7/28/2019 Chem rev VOL1 143_185

1/43

TWENTY-FIVE YEARS OF RESEARCH ON T H EYTTRIUM EARTHS1G. URBAIN

Pr o fes s o r o f C h e mi s t r y in the F a c u l t y o j Sciences at the UniijersztU of Paris 'The present article is a brief account of the investigations whichI have carried on for twenty-five years in the field of the yttrium

earths.These' investigations have been published in fragments invarious scientific periodicals printed in the French language. Icannot better acknowledge t o Yale University the great honorwhich it has conferred upon me than to write out for it an out-line of my work on the yttrium earths.No other natural group of chemical elements contains somany members as that of the rare earths. It comprises in itself

one-sixth of the known elements.If our knowledge of the rare earths is still limited, it is not be-cause the investigators who have engaged in the study of themhave been few. They have in fact been many, and among themare distinguished scientists, such as Berzelius, Mosander, Bun-sen, Cleve, de Marignac, Nilson, Auer vonwelsbach, Lecoq deBoisbaudran, and Crookes.One has only to run over the history of the rare earths to appre-ciate the considerable volume of work tha t has been done in th atfield, by more than tw o hundred investigators of various nation-alities. In spite of this multitude of workers, the problem isnot yet completely solved.

1 An essay prepared in connection with the dedication of the Sterling Chem-2 Translated from the French by Prof. and hlrs. P. E. Browning; approvedistry Laboratory.by Professor Urbain.

143C H E M I C A L REVIEWS, VOL. I , NO. 2

7/28/2019 Chem rev VOL1 143_185

2/43

144 G. URBAINFor a long time the scarcity of original material was an obsta-cle. Twenty-five years ago the minerals which contain the rare

earths were very expensive. It was impossible for anyone toobtain large quantities of them. Practically no other depositswere known than those of Norway, and the working of these,which occurred in rocky formations, was always difficult and oftenuncertain.Eut following the industrial development of thorium whoseoxide forms 99 per cent of the material used for mantles, in in-candescent gas lighting, further search was made, and importantdeposits of the rare earths were found: monazite sands in Amer-ica, gold- and diamond-bearing sands in South Africa and Aus-tralia, etc. During the past fifteen years it has becohe easyto obtain rare earths, a condition which should encourage prog-ress in this field of study. As a matter of fact, the lower mem-bers of the series, those most abundant and most easily prepared,have been the ones almost exclusively studied. The investiga-tors have for the most part colonized a territory already exploredand conquered.And despite the attraction of the veritable unknown, thereare few who have taken up this problem from its most interestingside, namely, the investigation of the elements which compose theyttrium earth group, the higher members in the rare earth seriesS3To be sure, the pursuit of this work meant going farther thanscientists such as Marignac, Cleve, Lecoq de Boisbaudran andDemarqay had gone. Merely to travel afresh the path whichthey had explored involved a long series of laborious separa-tions.From the experimental point of view, the study of the yttriumearths consists of two parts, of equal importance: (1)a chemicalpart , imposed by the necessity of separating the members of thegroup from one another; (2) a physical-chemical part, imposedby the necessity of characterizing the various elements of the

As to discoveries, they were most uncertain.

8 Since the death of Demarpay, in 1901, a few investigators only have taken upor carried on the study of the yttrium earths: Auer von Welsbach in Austria;Hoffmann and R. J . Meyer in Germany; Holmberg in Sweden; chiefly James inAmerica; and the author of this article in France.

7/28/2019 Chem rev VOL1 143_185

3/43

BESESRCH ON YTTRIUM EARTHS 145group by specific, measurable properties. These two parts arecomplementary, one to the other.However, that is only method of precedure, or techniqud.With materials such as these, technique is valueless unless it isemployed with adequate method and logic. It is here, doubt-less, that the principal difficulty of the problem lies.In order that the method which I used may be understood,I will begin by saying a few words regarding technique.

Anal ytica l procedure: The yttrium earths cannot be separatedas one separates the common elements. The latter belong tovarious families. They differ in a marked degree from oneanother and it is very easy to separate them by utilizing thedifferent chemical properties which facilitate their identification,The yttrium earths are members of one and the same family.Consequently their chemical properties are very similar. Formore than fifty years the group was regarded as a single elern-ent, for the different members are invariably associated in theirminerals, and always in practically the same relative proportions.hot only are their chemical properties almost identical, buttheir salts are isomorphous. Those who, knowing nothing of thismatter, have sought new methods of separation, have achievedonly new chemical methods of fractionation. We know thatfractionation is a process of separation which depends not uponany essential difference in properties, but upon differences inone property common to the constituents of a mixture. Upto the present time the separation of the yt trium earths from oneanother has been accomplished only by fractionation.h ractionation is always a laborious operation, even when twoelements only are concerned. Bu t when it is applied to theyttrium group, in which the existence of eleven members has

been demonstrated, the fractionation becomes extremely tedious.An illustration of this may be found in a fractionation which Ihave been carrying on for twelve years, in which the earthshave been treated systematically every day, by improvedmethods, and in which, despite this great amount of effort,certain separations are far from complete.

7/28/2019 Chem rev VOL1 143_185

4/43

7/28/2019 Chem rev VOL1 143_185

5/43

RESEARCH O N YTTRIUM EARTHS 147treatments the measure of the coefficients of magnetism. Thismethod, as we shall see later, is far superior to the atomic weightmethod, and especially to the solubility method recommendedby Marignac.W o r k i n g method recommended in these investigations. As inall experimental science, the progress of the knowledge of therare earths has followed the progress of its technique. Th eearly investigators were guided at first by observation of colora-tion. The group is in fact composed of members, some of whichform colorless salts, others salts of various colors. The studyof absorption spectra has made i t possible to express by num-bers the sensation due t o coloration; and by this scientific pro-cedure different compounds may be distinguished in mixtures,the colors of which differ very little.Absorption spectra are the easiest to observe, and this methodhas been recommended by the majority of investigators. Otherprocedures were not employed until later, but a few studentshave made use of them.Additional spectroscopic methods have increased the chanceof discoveries. But this situation has given rise to new difiicul-ties, which I will try briefly to indicate.Let us suppose that three independent investigators frac-tionate the same unknown earths. Suppose also that each ofthem takes as guide a different kind of spectrum; the first, forexample, the absorption, the second the spark, and the thirdthe phosphorescence spectrum.Experience shows us that the following fact results. Thefirst will distinguish a certain number, A , of absorption spectra;the second a number, B, of spark spectra; and the third a num-ber, C, of phosphorescence spectra. These investigators havediscovered, in all, A +B + C new spectra. Shall we, then,conclude that they have discovered A +B +C new substances?Certainly not, since one element may give a t the same time thesethree varieties of spectra. One element may thus have been dis-covered three times in succession.

Let us suppose th at A is the largest of these numbers. Therecannot have been less than A elements discovered. Bu t there

7/28/2019 Chem rev VOL1 143_185

6/43

148 G. URBAINmay have been more, for not all of the rare earths show necessarilythese three kinds of spectra in that par t of the spectroscopicfield which is easily observed; and moreover the three investi-gators may not have carried the separations sufficiently far toavoid the danger of confusion or of omission.

Unless these three investigators share with one another thematerials which they have treated, no attempt can be made toattribute to a restricted number of elements the different spectrawhich they have observed.And finally, even if these investigators should meet to com-pare resulta, they would find themselves confronted by so manyuncertainties, since they had not isolated pure substances, thatthe best they could do would be to set up a probable theory of thecoi:qmon origin of certain of their spectra only.6The study of the rare earths was in the sta te which I havejust outlined schematically when I began my investigations.The subject was in certain particulars well advanced, but onthe whole it was so confused that most chemists considered ithopeless. After a period of prep-aration,6 during which I familiarized myself with the processesrelated bo the subject, I undertook long, methodical treatments,taking pains t o observe, throughout, all the characteristics ofevery kind which my predecessors had described; to them Iadded those that I myself discovered later. So far as the quan-tity of material a t my disposal permitted, I forced myself to pur-sue the treatments t o the point where they were unnecessary,zt point which I determined by the invariability of the proper-ties observed.Kot only have I taken pains to prolong the treatments in cer-tain cases beyond the point necessary t o obtain pure products,but I have varied the treatments; for certain mixtures, undecom-

I began by making a general review.

The new method of X-ray spectra leaves no room for such uncertainty, but itis applicable t o the rare earths only where the separations have been carriedvery far, so that the observations can be made with much simplified mixtures;for elemelits having nearly the same atonic weight have lines which exactlyooincide.Doctorat e thesis, 1898, Recherches sur la SBparation des Terres Rares.

7/28/2019 Chem rev VOL1 143_185

7/43

RESEARCH ON YTTRIUM EARTHS 149posable by a method based on one principle (differences in solu-bility, for example), may be decomposed by a method based onanother (as differences in the basicity of th e oxides).When the different processes produced no further results, andthe properties remained invnriable, whatever the principle ofth e method used, I knew beyond all doubt what group of proper-ties belonged to the isolated element. Furthermore, I always,then pursued the study of the mixtures intermediate betweenpure elements, in order to bring to light, if possible, some prop-erty still unknown; and to this end I reduced these intermediatefractions to a minimum by means of long-continued treatment.

