July 5, 1963
COMMUNICATIONS TO THE EDITOR
in benzene. The resulting solution is then separated from the small quantity of benzene insolubles (less than 0.3 g.) by filtration and evaporated a t room temperature in a stream of nitrogen t o yield 11. Final drying is a t 120'. ,4nal. Calcd. for CI2Hl3CrO6P2: C, 55.25; H, 4.43; Cr, 10.0; P, 11.88. Found: C, 56.39; H , 4.36; Cr, 9.7; P, 11.93. Yields vary from 70 t o 95%. The intrinsic viscosity of I1 prepared in this manner is 0.6 to 0.7 in chloroform. Less rigorously controlled conditions lead to polymers with intrinsic viscosities from 0.1 to 0.5. Even unfractionated samples with viscosities in the range 0.12 t o 0.20 have number average molecular weights greater than 10,000 as determined by ebulliometry and vapor pressure osmometry in chloroform. Consequently, the higher-viscosity samples certainly have molecular weights of a t least several tens of thousands. Although any of the groups present in I1 could serve as bridging groups, the most probable structure contains a double-bridged backbone similar t o that suggested for Cr(AcCHAc) (OP(C6H&0)21 except t h a t a cis configuration is not required. The infrared spectrum of I1 contains absorption peaks characteristic of PO2 stretching with virtually the same frequency and absorption profile as found for polymeric Cr(AcCHAc)(OP(C6H5)20)2.1This is strong evidence t h a t the diphenylphosphinate anion is functioning in the same way in both polymers. Furthermore, the hydroxyl group and the water 0-H stretching vibrations can be identified separately a t frequencies which suggest they are normally coordinated groups. Thus infrared indicates that the hydroxyl groups are not bridging groups. The solubility of I1 in benzene or chloroform accompanied by marked swelling and the high intrinsic viscosity values are good evidence for the presence of linear chains as the predominating species with crosslinking only of minor importance. The indications are, then, that the repeat unit is
2019
debted to several of our colleagues for assistance with various experiments. Analytical data were supplied by our Analytical Department. RESEARCH A N D DEVELOPMENT DEPARTMENT A . J. SARACENO PENNSALT CHEMICALS CORPORATION B. P. BLOCK WYNDMOOR. PENNSYLVANIA RECEIVED APRIL25, 1963
The Electronic Structures of Square-Planar Metal Complexes. 111. High-Spin Planar Cobalt(1) and
Iron(I)'S2 Sir: The reaction between toluene 3,4-dithiol and a number of divalent transition metal ions gives complexes which have the general formula
I Formula I is established from analytical and conductance data in Table I. Magnetic susceptibility data in Table I indicate that the copper complex has S = 0, the nickel complex has S = l / 2 , the cobalt complex has S = 1 and the iron complex has S = 3/2. The e.s.r. spectrum of an acetone solution of Ni(TDT)2- gives = 2.08; the [ (nGHg)4N][Ni(TDT)2] powder gives g l = 2.045 and g , , = 2.193.3 The g-values show that the unpaired electron is mainly localized on the nickel. Further, the anisotropy in the g tensor is consistent with a d9planar situation. For comparison, planar Cu(acac)z has g l = 2.075 and g l l = 2.254.* TABLE I PROPERTIES M( TDT)z- COMPLEXES Complexes
Analytical
Found
57.17 H 4.36 [~CBH~)~AS(CH~)][N~(TDT)~] C 57.57 H 4.39 L~CaHs)aAs(CHz)I[Co(TDT)n] C 57.55 H 4.39 [~C~H~)~A~(CH~)][F~(TDT)I] C 57.82 H 4.41
56.79 4.97 57.49 4.37 57.00 4.58 57.18 4.40
[[CsHs)aAs(CHa)][Cu(TDT)zlC
Thermogravimetric analysis of I1 shows initial weight loss a t 375' with a step in the 410-430' region which corresponds to a 3-10Oj, weight loss. No polymer melt temperature has been observed up t o or well beyond the decomposition point. Surprisingly the polymer shows remarkable resistance to hydrolysis and other chemical degradation. For example, no change in intrinsic viscosity is oberved upon refluxing a suspension of the polymer in water for several hours. A cast film of high-viscosity I1 plasticized with 30% Aroclor 1254 has a tensile strength of over 1900 p.s.i. In addition to the diphenyl species described here we have also been able to prepare the analogs with phenylniethylphosphinate, dimethylphosphinate and cacodylate bridging groups. The mechanism by which this kind of polymer forms is not clear. It would appear that the intermediate I could be a polymer somewhat analogous to the phosphinate polymers involving zinc, beryllium and cobalt.? The oxidation step then may serve t o increase the oxidation state of the chromium and introduce the additional ligands. Alternatively polymerization may be involved in the oxidation step. Acknowledgment.-This investigation was supported in part by the Office of Naval Research. We are in-
fieif
Cafcd.
