used in the production of vegetable oils may conveniently be divided into four classes. First we have what may be called the preparatory machinery, the plant, that is to say, which deals with the seeds, nuts or fruit as received from the growers, and reduces them to a form suitable for treatment m the subsequent oil recoverj^ processes. Xext we have the presses wherein the material so prepared is crushed. Thirdly, there is the plant emplo}"ed when the oil is extracted by chemical solvents, either as an alternative to crushing or as supplementan,- thereto. Fourthly, there is the plant employed to refine the oil. To these four classes of oil mill machinen,' and plant a fifth has to be added. This is not so much concerned with the production of the oil as with the production and treatment of the cake. Of the machinery' m the preparatorj- class it may be said that there are three distinct divisions. In the first of these we have the machines and appliances used in the separation of the oil-bearing portion of the natural product from the non-oil beai-ing portions, or from secondary oil-bearing portions whicli it is desired to treat apart from the fii-st. The portions thus separated have in general to be prepared for the presses b}' j'rushing, shreddmg, and otherwise reducing them to a meal of sufficient fineness to present the mmute oil vesicles in the best form to the action of the press. The machines thus employed form the second of our three divisions. The third division embraces machines and apphances concerned with the manipulation of the meal just before it goes into the press. These manipulations include the heatmg of the meal to a suitable temperature and its rough moulding into slabs or cakes for insertion within the press. It will be imderstood, of cour,se, that this is a general outHne only, and that all oil-bearing vegetable products do not necessarily require the whole run of the appliances thus indicated. Thus linseed, rape seed, and similar small seeds do not entail the use of any preparatory machines of the first division. On the other hand, cotton seed frequent!}-, and palm kernels nearly always, require the use of machines of all three kinds. The machines of the first division are in general of a speciahsed nature, that is to say, they are in most cases each designed to deal with one particular class of seed. The machines of the second division are very similar among themselves whatever the seed or nut being treated. The machines of the third division do not varj' with the nature of the seed being dealt with, for by the time the seed reaches them it has lost all its outsta-.ding original physical features, and whatever it was to begin with is now in the form of more or less fine meal. We now pass to a description of the more oi less specialised preparatory machinery in use or designed for the treatment of certain important oil-bearing substances. Before domg so we desire to make two general remarks which may avoid occasion for misunder-standing. In the first place certain of the machines described below are suitable for treating substances other than that mentioned in the heading under which they appear. Secondly, the fact that any given machine is mentioned as being made by such and such a firm does not necessarily mean that it is made onlj- by that finn. INSERT FOLDOUT HERE PREPARATORY MACHINERY FOR COPRA AND LINSEED 13 Preparatory Machtneby for Copra. Tlie cocoa-nuts as gathered have first to be split open. The split nuts are then set in the siui, placed in a kiln, often of rude construction, or, as is now becommg the practice, deposited in a lightly built galvanised iron house beneath the floor of which steam-heating pipes are disposed. In either of these ways the moisture contained in the flesh of the nut, amounting to about half the original weight of the flesh, is driven off and the flesh itself becomes loosened from the shell. The dried flesh — copra — is then exported. Great care is necessary in carrying out the drying process, for the material can readily be spoilt by attempting to drive off the moisture too rapidly, with the result that the flesh is discoloured and the oil recovered from it is difficult to refine. On the other hand, the natural moisture in the flesh must not be allowed to remain unduly long in contact with the oil in the flesh after the nuts have been split. If this is permitted the water will hydrolise the oil, that is to say, it will enter into chemical combmation with it and split it up into two elements, namely, glycerine and free fatty acid. The splitting of the nuts was, and still is, frequently performed simply with a hammer. More scientific means are now, however, being introduced. In Fig. 1 we give the general ari'angement of a machine for the purpose made by Rose, Downs &, Thompson, Ltd., of Hull. This machine deals with nuts as gathered, that is to say, it treats them with their outer husks still on. By means of three circular l^nives having saw-like teeth, and spaced at 120 degrees apart, it cuts through husk, shell, and kernel, and divides the whole into three parts. The knives are momited on three shafts forming a triangle in plan, and geared together by bevel wheels. One of the Icnife shafts is driven by spur wheel and pinion from a belt-driven comitershaft canying a flywheel. The knives are 1 ft. 5 in. in diameter and run at about 25 revo- lutions per minute. A more or less conical sheet metal hopper is fixed over the knives. The loiives pass through openings in the sides of this hopper. Three bent plates or knees are attached to the inside of the hopper to act as guides for the nuts and lead them to the centre of the hopper. There they are caught by the Iviiives and are cut and carried do\viiwards to fall on to the base plate of the machme. About 2 h.p. is required to drive this machine. Its output may be returned at 2,000 nuts per hour. It will, in fact, split the nuts as fast as one man can feed them to it. The copra as received at the oil mill in this or some other coiuitry is in the form of lumps of considerable size. These have to be reduced to the form of " meal " by various shreddings and gi'indings. Before doing so, however, it is necessary closely to examine the material, for it is frequently found to contain an odd assortment of scrap iron, such as hammer heads, bolts, nuts, horseshoes, nai!s, etc. The native and other gatherers of palm nuts, copra, and so on, are paid by weight, and on occasion do not scruple to turn a dishonest penny. Many breakdowns of machinery have been caused by the undetected presence of sucji foreign matter in tlie material treated, particularly so in the case of mills handling copra and other sub.stances which as part of the preparatory process have to be gi'ound. To remove the objectionable substances resort is commonly made to hand picking the material before anything else is done to it. This process is slow and monotonous, so that it is not sui-prising that modern practice should call for some mechanical means of perfomiing the operation. An appliance of this nature, a magnetic separator, made by Rose, Dowtis & Thompson, Ltd., is illustrated in Fig. 2. The material to be treated is delivered on to a sloping sheet metal tray supported on four flexible spring rods A and rapidly vibrated by means of a connecting rod B, and short-throw crankshaft C. The shaking 14 THE PRODUCTIOX AXP TREATMENT OF A'EGETABT.E OILS action of tliis tray distributes the material luiifomily. and further causes the heavy iron ingi-edients to sink by gi-a-vity to the bottom of the batch. Sometimes the tray T~E Ekcineeb" S.«i>. Sc Fig. 2. — Magnetic Separator for Copra, etc. — Rose, Iiowns & Thompson. can be made in the form of a screen, when it will serve the additional purpose of separating any fine impurities, sand, chips, and so on. from the oil-bearing substance. HBS-^^ K ***1 1 1 L C r''^-^!!IB ^j^^lJUgj^^^m^^^^^ Fig. .J. — rreliminary Breaking Machine — Manlove, Alliott. The material leavuig the tray falls on to a power-driven magnetic barrel D, provided with several rows of studs, each row being momited on a separate commutator section of the barrel. The non -magnetic material is carried round the barrel and drops off PREPARATORY IMACHTNERY FOR COPRA AND LINSEED 15 at the front into a suitable hopper. The iron intruders are carried round to tlie back luitil they reach a point at which the section of the barrel to which thej' are adhering is automatically demagnetised for a moment to permit them to drop off. The size of machine illustrated has an output of from 1 to 1 J tons per hour. It is driven from the small pulley at the end of the shaker crankshaft, which shaft is rotated at 300 revo- lutions per minute. About a liorse-power is absorbed in driving the machine. The current required for the magnetic barrel is 7 amperes at 40 volts, and can be conve- niently supplied by a snaall belt-driven dynamo provided for the purpose. JMachines for Reducing Copra, etc. The reduction of copra as imported to the form of meal for the press requiies special consideration. The copra is generally received in pieces of such a size that it has to be reduced in three, four, or even five separate .stages. Usually these reductions are effected by means of rolls. For the first reduction, however, rolls may be dispensed Fig. 4.— Shredding and Crushing KoUs for Copra, etc.- -A. F. Craig. with and a preliminary breaking machine, Fig. 3, as made by Manlove, Alliott & Co., Ltd., of Nottmgham, u.sed m their place. This machine is claimed to be considerably cheaper than rolls both in first cost and in maintenance, and to rmi without attention so long as it is fed evenly. The casing of the machine is a cast-iron barrel ribbed internally to prevent the copra rotating as a ma.ss within it. The casing contains a power-driven segmental worm having a coarse pitch at tlie feed end and a finer 16 THE PRODUCTION AND TREATJfENT OF \TEaETABLE OILS pitch at the dehvery end. At the latter end there is fitted a hardened perforated steel plate through which the partially broken copra is forced bj^ the worm. A four-bladed knife revolves agauast the worm side of this plate anc^cuts the copra as it passes through the perforations besides assisting its passage through these holes. The worm shaft is fitted with Hoffmann ball thrust bearmgs. The perforated plate can be readily changed and one with smaller holes .substituted for it. This change pennits of the machine being used with e:[ual facility for the breaking of palm kernels. The material as thus disintegrated is next reduced a step further by means of rolls which shred and ciush it. A set of rolls suitable for this purpose, as made by A. F. Craig & Co., Ltd., of Paisley, is illustrated in Figs. 4 and 5. The example here represented has two pairs of rolls, the upper pair of which is fluted longitudinally, the lower pair being plain. The rolls, as is usual, are of chilled cast iron and are The Engineer" Fig. 5. — Shredding and Ciushing Rolls — Craig hydraulicallj'^ pressed on to steel shafts. One roll in each pair runs in fixed bearings, the other running in sliding bearings which are acted upon by relief springs disposed within cii-cular boxes on the frame sides. The force of these springs is adjustable to give the required degree of pressm'e between the I'olls. In a common size of machine, one capable of dealing with 15 cwt. of material per hour, the rolls ai'e 48 in. long. The lower rolls are each 14 in. in diameter, and run at about 130 revolutions per minute. The upper rolls are of different diameters and rotate at different speeds. The larger roll is 16 in. in diameter, and rims at about 110 revolutions per minute, the smaller roll being 12 in., and ruiming at about 47 revolutions. The peripheral speed of the larger roll is thus about three times that of the smaller. As a result, the partially reduced material falling from the hopper between the top rolls is shredded by the fluted surfaces. Falling between the plam lower rolls it is still farther reduced by a crushmg action. Each of the four rolls is provided with a scraper worldng against its lower portion. The feed hopper consists of a tiough formed over the smaller of the two top rolls. At the front a fixed plate extends from it to the surface of the roll. At the back a hinged plate, adjustable by means of one or two screws and hand wheels, permits the quantity of material passijig out of the hopper to be regulated. About 15 b.li.p. is consumed in driving a set of rolls of the size mentioned in this paragi'aph. PREPARATORY MACHINERY FOR COPRA AND LINSEED 17 A set of rolls made by Manlove, Alliott & Co., Ltd., for the same purpose aa the above is represented m Fig. 6. Li this case there are three pairs of chilled cast-iron rolls. All six are of the same size, each being 15 in. in diameter and 36 in. long The two top pairs are spirally fluted and are driven from the belt pulley at the right-hand end of the machine through double helical gearhig. The two lolls in each of these Flo. 6. — Rolls for further Reduction of Copra — Manlove, Alliott. upper pairs rotate at different speeds so that the action is a shredding and grinduig one. The two lower rolls are plain and are separately driven at equal speeds by the belt pulley at the left-hand end of the machine. The material at this point is rolled rather than ground. As before, the rolls are fitted with scrapers. The feed hopper is provided with an adjustable shutter and a power-driven feed roll which ensures the material being delivered evenly along the length of the rolls, an important item in successful working. These rolls are made in various sizes to treat from 12 to 20 cwt. of material per hour, and in their driving consume from 8 to 10 b.h.p. 18 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS The final reduction of the material to meal of the proper degree of fuieness is carried out in powerful rolls, of which an example made by Messrs. Manlove, Alliott is illustrated in Fig. 7. The action required is one of rollmg not of giindmg. The five rolls in the machine illustrated are stacked vertically, and are driven positively by means of a double helical gear-wheel at each end of each roll. The rolls rotate at equal speeds and are either plain or lightly fluted, both styles being often found in the same macliine at once. The material leaving the hopper at the top is guided in a sinuous course through the rolls by means of four inclmed plates let into the machme framework on alternate sides of the rolls. The lowest roll spindle carries two fast and Fig. 7. — Final Seduction Rolls — Manlove, Alliott. 1 wo loose pulleys. A simple belt shifter is provided which permits of the two belts being moved simultaneously. This avoids the risk otherwise present of causing damage by transmitting the drive entirely through the gearing at one end of the rolls. The output of these rolls varies from 10 to 20 cwt. per hour. Preparatory Treatment of Linseed, etc. Linseed, rape seed, and similar small seeds require verj^ little special preparation for pressing. Beyond screening to remove foreign matter such seeds have only to be crushed between rolls to convert them to meal suitable for pressing. An appliance for screening lin.seed and similar seeds, made by Rose, Downs & Thompson, Ltd., of Hull, is illustrated in Fig. 8. This machine consists of a cast iron casing containing a slowly rotating cylindrical screen into the interior of which the seed is delivered. Inside the screen is a fast-moving paddle which throws the seed against the interior surface of the screen. The seed is delivered to the machme at the PREPARATORY MACHINERY FOR COPRA AND LINSEED 19 orifice A, and is caiTied into tlie screen hj the action of a sliort worm fixed on the end of the paddle shaft. The screenhig surface is formed of perforated sheet steel and covers the screen framework from the line BB to the line CC. The screened seed fallmg through the perforations collects \vithin the vee-sectioned hopper formed by the lower walls of the main casmg and is carried by a rotating worm either to the outlet D or the outlet E, according to the formation given to the worm. The tailmgs fail to pass through the perforations of the screen and are delivered through the gap left beyond the line CC to an orifice F, divided from the orifice E by a partition. The machine is driven from the right-hand end of the paddle shaft. The dischargmg Fig. 8. — Screening Machine for Linseed, etc. — Rose, Downs & Thompson. worm is driven by l)elt from the left-hand end of the paddle shaft and itself drives the cyhndrical screen by chain and sprocket wheels. The machine illustrated has an output of 28 to 30 cwt. per hour and requires about 3 b.h.p. to drive it. The paddle shaft runs at 100 revolutions per minute, the screen at 12, and the discharging worm at about 42. On occasion a screen of this type is required to deal with seeds or other such material of varying size, and to separate a batch into two portions besides the tailings. For instance, the left-hand half of the screened surface may be perforated r['^ in. mesh, and the right-hand portion -^ in. The hopper is then divided by a partition such as at G, and the discharging worm is made in two corresponding portions, one right, the other left-handed. In this way the finer-sized seeds are delivered through the orifice D, the coarser at E, and the tailings at F. 20 THE PRODUCTION AND TREATMENT OF ^T:GETABLE OILS A set of rolls suitable for crushing linseed, etc., made by Manlove, Alliott & Co., Ltd., of Nottingham, is illustrated in Fig. 9. The rolls are five in nimiber, measure 16 in. in diameter by 42 in. long, are stacked vertically and are quite plain on the surface. As usual, they are ground with great truth and are forced on to their shafts by hydraulic pressure, being thereafter keyed at both ends. The lowest roll is driven at both ends and is provided mth two additional pulleys, from which belts are taken to similar sized puUej^s at each end of the third and fifth rolls. The bearings for all the rolls except the lowest are free to slide vertically in their housings. Consequently Fig. 9. — EoUs for Linseed, etc. — Manluve, Alliott. the pressure exerted on the seed being reduced increases with each step m its descent from the hopper. The second and fourth rolls are driven simply by the friction between them and their neighbours. The consequent sHp of these rolls is reUed upon to give the grinding action which, to a small extent, should accompany the crushing action of the machine. Means are provided whereby the two upper rolls may be held sUghtly raised so as to increase the feed. Additional means are also provided whereby, if desired, the dead weight of the rolls may be assisted by the action of tightening screws and springs. The capacity of the roils illustrated is about 15 cwt. of seed per hour. CHAPTER IV PREPARATORY MACHINERY FOR PALM FRUIT AND PALM KERNELS We now come to a section of our subject concerning which there is much discussion and variety of opinion. Before proceeding to make any remarks on it, we may direct the reader's attention to Fig. 10. In this we give an ilhistration which we have had Fig. 10. — Palm Fruit, I'oricarp, Nuts, Shells and Kernels. prepared for the purpose of showing the fruit of the oil palm and its component parts. At A the whole fruit, as gathered from the tree, is shown. The shape and size of the fi-uit are, it will be noticed, somewhat irregular. On the average the fiuit is about ll in. long and f in. maximum width. At B the pericarp is to be seen. The pericarji has a smooth outer suiface, but on the whole consists of a mass oi fibres snieaied with a thick yellow oil, or, more correctly, fat. At C the nuts which the pericarp surrounds are shown, while at Dand E respectively are indicated the fragments of the ntit sliells 22 THE PRODUmOX AND TREATirEXT OF \'EGETABLE OILS and the nut kernels. The sample of fniit from which the original jthotogiaph repro- duced in this engraving was prepared was Idndlv supplied to us by A. F. Ciaig & Co., Ltd.. Caledonia Engine Works, Paisley. We were fortunate in seeming this sample, as the fruit rapidly deteiiorates aft^r being gathered, and is therefore rarely seen in this countri. A bunch of pahn fruil as taken down fiom the tree is shown in Fig. 1 1, for the original of which we are indebted to Messis. Manlove. AUiott. Such a bunch may weigh round abotit 1-4J lb. and mer.sure ab ut 12 in. long by 10 in. across. From the pericarp B (Fig. lU) pahn oil is obtained, while fiom the kernels E a totally different oil, palm kernel oil. is recovered. Owing to the rapid deterioration which the pericaip suffers after the fruit is gathered, it is impracticable to ship the whole fruit to Europe for treatment. In general, therefore, the practice is to recover the palm oil from the pericarp at or near the planta- tions in Africa, and to send the nuts or kernels overseas for treatment at home. Treatmext or the Palm FsriT. L'ntU recently, and still to a con- siderable extent, the production of palm oil was in the hands of the natives. The method they use is crude. Xot only do they lose, by following it. from a half to two-thirds of the possible oO yield, but the oil obtained is apt to have developed in it elements which lower its commercial value. The fruit when ripe, is deliberately allowed to ferment in the presence of water, so that. the hard pericarp may lie softened and readily separated from the nut. The separation is effected simply by beat- ing the softened fruit to a pulp, and thereafter picking out the nuts from The pulp is then Ixiiled in water, and the oil. rising to the This method of working induces hydrolysis in the oil — that Fig. 11. — BuBch of Palm Fn. the mass by hand. top, is skimmed off is to say. the oil combines with water, and is changed from a neutral condition to an acid one by the breaking down of its constitution ;nto free glyc^erine and free fatty acid. Once hydrolysis is started it is liable to continue so that frequently palm oil is received at its European destination containing as much as 5(t per cent, of free fatty acids. The commercial value of the oil is proportionately reduced. Many attempts have been made to treat the fruit in a scientific manner, the direct- objects being to obtain the full jield of oil from the pericarp, and to do so without causing the oil to decompose, or. to use the technical tenn. to hydrolyse. It is obvious that to prevent hydrolysis the fruit must be subjected to a treatment which in no waj calls for its being placed in contact with water in any form. This, the ideal process, is commoiJy spoken of as the " dry '" method. Certain " diy "' methods, notably a German one. have been proposed, and have received some apphcation which have not come up to the ideal standard, for at some stage or other water or steam has been used to assist the recovery of the oil or the separation of the pericarp from the nut. MACHINERY FOR PALM FRUIT AND PALM KERNELS 23 Separation of the Pericarp. The chief difficulty imdoubtedly lies in the effective separation of the pericarp without waituig for it to soften, eitlier by natural deterioration or by fermentation in presence of water. What may be called a compromise process may first be described. The machinery for this process has been supplied by Manlove, Alliott & Co., Ltd., of Nottingham, and we are informed that good, if not ideally satisfactory results have been obtained with it. Under this method of worldng the fruit freshly gathered is taken to a machine provided with a revolving shaft, on which are mounted several bayonet-like luiives. Here the fruit with its nuts is cut and churned up into a pulp. The Engineer" Fui. 12. — Fairfax'.s Depericarping Machine. The mass, sufficiently reduced, is then placed in a cage pre.ss, and pressure applied to it until the nuts are heard to begin to crack. The oil which ffows away is, of course, of good quality, but in quantity does not represent the full oil content of the pericarp. The half-pressed material is therefore boiled up with water to recover the remaining oil, and to complete the separation of the pericarp from the nuts. The oil skimmed off the boilmg water is naturally of an inferior quahty to that running from the press. Palm oil is a very valuable substance, and would be still more so were it possible to obtain it in good condition in large and regular supplies. Consequently we find that much attention has been, and is being, devoted to the design of a satisfactory depericarping machine for palm fruit which will permit the whole pericarp to be treated by a truly dry process. To Messrs. A. V. Craig, of Paisley, we are indebted for the 24 J'HE PRODUCTION AND TREATMENT OF \T:GETABLE OILS particulars and Olustrations which we are enabled to give of their " Caledonia " dry process, and of the macliines designed to give it effect. The depericarpLiig machine used under this process is the patented invention of Mr. H. G. Fairfax, and is illustrated in Fig. 12. Li Fig. 13 we give a working drawing of the same invention, as carried out on practical lines hv Messrs. Craig. The macliine consists essentialh' of two parts, namely a rotathig table A (Fig. 12), and a rotating cover B, the former running quickly in one direction and the latter slowly in the opposite direction. The table carries a series of closely spaced curved blades or abraders C. The cover is formed with a series of wider spaced oppositeh^ cm^ved ribs D. The fruit is fed into the cover through the annulus F. and passing outwards is stripped of its pericarp by the blades C. The counter curvature cf the blades C and the ribs D has, of course, an important influence on the stripping action. The loosened peiicarp falls between the blades C and is at once ejected outwards by cen- Tme Engineer' Fio. 13. — Depericarping Machine^for Palm Fruit — A. F. Cruig. trifugal force over the edge of the table. Here it falls into the receptacle F, which, to facilitate the movement of the pericarp, is steam heated. It will be noticed that the steam jacket G also extends beneath more than half the effective part of the blades C. The nuts ure too large to pass between the blades on the table. Instead they are shot out through holes round the upstanding lip of the machine casing, and are thus collected separately from the pericarp. Any oil which may be set free during the stripping of the pericarp is shot against a screen H and flows away down the outlet J. The nuts as they leave the depericarpmg machine may have small portions of the pericarp still adhering within the irregularities of their shells. To recover such portions the nuts may be passed through the patented brush macliine illustrated in Fig. 14. This, like the depericarping macliine above described, is made by Messrs. Craig, of Paisley. It consists, in essence, of two cylindrical brushes, 3 ft. long and about 7g in. in diameter, revolving side by side at 300 revolutions per minute, beneath a casing or cover embracing the upper portions of the brushes. The brush spmdles rotate in bearings fixed to the casing, and are driven through bevel gearing from a shaft jour- nalled cross-wise on the maLu frame of the maciiine. The casing is also joumalled MACHINERY FOR PALM FRTIIT AND PALM KERNELS 25 to tliis cross shaft, so that it and, with it, the brushes may be set longitudinally to any desired inclination. It is fixed in position at the other end by means of bolts passed through one or otlier of a series of holes formed in a projection on the main frame. At the drivuig end of the machine the casmg is provided with a hopper, into which the nuts are fed. From this the nuts enter one or other of ten gi'ooves formed on the miderside of the casing. The in- clination of the casing causes the nuts to travel do\\Ti these grooves to the outlet end, and in so doing they are timied and brushed all over by the brush bristles, which form, as it were, the fourth side of the gi'ooves. The pericai"p fragments removed by the brushes fall into a hopper between the main frame uprights. The cleaned nuts emerging from the ends of the grooves are caught in a separate hopper. The macliine is designed nominally to deal with about 12 cwt. of nuts — say, 134,000 nuts — per hour.