In this way I have been xble to bring order, which I hopemay be definitive, into the first members of the ytt rium series.I have reduced to four the nurnber of elements between sama-rium and holmium, and I have shown that about twenty supposedelements whose separate existence had been considered possibleor even probable, are identical with europium, gadolinium, ter-bium or dysprosium. Previous to this research, none of theseelements had been isolated even reasonably pure. They wereso ill-defined that only one of them, gadolinium, had a place inthe international table of atomic weights, which, to some extent,assures chemical elements an ofccial existence. And never-theless, the existence of terbium had been announced by Mosandersixty years earlier.I set myself the task of submitting Marignacs ytterbium to

* analysis of this sort. It enabled me to discover two new elements,lutecium and celtium. I had given the name of neoytterbiumto the principal constituent of th e original ytterbium, in orderto leave to the illustrious Narignac, in the future, the credit ofhis fundamental discovery.SEPARATION O F TJ3E RARE EARTHS

The first problem that presents itself is the treatment of theminerals. Of these there is great variety. Some are phosphates,others are tantalo-niobates or silicates.

7/28/2019 Chem rev VOL1 143_185

8/43

150 G. URBAINIn a work surveying the whole field it was necessary to examinevery different minerals. It was known about 1895 tha t the dif-ferent rare earths are always found associated in nature, but itwas not known what are the relative proportions of each ofthese elements in each mineralogical specimen. It could onlybe hoped that these proportions differed enough to make it ad-visable to treat a particular mineral in order t o accomplish thespecial extraction of one of these substances. To this end Itreated, in turn, monazite sands, aeschinite, cerite, gadolinite,thorite, and later xenotime. These various minerals are verydifferent from one another, and each requires special treatment.We know tha t the group of rare earths is divided into twosub-groups : the cerium earths, which comprise the first membersaf the series, namely lanthanum, cerium, praseodymium, neo-dymium and samarium; and the yttrium earths, which comprisethe other members.The earths extracted from the minerals enumerated aboveare hardly distinguishable except by their content in elementsof each of these sub-groups. In general ways, the cerium earthsare more abundant than the yttrium earths, but each sub-grouphas almost always the same composition, so that the relativerarity of each of the rare earths varies only within narrow limits.That is certainly a notable fact. It probably cannot be ex-plained without resort to hypothesis. While I do not wish toenter upon that path, which I have always avoided as much aspossible, I will simply observe that all radio-active mineralscontain rare earths, though the converse is not necessarily true.But in some cases in which the rare earths are not accompaniedby radio-active substances the presence of helium, which is oneof the last products of atomic disintegration may be detected.

Debiernes observation that certain fluorides contain heliumled me to look fo r the rare earths in the fluorides (1). I thenproved that the phosphorescence of the fluorides must be attrib-uted in part to the presence of the rare earths, a fact which hadnever been suspected. These phenomena of phosphorescencehad been att ributed to mysterious causes or to unknown elements.

7/28/2019 Chem rev VOL1 143_185

9/43

RESEARCH ON YTTRIUM EARTHS 151Contribution to the study of t hor ium (2 ) . While I was studying

the treatment of the minerals which I have named, I experiencedthe difficulties which one always encounters not only in separat-ing t he rare earths from one another but also in separating themas a whole from the elements associated with them in their min-eral sources. Koteworthy among these associates was thorium,a very valuable element at that time (1 kilo of the oxide in theuntreated mineral was worth 1000 francs; extracted and purifiedit was worth, in 1914, 30 francs), and one whose separation fromthe true rare earths had been worked out only very imperfectly.I found in acetylacetone an excellent reagent for thorium. Theacetylacetonate of thorium is soluble in chloroform, which takesit from its aqueous solutions; moreover, this compound may bedistilled in vacuo. The corresponding salts of the rare earthsdo not act in the same way; they are difficultly soluble in chloro-form, and they decompose in vacuo under the influence of heat (3).The analysis of pure thorium sulphate, obtained by this methodenabled me moreover to determine the atomic weight of the ele-ment with an accuracy of the order of one part in a thousand (4).The rare earths, in separat-ing from one another by their differences in solubility alwaysappear in the same order, whatever the character of the salt sub-mitted to fractional crystallization. I have repeatedly empha.-sized this remarkable fact, and I have found in it a natural lawof seriation ( 5 ) .It may happen, however, that the series becomes reversedin some measure in the process of fractional crystallization ofcertain salts, such as the ethylsulphates or the simple nitrates ofthe type M(S03)35.Hz0.I can find nothing better with whichto compare this anomaly than the phenomenon of the maximumdensity of water, or the anomalous dispersion in which the spec-trum reverses. Could this law be extended to apply to othergroups of elements? The work of Locke upon the alums andthe sulphates of the magnesium series makes i t seem probable.Rut in groups which contain only a small number of ele,ments,every anomaly will mask the law; whereas in the case of the rareearths, which are very numerous, the anomalies do not weakenit in any way.

Law of seriation of the rare earths.

I

7/28/2019 Chem rev VOL1 143_185

10/43

152 G . U R B A I NL i m i t s of fract ionat ion. When the at tempt is made to separatetwo closely associated rare earths by any method whatever, itis always found that the progress of the separation is relativelyrapid a t first, but tha t the last traces of impurities are very dif-ficult to remove. Sometimes they cannot be completely removed;there is a limit to the separation.Poor rr-ethods are those which adxit a lirnit correspondingto very incor;plete separations. If the worker takes as his onlyguides insufficient characteristics: if he confines himself, for in-stance, to the determination of atomic weights or of solubilities,the constancy of the numbers which correspond to a limit may

deceive him as to the value of the results which he has obtained.Experirr,entally, the h i t of a fractionation, by crystallizationof isomorrhous salts, is in general a pure product, but it may byexception be formed of an indecomposable mixture. The mother-liquors have then the saxe composition as the crystals separat-ing from the solvent. This phenomenon is interpreted as likethat of two liquids which distil at a fixed temperature under Rgiven pressure, although the boiling-points of the pure substancesare quite different.A new method of separnt ion of the rclre earths: fractionation oft h e ethylsulphates ( 6 ) . These studies required a real period ofapprenticeship, during which I had to beconie familiar withthe compounds, with the methods of measurement, and alsowith the processes of separation which had been suggested upto that time, all of which I tried (7 ) .A good xany attempts were made to find some method ofseparation of the yttrium earths which should be better than theEethods of my predecessors, for it was certain that this branchof chemistry could make no real progress until the earths couldbe separated from one another.After four years of research I had succeeded in isolatingonly yttrium practically pure, of all the numerous elements whichcoapose the yttrium series; but I had found, in the crystalliza-tion of the ethylsulphates, a process of fractionation of the earthsof this group which was to be at least as fruitful as Auers methodof the crystallization of the double ainmonium nitrates had

7/28/2019 Chem rev VOL1 143_185

11/43

RESEARCH ON YTTRIUM EARTHB 153been for the earths of th e cerium group, in bringing about theseparation of lanthanum, praseodymium and neodymium.This method of the ethylsulphates is very general, and it ap-plies to the group of ihe yttrium earths as a whole. The firstmembers of the series crystallize first; and the various membersof the group coae out in the normal order of their seriation.Thus there is obtained by fractional crystallization a first, rela-tively quick classification, which pernits one later to employ,lilore niethodically and easily, other processes better adaptedto the particular separations of th e members from their closestassociates. But this procedure has one inconvenient feature; theethereal salts saponify in the mother-liquors, upon applicationof heat or merely in the course of time.

ANALOGIES BETWEEN BISMUTH AND THE RARE EARTHSGse of bismuth as an agent of sepamtion. If the methods offractionation permit t he obtaining of pure isomorphous com-pounds, intermediate fractions cannot be avoided. Such sepa-rations are not and indeed seem incapable of being quantitative.However, in certain cases Lacombe and I were able to trans-

form the approximate method of fractionation into a yuanti-tative procedure, and so to achieve sharp separations.The principle of the method is as follows:Let us suppose that to the mixture of salts to be separated byfractionation there is added an isomorphous salt of an elementwhich is of another group and easy t o separate from the orginalsalts by a quantitative procedure.If the whole is fractionated, the added salt will take a position,determined by its solubility, among the salts of the originalmixture. At a certain point, the progress of the separations maybe such that one of the fractions of the systematic treatmentwill be made up of this foreign salt in a sta te of purity. Thenby the elimination of the fractions above and below the foreignsalt. a clear-cut division will have been made between the saltsof the natural series. If the initial mixture contains only twoconstituents, the separation of these two will be complete. Theforeign salt has acted as agent of separation.

7/28/2019 Chem rev VOL1 143_185

12/43

154 G . U R B A I NTo make use of this method it was necessary to know of anordinary element enough like the rare earths to be able to playthis part of separator.We knew of several instances of isomorphism between therare earths and the alkali earths, but considerations of a purelypractical nature prevented our making the experiment with them,Our theory would perhaps have remained fruitless if Goste Bod-man, of Upsala, had not pointed out, in 1898, the isomorphismof bismuth nitrate and sulphate with the corresponding saltsof the rare earths. After several trials, we succeeded in provingthat bismuth could be used as a separating agent when it was

intercalated between samarium and europium.Demarqay had made an approximate separation of these twoelements by fractional crystallization of the double magnesiumnitrates.We proved first that bismuth nitrate can be combined withthe nitrates of the magnesium series to give, in an acid solution,salts absolutely identical in form and general properties withthe corresponding salts of the rare earths, These first resultsfurnished the subject matter of an article published under thetitle: On a new series of bismuth compounds (8).We described the following salts:

These various salts bore the closest possible resemblance to thesamarium and europium salts of the same type. Their solu-bility was between tha t of the samarium and that of the europiumsalts.We pursued the study of the remarkable analogy betweenbismuth and the elements of the rare earth family and showedthat it goes very far (9). In a general way bismuth, in acidicsolutiop, belongs by its isomorphism to the family of rare earths,and it should always be placed between samarium and europium.

