P. w.SELWOOb
2362 [CONTlUBUTIO# FROM THE
VOl. 5ci
F R ~ CCHEMICAL K LABORATORY OF PRINCETON UNIVERSITY ]
Magnetochemical Properties of Smaxkm* BY P. W
SRLWOOD
The purpose of this work was to clear up one ur two points in connection with the magnetic properties of samarium compound$, and to detetmine the electronic configuratiori of the divalent samarium ion.
The statement appertrs in the literature' that nitrogen may egect a surface reduction of samarium salts. This observation was verified a t a temperature of 450". But with the apparatus described, dehydration was complete a t 2.iOo, a t which temperature no reduction took place. The anhydrous samarium tribromide havmg been obtamed, pure dry hydrogen a t normal pressure was passed Experimental Part over the salt while the temperature was raised slowly to 740" and held there for five hours. Progress of reductid Preparation of Materials.-A quantity of samarium of the highest purity was put at the writer's disposal through could be followed by development of the intense chocolatebrown color characteristic of divalent samarium Comthe kindness of Professor B. S. Hopkins of the university of Illinois The cotnpounds on which Magnetic measure- pletion of reaction was indicated by discontinued evoluments were made were the oxide, sulfate octahydrate, tion of hydrogen bromlde The product was analyzed for tribromide and dibromide. With the exception of the both samarium and bromine afid was found in different dibrornide, preparation of these compounds involves no batches to run from 50 to 90% pure SmBrz, the remainder difficulty. Divalent compounds of samarium, however, being unreduced SmBrS for which a suitable correction have been prepared only a few times previously and in the magnetic measurements was made. Samarlum The . ~ ~ dibromide proved to he quite stable if kept perfectly dry, probably never in a high state of p u r i t ~ . * J < ~ preparation consists of two steps, first, dehydtatim of the preferably ih a hydrogen atmosphere. But on expbsure moist tribromide, and second, reduction with hydrogen to to inoist air &tomposttion was alrtrbst instantaneous the dibromidr Bath these steps were carried out con- Measurements taken twelve hours apart, however, on material kept in the magnetic balance gave no indication veniently in the same apparatus (Fig. 1). A stream of of decomposition. F -7r - -4, Maghetic hfeasurements.-The magnetic measurertlehts were done on e form of the G u y bafence previously described in detail.a.9 The same precautions and methods of handling the results were adhered to
Efxperimental Results Table I gives the magnetic susceptibilities of the compounds Sm&, Smz(SO&.8H20 and SrnBra at 20'. In this table xs is the specific (gram) susceptibility, XM the molar susceptibility and X h + + t the susceptibility per gram-ion of the trivalent ion corrected as usual for the diamagnetism of anion, crystal water and cation. Fig. 1.-A.pparatus for the preparation of samarium dibromide.
TABLE I Xg
oxygen-free nitrogen was saturated with kybtbbtufnic acid, dried m e r p h o s p h m s pentox)cle, and f3sssed oler fhe hydrated tribromkk a t a pressure of a few c'm. of mercury. The temperature was held for two hours a t go", an hour a t 1306,and was then slowly raised to 250' when dehydration was complete and the hydrogen bromide was swept out with a stream of pufe nittMen. The fe&aM product was kotl! anhydrous a#d Itee ffofa bade c6nfsCtmiristiOh (1) Presented a t the Cleveland meeting of t h e Amerkan CKemical Socieky, September 19, 1934. ( 2 ) Matignon and Cazes, Cornpf. rend., 140, 1637 (1905); 142, X:l (1906); Ann. chim. p h y s . . [SI 8, 418 (1906). (:i)Jantsch, Ruping and Kunze, 2. utiorg. d l g < t t i , Chrm., 161, !IO (1927). (4) Prandtl and I c o ~ l r. b d , 173, 205 (l928j (51 Klemm and Rockstroh, rbid., 176, 181 (1920) (6) J.intsch arid Skalla, ibid., 199, 391 (1930).