* The ten grooves hold at any one time 400 nuts. The nuts are in contact with the brushes for about 11 seconds each. By altering the inclmation of the brushes and casmg the output can be adjusted wdthin certain limits. The two brushes revolve m opposite directions, and are covered ^^^th " wire cloth " formed of leather, in which are fixed projecting wires as indicated in the sketch. Fig. 15. Treatment of the Nut.s. The pericarp thus recovered is pressed at once. The nuts as cleaned by the b.ushmg machme are dried either naturally or artificially to loosen the kernel within the shell. They have then to be cracked open and the kernel separated from the shell fragments. A cracking and separating machine, made by ^Messrs. Craig, of Paisley, is illustrated in Fig. 16. The hopper of this machine is vee-shaped in section, and is provided internally with an inverted vee- shaped surface, which divides the nuts into * These figures imply that the nuts run at about 11,000 to the hundredweiglit. They vary in size and sometimes number as few as 5,500 to the hundredweight. 26 THE PRODUCTIOX AND TREATMENT OF M^GETABLE OILS two streams. Each such stream passes down a pipe A east on the outside of a semi- cylindrical casing, which contains a drum driven at a high speed, about 1,000 revolu- tions per minute. The nuts fall into tl.e interior of the drums and are shot out by centrifugal force through slots in the drum periphery. Striking forcibly against the inner wall of the surrounding casings, the shells are cracked open, and Avith the kernels fall to the foot of the casmgs. whence they are conducted on to the shaking separator, disposed between the legs of the machine frame. The separator consists of an inclined _.a^ p^ Qrir^jr' |L ^ 3 B^^n— a Plan with Hopper S Platform removed "the Encineeb" Fig. 16. — Palm Xut-ciacking and Separating Machine — Craig. tray, the bottom of which is formed with a special surface. At the lower end it is overhmig on s\Ninging links, the pivot points of which can be varied to give the tray the required inclination. At the higher end it is journalled to a short-throw crank- shaft driven at 250 revolutions per minute. The bottom of the tray is not perforated. The shaking action, combined with the special construction of the bottom surface, results in the kernels being passed to one end of the tray while the shell fragments pass to the other. Falling over the ends they are collected in hoppers. The shells can be used as fuel, either under a boiler or in a gas producer. MACHINERY FOR PALM FRUIT AND PALM KERNELS 27 Li general the kernels .are shippefl to oil mills in Europe or elsewhere. Their preliminary treatment closely agrees with that accorded to copra, much the same shreddmg and reducing rolls being used to convert them to the form of meal. Certain considerations, however, have led the factories in Africa to contemplate undertaldng the work of recovering palm kernel oil within their own walls. The chief of these is the fact that pahn fruit is not available all the year round, so that during the " off " season, if palm oil alone is dealt with, the expensive presses and other plant must lie idle. By a little additional capital expenditure the factory can be readily fitted for -s^B^ -4^- i[^- |a Plan of Ground Fl, Dej)eri carpers Brfjsh Msrhmes Nut Crac±c_ Herrel S eparator ficdacrnQ Milt 'Shrr ddma Rolls 'Q Kettle _ . Accumulator Fig. 17. — Palm and Palm Kernel Oil Mill— Craig. treating the kernels, so making it possible to fill in the otherwise idle period, and to keep the staff together. A Palm and Palm Kernel Oil Factory. In Fig. 17 we give the general lay out of an African mill working on Messrs. Craig's " Caledonia "' dry system. The equipment of this mill includes preliminaiy screens for removing any sand or other material which the natives may be tempted to mix with the fruit, three depericarping machines, three brush machines, six combined nut- cracking and kernel-separating machines, three reducing mills, three sets of shredding rolls, three heating kettles, and three ciushing presses. The latter are of the bar cage t3'pe, to be described in a later chapter, and have each three .sections, namely, a preli- minary, an intermediate, and a tuiishing press. Equipment is also provided for sealing up the palm oil in tins as soon as it has been expressed from the pericarp. The design of this factory is such as to enable it to deal with about 50 tons of fresh fruit daily or with the kernels derived from about 100 tons of nuts. 28 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS As showmg the importance of oil palm fruit, we may remark that in 1913 the United Kingdom imported palm oil to the value of £2,326,842. In the same year Germany imported palm kernels to the value of £3,314,278, while other countries, including our own, took together kernels valued at £1,918,974. The outbreak of war greatly affected matters. In 1914 our imports of palm kernels were valued at £1,411,928, and in 1915 at about £2,500,000. Even so the palm fruit industry may yet be said merely to be in its infancy. CHAPTER V PREPARATORY MACHINERY FOR COTTON 8EED AND CASTOR SEED Cotton seed, as we remarked in our second chapter, is, so far as the oil niillmg industry is concerned, of two varieties, one being the black Egyptian seed, the husk of which as received is practically free from adhering cotton fibre, and the other the white American or Indian seed, to which quite a considerable quantity of cotton fibre may be adherent. The " white " seed is white merely by virtue of the adhering lint. In Fig. 18 wc reproduce a photograph of some samples of the two varieties of cotton l>"iG. IS. — American aud Kgyjitiau Cotton Seed. seed, kindly supplied to us by Rose, Downs & Thompson, Lt 1. At A the American seed is showii, and at B the Egyptian. The husks C and D respectively are hard and tough, the American being, if anything, harder and tougher than the Egyptian. The oil-bearing kernels E, F, are soft yellow or whitish bodies which can readily be crushed between the fingers. The American kernels are di.stinctly smaller than the Egyptian. The engraving is facsimile as to the size of the seeds. La this comitry it is a common custom in the production of cotton .seed oil simply to reduce the seed as received between rolls and then to press the resultant meal in the usual way. In this way the husks and, in the ca.se of the American seed, the adhering cotton lint pass into the cake. There does not appear to be any serious agricultural objection to this course, for cotton seed cake is in great favour as a cattle food. An excessive amount of lint in the case of the American .seed, such as is some- times fomid on seed that has been badly ginned, would, undoubtedly, lower the value 30 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS of the resultant cake. Moreover, the excess lint has a distinct commercial value as cotton. Hence, for two reasons it may well pay the oil mills handling American cotton seed to re-gin or de-lint it as a preliminary to treating it in the rolls and presses. De-linting. A cotton seed de-linting machine as erapljoyed at an oil mill is almost identical lith a cotton gin as employed by the cotton grower. In Fig. 19 we give a drawing the original of which was supplied to us by Rose, Downs & Thompson, Ltd., showing the general arrangement of a cotton seedde-linter. The seed delivered to the machine at A is admitted by a power-driven feed roller in an even stream into the seed box B. One wall of this box consists of a grating C through which project the tips of a large number of fine-toothed circular saws. These saws number usually 106, and are spaced apart on the shaft D by means of thin cast-iron distance washers. The saw cylinder runs at 375 revolutions per minute. The seed in the seed box is churned up by the saws, the teeth of which catch on the lint and remove it in great part from the seeds. The de- linted seed escajies from the shoot E under the control of a hinged regulating board, not shown in the drawing. The lint adhering to the saws is picked off the teeth by a circular brush mounted on the shaft F and revolving at about 1,360 revolutions per minute. From this brush the lint is deflected into a flue G by means of an air draught produced by a fan on the shaft H. The draught is regulated by the damper J and the handle K. From the flue the lint is delivered into a " condenser " L, a casing containing a revolving cylindrical cage of wire cloth, on which the lint collects as a roll and from which it is removed from time to time. This machine absorbs from 4 to 8 b.li.p., and can treat from 3 to 20 tons of seed per twenty-four hours, according to the nature of the seed and the extent to which it is desired to de-lint it On the average it may be expected that round about 20 lb. of lint will be obtained from a ton of seed. The presence of iron particles amongst the seed fed to the machine has to be guarded against, because of the very destructive effect such material would have on the saw teeth. It is, therefore, a common L Engineer" Plan Swain Sc. Fig. 19. — Cotton Seed De-linter -Rose, Downs. MACHINERY FOR COTTON SEED AND CASTOR SEED 31 32 THE PROBUCTIOX AXD TREATilENT OF VEGETABLE OILS practice to embody in the seed box a series of electro-magnets over which tha setd is compelled to pass before it reaches the saws. Decorticating Cotton Seed. Following American practice, it is becoming common in this countiy, in some cases, to remove the husks or cortex of the seed? before crushing and pressing them. In this way the kernels, or " meats '' as they are called, alone are pressed. The advantages of this practice lie in the freedom from discoloration of the oil, otherwise liable to be produce:! by the colouring matter in the husks, the improved quality of the cake, and the increased output of oil obtained from a press of given size. A decorticating machine for cotton seed, made by Rose Downs & Thompson, Ltd.. is shown in sectional elevation in Fis:. 20. The machine mav be described a-< Fig. 21. — Castor Seed — Pods, Beans and Kernels. consisting of a rotating barrel carrying ten knives crosswise on its periphery, and of a fixed " breast '" carrying three stationary luiives similarly disposed crosswise. The seed is fed on to the barrel from an overhead hopper by means of a power-driven fluted feed roll, working in conjunction with a hand-regulated shutter across the hopper mouth. The " breast "" is made in four sections, the divisions being coincident with the planes of the central lines of the three " breast "' knives. The seed falling on to the rotating barrel is caught between the rotating and fixed knives. The husks and kernels together are carried round to the lower edge of the " breast '" and are there collected. This machine is made in several sizes. That size illustrated has an output of about 10 cwt. per hour, and to drive it absorbs some 6 b.h.p. The knife barrel in this case runs at 1..500 revolutions per minute. Considerable mechanical interest attaches to the method adopted by the designers of this machine for carrying the ban-el blades. Three conditions have to be met. First, the knives have to be readily adjustable radially to suit possible variations in the size of seed deli%'ered to the machine for treatment. Secondly, the knives have to Ije easily removable, so that the\- may be taken out and sharix;ned. Thirdly, the MACHINERY FOR COTTON SEED AND CASTOR SEED 33 luiives must he fastened in some particularly secure manner, to withstand the centrifugal force on tliem arising from their high speed of rotation. To fulfil these requirements, no attempt, it will be seen, is made to fix the knives directly to the barrel itself. The barrel is simply slotted to allow the knives to pass through it. In each side frame of the machine a circular central hole is formed. Through these holes the ends of the knives project. Beyond each frame a flanged and slotted disc is fixed to tlie banel The Engineer" Fig. 22. — Castor Seed Sheller— Rose, Downs & Thompson. shaft. Tiie ends of the knives are carried through the slots in these discs and are gripped in the slotted heads of bolts radiating inwards from the disc flanges. In use, the flanged discs are enclosed within a stationary siieet metal casing to prevent accidents. A similar method is adopted for securing the " breast " knives in place. In this case the slot-headed bolts are attached to bosses projecting from the main framing. The further treatment of cotton seed requires no special remark. It is crushed, preparatory to pressing, in rolls closely similar to or identical with those used for linseed or even copra, see Figs. 4, 5 and 9, Chapter III. 34 THE PRODUCTION AXD TREA-OnrST OF VEGETABLE OILS "t-e E-.csEta" Fig. 23.— S: Thompson. differs in no essential respect from the foim used in many other industries. Edge ninners are to be found in employment in chocolate and confectionery factories, for mixing mortar, m paper mills, under the name of "kollei^angs," for reducing "broke" 42 THE PRODUCTION AND TREATMENT OF \T:GETABLE OILS paper to pulp, and elsewhere. An oil mill edge runner, as constructed by Robert Middleton & Co., of Leeds, is illustrated in Fig. 27. It consists, as usual, of two stones mounted on an axle, which is rotated in a hoiizontal plane by means of a vertical shaft, which is set somewhat nearer one of the stones than the other. In the case illustrated the stones are i ft. in diameter and 12 in. wide. The driving pulley nuis at about 100 revolutions per minute, and rotates the vertical spmdie through bevel reduction gearing at about 20 revolutions per minute. The action of the rimner depends as usual upon the sUpping which takes place between the edges of the stones Fig. 29. — Eeducins Mill and Cake Breaker — Robert Middleton. and the 1 :cd stone on which they run. A pair of sweepers is carried romid with the vertical shaft.' These sweepers can be adjusted to guide the material beneath the stones or to turn it outwards so as to discharge it through an outlet door in the pan sun-ounding the bed stone. In the case of machines intended for reducing linseed, etc., the runners are commonly made of selected hard grit stone. For reducing olives, a common apphcation of the edge runner, they are usually made of granite. The edge runner illustrated was designed to deal with about 4 tons of oHves per day, and absorbs in its driving about 8 h.p. Other alternatives to the ordinary rolls take the form of special grinding or reducing miUs and disintegrators. SOME SPECIAL FORMS OF REDUCTION MACHINERY 43 Grinding and Reducing Mills. A special form of grinding mill suitable among other things for finely grinding palm kernels and copra, is illustrated in Fig. 28. This machine is made by Rose, Dowiis & Thompson, Ltd., of Hull. It contams two pairs of finely fluted rollers, the bearmgs of which are acted upon horizontally by springs which peimit the lollers to " give and take " with the feed. The material is fed from the hopper at the top to the openuig between the first pair of rolls, and thence passes through an intermediate hopper to the second pair of rolls. The special feature of the mill lies in the provision, beneath one of the first pair of rolls and beneath both of the lower pair, of concave plates between which and the associated roller the material must pass before it proceeds farther on its course. The position of the upper concave plate is adjusted by means of two screwed rods provided with springs and hand wheels. The two lower plates are similarly held up to their work and adjusted by means of weighted levers. The discharge of the material takes place simultaneously from each side of the lower pair i p-/-4g Auto. Feed 100 R PM. "The Engineer" 2 6i H FUi. 30. — Ili^^integl•ator — Rose, Downs & Thompson. Swain Sc. of rolls. The machine illustrated has an output of from 30 to 40 cwt. of palm kernels or copra per hour. Its belt pulley runs at 300 revolutions per minute. About 24 b.h.p. is required to drive it. Its I'olls are 16 in. long and 12 in. in diameter. The object of fitting the concave plates beneath the rolls of this machine is clear. Each concave is in its effect equivalent to the provision of an additional roll or pair of lolls. Thus, in the case of the machine illustiated in Fig. 28, the material is reduced to the same extent TREA'nrENT OF ^t:getable oils top casting. To the inner side of each of the four columns of the press a flat, square- edged runner or giiide is pinned. The width of the plates is a loose fit between these guides, while sqtiare-headed studs, screwed into the jjlate edges, engage the outer faces of the guides, and prevent the plates moving length^vise. This method of supporting the plates secures the required condition, namely, that the plates, when Fig. oH. — Batterv ul iinu .lUtrn-.^Uiencan i'resse? — ManloTe, Alliott. pressed upwards, should close together without friction, or at least without cumulatiTe frictional resistance. Were this condition seriously departed from there would be a danger of the upper cakes Ijeing less thoroughly pressed than the lower. The press plates are corrugated in the manner shown in the engraving. These corrugations, as well as the longitudinal ridges which are raised on the plate, are intended, as far as possible, to prevent the meal from spreading when the pressure is applied. If a brand mark is required on the finished cake the desired letters, etc., are OIL PRESSES— ANGLO-AMERICAN TYPE 59 raised or sunk on one side of each press plate. The press plates are, in the instance illustrated, of steel. They are frequently rolled to the required formation. Occasionall\- they are built up from steel plates. When a brand mark is required on them they are commonly made of malleable cast iron. The same presses may not, however, always be used for one class or quality of material. Under these circumstances, to avoid having to change the i)lates to obtain merely a different brand mark, plates are made in which the brand mark is formed on a removable portion. The ram of the press illustrated in Plate I. is 16 in. in diameter. The working pressure is 2 tons per square inch, so that the total force exerted is some 400 tons. This gives about § ton per square inch a.s the pressui'e exerted on the meal cake. The hydraulic pressure is some- times transmitted to the ram cylinder by means of water alone, or of water mixed with glycerine, to prevent the liquid from freezing too readily. Fre- quently, however, the working fluid preferred is oil, and, if possible, oil of the same nature as tliat being extracted, the reason being that any leakage of tiie working fluid from the cylinder into the tray catching the expressed oil is thus rendered harmless in its effect upon the oil being recovered. Anglo-American presses are usually arranged in oil mills in sets of four, as shown in Fig. 39, where a battery, made by Messrs. Manlove, Afliott, is illustrated. The fourth press in this vie^ is represented without its plates. The presses are eni irely separate. Some- times, however, they are to be found provided with a common gutter or tray for catching the oil. They are worked separately but in unison. Thus, while one press is being charged another is having the pressure applied to it, a third is standing under the pressure, and the fourth is l)eing unloaded. The pres.ses illustrated in Fig. 39 are of the same size and general design as that repiescnted in Plate I. One secondary point of difference is to be noticed in tiie arrangement made for sup])orting the plates when the ram is lowered and the press is leady for chaiging. Instead of the links shown in the drawing, the long edges of tlie plates are formed with two projecting ears. The gap between these ears is the same on all the plates, and fits on to a vertical flat bar fixed between the toij and bottom castings on each side of the press. The breadth of tlie ears decreases from plate to plate downwards, so that they may pass farther and farther down between a pair of inclined bars similarly fixed, and ^\ith steps cut on their facing edges. This " laider "' airangement, as it is called, has the advantage that it dispenses with the need for any additional means of fixing the position of the plates. No runners are required on the insides of the four cohunns, for the inclined bars defuiitely fix the position of the plates crosswise, while the vertical ^0. — Small AuKlo-Americau Press, 60 THE PRODUCTION AND TREATMENT OF \^GETABLE OILS bars fix their position lengthwise. Instead of solid ears four pins are sometimes to be found on the long edges of the plates fulfiUmg the same f miction. The above examples may be taken as representing the standardised design of Anglo-American presses. This standard design, it may Ije repeated, uses a ram 16 in. in diameter, and a working pressure of 2 tons per sc^uare inch on the ram. It turns out at one pressing sixteen cakes, each weighing from 10 lb. to 12 lb. Various other sizes of press are, however, made, rangmg from a twelve-cake to a twenty-cake press. The pressure emploj^ed in these presses is in general the standardised 2 tons, and as a rule the diameter of the ram in inches is equal to the number of cakes made at one pressing. For special pui-poses presses using 3 tons per square inch are made in the larger sizes. Small Akglo-Ajierican Plaxt. As a contrast to these regular-sized presses, we illustrate, in Fig. 40, an Anglo- American press, by Messrs. ilanlove, AUiott, which is built to a very small scale. The engraving shows a complete self-contained plant, comprising a set of four-high chilled reducing rolls, a heatmg kettle with steam jacket, meal-moistening an-angement and agitating gear, a meal-moulding machine, and a twelve-cake press, operated by hydrauhc power. The plant is capable of dealing with about a ton of seed — containing not more than 3.5 to 40 per cent, of oil — per day of eleven hours. With the exception of the moulding machine, which is operated by hand, the entire plant is driven from the belt pulley at the right-hand end of the overhead comitershaft. The hydraulic power for the press is supplied by a horizontal pump mounted on the press head, and driven from the kettle agitator shaft by means of an excentric and a connecting rod. It will be noticed that the press plates in this example are coupled together by two sets of " lazy tongs."' About 8 to 10 b.h.p. is required to drive the whole plant. For treating verj' oily material similar smaU-sized plants are made with a press of the cage type. In both forms two presses equivalent in output to the one shown in Fig. 40 are sometimes suppUed, so that the whole plant may be run more or less continuously. Limitations of the Anglo-American System. The Anglo-American type of press is undoubtedly an efficient piece of machineiy, and possesses certain well-marked advantages. Thus, it is comparatively simple and straightfonvard in design. Running in conjunction with a modem meal-moulding machine it is easily and quickly loaded. It is equally easily unloaded after the meal has been pressed, although in this connection it is to be remarked that the stripping of the press bagging from the cakes taken from the press may involve considerable labour, so much so that in some mills it has been thought advisable to install special machines which permit the stripping to be performed mechanically instead of by hand. At the same time the general design of this type of press is not altogether free from disadvantages. One obvious drawback lies in the fact that the cake of meal is pres.sed only on its two faces and not simultaneously roimd its edges. This defect is partially compensated for by corrugating the press plates in the manner we have explained, so as to prevent or reduce the tendency of the meal to spread. This expe- dient is more successful with some materials than with others. Thus, if the material being cnished is castor seed, copra, palm kernels, or such like substances of a very oily nature, the mobility of the material, arising from its high oil content, commonly results in the meal spreading excessively, however the press plates may be shaped. If the spreading is excessive more oil v>i\\ be left in the cake than is desirable OIL PRESSES— ANGLO-AMERICAN TYPE 61 or profitable, so that the cakes will probably have to be again reduced to meal, and pressed a second time. Second expression oil does not, however, command as liigh a price as first expression oil. Consecpiently, as a general rule, it is the oil seed crusher's endeavour to extract as much oil as possible at the first expression. A second disadvantage of the Anglo-American system of press is, like the first, of importance onlj" when very oily material is being handled. In forming such material into rough cakes in the moulding macliine it is difficult to carry the compression as far as it should go without expressing some of the oil from the meal. If this undesirable expression of oil is to be avoided, the rough cakes put into the main press must be less compressed, and consequently thicker than usual. As a result the press has to be made taller by a corresponding amount, and the movement of the ram has to be increased in order to make good the deficiency in the preliminary compression of the meal. Under the high hydraulic pressures in use the latter item reacts unfavoiuably on the upkeep charges of the press generally, and of the ram and valves in particular, and, in addition, increases the amount of jiressure fluid used at each movement of the ram. It will be gathered from these remarks that the Anglo-American type of press is best adaj^ted for dealing with seed, etc., containing a moderate amount of oil. For very oily seeds a press is required which, in the first place, supports the layer of meal round its edges while pressure is being applied to its faces, so that the meal may be evenly pressed and prevented from spreading, and which, in the second place, will be able to work without the assistance of a separate preliminary moulding machine. Sucli an appliance is the box cage type of i^ress, which, in several different forms, has recently come into extensive employuient, following ujjon a great increase in the amount of veiy oily seed, etc., received for treatment. In our next chapter we will illustrate and describe tj'pical examples of cage presses. For the present we need only say that the cage tj^pe of press is itself not wholly free from disadvantages peculiar to its design. CHAPTEE IX OIL PRESSES— CAGE TYPE Ha\tng described and disctissed in our preceding chapter the construction, working, and limitations of oil presses of the Anglo-American type, we wUl now proceed to deal similarly with pres.*es of the cage type. Cage Peesses. In contrast with the Anglo-American type of press, the design of which has reached a notable degree of standardisation, the cage type of press is made in many forms which are sufficiently distinct to merit some sort of classification were such a course likely to lie of any value. On examination, however, it will be found that the different forms differ more as regards the general arrangement of the presses, less in the details of their design and not at all in their princ.ple of action. As illustrating the design of a cage press and its method of working, we give in Plate II. the reproduction of a drawing — specially prepared for us by Messrs. Manlove, Alliott, of Nottingham — sho^ving the construction of a cagp press as made by this firm. Like the Anglo-American press by the same makers, described in our previous chapter, this cage press is provided with a cast-iron head, four forged -steel columns with buttress threads, a cast-steel cylinder and a cast-iron ram. The bottom of the press is formed solidly with the cylinder — a practice frequently followed also in the design of Anglo-American presses — and is therefore of cast steel. A circidar cast- iron oil tray, provided on its underside with four bosses through which the columns pass, rests on top of the cylinder casting. The ram rises through a gland at the centre of this tray. Just above the tray four split muffs a*ie bohed round the columns to provide stops whereon the cage in its lowest position may rest. The CoxsTRCcnox of the Cage. The cage consists, first, of a cast-steel top piece and a cast-steel bottom piece, bored centrally and formed with four ears or comers suitably concaved to fit on to the press cohunns. Between these two castings are arranged a number of vertical rolled-steel bars, the ends of which are nicked to fit into an annular recess in the top and bottom castings, as clearly shown in the sectional elevation given in Plate II. The bars are of T section with, however, the horizontal limb reduced to a mere fillet on either side. The bars are 2 in. deep, while the fillets measure f in. The faces of the fillets are machined very lightly so as to leave alternate high and low portions at 5 in. centres throughout their whole length. The bars when assembled thus bear against their neighlx)urs at every .1 in. of their lengths, while between these bearing points narrow spa«-s are left. The width of these spaces is made to suit the class of seed to be pressed, and ranges from 47?,to in to c\ in These spaces have to be sufficiently large to allow the expressed oil to flow away freely through them, but not so large that the meal also can pass out of the cage. Externally the bars are stiffened by a series of weldless steel rings. Three vertical tie reds lying jjist outside the rings imite the top and bottom castings of the cage. On to these are slipped a INSERT FOLDOUT HERE OIL PRESSES— CAGE TYPE 63 number of ferrules which, fitting between the rings, hold these at the proper distance apart. Surrounding the bars and rings there is a cylindrical lagging of sheet steel united tu the toj) and bottom castings of the cage. This lagging prevents the expressed Fig. 41. — Two Small Cage I'lx -Mauluvn, Alllolt. oil from splashing and helps to guide it into the collecting traj' belo\N. In Fig. 41 we give a view of two small cage presses made by Messrs. :\Ianlove, Alliott. This engraving helps to make clear the construction of the cage. The diameter of the cage is equal to that of the ram, which is 16 in., just as it was in the case of the Anglo -American press dosciibcd in the preceding chapter. 