7/28/2019 Chem rev VOL1 143_185

13/43

RESEARCH ON YTTRIUM EARTHS 155Thence we were able i o work out the theory described above,and to obtain a t least one sharp separation in the series of the

rare earths.Experimentation has confirmed our conjectures.A sharp separat ion in the rare ea rth series (10). If a large quan-tity of the double nitrate of magnesium and bismuth is addedt o the mixture of double nitrates of magnesium and the rareearths, and the whole is methodically fractionated, the puredouble nitrate of magnesium and bismuth is obtained in one ofthe fractions after several months of daily crystallization. If,next, the bismuth present in the other fractions is eliminated,by means of its special reactions, a clean separation betweensamarium and europium is established. As may be seen, theproblem resolves itself into the obtaining of a pure salt of bis-muth by means of fractionation. Samarium acts as its impurityin the preceding fractions, and europium in the succeeding. Bis-muth establishes a line of demarcation between the earths ofthe cerium group (members loiver in atomic weight thaneuropium)and those of the yttrium group.A second sharp separation by the use of bismuth as sepsrat ingagent. Although bismuth is always inserted between samariumand europium, I discovered later the way to intercalate i t betweengadolinium and terbium, and thus sharply to separate thesetwo elements. I stated in a previous paragraph (law of seria-tion of the rare earths) that the series may in certain cases sufferreversals. That is what happens when the simple nitrates crys-tallizing with 5 molecules of water are fractionated in nitricacid. The series reverses at gadolinium, whose normal placeis between europium and terbium.Of all the simple nitrates crystallizing with 5 molecules ofwater, that of gadolinium is, indeed, least soluble; so that bythis method of fractionation europium and terbium appear againafter gadolinium.If the nitrate of bismuth is added to the mixture, it crystal-lizes in the same way after that of gadolinium, and it keepsnevertheless its normal position between europium and sama-rium. Bu t if the mixture has been freed from these two earths

7/28/2019 Chem rev VOL1 143_185

14/43

156 G. URBAINby previous treatments, the bismuth intercalates itself betweengadolinium and terbium. This separation is long and laborious,and several years were required to bring i t about.C n n new method f o r the isolation of terbia (11). On the otherhand, Lacombe and I have shown that in the separation of therare earths bismuth may be advantageously employed in anotherway:C n the use of bismuth us agent of sepcration in the rare earthseries (12). The difficulty experienced at times in bringing aboutcrystallization of very soluble salts in syrupy solutions is wellknown. If a certain amount of much less soluble salts of thesame type is added to these syrups, the latter salts readilycrystallize; but, because of their isomorphism, in the process ofcrystallization they carry over in a state of solid solution a cer-tain proportion of the salts which do not crystallize spontane-ously in the original syrups.Now, though the first members of the yttrium group crystallizereadily as double magnesium nitrates, the following memberscrystallize with great difficulty; and beginning with terbium theyare absolutely uncrystallizable. The addition of double ni-trates of bismuth to these syrups made it possible for us to ex-tract the majority of the earths most closely associated withterbium, and in this way to free them of the greater part of theyttrium, which is very abundant in this series and which con-stitutes one of the principal difficulties in the isolation of therarer elements which always accompany it.

SPECIAL METHODS OF SEPARATIONSeparat ion of europ ium f rom gado l in ium. The solubility ofthe double nitrate of magnesium and europium differs very little

from that of the double nitrate of magnesium and bismuth, butis slightly greater.In the course of the fractionation, the mass of bismuth rap-idly drives the gadolinium into the last fractions; and at theother end a series of fractions is obtained consisting almost ex-clusively of bismuth, with no gadolinium but with a trace oreuropium. This element is extremely rare compared with samar-

7/28/2019 Chem rev VOL1 143_185

15/43

RESEARCH ON YTTRIUM EARTHS 157ium and gadolinium. By elimination of bismuth and magnesiumfrom th e fractions, pure europium is obtained at the head of theseries (13) .Separat ion of g a d o l i n i u m f r o m t e r b i u m . The double mag-nesium nitrates of t he earths rich in terbia arelmuch more solublethan magnesium nitrate itself. Consequently th e simple ni-trate has a marked tendency to form. The relation of the con-stituents necessary to the formation of the double salts is changedwhen this takes place, and therefore the fractionations becomevery difficult to carry on. The situation is changed in the caseof the double nitrates in which magnesium is replaced by nickel,salts which, except in color, very much resemble the other.X-ickel nitrate, being more soluble in nitric acid than magnes-ium nitrate, manifests less tendencyto form; and if one takes painsto start the crystallization vith a few particles of double nitrateof bismuth and nickel, double salts free from si aple salts arealways produced.In this new method, gadolinium is concentrated in the firstcrystals (14).

Separat ion of t e rbium r o m dyspros ium. During the whole proc-ess of the crystallizations of the double salts, hitherto described,crystallizations were also made of the yttrium ethylsulphatescontaining the earths of this group, salts of greater stability andmore soluble than the ethylsulphates of the erbium earths.A good separation of terbium from dysprosium may be broughtabout by means of these crystallizations. They eliminate yttriumin the last fractions, with erbium first and holmium following.When t he results obtained by these treatments seem in dangerof being affected by saponification of the ethylsuphates, onemay convert these compounds into simple nitrates and addthe rare earth nitrates which have resulted from crystallizationof the double nickel salts and which have subsequently beencompletely freed from nickel, care being taken to bring togetheronly products of sLxi1ar composition.The methodical fractional crystallization in nitric acid of thesimple nitrates having 5 molecules of water of crystallizationmay then be carried on. This prolonged treatment brings aboutthe separation of terbium from dysprosium (15).

7/28/2019 Chem rev VOL1 143_185

16/43

158 G. URBAINSeparat ion of d y s p r o s i u m f r o m h o l m i u m . This separationis extremely long and arduous.T o accomplish it, it was necessary to continue the crystal-lizations of the simple nitrates of the two earths over a periodof more than two years. At the head of the fractionation seriesdysprosium is obtained; holmium goes over into the followingfractions with yttrium. These methods of separation havebeen described in detail in a general article entit led: Inves-tigations on the separation of the rare earths (16).While carryingon the fractionation of the earths closely associated with yttrium-holmium and erbium-I undertook, about 1905, the study ofMarignacs ytterbium, of which I had been able to prepare some50 grams.For a long time the earths of this group had engaged my at-tention, and as far back as 1901, in collaboration with my brother,

I had announced a method which effects a comparatively rapidseparation of ytterbia and erbia: On the isolation of yttria,ytterbia and the new erbia (17).We had shown that the old method of heating the nitratesgives excellent results when the earths submitted to the treat-ment have been previously freed from earths of the terbiumgroup by the ethylsulphate method.The crystallization of the nitrates, although less efficacious,was recognized as less laborious (18), and it is to this methodthat I had recourse when I set about the breaking up of ytterbiuminto its constituents, an undertaking which led me to dis-cover first lutecium, then celtium.The work of fractionation was most arduous. At certainperiods my assistants and I completed as many as 800 crystal-lizations in one day.

Separat ions of the other earths of the series.

CHARACTERISTICS O F THE YTTRIUM EARTHSAt this point i t is advisable to become more definite respectingthe general and necessarily somewhat vague statements made atthe beginning of this article.

7/28/2019 Chem rev VOL1 143_185

17/43

RESEARCH ON YTTRIUM EARTHS 159In 1898 the subject of the rare earths was in great confusion.There was an abundance of documents of very unequal value.

In these, truth and erroi were closely associated, and therewas no way to distinguish with certainty one from the other.It was even quite difficult to separate facts from the hypothesesand interpretations which, more often than not, needlessly en-cumbered them.Moissans treatise on inorganic chemistry, which appearedin 1904, contains a historical sketch of the subject, under mysignature; this outline covers a centurys work on the par t of thechemists, and shows how the problem, stated simply at first,became by degrees so complicated that even the specialiststhemselves sonietimes lost their may.A thorough knowledge of the history of the subject wouldbe necessary t o follow my investigations in all their details;but I will confine myself to giving a brief sketch ofthe work of the more important of my predecessors. This pro-cess of selection brings before us the results obtained by Lecoqde Boisbaudran, Crookes, Demarpy, and Exner (of Vienna).In the very brief outline which follows, I will show that theirwork was not all on the same plane.

Sir William Crookes studied chiefly phosphorescence spectra.His skill was rewarded by the discovery of one of the mostbeautiful phenomena th at it is possible to observe in the vacuumtubes that bear his name.But he assumed, a priori, that the phosphoyescence spectrabehave like the spark or absorption spectra. This assumptionled him to consider as new substances the yttrium earths, whichemit distinct phosphorescence. But since all these earths gavehim the same spark spectrum, that of ytt rium, and since theycould not at the same time be yttrium and elements distinctfrom one another, he formed the conception of the meta-elements,a paradox which consists in looking upon an element character-ized by a constant spark spectrum as being composed of scarcelydistinct atoms which the fractionation process differentiates,because of the difference of their phosphorescence spectra, butwhich are generally reduced to a single band.

