1882
C6H6.A$1A1,. In the former, this asymmetry mani'"est,d itself in statis!ically disordered Ag positions, but it is unambiguous in the latter. It is not likely that molecular packing would be the cause of this asymmeti y in metal-carbon distances, because the the packing is quile different in C,H6.AgC104, C6146' CuA.CI4 and C6H6.AgA1C14. A similar asymmetry in Cu-C distances has been observed in a coFper(1)olefin c ~ m p l e x . ~ ' We suggest that this asymmetry between the closest carbon-to-metal distances is a fundamental property of metal ion-aromatic complexes. Further, we suggest that the reason for this asymmetry is a compromise between the acceptor properties of the metal ion, or coordinated metal ion, and the donor properties of this same ion. That is, if the acceptor orbital were a 4sor an sp3-hybrid orbital, the most advantageous position would be at the point of greatest electron density of the ring, directly above one of the carbon atoms. On the otiiir hand, using inner d orbitals for the metal ion as donor, the most likely position would be above and symmetrically between two carbon atoms of the aromatic system. Hence, a compromise between these
two effects is reached and unequal metal-carbon distances result. However, in platinum- and palladiumolefin complexes, there has been no evidence for different metal-to-carbon distances, but it is quite likely that details of the bonding may be quite different for olefin complexes. Although the bond distances alternate in lengths around the ring and suggest a cyclohexatriene system, the errors are sufficiently large that this variation of bond distances may not be real and caution should be applied to any interpretations based on C-C distances in this complex. It is, in fact, an interesting question as to why the complex forms at all. In the presence of chlorine donors it is surprising tht the metal-aromatic bond is preferred to M-C1 interactions. The answer may be that in the packing of anhydrous CuAIC14 large voids remain, and it becomes energetically favorable to form metal ion-aromatic bonds over metal ion-chlorine bonds. We plan to investigate the crystal structure of anhydrous CuA1C14 in the near future. Acknow edgment. We wish to acknowledge the financial support of NSF Grant GP-1575.
Cage Compounds Containing the Trirhenium( 111) Cluster: Re,Br, ( AsO, ) (DMSO)31 F. Albert Cotton and Stephen J. Lippard2 Contribution f r o m the Department of Chemistry, Massachusetts Institute 03 Technology, Cambridge, Massachusetts 02139. Received January 12, 1966
Abstract: The preparaticm and molecular structure of Re3Br,(As04)2(DMS0)3, a representative member of the new class of compounds Re3X3(M03 &Ln,are discussed. From the visible spectrum, it is apparent that the trirhenium(111) metal atom cluster occurs in this compound. Further insights into its structure are obtained from a detailed analysis of the infrared spectrum. Apparently, two tridentate arsenate ions have replaced the six axial halide ions in the Re3Brsmolecule to form a cage which in-orporates the cluster. The solvent (DMSO) molecules are thought to occupy nmbridging equatorial sites in the cluster.
T
he existence of the trirhenium(II1) metal atom cluster in complexes prepared from rhenium(II1) chloride and rhenium(II1) bromide has been well established. 3-1 Studies by Robinson and Fergusson" have shown that, of the twelve halogen atoms (six (1) Supported by the U. S . Atomic Energy Commission. (2) National Science Foundation Postdoctoral Fellow, 1965-1966. (3) J. A. Bertrand, F. A. Cotton, and W. A. Dollase, J . Am. Chem. Soc., 85, 1349 (1963); Inorg. Chem., 2 , 1106 (1963). (4) W. T. Robinson, J. E. Fergusson, and B. R. Penfold, Proc. Chem. Soc., 116 (1963). (5) J. E. Fergusson, B. R. Penfold, and W . T. Robinson, Nature, 201, 181, (1964). (6) F . A. Cotton and J. T. Mague, Inorg. Chem., 3, 1094 (1964). (7) F. A. Cotton and J . T. Mague, ibid., 3, 1402 (1964). (8) F. A. Cotton and S . J. Lippard, ibid., 4, 59 (1965). (9) F. A. Cotton and S . J. Lippard, J . Am. Chem. Soc., 86, 4497 (1964). (10) F. A. Cotton, S. J. Lippard, and J. T. Mague, Inorg. Chem., 4, 508 (1965). (11) B. H. Robinson and J. E. Fergusson, J . Chem. Soc., 5683 (1964).
Journal of the American Chemical Society
88:9
1 May
5, 1966
axial, three equatorial bridging, and three equatorial nonbridging) in the Re3X123-ion, only three, presumably the equatorial bridging ones, are not subject to exchange with thiocyanate or radioactively labeled halide ions. Apparently, then, the stable smctural unit in the trirhenium(II1) metal atom cluster compounds is the Re3X3group (I).