64 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS The pressure in the present instance is 3 tons per square inch as compared with 2 tons employed in the cylinder of the Anglo-American press. This full pressure acts on the meal, whereas in the Anglo-American press the pressure on the cakes is much less than the pressure in the cylinder. The pressure, in fact, is, as we stated, but I ton per square inch, whereas in the cage press illustrated in Plate II. it is 3 tons per square inch. The employment of such a hig'.i pressure as this on the meal in an Anglo-American press would be next to impossible, for it would cause even the least oily of meal to spread excessively between the plates. Its adoption in the cage press has only been made possible by reason of the care and thought which have been given Fig. 42. -Cage Press and Kettle in a Mill. to the construction of the cage. It may be taken that the outward pressure on the bars of the cage is practically 3 tons per square inch, for the meal during the pressing acts almost like a liquid forced out of a cylinder through a restricted orifice. Under these conditions the greatest care is necessary in designing the cage to ensure that the bars shall not twist or bend and that the spaces between them shall remain constant. How this is secured we have explained. We need only add that the cages of similar presses made by other firms are constructed on the same principle. Immediately above the cage, as shown in the sectional elevation in Plate II , a cylindrical cast-iron head is slimg from the underside of the top casting of the press. This head is moimted on a four-wheeled carriage which runs on a pair of fixed rails. The movement of the head is effected by means of a hand wheel and pinion engaging with a rack on the head The rack is sunk into the top of the head so that a plain OIL PRESSES-CAGE TYPE 65 bearing siu-facc may be provided between the head and the underside of the top casting. Four square-lieaded studs fixed to the rails form stops which limit the movement of the head. By these means the head can be brought directly over the cage when every- thing is leady for pressing, or removed to one side to facilitate the loading or unloading of the cage before or after pressing. Method of Working of a Cage Press. In Fig 42 \\L' give a view of a press of the type shown in Plate II. as actually arranged in an oil mill. It will be noticed that the press is disposed about half abovo and half below the working floor level, and that a heating kettle is placed close beside it. \\'hen the cage is resting on its bottom stops the top sui-face of its upper casting is level with the surface of the plate, hung beneath the kettle, on which the strickling bo.\ slides. With the movable press head nni back out of the way the strickling box with a charge of meal can thus be pulled over to discharge its contents into the cage. Before this is done, however, a circular steel plate is dropped into the mouth of the cage so that it maj' come to rest, a short distance down, on four catches projecting through the walls of the top casting of the cage. A circular sheet of press bagging is placed on top of the plate. Thereafter the meal is striclded in a layer into the cage mouth. The plate catches are mounted on a ring so that they may be withdrawn simultaneously to allow the plate, bagging and layer of meal to drop down on to the head of the ram. The fall allowed is not great, however, for the ram to begin with is run up almost to the top of the cage, and as the loading proceeds is allowed to descend slowly to keep pace M'ith the formation of the layers of meal. The layers thus formed differ from the cakes placed between the jjlates of an Anglo-American press in the fact that the meal is cpiite uncompressed. To take full advantage of the capacity of the cage, therefore, strickling is continued rmtil the ram reaches the bottom of its stroke. In this condition about haK the depth of the bottom casting of the cage is filled with meal. When jiressing commences this meal is at an early stage forced upwards into the cage proper and there, partially at least, makes good the reduction of voliune suffered by the general body of the meal. When the cage is fully charged the movable head is run back over it. The end of the head is turned to a good fit with the bore of the cage. Just before pressure is applied to the meal by the ram, ^^ressure is admitted to two auxiliai-y ram cylinders — see A, Fig. 41 — which, acting beneath the lower casting of the cage, hft the cage a short distance upward so as to cause its mouth to 2Jass on to the cyUndrical liead and so close the joint. Pressure is then admitted to the main cylinder. As the expression of the oil proceeds the cakes of meal become bound tightly against the walls of the cage. The friction thus developed round their edges is sufficient to lift the cage still further on to the head as the compression of the meal increases. In other words, no ])rovision is made to bring the cage as it rises up against a dead stop. This is an important point. Were such a dead stop in existence the friction round the edges of the cake would result in the cakes being subjected to an effective pressure whicli would decrease from cake to cake upwards. As it is the cage " floats "' with the meal, etc., inside it, and the effective pressure is the same on the top and bottom cakes. It is not, however, necessarily the same towards the middle. When the cakes have stood for a sufficient length of time under pressure, the main ram is set to exhaust until the cage is lowered on to the bottom stops. The auxiliarj' hydraulic cylinders are arranged to act as buffers for the cage so as to biing it cpiietly to rest. The movable press head is then run out. In its outmost position it does not clear completely the face of the top casting of the cage, and therefore forms 66 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS a stop which will for the time being prevent the cage from rising. Two additional stops are ijrovided for the opposite side of the top casting of the cage. These two stops consist of half caps -which can be swung roinid on the pillars of the press. With these three stops in action, pressure is once more admitted to the main hj-draulic cylinder, so that the ram. rising, may force the cakes out of the cage. To facilitate the ejection of the cakes the bore of the cage is slightly tapered, so that its diameter at the top end is a small fraction greater than at the lower end. Thus a slight upward movement of the cakes, etc., in the cage is sufficient to relieve the binding pressure round their edges. In general the cakes produced in a l^ress of this type are again reduced to meal whicli, after being heated, is expressed a second time. A cage jjress can be used for this second expression, but as the material has. now had the bulk of its oil removed, it can quite convenientlj' be treated in an Anglo- American press. .Vlterxative Cage Press Arrange- ments. The press just described is — as shown in Fig. 42 — arranged to work by itself in conjunction with a separate meal-heating kettle. Very frequently, however, cage jiresses are worked in pairs or in sets of three, the two or the three presses in each set being quite indepen- dent, except in so far as they are fed from a common kettle. Li Fig. 43 we give a view of a two-press set, made by A. F. Craig & Co., Ltd., of Paisley. The general arrangement of this set is shown in Fig. 44. This engraving incidentally indicates the nature of the foundations required for an oil press. The presses, except for one or two obvious minor differences, arc similar in design to that already descriljed. The method of charging the cages is, liowcvcr, quite Fig. 43. — Twin Ciige l're»ses — t'laig. different. A heating kettle is arranged over the presses, being supported partially on the press heads and partially on an extra pillar. The bottom of the kettle is pro- vided with two outlet holes which register with a hole formed at the centre of each press head — see Fig. 44. Each outlet is controlled by a pair of shutters, one above and one below the press liead. These two shutters are comiected so as to be operated simultaneously. Thus a double movement of the control liandle fills the hole in the press head with meal and then discharges this measured quantity into the cage. An alternative arrangement by Manlove, Alliott & Co., Ltd., is indicated in OIL PRESSES— CAGE TYPE 67 Fig. 45. This arrangement is particularly suitable where large presses are required. It consists of a battery of four presses, a separate compressor and extractor press and a power-driven travelling carriage. The compressor and extractor press is i^rovided with a movable head, as in the case of the presses described above, so that its cage The Engineer" Swain Sc. Fig. 44. — Twin Cage Presses — Craie;. may readily be charged with layers of meal from the adjacent kettle. The cage may be regarded as being in two parts, the lower of whicli is fixed, in so far, at least, as the position of its centre line is concerned, while the upper part can ho run out on to cross rails on tiie top of the travelling carriage. The strickling proceeds until both parts are filled with layers of meal. Pressure is then applied in the ram cybnder so as to com- 68 THE PRODUCTION AXD TREATMENT OF VEGETABLE OILS press all the meal into the upper movable part of the cage. When this is accomplished the pressure is released, and the movable part, of the cage is run out on to the travelling carriage, which then transports it to one of the four main presses. The travelling carriage is provided with two sets of cross rails so that it maj^ support a cage ready for pressing, while giving accommodation for the reception of a cage the pressing of which has been completed. On the return joumej-, therefore, the carriage brings back to the preliminary press a cage from which the cakes are ready to be extracted. This extraction is performed at the preUminary press in the manner indicated ah-eady for the single press by the same makers. '^^^^■^^^ "Tilt Engineer" Fig. 45. — Battery of Four Cage Presseb — Maulove, Alliott. The preliminary press used in this system of working is, we think, to be regarded properly as the equivalent of the moulding machine required with an Anglo-American type of press. Its adoption has the distinct advantage that tlie main presses can be made with fixed, and not sliding heads. Further, it will be gathered that as the " slack " in the meal is taken up in the preliminary press, the movement, and therefore the length, of the rams in the main presses can be made quito siiort. Revol%tn'g Cage Press. Another interesting and important alternative arrangement of working cage presses in groups lies in the adoption of a rotarj' principle. An example of the apphca- INSERT FOLDOUT HERE OIL PRESSES— CAGE TYPE 69 tion of this principle is illustrated in Plate III., where we show a revolving cage press made by A. F. Craig & Co., Ltd., of Paisley, under Craig and Morfitt's patent, A photograpli of such a press is reproduced in Fig. 46. Fig. 4G. — Revolving Cage Press — Craig. The fundamental featuic of the design hcs in the provision of three cages arranged with their centres at the apices of an equilateral triangle, the whole being rotatable as a block round an axis passing through the centre of the triangle. Corresponding to the three cages there are three fixed press heads ?,nd three hydraulic cylinders and rams, arranged with their centres at the apices of an identical equilateral triangle. 70 THE PEODUCTIOX AaD TREATMENT OF \'EOETABLE OILS During a complete rotation of the cages about their common axis each cage passes in tuni between each press head and its corresponding hydraulic ram. The method of working is to fill cage A — see Plate III. — with meal from an adjacent kettle, while it stands beneath one of the press heads, and to give the meal in it at this point a iJieiiminary com^jression. The cage system is then rotated clockA\ise through 120 degrees, so as to bring cage A beneath the second press head where the meal is subjected to an intermediate compression, and so as to bring cage C into the position formerly occupied by cage A. Cage A is allowed to stand under pressure while cage C is being emptied and recharged ^^"ith meal. Thereafter a further rotation of the cage system through 120 degrees brings cage A beneath the third press head where its meal receives the final compression. Meanwhile, cage C is receiving its intermediate compression beneath the second press head, while cage B beneath the first press head is Ix-ing emptied and recharged. A final rotation of the cage system brings cage A back again beneath the fii'st press head for emptying and recharging. The meal, it \\ill be seen, is pressed in three sepai-ate .stages. It will also be gathered that the arrangement secures practically continuous working. It is stated that even"thing about a set of these presses can be worked by one \niskiUed man with the assistance of a boy. The detailed design of the press is noteworthy. The meal-heating kettle is of the usual type, and is fitted with the usual means of stirring, heating and moistening the meal. It is sujaported partly on the first press head and jjartly on two separate colunnis. The strickling box slides on a surface with guiding edges formed on top of the fii-st press head. A circular hole equal in diameter to the bore of the cage is formed in the press head, and is provided with four catches, o^jerated simultaneously, for temporarily supporting the usual steel disc and cireular piece of press cloth on to which the meal is deposited from the strickling box. The withdrawal of the catches allows the plate, cloth, and layer of meal to fall on to the ram head. Dm-ing the charging operations, the ram starting from its highest position is allowed slowly to faU. Oliarging is continued until not only the cage, but the compression chamber — D in Plate III. — beneath it is completely filled. ^Mien this stage is reached, a run-out shde, operated by racks, pinions and hand wheel, is moved back to close the opening in the press head from the underside. Tliis shde is provided with a shallow circular boss turned to fit the bore of the cage. ^Yhen matters are in this condition, two auxihaiy hydraulic rams are brought into action beneath the compression chamber wliich. lifting this chamlier and the cage. clo.-*e the joint between these two parts and also the joint between the cage and the run-out slide beneath the press head. Pressure is then appUed beneath the main ram which, rising through the compression box. compresses the meal entirely into the cage. A cei-tain amomit of oil is forced out of the meal at this stage, and is caught iji a tray beneath the press. Even before the meal is entirely pushed out of the com- pression box, the pressure may be sufficient with some seeds to express a portion o.' the oil from the meal. For this reason the top end of the compression chamber is finely perforated, so that the oil expressed may escape readily. "\Mien the pressure of the main ram is relieved, the meal layers tend to expand a little. To obviate any trouble which this expansion might cause when it comes to rotating the cages, the meal is compressed further into the cages than would be necessary were expansion absent. The compression suffered by the meal in the preliminary press is sufficient to bind the buUc of it witliin the cage, so that when the ram falls it remains there. The steel plate and the press cloth at the foot of the cage and one or two of the lowest layers of meal require, however, to be supported when the ram is lowered and while OIL PRESSES— CAGE TYPE 71 the cage is being *i ■1 Wmi 1 -k^^^^^^^lal m ^ isl ly ^^^3BI ^^^^^1 ■M _. ■ - '^w ^^^^H iH^^S...:.^S Fig. 52. — Accumulators in aii Oil Mill — Manlovo, Alliott * and head room than is strictly economical. When, however, the number of presses in tlie mill exceeds a certain figure, a point is reached at which both first cost and floor space will be economised by providing accumulators — worked by one or two pumps — rather than separate pumps for each press. Practice implies that this point is in general regarded as having been reached, when there are eight or more presses in the mill. When accumulators are installed, it is usual to find one arranged for a high -pressure supply working in conjunction with one or more for a low-pressure supply. An additional advantage of using accumulators in large mills lies in the fact that the low- pressure supply can be utilised for working the meal-moulding machines, if the presses arc of the Anglo-American type, the hydraulic cake-trimming machines referred to THE GENERAL ARRANGEMENT OF OTL :MTLLS 79 later on in this chapter,, and any hydrauHc lifts with wliich the mill may be equipped. It is frequently stated that a still furthei advantage attending the use of accumulators is to be found in their " safety-valve action," for in general it is impossible to admit to the press cylinder and its connections a pressure greater than that for which the accumulatoi' is deliberately loaded. Tiie implied danger, however, need be no greater with a properly arranged system of independent pumps than it is with accumulators. The accumulators used in oil mills are in no essential i-espect diffei'ent from those used for other purposes. The low-pressure accumulator — see Fig. 51 — may be loaded with cast-iron weights, but usually it is loaded similarly to the high-pressure accumulator sho\\^l also in Fig. 51, namely, by means of a mild steel plated case filled with slag, sand, or other material. The cylinder is of cast iron, formed externally with two ribs, on which the weight case is guided. The weights are hung from a cast-iron cross- head fixed to the top of the ram. Sometimes matters are reversed, the ram being fixed to the floor and tlie cylinder, with the weight case attached to it, s'iding on the ram. The low-pressure accumulators are commmly designed for a pressure of 500 lb. to 600 lb. per scpiare inch. Their rams may have a diameter of from 8 in. to 15 in., and a stroke of from 8 ft. to 12 ft. The high-pressure accumulators give a working pressure of 2 or 3 tons per square inch, accord- ing as the presses are of the Anglo- American or of the cage type. Their rams varj' in diameter from 2|- in. to 5 in., and in stroke from 5 ft. to 12 ft. Li Fig. 52 we give a view of three accumulators by Manlove, Alliott & Co.. Ltd., as erected in an oil mill. Here thei-e are two low-pressure accumulators, each with a stroke of lU ft., and one high pressure accumulator with a stroke of 5 ft. — Accumulator I'liiiip — Maiilovo, Alliott. Accumulator Pumps. TJie pumps employed in coimection with the accumulators in an oil mill arc usually of a vertical reciprocating belt-driven type. An example by Manlove, Alliott & Co., Ltd., is illustrated in Fig. 53. The cast-iron base is in the form of a tank into which the hydraulic fluid from the press exhausts and from which the pump draws its supply. A bridge is cast acro.ss the top of the tank, and on this are fixed the side frames carrying the driving shaft. The side frames are fixed to the tank by means of bolts, which extend right from the caps of the driving shaft bearings to the underside of the bridge. The driving shaft carries an excentric at each end. The cxcentric rods arc of cast steel and each reciprocates a crosshead to which two rams are attached. The rams 80 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS work within forged steel blocks fitted into the top of the tank. The suction valves and the delivery valves are fitted with renewable seats of nickel steel. The delivery valves are, as shown in the engraving, arranged in two steel blocks so that access may readily be had to them. The chief point of importance to pay attention to in the design of such a pump as this is the accessibihty of those parts liable to wear or get out of order. Oil mills usually are run both night and day, so that any repairs required to the machinery have to be effected in the short meal time stoppages. Accumulator Relief Valves. As usual, means have to be provided whereby the pumping up of the accumu- lators is stopped when their rams reach a certain height. Elsewhere this is frequently effected by automatic means which stop the pumps. Li oil mills, however, it is customary to keep the pumps running continuously and, when required, to deflect tlieir delivery back through a relief valve into the supply tank. An automatically worked relief valve arrangement by Messrs. Manlove, Alhott is illu.stratcd in Fig. 54. When the weight case of the accumulator reaches its prescribed height, a bar on its crosshead, shown at A in Fig. 52, strikes the end of the lever B (Fig. 54), and pushes it upwards. The opposite end of this lever is connected by a chain, etc., in the manner shown to the weighted lever of the relief valve. This lever is therefore moved up. It is held up by the action of the cam lever which, moving out under the influence of a spring plunger, engages the pin shown on the side of the lever B. When the accumulator falls again the bar A (Fig. 52) strikes the now projecting cam lever C, moves it in, and allows the weight of the lever D to pull down the lever B and so resets the arrangement. When the lever D is raised the delivery from the pump is deflected from the accumulators back to the supply tank. When the lever falls again the deUvery to the accumulators is resumed. The actual raising of the weighted lever is not effected by the chain and rod, but by a plunger beneath it. A by-pass supply of low-pressure fluid is admitted beneath this plunger by a piston valve E operated when the lever B is raised. During the resetting of the arrangement this piston valve acts as a dashpot and prevents the relief valve from being dropped violently on to its seating. Fig. 54 -Belief ^'alve i)etaD8- .rUliott. Manluve, THE GENERAL ARRANGEMENT OF OIL MILLS 81 Cake-trimming Machines. Cake-trimming machines, as will have been gathered from what has already been said, form quite important items in the economy of oil mills. They are made in a variety of forms. A simple arrangement for paring the edges of straight-sided cakes, made by Robert Middleton & Co., is shown in Fig. 55. This machine is attended by two youths and can pare two cakes at a time. It comprises two knives moved by power along the edges of a slot at the centie of its table. The oily parings fall into the slot, where they are caught in a trough and are broken up and moved forward to the spout by a series of steel conveyor knives mounted on a power-driven shaft within the trough. A similar type of machine is made by Rose, Do^vns & Thompson, Ltd., except that all the driving gear is carried on two standards bolted to the table top, the idea being that in this way the woi"ldng parts camiot become clogged with oily cake parings. A machine of this type is indicated in the mill arrangement, Fig. 47. Various designs of automatic cake-paring machines are now coming into use. The Engineer" ***"• So Fig. 55. — Cake-Paring Macbiue — Robert Middleton. Fig. 56. — Hydraulic Cake-Paring Machine — Manlove, Alliott. In a tjTjical example the cakes, one by one from a pile, are moved forward against two knives set at the desired distance apart. Their movement is then continued at right-angles to the first traverse, so that the two remaining edges of the cakes may be 82 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS passed between a second pair of knives. Such a macliine can readily deal with thirteen cakes per minute. There is sometimes a Utile difficulty in getting a clean-cut edge with the abofe machines. Further, their operation calls for a certain amoxmt of skill and judgment, and the cutting knives have to be carefully looked after. These considerations have led to the introduction of hydraulic parmg machines. An example of this class, made by Manlove, Alliott & Co., Ltd., for paring the roimd cakes obtained from cage presses is illustrated in Fig. 56. In this machine the cakes, one at a time, are pushed against stops on the table imdemeath a power-diiven revolving knife having saw-like teeth. When it is thus in position pressure is admitted to a hydi-aulic cylinder beneath the table. The ram head forces the cake against the revolving knife, which trims the cake b\" a combined shearmg and cutting action. The cake, after being trimmed, remains within the circular knife, being held there by automatic catches. As succeeding cakes are trimmed the pile rises into the hollow top of the macliine until, -with each fresh cake fed to the knife, a trimmed cake is ready to be removed from the top of the pile. Tiie ram is conveniently worked from the low-pressure accumulator, but if there is no such source of hydraulic supply the machine can be designed to utiUse steam pressure. Similar machines are made for paring Anglo-American cakes. In these, of course, the action is one of pure shearing, the cakes being forced up against four fixed knives arranged at the edges of a quadrilateral opening in the machine head. CHAPTER XI EXTRACTION OF OILS BY CHEMICAL SOLVENTS We now pass on to describe the second method of recovering oils from vegetable substances, namely, their extraction by means of chemical solvents, such as benzene, ether, chloroform, carbon disulphide, carbon tetrachloride — CCI4 — and tetrachlore- thane — C2CI4H2. The idea of using solvents for this purpose is by no means a recent one. It was introduced as a practical process as long ago as 1843, by Fisher, of Birmingham. It is, nevertheless, true that only recentlj^ has the process come to be extensively adopted, for it has had to struggle against the prejudices inherited from its earlier and admittedly impeifect working. These prejudices are not yet by any means dead, and even in text-books of high standing statements are to be found concerning the results of the process which seem to be based on misinformation as to its modern state of development. Objections Alleged Against the Process. Before proceeding to describe modern examples of solvent extraction plant it is very desirable that we should deal with tlie objections which have been and still are, urged against the process. We may preface our remarks under this head by saying that in brief the process consists of allowing one of the solvents named above to perco- late through the seed or meal in a closed vessel, heated or cold, of draining off the solvent and the oil which it has dissolved from the seed, of transferring the liquid to a heated still, and of then driving off the volatile solvent so as to leave the oil behind. The solvent driven off is condensed and used repeatedly. The objections alleged against the process fall into three principal categories. In the fir.st place it is argued that it extracts the oil so effectively from the seed that the residue of meal is next to useless as a cattle food and, at best, is fit only for manure. Secondly, it is stated that it is difficult or impossible entirely to eliminate all trace of the solvent used both from the oil and the residue of meal, so that the oil is made unfit for edible purposes and fit only for soap-making and kindred uses, while the nauseous taste or poisonous action of the .solvent left in the meal provides a second reason why such meal should not be fed to cattle. In the third place, not one of the solvents used, it is .said, is free from technical objections. Thus ether and chloroform are far too expensive to permit of their