7/28/2019 Chem rev VOL1 143_185

18/43

160 G. URBAINThis idea, incorrect as applying to the chemical elements ofthe rare earth group, has been recognized as correct in the realm

of the isotopes, to which the rare earths, as they are consideredin this article, are not comparable.Moreover, Crookes proposed, one after another, various in-terpretations of the beautiful phenomena tha t he had discovered;but if his theories are not always tenable, the facts that he ob-served are of unquestioned value, and deserve to command at-tention.We shall have to describe precisely the nature of the elementsor meta-elements tha t he designated by G,, GB,. . . GO,Gs,. . .etc., the bands of which may be observed in the visible spectrum,and of the phosphorescent elements that he designated as vie-torium, ionium and incognitum, the bands of which may beobserved in the ultra-violet region.Were they new? Had they not already been described,thanks to some other characteristic? Were they even reallydistinct from one another?Lecoq de Boisbaudran thought that Crookes visible bands inthe phosphorescence spectrum of yttria belonged to earths thathe had previously designated by the characters 2, and Zp; hehad observed the phosphorescence of these earths in connectionwith the interesting phenomenon that he had discovered, in thecase of certain solutions, by reversing the direction of the non-condensed induction spark.But what were the elements 2, and 2, themselves? Lecoqde Eoisbaudran, indeed, told us that Zp is concentrated in thedark-colored terbias, and 2, in the light-colored terbias, whichare less basic than the dark. But this subject of the terbiaswas most complicated, for the same author had distinguishednmong the dark terbias an element 26, characterized by anabsorption band, and in the light terbias an element Z charac-terized by a spark spectrum, and another which he namedd j prosium, characterized by one part only of the spectrumof the original holmium ( X of Marignac and Soret). iI7hat werethe relations arcong these elements, or rather among thesespectroscopic characteristics?

7/28/2019 Chem rev VOL1 143_185

19/43

RESEARCH ON YTTRIUM EARTHS 161Demarqay, who had developed a new spectroscopic technique,added new properties to the existing abundance of characteris-

tics, and new designations to the large number of tentativenames. i h o n g the thousands of lines in the ultra-violet field,he distinguished a dozen which he attributed t o an element r,and as many more to an element A. The first is to be found inthe dark terbias, the second in the lighter earths, those tha tappear later in the series.Exner described an arc spectrum of terbium. But whatexwtly was this terbium? The descriptions of the successiveauthors who had discussed the question since it was first pro-pounded, sixty years earlier, were hardly in accord. To Mo-sander, who discovered it, it was an element to which was duethe yellowish color of the yttrium earths between white yttriumoxide and pink erbium oxide. To Marignac it was orange-colored earth, whose place was between gadolinium and earthX (original holmium). T o Lecoq de Boisbaudran there wasno terbium, but there mere terbias, and he thus designated allthe yttrium earths of this group having colored oxides, the colorvarying from yellow (light terbias) to dark brown (dark terbias).In these terbias he observed m a n y spectroscopic characteristicsbelonging t o elements which might be distinct.The spectroscopist Exner also described an arc spectrumof holmium. But this holmium, prepared by Cleve, who sonamed the X of Marignac and Soret, contained as impuritiesmost of the yttrium earths and considerable quantities of yttriumand terbium. It corresponded to the purer, light-colored earthsin which Lecoq de Boisbaudran observed not only true holmium,but also dysprosium, 27, and 2,. Exner distinguished XI, X, , Xain Cleves holmium. Such, in r6sum6, is the history of the earthsforming the heart of the yttrium series.On the other hand, the history of the first members of theyttrium series, europium and gadolinium, was not yet free fromobscurity; and the individuality of these two elements stillpresented many uncertainties.Mosander had distinguished only these three earths in yttria. But he hadgiven the name of erbium to that which, in consequence of confusion, was latercalled terbium.

7/28/2019 Chem rev VOL1 143_185

20/43

162 G. URBAISCrookes was the first to observe one of the characteristicsof europium : it was a band of orange-colored phosphorescence

which he at first thought an anomalous band resulting from thecombined phosphorescence (? ) of yttrium and samarium, andwhich he later attributed to a meta-element, S , of samarium.Some years later Lecoq de Boisbaudran observed, by thismethod of reversing the sparkwith solutions, an analogous bandof fluorescence which he attributed to an element Zr. Hestated tha t this element was probably not without relation to theanomalous band of Crookes.Demarpays europium showed all these characteristics, andDemarpay doubted the simplicity of this element, not only be-cause of the previous experiments of Lecoq de Boisbaudran, butalso because europium exhibited an absorption spectrum rich inlines, but in general extremely faint; and because its sparkspectrum showed, besides numerous strong lines, a background ofextremely numerous but very faint lines. He considered thesefacts as so many spectroscopic anomalies in conflict with allthat he had so carefully observed in the pure earths of the ceriumgroup.There, assuredly, was an array of presumptions, but no directproof of the complex character of europium.The homogeneity of gadolinium was not debated, but myown research was to bring up, in connection with this element,an unexpected problem, which I myself stated in the followingway: is the element victorium, whose discovery Crookes hasrecently announced, a constant impurity of gadolinium, or hasit been confounded with that element?In order briefly to indicate the degree of our knowledge ofCrookess victorium, I cannot do better than to transcribe herewhat I wrote on the subject in my article. The Rare Earths,for Moissans treatise on inorganic chemistry.

In 1898, Crookes announced the existence of a new element inter-mediate between yttr ium and terbium. This element is characterizedby a particular phosphorescence spectrum, and its atomic weight isabout 118. ,4t first Crookes called this subetance monium, but later,in 1899, victorium. Further research is needed to establish the exact

7/28/2019 Chem rev VOL1 143_185

21/43

RESEARCH ON YTTRIUM EARTHS 163relation of victorium to the terbias of Lecoq de Boisbaudran and t othe earth A of Demarpay.

In consideration of the facts th at I have since established,namely, the identity of victorium and gadolinium (19) and theidentity of dysprosium and the element A (20) ; and further ofth e fact that terbium separates gadolinium and dysprosiumin the series; it becomes manifest that the information given outby Crookes was very inaccurate. But what he did was importnnt;the discovery of victorium was the confirmation of the investiga-tions which this scholar had been carrying on for fifteen yearson the phosphorescence of the rare earths. Victorium wasthe first element which this method of spectroscopic observationled him to discover.8The problems to be solved recalled those confronting the chem-ists who, in the dawn of chemical science, devised proximateanalysis :mixturesmanifested properties corresponding to differentsensations-taste acid, color green, for example. Those chem-ists had at first considered these characteristics separately, asone may legitimately do for light, heat and electricity, whichare produced by material means, and they sought to separatethe acidity principle from the color principle.It was only the constant association of characteristics so dif-ferent that led to the idea of a definite substance, an idea thanksto which modern proximate analysis differs from the decom-position of the mixtures of the alchemists.In the study of the yttrium earths it was necessary to transferthe problem from the field of proximate to tha t of elementaryanalysis; in the terbium group there had been announced aprinciple of color (Mosanders terbium), principles of phosphor-escence (Gp, Gs. , . Z,, Zp. . . ), of absorption (26,dyspro-sium), of emission (r,A , 2,. . . . X 2 . . . ) as varied as pos-sible. In order to arrive at a definite chemical element, the con-stant association of certain of these principles, to the exclusion ofthe others, must be established.

8 I n the observation of his anomalous band, Crookes had really discoveredeuropium, but he had interpreted his observations in a very different way.

7/28/2019 Chem rev VOL1 143_185

22/43

164 G . URBAINThe exclusion was the only part of the problem correspondingto positive experiments : methodical fractionations made success

possible there. But to prove the constant association of simul-taneous characteristics, a method of experimentation must beadopted which should not be limited to negative experiments.The rare earths must be obtained pure before i t would be possi-ble to state precisely the laws obeyed by their spectroscopicproperties, which had been shown to be so different and so differ-ently preceptible.I venture t o emphasize this point. The problem was not somuch to obtain pure products, as products in which thedifferent characteristics should coincide absolutely, withoutvariation, during definite and sufficiently prolonged stages reachedin the course of fractionation; stages of constant atomic weights,of constant and constantly associated spectra of different kinds.If a product at a given stage has withstood a variety of tests;if at similiar stages, products obtained from earths which havebeen extracted from different minerals coming from differentsources, are identical with one another and with the originalproduct, there is justification for considering the inaterial at thisstage as formed of a single element (except for the coinmon ele-ments with which it is combined in a salt), conforming to thedefinition of an element; of which one may only presume that itwill not sometine be resolved into a still simpler body.gI wondered for a long time whether this experimental reasoningdid not in some degree beg the question. I established, forexample, certain laws of cathodic phosphorescence, by the use ofspecimens taken at the definite stages of fractionation, and Iseemed thenceforward to be depending on these laws to provehomogeneity of these same stages. If that were really so , Ishould perhaps have made a vicious circle. In reality, I by nomeans absolutely proved that europium, gadolinium, terbiumand dysprosium are homogeneous. The complexity of a bodymay be proved, but the opposite is incapable of proof. The

9 These considerations have become of historic interest only, now that thechemical elements have been defined by their atomic number (Moseley), whichmay be obtained by means of the X-ray spectra.

7/28/2019 Chem rev VOL1 143_185

23/43

RESEARCH ON YTTRIUM EARTHS 165simplicity of the elements will always be a hypothesis. Whatmy method permitted me to demonstrate positively is that thevariations formerly observed among certain characteristics areapparent variations only. After putting together, sornewhatreluctantly, substances which it had cost so much labor to sepa-rate, I showed that variations in the proportions of the mixturesbrought about the combination of all those appearances whichwere wrongly supposed to be due to the beginnings of separation,obtained by fractionation.It is evident tha t if certain peculiarities presented by iron, forjnstance, should be explained by the supposition that it is complex,that hypothesis might remain a dead letter up to the time v;hentw7o dissimiliar irons were really obtained.