X 1W
SmPa
8.60
%%.r
3.49 2 34
Sml(SO& 8H10
Xp X W
1950 972 1710
Xs,+++
X 10e
1080
1060 1020
Even in those rtue earths havifig an appreti&ie orbital eompdheM Os maghetic motnew, the 4f eEem6n shell is &Well i shieIbe'd Etom ehterrlb1 ififftla& fiat &i#eren€ compottfids of the sa& element may be ekpected to have approkinmtely the Same magnettc suweptibility per garn-ion. Klernni arid Rockstroh5 haw, howc.ver, reported a wide divergrncr amounting to .-)O( lwtwerri (7) Owrnh, Hdlke rnd Kremers, 'Inis 18) Selwood, rbrd , 66, 3161 (1933). (9) Selwood, tbrd , 66, 4869 (1933)
IoiiHhAi
42, 51.) ( I ~ C L O )
MAGNETOCHEMICA~L PROPGRTIES OF SAMARIUM
Nov., 1934
samarium oxide and tribromide. It seemed worth while, therefore, to repeat their measurements. Table I1 gives their values together with tbe writer's and those of two other recent workers in the field. TABLE I1 PER GRAM-ION X loE AT 20' SUSCEPTIBILITIES SmiOa
Klemm and Rockstroh Freedlo Williamsll Selwood
SmBra Sm(SOi)r~8EfrO
..
990
1560
1060
1050
.. ..
1030
1060
1020
..
..
As the agreement is satisfactory between values reported by the last three workers named, it seems safe to conclude that the discrepancy with regard to samarium tribromide does not exist. Table I11 gives data obtained over a temperature range of about 300' from the compound samarium dibromide, corrected of course for tribromide content. For purposes of comparison the writer's previously published data on trivalent europium are included.
2393
trivalent ions of samarium and europium, the data being taken from the work of Freedloand the The dotted lines indicate ~ r i t e r respectively. ,~ theoretical extension of the experimental data to absolute zero.la The experimental points shown are for divalent samarium, and, as may be seen, are almost the same as for trivalent europium. This is the result to be anticipated from the Sommeifeld-Kossel rule which states that ions with e q d numbers of electrons sften have very similar properties. The result i s af added interest because both samarium and europium have decidedly anomalous temperature coefficients of magnetis susceptibility.
TABLE I11 *E.
373 293 223 153 83
'g
X lo6
15 2
17 19 22 24
2
6 9 6
X M X 10'
4710 5330 6080 7100 7630
xS,++
x
10,
4800 5420 6170 7190 7720
xE,ttt
x 101
4630 (343O) 4940 5480 6060 6370
The only other measurement reported for divalent samarium is that of Klemm and Rockstruh5 also on the dibromide, a t the single temperature 20'. They give xM = 7100 X 1W6,a very considerable divergence from the writer's value. It may be of interest to add that the reflection spectrum of solid samarium dibromide apparmtly consists of a diffuse band throughout the visible bearing no resemblance whatever to the discrete bands characteristic of most trivalent rare earth salts. Discussion of Results Figure 2 shows "effective" Bohr magneton numbers plotted against temperature. These numbers are given by Pr ff. =
VmmV
where x is the susceptibility per gram-ion, k the Boltzmann number, L Avogadro's number and /3 the Bohr magneton 0.9174 X e. m. u. No attempt has been made to evaluate the molecu lar field constant. The solid lines are for tht (10) Freed, THISJ U U R N A L 62, 2702 (1930) (11) Williams Phys R e v , 12, 158 (1918)
0
I
2
3
4
/Jeff.
Fig. 2.--Effective Bohr magneton numbas for the ions Smttt, Eutt+ and Sm++: experimental points 0 are for Smt+.
