Вы находитесь на странице: 1из 5

Highlights

Bond Theory
DOI: 10.1002/anie.200504322

To Boldly Pass the MetalMetal Quadruple Bond


Udo Radius* and Frank Breher*
Keywords: arene ligands bond theory metalmetal interactions multiple bonds Dedicated to Professor Hansgeorg Schnckel on the occasion of his 65th birthday

The

theory of chemical bonding is intriguing and of fundamental importance. Bonding concepts were developed and rules were established over the years to describe and subdivide molecular compounds consisting of either single, double, or triple bonds. The latter was believed to be the highest accessible bond order for a long time. Although compounds that contain quadruple bonds between transition metals were already prepared in the 19th century, for example [Cr2(mO2CMe)4(H2O)2]),[1] it was not until 1964, when Cotton et al. reported on the crystal structure of K2[Re2Cl8]2 H2O featuring a surprisingly short ReRe distance of 2.24 ,[2] that such a quadruple bond between two transition-metal atoms was confirmed unequivocally. The [Re2Cl8]2 ion has become the prototype for this type of complexes, and a new era of inorganic chemistry and a rich chemistry has evolved around this class of transition-metal complexes.[3] The well-known qualitative description of a s2p4d2 quadruple bond is elegant and deceptively simple (and familiar), but it is based on a oneelectron model and the inherent assumption that four bonding orbitals are doubly occupied. Today we know that this is not entirely true for weak intermetallic bonds. The small overlap between d orbitals results in a relatively weak interaction, thus making a simple

[*] Priv.-Doz. Dr. U. Radius, Prof. Dr. F. Breher Institut fr Anorganische Chemie Universitt Karlsruhe (TH) Engesserstrasse 15 76131 Karlsruhe (Germany) Fax: (+ 49) 721-608-8440 E-mail: radius@aoc1.uni-karlsruhe.de breher@aoc1.uni-karlsruhe.de

molecular orbital (MO) description of the quadruple bond inappropriate.[4] A more accurate, but less transparent description of the quadruple bond requires an approach that goes beyond single configuration methods inherent to simple MO formalisms. To properly describe systems already in the ground state, treatments must include correlation effects. A bond order analysis of [Re2Cl8]2 relying on CASSCF (complete active-space self-consistent field) calculations,[5] for example, has shown, that the ReRe bond has an effective (calculated) bond order of 3.2 and the net bond order contribution of the d bond is about 0.5. The reason is mainly a partial occupation of the antibonding d* orbitals. Whether it should be called a triple or a quadruple bond or whether the d bond should be called a weak bond or half a bond is more or less a matter of definition of the term bond or bond order. The alternatives here are to describe the bonding as a weak quadruple bond (four orbitals with relevant overlap) or as a bond involving four electron pairs with an effective bond order of about three. Various orbitals contribute to a different extent to metalmetal bonding, as demonstrated by intriguing compounds such as the silicon analogue of an alkyne, RSiSiR, reported by Sekiguchi et al. or Wiberg et al.[6] and the corresponding germanium, tin, and lead compounds REER (E = GePb) comprising bulky aryl ligands (Ar and Ar*)[7] described not long before by Power et al.[8] For the REER compounds, the term alkyne analogue does not imply that each of the three valences available for the Group 14 element contribute equally to chemical bonding to retain a triple bond featuring an integer bond order of three. As a result of the