use commercially. This .seems to be a sound contention. Carbon tetrachloride, it is urged, is aLso expensive and is apt to exercise a poisonous action on the workers attending the recovery plant. Further, while admittedly non-inflammable, it suffers from the great disadvantage that it very readily attacks metals. Regarding carbon disulphide, it is argued that while it is a very good solvent it is difficult to obtain pure and that it is apt to impart even to soap made from oil extracted with it an unpleasant sulphurous smell. It is further very readily inflammable, and, like carbon tetrachloride, exercises a poisonous action on those working with it. Again.st benzene the chief objection levelled is its inflam- mability. In addition it is stated to be the most difficult of all the solvents to eliminate from the oil and meal. Tetrachlorethane is a solvent of recent introduction. So far as we know it is not as yet in extensive use for the extraction of oils. It seems, however, 84 THE PRODUCTION AXD TREATMENT OF VEGETABLE OILS to possess certain features which may in time lead to its wide adoption. Thus, its commercial production appeais to be simpler than that of carbon tetrachloride, while its action on metals is much less. It must be admitted that certain of the above-mentioned objections are perfect'y sound when applied to the process as carried out with old-fashioned apparatus — frequently of German origin — and using carbon disulphide as the solvent. As applied to modem British -made plant using benzene, as in the case of the system to be described, they appear to be quite out of date and in direct conflict with established fact. It is undoubtedly true that the process, as it can now be carried out, is rapidly being adopted on an extensive scale, a circumstance which seems to afford conclusive evidence that the objections summarised above are now recognised as being no longer vaUd. The Objections Refuted. In refutation of the objections urged against the process it may Ije directly stat-ed that extracted meals are daily being used in large quantities both in this country and abroad as food for cattle, while a mimljer of plants are at work in tliis country using the chemical solvent extraction process and producing nothing but oil of edible quality, as, for instance, oils which are used in the manufacture of first-grade margarine. Here we have evidence that all traces of the solvent used can now be ehminated, both from the oil and the meal. Whether or not the entire absence of oil in extracted meal lowers the value of the residue as a foodstuff is a very debatable point. There are distinct indications that a marked percentage of oil in a cattle food is not quite as great an advantage as it was at one time beheved to Ije. This seems to be recognised by many cattle-feeders themselves and is supported by the views expressed in the recent report of the Government Committee on Oil Seeds, which views tend to the recommendation as a cattle food of extracted meal even when next to entirely free from oil. In explana- tion of this it may be pointed out that while oil is a heat former it is the albumenoids in the material that count from the actual food or flesh -forming point of view, and that extracted meal is richer in these albumenoids than the cake procured by pressing the same seeds. Apart from this question it is to be noted that no oil cake is fed by itself to cattle. It is diluted with bran or other substance. The " other substance "' may very well be extracted meal, which may l>e mixed with the cake to give a foodstuff of the desired oil content. In any event the argmnent against the extraction process, which is based on the deficiency of oil in the residue, entirely falls to the ground when we observe that under modem conditions the operator using this process can arrange to leave as much or as little oil in the residue as he may desire. Advantages of the Pkocess. From the technical point of view the chief advantages attending the adoption of the process he fii'st in the comparative simphcity and cheapness of the plant required ; secondly, in the small amovmt of power absorbed ia driving the plant ; and thirdly, in the fact that the labour demanded for its attendance need not be highly skilled. From the commercial point of view its full advantages can otdy be assessed by a careful study of certain factors which vary from place to place and from time to time. If it be a question whether the press or extraction system shall be adopted, everything turns upon whether or not the seed to be treated yields a residue which, quite apart from the process of recovery used, is in demand as a cattle food. Thus rape seed, even when treated by the crushing process, is not greatly valued as a cattle food. In such cases the only product primarily to be considered is the oil. This naturally EXTRACTION OF OILS BY CHEMICAL SOLVENTS 85 points to the adoption of the solvent extraction process as the better method of treating such material in view of the considerably higher yield of oil which it secures. A secondary consideration points in the same direction. If the residue of the seed is unsuitable as a cattle food, its only other important outlet is as a manure or fertiliser. Press cake has to be broken up and reduced again to meal before it can be used for this purpose. Extracted meal is suitable for it as soon as it is taken out of the extractor plant. Far more important than this, however, is the fact, now well established, that grease or oil in a fertiliser prevents the soil foods from being absorbed by the soil for, if present, it acts to defend the fertihser against the attacks of those organisms which convert the constituents of the fertiliser into immediate soil foods. Clearly, then, the extraction process, ehminating as it can be made to do practically all oil from the residue, has very great claims to attention when the residue has to be used as a fertiliser. If the seed residue, on the other hand, is suitable for cattle-feeding purposes, the first point to consider is whether there is a local market for it in this form. It may well be that there is not, and that, in view of the cost of shipping the residue to the nearest market, the balance is in favour of using the residue as a manure. Here again the adoption of the solvent extraction process is indicated as desirable. The conditions here touched upon arise very often when the recovery of the oil in the neighbourhood where the oil-bearing seed is grown is under consideration. This practice is desirable in itself, for the seed, being fresh, will almost certainly produce a better oil than it would after deteriorating during its journey to some distant factory. There is, however, probably no local or conveniently adjacent market for the residue as a cattle food, and this, up to the present, has led to the shipping of enormous quantities of oil-bearing seed for treatment in this and other countries remote from the country growing the seed. By adopting the solvent extraction process the grower can save freight charges by shipping nothing but oil, and can dispose satisfactorily of the residue by using it as a manure on his own plantations. Commercial Aspect of the Process. We thus see that the solvent extraction process has distinct claims to attention when : (a) The residue is not usable as a cattle food by reason of the nature of the seed itself, and when (6) The residue, although suitable for cattle feeding, is not u.sable in this way by reason of there being no market for it situated conveniently near the mill. A third case arises, namely, when (c) The residue is usable as a cattle food, and can be conveniently disposed of as such. The.se conditions are met with, for example, when it is a question of treating linseed or cotton seed in this country. Which process it is best to adopt under these circumstances is a matter for very close study. Several factors are involved. But in investigating the matter a certain line of argument frequently advanced by tho.se interested in the solvent extraction process should not be too readily accepted. According to this argument it is poor policy to dispose of oil as a constituent of oil cake fetching £12 to £18 per ton, when, by extracting it completely, it can be sold for £40 to £60 per ton. This argument appears to be fallacious, in so far as it overlooks the fact that it is the custom of the oil-seed-crushing industry to charge for the oil cake in such a way that the oil in it reaps the same price as the bulk of the oil separated from the seed. Thus a 86 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS ton of linseed containing 40 per cent, of oil originally, after being crushed, appears roughly as — £ i. d. 747 Ih. oil : value, at £o3 per ton .. .. ..17134 1,493 lb. cake : value, at £19 per ton 12 13 4 2,240 30 6 8 Cake : 1 per cent. oiL 149 lb. oil : value, at £53 per ton 3 10 8 1,344 lb. dry residue : value, at £15 4s. bii. per ton . . 9 2 8 Similarly, a ton of undecorticated Egyptian cotton seed containing 24 per cent. »)t oil originally wiU, after crashing, appear as — £ s. ./. a4S lb. oil : value, at £53 per ton . . 8 4 11 1.S92 lb. rake: value, at £15 10s. per ton 13 1 9 2,240 21 6 10 Cake : 10 per cent. oil. 189 lb. oil : value, at £53 per ton 4 U 4 1,703 lb. dry residue : value, at £1 1 7s. per ton 8 12 5 1.892 13 1 9 Clearly, then, so far as the money reahsed hyhis products is concerned, it does no( matter to the oil crusher how much or bow Httle oil he leaves behind in his cakes He gets the same price for the oil whether he recovers it or allows it to remain in tlie cake. Were he to adopt the solvent extraction process he would not reahse a penny more for the oil contained originally in the seed. At the present moment in this country linseed and cotton seed are crushed rather than extracted, because a demand exists for linseed and cotton seed press cake contain- ing a considerable percentage of oil. Rightly or wrongly, little or no demand exists for linseed and cotton seed extracted meal. On the other hand, rape seed is extra.cted rather than cnished, because no demand exists for rape seed press cake. The oil left in such cake would represent a sheer loss, for the cake could not be sold at a highei figure than the extracted meal. In addition, the oU left in the cake would, as we have already stated, lower the manurial value of the residue. By way of conclusion to this brief discussion of the relative merits of the two processes, we need only remark that they should not Ije regarded necessarily as rivals. The solvent extraction process has a very distinct field of its own. Worked side by side with the crushing process, so as to recover the last portion of oil from the seed, it is of veiy great value in ceitain particular cases, as, for example, when the material to \ye treated is olives. As a direct alternative to crushing its importance is rapidly increasing. When the true value of extracted meal as a cattle-feeding stuff l^ecomes more generally recognised the rivaln,- of the process with the crushing method will no doubt tmdergo great development. The Ideal or the Process. The ideal solvent extraction process, it can be said, should seem* the complete recovery of all the oil in the seed — or as much of it as it is desired to recover — in one stage, and should leave the residue of the seed in a drj" state. It may be remarked that certain extraction processes fall short of this ideal, in so far as the meal after extraction has to be separately dried. INSERT FOLDOUT HERE EXTRACTION OF OILS BY CHEMICAL SOLVENTS 87 Preparation of the Material. Palm kernels, copra, soya beans, and similar materials are prepared for tbe extrac- tion process in precisely the same Avay as for crushing, the only difference being that the flesh need not be reduced or shredded to quite the same degree of fineness. Seeds such as rape seed, linseed, etc., need only be lightly rolled. Cotton seed, castor seed beans, and similar material commonly decorticated before being ciushed can, if desired, be extracted in an undecorticated state, the seed being simply rolled so as to break the cortex. The saving of the expense of decorticating results in considerable economy if the residue is to be used as a fertiliser, or if the skin or shell of the seed being treated possesses, as is sometimes the case, a distinct feeding value. The "Scott" Extraction Plant. One of the best-known forms of solvent extraction plant is that working on the " Scott " system, and made by George Scott & Son (London), Ltd., Kingsway House, Kingsway, London, W.C. LTnder this system the solvent commonly used is benzene. Benzene — or benzol, as it is stiU frequently called in commerce — is, when pure, a colourless liquid having a specific gravity of about 088 at 15° C, and boiling under normal pressure at about 80° C. It is very slightly soluble in water, but is soluble .n alcohol, ether, carbon disulphide, etc. On the other hand, it very readily dissolves resins, sulphur, phosphorus, fats, oils, and many alkaloids, and other organic compounds. Two features of the " Scott " system may here be set down. In the first place, the extraction is performed in the cold, thereby practically eliminating all dangei arising from the inflammable nature of the solvent used. Secondly, the extraction is effected partly by the solvent in liquid form and partly by it in the form of a vapour. In this respect, the system differs from others. In general the solvent is wholly in tiie form of a liquid, although it is evident that when hot extraction is adopted the solvent admitted as a liquid must at least in part become vaporised. The " Scott " system, therefore, may be said to combine the advantages of hot extraction with the safety of cold extraction. In Plate V. we reproduce a drawing, specially prepared for us by Messrs. Scott, showing in diagrammatic form the plant used under the " Scott '" system. Figs. 57, 58 and 59 show views of actual installations, while in Fig. 60 a small extraction plant suitable for trial and similar purposes is represented. Referring to the line engraving it will be seen that each extractor is fed with meal through a door at the top from an overhead hopper. The doors are, as indicated in Fig. 58, provided with hinged bolts, so that they may, when the extractor is charged, be readily and tightly fastened down. The hopper system of feeding the extractors economises labour, but entails the erection of a fairly lieavy superstructure. In large mills it is sometimes foui.d conven ent to dispense with hoppers and to provide instead a conveyor with a suitable oil-take to each extractor. This method has an additional advantage over the hopper system, in that by its adoption it is readily possible to feed the extractors with a mixture of seeds in any required proportion. With many materials it is desirable that the mass in the extractor should be agitated while the solvent is at work. Figs. 57 and 58 and the diagram represent plants pro\ided with agitating gear driven by means of a belt prdley, worm and worm wheel. The plant shown in Fig. 59 has no agitator. When the extraction process is completed the discharge doors near the foot of the extractors are opened so that 88 THE PRODUCTION AND TREA-mENT OF ^'EGETABLE OILS the agitator may deliver the residue of the meal on to a conveyor which runs past the doors. This residue, it is to be noted, is, under the " Scott '" method of working, quite dry and can, if required, be fed directly to cattle or horses, if the seed being treated renders this practicable. Gen'eeal ^Method or Wobkixg. During the period of extraction, the solvent, with the oil it ha* dissolved, is drained off from the foot of the extractor through the pipe A. Passing along the pipe B it reaches a stream -heated tubular vaporiser. Here a portion of the solvent is driven Fig. 57. — Benzene Solvent Extraction Plant — Scott. off as vapour, and rising up the pipe C this portion enters the extractor at the top to act, as we have explained, in conjunction with the solvent admitted as a liquid. The remaining portion of the solvent with all the dissolved oil leaves the foot of the vaporiser at D, and ilowing along the pipe E reaches a pump which lifts it up into a stUl-feed tank. Leaving this by way of the pipe F the liquid flows through a heater-condenser — or " heat exchanger "" — and so reaches the continuous still, appearing Hke a column on the right of the engraving. The construction of this still will be referred to shortly. For the time being it is sufficient to say that it completely drives off the solvent from the oil. The finished oil leaves the still at the foot as indicated. The solvent vapour finds its exit at the top through the pipe G. Flowing through the heater-condenser it is partially con- densed by the contra-flowing liquid passing to the still, and, at the same time, assists the work of the still by pre-heating the incoming supply of liquid. Leaving the EXTRACTION OF OILS BY CHEMICAL SOLVENTS 80 Fio. 58. — Scott Solvent Extractor with Asitatinj? Gear. Il^fcl^^w^ f ^ ] Lj na^s ^^^>i ■Ir^^ 1 - ^^ '^^^Wi^ V l-'io. .')9.— Siutt S,,lviiit Extnietor without A-ita Itatlli'- Ij.j.ll. 90 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS heatei' condenser the partially condensed solvent vapour is reduced completely to liquid in two condensers. On the way through the still, as we shall see presently, it has picked up some water. It is, therefore, taken by way of the pipe H to a water separator. The action of this separator depends upon the difference between the specific gravities of the solvent and water. The water flows off at the pipe J. The liquid solvent passes along the pipe K into a store tank ready for re-use. It will thus be seen that of a given amount of solvent introduced into the extractor a portion is returned directly to the extractor as vapour, and a portion is delivered into the store tank ready for re-use. The former portion emerging from the extractor as liquid containing oil in solution again reaches the vaporiser. Part of it is returned once more as vapour to the extractor, and the remainder, passing through the still, is cleaned of dissolved oil and joins the first portion of the original charge of solvent in the store tank. Obviously, as time goes on, rmless something is done, practically all the original charge of solvent will be found in the store tank ; no vapour worth speaking of \nll be foimd ascending the pipe C, and the extraction process will come automatically to a standstill. This condition may or may not correspond \^ith the complete recovery of the oil from the seed or with the degree of recoveiy desired. If it does not, a fresh quantitj' of clean solvent is passed into the extractor to complete the process or carry it a stage further. The maimer in which this fresh cpiantity is introduced is the same as that in which the original amount of solvent is admitted into the extractor at starting up. It is conducted as foUows : — The workman temporarily closes the valve L and opens the valves M and X. Clean solvent from the store tank now flows down the pipe P to the vaporiser. Partly as vapour it rises up the pipe C to the extractor, and partlj' as Uquid it flows out at D to the pump which, lifting it, sends it along the pipe Q past the valve X into the extractor at R. When sufficient fresh solvent has thus been introduced, the valves are reset and the former process is resumed. It is found that when the extraction of the oil from the meal is nearlj- completed, the solvent drawn off from the extractor contains very little oil. It is not economical to pass this poor liquid into the still. It is therefore sent into a ' half-spent ' solvent tank — not shown in the diagram — and is re-used as the first charge of solvent for a fresh charge of meal. -Small Solvent I'laut. EXTRACTION OF OILS BY CHEMICAL SOLVENTS 91 Cleaning the Meal of Solvent. When the exti'action is quite completed, and before the discharge doors are opened, the valves L and S are shut dowii and the valve T is opened. Steam, in a dry condition but not supei'heated, is then admitted to the extractor through the valve U. Blo\\'ing through the meal this steam carries off all traces of tlie solvent from the meal. The steam and solvent vapour rising up the pipe V reach the condensers, and traversing the water separator as before, are passed respectively to waste and to the store tank. Alternative Method of Working. The above description relates to the M'orking of one extractor. The other extractor is worked similarly, the vaporiser and its connecting piping being duplicated for this purpose. Our description, further, covers only one method of working the plant. Modifications are provided for. Thus the vaporiser, once the original charge of solvent has been introduced into the extractor, can be completely cut out, so that the extraction may be performed entirely by liquid solvent. To achieve this the valve L is held closed and the valves W and the valve N opened. The solvent, with its dissolved oil now reaches the pmnp without passing through the vaporiser, and is sent back to the extractor along the pipe Q. After it has been circulated through the meal a sufficient number of times it is sent into the still feed tank by closing the valve N. Safety Valves. With ))lant of tliis nature it is very important to provide safety valves at all points where pressure might coneeivably accumulate, and at the same time to provide means whereby this pressure may be relieved without allowing any of the inflammable solvent vapour to escape into the atmosphere. The points at which excess pressure might possibly accumulate are in the vaporisers, in the extractors, and in the still. Safety valves are therefore provided at X, Y, and Z, respectively. It will be noticed that the two safety valves X are connected by a horizontal pipe having union with the pipe V, up wliich the cleaning steam passes at the termination of tiie extraction process. Excess vapour passing the safety valves X does not escape into the atmosphere, but into this horizontal pipe, and so reaches the condensers. A similar arrangement is provided for the safety valves Y. The safety valve Z for a similar reason is arranged on a by-pass bridging the stop valve for the still, and delivers any excess vapour e.scaping past it through the heater-condenser into the main condensers. There is little danger of any accumulation of pressure within the conden.sers or store tank. But, in any event, the water separator acts as a seal to both, and therefore as an emergency pressure-relieving device. On the top of each vapour pipe C a deadweight safety valve is provided with outlet direct to the atmosphere. This is a purely precau- tionary measure. The valve is .set to a few pounds above the release pressure of the valve X, and is intended to come into use should the latter valve, for any rea.son, fail to act. So far as Messrs. Scott know these deadweight safety valves have never yet on any of their plants been called upon to fulfil their function. The Continuous Still. The separation of the solvent from the oil is begun in the vapox'iser. This is done simply to take incidental advantage of the steam required to generate the vapour for the extractor. The main and final separation takes place in the still. This separation is a most important feature of the process, for on its completeness must largely depend the commercial value of the oil recovered. Very frequently this separation has been 92 THE PRODUCTION AND TREATMENT OE \rEGETABLE OILS attempted in stills of the pot or bulk charge type. This naethod of working occupies a considerable amount of time, a fact imjiortant in itself and also in its bearing upon the effect which contact with heat for a prolonged period has upon most oils. Li addition, towards the end of the operation there is little solvent to remove, so that during this time the steam used to drive off the solvent cannot be used with full efficiency. Thus fuel is wasted, and an unnecessary tax is placed on the condensers which collect the steam and solvent vapour. The still shown in the diagram is of a form recently patented by Mcs.srs. Scott. A number of these stills are already satisfactorily at work on the production both of edible and of trade oils. The " Scott " still is divided into a number of sections, each of which is a still by itself. The oil and solvent mixture heated in the " heater- condenser " to approximateh' the distilling temperature enters the top section of the still and passes downwaids in turn through each of tlie others. In so doing it comes into direct contact with an ascending current of steam admitted below the bottom section of the still and bafHed in such a way as to cause it to take a tortuous course through the descending Uquid. As the steam ri.ses it liberates the solvent as vapour, which vapour assists the steam in distilling the solvent from the liquid passing through the next liighest section of the still. It will be seen that under thi.s method of worldng the freshest steam is caused to act upon the liquid with the least amount of solvent in it, that is to say, in the liquid at the time when it contains those last traces of solvent which are the most difficult to remove. At the top of the still the steam, partially used up, is given the easiest work to do, namely to attack the liquid when it is richest in solvent, and therefore has tlie lowest boiling-point. Working Charges. It is claimed for the " Scott " system that verj' little labour is required to run the plant. The pioportion which the labour charges will bear to the other working costs depends, however, on the size of the plant, for while the size varies, the number of men required to operate it remains constant. It is stated that the vexy largest plants consisting of many extractors can be operated by two men. Economy of steam consumption is another pouit connected with the plant to which the makers call attention. The coal required per ton of raw material, we are hiformed, may be set down as from 2 to 3 cwt. The only other item of working costs to be considered relates to the solvent. It is fomid that in operation a certain amount of solvent disappears ; where it goes to is by no means clear. This loss may be returned at Ig gallons per ton of material treated. It is, perhaps, worth adding that the solvent extraction process has to-day a very wide field of application outside of the vegetable oil industr}\ It is bemg employed for the extraction or recovery of grease, oil, or fat, from many miscellaneous substances, such as wool waste, bones, leather scrap, rags, factory sweepings, and refuse of all sorts. CHAPTEK XII THE REFINING OF OILS We now come to a section of our subject concerning wliich a great deal of secrecy is commonly exercised. Oil refining is usually carried out in works quite separate from the mills producing the oil. It may, in fact, be properly regarded as constituting an industry by itself. It requires the possession of a considerable laiowledge of chemistry, for each oil in general has to be treated in a special manner. The refining may be carried out to vaiying degi'ees of completeness. According to its degree, so does the enhanced price obtained for the oil vary. As a rough guide, however, it may be said tliat refining increases the value from, say, £5 per ton, as in the case of rape oil, to anything up to £10, as in the case of cotton seed oil. A perfectly pure oil is a definite chemical body. It may be regarded as being formed by the union of a molecule of glycerine with a molecule of fatty acid accom- panied by the withdrawal of a molecule of water. I The glycerine is definitely constant from oil to oil. The fatty acid varies from oil to oil, and by its variation gives the oil its characteristics. All piu'e oils, such as we are for the moment considering, are probably identical, in so far as they are colourless^ odourless-a^wLiaeteless. Crude oils differ from pure oils in three principal respects. In the first place, they may be coloured. The colouring matter is derived either from the fleshy portion of the seed from which the oil is recovered or from the husk of the seed, if this is crushed along with the fleshy portion. Secondly, crude oils contam vegetable fibrous matter or mucilage or other foreign bodies crushed out of the seeds along with the oil. Such mucilage is simply suspended mechanically in the oil. Thirdly, they may contain free fatty acid and free glycerine, caused by some portion of the oil absorbing water and spHtting up. This splitting-up process, or hydrolysis, as it is called, is fi:equently caused by careless or crude methods of manufacture, as in the case of palm oil. Even, however, with the most careful manufacture, some fatty acid is nearly certain to be present in the crude oil, the rea.son being, apparently, the hydrolysis of the oil by natural proces.ses in the seed itself before it is crushed — possibly even before it is gathered. The presence of free glycerine in an oil is rarely objectionable, for it is colourless, tasteless and odourless and stable. The presence of free fatty acid is nearly always objectionable, for to such may usually be attributed the characteri.stic taste and smell of an oil, while in addition, its decomposition turns the fat or oil rancid. The possibility of such acid being present is the prime reason why vegetable oils are not in favour as lubricants. In addition to the removal of mucilage, of colouring matter, and of free fatty acid, oil refining frequently includes a fourth class of operation. On a cold day certain qualities of olive oil will be noticed to throw do^^^l a flocculent whitish deposit. Cotton seed and other oils likewise become cloudy when the temperature falls. This deposit is " stearine " — or " margarine," as it is frequently and somewhat imfortmiately called — and its removal is desirable in certain circumstances, notably so if the oil is to be used for bummg, lubricating or edible purposes. The " stearine " itself is a valuable substance when isolated, and is made use of in the manufacture of candles, margarine, margarine cheese and lard substitute. The processes employed in oil refining are either mechanical or chemical, or a 94 THE PRODUCTION AND TREATMENT OF \^GETABLE OILS combination of both. Thus mucilage is removed mechanically. Bleaching is in general efifected chemically, but is frequently accomplished by what is really a mechanical process. Free fatty acids are removed by a chemical reaction. " De- margarination "' is most frequently effected by physical processes. Preliminary Refining in the Oil Mill. , A certain amount of preliminary refining is commonly conducted on the oil before it leaves the oil mill. This refinmg aims only at the removal of the mucilage, etc., in the oil. Formerly, it was conducted simply by storing the oil for prolonged periods, sometimes extending to years, in storage tanks, wherein the foreign matter gradually Fig. G1.