MAGNETIC SUSCEPTIBILITIES OF ReC16, ReC13, T c C I ~ AND ilfoClj*
Jan. 5, 1959
t>
the following equation, to which the actual mmole quantities of reactants and products are appended. Sbz(CH3)I 2HC1---+ 2(CH,)zSbCl Hz 0.447 0.885 0.892 0.414 The only assumption here was the identity of the colorless liquid (CH?)2SbCl; its mercury-reactivity made a mol. wt. determmation difficult. Two samples of the bistibine were fully analyzed by carrying on the HCl reaction for 15 hr. a t 250". The HCl requirement was determined by measuring the initial amount and the remainder. The methanehydrogen mixture was analyzed by CuO-combustion. The SbC13 was determined by oxidimetric titration, using standard KBr03 solution. The results are summarized in Table V, listing all products in mmoles. TABLE V ;ZNALPSISOF Sb2(CH&
(CHs)zSbCl, 18.95%. Thus it seemed that the desired cleavage occurred, but the desired (CH3)SbBClz appeared only in the form of its decomposition products, some of which were not intelligible. In a third experiment on this cleavage the boron trichloride was employed in excess and the mixture was heated for 42 hr. a t 60'. In this case the main effect was a methylation of the boron trichloride: 4.696 mmoles of Bel,, reacthg with 3.810 mmoles of. Sbz(CH&, gave 4.318 mmoles of (CH3)2BCl, representing 95% of the unrecovered BC13 (4.541 mmole) and 57% of the methyl groups. Neither the oil nor the sublimate appeared; hence the black solid could contain 0.31 mmole of SbC13and a material empirically formulated as Sb(CH&.de. Thermal Decomposition.-A 0.433 mmole sample of Sbz(CH3)rwas made by the nearly quantitative conversion of 0.883 mmole of (CH&SbH, and withstood one week of heating in a sealed tube a t 100:. Incipient decomposition was observed after 17 hr. a t 160 : a trace of metallic mirror Sample HCl used H2 C H4 SbCls had appeared. After 20 hr. a t 200°, the mirror had inirng.) Found Calcd. Found Calcd. Found Calcd. Found Calcd. creased and the yellow liquid had become colorless. The 182.0 3 . 5 9 8 3.600 0.598 0 . 6 0 0 2.390 2.400 1 . 2 0 1 1.200 yield of (CH3)3Sb(mol. wt. 166.8; v e t . 31 mm. a t 0 " ) was ? 4 2 . 2 4 . 7 8 6 4 . 7 8 8 0.795 0.798 3 . 1 7 4 3 . 1 9 2 1 . 5 9 5 1 . 5 9 6 0.571 mmole. The metal weighed 32 mg. (0.263 mg. atom This very clean demonstration of the reaction Sb*(CH3)4 Sb). These results cleanly indicate the equation 3Sb24(CH&Sb, which requires 0.289 mg. (CH3)4-+ 2Sb f 6HC1 -+ H2 6CH4 2SbC13 dispels any possible doubt of the identity of the bistibine, the formation of which was atom of Sb and 0.578 mmole of (CH3)aSb. Grease-catalyzed Decomposition.-The bistibine proved the most direct evidence of the identity of dimethylstibine. No such cleavage of Sbz(CH3)*occurred when HzS was to be unstable when stored a t room temperature with the used instead of HC1: there was no reaction during 18 hr. a t vapor in contact with a stopcock lubricated with Apiezon T grease. The yellow liquid slowly formed a white solid looD. Cleavage by Boron Trichloride .-The fairly clean low- while the remaining liquid turned bright red. After two temperature cleavage of Sb2(CH3)aby HC1 encouraged the days the liquid had turned colorless and the solid was black. hope of making (CHa)zSbBClzby the action of BC&. How- The yield of (CH3)SSb now was 0.513 mrnole per mole of the ever, two experiments a t room temperature gave only an original bistibine. Hence the non-volatile black solid had adduct empirically formulated as Sb2(CH3)4.0.83BC13. the empirical composition Sb(CH&.as. One sample was heated in vacuo to 100 , giving a 10% yield Acknowledgment.-The generous support of of (CH3)zBCl (proved by hydrolytic analysis), a black solid, this research by the Office of Naval Research is a pale yellow oil and a white sublimate. After destruction of the oil by shaking with mercury, the white sublimate was gratefully acknowledged. Reproduction of this analyzed, showing two B per, C1. The other sample .of the paper in whole or in part is permitted for any adduct was heated only to 60 , giving a little (CH$)zBCl,the purpose of the United States Government. black solid, and the slightly yellow oil without the sublimate. A chloride analysis on the oil gave 18.61%; calcd. for Los ASGELES 7, CALIFORNIA