I From the known geometry ( c f . ref 3-8) and net charge ( + 6 ) of I, it seemed likely to us that anions such
1883
P
H
Figure 1. Perspective view of the proposed structure of the molecule Re8Br3(As04)2(DMS0)3.
as ASO,~-, in which three oxygen atoms define an equilateral triangle of edge -2.5 A, would be ideally suited for forming neutral adducts with I of the type Re3X3(A~04)2.Thus we began an extensive investigation of the reaction of the rhenium(II1) halides, Re3C19and Re3Brg,with the M033- or Mod3- anions, where M = As, P, etc. In this paper, we discuss the preparation and molecular structure of Re3Br3(As04)2(DMSO), as a typical member of the series.
Frequency i n crn.”
x IO3
Figure 2. Visible suectra of compounds thought to contain the trirhenium(II1) metai atom cluster-dissolved in-strong donor solvents: solid line, Re3Br3(As04)Z(DMS0)3 in DMSO; dashed line, Re3Br9(C5H5N)3 in pyridine.
mixture containing silver halide and the desired cagecluster addition compound precipitates. The great insolubility of both products makes their separation quite difficult. Only by using strong donor solvents, Experimental Section such as DMSO, dimethylformamide (DMF), methanol, Preparation of Re3Br3(AsO&(DMS0)3. To the red solution or pyridine, which presumably are capable of “solubiformed by the addition of 1.0 g (0.8 mmole) of rhenium(II1) brolizing” the trirhenium(II1) cluster by reacting to fill one mide to 50 ml of acetone10 was added 1.0 g (-2 mmoles) of silver arsenate. The heterogeneous mixture was vigorously agitated for or more of its nonbridging equatorial positions, can the 4 hr. Subsequent filtration yielded a solid product mixture and a separation be effected. The compound obtained by clear filtrate. The solid was then extracted with 50 ml of dimethyl subsequent reprecipitation from the donor solvent sulfoxide (DMSO) containing a few drops of 48 HBr for 15 min medium by the addition of a second solvent (e.g., on the steam bath (-90”). The resultant mixture was filtered to tetrahydrofuran; vide supra), then, invariably contains give a white solid (AgBr) and a deep red filtrate. To the latter was added ca. 300 ml of tetrahydrofuran. After cooling the solution to bound solvent molecules. 0’ and scratching the walls of the container with a glass rod, a The preparation of R~,B~,(AsO~)~(DMSO), described flocculent red solid separated out and was collected by filtration. in the Experimental Section of this paper, has been The yield, after drying at 100” in vacuo over P205 for 24 hr., was repeated several times in our laboratory, and this 0.5 g (-50%). Anal. Calcd for Re3Br3(As04)2[(CH3)2SO]3: Br, 18.3; C, appears to be the only compound which consistently 5.50; H, 1.38. Found: Br (average of two independent preparacontains three bound solvent molecules. Work now in tions), 18.5; C, 5.78; H, 1.47. progress indicates that others, such as Re3C13(A~04)2Physical Measurements. Infrared spectra were taken with both Re3X3(As03)2(DMSO)z, Re3C13(As03)2the Perkin-Elmer Model 337 and 521 spectrometers on samples (DMF)2, and Re3Br3(P04)2(DMSO)2, have also been obmulled in Nujol or pressed into KBr pellets. Visible spectra were run in DMSO on the Cary 14 spectrophotometer using 1-cm tained in similar reactions and will be reported later. At matched quartz cells. A 114.6-mm Debye-Scherrer rotatingpresent, we intend to show that, as a representative sample powder camera and Cu K a radiation filtered through nickel were used in the attempt to get a powder pattern of Re3Br3(AsO4)2- example of the entire class of compounds Re3X3(MO, or &Ln, the physical properties of Re3Br3(DMS0)a. Computations. Force constants used in the analysis of the in(As04)2(DMS0), are consistent with the structure frared spectra were computed using a computer program‘* which shown in Figure 1. adjusts the constants to give the best least-squares fit of the calPhysical Properties and Molecular Structure of culated frequencies to the measured ones. Details of the calcuiaRe3Br3(As04)2(DMS0)3.X-Ray powder diffraction tion are discussed below. studies indicate that the Re,Br,(AsO4h(DMSO), obtained as described above is amorphous. Although Results and Discussion this is not surprising in view of the method of preparaPreparations. When solutions of rhenium(II1) chlotion, it does preclude for the present a direct singleride or bromide in acetone are vigorously shaken with crystal X-ray structural investigation. On the basis of the silver salt of PO4+, AsOd3-, or ASO,~-, a product infrared and visible spectral information, however, it appears that the structure is probably the one shown in (12) J. H.Schactschneider and R. A. Snyder, Specrrochim. Acto, 19, 117(1963). Figure 1. Cotton, Lippard
I ResBrdAs04)ADMSO)s
1884 0.o
I
