Thulium, Tm = 168·5

The thulia isolated in 1879 was described by Cleve[384] as pale rose in colour; in the following year, having obtained it in larger quantity, he found that it was white, and dissolved in acids to form colourless solutions which showed absorption bands in the red and blue. The spectra of the thulium compounds prepared by Cleve were examined by Thalèn,[385] who concluded that a new element was certainly present, though it had not been freed from ytterbium and erbium. Incidental observations on the new oxide were made by various investigators, but no extensive researches were carried out upon it until 1911, when James[386] published an account of the separation and purification by the bromate method, stating that after some 15,000 operations, his products remained unaltered; he gives, however, no spectroscopic determinations, though part of his material, spectroscopically examined by Sir William Crookes, was described as ‘Very good thulium, with a trace of ytterbium.’ In the same year Auer von Welsbach[387] published an account of a spectroscopic investigation, as a result of which he concludes that thulium is a mixture of at least three elements, of which the second, Tm II, agrees fairly well in properties, so far as the two accounts allow of comparison, with the thulium of James.

Thulia is described by James as a dense white powder, with a greenish tinge, which ‘emits a carmine coloured glow, when carefully made to incandesce.’ The salts have a greenish tint, very susceptible to traces of erbium; addition of erbium compounds turn the solution first yellowish-green, then colourless, and finally pink. von Welsbach describes Thulium II as forming an almost white sesquioxide, which, when heated in the flame, gives a purplish light quickly succeeded by a splendid characteristic glow; the salts are pale yellowish-green by daylight, emerald-green by artificial light, the colour being almost complementary to that of erbium salts. In solution, salts of Tm II give the bands at 685 and 464 ascribed by James and other workers to thulium.

Until further researches on these interesting results are published, the elementary nature of thulium cannot be considered definitely settled; it appears probable, however, that homogeneous salts of a definite element were obtained by James. The following salts are described by James (loc. cit.).

The chloride, TmCl₃,7H₂O, separates at ordinary temperatures from the concentrated solution of the oxide in hydrochloric acid as greenish crystals, very soluble in alcohol and water. The bromate, Tm(BrO₃)₃,9H₂O, forms pale bluish-green hexagonal prisms, isomorphous with the analogous salts of the group. The sulphate and nitrate separate as the octohydrates. The precipitated oxalate has the formula Tm₂(C₂O₄)₃,6H₂O, and is soluble in excess of alkali oxalate. The acetylacetone derivative was prepared by dissolving the precipitated and well-washed hydroxide in alcoholic acetylacetone; it recrystallises from absolute (?) alcohol as the dihydrate, Tm₂(C₅H₇O₂)₆,2H₂O. The phenoxyacetate, Tm₂(C₆H₅·O·CH₂·COO)₆,6H₂O, was obtained in a similar manner by addition of the hydroxide to a solution of phenoxyacetic acid in dilute alcohol.

Atomic Weight.

—Cleve gave the value 170·7 for this constant, but his material was very impure. In a footnote to a paper published in 1907, Urbain[388] pointed out that the value could not be above 168·5. Analyses of the salts prepared by James agree fairly well with the theoretical values calculated on this basis, but a systematic determination with pure material has not yet been made. The International Committee (1912) have adopted the value 168·5.

The first indication of the complexity of Marignac’s Ytterbium was furnished on spectroscopic grounds by Auer von Welsbach in 1905;[390] he showed that a separation could be effected by the fractional crystallisation of the ammonium double oxalates from concentrated ammonium oxalate. Three years later[390] he published a full account of his method, gave atomic weight determinations, and mapped the spectra of the two new elements. In 1907, Urbain[391] independently effected a separation by the fractional crystallisation of the nitrates from nitric acid, and proposed the names Lutecium (from the old name for Paris) and Neoytterbium for the elements.

The two new elements resemble one another so closely in chemical properties that the account given by Astrid Cleve in 1902[392] of the compounds of the old ytterbium applies in practically every detail to the new elements. The oxides are white, and yield colourless salts, showing in solution no absorption bands in the visible region.