+
+
+
+
+
[CONTRIBUTION FROM
YENABLE
HALLO F CHEMISTRY, UXIVERSITY O F NORTH CAROLINA]
Chemistry of Rhenium and Technetium. 11. Magnetic Susceptibilities of ReCl,, ReC13,TcCL and MoCl, B Y KERROK N O XAND ~ CHA4RLESEUGENECOFFEY RECEIVED JULY 18, 1958 The magnetic sdsceptibilities of ReCls, ReCIJ, TcC14 and MoClb have been measured by the Gouy technique from liquid nitrogen at room temperature. The pentachloridedj follow the Curie-Weiss relationship with these constants: ReC15, peff = 2.21, 4 = 164'; MoCls, perf = 1.52, A = -23 . The l / x vs. T plots for ReCh and TcClr are not straight lines. Exchange interaction between the magnetic moments probably is significant for all of these compounds.
Introduction Magnetic susceptibilities have proven t o be useful in inorganic chemistry for interpretations of the electronic structures of compounds, especially those of the first transition series. Not so much is known about the theoretical magnetic behavior of compounds of the heavier transition elements, partly because of a lack of data on a sufficiently wide variety of compounds. We have measured the susceptibilities of several chlorides in order to investigate the magnetic behavior of compounds in this region of the periodic table. Experimental The methods described previously2 were used to prepare ReCls, TcC14 and M o c k . Rhenium trichloride was pre(1) Bell Telephone Laboratories, Inc., M u r r a y Hill, New Jersey. (2) K. Knox, S. Y. Tyree, Jr., e l a l . , THIS JOURNAL, 79, 3358 (1957).
pared by (1) the thermal decomposition of ReClb a t 250" and (2) the reaction between ReC15 and Re a t 450" in a sealed tube. Based on the conversion of ReC15, (1) gave yields around 5070 whereas (2) yields around 85%. After purification by vacuum sublimation a t 450°, the products of both reactions were the same. Analysis of several different preparations by essentially the same procedure used previously for ReClS2gave these average results with their standard deviations
'2 Re
(theory 63 66) 63 66 =I= 0 59
'z36C1 25(theory 36 34) i 0 28
Cl/Re 3 00 + 0 01
The compounds were loaded into calibrated Pyrex tubes in the dry box and weighed. The stoppered tubes were sealed off immediately after removal from the dry box. The magnetic susceptibility measurements were made by the Gouy method in fields of up to 5000 oersted in an apparatus previously d e ~ c r i b e d . ~None of the compounds had a field dependent susceptibility within experimental error. The reproducibility of the results on a single sample (3) C. E. Coffey, Ph.D. Thesis, University of North Carolina, Chapel Hill, North Carolina, 1956.
c,
KERROKNOX.ZKD CH.ZRLES EUGENE COFFEY
was 1% or better. The main source of error, however, lies in the purity of the compounds. An idea of the accuracy of the measurements may be derived from the results on three different preparations of MoC16, which had a spread of about 47,. Thus, the accuracy in x is probably a few per cent.
Results and Discussion The corrected molar magnetic susceptibilities of ReC15, ReC13, TcC14 and LfoCIb are given in Table I. The diamagnetic corrections used are those of Klemm, recommended by S e l ~ o o d . ~ In ” addition, these various values were interpolated from those given in ref. 4: Tc(IV), 18 X Re(V), 21 X Along with each value of x is given the standard deviation of many measurements on the same sample. For MoC16, the susceptibilities of three different samples, I, I1 and 111,are given separately. TABLE I CORRECTED hforAR h,fACNETIC SUSCEPTIBILITIES O F ReClv RcC1, TcCli ASD hloC!i Com-
pouncl
T,OR.