CATHODIC PHOSPHORESCENCE O F THE YTTRIUM EARTHSOn either side of a stage of constant properties, that is, in theintermediate fractions comprised between pure products, novariations haire ever been observed in the absorption, the spark,or the arc spectra which could not be attributed to differences inperceptibility in these different groups of spectra, but such was

not the case with the phosphorescence spectra.Crookes had considered the phenomena of phosphorescence ascharacteristics of elenents which the other spectra were unableto reveal. Lecoq de Boisbaudran held the view that phosphores-cence was apparent only in case of mixtures, and that it was dueto the presence of elements which other spectroscopic characteris-tics could indicate equally well. These two noted scientists hadengaged in a long dispute which served only to emphasize theirdisagreement. We shall see that the view of Lecoq de Bois-baudran was correct. &evertheless the several scientists whohad taken up this problem agreed with Crookes, possibly becausethey had worked with mixtures that were too impure.I was obliged to take up this particularly delicate question.In the effort to solve certain problems relating to the history of therare earths I engaged in the study of a more general subject, thatof the phosphorescence of solid solutions. I succeeded in com-

7/28/2019 Chem rev VOL1 143_185

24/43

166 G . URBAINpleting the task begun by Lecoq de Boisbaudran. I will sum-marize briefly the notes that bear upon this matter:

O n the law of o p t i m u m of phosphorescence in bin ary phosphores-cent systems (21) . The phosphorescence of solid solutions is aproperty of dilute material:1. As Lecoq de Boisbaudran has stated, differing from Crookes,pure substances have no perceptible phosphorescence ; strongphosphorescence is always the result of a mixture of at least twosubstances (binary phosphorescent system) ; one acts as phos-phorogen and the other as diluent.2. In a binary phosphorescent system, the optimum of phos-phorescence always corresponds to small quantities of thephosphorogen. In mixtures of pure rare earths and lime, thisoptimum, never sharply defined, corresponds to the content ofpure rare earths, of the order of one part in a hundred.3 . The law of the optimum is general. It applies equally to thecommon elements and the rare earths.4. The coloration of the phosphorescence, as well as its spec-trum, varies with the dilution of the phosphorogen.

5 . We m-ay reproduce, with preparations made from puresubstances, the marked spectroscopic variations observed in theintermediate mixtures which fractionations of rare earths give.Such variations do not prove, then, that by fractionation anelement, the final term arrived at by analysis (Lavoisier), hasbeen broken up into several constituents.6. The law of the optimum becomes much more precise, ifinstead of being applied t o the phosphorescence as a whole,-it isapplied singly t o each of the bands whose sum total makes up thespectrum.The correct statement of the law is the following:In every binary phosphorescent system in which the relativecontent of phosphorogen and diluent is made to vary, it isproved :1. That each phosphorescent band passes through an optimum.2 . Tha t the optima of the different bands do not necessarilycoincide, although they always correspond to relatively weakproportions of the phosphorogen.

7/28/2019 Chem rev VOL1 143_185

25/43

RESEARCH ON YTTRIUM EARTHS 167The cathodic phosphorescence o f the rare earths (22) . Thefirst part of this article sums up our present knowledge of thephosphorescence of solid solutions. The cathodic phosphores-cence of the rare earths is here more especially studied.Experiments showed that the reversal spectra of Lecoq deBoisbaudran should be considered as cathodic phosphorescencespectra of liquid solutions, and that this phenomenon should beattributed to the ions of the phosphorogen. Generalizing fromthis result, we may assume that the same should be true in thecase of solid solutions.There follows a summary of my general observations upon the

phosphorescence of the fractionated earths.Of the first four members of the yttrium series which I con-sidered to be pure, none showed perceptible phosphorescence.On the other hand, the interdediate mixtures were in generalphosphorescent.Only the earths comprised between terbium and dysprosiumdo not glow under the influence of the cathode rays. This resultis due to the fact that the colorless earths alone may act asdiluents. To be more exact, the earths which give a phosphores-cent spectrum cannot play the part of diluent to those earthswhose phosphorescent spectrum is in the same spectroscopicregion.For the first members of the yttrium series, gadolinium alonemay act as diluent ; europium-bearing gadolinium emits "abeautiful red light and terbium-bearing gadolinium a beautifulgreen light.

For the higher members of the series, yttria (oxide of yttrium)acts as diluent. The yttrium earths a t the beginning of theseries, those carrying dysprosium, give a yellow phosphorescence;those at the end of the series, carrying erbium, a greenphosphorescence. Theirvariations, as a function of their concentration, were the object ofspecial study. The question was, really, to find a correct inter-pretation of a phenomenon which a superficial analysis mightattribute to a possible complexity of earths capable of phos-

Each of these phosphorescences was described in detail.

7/28/2019 Chem rev VOL1 143_185

26/43

168 0.URBAINphorescence. It was certainly such variations that suggested toCrookes his theory of meta-elements.

Confining himself to the analytical side of the research, thisscientist thought that these variations in spectra were due to thebeginning of a separation of distinct elements. The syntheticstudy of mixtures prepared from the beginning from pure sub-stances showed that this is not the case; the relative intensity ofthe bands varies only with the concentration. In the con-secutive parts of a fractionation the concentration necessarilyvaries, and consequently the relative intensity of the bands of thephosphorogen. This variation is not due to a separation of thesupposed constituents of the phosphorogen : it siniply appearsto be.If, as an unverifiable theory, the spectroscopic variations ofbinary systems may be considered as resulting from elementarycomplexity, they may with equal reason be attributed to com-plexity of structure or to movement of atoms, molecules, or evenof the ions which we suppose to constitute all matter; but thereis not sufficient justification for the assertion that variations incathodic phosphorescence, observed in course of the fractionationof the rare earths, result from a separation of the constituentswhich the first hypothesis supposes, unless the fact of a separationhas been confirmed by the study of other spectra.'Iese conclusions were reached only after long consideration,and it was in order to clear up a number of doubtful points thatI undertook a careful study of the optima of phosphorescence.To avoid all uncertainty, I abandoned temporarily the rareearths and turned my attention to the corninon elements, whosesimple character could hardly be doubted. With the assistanceof Bruninghaus I studied with great care the phosphorescence of

the nianganese-lime system. The chief difficulty of the problemconsisted in obtaining pure lime, free from phosphorescence. Butall conimercially pure lime contains some manganese, which,under the action of the cathode rays gives it a beautiful orangephosphorescence. In order to obtain lime free from phos-phorescence we were obliged to carry on daily for three monthssixteen fractions of the pure commercial nitrate of calcium. It

7/28/2019 Chem rev VOL1 143_185

27/43

RESEARCH ON YTTRIUM EARTHS 169is only then tha t me obtained, in the head fraction, lime free frommanganese. Next, we mixed this lime by chemical methods withknown proportions of the oxide of manganese and we determinedthat the luminous optimum corresponded to the proportion ofone per cent of manganese oxide. Above and below that amountthe phosphorescence was more and more feeble. The colorationof the phosphorescence of the mixtures varied slightly, bu t wewere not able to determine whether o r not t ha t variation was dueto some inipurity, since despite all our efforts our reagents werenot epectroscopically pure, and moreover the spectroscopicanalysis of the phosphorescence showed only a single band, verybroad and with ill-defined outlines.In case of the rare earths, on the contrary, the phosphorescencespectra are always made up of narrow bands, in general sharplydefined and numerous.In order to analyze the phenomenon, we had recourse to anearth of the cerium group, samariuir,, which Lacombe and I (23)had successfully isolated nearly quantitatively, and which, afterpurification, had been pronounced spectroscopically pure byProfessor Eberhard of the Potsdam Observatory, after a verycomplete study of it by means of a grating of wide dispersion.It was shown by previous researches of Demarsay thatsamarium was an element defined in the same way as any commonelement. The use, then, of such material gave all the desirableguarantees, and could serve as the starting-point for very delicateexperiments. Now, the study of the samarium-lime systemshowed spectroscopic variations which could leave no doubt as totheir origin.I t was necessary to reject definitely Crookes hypothesis of themet a-elements.

The phosphorescence of the different members of the fractiona-tions m-as then studied with lime as the diluent. This methodhad the initial advantage of avoiding the necessity of mixing therare earths, after all the trouble of separating them. Moreoven.,by t he use, in each mixture, of th e rare earth in the proportion of1 per cent only, which corresponds to the optimum in the case ofpure earths, the spectra of phosphorescence presented n o dilEculty

7/28/2019 Chem rev VOL1 143_185

28/43

170 G. URBAINof interpretation; all the consecutive members of the fractiona-tion of each pure earth showed the same phosphorescence spec-trum, just as they showed the same absorption, spark, and arcspectra. These spectra agreed exactly, and there was no doubtwhatever that they characterized the element, in the same waythat the other kinds of spectra characterized it.This new method of analysis of the rare earths removed alldifficulties relative to the phosphorescent elements. I was ableto identify the different meta-elements of Crookes as elementswhich I had isolated, and to give a very detailed description of thephosphorescence spectra of the different yttrium earths :

O n victorium and the ultra-violet phosphorescence of gadol in iumO n the cathodic phosphorescence of e u r o p iu m (25).Cathodic phosphorescence o f euro piu m di luted in l ime ( 2 6 ) .Spectra of cathodic phosphorescence of terb ium and of dyspros ium

diluted in l ime (27).O n he n a tu r e of certain phosphorescent elements and meta-elements

of Sir W . Crookes ( 2 8 ) .In connection with this subject, the phosphorescence of therare earths diluted in calcium sulphate and in calcium fluoridewas described; and we showed by the examination of alarge number of fluorites of many colors and origins that thefluorites practically always contain rare earths.