The oxides, R₂O₃, though perfectly white, are coloured yellow or brown by the faintest traces of thulium. They are attacked by acids only slowly in the cold, but dissolve readily on warming; lutecia is slightly the less strongly basic. The chlorides crystallise with six molecules of water, and are extremely soluble and deliquescent; when heated in a stream of hydrogen chloride, they form oxychlorides of the type ROCl. The platinocyanides crystallise with 18 molecules of water, and have the characteristic appearance of the analogous compounds of the yttrium elements. The sulphates crystallise at all temperatures as the normal octohydrates, and are moderately easily soluble in water; conductivity measurements show that they are partially hydrolysed in solution. The nitrates crystallise from concentrated aqueous or nitric acid solutions as the tetrahydrates; by evaporation of the aqueous solutions over sulphuric acid, the trihydrates are obtained. These compounds are anomalous among the rare earth nitrates, by reason of their low water content. The neutral carbonates are thrown down by ammonium carbonate as the tetrahydrates; if a stream of carbon dioxide be led into aqueous suspension of the hydroxides, basic carbonates of the formula R(OH)CO₃,H₂O, are obtained. The oxalates are precipitated as the decahydrates; they are readily soluble in excess of alkali oxalate.

Many other salts of the old ytterbium have been prepared.

Atomic Weights.

—The values determined by Urbain (loc. cit.) for the fractions obtained by the nitrate method gave the number 170·1 for the least soluble fraction free from terbium, and 173·4 for the most soluble fraction. Auer von Welsbach (loc. cit.) obtained the values 172·9 and 174·2 for the least soluble and most soluble fractions from the double oxalate crystallisation respectively. More recently[393] he has determined these constants with highly purified material, employing a modified method. The weighed anhydrous sulphates are transformed into the oxalates, which are then ignited to the oxides. He obtained the values Yb = 173·00, Lu = 175·00.

The values adopted by the International Committee are Yb = 172·0 and Lu = 174·0.

Spectra.

—The spark spectra are of more use in distinguishing the two elements than the arc spectra. The spark spectrum of the old ytterbium was mapped by Exner and Haschek,[394] and of the two compounds by both discoverers (loc. cit.). See also Eder and Valenta.[395]

The arc spectra have been mapped by Eder and Valenta (loc. cit.) and by Exner and Haschek; the latter authors give as the most intense lines the following:

Celtium

The separation of Marignac’s ytterbium into the two elements described above was accomplished by Urbain with the yttria earths extracted from xenotime. In carrying out the same process with the ytterbia earths from gadolinite, that author[396] obtained from the mother-liquor an earth for which the coefficient of magnetisation was found to be 4·1 × 10⁻⁶; lutecia has a coefficient three to four times as great. A spectroscopic examination revealed the presence of lines which did not correspond with those of any known body, and Urbain considered that a new element, for which he proposed the name Celtium, with the symbol Ct, must be present. Lutecia from xenotime shows no trace of the new element.

Spectroscopic evidence for the existence of a third ytterbium element had previously been brought forward by Auer von Welsbach[397] and also by Exner and Haschek.[398]

The new element appears to be intermediate between lutecium and scandium, and therefore may be expected to have a higher atomic weight than the former element. Its chloride is more volatile than that of lutecium, less volatile than that of scandium; its hydroxide is more feebly basic than that of lutecium, but more strongly basic than that of scandium.