— Filter Press for (_)il — Mauluve, AUiott. fell to the bottom, leaving the clear oil on top. Modern practice now makes use of filter presses, and so very greatly economises both time and space. Filter Presses. The filter press is used in many industries for effecting the separation of solids from liquids. It is made in several modifications, but always follo^^•s the principle of forcing the liquid to be filtered through a layer of cloth, swansdown or twill. A filter press suitable for use in an oil mill is illustrated in Fig. 61. It consists primarily of a series of cast-iron plates formed with a lug at each side, which lugs support the plates on a pair of steel rods— usually circular — extending between the two fixed ends or pedestals of the press. In the form of plate shown in Fig. 62 an edge is raised up all round the peripheiy of each face, so that when two plates are brought together the dished centres form a chamber between them. Through the centre of each plate a circular feed hole is formed for the oil. The method of working will, perhaps, be understood with the help of the sketch. Fig. 63. For each plate there are provided two filtering cloths A, B, formed with central holes and united round the edges of these holes by means of a short cylinder or ring of cloth C. The cloth A can readily be rolled THE REFINING OF OILS Fig. 62.— Filter Press Plate. up, slipped through the central hole in the plate and spread out flat on the other side. The two cloths to facilitate the assembly of the plates are then held at their upper edges by means of clips D passing on to a rib formed across the top of th^ plate. When all the plates, thus clothed, have been assembled in the framework of the press, the pinion E (Fig. 61) is rotated by means of a tommy - bar inserted in holes round its flange. This pinion engages with a rack extending from the sliding head F, go that the action results in the press plates and their cloths being closed up together. The final closure of the plates is effected by turning down the half -covers G and screwing up the hand -wheels H by means of levers. The chambers between the plates are thus sealed by nipping the cloths between the raised edges. Crude oil is now pumped into the press at the right-hand end, and flow- ing throxigh the central feed holes fills all the chambers between the plates. Under the pressure of the oil the filter cloths are pressed backwards until they meet the support of the plates. The faces of the plates, as shown in Fig. 62, are formed with vertical grooves connected by shoi-t horizontal grooves, so that the oil filtering through the cloths may trickle downwards into a gutter formed along the jilate just above the lower raised edge. From this it is conducted through three holes into a central passage waj'^ J (Fig. 63), and so through cocks K (Fig. 61) into a collecting trough. Instead of joining the two cloths for each plate by a ring of cloth, the cloths may be entirely separate. The central holes in the cloths are then nipped to the edge of the feed hole in the plate by mean.s of a clip either of a screw or bayonet- fastening type. Yet another alternative metliod, one finding considerable favour, is to form the raised edges of the plates as a separate frame having lugs, like tlio.se on the plates, for their independent support on the two horizontal bars. This design of press is that actually shown in Fig. 61. The feed holes, as indicated at L in Fig. 63, are in this form placed near the upper edge of the plates, and are continued through the loo.se frames M. A hole N conducts the oil from the feed passage into the chambers between the plates. Tlie cloths are placed between the loose frames and the plates. In this way the edges of the plates and the feed holes are sealed simultaneously when the press is sci-ewed up. This design has certain advan- tages when it is desired to remove in one piece the cake left on the cloths. The Ehgineeb" Fig. 63.— Two Form Swain Sc. of Filter Plates. 96 THE PRODUCTION AND TREATMENT OF ^"EGETABLE OILS The method adopted for closing the press plates also raries a good deal. Thus it may, in addition to the manner shown in Fig. 61, be effected by means of a central screw or a compressed air cylinder, or a hydraulic ram. In connection with the feeding of the press with crude oil considerable attention has to be devoted to the fact that the filtration towards the end of the operation becomes slower, so that a lessened feed is required. If the press is fed by means of a belt -driven pump, a rehef valve should be provided on the feed i^ipe. so that with the pump rmining uniformly some of the feed may be by -passed when the speed of filtration falls off. A steam-driven pump ^_^ ^^^^^^^^^fiVVS. r« ^^1 ^^. y ! ' ^•i.,..^ W y Fig. ii4.— ■\Vashin; Machine for Filter Cloth> — Miiilove, Alliutt. can be itself regulated to suit this requirement, and therefore does not require the provision of a relief valve. A better method than either seems to be the adoption of a forcing ram worked by compressed air. The flow in this case is stated to be entirely seK-ad justing. As usually sujiplied these filter presses may have am-thing from six to forty-five chambers, the dimensions of the plates varying from 13 in. to 40 in. square. The thick- ness of the cakes left in them is from 1 in. to If in. Little can be said as to the out;>r.t, for this varies from oil to oil, and with the one oil, according to its previous treatment, and whether or not it is filtered hot or cold. As a guide, however, it may be said that a press with twenty -four chambers and plates 25 in. square — giving a total filtering area of 208 sq. ft. — may be expected to filt-er in twenty-four hours 140 cwt. of Unseed cocoa-nut, or fresh olive oil : 100 cwt. of crude cotton oil : 7'^ cwt.. of crude i-ape or THE REFINING OF OILS 97 stale olive oil ; or 16 cwt. of castor oil. The length of time for which the press \vill work without being opened for cleaning also depends upon the nature^of the oil being filtered. In the case of linseed oil the' press may be run continuously for about a week. It would then be allowed to stand with the pressure removed from it for, say, three hoars, at the end of which time it would be opened up and the cakes formed in the chambers removed. Thereafter it is ready for a further run. Filter Cloth Washing Machine. Occasionally, and particularly when the production of edible oils is in ques- tion, it is desirable to remove the press filter cloths and wash them. A washing- machine for tills purpose is illustrated in Fig. 64. In this machine the cloths are treated in a hot dilute solution of caustic soda which, combining with the oil, produces soap and so cleanses the cloths from mucilage and dirt. The machine has an outer casing of galva- nised steel fixed to two cast-iron ends. The internal rotary washing compart- ment is constructed of hard-rolled brass plates perforated from the inside in a special manner so as to avoid the creation of burrs. Five lifters or rubbers are provided inside the drum, while in the larger sized machines, such as that illustrated, there is a central partition. The outer casing and the inner drum are both provided with segmental doors sliding in brass guides. Hand-turning gear is provided for bringing the two sets of doors into alignment for loading and unloading purposes. The shaft carrying the washing compartment jjasses through glands in the cast iron ends of the casing, and is supjjorted externally in adjustable roller bearings. It is driven through a silent rocker chain from a shaft at the back of the machine. This shaft carries two loose pulleys for crossed and oi)en belts, and a fixed belt pulley. A wonn and a worm wheel gear is pro\idcd automatically to move each belt alternately on to the fixed central pulley, so that tiie direction of rotation of the washing compartment may, during a run, be reversed at regular intervals. Steam and hot and cold-water valves arc arranged on the casing in order that the cloths may be washed, boiled, and rinsed. A full bore waste outlet is also provided. After the clotlis have been removed from the machine they are placed in a centrifugal hydro-e.xtractor, which removes the bulk of the water. Thereafter they may be thoroughly dried, if thought necessary^ in a steam-heated hot-air rotary drying machine. Fk;. ()0. — t't'iitrihisral lOxtractoi- lor " Foat^ ••8 THE PRODITTION AND TREATMENT OF ^^:GETABLE OILS The Tkeatmext of " Foots." In all oil mills, ■whether the presses in use are of the Anglo-American or the cage type, a considerable amotmt of meal saturated with oil escapes from the press and DriV.ne Sfi3. t Fig. 66.— Cotton-Seed Oil Eetinerv — Maiilove. Alliott. accumulates in the tanks in which the presses stand, the oil dishes, and so on. This material is knowii as " foots," and to avoid waste is treated so as to separate the bulk of the oil from it. A common method of effecting this separation is by the employment of a ■■ centrifugal "" such as is shown in Fig. 65. This machine consists of a cast-iron THE REFINING OF OILS 99 casing enclosing a basket of tinned-steel wire with a pressed-steel top ring and bottom. The basket is mounted on a vertical shaft supported at the top and bottom on ball bearings, and driven through friction cones from a horizontal cross-shaft. The peri- pheral speed of the basket is usually about 9,000 ft. per minute. The slurry is placed within it, and in a very short time the bulk of the oil is driven off out of the basket. This oil may be mixed with that extracted in the usual way or sold separately. The residue in the basket is foimd to contain about the same percentage of oil as does the original seed before pressing. It is therefore returned to the meal kettle and worked back to the press with the fresh meal. Li addition to the preliminary separation of mechanical impurities carried out as mentioned above, the refining of oil comprises the removal of free fatty acids and of bleaching to get rid of the colouring matter. Broadly, it may be said that the removal of the free fatty acids is necessary if the oil is to be used for edible purposes, and that bleaching is- desirable if it is to be used for the manufacture of paints or varnishes. Removal of Free Fatty Acids. By removing the free fatty acids from the crude oil, the oil is deprived of the elements which give it its characteristic odour and taste, and which render it liable to decomposition. At the same time its colour will probably be improved, for the free fatty acids are a cause of discoloration in addition to the colouring matter absorbed by the oil, during its extraction, from the husks of the seeds. The standard method of removing the fatty acids is to treat the crude oil with caustic soda solution, carefully regulated in strength and amount, and at a carefully regulated temperature. The soda solution combines with the free acids to form a soap, but is not sufficient in amount to go farther and saponify any material amoimt of the neutral oil. It is here to be noted that more caustic soda has to be added to the oil than is theoretically necessary to neutralise the percentage of free acid revealed by analysis in the crude oil. The surplus soda does not, however, attack the neutral oil miless of course it is permitted to be present in an altogether excessive amount. The reason, both for the procedure and of the result, lies in the fact that the action between a given amoujit of caustic soda and a given amomit of oil will cease at a point, short of completion, at which a state of equilibrium is established between the amount of soap formed and the amount of oil and of caustic soda .still left micombined. The point in question is influenced by the temperature at which the reaction is conducted. On the neutraUsation of the free fatty acids being completed, there is thus left in the refining kettle a mixture consisting of soap, acid-free oil and caustic soda in solution. This mixture is allowed to stand for some hours to permit the soap, soda solution and any mucilage or albuminous matter to sink to the bottom, while the oil rises as a clear hquid to the top. The clear oil is then dra\^Ti off for further treatment. The residue at the foot of the kettle, containing as it does, a certain amount of neutral oil besides the soda, etc., is removed separately, and is sold to the soap maker as "soap stock." To facihtate the settling out of the soap, etc., from the oil, salt is sometimes thrown into the kettle, for soda soap is insoluble in salt water. The clear oil has next to be wa.shed with water to remove all traces from it of the soda. Thereafter it is treated in a vacuum still to drive off any volatile fatty acids which may linger in it, as well as the last traces of moisture left in it by the washing process. A vacuum still is used in order that the volatile acids and the moisture may be driven off at a temperature below that which will deleterioiisly affect the oil. 100 THE PRODUCTION AXD TREATMENT C»F ^^:GETABLE OILS Bleachtsg. The oil may or may not now have to be bleached. If it is to be used for edible purposes, it is desirable that it should be bleachetl by means of fullers earth or such- hke absorbent material. For other purposes chemicals liberating chlorine or oxygen may be used. Treatment with fuller's earth, animal charcoal, etc.. not only helps to bleach the oil : it also assists in deodorising it. The process consists in thoroughly stirring ( 1 r^ ( 3kj j r I i ir J np Fig. 67. — Vacmim Pan and Condenser. the dry absorbent powder into the oil when gently heated and, after agitation for a short time, in passing the liquid through a filter press such as we have described above. The earth or charcoal with the absorbed colouring matter is retained on the filter cloths, while the clear oil is drawn off. The filter press is usually arranged to permit steam to be blown through it after filtering is completed. In this way the cakes are washed free from oil, so that on the press being openetl the earth falls out as a powder. It may surprise some to learn that oils can be bleached in the above purely mechanical manner. The explanation of the matter lies in the fact that the colouring THE REFINING OF OILS lOl substance in the original seed is. in general, in the form of a powder, and passes as such into the oil. . It can therefore be absorbed and held back by the earth or charcoal. It will be noticed that this method bleaches the oil by the direct removal of the colouring matter. A similar end is achieved by the sulphuric acid method, which is applied occasionally for bleaching certain oils. The acid dehydrates or chars the colouring matter and other impurities, and causes them to coagulate, so that they may readily be removed by tilti'ation or sedimentation. This treatment incidentally secures the removal of any moistui-e in the oil, by reason of the strong attraction for water posse.ssed by sulphuric acid. As, however, some acid may remain behind in the bleached oil, the method is not usually adopted if the oil is to be used for edible or lubricating purposes. Bleaching by means of chlorine or oxygen does not secure the removal of the colouring matter. The colouration is destroyed by the oxidation of the colouring matter, but this, when oxidised, is allowed to remain behind in the oil. The chemicals used are, in general, such as to render the process unsuitable for application to the treatment of an edible oil. In most cases the oxygen or chlorine is generated by chemical reaction within the oil itself. Thus bleaching by means of oxygen may be effected by adding to the oil manganese dioxide and sulphuric acid. Similarly, chlorine may be generated by adding bleaching powder and hydrochloric acid. In one case manganese sulphate, and in the other calcium chloi-ide, is left behind in the oil, and has subsequently to be removed by washing. Further, in both cases the reaction of the chemicals results in the formation of water. Many other methods of bleaching oils by means of chemicals, or otherwise, are practised or have been proposed. It is not necessary for us here to discuss these, for they belong more to the chemical than to the engineering side of our subject. We need only remark that one of the oldest and one of the very best methods is by exposing the oil to the action of sunlight and air. This process results in the natural oxidation of the colouring matter, and is extensively adopted in the case of linseed, poppy and walnut oils, as used by artists. It is. of course, a very slow method. Recently, the bleaching of oils by means of ultra-violet rays has attracted some attention. Cotton Oil Refining. The arrangement of a typical refinery for treating cotton-seed oil is reproduced in Fig. 66. The oil in this case is fir.st heated by steam in a mixing tank A until it reaches a temperature of about 140° F. Thereafter the oil is violently agitated by means of compressed air. the temperature, meanwhile, being kept as near 140° F. as possible. During the agitation cau.stic soda solution from the tanks C, D, is run into the mixing tank. As this solution, being heavier than the oil. tends to sink to the foot, care is necessary if it is to be brought properly into intimate contact with the oil. This is secured by distributing the solution evenly over the surface of the oil, and by the vigorous agitation to which the contents of the mixing tank are subjected. When it has been ascertained by testing samples that sufficient caustic soda has beg added to neutralise the acid reaction of the oil, the charge is allowed to standand settle in the mixing tanly The settling is usually sufficiently complete at the end of about twelve hours to permit the clear supernatant oil to be drawn off and passed into the washing tank H. Li so doing, gi'eat care ha.s to be exercised that none of the residue is passed off with the clear oil. This residue is ultimately drained into the mucilage tank G. In the washing tank the oil is gently heated and washed with water to remove the caustic soda solution remaining in it. The water is distributed uniformly over 10-2 THE PRODUCTION AND TREATMENT OF \EGETABLE OILS the surface of the oil, which, as before, is violently agitated by means of compressed air jets. On aUowing the charge to settle, the oil rises to the top. The water con- taining the soda in solution sinks to the foot of the tank and is drawn off. For the production of the best edible oils two or three washings may be required. Fig. 6S. — Cocoa-nut C»il Ketinerr— Man.ove. .Uliott. If the oU is for edible purposes — say for the manufacture of margarine or lard substitute — it will either be passed without being bleached into the vacuum pan N or will be bleached by the fuller s-earth method already referred to. Oils for other than edible purposes are passed from the washing tank into the bleaching tank B. THE REFINING OF OILS 103 Here they are agitated in the usual way and are subjected to the joint action of hydro- chloric acid delivered from the cast-iron tank E, and of bleaching powder solution dra\m from the slate tank F. When bleaching is completed, the charge is returned to the tank H, wherein the bleaching chemicals and the salts formed by them are washed out of it. The procedure may be slightly varied by passing the oil direct from the tank A into the bleaching tank. This avoids the fii'st washing, but I'esults in a certain amount of acid being wasted in the neutralisation of the caustic soda solution remaining in the oil. This neutralisation, it may, however, be noted, results in the production of sodium chloride, the presence of which in the oil is by no means harmful, but frequently of assistance. The oil dra\ra from tlie washing-tank is now passed into the vacuum pan N — showTi separately in Fig. 67. Here it is mechanically agitated and heated under a vacuum, so as to drive off the moisture and any free volatile fatty acids which maj' yet remain in it. The expelled products are caught in the condenser P. If the oil is an edible oil, and if it is required in a bleached condition, some refiners combine the fuller's-earth treatment with the treatment of the oil in the vacuum pan. On leaving the pan, the oil is. in such a case, passed through a filter press, whereafter it is ready for the market. Cocoa-nut Oil Refinixg. A small refinery for cocoa-nut, palm kernel and similar oils is illustrated in Fig. 68. The procedure in this case is, in prmciple, similar to that followed in the cotton oil refinery described above, except that no provision is made for bleaching the oils chemically since they are here intended solely for edible purposes. The oil. as received, is first passed through a filter press A to remove mucilage, etc., and is thence run into a storage tank B. From this it is passed by gravity into the refining tank C situated on the floor below, where it is heated, agitated, and treated with caustic soda solution from the tanks D in the usual way. After settling, the mucilage and other residue is dra^vii off into the pitch-pine tank E situated on the ground floor, while the clear oil is passed into the tank F on the first floor. In this tank F the oil is washed with hot water from the tank G and is, in addition, heated by means of a steam coil. The tank F. in fact, not only .serves for washing the oil, but also acts as a prehminary still for driving off a certain amount of the volatile free fatty acids which yet may linger in the charge. By means of a rotary pump H. the charge in the tank F can be sent back to the refining tank C. so as to be returned to the tank F for further washing and heating. The oil, previously treated with fuller's earth or not. as is thought desirable, is passed from the preliminary still F through a second filter press J. and thence into a finishing still or vacuum pan K of the design illustrated already in Fig. 67. Demargarin.vtiox. Certain oils, notably cotton-seed and olive, as we have already remarked, throw downi a deposit of " stearine " when the temperature falls below a certain point. Chemically, an oil is formed bj- the union of a fatty acid with glycerine accompanied by the withdrawal of a certain number of atoms which, taken together, constitute water. The body fonned by such a union is knovm as a glyceride. A glyccride is thus an oil, but no actual oil, so far as we know, is formed of one and only one glyceride. Stearine is a glyceride, being formed by the union of stearic acid with glycerine. Palmitine — palmitic acid and glycerine — is another. And there are many more, 104 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS such as oleine. linoline, linolenme, and so on. These gh'cerides solidify at different temperatures. Thus, of those mentioned, stearine and pahuitine may be said to have relatively high solidifying points, and oleine, linoline and linolenine, relatively low soUdifying points. Taking the particular case of cotton-seed oil, we find that this oil consists prmcipally of a mixture of palmitine, oleine and linoline. When the tempera- ture falls, the palmitine solidifies out, while the oleine and linohne are still liquid. The " stearine " deposited by cotton-seed oil is, therefore, not stearine, but palmitine. From other oils — for example, from olive oil — it may consist of a mixture of true stearine. palmitine. and other glycerides solidifying at a relatively high temperature. ^^ T^^^^H^^^^^^^^^^^^fek^^^^^^^^SiK^^^^HRSB F:o. 69. — Stearine Presses for Demargarinating Oil. The extraction of the " stearine " is an important operation. ])articularly in the case of cotton-seed oil. This oil. after being " demargarinated," is knowii as " winter cijl," because it •will not throw dov,ii a deposit or become cloudy at temperatures / normally occurring in winter. A usual method of carrjing out the demargarmation is to cool the oil artificially luitil the " stearine " portions solidify, and then to pass I the whole through a filter press. A slightly different method consists of completing the freezing of the whole oil in flat pans, wrapping the frozen cakes in bagging and pressing them in a hydraulic press. Under the pressure, the portions of the oil havuig the lowest freezing-point. Hquefy. and are forced out and drain away. A set of stearine presses, suitable for this method of working, is illustrated in Fig. 69. The presses differ considerably from those of the Anglo-Amei-ican type used for crushing seeds. Each is provided with a ram 12 in. in diameter and suitable for a working pressure of 2 tons. Asquare table is formed at the head of the ram. and on to this a four-wheeled carriage can be rim on I'ails from either side of the press. The carriage is provided with catches, THE REFINING OF OILS 105 which can be hhiged down to engage the columns of the press, and with two vertical guide bars, which, when the ram rises, enter holes in the press head and so hold the carriage steady. The expressed oil is caught in the box-like carriages. There are two carriages for each press, so that one may be filled while the other is under pressure. Special provision is made to ensure that the pressure shall be applied very slowly. The refrigerated cakes are pressed between steel plates. These are sufficiently large to accommodate four cakes each. All the machines and plant illustrated in this chapter represent the practice of Manlove, Alliott & Co., Ltd., Nottingham. CHAPTER XIII THE HYDROGENATION OR HARDENING OF OILS Fatty vegetable and animals oils may be described as consisting of a glycerine part and an acid part. The composition of the glycerine part is constant. The compo.sition of the acid part vaj-ies from oil to oil. and is characteristic of any one oil, or of any one gi-oup of oils. Several important vegetable and animal oils contain an acid part having the general chemical formula C„H2„0.