gradually increasing, nonlinear, transbent geometries on descending the group, a considerable weakening of one component of the degenerate p bonding was suspected. Although the structural data, as well as quantumchemical calculations point towards bond orders approaching three for Si, approximately two for Ge and Sn, and one for Pb, the correlation of bond length and bond order is questionable and should be treated with care, especially when counterions are involved. This has led to controversial discussions in the past, for example, whether [Ar*GaGaAr*]2 (formally isoelectronic to neutral RGeGeR) is the first experimentally proven galliumgallium triple bond or not.[9] Although not capable of distinguishing between covalent and electrostatic contributions, the best experimental tool for a classification of such bonds might be the force constant of the metalmetal bond.[10] A milestone for multiple-bond chemistry involving transition metals was recently reported by Power and co-workers, who impressively succeeded in isolating a stable compound with fivefold bonding between two chromium(i) centers.[11] To achieve the highest bond order possible in an isolable compound, the number of ligands has to be minimized, since their binding reduces the number of valence orbitals available to form metalmetal bonds. Furthermore, the steric requirements for the ligand system is of crucial importance since large ligands prevent oligomerization and undesirable bridging motifs are usually disfavored for steric reasons. Relying on their background in main group chemistry,[8] Power et al. have synthesized and characterized a dinuclear metalmetal-bonded complex with one ligand per metal atom in the ligand
Angew. Chem. Int. Ed. 2006, 45, 3006 3010

3006

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angewandte
sphere. The reduction of [{Cr(mCl)Ar}2][7] with a slight excess of potassium graphite afforded the thermally robust complex [ArCrCrAr] (1) in 41 % yield. Following the simplified picture developed for [Re2Cl8]2, the bonding in a (hypothetical) linear complex [RCrCrR] (R = monoanionic ligand) can be described as a quintuple (tenelectrontwo-center) CrCr bond formed by a fivefold overlap between metal d orbitals (Figure 1). Five electron core geometry with a CrCr bond length of 1.8351(4) . The chromiumchromium distance is slightly longer than the bond length reported for a quadruply bonded, but ligand-bridged CrII dimer [Cr2(m2-OMe-5-MeC6H3)4] (1.828(2) ), which has the shortest metalmetal bond distance observed so far.[12] The atoms C1, Cr1, Cr1A, and C1A of the central Cr2C2 unit in 1 are aligned in a plane, but deviate significantly from linearity (angle Cr1A-Cr1-C1 102.78(3)8), adopting a trans-bent structure. Each chromium atom is bonded to the ipso-carbon atom (Cr1C1 2.131(1) ) of the Ar ligand and through a somewhat weaker interaction (Cr1C7A 2.2943(9) ) to an ipso-carbon atom of a flanking dipp ring (dipp = C6H3-2,6-iPr2) of the ligand. The interaction of the dipp moiety with the chromium atom can also be considered as that with a distorted h6-coordinated arene ligand; the chromiumcarbon distances range from 2.29 to 2.97 . The simple bonding model given above, therefore, has to be adjusted to the geometry observed. Additionally, mixing of orbitals can occur due to the lower symmetry of the complex. DFT single-point calculations on the experimentally verified structure presented by the authors support the view that there are five orbital interactions (one s, two p, and two d) between the CrI ions (Figure 3). HOMO and HOMO1, which differ in energy by 0.41 eV, correspond to d bonds, and the LUMO, which is 2.01 eV higher in energy than the HOMO, corresponds to a d* orbital. The chromiumchromium s-binding orbital (HOMO2) that emerges from the single-point DFT calculations on 1 is comparatively high in energy, above the two p-binding orbitals (HOMO3 and HOMO4). This might be due to significant orbital mixing and/or due to reduced overlap along the chromium chromium axis. Although the authors provided first results on CASSCF calculations in the Supplementary Material of the original article, the meaning of this feature for metalmetal bonding remains to be evaluated in a more thorough theoretical treatment of 1 or models of this compound. Before entering into a further discussion of the bonding in complex 1, it is most instructive to briefly outline the

Chemie

Figure 3. Electron density surfaces and energies for the CrCr frontier orbitals in [ArCrCrAr] (1)[11a] (reprinted with permission from Science 2005, 310, 844. Copyright 2005 AAAS).

Figure 1. Schematic MO picture for linear [RCrCr-R].

pairs play a dominant role in holding the metal atoms together, but it does not necessarily imply that the bond order is five or that the bonding is very strong. As pointed out for [Re2Cl8]2, the ground state of the molecule possibly involves mixing of other higher energy configurations with less bonding character. The molecular structure of [ArCrCrAr] (1) (Figure 2) reveals a planar

Figure 2. Molecular structure and numbering scheme of [ArCrCrAr] (1).