Urbain (loc. cit.) gives the following as the principal lines in the spectrum; strong lines are denoted by a single, very strong by a double, asterisk:

Yttrium, Yt = 89·0

Since the separation of yttria proper from the old yttria earths by Mosander, in 1842, the individuality of yttrium has been well established. The yttria of the workers of the sixties and seventies, to judge from the atomic weight determinations, must have been very impure, but no doubts were raised as to its homogeneity. By examination of the cathode luminescence spectra, Crookes[399] concluded that the oxide was of a complex nature; Lecoq de Boisbaudran, however, showed that the phenomena observed by Crookes were due to traces of impurity in his material, a conclusion confirmed by the work of Baur and Marc.[400]

The oxide is the most strongly basic of all the yttria earths; in the basicity methods of separation, therefore, it collects in the end fractions, and is easily separated from the erbia and ytterbia earths by the nitrate fusion and similar processes. The terbia earths, however, which are comparable to it in basic strength, cannot be easily separated by such methods; processes of fractional crystallisation are very convenient in this case, since yttrium falls, with regard to the solubility of its simple salts, among the erbium group—between holmium and erbium generally—which is easily separated from the less soluble terbium elements. The separation of yttrium, therefore, affords an example of the combination of methods of both kinds.

The methods for the separation and purification of yttrium have recently been exhaustively examined by Meyer and Wuorinen.[401] They consider the chromate method suitable only if the terbium elements have already been removed. The ethylsulphate method is said to be tedious, whilst the ferrocyanide method indeed effects very rapid concentration, but with great loss. For purposes of concentration they find the most suitable method in the fractional hydrolysis of the phthalates; these salts are soluble in cold water, but hydrolyse when the solution is warmed, the most positive elements remaining of course longest in solution. For the final purification, they recommend fractional precipitation of the iodate from nitric acid solution; yttrium iodate being more soluble than the iodates of the erbium and ytterbium group, the latter collect in the first precipitates.

Pure yttria is quite white, and gives rise to colourless salts, which in solution show no absorption spectrum in the visible region. A very large number of yttrium compounds have been prepared, of which sufficiently detailed accounts have been given in the general description of rare earth compounds. For an exhaustive treatment, the reader is referred to Abegg’s ‘Handbuch.’

The metal has probably not been obtained in the pure state; impure yttrium has been obtained by Winkler[402] by the action of magnesium on the oxide, and by Cleve[403] by the action of sodium on a mixture of the chloride with common salt, and by electrolysis of the mixture of fused chlorides. It is described as a greyish metal, resembling iron in appearance; it oxidises in the air and readily decomposes boiling water. The hydroxide is thrown down as a gelatinous precipitate by alkalies; ammonia throws down basic salts, but in presence of hydrogen peroxide an hydrated peroxide is obtained. The oxide absorbs carbon dioxide from the air, and liberates ammonia from ammonium salts.

The anhydrous chloride has been prepared by many authors; it melts at a relatively low temperature, 680°, and is the most easily volatilised of all the rare earth chlorides. After fusion, it forms a mass of brilliant white lamellæ.[404] It is characterised by the ease with which it dissolves in pyridine. From aqueous solution it separates as the hexahydrate, YtCl₃,6H₂O, which melts at 160°. The bromide separates from solution as the enneahydrate, YtBr₃,9H₂O; the bromate[405] also separates with 9 molecules of water of crystallisation.

The nitrate cannot be obtained anhydrous; the normal hydrate, Yt(NO₃)₃,6H₂O, loses 3 molecules of water at 100°, but further heating converts it into basic salts. A basic nitrate, 3Yt₂O₃,4N₂O₅,20H₂O, is described by James and Pratt[406] as stable at ordinary temperatures, and in contact with solutions of the normal nitrate. The sulphate octohydrate is isomorphous with analogous compounds of the rare earth elements, and with the selenate, Yt₂(SeO₄)₃,8H₂O; the latter compound can also form an enneahydrate. The phosphate, YtPO₄, occurs in nature in the mineral xenotime, and has been obtained in the laboratory in the crystalline form; many other phosphates have been prepared. The platinocyanide, Yt₂[Pt(CN)₄]₃,21H₂O, has the characteristic red colour with greenish-blue fluorescence.

Many organic yttrium salts have been prepared by James and Pratt[407] and by Tanatar and Voljanski.[408]

Atomic Weight.