(B.M.) Diamagnetic
h' 85
1.89
91
3.27
96
4.39
94
For 0.0001 M solutions in nitromethane at 25'; expressed in crn.a mole-' ohm-'. a
The electronic spectra of the M(TDT)2- complexes in acetonitrile show the following weak bands, which may be assigned to d-d transitions: M = Ni, 7300 ern.-' ( E 190); M = Co, 9300 cm.-l ( e 58), and a shoulder, indicative of a maximum a t 11,300 cm.-' ( e -100); M = Fe, 7100 cm.-l ( E 45), 7900 cm.-' ( e 5 5 ) , 9200 cm.-' ( E 65) and 10,300 cm.-l ( E 95). The X-ray powder patterns show t h a t the [(CeHb)ZAs(Me)][M(TDT)Z]complexes are isomorphous. We conclude from the above evidence that the complexes are planar and are composed of M f and two dithiolate radical anion moieties. The unpaired electron which each radical anion would possess must be paired (1) Paper 11: E. Billig, K. Williams, I. Bernal and H. B. Gray, Inorg. Chem.. in press. (2) T h e support of the National Science Foundation is gratefully acknowledged. (3) Details of these experiments and e.s.r. experiments on similar Cu2 * complexes will be published by H. B. Gray, E . Billig and I. Bernal. (4) B. R. McGarvey, J . P h y s . Chem., 6 0 , 71 (1956). Also, planar Cu(sa1icy1aldimine)z gives g, = 2.04, gy = 2.05 and g, = 2 20: A. H. Maki and B. R.McGarvey, J. Chem. Phys., a9, 35 (1958).
COMMUNICATIONS TO THE EDITOR
2020
with its partner in the planar complex in one of the Amolecular orbitals of the c ~ m p l e x . ~ The Co(TDT)z- anion is the first example of a highspin, planar ds metal complex. The exceptionally weak field of the T D T ligand is indicated from the position of the first band in Ni(TDT)2-, which is tentatively assigned t o the dxy + dX2--2 transition. For a AI (xy + x 2 - y2) in the neighborhood of 7000 cm.-l, the planar Co+ complex is predicted to be high-spin, even in the absence of axial interactions6 Since the electronic spectrum of [(C~H&AS(CH~)] [ C O ( T D T ) ~is] the same in acetone, acetonitrile and pyridine, axial effects probably are of no importance in this case The M(MNT)2- complexes, which we reported earlier,' on oxidation give M(MNT)z- complexes,s which apparently are composed of M + and M N T radical anion m ~ i e t i e s . ~ Of interest is the reaction of Fez+ and MNT, which gives Fe(MNT)2- directly. The Fe(MNT)2- complex has S = 3/2, with d-d electronic bands a t practically the same energies as planar, high-spin CO(MNT)~*-.Thus Fe(MNT)2- and Fe(TDT)s- are probably the first examples of high-spin Fe+ in a planar situation. It is important t o note that the ligand field strength of MNT is the same as a dianion with d7 Co2+ and as a radical anion with d7 Fe+. (5) Evidence has been presented t h a t the Ni
({'if:)
complex
is planar and contains Nil and two radical anions. The two odd electrons are supposedly paired in the a-molecular system of the complex; see G. N. Schrauzer and V. Maywea. The . . J. A m . Chem. Sac., 843 3221 (1962). +
)m : pelx
can be reduced t o h'i
(";/..)
:which
\S/'\Ph 2 has S = 1/z and shows a highly anisotropic e.s.r. spectrum: see ref. 8. The reduced complex may be electronically similar t o N ( T D T ) z - . (6) C. J. Ballhausen and A. D. Liehr, i b i d . , 81, 538 (1959). suggested t h a t high-spin, rigorously planar ds Nil is possible for a AI less than 10,000 cm. (7) The RII(MNT!zZ- complexes have been shown t o consist of Ma+ and +
R. Williams, I. Bernal and E. Billig, i b i d . , 8 4 , 3596 (1962). ( 8 ) For M = Ni, Pd and P t , see A. Davison, N. Edelstein, K. H. Holm and A, H. Maki, i b i d . , 85, 2029 (1963). We thank Dr. R. H. Holm for allowing us t o see a preprint of these results prior t o publication. (9) Thus for the coplanar ring system, a slable eleclronic siluolion exists for eilher ten or twelve eleclrons in the s-molecular orbitals composed of the nine p ( a ) valence orbitals. We are now carrying out a complete molecular orbital analysis of these systems.