Angew. Chem. Int. Ed. 2006, 45, 3006 3010

bonding situation in the dichromium dimer, Cr2, which has attracted considerable interest in recent years.[13] By continuing with the concept of minimizing the number of metal ligands to maximize the number of free metal valence orbitals available to form metalmetal bonds to an extreme, even larger bond orders than five should be feasible. The bare dimers of the open dshell transition metals provide an opportunity to examine multiple bonds between metal atoms in the absence of ligand effects. Among the dimers of the first transition series, Cr2 potentially provides one of the most intriguing examples of multiple metalmetal bonding. Since the chromium atom has a high-spin 7S ground state (a (3d)5(4s)1 valence electron configuration with one electron in each of six valence orbitals), closely spaced energy levels result. The spin pairing of two ground-state atoms results in a 1Sg+ Cr2 molecule with a valence electron configuration (4ss)2(3ds)2(3dp)4(3dd)4 comprising two s, two p, and two d bonds giving formal chromiumchromium bond order of six. The complexity of the bonding in Cr2, however, was for a long time a challenge for ab initio quantum chemistry because of this hextuple bond and the unusual shape of the Cr2 potential energy curve (Figure 4). www.angewandte.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

3007

Highlights
tion between antiferromagnetic coupling and bonding, however, is in this case not clearly defined. As with Cr2, the chromiumchromium bond in the complex 1 presented by Power et al. might alternatively be described as a quadruple bond with two antiferromagnetically coupled electrons residing in chromium-localized orbitals. Magnetic measurements on [ArCrCrAr] revealed a temperature-independent weak paramagnetism of 0.000112(5) emu per mol Cr, which is consistent with strongly coupled paired electrons (S = 0) and an first excited state (S = 1) relatively high in energy, without a significant population of S > 0 states at room temperature. The most common metrics used to gauge the quality of calculations of the electronic structure of quadruple bonds are the geometry-optimized metalmetal bond length and the dd* excitation energy. The electronic absorption spectrum of [ArCrCrAr] displays a broad absorption at 488 nm, which lies in the range observed for dd* transitions of compounds with metalmetal quadruple bonds. Without doubt, the analysis of the bonding situation in [ArCrCrAr] as well as those for the iron and cobalt analogues, briefly mentioned by Power and co-workers, will be an interesting topic among theoreticians. It is very interesting to note that the structurally related [ArFeFeAr] (+ 4 electrons) and [ArCoCoAr] (+ 6 electrons) dimers have much longer metalmetal contacts (2.53 (FeFe) and 2.80 (CoCo)). Assuming the classical bonding picture, this would imply an increase of the bond length of approximately 0.7 due to occupation of two antibonding d* orbitals. A bonding distance as long as 2.53 would result for a formal iron iron triple bond! Going one step further to f block elements, the uranium atom holds, similarly to a chromium atom, six electrons in its valence shell. However, whereas chromium has exactly six valence orbitals, there are 16 such orbitals close in energy available for uranium; that is the seven 5f, five 6d, one 7s, and three 7p orbitals. Gagliardi and Roos recently presented calculations on hypothetical diuranium U2 using CASSCF calculations.[14a] As the authors stated, the maximum flexibility for describing electronic structures and the capability of handling arbitrary spin as offered by the CASSCF method is important, because we cannot assume anything concerning the final number of paired electrons in U2. Indeed, when considering heavy actinide elements, metalmetal bonds proved to be particularly complicated. Although the ground-state electronic configuration of an uranium atom is (5f)3(6d)1(7s)2 (quintet ground state), the energy cost of unpairing the 7s electrons is low, revealing in principle six electrons available for each U atom to form chemical bonds. The bonding situation for uranium, however, is considerably more complex than the situation found for Cr2in fact unique, as reported by Gagliardi and Roos.[14a] Their calculations revealed that three (two-electrontwo-center) electron pair bonds are formed by hybrid orbitals with predominantly 7s and 6d character, that is one s-type and a degenerate set of p-type orbitals (Figure 5, first row). Furthermore, two singly occupied orbitals of s-type (6dsg) and of d-type (6ddg), which show mainly d-orbital character, give rise to two oneelectrontwo-center bonds between the two uranium atoms. Two singly occupied orbitals of d- and p-type (5fdg/5fpu and 5fdu/5fpg, respectively), with predominantly f character, form two additional one-electron bonds. Finally, two electrons occupy a localized 5f orbital (5ffu and 5ffg) equally distributed over both uranium atoms with the electron spins aligned parallel. To summarize, the bonding in U2 arises from three twoelectrontwo-center electron pair bonds (s + 2 p), four one-electrontwo-center bonds (s + p + 2 d), and two localized electrons. For U2, all single spins are predicted to be parallel (ferromagnetically coupled) and the S = 3 septet state is the ground state of the molecule.[15] The calculated (CASPT2-SO, including spin-orbit coupling, SO) equilibrium bond length of (2.43 0.05) and a harmonic vibrational frequency of 265 cm1 suggests comparability of the strength of the U2 bond to that of other multiple bonds between transition metals. Overall, the unprecedented ground state of U2 is expressed as s2p4s1d1d1p1 f1 f1 (with f1 = localized 5f orbital, S = 3, L = 11, ground state
Angew. Chem. Int. Ed. 2006, 45, 3006 3010