RcC16
301 253 196 150 79
Rea1
305 2n0 105 115
78 TcCI,
306 2 53 1% 1Al 78 hloCl6 305 I 305 I1 305 I11 106 I 106 I1 19G 111 78 I 78 I1 i 8 111
21%
x
106
1315 i 4 1438 A= 3 1697 -C 4 19-44 i 23 2064 + 8 ‘lo1 1 420 rt 1 449 f 1 482 f 1 552 & 2 4939 f 32 6’775 =I=77 8338 + 15 11,175 =t 82 17,348 rt 435 1019 f 6 1077 i 4 982 f 4 1640 + 5 1730 f 0 1590 i 3 55A8 f 14 5663 =t 26 4755 i 9
XM
X 108 (lit.)
1225(293°K.)4 1485(105OK.)* 1925(9O0R.)‘ 76(293”R.)
860(281°K.),6 965(293’ K.),6 1090(289”K.)’ 1200(198°R.),61690
1980K.)~ 3600(90°R.),6 4460 (SoOK.)*
The reciprocals of the susceptibilities given in Table I are plotted vs. absolute temperature in Fig. 1. I t is seen that ReClS obeys the CurieWeiss law down t o a temperature slightly below 150’K. and then 1 / x ~begins to drop below the Curie-Weiss value, X M becoming larger than expected. The deviation here is small and just outside the experimental error.~Using only the first 4 points, with peff = 2.83 d / x ( T A), an effective magnetic moment of 2.21 =t0.002 Bohr magnetons is calculated for ReC& with a Weiss constant of 164 f 5’ (the standard deviations here represent the
+
(3a) P. W. Selwood, “Magnetochemistry,” 2nd Ed., Interscience Publishers, Inc., Xew York, N. Y., 1956, p. 78. (4) W. Schiith and W. Rlemm, 2. anorg. allpem. Chem., 220, 193 (1931). (5) W. Sucksmith, P h i l . M a g . , 14, 1115 (1932). ( 6 ) TV. Klemm and H. Steinberg. 2 . anot’g. allgem. C h e m . . 227, 193 (19 ;(i). (7) 13. ‘111. l ~ j a l ~ b wZ‘ioc. , A c a d . S c i . A;,lslt,idom, 36, 693 (1932).
ITOl.SI
U’S of the constants of the least-square straight line, given the U ’ S of the 4 points). These values are in fair agrcenient with those of Schuth and K l e m ~ n , ~ who reported peff = 2.3 with A = 2G5’. The differences are probably due to different amounts of impurities introduced in handling the compound, which is very reactive. The moments agree quite well since they depend only on the slope, while the A’s and actual values of the susceptibilities, compared in Table I, are not in such good agreement Each sample of YloC15 came very close to a straight line over the whole temperature range. The average values of peff and A are 1.52 f 0.03 and - 2 3 5’ (the standard deviations here represent the a’s of the average p and A of the 3 different samples). Again the values are in fair agreement with the previous r e s ~ l t s . j -of~ which Klemni and Steinberg’s are probably the best. Their data, treated by the Curie-Weiss relationship, give peff = 1.5 with A = -30’. Molybdenum pentachloride would be an interesting substance t o take down to lower temperatures because there i t must become ferromagnetic or abruptly anti-ferromagnetic. For ReC13, our values are not in agreement with previous work although the main features of the behavior are the same. The present measurements show a low paramagnetic susceptibility only slightly dependent on temperature. Schuth and Klenim4 found a net diamagnetism, probably temperature independent, which becomes slightly paramagnetic when corrected for the diamagnetism of the atoms. PerakisS found a temperature-dependent paramagnetism for ReBrs and questioned Schuth and Klemm’s values, but he does not give quantitative results The 1’xL1 F I T , T plot is more markedly concave toward the temperature ayis than for ReClj The two high temperature points in a crude Curie-11-eiss calculation give p e f f = 2.04, A = 985‘. Technetium tetrachloride has a high molar susceptibility which is strongly temperature dependent. The l / x . v s . T plot is again curved but for this compound in the direction of the l/x axis. The two high temperature points give perf = 3.14 with A = -57’. h comparison of these results with various calculations for the free metal ions is given in Table 11. The fourth and fifth columns give the ground state and moment calculated on the assumption of ( j , j ) coupling between the electrons, which is the rule for heavier elements, and multiplet spacing wide with respect to kT. The g values are those of the free metal ion.9 The last four columns assume (L,S) coupling and give the ground state and the moments for wide multiplets, narrow multiplets and narrow multiplets with the orbital contribution quenched, i.e., spin only, in that order. As can be seen, none of these simple theories is wholly satisfactory. I t is interesting that (j,j) coupling with wide-multiplet spacing is a close approximation for ReC15. Of course, there is no question of coupling for Mo(V) with only one electron, but the wide multiplet spacing value is much better than
*
(8) U Perakis, J P h y s Rodkum,15, 191 (1954). ((1) See c p , I€ E n’hite, “1ntroduct:on to Atomic Spectra,“ A I c (.r?i\-IIill Boo6 Co , V e w York hT Y , 103%. pp 287, 442.