^. Among these we have butter fat, cocoa-nut oil, palm oil, palm kernel oil, lard, tallow, and various '" butters." such as cocoa, mace and nutmeg butters. It will be noticed that the oils and fats mentioned are in general characterised by the possession of a thick consistency ; that is to say, they have high melting-points, or, in other words, they are naturally "' hard."' Man}' other important vegetable and animal oils differ from those just mentioned in that their acid parts fail to tit the general formula quoted to the extent of two, four, six or eight atoms of hydrogen. Thus, in rape oil. and certain fish oils, two hydrogen atoms are missing. Four are absent in the acid parts of soj-a bean oil and cotton-seed oil. In linseed oil six atoms are missing, and in certain liver and blubber oils eight atoms are awanting. All these oils, it will be noticed, are in general characterised by the possession of a liquid consistency. In passing it should be observed that castor oil does not appear under either division. This oil is exceptional, for its acid part contains not two, but three atoms of oxygen. The oils of the second division — or. to be quite exact, the acid parts of these oils — are termed unsaturated, for they are, theoretically at least, capable of taking up and combining with additional atoms of hydrogen. In practice, however, under normal conditions, hydrogen, even in the na.scent state, is quite without action on fatty oils. The great commercial value attaching to the power of being able to convert an unsaturated into a .saturated oil has led to mucii investigation of the matter. It has been discovered that the addition of the hydrogen atoms can be effected if the oil is suitably treated with hydrogen gas in the presence of finely divided nickel or palladium. Each of these metals acts as a catalyst, and is left unaltered after the hydrogen has been taken up by the oil. These are the broad chemical aspects of the process we are now discussing. Wherein lies its commercial applicability I The answer to this question can be given in a general statement. For the purposes of modern industry the world's sujjply of natural fats is deficient, while the supply of liquid oils is superabundant. The liydro- genation process permits us to make good the deficiency by converting some of the superabundant liquid oils into hard fats. As an instance of the commercial applicability of the hydrogenation jn'ocess. we may look for a moment at the .soap-making industry. The ideal substance for tlie .soap maker to work with may be said to be tallow. It is a firm substance, and yields a firm soap such as we are accustomed to. Tallow, howevei-. is expensive, and is obtainable only in strictly limited amounts. The soap maker accordingly falls back upon some of the harder oils, .such as cocoa-nut oil, palm oil. and palm-kernel oil. THE HYDRC^GEXATION OR HARDENING OF OILS 107 These oils are also expensive and are in increasing demand in other industries. If, however, the soap maker tries to replace them with one or other of the abundant naturally liquid oils, such as whale oil, soya-bean oil, and so on. his product loses greatly in quality, and is apt to be a soft, sticky mass, mausable or unsaleable as soap for many purposes. By hardening these oils before using them in the soap kettle, he obtains a substance practically identical with tallow without affecting the yield from them of that valuable by-product of the soap-making industry, glycerine. The hydrogenation process thus throws open to the soap maker a wide range of oils which otherwise would be next to useless for his purpose. Similar remarks applj' to the candle-makmg industry, which, again, calls largely for fats rather than oils. For certain edible productions, notably margarine and chocolate, fats are now in demand to a greater extent than can be conveniently met from natural sources of supply. Whatever may be the case to-day — it is verj' difficult to find out exactly how matters do stand at present — it is certain that artificialh- hardened oils ^\'ill soon be in extensive and acknowledged use for edible purposes. Just for the moment there is a feeling of vmcertainty as to this employment of them, for it is not yet settled how far the possible presence in the hardened oils of a small amomat of the nickel or other catalyst is harmful to the human constitution. The chief oil hardened at present is whale oil. Increasing quantities of cotton- seed, linseed, soya-bean, cocoa-nut, and other oils are. however, also being subjected to the process, so that the subject is one quite properly faUing within the scope of this volume. With regard to the hardening of cocoa-nut oil, a word of explanation is no doubt desirable. This oil is just on the border line between the true oils and the true fats. It is one thing in one part of the world, and the other in another part. All its acid part is not saturated, but contains portions of imsaturated acids. It is therefore capable of absorbing a certain amomit of hydrogen, and so becoming harder than it is normally in this climate. Hardened oils are white, tasteless, odourless, substances of tallow-like consistency. Theoretically at least, they should all be identical, whatever may be the particular oil started %rith, and in practice such identity seems to be attained, at least in the oil as freshly hardened, but there is some uncertainty whether a hardened oil if kept long enough will or will not develop some characteristics of its parent. Thus, hardened whale oil may, sooner or later, develop a fi.shy smell, and hardened cocoa-nut oil the characteristic smell of cocoa-nuts. In practice, however, the oils are usually hardened at the soap works, or wherever else they are to be used, or are otherwise employed with but little interval between being hardened and being treated in industrial processes. The point is of importance, for there are distinct signs that, in the near future, certain oils will be hardened before shipment to this countr}^ Thus the process is attracting considerable attention from the soya-bean oil producers in Japan and Jlanchuria, the idea being that hardened oils may be sliipped and carried without ri.sking that loss through leakage, etc., which is a serious item in the shipment of liquid oils. The Technology of Oil Hardening. Coming to the technology of the process we find that success is dependent primarily upon two circumstances, first, the careful preparation of the catalyst, and, secondly, the use of very pure hydrogen. Veiy little variation of procedure may quite readily result in an entire failure to harden the oil. The catah'st commonly used on a commercial scale is metallic nickel prepared in a finely-divided .state by chemical precipitation. Once made it must be kept Ids THE PRODUCTIOX AND TREATNEEXT OF ^'EGETABLE OILS rigorously apart from certain other substance*, notably air. moisture, sulphur, arsenic, carbon monoxide, methane, etc. These substances oxidise or otherwise react on the metaUic nickel, and quite destroy its catalytic action. Thus it is stated that a tenth of 1 per cent, of sulphuretted hydrogen, if present in the hydrogen used in the process, will prevent the hydrogenation of the oil. The effect of these substances on the catalyst is felt in three directions. First, a* \re have said, it means that the hydrogen used must be very pure, and free especially from moistm^ and sulphur compounds. Secondly, the oil to be hardened must be thoroughly freed as a prelimLuary from the moisture which, when received, it is certain always to contain. Thirdly, in preparing the catalyst a stage is reached when it must be treated and handled out of contact with the atmosphere. Given the satisfactory- attainment of these conditions the process is simple. The oil with the catalyst added is heated m an atmosphere of hydrogen inside a closed vessel — an autoclave — fitted with a mechanical agitator. The oil and hydrogen are brought into ultimate contact and at the end of three to four hours the absorption is found to be complete. The temperature at which the work is carried on is of great importance. It appears that for any given pressure of hydrogen inside the autoclave there is a definite temperatiu^e which must be reached before the absorption begins. At atmospheric pressure this temperature appears to be about 250' C. In practice such a temperature would almost certainly result in the hardened oil being discoloured. To avoid this some temperature approximating 2(H.>' must be used. The pressure of the hydrogen has to be increased above atmospheric as the temperature is decreased. A normal working condition is a temperature of ITO"^ to ISO' C. in conjunction with a pressure of 70 to 80 lb. per square inch. When the absorption is complete the oil is nm out -pical hydrogenation plant on the Lane system, and in the next^ — because of the vital importance to the success of the hardening process of an inexp«isive supply of pure hydrogen — to describe the Lane system of generating hydrogen. In the engraving (Fig. 70), we reproduce the general plan of an oil-hydrogenising plant erected to Mr. Lanes designs. As set to work in the first instance, this factory- has a capacity for treating I ton of oil per hour, but throughout provision is made for trebling the plant and the output. In this chapter we are concerned solely with the lower portion of the plan — ^the oil treatment department. The upper portion — more than 50 per cent, of the whole — represents the lay-out of the hydrogen-producing plant which we will deal with in otir next chapter. INSERT FOLDOUT HERE THE HYDROGENATION OR HARDENING OF OILS 109 Reception and Desiccation of the Raw Oil. The raw oil is received at the works at the point marked A in the lower left-hand corner of the plan. The fir.st operation is to remove the oil in a cleanly and thorough manner from the barrels and to pass it into steel storage tanks. As the raw oil may be naturally thick — such as is the case if cocoa-nut, palm, or palm-kernel oil is being treated — means have to be provided for heating it so that the barrels may be properly emptied and the oil in the storage tanks may be kept sufficiently liquid to be pumped on to the next stage. The means provided consist of steam jets for heating the barrels and flat steam coils at the bottom of the tanks to preserve the contents in a liquid state. The tops of the .storage tanks are open, and across their mouths is erected a wooden .stage in which grilles are formed on to which the barrels are emptied. The next stage consists in thoroughly drying the oil so as to meet the requirement for success, mentioned above, that the cataly.st should not be brought at any time into contact with raoi.sture. The desiccation is performed in two stages. The oil is first pumped from the storage tanks into open preliminary heating ves.sels circular in section and having conical bases. These vessels are fitted with mechanical agitating gear and with steam heating coils. In them the oil is freed of the gi-eater part of its moisture. To secure the final and complete desiccation the oil is pumped into vacuum pans consisting of circular sectioned vessels with domed tops and conical bases, and containing a heating coil and mechanical stirring gear. The domed top of each pan is provided with an inlet connection for the oil and a connection to a vacuum pump. The outlet for the oil is through a cock at the foot of the conical base. Steam, fluid level and vacuum gauges, and thermometers are fitted in comiection with the pans. The oil leaving the pans is now ready to be brought into contact with the hydrogen in the autoclaves, but before proceeding to de.seribe these we will deal with the preparation of the catalyst. Preparation of the Catalyst. The catalyst employed in the Lane process is finely divided metallic nickel. It is received at the works in the form of nickel sulphate in crystals. The first step in its preparation consists of making a solution of the nickel sulphate and another solution of sodium carbonate. This is done in the two tanks B, C, respectively. Each of these tanks is fitted with an open steam jet to facilitate the preparation of the solution. They are erected over a third tank provided with means for mixing the two solutions when they are turned into it. The result of this mixture is the precipitation of insoluble nickel carbonate and the passage of sodium sulphate into solution. Previous to the admission of the two solutions a quantity of finely divided refractory neutral material is placed in the mixing tank. In practice this material is usually kieselguhr — that is to say, infusorial earth consisting of siliceous diatom fossils. Its fimction is to act as a carrier for the nickel. The mother liquor, the precipitate of nickel carbonate, and the kieselguhr are drawn off from the mixing tank and piun])ed througli a filter press of the type described in our preceding chapter. AVhen the filtering is completed the nickel carbonate and the kieselguhr are found consolidated on the filter cloths as cakes. These cakes are thoroughly dried in hot-air stoves, and thereafter are reduced to powder by means of an edge runner. The carbonate has now to be roasted or calcined so as to reduce it to the form of oxide. Thereafter comes the very delicate operation of reducing the oxide to the metallic form. This is effected bj- heating the oxide in contact with hj'diogen — which must be quite free from air — at a certain temperature. One form 110 THE PRODUC'TIOX AX'D TREATMENT OF VEGETABLE OILS of the apparatus employed is contained ^^•ithin a heat -insulated vertical ease to which the pulverised material is fed automaticallj- at the top. while the hj-drogen is admitted at the foot. Inside the case there is provided a series of slowly reciprocating grids or sieves. The movement of these constantly exposes fresh portions of the substance to the action of the hydrogen, and at the same time determines the rate at which the substance falls through the case. The apparatus is heated by the hydrogen itself, the gas before its admission being heated to the requisite temperature in an external superheater or stove. In this jjarticular form of reducing apparatus the reduction of the oxide to the metallic form is effected in the lower portions of the case. In the ujjper portion the material — fed to the case in the form of the carbonate — is calcined to the oxide. After leaving the lowest grid the reduced material accompanied by the kie.selguhr, must not. of course, be permitted to come into contact with the air. It is therefore caused to fall into a tank of oil of the same kind and quality as that to be hardened. After thorough mixing the black oily preparation is ground to a suitable consistenc}^ and is then finallj' ready for admission to the autoclave along with the oil. The Laxe Autoclave. An autoclave, designed according to Mr. Lanes patents, is illustrated in section in Fig. 71. It is a cylindrical upright vessel, closed top and bottom, and surrounded by an outer jacket of fire-brick to constitute a flue for the gases of a separately fired furnace. The upper half of the vessel is occupied by agitating gear consisting of a series of square beater discs A mounted on a power-driven vertical shaft, and an equal number of metal plate cones B formed with square holes at their centres, and fixed relatively to the walls of the vessel. The lower half is, in the working condition, occupied by the oil to be tieated mixed with the catalyst. A pump C draws the oil from the foot of the vessel and discharges it continuously on to the uppermost of the cones B. Falling from this on to the first of the beater plates A it is shot off against the walls of the vessel, and is discharged through the opening in the second cone on to the second beater plate. Before it returns to the bottom half of the vessel the oil is thus thoroughly churned up in the atmosphere of lijdrogen under pressure which fills the upper half of the vessel. The oil and catalyst are introduced at D and the hydrogen at E. At F a connection to a vacuum pump is provided whereby, as a preliminary to the introduction of the catalyst and hydrogen, the air in the vessel can be removed. It has been found, as the result of practical experience, that the oil in the lower part of the vessel is apt to suffer from being exposed too long in contact with the hot Lane Autoclavf INSERT FOLDOUT HERE THE HYDRO(iP:NATU)X OR HARDENING OF OILS 111 surrounding walls. To overcome this Mr. Lane, in his most recent designs, extends the agitator shaft to the foot of the vessel, provides it with a beater or paddle, and surrounds it with a cylindrical jacket. The oil is thus circulated from the paddle up the annular space between the jacket and the walls of the autoclave, and down again tlirough the jacket to the paddle. In Plate VI. we give the general arrangement drawing of a Lane autoclave provided with this improvement. If matters are properly regulated the pressure inside the autoclave, as the hydrogen is pumped in. is seen to rise at first. On reaching a certain point, depending upon I ^T^HI ■ ^HH^ ^mf^i'mi^t^WiXt^^ '^SBS^KKKtk % fM 1 \ IS J r ?^er cwt., carbonate of soda at 55s. per ton. £ s. 16 14 CHAPTER XIV THE GENERATION OF HYDROGEN FOR OIL HARDENING PURPOSES So important to the successful working of the hydrogenation process of hardening oils is an inexpensive commercial method of obtaining pure hydrogen that we need make no excuse for devoting a separate chajater to the subject. There are, of course, various methods of generating hj'drogen on a commercial scale, one well-luiown one being the Linde-Frank-Caro process, which extracts the hydrogen from water gas by liquefying the nitrogen, carbon monoxide, etc., in a liquid air condenser. This process is worked in this country by the British Oxygen Co., Ltd., and, we understand, yields a gas which is suitable for hydrogenating oils. The process we propose here to deal with exclusively is, as we mentioned in our preceding chapter, that which has been develoijed within the past fourteen years by Mr. Howard Lane, of the Laboratory, Ashford, Middlesex. The basis of this process is the oxidation of metallic iron by means of steam, the oxygen oi the steam entering into union with the iron, and the hydrogen being set free. Proposals on this basis have been numerous — probably more numerous than those under any other system of producing hydrogen — but in many instances success with Ihe method has been confined to the laboratory. The undoubted commercial success which Mr. Lane has achieved with the process is due very largely, but not, as we will have to explain, wholly, to the attention which he has given to the design of the details of the plant used. The iron, under the Lane system, initially .supplied to the hydrogen retorts is calcined spathic iron ore, the jjurest form in which ferrous carbonate, J'eCOg occurs in Nature. This substance, when subjected to heat, speedily parts with its carbon dioxide, and becomes converted to a porous mass of ferrous oxide, FeO. So converted it is packed within the hydrogen retorts. The working of the process calls for the alternate reduction of the ferrous oxide to metallic iron by means of a combustible gas, and the conversion of tliis metallic iron back to ferrous oxide by means of steam. The combustible gas used for (he reduction may in a small plant be ordinary town's gas, but on a large scale purified water-gas, generated at the site, is undoubtedly to be preferred on the score of economy. When water-gas is used it is purified by the removal of the sulphur dioxide, hydrogen sulphide, carbon dioxide, moisture, and other impurities, which, as made in the producer, it contains. As admitted to the ferrous oxide in the hydrogen retorts, it therefore consists of about equal quantities of hydrogen and carbon monoxide. The reduction of the ferrous oxide to metallic iron is accom- plished at the expense of these two constituents, which are converted respectively into moisture and carbon dioxide. The reduction being complete, the supply of purified water-gas is shut off and steam at a low pressure is admitted to the retorts. The earlier portions of the hydrogen, which immediately starts to come off, are sent elsewhere than to the hydrogen holder, for they are impure to the extent that they carry with them the reducing gas, the water vapour, and the carbon dioxide, lingering in the retorts as a result of the previous reduction process. Three practical points must now be noted, for tiiey lie at the basis of Mr. Lane's method of working. In the first place, it has been found that the reduction of the 114 THE PRODUCTION AXE) TREAT.ArENT OF VEGETABLE OILS material in the retorts occupies about twice as long as the oxidation. Accordingly Mr. Lane divides his retorts into three sections, two of which are " reducing," while one is " oxidising." In the experimental plant at Ashford we found that the control valves were being operated everj" ten minutes, so that each section of the retorts was producing hydrogen for ten minutes in every half hour. In the second place. ^Ir. Lane has found a difficulty which previous workers with this process have also met, and wluch has been responsible for its being commercially impracticable, or for its being deemed so, in more than one instance. The difficulty is that, after a time, the iron gradually loses its activity', and in the end practically fails to react with the oxygen of the steam. The trouble, Mr. Lane has discovered, arises from the fact that it is not possible entirelj' to free the water-gas, or other reducing gas used, from sulphur, carbon dioxide, and other impurities. The.se impurities either combine with the iron or collect within its pores, so reducing and finally stopping, its activity. To overcome this ^Ir. Lane arranges that, at stated intervals, the working of the retorts is interrupted momentarily while air is passed through them backwards. This bums out. or otherwise removes, the impurities collected in the iron. Thirdly, the water-gas, or other reducing gas, it has been fomid, must considerably exceed in amount that theoretically necessary to efiEect the reduction of the iron oxide in the retorts. The gas leaving the retorts during the reduction period is thus mialtered water-gas, carrying with it the moisture and carbon dioxide resulting from the oxidation of a portion of the volume entering the retorts. This gas would represent a con.siderable loss but for the fact that, after removing the moisture in it, it may be deflected and used for firing the retorts. Li Fig. 70 (Chapter XIII.) the general arrangement is given of the hjdrogen- generating plant attached to an oil-hydrogenising factorj' erected to ilr. Lanes designs. The plant consists of tliree principal items, namely, (a) a hydrogen retort furnace containing the iron-working .substance which is alternately oxidised bj' the steam delivered from {b} a boiler, and reduced by the products delivered from (c) a water-gas generator. Added to these there are (d) purifiers for the water-gas, and for the hydrogen (c) holders for the two gase^. (/) compressors for the hj-drogen, and (g) reservoirs for the storage of the compressed hydrogen. Water-gas Generators an'd Pcrifif.rs. It is, we tliink, unnecessarj" for us here to enter into a description of the water-gas generators supplied with the plant. Although they embody in their design certain details representing improvements of ilr. Lane's own invention, they are in principle, and in action, similar to all other water-gas generators. Further than this they do not form an essential feature of the plant, for other combustible gases — for example, town gas — can, as we have remarked, take the place of water-gas. It is sufficient for us to say that the generators are supplied with air from a tuibine-driven blower, and with steam from the .same boiler as that supplying steam to the hydrogen retorts. The water-gas generated has, on the average, a calorific value of from 280 to 300 B.Th.U.s, and in the raw state may be said to have roughly the following composi- tion : — Hydrogen, 49 per cent. ; carbon monoxide, 43 ; methane, J ; carbon dioxide, 4 ; together with nitrogen, sulphur dioxide, hydrogen sulpliide, moisture, and dust and other mechanical impurities. The gas, on leaving the generators, passes through a superheater, where jt exchanges some of its heat with the steam flowing from the boiler to the generators. Thereafter it is led to a scrubber, where it is cooled and washed with water to deprive it of its dust. It is then taken to a gasholder. The gas as required is withdrawn from the holder by means of a " booster " or INSERT FOLDOUT HERE GENERATION OF HYDROGEN FOR OIL HARDENING PURPOSES 115 compressor and passed along to the purifiers. The booster is driven by a small steam engine, which is controlled by the pressure of the gas in such a way, that as tlie resist- ance of the purifiers increases so does the pi'essure of the gas. A by-pass is provided in order that the gas may be sent, if necessary, straight to the purifiers at the gasholder pressure. This is sometimes convenient, as, for example, when the booster has to be cleaned or repaired or when the hydrogen retorts are being run banked. The water-gas purifiers for the installation represented in Fig. 70 are four in number. They serve in the usual way to remove the sulphur dioxide, hydrogen sulphide, carbon dioxide, moisture, etc., from the water-gas. They are controlled by a centre valve of special construction, packed with hard fat, to prevent leakage, on the principle of the Stauffer gi'ease ctip. This valve is actuated in such a way that one of the purifiers is always in reserve, while the gas passes in sequence through each of the remaining three. The crudest gas always enters the foulest purifier, and leaves from the cleanest. At intervals, the foulest producer is switched off for cleaning and recharging, while the stand-by purifier is brought into action at the other end. In this way all four purifiers are cut out and cleaned in turn without interfering with the continuous purification of the water-gas. The gas, after leaving the purifiers, is ready to be passed into the hydrogen retorts. Hydrogen Retort Furnaces. The general arrangement of the Lane hydrogen retort furnace is represented in the drawings given in Plate VII. Before describing the construction and mode of action of the furnace, we would repeat what we remarked above, namely, that it takes twice as long to reduce the ferrous o.xide to metallic iron with the water-gas as to oxidise the iron with the steam. In other words, the time spent in preparing a given weight of material for the production of hydrogen is twice as great as the time occupied in the succeeding step during which the hydrogen is being generated. The furnace con.sists, primarily, of a brickwork casing containing, in the size illustrated, thirty-six vertical, cast-iron, pipe-like retorts. The top ends of the retorts are fianged and jirovided with covers for removal when the retorts have to be recharged with ferrous material. The spent material is removed through similar covers at the foot of the retorts. The thirty-six retorts are arranged in two groups, each containing two rows of nine retorts each. This division is of no practical significance. What is, however, impoitant is the division of the thirty-six retorts into three groups P, Q, R, each grouj) containing three of the retorts in each longitudinal row. While the gi'oups P and Q are " reducing," the group R is " oxidising." After running thus for a certain length of time, the gi'oup Q is changed over to " oxidising " and the gi'oup R to " reducing," the group P remaining at the reducing setting. Thereafter P is set to oxidise, and Q and R to reduce. In this way the generation of hydrogen — from the oxidishig group — is made continuous, while the double time required for reducing is allowed to each group. In order to facilitate our description, the three groups of retorts are, in the diagram of the plant given in Fig. 73, repre.sented as three .single retorts P, Q, R. Across the front of the furnace and external to the brickwork, there run six horizontal pipes A, B, C, D, E, F. The top end of the retort P is connected as at G to a valve H on the pipe A, and the bottom of the same retort as at J to a valve K on the pipe F. The top and bottom ends of the retort Q are similarly connected to valves on the pipes B, E, respectively, and the top and bottom ends of the retort R to valves on the pipes C, D. The three pipes A, B, C are coimectcd at each end to vertical pipes L, M, and 110 THE PRODITTIOX -\XD TREATMENT OF VEGETABLE OrL'< the three pipes D, E. F to two other vertical pipes X, S. The valves H,. K are inter- coimected, so as to be operated together. The two other pairs are similarly connected. The pipe X is connected with the water-gas supply. With the valve setting indicated in the diagram, the retorts P. Q are receivmg water-gas from the left-hand portions of the pipes F. E respectively. The gas rising up the retorts is reducing the ferrous oxide in them, and with the moisture and carbon dioxide, resultuig from the reaction, is passing away by the left-hand portions of the pipes A. B to the pipe L. From this pipe it may be sent wholly into the furnace for heating the retorts — ^no other fuel being necessary. As we have already said, it is not practicable to work Gas Pun Fie r Hydrogen 73. -Lhiifrram of the Lane Hvdrofjeu Eetort Kurnace. with ju.st sufficient water-gas to reduce the charge of ferrous oxide in the retorts. An excess is required, but, as will now be understood, the excess amomit in Mr. Lanes plant is subsequently usefully employed. At one time Mr. Lane utili.