—The numbers obtained by the investigators who have determined this constant vary to such an extent that considerable uncertainty attaches to the value, 89·0, at present accepted by the International Committee. The determinations carried out prior to 1870 gave such diverse results that they are of little use in fixing the constant; since that date, all the investigations, with the exception of the most recent, have given values below 90, the sulphate method being generally employed.

Cleve and Höglund,[409] in 1883, carried out six determinations by the synthetic method; their results were concordant, and gave the mean value 89·57. Brauner considers this result if anything too low, as traces of undecomposed acid sulphate may have been present in the anhydrous sulphate. The same method was employed again by Cleve in 1884;[410] the mean of twelve very concordant results gave the number 89·11.

Much stress is laid by Brauner[411] on an unpublished determination of Marignac, carried out with material entirely free from terbia, which gave the value 88·88. H. C. Jones in 1895[412] carried out two series of determinations with material purified by Rowland’s method, i.e. precipitation with potassium ferrocyanide;[413] the results in both series were very concordant, the synthetic method giving the value 88·95, the analytical method the value 88·97. This work has been taken by the International Committee as the basis for the accepted value. According to Brauner, the ferrocyanide method does not give perfectly pure material.[414]

Egan and Balke[415] have recently found the ratio Yt₂O₃ : 2YtCl₃ to be very suitable as a basis for atomic weight determinations; the oxide is converted into the anhydrous chloride in a quartz flask. In a preliminary experiment, they obtain as a mean of three consistent determinations the provisional value 90·12; the yttria employed was considered to contain not more than one-half per cent. of erbia.

Recent work by Meyer and his co-workers[416] indicates that the accepted value is too high. Preliminary work with the synthetic sulphate method gave the values (corrected) 88·71 and 88·73; the mean value of six analytical sulphate determinations, made on material carefully purified by the iodate method, was 88·75, the extreme values being 88·71 and 88·76. They consider that the true atomic weight is 88·7, the value of the second decimal figure being a little uncertain.

Scandium, Sc = 44·1

The scandia obtained by Nilson in 1879 was isolated from the minerals gadolinite and euxenite; it consisted very largely of ytterbia, as shown by spectrum examination[420] and by atomic weight determinations, which gave the value 90. In the same year[421] Cleve prepared the oxide in a much purer state, using as his source the minerals gadolinite and keilhauite; he described several salts, carried out atomic weight determinations by the analytical and synthetic sulphate methods, and showed that scandium corresponds with the Eka-boron of which the existence was predicted by Mendelejeff in 1871.[422] Starting from a large quantity of euxenite, Nilson[423] in the following year prepared several grams of approximately pure scandia, which contained only traces of ytterbium.

The investigation of scandium, which occurs only in extremely small quantities in the minerals employed by Nilson and Cleve, and was therefore believed to be exceedingly rare, was not continued until 1908, when Sir William Crookes[424] made a systematic investigation of a large number of minerals in order to find a convenient source of the element. He showed that scandium is present in many rare earth minerals, and selected as the most suitable for the extraction of the element a complex mineral named Wiikite, some specimens of which he found to contain over 1 per cent. of scandia (see p. 70). The mineral was decomposed by fusion with potassium hydrogen sulphate, and scandia extracted from the rare earths by the nitrate fusion. The separation effected on these lines was very thorough, Crookes considering a specimen of scandia unsatisfactory if it showed any trace of the dominant ytterbium line, 3694·344, on an over-exposed plate, or if it gave an atomic weight for the element higher than 44·1.

A systematic investigation of the common rocks and minerals for scandium was carried out by Eberhard in 1908, as a result of which processes for the extraction of the oxide from wolframite were worked out by R. J. Meyer (see pp. 3, 131). Wolframite is a tungstate of iron and manganese, containing, in addition to other oxides, small quantities of the rare earths, of which considerable proportions are found to be scandia. The mineral is fused with soda in the usual way, and the rare earths concentrated by the oxalate method. Scandium is precipitated as the fluoride by addition of sodium silicofluoride to the boiling acid solution, and purified by precipitation as the double ammonium tartrate.[425]

Whilst the researches of Crookes and Eberhard have shown how widely distributed the element really is, the minerals which they found richest in scandium still contained extremely small quantities of the oxide. The discovery of the mineral Thortveitite (see p. 44), which contains about 37 per cent. of scandia, is therefore of the greatest scientific interest, and will doubtless allow of a very searching examination of the properties of this interesting element.