MAGKETIC SUSCEPTIBILITIES OF ReC16, Re&
Jan. 3, 19.59
7
TcC14AND hloCl:,
TABLE I1 MAGNETIC MOMENTS OF hfETAL CHLORIDES
-
(j.j) Coupling
Compound
ReC16 ReCb TcCli MoC15
fief
(ohsd.)
2.21 2.01 3.14 1.52
Electron structure
5d2 5d4 4d 3 4d
Grou 11d state
gdXiTC
(3/2,3/2)z (1/2,1/2,1/2,1/2)0 (3/2,3/2,3/2) '/2 (3/2) 3 / ~
1.96 0.00 1.55 1.55
-
(L,S) Coupling Ground state aE'2
'Do
w/, 'D3/2
d/eS(S
+ 1) +
~d7FFii ZETT 1.63 0.00 0.77 1.55
4.47 5.48 5 20 3.00
d/4s(s+ 1) 2.83 4.90 3.87 1.73
even the spin only. The d4 configuration of Re- however, for exchange-interaction to be negligible. (111) makes it the most likely t o have multiplet 2500 spacing comparable to k T ,lo which could account for the wide discrepancies in Table 11. For 2250 TcC14, the spin-only value seems the best, although the interaction of the moments is probably complicating the picture. 2000 Deviations from the free-ion moments can be attributed t o a t least two factors: (1) crystal-field 1750 splitting of the ground state, and (2) exchange interaction between the magnetic moments. Since the structures in the solid state of the compounds 1500 under discussion are not known, i t is impossible to 1 attempt a complete theoretical treatment. The XtJ 1250 theory of Kotani" has been applied with some success to compounds of the heavier transition eleMOCLs ments.l2 I t does not fit the present compounds, 1000 however. The molybdenum and technetium chlorides do not follow the theoretical behavior while 750 the rhenium chlorides have the qualitative behavior of the theory only. For ReC& the spin-orbit coupling constant varies between 9000 and 500 15,000, improbably high values, while for ReC& it varies between 1000 and 1800, improbably low val250 ues. I t is probable that these compounds do not meet the requirements of Kotani's theory: (1) cubic symmetry for the crystal field and (2) negliI ! I ! C 50 100 150 200 250 300 2 ;o gible exchange coupling between magnetic ions. TEMPERATURE IN DEGREES KELVIN. The magnetic behavior of these compounds is in Fig. 1.-Reciprocal susceptibility us. temperature. line with the general trend for compounds of the heavier transition elements, which exhibit low moAcknowledgments.-This work was financially ments compared to the first transition series. supported by the U S . Atomic Energy Commission These chlorides are not magnetically dilute enough, whose assistance is gratefully acknowledged. Appreciation is also expressed to the Oak Ridge Na(10) J. H. Van Vleck, "The Theory of Electric and Magnetic Susceptibilities," Oxford, London, 1932, p. 312. tional Laboratory for the loan of the technetium. (11) M Kotani, J . P A y s . SOC.J a p a n , 4, 293 (1919). We wish to thank William A. Wood for assistance (12) R. I3. Johannesen and A. R. Lindberg, THISJ O I J K N A I . , 76, 5349 with the susceptibility measurements. (1954); A. Carnshaw, B. K. Figgis, J. Lewis and R. S. S y h o l m ,
-.
N O ~ I L I179, . P , 1121 (1957).
MURRAY HII,L,NEWJERSEY
