RARE EARTHS  , in

chemistry, the name given to a group of oxides of certain metals which occur in close association in some very rare minerals . Although these metals resemble each other in their chemical relationships, it is convenient to subdivide them into three groups: the cerium, terbium and ytterbium groups . The first includes scandium (Sc, 441.1), yttrium (Y, 89.o), lanthanum (La, 139•o), cerium (Ce, 140.25), praseodymium (Pr, 140.6), neodymium (Nd, 144.3), and samarium (Sa, 150.4); the second includes europium (Eu, 152.0), gadolinium (Gd . 157.3), and terbium (Tb, 159.2); and the third includes dysprosium (Dy, 162.5), holmium (Ho, ?) erbium (Er, 167.4), thulium (Tm, 168.5), ytterbium or neoytterbium (Yb, 172.0), and lutecium (Lu, 174.0); the letters and numbers in the brackets are the symbols and atomic weights (inter-national) . Although very rare, a large number of minerals contain these metals; they are chiefly found in Scandinavia, parts of the Urals, America and Australia, generally associated with Archean and eruptive rocks, and more rarely with sedimentary deposits . They are usually silicates, but many complex tantalates, niobates, phosphates, uranates and fluorides occur . The chief mineral species are monazite, a phosphate of the cerium metals, containing thorium (this mineral supplies the ceria and thoria employed in making incandescent gas mantles); cerite, a hydrated silicate of calcium and the cerium metals; gadolinite, a silicate of beryllium, iron, and the yttrium metals; samarksite, a niobate and tantalate of both the cerium and yttrium metals, with uranium, iron, calcium, etc.; and keilhauite, a titanosilicate of yttrium, iron, calcium and aluminium; other species are fergusonite, orthite, aeschynite, euxenite and thorianite . The chemistry of this group may be regarded as,beginning with Cronstedt's description of the mineral cerite from Bastnaes in 1751, and the incorrect analyses published by T . O . Bergman and Don Fausto d'Elhuyar in 1784 . Ten years later Gadolin investigated the mineral subsequently named gadolinite, which had been found at Ytterby in 1788 by Arrhenius . This discovery of a new earth was confirmed by A .

G . Ekeberg in 1799, who named the

base yttria . Cerite was examined simultaneously by Klaproth in Germany and by Berzelius and Hisinger in Sweden; and a new base was discovered in 1803 which the Swedish chemists named ceria . Both these oxides have proved to be mixtures . In 1839 Mosander separated " ceria " into true ceria and an earth which he termed lanthana (Gr . XavO0.w€u', to lie hidden), and in 1841 he showed that his lanthana contained another base, which he called didymia (Gr . ScSuµoL, twins) . This didymia was separated in 1879 by Lecoq de Boisbaudran into a new base, samaria, and a residual didymia which was shown by Auer von Welsbach in 1885 to consist of a mixture of two bases, praseodidymia and neodidymia; more-over, samaria was split by Demarcay in 1900 into true samaria and a new base named europia . In 1843 Mosander also split 9TO yttria into two new bases which he called "erbia" and "terbia," and a true yttria, but in 186o N . J . Berlin denied the existence of Mosander's " erbia," and gave this name to his "terbia." The new erbia has itself proved to be a mixture . Marignac in 1878 separated an ytterbia which was split by Nilson in 1899 into scandia (the metal of which proved to be identical with Mendeleeff's predicted eka-boron)and a new ytterbia, which, in turn, was separated by Urbain in 1907 into neoytterbia and lutecia (C .

A. von Welsbach proposed for these elements the names aldebarianum and cassiopeium) . Berlin's erbia was also examined by Soret in 1878 and by Cleve in 1879; the new base then isolated, Soret's X or Cleve's holmia, was split by Lecoq de Boisbaudran in 1886 into a true holmia and a new

oxide dysprosia . The same erbia also yielded another base, thulia, to Cleve, in 1899, in addition to true erbia . The original erbia of Mosander was confirmed by M . A . Delafontaine in 1878 and renamed terbia; this base was split by Marignac in 1886 into gadolinia and true terbia . These relations are schematically shown below; the true earths are in italics, mixtures in Roman . Ceria Ceria Lanthana 1 Lanthana Didymia Samaria Samaria Europia Yttria 1 I I Yttria Erbia Terbia (Mosander) (Mosander) Terbia Erbia (Delafontaine) (Berlin) Terbia Gaholinia Ytterbia Thulia Soret's X Erbia Holmia Scandia Ytterbia Holmia Dysprosia Neoytterbia Lutecia Methods of Separation.—The small proportions in which the rare earths occur in the mineral kingdom and the general inter-mixture of several of them renders their efficient separation a matter of much difficulty, which is increased by their striking chemical resemblances . While it is impossible to treat the separations in detail, a general indication of the procedure may be given . The first step is to separate the rare earths from the other components of the mineral . For this purpose the mineral is evaporated with sulphuric or hydrochloric acid, or fused with potassium bisulphate, and the residue extracted with water . The solution of chlorides or sulphates thus obtained is treated with sulphuretted hydrogen, to remove copper, bismuth and molybdenum, and the filtrate, after the ferrous iron has been oxidized with chlorine, is precipitated with oxalic acid .

The oxalates (and also thorium oxalate) may be converted into oxides by

direct heating, into nitrates by dissolving in nitric acid, or into hydroxides by boiling with potash solution . The thorium may be removed by treating the nitrate solution with hydrogen peroxide, and warming, whereupon it separates as thorium peroxide . The next step consists in neutralizing the nitric acid solution and then saturating with potassium sulphate . Double salts of the general formula R2(SO4)3.3K2SO4 are formed, of which those of the cerium group are practically insoluble, of the terbium group soluble, and of the ytterbium group verysoluble . The sulphates thus obtained may be reconverted into oxalates or oxides and the saturation with potassium sulphate repeated . To separate the individual metals many different methods have been proposed; these, however, depend on two principles, one, on the different basicities of the metals, the other, on the different solubilities of their salts . Bahr and Bunsen worked out a process of the first type, which utilized the fractional decomposition of the nitrates into oxides on heating . The mixed oxalates are converted into nitrates, which are then mixed with an alkali nitrate to lower the melting-point, and the mixture fused . The nitrates decompose in order of the basicities of the metals, and after a short fusion the residue is extracted with boiling water, and the basic salt which separates when the solution is cooled is filtered off . This'contains the most negative metal; and the filtrate, after evaporation and a repetition of the fusion and extraction, may be caused to yield the other oxides . A second method, based on the same principle, consists in the fractional precipitation by some base, such as ammonia, soda, potash, aniline, &c . The neutral nitrates are dissolved in water, and the base added in such a quantity to precipitate the oxides only partially and very slowly .

Obviously the first

deposit contains the least basic oxide, which by re-solution as nitrate and re-precipitation yields a purer product . To the filtrate from the first precipitate more of the base is added, and the second less basic oxide is thrown down . By repeating the process all the bases can be obtained more or less pure . Many processes depending upon the different solubilities of certain salts have been devised . They consist in forming the desired salt and fractionally crystallizing . The mother liquor is concentrated and crystallized, the crystals being added to the filtrate from a re-crystallization of the first deposit . These operations are repeated after the manner shown in the following scheme; the letter C denotes crystals, the M.L mother liquor, whilst a bracket means mixing before re-crystallization .