In the previously reported case of divalent europium9 it was shown that the spectral term must be 8s and the electronic configuration identical with that of trivalent gadolinium because the low value obtained for A, the molecular field constant, in the W e i s law x = C / ( T $. A) excluded any other possibility. For samarium, the anomalous temperature coefficient Fenders it impossible to determine A with any degree of accuracy, and it is doubtful if it would have any significance even if obtained. The evidence in the present case is, therefore, to some extent circumstantial, but in view of the correspondence xsrn*+ = xEu+++ at all temperatures investigated, it is probably safe to say that the electronic configurations of these two ions are also identical. This is the conclusion reached by Klemm and Rockstroh on much less complete evidence. Thc arrangement of electrons in trivalent europium (12) Van Vleok, "Theory of Electric and Magnetic Susceptibili ties, Oxford University Press, 1932.
P. W. SELWOOD
2394
from the 4f shell out is 4fe5s25p6. In divalent samarium it is probably the 6s and one of the 5d electrons which are lost 4f65s25p65d(26s).The remaining 5d electron does not, however, stay in that position, but migrates back to the 4f shell. Its ready removal again is attested by the extremely easy oxidation of the ion. It has been claimed by Prandtl and K0g14 that divalent samarium does not exist, and that its supposed compounds are merely solutions or compounds of the normal salts with the free metal. But the magnetic evidence shows beyond all question that such cannot be the case. Table I V shows all the known valence states of all the rare earths arranged so that ions having equal numbers of extranuclear electrons are on the same row. The table also indicates the normal spectral term probably associated with each ion, together with the number of 4f electrons present. These spectral terms were originally assigned to the trivalent ions only. It has now been shown that they are equally applicable to the divalent ions of samarium and europium, and most likely to the other non-trivalent ions indicated. TABLE IV ISOELECTRONIC ARRANGEMENT OF THE RARE EARTH IONS
No. of
electrons Term
54 55 56 57 58 59 60 61 62 63 64 65 66 67 68
Divalent
4fO'S 4f2
4fa %/, 4f4614.. .
4f78S 4f8 "lie 4fg 6H~j/2 4f10613 4f11 %/, 4f12 8Ha 4f1$2Fr/2 4f" 'S
Quadrivalent
5
La+++ Ce + + + Pr+++ Nd+++ I1
4f1 2Fs/,
4f6 6 H a / , . . . 4 f g 7 F o ...
Trivalent
ce++++
Pr+
+++
Sm+++ Sm++H E u + + + EU++H G d + + + Tb+++/Tb++++ Dy+++ Ho+++ Er +++ Tu+++ ?
H y b + + +
Y b + + /LU+++
+ ++
Vol. 56
The capacity of certain rare earths to form nontrivalent ions is very readily seen from this table to consist of a striving to reach the normal condition of lanthanum, gadolinium or lutecium. To be sure, in some cases such as praseodymium and samarium, the atom is not able to go all the way, and this condition is characterized by intense color and chemical instability. The question now arises: what is it that is so desirable about the normal states of lanthanum, gadolinium and lutecium? The answer must obviously be that these three elements alone are in S states, they have no multiplet structure, the orbital component of magnetic moment in each is lacking. On the basis of such consideration, K l e m ~ n ' ~ was able some years ago to predict the existence of divalent ytterbium. Another possibility concerns thulium. It will be noticed that thulium stands in precisely the same relationship to ytterbium as does samarium to europium. It seems probable] therefore, that thulium will form a series of intensely colored, unstable, divalent compounds, analogous to those of samarium. Thulium is, unfortunately, so rare, that it has not as yet been possible to put this prediction to experimental test.
Summary Magnetic susceptibility measurements have been made on the compounds Smz03, Sm2(S04)3* SHzO, SmBrs and SmBr2. Divalent samarium is shown to have a definite existence, and confirmation is not obtained for a discrepancy reported by Klernm and Rockstroh in the susceptibilities of Smz03 and SmBrs. The SUSceptibilities and hence electronic configurations of divalent samarium and trivalent europium are shown to be identical. PRINCETON, N. J. RECEIVED SEPTEMBER 21, 1934 (13) Klemm, 2. onorg. digem. Chem., 184, 345 (1929)