Whilst the low atomic weights of scandium and yttrium place them, to some extent, apart from the other rare earth elements, the latter element at least is so closely allied in properties to the other members of the group that yttria is one of the typical oxides of the family. Scandium and its compounds, however, present many peculiarities of behaviour when compared with the typical members, on the grounds of which Urbain[426] has contended that scandia should not be classed among the rare earths at all. Whilst this contention is perhaps rather extreme, especially in view of the fact that in nature scandia always occurs with other yttria oxides, it must be admitted that in many respects the element is anomalous. The oxide is the weakest base of the whole group, yet the oxalate is comparatively readily soluble in mineral acids (compare p. 132), and the potassium double sulphate is almost insoluble in potassium sulphate. The sulphate is altogether exceptional in that it is very easily soluble in water, and crystallises out with 6 molecules of water of crystallisation. The fluoride and the carbonate both dissolve readily in excess of precipitant, whilst sodium thiosulphate precipitates a basic salt from neutral solutions.

Meyer has pointed out the close resemblance between beryllium and scandium. The oxide and salts are colourless; the latter have a peculiar sweet astringent taste, and readily yield basic salts.

The hydroxide, Sc(OH)₃, is thrown down by alkalies as a bulky white gelatinous mass; the oxide is a white powder, less readily soluble in dilute acids than most of the rare earths. The fluoride is important on account of its insolubility in mineral acids, which exceeds that of all the other rare earth fluorides, and approaches that of thorium. It is thrown down from neutral or acid solutions by addition of hydrofluoric acid or a soluble fluoride; if the solution be boiled, a soluble silicofluoride will also precipitate scandium fluoride, though no precipitate is obtained in the cold. This behaviour is due to the ease with which the silicofluoride is hydrolysed at high temperatures, according to the equation:

Sc₂(SiF₆)₃ + 6H₂O = 2ScF₃ + 3SiO₂ + 6H₂F₂

and is of great value in separating scandium from the other earths. The fluoride is extremely resistant to acids, being completely decomposed only by fused bisulphate. In the absence of acids, the freshly precipitated fluoride dissolves in excess of concentrated alkali fluoride, forming double salts; in this behaviour, scandium resembles zirconium, but differs from thorium and the cerium and yttrium elements.

The chloride separates from solution at ordinary temperatures as the dodecahydrate, Sc₂Cl₆,12H₂O, which loses 9 molecules of water when kept for six hours at 100°. The trihydrate Sc₂Cl₆,3H₂O, is converted into scandia at a red heat, with the loss of 6 molecules of hydrogen chloride. The iodate, Sc(IO₃)₃,18H₂O, is obtained as an almost insoluble white crystalline powder by addition of ammonium iodate to a salt in solution; hydrates with 15, 13, and 10 molecules of water are known, and at 250° the anhydrous compound is obtained. It resembles the iodates of the cerium and yttrium group in being soluble in strong nitric acid, but the separation of thoria and scandia by this method is tedious and unsatisfactory.[427]

The platinocyanide, Sc₂[Pt(CN)₄]₃,21H₂O, was obtained by Crookes[428] by double decomposition of the sulphate with barium platinocyanide, in crimson monoclinic prisms, with a green fluorescence. It dissolves in water to a colourless solution. Orlov[429] shows that it can occur also in a second form, stable at higher temperatures; this is yellow, with a blue fluorescence and crystallises with 18 molecules of water. The two modifications resemble respectively the platinocyanides of the yttrium and of the cerium elements; in this respect, therefore, scandium occupies an intermediate position between the two groups.