(24) .

THE MAGSETISM OF THE RARE EARTHSThe properties which serve to define the elements should beatomic in character. In other words, they should be independent,to a certain extent, of the nature of the combinations in which the

elements exists.The spectra and the atomic weights are of this nature; also,these properties are necessary and sufficient to define an elementand to establish, in some sort, its status. Thereare other properties which may be considered as atomic, forexample, radioactivity. But this property is possessed, a t leastAll the elements give spectra and have atomic weights.

7/28/2019 Chem rev VOL1 143_185

29/43

RESEARCH ON YTTRIUM EARTHS 171in a perceptible degree, by certain elements only. In particular,the rare earths are not radioactive.Between exclusively atomic properties and molecular propertieswe may imagine transitions; some properties are found, althoughin different degcees, in the majority of the compounds of an ele-ment.Indeed, according to a law of Wiedemanns, which is, however,only approximate, all the compounds of a paramagnetic elementhave the same molecular magnetism.As a matter of fact, all the combinations of a paramagneticelement do not follow this rule, but the law is essentially true forall the salts of a given element, with the reservation th at it doesnot forni complex salts, in which the chemical and physicalreactions would be masked.Kow, the rare eaPths do not form complex salts, and the greaternumber of them have been found to be paramagnetic. This thenis a property which in the case of the rare earths has the charac-teristics of an atomic property. Kow, because of the work of P.Curie, the coefficients of magnetism are easily measurable. Thesemeasurements are capable of rendering a service, in the study ofthe yttrium earths, comparable t o the determinations ofatomic weight. According to the determinations of certainphysicists, notably those of Stephan Meyer, of Jienna, differentrare earths possess widely differing magnetic properties, while theatomic weights remain grouped within quite narrow limits.While these determinations were based upon products whichwere, in general, rather impure, they conveyed a useful sugges-tion, and I set myself the task, with the aid of Jantsch, of deter-mining the coefficient of magnetism of the different rare earthswhich I had already been able to isolate in the pure state:

The results obtainedare of especial interest; certain rare earths are diamagnetic (nega-tive coefficients of magnetism) ;bu t most of them are paramag-netic (positive coefficients of magnetism). The different valuesranged from -1 x to 300 x lo+. The difference for twoclosely associated earths was considerable; the ratio of theirmagnetic coefficients was frequently as high as 1 to 6, while their

Such is the case with magnetic properties.

O n the magnet i sm o f the rare earths (29).

7/28/2019 Chem rev VOL1 143_185

30/43

172 G. URBAIKatomic weights did not differ more than two or three units forvalues of the order of 160.

Although the determination of the atomic weights requiredgreat precision, the determination of the magnetic coefficientscould be made by the remarkably simple method devised by P.Curie and Cheneveau. While the determination of an atomicweight requires ten days, that of the coefficient of magnetismrequires only a quarter of an hour. The new quantitativetechnique was thus an ideal guide for treatments which requirenumerous and frequent checks.Such are the practical considerations, which have so great abearing on the final results of investigations consisting of thedefinition and isolation of the successive members of this per-plexing yttrium group.From a theoretical standpoint the results of these researches arenot less worthy of attention :The earths of the cerium group-lanthamum, cerium, praseo-dymium, neodymium and samarium were sharply distinguishedfrom those of the yttrium group-europium, gadolinium, terbium,dysprosium, holmium, erbium, thulium and the original ytterbium.Yttrium, which is diamagnetic, occupied a place by itself.In each of these series, the magnetism, a t first very faint or evennegative, gradually increased, attained a maximum and then

I decreased. From this point of view, the yttrium series, more *complex than that of cerium, was notable also on account of theunexpectedly high degree of magnetism of its members.Not only did the entire rare earth series show two magneticmaxima, but bismuth, which Lacombe and I showed as taking itsplace between the cerium and the yttrium groups by virtue of thechemical properties of its salts, took the same place by virtueof its magnetic properties.Before making, with the assistance of Jantsch, a systematicstudy of the earths of the series as a whole, I had made severalmeasurenents of dysprosium, and I had been struck with thehigh paramagnetism of that element, whose compounds arefifteen times as paramagnetic as those of iron:

7/28/2019 Chem rev VOL1 143_185

31/43

RESEARCH Oh- YTTRIUM EARTHS 173On the ultraviolet spectrum of dysprosium and the remarkable

magnetic properties o f that element ( 3 0 ) . Since that time I havealways followed the progress of my fractionations, from a quanti-tative point of view, by measuring the coefficients of magnetism.This method, which gives the service generally required ofgravimetric analysis, can moreover be combined with it to ad-vantage fo r the solution of certain problems of exact elementaryanalysis. I gave an instance of this and described the newmethod of analysis in a m-ore recent note : Inmixtures, the magnetism of the rare earths is, Like the mass, anadditive property.Since, moreover, the coefficients of magnetism of the rareearths are not proportional t o their atomic weights, we maydemonstrate, by the aid of appropriate diagrams, the existence ofan intermediate earth between two end fractions containing puresubstances.The principle of this method of research upon the elementsrecalled, in a certain measure, the principle of the method used ininvestigations upon radioactive substances. I applied this, byway of illustration, t o the earths existing between dysprosiumand pure yttrium, and I demonstrated by a study of the magneticproperties of the successive members of the fractionations thatthe holmium contained in the mixtures could not have dis-appeared. Along with these investigations of magnetism Isucceeded in establishing the fact that Marignacs ytterbrium isnot an element, and I uniformly followed the progress of myfractionations by means of measurements of magnetism.

O n the magneto-chemical analysis o f the rare earths (31) .

THE ATOMIC WEIGHTS O F THE RARE EARTHSThe chemical elements are not officially recognized until theyappear in the table of the International Atomic Weight Com-mittee. This decision, which was made a t the InternationalChemical Congress of Paris, in 1900, is justified by the importantpart which the atomic mass plays in the mutual relations of theelements.

7/28/2019 Chem rev VOL1 143_185

32/43

174 G. URBAINDeterminations of atomic weights are of the greatest interest.From the beginning of these studies I have been obliged to pay

much attention to this delicate matter. In the course of my workI have made several hundred determinations : The rareearths comprise the following sixteen elements, arranged in theorder of their increasing atomic weights:Revision of the atomic weights of the rare earths (32).

ScandiumYttriumLanthanumCeriumPraseodymiumKeodymiumSamariumEuropium

GadoliniumTerbiumDysprosiumHolmiumErbiumThuliumNeoytterbiumLutecium

Among these elements, holmium, thulium, erbium, neoytter-bium, and lutecium are still little known. The isolation ofeuropium, terbium, and dysprosium, in a state of purity, is ofvery recent date. There is as yet only a small amount of informa-tion concerning their atomic weights. These values thereforeshould be considered as first approximations, although they prob-ably are not far from correct, especially those which agreed forseveral consecutive fractions of the series by means of whichthese earths were isolated.In fact the constancy of the atomic weights in successive stepsof a methodical and lengthy fractionation of a pure earth of thisgroup permits the establishment of a value to a high degree ofprobability. ,4value determined solely from a specimen selectedfrom among the members of a fractionation is, in general, a pointchosen arbitrarily on a continuous curve, unless it is establishedthat this curve becomes a straight line representing constantatomic weights. It is therefore absolutely necessary in the caseof elements which cannot be sharply separated from one anotherby any method, and which can be purified only by progressivetreatments, to prove the constancy of the atomic weight by pro-longing the fractionation beyond the point which seems necessaryto obtain pure products.

7/28/2019 Chem rev VOL1 143_185

33/43

RESEARCH OX YTTRIUM EARTHS 175If this consideration, essential in the determination of theatomic weight of a rare earth, is neglected, and if measurement is

limited to the fraction which appears to be the purest, the valueobtained has but a limited claim to probability. Many similarmeasurements made by different investigators would have toagree before they would lose this character. A comparison ofthe determinations given below is enough t o show that thisgenerally holds true.The history of t he atomic weight of terbium offers indeedagood illustration of the foregoing statement. These are thevalues successively announced :DelafontaineSmithDelafontaineMarignacLecoq de BoisbaudranLecoq de BoisbaudranKruss and HofmannMarcFeitMelle PotrazG . Urbain

(1864) . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 3 . 0(1878) , . . , , 1 5 3 . 7(1878) . . . . . . . . . . . . . . . . 1 4 7 . 0(1878 . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 4 8 . 5(1886) . . . . . . . . . . . . . . . . . .

. . . 1 4 8 . 9 - 1 6 2 . 2(1902) . . . . . . . . . . . . . . . . . . . . . . . . . . . 151.9-161.2(1905) ........................... 1 5 8 . 6(1905) . . . . . . . . . . . . 1 5 4 . 0(1906) . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 5 9 . 2

A table like this needs no comment.In the case of the common elements, which may be obtainedfairly easily in a degree of purity superior to that which theaccuracy of the measurements permits, the whole effort of theexperimenter is necessarily directed toward everything that mayincrease that accuracy. In the case of a rare earth, of which ithas been practically impossible until recently to determine thedegree of purity, the effort to secure great accuracy was evidentlysomewhat illusory. Great accuracy has no meaning unless thereis the assurance that the inaccuracy due to the presence ofimpurities does not much outweigh that which is inherent in themethods of measurement themselves. Success is possible onlyif the successive members of a fractionation give identical resultswith an identical method of measurement.In every other case, the influence which impurities may produceis a matter of judgment, or of rough spectroscopic comparisons,

C H E N I C A L REVIEWS, VOL. I, NO.