^ed the excess gas passed through the retorts, partly for firing the furnace and partly by retummg it to the reducmg gas purifiers, as indicated in the diagram, so as to make it available for a .second passage through the retorts. This plan has been given up, for it was fomid that the volume of carbon dioxide coming off with the excess gas was such as seriously to overtax the capacity of the purifiers. Mr. Lane now prefers to utilise tlie excess gas either wholly for firhig puj-poses or partly for firing and partly for estabhshmg a reducing envelope for the combustion chamber in which the retorts are set. the object being to minimise the wear of the brickwork. GENERATION OF HYDROGEN FOR OIL HARDENING PURPOSES 117 The pipe M is connected with a supply of steam, ^^''ith the valve setting repre- sented in the diagram, the retort R is receiving steam from the right-hand portion of the pipe C. Tliis steam passing downwards becomes decomposed, oxidising the metallic iron in the retort and setting free hjalrogen. Tlie hydrogen leaving the bottom of the retort reaches the right-hand portion of the pipe D and so passes into the pipe S, whence it is conducted to a purifying plant and a gasholder. It will thus be seen that simply l^y the operation of two of the three connected pairs of valves, every ten minutes or so, the plant is capable of giving a practically continuous output of hydrogen gas. Two practical points have, however, to be noted. hi the first place, when any one of the retorts is changed over from "" reducing " to ■■ oxidising," it is at the moment of the change filled with water-gas carrying a certain percentage of moisture and carbon dioxide. The first portion of hydrogen formed is therefore bound to be contaminated with these substances. To avoid passing this impure gas into the pipe S, it is arranged that the valves, while they can be operated simultaneously in pairs, as stated, can also be operated separately. Thus, when the reducing period in, say, the retort P is completed, the valve H is operated to admit steam to the top of the retort, while the valve K is for the moment left untouched. The steam being at a higher pressure than the water-gas, passes down the retort and becomes converted to hydrogen. This hydrogen mixing \^ith the water-gas in the retort causes the latter to flow back into the pipe F and the pipe N. Tho impure hj'drogen then passes with the fresh water-gas into the retort Q and the retort R — now set for reducing — and is therefore not wasted. In a very short time the hydrogen generated is sufficiently pure to permit the valve K to lie operated so as to allow the retort P to take up its projjer function. When jjassing from "" oxidising '" to " reducing," the retort is at first filled with pure hydrogen. This, beyond rei^resenting a small waste, is of no significance, as the incoming water-gas will merely be enriched in hydrogen to a iDro^Jortionate extent. The pair of valves can, therefore, be operated simidtaneously when the change from oxidising to reducing is being made. In the second place, as we have already said, the ferrous material, unless revivified in some way, very soon loses its activity and fails to decoraijose the steam. This, phenomenon, Mr. Lane has found, is due to the dejiosition on the iron of sulphur and other impurities which, even with very careful purification of the water-gas, accumulate in the retorts during successive periods of reduction. The practical cure devised for the trouble is at intervals to blow air through the retorts, so as to burn out the accumu- lated impurities. To effect this, the three-way cock T. Fig. 73, is turned to shut doM'n the supply of water-gas and to open a branch pipe leading from a fan or other blower. The three-way cock U on the excess water-gas outlet pipe is also turned so as to close this pipe and open a branch pipe leading to the atmo.sphere. The air from the fan, if the retort valves are placed in the " reducing " position, then passes up through the retorts, and with the sidphur dioxide and other products derived from the impurities in the ferrous material blows off into free space. The action of the wat*r-gas on the ferrous oxide during the reducing pei-iod results, as we have said, in the excess water-gas passing off being laden with moisture and contaminated with carbon dioxide. For efficient combustion that portion of the water-gas used for firing the furnace should not lie heavily laden with moisture. Accordingly, somewhere at or near the ]ioint V (Fig. 73), the excess water-gas is taken oil to a condenser and returned. The po.sition of the condensers relatively to the furnaces is indicated in the plan given in Fig. 70. With a little study of the drawuigs given in Plate \M1. the lines on which the lis THE PRODUCTION AND TREATMENT OF ^^:GETABLE OILS GENERATION OF HYDROGEN FOR OIL HARDENING PURPOSES llf) design of tlie hydrogen retort furnace is carried out in practice will now readily be understood. Several points, however, may usefully be called attention to. The retorts are of cast iron and are 9 in. in internal diameter, li in. thick, and 9 ft. 9 in. long. Each fits into a base socket and seats therein on a joint of asbestos. The three groups of retorts P, Q, R, as sho\^ii in the plan, are each divided into two equal sub-groujjs. Six pipes — see the side elevation — run horizontally along the two sides of the furnace exterior. To each of these jsipes the top — or the bottom — ends of a .sub-group of the retorts are connected. Each of the six pipes on one side of the furnace is connected to the corresjjonding pipe on the other side by a horizontal pipe extending across the front of the furnace. This front pipe in each instance is inter- rupted at a suitable point to couple up with the two flanges F, G of the revensing cock — see Fig. 74. The top ends of the twelve retorts in each group are thus connected to one such reversing cock, while the bottom ends of the same twelve retorts are connected to a second reversing cock situated directlj^ in line with and below the first, as shown in the front elevation in Plate VII. In front of the six front pities referred to, and comiected to the flanges H, J of the reversing cocks, lie the six pipes represented in the diagram Fig. 73, at AB — F. The vertical pipes L, M, N, S in the diagram are clearly shown in Plate VII., the only point to notice being that in practice the pipes L and M are respectively united to the pipes N and S, and are not separated therefrom. A blank, however, is interposed between the flanges of each pair. With the cock plugs turned anticlockwise through about 30 degrees from the position shown in the plan, Fig. 74, the ports A and B are opened, and reducing gas is sent upwards through the retorts. A 60-degree movement of the plug in the clockwise direction from this position opens the ports A, C, and causes steam to pass downwards through the retorts. It will be noticed that a fourth and fifth port are formed in the body of the reversing cock, and that when the cock is in the central position shown in the engraving these two ports are open to one another. The flange K in each of the three lower reversing cocks is open to the atmosphere. In the three upper cocks it is connected by a vertical pipe to a horizontal pipe extending across the top front edge of the furnace casing. At one end of this horizontal pipe an ejector is fitted. By turning the reversing cocks into the central position, and setting the ejector to work, air is drawn upwards through the retorts for the purpose of burning out the impurities which, in time, accumulate on the ferrous material. The use of an ejector in this way instead of a fan, as indicated in Fig. 73, has certain obvious advantages, and is now Mr. Lane's standard practice. It will be noticed from the front elevation in Plate VII. that the spindles of the three upper reversing cocks are extended down to the level of the lower cocks so that the handles of each pair are brought close together. The two cocks of each pair can thus be moved simultaneously as when passing from '" oxidising "' to " reducing," or separately as when the impure hydrogen has to be blown momentarily into the water-gas pipes at the commencement of the oxidising periods. During the normal running of the furnace the excess water-gas is, as we have said, made use of in part for firing the retorts. At the commencement of a run a valve near the reducing gas inlet is closed and another one on the same pipe is opened. This enables the furnace to be fired with water-gas taken direct from the supply main. These means are also called into use during the slack periods, when the generation of hydrogen is interrupted. The firing may he reduced during such periods, but it is not desirable that it should be totally stopped. Generally Mr. Lane recommends that the plant should be run continuously day and night ; but if this is impracticable, he recommends that the temperature of the furnace should be 120 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS kept as nearlj- equal as possible, for this reduces the wear and tear on the furnace work. The plant illustrat-ed in Plate VII. gives an output of about 3.500 cub. ft. of hydrogen per hour. For smaller plants having hourly outputs of, say. 250 to 1,000 cub. ft. ordinary town's gas is conveniently used for i-educing the ferrous material and for firing the retorts. The small experimental plant at ilr. Lanes laboratory — see Fig. 75 — is operated in this manner. But for plants above such outputs up to the largest size — say. 10,000 cub. ft. per hour — it is distinctly economical to install with them their own gas producers. The puritj' of the gas generated by Mr. Lanes process is guaranteed by him to be from 99 to 991 per cent. In practice, however, this, we are informed, is exceeded, the purity reaching as high as 99| per cent. The purification of the hydrogen after it leaves the retorts con-^ists of passing it through a scrubljer. Fig. 75. -Experliucntal llyilrogen Plant. where it is washed with water, and then through ijuiifiers in which lime is employed to remove minute traces of such impurities as sulphur. After purification the gas is passed into a holder, whence it is withdrawn as required, compressed to a pressure of anjiihing up to 3,000 lb. per square inch, and stored in a batteiy of weldless steel cylinders. From these it is allowed to expand at the proper pressure into the oil hydrogenising autoclaves. The cost of producing hydrogen by this method is difiicult to state, for it depends almost entirely upon the local prices of fuel and labour. It may. under normal con- ditions, be expected in the average case to varj- from 3s. dd. to 7s. Gd. per 1. 000 cub. ft. In some cases, however, it may be as low as 2s. (id., or less actually than the cost of town's gas in the London area. Before leaving this account of Mr. Lanes apparatus, it ma}' perhaps be stat«d that his oil hydrogenating plant was the outcome of the success which attended his efforts to produce pure hydrogen in large quantities under commercial conditions. It would appear likely that in the near future Mr. Lanes hydrogenising plant may be applied to substances other than the classes of oil named. A veiy promising, and, if GENEHATIUN UF HYDROGEN FOR OIL HARDENING PURPOSES 121 successful, a very important application of it, lies in its use for hydrogenising mineral oils. It may yet be possible to synthetase petrol by its means. Mr. Lane has already succeeded in devising apparatus whereby he can cause hydrogen to combine with acetylene, CM.,, to produce ethylene, C.Hj, a gas which can be liquefied at a tempera- ture of 0" C. by a pressure of 41 atmospheres,, and which possesses great energy as a motive-power fuel. CHAPTEE XV THE MAXITACTURE OF SOAP SoAP-siAKDTG provides a very important industrial outlet for the employment of vegetable oils, although, of course, the soap maker also uses large quantities of animal oils and fats. Chemistry of Soap-m.\ki>g. As we remarked in a preceding chapter, fatty vegetable and animal oils may be considered as consisting essentially of a glycerine part and an acid part. To the soap maker the acid part is the portion of prime importance. Li the process of manufacture the acid part is caused to unite with an alkali, the glycerine pai't beuig in general left over as a bj^-product. It is, of course, a very valuable bj'-product, particularly at the present moment, and, as a consequence, we fuid it an increasingly common practice, particularly on the Continent, to recover the glycerine from the oil by special processes in deglj^cermising works, which carry on their industry quite apart from that of the soap maker. Under these conditions the soap maker works with the by-product of another industry, namely, the fatty acid stock discarded from the deglycciinising works. With the plant employed in the latter works we do not propose in this chapter to deal. Our attention will be devoted solely to the manufacture of soap from undivided oils and fats. When an acid (sa3', sulphuric acid) is caused to act on a metal (.say, copper) a salt (copper sulphate) is produced. If the acid is the fatty acid contamed in a vegetable or animal oil or fat, and if the metal is either sodium or potassium, the salt produced is known as a soap, a hard soap if sodium is the metal and a soft soap if it is potassium. Other soaps are possible and are made. Thus practical uses are found for soaps obtained by substituting for the alkali metals either iron, nickel, cobalt, zinc, magne- sium, aluminium, copper or mei'curj'. These " soaps " are, in general, in.soluble in water, and are used for such purposes as waterproofing agents for canvas, as " driers " to be added to boiled oil or varni.sh, as constituents of anti-fouling compositions for ship bottoms and so on. We need say nothing more about the manufacture of these " soaps " than that they are made similarly to ordinary soap, or by emplonng such soap as a basis for decompo.sition. Pure hard soap is thus the fatty acid salt of the metal sodium. It should be perfectly neutral. It contains none of the glycerine of the oil or fat from which it was formed. Pure soft soap is the neutral fatty acid salt of the metal potassium. In its commercial production, practice is diAnded as to whether or not it should be freed from the glycerine of the oil or fat u.sed in its manufacture. It seems to be established that if the glycerine is removed the quality and ajjpearance of the soap suffer, and accordingly it is quite a common practice to allow the gljxerine to remain in the .soap. Soap Boiling. There are two distinct methods of making hard or soda soap, namely, the hot and the cold processes. The latter has a restricted application, and is not of sufficient THE MANUFACTURE OF SOAP 123 importance to be considered here. Under the hot process the sodium is presented to the fatty acid of the oil or fat in the form of an aqueous solution of caustic soda, NaOH. This solution is added gradually to the oil or fat in a soap kettle, the whole being kept boiling. A typical soap kettle, made by W. J. Eraser & Co., Ltd., of Dagenham, Romford, Essex, is illustrated in Fig. 76. It is built up of mild steel plates, and contains several separate steam heating coils and a swivelhng outlet pipe with a chain hoist, whereby the soap, when formed, may be drained off. At the foot of the kettle an outlet is provided for the liquor separated from the soap. The size of these kettles may vary from 8 ft. diameter b}^ 8 ft. deep to 13 ft. diameter by 14 ft. deep, and their capacity from 5 to 25 tons. The boiling, it will be seen, is carried out at atmospheric pressure. This is the common practice. A recent improvement consists in conducting the operation under about 100 lb. of pres-sure in a closed vessel some 4 ft. in diameter by 8 ft. high. The soda solution is added gradually to the oil or molten fat in the kettle. If it is added too rapidly the saponification process is retarded. On the other hand, the total amount of soda solution mixed with the oil or fat must be more than the quantity theoretically necessary completely to saponify the substance. An excess is required, because if the theoretical amoimt only is used a point is reached at which the soap formed up to that point, the oil or fat yet remaining to be saponified and the alkaline solution corresponding to this quantity of oil or fat will establish a balance. Fig. 76.— Soap Kettle— Fraser. When the boiling operation is completed, the kettle contams, first, soap, and secondly, water, in which are dissolved the surplus caustic soda and the glycerine set free from the oil or fat. Various impurities from the caustic soda and some animal or vegetable tissue or other non-saponifiable matter from the oil or fat used will also be present. The mass in the kettle, for the moment, is a more or less clear homo- geneous substance. Soap, however, is scarcely, if at all, soluble in a solution of salt. Accordingly, dry common salt is shovelled into the kettle, and the whole contents are thoroughly boiled up again. The salt entering into solution causes the soap on cooling to separate out on the surface. The aqueous Hquor below the soap containing caustic soda, salt and glycerine in solution is run off through the bottom of the kettle and sent to the glycerine recovery department. The soap layer is now boiled up again with water and again salted out. The aqueous liquor is run off and the boiling and salting process repeated a third time. Thereafter, the soap left when the third liquor is drained off is given a fuial boilmg with water in order to hydrate it to the correct degree. It is not, however, subsequently salted, but is allowed to stand for some few days imdisturbed. At the end of this time it is fomid to have separated into three layers. At the foot there is a small layer of alkaline liquid. Intermediately, and amounting to about a third of the whole ma.ss in the kettle is a layer of dark- coloured soap called tlic " nigj-e." This substance contains traces of caustic soda and 124 THE PRODUCTION AND TREATMENT OF VEC4ETABLE OILS salt solution, and owes its darl< colour to the ]3resenoe in it of soaps of iron, co}>per. and other metals. Above this is the "' neat " soap which, being practically pure and neutral, is in a condition to be used. The " nigre,"" after removal, is boiled and salted and otherwise treated for the recovery of its valuable portions. Crutching. The " neat " soap is, as we have said, in a condition to be used. Li nearh' every case, however, it is passed into a " crutching " machine, wherein colouring, scenting or other matter is added to it. Among such other matter are various " fillei-s," such as clay, talcum, chalk, barytes, seed husks, asbestos, magnesium salts, and starch. These substances increase the weight of the soap, and are frequently regarded as adulterants. In some cases the soap is " filled " with either the borate, carbonate or sihcate of soda. These fillers have themselves distinct cleansing propeities, so that their addition is not strictly to be classed as adultei'ation. A crutching machine made by E. Timmins & Sons, Ltd., ■Runconi, is shown in Fig. 77. It consists of a double - walled steam-jacketed cylindiical vessel containing a vertical power-driven shaft, from which four or more beater arms extend horizontally. Six or more fixed arms springing from the inner surface of the vessel co-operate with the rotating a r m s . Very frequently t h e 1* ^ idiO 1 Fig. 77. — Crutching Machine — Timmins. crutchers are arranged in paLivs, as showii in Fig. 78, where a twin set, made by R. Daglish & Co., Ltd., of St. Helens, is represented. The practice here indicated of driving the machines by an attached single-cylinder steam engine is quite usual, for it permits the exhaust steam from the driving engine readily to be utilised in the jackets of the crutchers. In Fig. 79 we illustrate in cross-section a steam-driven crutcher made by George Scott & Son (London), Ltd. Soap Frames. The soap, while still hot. is run out of the crutching machines into moulds or " frames," where it is allowed to cool and set. A typical soap frame, made by Messrs. Timmins, of Runcorn, is showni in Fig. 80. These frames have removable sides, so that they may be knocked down when the soap has solidified. They are frequently made of cast iron, but mild steel is now being commonly em])lo}ed. The capacity of each is anything from 3 to 10 cwt. of soap. THE MANUFACTURE OF SOAP 1^5 Slabbing and Cutting. The slab of soap, as taken from the frame, has to be cut up into bars, and these bars have, commonly, to be again cut into tablets. The original slab is first subdivided into several slabs of lesser thickness, and each of these is cut up into bars by means of a machine, such as that .«ho'^^^l in Fig. 81. The machine illustrated is made by Messrs. Timmins, of Runcorn, and has a flat table whereon the divided slab rests. By means of a hand-wheel, crank discs, links, levers, and a guided cros.spiece, the slab is pushed forward beneath a fixed bridge, from which a number of equally spaced piano wires extend vertically to the surface of the table. Means are provided for Via. "S. — Twill Crutchiii^ Machines — Daglish. adjusting the tension in the wires. The bars thus formed, if they are to be further divided into tablets, are taken, separately, to a cutting machine of the type illustrated in Fig. 82. The machine illustrated is made by Messrs. Daglish, of St. Helens. In the block of wood A a number of vertical saw cuts are formed, while along the top a vee-sectioned recess is provided for the reception of the bar of soap. The frame B is pivoted on a rod C at the back of the machine and carries a number of equally spaced piano wires, which register with the saw cuts in the block A. With the bar of soap in place the frame is simply pressed dowii by hand. Drying. cut into tablets contains round about 33 per cent, of water, The soa[) as thu and for this reason is comparative!}' soft and sticky. It is customary, therefore, to 12G THE PRODUCTION AXD TREATMENT OF \'EGETABLE OILS subject it to a dr\ing treatment, in oi-der to form a crust of hard soap roun'l the soft interior. By so doing further evaporation from the body of the soap is retarded and the weight is preserved. In addition, the dn-ing of the crust is eseential, if, as is frequently the case, the tablets after cut- ting, have to be pressed. It is impossible to carry out this pressing if the criLst is not hai-d, for the sticky soap is bound to adhere to the press dies. Practice as regaid-; diying has recently undergone a change. Formerly, the soap was dried simply by l^lacing it in a room heated by steam pijjes or coils. The improved modern method makes use of a warm air blast. Apparatus for this purpose, made by Messrs. Eraser, of Dagenham. Essex, is illustrated in Fig. 83. It consists simply of a steam heater through which air is driven by an attached fan into a wrought iron easing provided with hinged doors and containing several tiers of galvanised iron wire trays for holding the soap. STASrPIXG. Fig. 79. — S^team Driven Crutcher — -Scott. their api>earance. A hand-machine for this purpose, made by ilessrs. Daglish. of St. Helens, is illustrated in Fig. 84. This machine is capable of dealing with either 1-lb. or } lb. tablets. It con- sists of two balanced fly-wheels united b3' a crosspiece with handle, and operating, through an arm. a phmger. which works within a sleeve pivoted at its lower end to the frame of the machine. The bottom of th? mould box is loose, and is designed to rise on the upstroke of the plunger, so as to eject the soap from the mould. The tablet is frequently stamped twice, once on each face. CHIPriNG AXD MlLUXG. Soaps prepared as described above are suitable for many purposes, notablv for lamidry and similar work. They are liable, however, with time to lose weight by shrinkage and othenvise to deteriorate The rough tablets are, after being dried, verj- commonly stamped to improve Fig. hi.— Soap Frame — Tin.- For the production of the best quality of toilet soaps, the process known as milling is resorted to. The first step in this process THE MANUFACTURE OF SOAP 127 is to reduce to cliips the soap as taken from the frames. Tlie slabs are first cut into bars and paiiially dried. Thereafter, they are taken to a chipping machine, such as Fig. si. — Slab Cutting Machine — Timmiiis. that shown in Fig. 85, which illustrates a double-sided machine made by Joseph Baker & Sons, Ltd., of Willesden Junction, London. The bars of soap are placed in the Fk;. Sli. — Bar Cutti shoots shown, so that their ends may come in contact with the blades of the rapidly revolving cutters disposed within the casings. The chips fall from the foot of the casings on to trays supported on the angle-iron runners shown. The thickness of the chip.s can be regulated to suit requirements. 128 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS The chips are next dried until thej' contain round about 10 per cent, of water. For this operation the driving jilant ilhistrated in Fig. 83 is suitaljlc. Colouring and Fig. S3. — Soap Di-yin^sr Plant — -Fraser. scenting materials are then added to tlie dried chiiJs, and the whole is ground up in a toilet soap mill. A machine of this description made by INIessrs. Baker, of \\'illesden Junction, is illustrated in Fig. 86. This mill consists of five granite, or sj'^enite, rollers very carefully ground to truth. The rolls are 31 in. long, the four lower rolls being 13 in. in diameter, and the top roll 19 in. The lowest roll and the third roll run at a relatively slow speed. The second and fourth rolls run at about twice this speed, while the top roll runs at about four times the speed of the lowest roll. A double hopper is arranged in front of the rolls. The soap chips are fed into the lower division of this hopper and pass thence to be ground between the differentially moving rolls. As the material comes round the fifth roll, it is scraped off into the upper division of the hopper. Wien the whole batch has accumulated in this division a shutter at the foot is with- dra\\n. and the charge allowed to return to the lower division for a second pass through the mill. From .. four to eight passages are frequently ■''" "" '■ given to the material, the number depending upon the quality desired in the resulting product. At the end of the last pass the soap is scraped off in the form of thin ribbons from the back of the top roller. Fig. S4. — Stamiiing Macliiiic- The manufacture ok soap 129 I'l'-i. i)(j. — MilliiiL' Mauhiiie— Baker. 