Figure 3. The broad band at -925 cm-l is partially due to the S=O stretching frequency of coordinated DMSO, as previously assigned in similar complexes of the rhenium(II1) halides with sulfoxides.1° When the Re3Br3(As04)2 cluster was prepared using hot dimethy!formamide (DMF) or methanol instead of DMSO to separate it from silver bromide (vide supra), however, compounds containing D M F or CH30H instead of DMSO in the nonbridging equatorial positions of the trirhenium(II1) cluster were isolated. The 400-1200cm-' regions of the infrared spectra of these adducts also appear in Figure 3, from which it is clear that a broad band, centered at approximately 900 cm-l, is characteristic of the R e ~ B r ~ ( A s group. 0 ~ ) ~ It can also be m! 1 I I 1 I I I seen from the figure that strong bands occur at 740, '1200 1100 1000 900 800 700 600 500 400 Frequency (ern:') 725,550, and 539 cm-l as well as a weaker band at 572 cm-' (shoulder). These bands are, in all probability, Figure 3. Infrared spectra of Re3Br3(As04)z(DMS0)3 mulled in not due to coordinated solvent, for, as evident from Nujol (solid line) and pressed into a KBr disk (dashed line); and, Figure 3, they occur at the same frequency for L = of Re3Br3(As04)2Ln mulled in Nujol, where L = dimethylformamide (open circles) or methanol (closed circles). DMSO, DMF, and methanol. As further evidence for this point, the infrared spectrum of a Nujol mull of the crude reaction product of Re3Bra and Ag3As04 was Previous experiments in our l a b o r a t o r i e ~ and ~ ~ ~ ~ taken before the addition (vide supra) of a donor solelsewhere' ' have established that, in the visible specvent. Strong bands in the 700-725 and 540-550-cm-' trum, two bands at ca. 12-13,000 and 19-20,000 cm-', regions were observed. That the band at 725 cm-' having relative intensities of 1 :3, may be found for all does not result entirely from Nujol was verified by compounds known or thought to contain the Re3X3 running the spectrum of a sample of Re3Br3(As0& group dissolved in solvents such as chloroform, acepressed into a KBr disk (Figure 3). tone, or acetonitrile. In more strongly donating solNow, let us presume that the structure of Re3Br3vents, such as dimethyl sulfoxide or pyridine, the low(ASO~)~(DMSO), is as shown in Figure 1. It is then a energy band becomes considerably broadened. l 3 The reasonable assumption that only the As-0 and Re-0 visible spectrum of Re3Br3(As04)2(DMS0)3,1.5 X stretches of the ( O A S O ~ ) fragment ~ R ~ ~ will contribute to M in DMSO, appears in Figure 2. For comparithe infrared spectrum in the region above 400 cm-l. son purposes, we have reproduced14 the spectrum of If, in addition, there is no coupling from one-half of the Re3Br,(C,H5N)3 in pyridine. From the similarity of molecule to the other, then, under C3" symmetry, there these spectra, it seems likely that, in the adduct of are seven As-0 or Re-0 oscillators. Using this model, rhenium(II1) bromide and silver arsenate, the Re3Br3 and neglecting Re-0 to Re-0 and Re-0 to As=O inmoiety of the initial Re3Brgremains intact. teractions, a secular equation was constructed and Although the visible spectrum of Re3Br3(As04)z- factored by symmetry (Table I). l5 is perfectly consistent with the structure proAs a first approximation, values of 6.0 mdynes/A for posed in Figure 1, other structures may be drawn which the As=O and 4.0 mdynes/A for the As-0 and Re-0 also contain the Re3Br3group. Thus, for example, a force constants, and 0.5 mdyne/A for the As=O to polymer such as that shown below (11), where the As-0, As-0 to As-0, and As-0 to Re-0 interaction arsenate groups form bridges between neighboring constants were assigned. The secular equation was then trirhenium(II1) clusters, might conceivably occur. solved and the infrared spectrum assigned as follows: v A ' A s = 0 = 900 cm-l; = 740 cm-'; V ~ A ~ - = O 725 cm-'; v A 1 ~ . + -=0 539 cm-'; v E R e - 0 = 550 cm-l. Finally, with these assignments, the three diagonal force constants were adjusted by the least-squares method to give the best fit to the observed frequencies, while the off-diagonal (interaction) constants were fixed at 0.5 mdyne/A.16 The final values of the force constants together with a list of observed and calculated frequenI1 cies, calculated after the last cycle of least-squares refinement, may be found in Table II.16a While a structure such as that shown in I1 seems unlikely on the basis of steric requirements, to lend further (15) See, for example, E. B. Wilson, J. C. Decius, and P. C. Cross, support to the structure proposed (Figure l), we have "Molecular Vibrations," McGraw-Hill Book Co., Inc., New York. undertaken a rather thorough infrared spectral analysis N. Y., 1955. (16) To test the validity of this procedure, we subsequently held the of the compound, the details of which follow. refined diagonal force constants fixed and varied the off-diagonal ones. The 400-1200-cm-* region of the infrared spectrum of They did not vary by more than 1 0 . 1 3 mdyne/A from the assumed value of 0.5 mdyne/A (cf.Table 11). Re3Br3(As04)2(DMS0)3, mulled in Nujol, is shown in 1
1
I
I
I
I
(13) This broadening has been ascribed14 previously to a time-averaged incomplete occupation of the coordination sphere of the rhenium atom cluster. (14) S. J. Lippard, Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, Mass., 1965.