7/28/2019 Chem rev VOL1 143_185

34/43

176 G . URBAINand there is thus introduced into the determinations a class oferrors which defy the resources of the calculus of probability.

It follows from the foregoing considerations tha t in the presentsta te of the case the only determinations which deserve acceptanceare those which have been made upon products sufficiently puri-fied; for the analysis and synthesis of sulphates, applied in thiskind of measurements, give rise t o errors of the second order incomparison with those which come from the impurity of theproducts.ISOLATION AND DEFINITION O F THE EARTHS OF THE YTTRIUM GROUP

SamariumOn samarium; i ts isolation and atomic weight (33). Samarium isthe last member of the cerium series.We published a method, already alluded to for its exact andquantitative separation from europium, the first member of theyttrium series.After making a suitable purification, we obtained samariumspectroscopically pure, a result up t o that time unknown inresearch of this kind. Samarium behaved like an element

throughout our treatments; its characteristics remained inva-riable.Subsequently I studied the subject of the phosphorescencespectra of samarium, and demonstrated that the constituents ofsamarium discovered by Crookes had no real existence.Europium

On europium; i ts isolation and atomic weight (34). We deter-mined a constant atomic weight for europium extracted frommonazite sands. This element was subjected to a long fractiona-tion after its separation from samarium, which precedes it inthe rare earth series, and from gadolinium which follows it.Later I carried on a similar study of the other properties ofeuropium, and especially of its absorption spectrum, which, inthe spectroscopic field earlier studied by Demarqay, was extremelyfaint. I observed far over in the ultraviolet field some very

7/28/2019 Chem rev VOL1 143_185

35/43

RESEARCH ON YTTRIUM EARTHS 177pronounced absorption bands belonging to tha t element. Thechief consideration that had led Demarpay to doubt the homo-geneity of europium existed no longer. The absorption, spark,and arc spectra of europium remained constant throughout theduration of our fractionations.Europium extracted from monazite, from gadolinite, fromxenotime and from the very radioactive rare earths of pitchblendemanifests identical characteristics (35).Nevertheless these results were not entirely satisfactory.Crookes, to whom I had given a sample of europium, had claimedthat this earth was not the same as his S8.

I knew very well that Crookes had observed up to that timeonly fairly complex mixtures; but the profound spectroscopicdifferences which he described between the phosphorescence of SSand that of europium left room for doubt.I made up my rdnd t ha t the anomalies pointed out by Crookesdepended upon laws of cathodic phosphorescence not yet under-stood. Experiments justified that point of view, and I was atlength able to assert that Sa (Crookes), and also Zg (Lecoq deBoisbaudran) are really identical with europium (2, of Lecoq deBoisbaudran) (36). Gadol in ium

The results of those studies that had to do with the isolationand definition of gadolinium have been outlined in the followingnotes:On gad olini um; purijicatiol? and atom ic weight (37).O n a ne w spe ctr um observed in gadol in ium ( 3 8 ) .O n victo rium and the ul traviolet phosphorescence of gadol in iumGadolinium was defined very incompletely on account of itscolorless salts and the absence of absorption spectrum. De-marpay, who had obtained i t in a reasonably pure condition, haddescribed only a spark spectrum. The nun ber 156 had beenaccepted as its atomic weight, after the very interesting in-vestigations of Benedicks, although his gadolinium always carrieda little terbium and europium.

(39).

7/28/2019 Chem rev VOL1 143_185

36/43

178 G . URBAINAfter having fractionated gadolinium well beyond the pointnecessary to obtain it pure, I performed a very extended series

of constant atomic weights. The measurements raised by 1.3the value 156 then recognized by the International AtomicWeight Committee. This correction seems small. In realityit is considerable, in view of the atomic weights 152 and 159.2,those of europium and terbium, which are the common impuritiesof gadolinium.In exploring the spectroscopic regions well over in the ultra-violet, I found that this earth gave a new absorption spectrum.This characteristic belongs exclusively t o gadolinium.The spectrum of phosphorescence assigned by Crookes to vic-torium belongs also to gadolinium alone. But Crookes was notready t o admit the identity of his victorium with gadolinium.I then attempted in vain by new methods of fractionation toobtain either gadolinium free from victorium, or victorium freefrom gadolinium. As the Yesult of these experiments and of acritical study of the work of Crookes, I arrived at the conclusionthat victorium with its atomic weight of 118 must be a mixtureof gadolinium and yttrium (40). Finally, by mixing suitableproportions of yttr ium and gadolinium, I obtained an earth

which possessed all the properties assigned by Crookes tovie t orium.Terbium

In 1846, Mosander had observed that the yttrium earths con-tained in his fractions between the pink oxide of erbium and thewhite oxide of yttrium had a yellow coloration. He attributedthis coloration to the existence of a new element, which wasnamed terbium. Some years later, Delafontaine obtained thesame results and confirmed the conclusions of Mosander.Nevertheless there mas a question as to the existence of ter-bium, for there was nothing to prove that the yellow colorationwas not due to the presence of some element already known.Thus, in 1866, Bunsen concluded from his researches upon theyttrium earths that their yellow color chould be attributed to atrace of cerium. I n 1878, Cleve and Hoglund confirmed theconclusions of Bunsen.

7/28/2019 Chem rev VOL1 143_185

37/43

RESEARCH ON YTTR IUM EARTHS 179However, Delafonatine maintained that the conclusions ofMosander alone were correct, and requested Marignac to verify

the foundation of his claims. Marignac furnished incontestibleproofs of the existence of terbium, bu t he did not succeed inisolating pure terbia. He con-sidered what was then called terbia to be really a group of ele-ments. By the study of reversal spectra, he distinguished twosubstances, which he provisionally designated as 2, and 20.Furthermore, he observed in the light-colored terbium earths aspark spectrum which he attributed to an element Z,, and in thedark-colored terbium earths an absorption spectrum which heattributed to an element Z . Finally, Demarsay attributed t o anelement r a spectrum of ultraviolet lines observed with theearths of the group. In fact the whole question since the time ofMosander was singularly complicated, and it had not been possi-ble to isolate any substance having constant properties.

I succeeded in separating this very dark-colored earth withconstant characteristics and an atomic weight of 159.2. This isthe earth which, owing to thelstrong coloring power of its oxide,gives t o the yttrium earths the yellow color which led Mosanderseventy-seven years ago to suspect it s existence.

In 1885 Lecoq de Boisbaudran took up the matter.

This subject was discussed in the following notes:O n a n yt tr iu m earth closely associated wi th gado linium (41).On the isolation of terb ium (42).At om ic weight and spark spectrum of terb ium (43).Cathodic phosphorescence spectra of terb ium and dysprosiumO n he n a tu r e of certa in phosphorescent elements and meta-elementsO n some compounds of t e rb ium and dyspros ium (46).The isolation of terbium did much to clear up the confusionthat existed with regard to the earths of this group, for it enabledme to identify as terbium a certain number of substances whichwriters had designated by provisional symbols and occasionallyeven by new names:

diluted in l ime (44).of Sir Vi. Crookes (45).

7/28/2019 Chem rev VOL1 143_185

38/43

180 G . URBAIN

Terbium.........Z 8 . ..............G 8 . .............Z 6 . . ............Ionium. .........Incognitum.. ....r ................

CHARIICTERIWICBIAME8 O R 87YBOIBYellow color of oxidesReveranl spectrumVisible phosphorescence spectrumAbsorpt ion spectrumUltraviolet spectra of phospho-rescenceBpark spectrum

AUT HOB I T I E S

MosanderLecoq de BoisbaudranCrookesLecoq de BoisbaudranCrookesDemargay

D y s p r o s i u mPure dysprosium, which in the yttrium series immediately

follows terbium and precedes holmium, was likewise isolated forthe first time somewhat later.Lecoq de Boisbaudran had so named an earth characterized byan absorption spectrum which came from the splitting up of anabsorption spectrum observed by Soret in some of Marignacsfractionation products. Sorets spectrum had been attributed tothe elements X, which Cleve named holniium.Lecoq de Boisbaudrnn observed the presence of dysprosium inmixtures only, and he could not observe the color of its salts.I isolated a few grams of dysprosium characterized by constantproperties and by salts of a pale yellowish-green color.This subject has been outlined in the following notes:O n the isolation and the various atomic properties of dyspros iumCathodic phosphorescence spectra of terbium and dyspros iumAto mic we ight s of dyspros ium (48).On the nature of certain phosphorescent elements and meta-O n some compounds of t e rb ium and dy spros ium (46).O n the ultraviolet spectrum of dysprosium and the remarkablemagnetic properties of that element (30).These investigations also showed that a certain number ofsubstances whose individual existence had been considered aspossible or probable are identical.

(47).diluted in l ime (44).

elements of Sir W . Crookes (45).

7/28/2019 Chem rev VOL1 143_185

39/43

RESEARCH ON YlTRIUM EARTHS 181

Dysprosium.. ..........A.. ...................2 a.. .................2 7 . . . . . . . . . . . . . . . . . .G 6. . . . . . . . . . . . . . . . . .XI . . . . . . . . . . . . . . . . . . .