DEPARTMENT OF CHEMISTRY COLUMBIA UNIVERSITY NEWYORK27, NEWYORK RECEIVED APRIL 23, 1963
HARRY B. GRAY E. BILLIG
An Unusual Boron Exchange Reaction' Sir: We wish t o report an unprecedented type of exchange reaction involving complete isotopic substitution in a Blo framework. Previous studies of exchange and interconversion processes in boron hydrides and boron hydride ions containing ten or more boron atoms have concentrated almost exclusively2 on exchange and tracer experiments using deuterium. Deuterated decaboranes have been prepared specifically labeled in a variety of positions by exchange of (1) Interconversion of Boranes. V I I . For paper V I of this series see D. F. Gaines, R. Schaeffer and F. Tebhe, I n o r g . Chem., 2 , 526 (1963). (2) Absence of boron exchange between LOB labeled diborane, 'OBzHa, and normal decaborane, BloHu, has been demonstrated: I. Shapiro and R . E. Williams, J . A m . Chem. Sac., 81, 4787 (1959).
Vol. 85
decaborane with d i b ~ r a n e - d ~ deuterium ,~ ~ x i d e , or ~!~ deuterium chloride,8and by reaction of the monosodium salt of decaborane, NaBloH13, with deuterium ~ h l o r i d e . ~ Although within the limits of these experiments decaborane and the closely related salt NaBloHl~have appeared immune t o framework disruptions, the initial results in the present study of skeletal rearrangements in the higher boron hydrides indicate that diborane undergoes complete boron interchange with NaBloHla in ethereal solution. A sample of NaBloH13was prepared4,' by reaction of decaborane (0.778 mmole) of normal (20% log) isotopic content with sodium hydride in diethyl ether solution. Hydrogen evolution according t o the equation BloH14
+ NaH -+NaB10Hla + Hz
(1)
was complete (97.3% of calculated) within 25 min. a t room temperature. The solution was filtered under vacuum t o give a pale yellow filtrate. Diborane, 96.470 O ' B (4.58 mmoles), was condensed into the flask and the coiitents were warmed t o 0 ' and shaken for 1 hr., during which time no hydrogen formed. By mass spectrometric analysis the loB content of the recovered diborane was shown t o have decreased t o 62.7%, in close agreement with the value calculated (61.9% log, taking into account the incomplete hydrogen evolution for (1)) for diborane exchange with all ten borons of the NaBloH13. Boron-10 enriched decaborane was recovered by treatment of the solution with hydrogen c h l ~ r i d eaccording ~~~ t o the equation NaBloHlt
+ HC1+
BloHIA
+ NaCl
(2)
With suitable recorder amplitude variation, the llB n.m.r. spectrum of the loB enriched decaborane w;bs nearly coincident with an isotopically normal decaborane spectrum, in agreement with the mass spectral determination of participation of all boron positions. Although the NaBloH13structure has not been determined, it is probable, from the n.m.r. spectrums and the ready interconversion of decaborane and the salt by (1) and (2) (presumably involving primary removal and replacement of a bridge proton4), t h a t the boron skeleton is similar t o that of decaborane. With the implication that, in the absence of diborane, the boron framework of NaBloH13 undergoes neither dissociative nor internal rearrangement,4 i t therefore seems more probable that in loBexchange the borane-lOB (lOBH3) adds to the normal BIOH13-, forming an exchange intermediate of over-all composition BllH16-, capable of losing an isotopically normal borane to regenerate B10H13- containing the labeled boron in a position different from that originally occupied by the leaving boron. Considering only the boron framework, a model meeting these requirements can be formed by adding a boron t o a decaborane-like cage between the 5 and 10 positions9 t o produce an icosahedral fragment with 11 of the 12 positions occupied, similar to the boron structure proposed10 for the known" BllH14- ion. If the labeled boron is capable of losing its identity among the five top borons, perhaps by tautomerism of the ten (3) J . J. Kaufman and W. S. Koski, i b i d . , 78, 5774 (1956). (4) J . J. Miller and M. F. Hawthorne, ibid., 81, 4501 (1959). (5) I. Shapiro, M. Lustig and R. E. Williams, i b i d . , 81, 838 (1959). (6) (a) M . F. Hawthorne and J. J. Miller, ibid., 80, 754 (1958); (h) J. A. Dupont and M. F. Hawthorne, i b i d . , 8 4 , 1804 (1962). (7) W. V. Hough and L. J. Edwards, Abstracts of Papers, 133rd Kational Meeting of the American Chemical Society, San Francisco, Calif., 1958, P. 28-L. (8) P;.J. Blay, K. J. Pace and R. I.. Williams, J . Chem. SOL.,3416 (1962). (9) The numbering system used is reproduced in ref. 5 . (10) E. B. Moore, L. L. Lohr and W. N. Lipscomb, J . Chem. P h y s . , 36, 1329 (1961). (11) V. D. Aftandilian, H . C. Miller, G. W. Parshall and E. L. Muetterties, Inorg. Chem., 1, 734 (1962).