Journal of the American Chemical Society
/ 88:9 1 May 5, 1966
(16a) NOTEADDEDIN PROOF. Since the submission of this article, a paper has appeared in the literature (A. Merijanian and R. Zingaro, Inorg. Chem., 5, 187 (1966)) in which a value of 6.9 mdynes/A is given for the force constant of the "As-0'' bond in several trialkylarsine oxides. This compares reasonably well with our result (6.34 mdynes/A) for the As=O bond in ReaBra(AsOa)z(DMS0)~.
1885 Table I. Information Pertaining to Calculation of MO Stretching Modes (M = Re, As) in OAs03Rea"Half-Molecule" Group of C3, Symmetry. Assumed Molecular Geometry and Definition of Internal Coordinates and Force Constants
Constant
0
~
~~
Symmetry species"
.
Force constants Interaction
~
Factored secular equationsb lFllGl - lElX = 0
Symmetry coordinates
pocos 110"
PO cos 110"
R = Raman active; IR = infrared active. quency in cm-1; see ref 15.
b
PRe
+
PO
lEl = unitarity matrix; p = reciprocal mass; X = 5.889 X 10-'(v2), where Y is the fre-
Table II. Observed and Calculated Frequencies for the Infrared Spectrum (400-1200-cm-1 region) of Re3Br3(As04)n(DMSO)a Mulled in Nujolo
derived from the structure proposed in Figure 1. Only the weak band (shoulder) at 572 cm-' cannot be assigned and, possibly, this may be the overtone of a Calcd,b band at much lower energies. Furthermore, it now Obsd, cm-1 cmAssimment seems rather unlikely that the compound has a structure such as I1 above, in which the low symmetry would be 900 (strong, broad) 899 As=O; At 740 (strong) 760 AS-0; Ai expected to result in a much more complicated spec725 (strong) 708 As-0; E trum in the 400-1200-~m-~region. 572 (weak,sh) ... ? In summary, then, while direct X-ray structural con550 (strong) 547 Re-0; E firmation is, at present, unattainable, spectral studies 539 (strong) 540 R e - 0 ; Ai strongly suggest that Re3Br3(As04)2(DMS0)3 (and, by The S=O stretching frequency at -925 cm-l is not included. inference, probably all complexes in the class Re3X3* Calculated as discussed in the text using values of 6.34, 4.19, (MO, or 4)Ln) has the structure shown in Figure 1. and 2.89 mdynes/A for Fa, Fb, and F,, respectively, and 0.5 mdynel A for all off-diagonal (interaction) force constants. By varying the The visible spectrum verifies that the trirhenium(II1) interaction constants (see footnote 16), a better fit of the A, and E metal atom cluster is still intact and the infrared specAs-0 frequencies could be obtained, the calculated values being 742 trum is consistent with, though it does not positively and 724 cm-', respectively. The final interaction constants were prove, the existence of a high degree of molecular then 0.57, 0.37, and 0.47 mdyne/A for Fab, F b b , and Fbo, respectively (see Table I). symmetry (C3v)for the OAs03Re3fragment. Further investigations of other compounds of this class are in progress and will be reported in due course. From Table 11, it is apparent that the five major bands in the infrared spectrum of Re3Br3(A~04)2- Acknowledgment. We are grateful to the MIT (DMSO)a (Figure 2), exclusive of the S=O stretching Computation Center for access to the IBM 7094 frequency, may be accounted for using a simple model computer. 0
Cotton, Lippard
1 Re&a(AsO&(DMSO)a