130 THE PRODUCTION AXD TREATNIEXT OF \T:GETABLE OILS THE MANUFACTURE OF SOAP 131 The mass of ribbons is next transferred to a squeezing machine or " plodder," of which an example made by Messrs. Fraser, of Dagenham, Essex, is illustrated in Fig. 87. This machine squeezes the soap through a perforated die plate A, so that it comes out in the form of small round threads. Thereafter, the die plate is removed and the material is once more passed through the machine. As it passes through the die plate B at the end of the cone-like mouth-piece, the soap is squeezed into a solid bar, which is received on the table C and thence passed to a cutting and stamping machine. The squeezing of the soap through the die plates is effected by a short worm rotating at about 20 revolutions per minute at the foot of the hopper, and arranged coaxially with the conical mouthpiece. This worm is fed with soap from the hopper by the action of a finger shaft D within the hopper, and driven by gearing from the worm shaft. A heating jacket is provided round the worm, to faciU- tate its work on the soap. When the machine is stopped at the end of a run, the cone is still filled with soap. To remove this the cone is hinged so that it may be swung downwards and clamped within the bracket E. The die plate B having previously been removed together with the cover or cap, which holds it in place, the hand-wheel F is operated so that the piston-like head formed on it may rise within the cone and eject the soap upwards. The output of this machine is from 3 to 5 cwt. per hour. A modification is sometimes to be found in use. In this the squeezing machine is com- bined with the mining machine. It is very doubtful if such a combination is as satis- factory as keeping the two machines apart. CHAriER XVI GLYCERINE RECOVERY AXD REFINING AND THE SPLITTING OF OILS Glycerine and a fattj- acid are, as we have remarked, the two essential parts of every animal or vegetable oil or fat. It must not, however, be thought that such an oil or fat consists simply of a mixture of these two substances. In reality neither glycerine nor fatty acid should exist separately as such in a neutral oil or fat. If they do, particularly if free fatty acid is present, we have a sign that the oil or fat has suffered some decomposition. The matter may be put with advantage in a popular May without introducing cumbersome chemical fonuulse. A molecule of oil consists of a molecule of glycerine — less an atom of hydrogen and an atom of oxygen — and a molecule of fattj' acid — less an atom of hydrogen. It will be noticed that the missing atoms together constitute a molecule of water. If this molecule of water can be added to the oil under suitable conditions then the molecules cf glycerine and fatt}- acid will be made complete and will separate from one another. If, instead of water, HOH, we add a molecule of caustic soda, NaOH, the glycerine is again made complete, but the fatty acid molecule receives a sodium instead of a hydrogen atom, and separates not as a fatty acid, but as a soap. Lime, Ca(0H)2, acts similarly and jields glj'cerine on the one hand, and a lime soap, insoluble in water, on the other. Caustic potash, KOH, also acts in the same manner, giving glycerine and soft soap ; and so on for other hj'droxides. The splitting up of vegetable and animal oils into glj-cerine and fattj- acid forms an important branch of industrj-, and is carried out in a variety of ways. Thus it can be directly effected by subjecting the oil to the prolonged action of superheated steam — a fact which explains wh}- animal and vegetable oils are not so popular as mineral-oils for lubricating parts of machine ly and engines subjected to high tempera- tures. The .splitting up can also be achieved bj- treating the oil with lime, drawing off the glycerine thus set free, and treating the lime soap further with sid^ihuric acid to convert the soap into fatty acid with the liberation of calcium sidphatf . There are several other important proces.ses of carrying out the work. Their object is, of course, to obtain the valuable glycerine bj- a direct method, and to recover the fatty acid as a Ijy -product, which maj- be sold to the soap maker or candle maker. The soap maker, as we have stated, very frequently prefers to work with the ■^^hole oil ( r fat and not with the fatty acid by-prcduct of the de-glyceriiiising works. He prefers to do so because the glyceiine thus comes under bis own control, and forms a valuable adjunct to his business. Apart from this it is beUeved that the recover}- of the glj^cerine is more complete if performed after the soap has been made than it is if the oil or fat is spht beforehand. Again, the recover}- of the glycerine if perfoimed at the soap works enables the soap maker also readily to recover the salt \^hich he uses to separate the soap in the kettle. Finallj-, it is stated that the production of soap from fatty acid stock requires much more skill thin is necessarj- if an unsplit oil or fat is used. In this chapter we propose to deal, first, with the recover}' of the glycerine .set free as a result of the soap making process ; secondly, with the splitting of oils and fats as carried out in de-glycerinisiug works and elsewhere and finally with the refining of crude glycerine. GLYCERINE RECO\rERY AND REFINING AND SPLITTING OF OILS 133 Glycerine — or glycerol, to give the perfectly pure body its proper scientific name — ■ is present in all vegetable and animal oils and fats to the extent on the average of about 10 per cent, by weight. It is, of course, known to the public as a colourless, odourless, sweet-tasting, syrupy liquid, but its fluidity appears to be due to its impurity. Pure glycerine is a solid at all temperatures up to about 17' C, at which point it melts. In its common form it is a liquid weighing about 1} times the weight of an equal volume of water. It is combustible, and burns to water and carbon dioxide. Its boiling- point, like that of water and other liquids, depends, of course, upon the pressure to which it is subjected. At normal atmospheric pressure it boils at 290° C. and in so doing suffers some decomposition. Its distillation cannot therefore be satisfactorily performed except under a reduced pressure. At an absolute pressure of 1 lb. per square inch it boils at 210° C, and at one-tenth of a pound per square inch it boils at 163^ C. It can be mixed with water in any degree, but is insoluble in benzene, carbon disulphide, and oils. The two first -named substances, as we have seen, readily dissolve oil. They will not, however, dissolve glycerine when separated from the fatty acid combined with which it forms an oil. On the other hand, glycerine itself is a very ready solvent for a large number of substances, rivalling, and at times surpassing, water in this respect. Among such substances are many metallic salts and halogen compounds, certain metallic oxides, caustic alkaUes. and various metallic soaps, that is to say, soaps in which the sodium or potassium is replaced by other metals such as iron, magnesium and calcium. Recovery of Crude Glycerine from Soap Works Spent Lyes. All the facts we have just mentioned have an important bearing on the problem of recovering the glycerine from the " spent lyes " of a soap works. The spent lyes run off from a kettle in which hard, i.e., soda, soap, has been made, consist of water in which various bodies are dissolved, and with which small proportions of various insoluble substances are mixed. They contain first of all nearly the whole of the glycerine combined in the original oil or fat from which the soap has been made. This constituent may amount to, say, about 6 or 7 per cent, of the whole, and is, of cour.se, dissolved in the water. Next in imi)ortance comes the salt — sodium chloride — which in the soap-making process is thrown into the kettle to cause the soap to rise and separate itself from the rest of the contents. This salt is dissolved in the solution of gljcerine and water, and is present in sufficiently large quantity to make its recovery from the lye an important element in the economy of the soap works. The lye also contains in solution a small amount of the caustic soda used to saponify the oil or fat, for it is impossible to work with just that amount of soda which is necessaiy to effect the saponification of the given quantity of oil or fat. The excess soda is dis.solved in the glycerine-salt solution.- The glycerine, salt, and soda are the three chief consti- tuents of the aqueous lye, and would be the only constituents if everything were theoretically perfect. In practice, however, the oil or fat, the salt and the soda, are never pure or anything like it. The oil or fat is sure to contain mucilage and albuminous matter, while the soda and salt between them contribute various chemical impurities, such as metallic salts and sulphates, sulphides, and other bodies. These pass into the lye unaltered, or in combination. In addition, the lye nearly always contain- a small amount of soap, for sodium soap is not completely insoluble in salt water. The spent lye is thus a very complex substance. For many years it was regarded as practically useless, it being held that the cost and trouble of recovering the glycerine from it were too great to make the undertaking pay. Glycerine in those days was in very limited demand. With the invention of djTiamite and nitro-glycerine, substances 134 THE PRODUCTION AND TREATMENT OF "\'EOETABLE OILS which to-day afford by far the greatest outlet for glycerine, the circumstances were altered, and great attention came to be jiaid to the recovery of glycerine by soap makers. "The EHGiNctr." Swain Sc. Fio. 88. — Single Effect Vacuirm Evaporator for Concentrating Crude Glycerine — Scott. Its enhanced value then made it profitable to devote considerable pains to its recover^', and to this was added as an incentive the practicability of recovering the salt simul- taneously from the lyes. To-day, soap makers, as a ride, have modified their soap- making practice to the end that the glycerine may be recovered more readily and in GLYCERINE RECOVERY AND REFINING AND SPLITTIN(i OF OILS 135 a purer form than used to he the case. In particular thej' have largely abandoned the use of certain crude saponifying chemicals in favour of others which are less likely to contribute undesirable impurities to the lye. Flu. S!l.- ■Iliit N'aeuum Kvapomtm tor I'oucentriitiiiL; (ilj'cerine — Scott. Purifying the Lye. The first step in the treatment of the lye is to acidify it. This is commonly done by running it into a tank and adding hydrochloric acid to it. The result of this is that the free caustic soda is converted into common salt and water, while any soap dissolved in the lye is decomposed into free fatty acid and common salt. S'multa- neously, iron sulphate, aluminium sulphate or common ahun is added to the lye. This combines with the free fatty acid to form a metallic soap, which, being insoluble, is precipitated. At the same time, these chemicals coagulate and precipitate the 136 THE PRODUCTION AND TREATMENT OF VEGETABLE OILS albuminous and other colloidal matter in the lye, just as they do in the case of their application to sewage purification. The treated lye is then i:)assed through a filter press and sent into a second tank. It consists now, primarily, of water, glycerine and common salt, with the excess of hydrochloric acid and ferric or other sulphate added during the pieceding treat- ment as im2:)iirities. Caustic scda is, there- fore, carefully added to it until it becomes neutral by the conversion of the hydro- chloric acid into salt. The soda also acts on the ferric or other sulphate, the result of the reaction being the precipitation of insoluble iron hydroxide and the formation of sodium sulphate which passes into solution in the glycerine, salt and water lye. The liquid is now once again filtered and is passed into a third tank, whence it is withdrawn as required for further treat- ment. CONCENTEATIOX. This further treatment consists of con- centrating the liquid by the evaporation of its water portion. As the concentration j^roceeds the salt, or the bulk of it, is thrown out of solution and can ultimately be collected and used again in the soap kettle. It is clear that all the salt camiot be removed fiom the Ij'e simjjly by evaporation of the water. Even if the evaporation were carried to completion there would still remain a fair amount of salt dissolved in the glycerine left behind. As a fact, the liquid resulting fiom the evaporation is what is known as crude glycerine, and at the best consists of, say, 80 per cent, of pure glycerine and about 10 per cent, of salt, the remainder being water and certain chemical imj^urities. The plant emjjloyed for evaporating the treated lye at one time consisted of fire-heated pans. These were succeeded by open air steam-heated ves.sels. The fact, however, that high temperatures or prolonged heating reacted inifavour- ably on the glycerine, was soon recognised, and as a result, vacuum evaporators were introduced. These not only effect the evaporation quickly and at a reduced temperature, but economise fuel by permitting exhaust steam to be used for their heating. A tyi^ical examj^le of a modern single effect vacuum evaporator for the recovery of crude glycerine, as made by George Scott & Son (London), Ltd., Kingsway House, Kingsway, W.C. 2, is illustrated in Figs. 88 and 89. The liquid having been filtered Fig. 90.. -Removing Salt from a Yaciuim Evaporator. Fig. 91. — Automatic Salt Dischargin GLYCERINE RECOVERY AND REFINING AND SPLITTING OF OILS 137 from the second treatment tank A into the third tank B is drawn up by the vacuum into the evaporator C — Fig. 88. This vessel is provided with a tube plate near the top and near the bottom. Between these plates extend a number of vertical tubes up which the licpiid is caused to rise. The space outside the tubes and between the tube plates is filled with heating steam. The tubes are of two diameters, and are so arranged as to promote a vigorous and uniform circulation without recourse to mechanical means. It is essential to have a good circulation in these evaj^orators, for otherwise the salt, as the evaporation proceeds, \^ill deposit on the interior of the tubes and restrict or choke them, instead of falling, as it is intended to do, into the conical end of the evaporator below the lower tube plate. A good circulation may furtlier be relied upon materially to reduce the chance of the liquid " frothing " and boiling over. To eliminate all danger from this cau.se, however, Messrs. Scott fit a "■ catch-all "" D or trap with internal baffles to intercept the overflow and return it to the evaporator. The vaporising space above the upper tube plate is connected by a pipe to a vacuum pump E of special design. The pump plunger works beneath a body of water in a tank : by the displacement of this water the steam is drawn over from the evaporator and delivered through a jet condenser. The lower conical end of the evaporator is in communication with a vessel F, into which the salt falls as the evaporation proceeds. When the salt vessel is full it is isolated from the evaporator by means of a sluice valve. The further precipitation of salt is allowed to accumulate in the conical end of the evaporator until the vessel Fhas been cleared. Before the door of the salt vessel is opened a valve on the pipe G is ojierated to place the vessel for a short time in communication with the vacuum inside the evaporator. The salt resting on a metallic filter inside the vessel is thus drained of nio.st of the liquid adhering to it. which liquor is returned to the evaporator. Steam is now turned on into the salt vessel, so as to wash and dry the salt as far as possible. The washings are returned by the pipe G to the evaporator. The door of the vessel can then be opened — see Fig. 90 — the salt removed, and the vessel once again put into communication with the evai^orator. The salt thus removed is comparatively dry, and can be re-used immediately in the soap kettles. For large plants an automatic arrangement is frequently fitted by Messrs. Scott, in place of the vessel F, whereby the salt is discharged continuously. This device is illustrated in Fig. 91. Its construction is simple and obvious. In this case the salt is discharged moist and saturated with liquor, and is immediately dried and washed in a centrifugal machine. In Fig. 92 we give a view of a large glycerine recovery plant capable of dealing with 500 tons of spent lye per day. This plant is fitted with the automatic salt-extracting arrangement referred to, and witii mechanical means for convejdng the salt to and from the centrifugals. The evaporation of the liquor and the extraction of the precij^itated salt are proceeded with until, as we have said, the liquor shows a concentration representing an 80 per cent, content of glycerine. This condition is judged by noting the tempera- ture of the liquor in the evaporator, for as the water is eliminated the boiling-point of the liquor left rises. At any given pressure above or below atmospheric, there is a definite boiUng point for each and every strength of liquor. In the neighbourhood of 80 per cent, concentration the boiling-point ri,ses bj' about 1° C. for each 1 per cent, increase in the concentration. It may be remarked that even at atmospheric pressm-e the boiling point of an 80 per cent, solution of glycerine in water is no more than about 120° C, and under a vacuum it is, of course, still less. Hence exhaust steam, if available, will in most cases be quite sufficient for heating the evaporators. GLYCERINE RECOVERY AND REFININC AND SPLITTING OF OILS 139 When the concentration has reached the desired degree the vacuum pumi) and jet condenser are closed down, and the ciude gl\'cerine is lun off into stoie tanks. A fresh cliarge of liquor, \\hich in the meantime has been treated chcmical'y in the mamier described above, is immediately introduced into the evaporator. The apparatus described so far is of the " single effect " tyjie, and is suitable for use where the supply of exhaust steam is abundant. Where it is not, and where fuel Fig. 93. — Double-effect Vacuum Evaporator tor Concentrating Crude CJlycerine. is expensive, it is usual to employ a double effect evaporating plant of the type illus- trated in Fig. 93. In such a ca.se live steam is supplied to the first evaporator onlj'. This evaporator is worked at a pressure not much less than atmospheric, so that the vapour developed in it may be .sufficiently hot to be utilised as the heating fluid for the second evaporator. This second evaporator works at a high vacuum. The liquor receives a preliminary concentration in the first evaporator, and is then passed on for final treatment ii the .-^ecoiid. I'stuillv nearly all, if not the whole of the .salt is 140 THE PRODUCTIOX AXD TREATiMENT OF VEGETABLE OILS deposited in the salt vessel of the second evaporator. The first, however, is also fitted \\ith a salt vessel, so that either evajjorator may be run on the single-effect jirinciple should repairs to one unit or anj' other cause render this desirable. The jilant shown in Fig. 92 — already referred to — consists of four double-effect evaporators. The presence of the salt in soap makers" waste Ij-es is undoubtedly a disadvantage when it comes to the problem of recovering the glycerine. The crude gljxerine. as we have said, is bound to retain a considerable percentage of the salt. Even the subse- quent refining of the glycerine by distilla- tion may not entirely eliminate it. With tlie increased demand for pure glycerine which has arLsen with the development of high explosives, more and more attention has come to be paid to alternative methods of obtaining the crude product, methods which do not involve the glycerine being brought into contact with salt at anj' point or with more than a small amount of any other chemicals. The Splittixg of Oils and Fats. The " splitting '' of oils and fats can be performed in several ways. Roughly stated, the object aimed at is to make each molecule of oil take i\p a molecule of water, so as to form a molecule of glycerine and a molecule of free fatty acid, or, to speak scien- tifically, to ■' hydrohse " the oil. We have already mentioned in a previous chapter, that the hj'droh'^^is of an oil once started is liable, if \^'ater be present, to continue automatically until a very considerable pro- portion of the oil is converted into a mixture of glycerine and free fatty acid. In the case of palm oil, for example, the prevention of hj'droh^sis is very difficult if not impossible. We are now dealing with an aspect of affairs in which the encouragement of hj^drolysis may be said to be the direct object in view. 04.- — < 111 and Fat Splitting Autoclaves Scott. Autoclave Proces.'*. The two principal methods of .sphtting oils are the autoclave method and the Twitchell process. A pair of autoclaves for oil or fat splitting by ^Messrs. George Scott & Son is illustrated in Fig. i)4. These are simjjly cylindrical heating vessels and as shown, are usually not provided with agitating gear. The fat or oil is introduced into the autoclave together with 1 or 2 per cent, of some base, such as lime, magnesia, barium oxide or — very commonlj' to-day — zinc oxide. Steam at a pressure of. say, 150 lb. is then admitted to the autcclave. the pressure being maintained for from four to six hours. The small amount of chemical base used is sufficient to start the decom- position of the oil or fat into glj-cerine and a lime, magnesia, etc., soap. This, once GLYCERINE RECOVERY AND REFINING AND SPLITTING OF OILS 141 started, induces hydrolysis, and the rest of the oil or fat taking up water from the steam becomes converted to glycerine and fatty acid. Thereafter the contents of the autoclave are treated with a small amount of acid to decompose the soap formed by the base into fatty acid and a lime, magnesium or other salt. The fatty acid and the glycerine are then separated by taking advantage of their difference of specific gravities. The glycerine contains much water — it is known at this stage as " sweet waters " — and is filtered and concentrated. The concentration is effected in an evajiorator similar to that described above in connection with the treatment of soap makers' lyes. No salt extracting details are, however, jirovided, for no salt is preseiit in the liquor. About 95 j'er cent, of the oil or fat originally introduced into the autoclave is on the average converted by this method, although at times as much as 98 jjer cent, can be completely split. The remainder passes away unchanged with the fatty acid. The Twitchell Process. An important alternative method to the above is Twitchell"s process. If oleic acid — a fatty acid occurring in many oils and fats — and benzene, naj^hthalene, or certain other bodies are mixed and treated with sulphuric acid a certain compound results, known as Twitchell's reagent This compound may popularly be said to have the power of fermenting oils and fats when boiled with them at atmospheric pressure, for it readily hydrolises them into glycerine and fatty acid. The fact that the action is satisfactorily effected at atmospheric pressure gives the Twitchell process certain advantages over the autoclave method. In particular, it permits the process to be conducted in wooden vessels. In Fig. 95 we give the general arrangement of a splitting plant on the Twitchell system, erected by Messrs. George Scott & Son. To secure success with this process the oil or fat must first be freed from iron, lime, and other impurities. Accordingly, it is initially boiled with sulphuric acid in the lead-lined wooden vat B. The coagulated impurities sink to the bottom, and the clean oil is drawn off from a point near its surface level and is pas.sed into the " saponifying vessel "' C. Here it is mixed with from one-third to one-half of its weight of distilled water drawn from the tank A, and with from J to 2 per cent, of the Twitchell reagent. The charge is then agitated and boiled for a period extending up to twenty-four hours by means of steam delivered direct into it from a perforated coil within the vat. A close-fitting wooden cover is provided for the vat which, while allowing steam to escape, prevents the free access of air to the charge. This is desirable, because the hot fatty acid set free from the oil is liable to darken in colour in the presence of air in excess. When the boiling is completed the charge is allowed to stand until the fatty acid portion rises to the top and the glycerine and water portion sinks to the bottom. The latter is drawn off into the tank D, where it is neutralised with lime water and allowed to settle. The fatty acid portion may be boiled up again with water to extract the last traces of free glycerine. Any sulphuric acid in it is neutralised with barium carbonate, the addition of which to the charge results in the precipitation of barium sulphate. The neutralised glycerine water is pumped through the filter press E into the tank F. The separated fatty acid is drawn off from the tank C by the pump G. From the tank F the glycerine water is passed into the evaporator H. This is of similar construction to the vacuum evaporators used for concentrating soap makers" crude glycerine, except that no salt- discharging det.ails are fitted to it. The vapour drawn off from the evaporator down the pipe J bj' the vacuum pump L is condensed by the water injector K and is sent as distilled water into the store tank A. Storage tanks for the partially concentrated U-2 THE PRODrCTIOX AXD TREAT>rENT OF \-EGETABLE OILS GLYCERINE RECOVERY AND REFINING AND SPLITTING OF OILS 143 144 THE PRODUCTION AXD TREATMENT OF \'EGETABLE OILS glycerine drawn from the evaporator are indicated at yi, M. In these the glycerine is allowed to settle and deposit any sediment it may hold. The partially concentrated glycerine may be returned for further concentration to the evaporator, and is then finally discharged at Q. P is a sampling cock. Glycerine Refining. The ciiide glycerine recovered from soap makers' lyes after concentration may contain iij) to about S(i per cent, of pure gh^cerine. The remainder consists of, say, 10 per cent, of water and 10 per cent, of salt and other impurities. The crude glycerine derived from the autoclave or TwitcheU process of splitting oils or fats contains on the average about 85 per cent, of glycerine. The remainder is largely water, but tliere is also present a considerable amount of organic and inorganic impurities. To a certain small extent the crude glycerine obtained by either of these methods is used directly ; but for the two chief outlets for the substance, namely, in the manufacture of high explosives and in pharmacy, it is essential that the impurities and the water should be practically eliminated. This eUmination is effected by distilling the crude glycerine under vacuum followed by concentration. The refuiing or distillation of glycerine is practically an industry by itseK. Usually, for instance, the soap maker does not carry his work beyond the stage of recovering the crude glycerine. This he disposes of to the glycerine refineries. Even some of the largest producers of crude glycerine regai-d it as their final market product, and do not attempt to refine it themselves. As we have said above, glycerine distils mider atmospheric pressure at 29(»° C, and in so doing siiffers some decomposition. It camiot therefore be satisfactorily distilled at ordinary pressure bj' means of dry, external heat. In practice the method adopted is to heat it in a vacuum by, and in the presence of. superheated steam. The crude charge thus distils without decomposition of the glycerine, but the distillate, it must be noted, is not pure glycerine. It consists of glycerine vapour accompanied by water vapour, and the vapour of any of the impurities in the crude charge which are volatile. Among the latter we may include common salt, for if this body is not actually volatile at the pressure and temperatures employed it would appear that it is carried over in the distillate mechanically with the water vajjour. and is found in the condensed distillate. The ijrocedure adopted for glycerine refining is to condense the distillate in several difterent fractions. Those conden.sed at the highest tempera- tures will be purest and richest in glycerine. As the condensing temperature becomes less the percentage of glycerine in the condensate falls, until in the last condenser the condensate consists of Uttle more than water contaminated with various chemical impurities. The diagrammatic an-angement of a glycerine refining plant b}' Messrs. George Scott & 8on is given in Fig. 96, while in Fig. 97 we give a view taken in a glycerine refinery fitted up by the same firm. Referring to the diagram, A is a steel still. Lito this the crude glycerine, previously heated for preference, is introduced until the still is about half full. The remainder of the charge is added as the distillation proceeds. The still is fixed close to a furnace B, the prime object of which is to fire the superheater C, which supphes the still with steam. Incidentally the waste gases from the furnace are used to assist in maintaining the temperature of the still contents. The super- heated steam is admitted both to closed and open coils inside the stiU. The bulk of the distillation, however, is effected by the steam issuing direct into the charge from the open coils. At D is indicated the inlet for the charge of crude glycerine, and at E is shown the discharge cock for the '• still bottoms," that is, the residue left after the (iLYCERINK KECOVEllY AND REFIiNINU AND SPLITTING OF OILS 145 distillation is over. The glycerine and steam vaijours leave the still by the pipe F, and ])ass into the cooling battery G. In the case of the plant re^J resented in the diagram the distillate can be condensed in nine different fractions. From six to nine fractions are usnal. The cooling battery G gives seven simultaneous fractions. It consists of seven intermediate receivers H and seven final receivers J. The inter- mediate receivers arc connected in pairs by means of six series of air-cooled bent pipes K. Radiation and atmospheric convection result in the establishment of a temperature Via. 96. — Glycerine Eefinin'' and Conceiitratin t / J'c/O •: i