I AUTROBITIEBHARACTBBISTICI.IA M E I O R BYMBOLdAbsorption spectrumUltraviolet spark spectrumReversal spectrumVisible spark spectrumPhosphorescence spectrumUltraviolet arc spectrum

Lecoq de BoisbaudranDemargayLecoq de BoisbaudranLecoq de BoisbaudranCrookesExner and Eascheck

Holmium, erbium and thuliumI have not succeeded up to the present time in obtaining in apure state the elements of the group; nevertheless I have been

able to determine some of their characteristics (49).THE CONSTITUENTS OF MARIGNAC'S YTTERBIUM

Discovery of celtium, Neoytterbium and luteciumThis is perhaps the most important of my investigations uponthe rare earths. It resulted in the discovery of two elements:lutecium and celtium. But i t is not yet completed.10In 1878 Marignac had succeeded in separating from erbia,

whose salts are pink, a colorless substance of the atomic weightof 173, to which he gave the name ytterbium.In 1904I began the fractionation of an ytterbium correspondingapproximately to the description of that element given byMarignac. My object was to assure myself of the constancy of itsatomic weight, and to determine its value in case I ahould findit constant.In 1607 I published my obmrvations:Starting from that part of the fractionation where I observedonly faintly the absorption spectrum of thulium, never verybright, the atomic weights, instead of being constant, varied in avery regular fashion from 170 to 174. Ytterbium was not anelement.

1 0 I obtained with the collaboration of Blumenfeld a neoytterbium scries offractions containing twelve members of identical properties; but there is here alimit of fractionation, for this neoytterbium cctrries still a small amount of lute-cium, which doubtless raises its atomic weight of 173.5.

7/28/2019 Chem rev VOL1 143_185

40/43

182 G . URSAINThe discovery of lutecium is certainly due to the method ofwork the principles of which I explained in the early part of this

paper.In the early members of the fractionation the presence ofthulium lowered the atomic weight. The higher atomic weightof the final members of the series might be caused by the pres-ence of thorium (atomic weight 233). I removed completelythe very small trace of it. But this trace was not influencing theatomic weight determinations, as I ascertained. Then I com-pared the arc spectra of t he fractions at opposite ends of theseries, taking care to photograph the spectra upon the sameplate, one above the other. The fraction having the higheratomic weight showed strong lines, which were either absent orvery faint in the fraction having the lower atomic weight.I then supposed that this difference was simply a physicalphenomenon due to the arc, and I compared in the same way thespark spectra, which showed exactly the same peculiarities.Next I made use of a device of Lecoq de Boisbaudran, whichmade it possible to obtain with ytterbium solutions a sparkspectrum not of lines but of bands. It was by means of thisspcctrum that Marignacs earth had been characterized by Lecoqde Boisbaudran, to whom Marignac had given his preparation.A comparison of the extreme ends of the ytterbium seriesshowed that a band of 3 , called y by Lecoq de Boisbaudran, wasthe most pronounced in the ytterbium of higher atomic weight,and was entirely wanting in the earth of lower atomic weight.I gave theelement having the band y and the new spark and arc spectra thename lutecium.A new element: lutecium, resulting from the splitting up ofMarignucs ytterbium (50). One month later, Auer von Welsbachannounced to the Vienna Academy of Sciences the same results&s mine.He tried to bring up the question of priority; I firmly contestedthat claim (51).An additional note, On lutecium and neoytterbium (52) ,containedall the atomic weight determinations of the successive members

There was noting to do but to yield to the evidence.

7/28/2019 Chem rev VOL1 143_185

41/43

RESEARCH ON YTTRIUM EARTHS 183of my fractionation, as well as the measurements which relatedto the coefficients of magnetism.As stated above, I had been stopped in my fractionations bylack of material. It became necessary to prepare a fresh supply,and therefore to begin new7 treatments. With this in view, Iturned to earths extracted from gadolinite (my previous work hadbeen conducted upon earths extracted from xenotime).Bourion and Maillard were good enough to give par t of theirtime to this undertaking, in which we made use of new methods:Extract ion of lu tecium f r o m the ear ths of gadolinite ( 5 3 ) . Sub-sequently I added these earths to those previously prepared, sodoing it as not to lose the benefit of the fractionations of both.The progress of the separations was followed by magneticmeasurements.

Cel tiumThe oxides contained in the final mother-liquors, which refusedto crystallize, showed properties which I could interpret only asdue to the presence of a new element; this I called celtium.My investigations continue to be halted by lack of material,

and it was impossible for me to submit the case of celtium to therigorous method which made possible the identification, withoutuncertainty or possible mistake, of the different yttrium earthswhich are the subject of this paper.Meanwhile, Moseley, having discovered his celebrated law,advanced the hypothesis that celtium was perhaps identicalwith the unknown element of atomic number 72. Investigationswhich Moseley and I made on this subject in June, 1914, yieldedno results. But de Broglie having perfected the original methodof X-ray spectra, the product containing celtium was turned overto him; and Dauvillier, his pupil and collaborator, found in it , in1922, lines characteristic of element 72. These lines gave toceltium, up to that time ill defined, a correct definition.Coster and Hevesy, who in 1923 found the same spectrum inzirconia, gave the same element the name Hafnium. Dauvillierand I should have established our rights of priority, and claimedfor element 72 the name celtium, which we agreed to give it.

7/28/2019 Chem rev VOL1 143_185

42/43

184 0. U R B A I NREFERENCES

(1) URBAIN:Compt. re nd . 143, 825 (1906);Urbain and Scal, i b i d . 144, 30 (1907).(2) URBAIN:Bull. so r . chim., 16 , 338, 347 (1896).(3) URBAIN:bid . , 17, 98 (1897).(4) URBAIN:Doctor thesis, page 43.(5) URBAIN: . chim. phys., 4, 64 (1906).(6) URBAIX:Comp. rend. , 126, 835 (1898).(7 ) URBAIX:Doctor thesis.(8) URBAIN:Compt. rend. , 138, 792 (1904).(9) URBAIN: . chim. phys., 4, 105 (1906).

(10) URBAIN: ompt. re nd . , 137, 792 (1903).(11) URBAIN ND LACOMBE:ompt. re nd . , 149, 37 (1909).(12) URBAIN:Compt. re nd . , 138, 84 (1904).(13) URBAIN ND LACOMBE:ompt. re nd . , 138, 627 (1904).(14) URBAIN:Compt. re nd . , 140, 583 (1905).(15) URBAIN:Compt. re nd . , 141, 521 (1905); i b i d . 142, 785 (1906).(16) URBAIN: . chim. phys., 4, (1906) (July 12).(17) URBAIN:Compt. re nd . , 132, 136 (1901).(18) URBAIN: . chim. phys., 4, (1906) (Feb. 26).(19) URBAIN:Compt. re nd . , 141, 954 (1905).(20) URBAIN:Compt. rend. , 142, 785 (1906).(21) URBAIN:Compt. rend. , 147, 1286 (1908).(22) URBAIN: nn. chim. phys., (8), 18, 222, 289 (1909).(23) URBAIN XD LACOMBE:ompt. re nd . , 138, 1166 (1904).(24) URBAIN:Compt. rend. , 141, 954 (1905).(25) URBAIN:Compt. rend., 142, 205 (1906).(26) URBAIN:Compt. re nd . , 147, 1818 (1906).(27) URBAIN:Compt. re nd . , 143, 229 (1906).(28) URBAIN:Compt. rend. , 146, 1335 (1907).(29) URBAIN ND JANTSCH: Compt. re nd . , 147, 1286 (1908).(30) URBAIN:Compt. rend. , 146, 922 (1908).(31) URBAIN:Compt. rend. , 160, 913, (1910).(32) URBAIN: ull. sot. chim., (4) 6 , 133 (1909).(33) URBAIN ND LACOMBE:ompt. rend., 148, 1166 (1904).(34) URBAIX ND LACOMBE:ompt. rend. , 138, 627 (1904).(35) URBAIN: . chim. phys., 4, (1906) (July 12).(36) ORBAIN:Compt. re nd . , 142, 206, 1518 (1906); J. chim. phys., 1 . c .(37) URBAIN:Compt. re nd . , 140, 583 (1905).(38) URBAIN:Compt. rend. , 140, 1223 (1905).(39) URBAIN:Compt. rend. . 141, 954 (1905).(40) URBAIN: . chim. phys., 4, (1906) (Oct. 27).(41) CRBAIN:ompt. re nd . , 139, 736 (1904).(42) URBAN:Compt. rend., 161, 521 (1905).(43) UnnAIr;: Compt. re nd . 142, 957 (1906).(44) URBAIN:Compt. rend. , 143, 220 (1906).(46) URBAIN ND JASTSCH:Compt. re nd . , 146, 127 (1908).(45) UnBAIK: COmpt. rend. , 146, 1335 (1907).

7/28/2019 Chem rev VOL1 143_185

43/43

RESEARCH ON YTTRIUM EARTFIS(47) U R B A I N : ompt. rend. , 142, 785 (1906).(48) URBAIN ND DEMENITROUX:ompt . rend. , 143, 598 (1Doa).(49) U R B A I N :. chim. phys. , 4, (1906) (Feb. 26).(50) U R B A I N : ompt. rend., 146, 759 (1907).(51) U R B A I N : h e m . Ztg. (1908) (Aug. 1) .(52) U R B A I N : ompt. re nd . 146, 759 (1907).(53) URBAIN, OURIONND MAILLARD:omp t . re nd . , 149, ln (1909).

185