Figure 4. Experimental and calculated potential energy curves for Cr2 using CASSCF and CASPT2 according to Roos[13d] (reproduced from reference [13d] with permission of the Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic).

At a short internuclear CrCr distance (re = 1.6788 ), the molecule is multiply bonded and exhibits a relatively deep potential. As the atoms separate, the rapid loss of bonding between the compact 3d orbitals is partially compensated by increased interatomic exchange stabilization, as well as by 4s4s bonding. As a result, the potential curve rises more slowly with increased internuclear separation than would be expected, which renders the curve highly anharmonic with a shelf or a shallow minimum at intermediate chromiumchromium separations. The nature of the Cr Cr interaction changes qualitatively with increased internuclear separation, from multiple 3d3d bonding at short distances to single 4s4s bonding (with the 3d electron cores of the two atoms antiferromagnetically coupled to give a singlet state) at long distances. The Cr2 potential provides a beautiful illustration of this phenomenon. Using highly correlated CASSCF and CASPT2 calculations, Anderson and Roos et al.[13dh] gave a qualitatively correct description of the Cr2 potential energy surface and calculated a total chromiumchromium bond order of 4.4 at equilibrium distance. Accordingly, the chromiumchromium bond consists of one s bond of primarily 4s and 3d contributions, two p bonds totally 3d in character, two d bonds totally 3d in character, and one antiferromagnetically coupled electron pair. This assignment avoids the rather counterintuitive description of two chromiumchromium s bonds. The distinc-

3008

www.angewandte.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angewandte
which is in fact more or less a matter of definitionnot trivial if reasonable at all. The decrease of orbital overlap and increase of nonbonding lone-pair character for molecules of multiply bonded main group elements, and hence the deviation from planarity (for R2EER2) or linearity (REER) has shown that the concepts developed for elements of the second period are certainly not appropriate to describe the nature of an elementelement multiple bond. The involvement of d orbitals certainly complicates the situation due to moderate overlap for d-type d orbitals. The work of Power et al., in unraveling the quintuple bond in trans-bent [ArCrCrAr] (1), has added a further dimension to the concept of multiple bonding in transition-metal chemistry. It is likely, however, that the bonding situation in [ArCrCrAr] is not fully understood and that the trans-bending of this complex will add an additional level of complexity to the analysis of bonding. The current state of the bond description for 1 is that five pairs of overlapping orbitals are more or less involved in metalmetal bonding; a situation that chemists usually would describe as a quintuple bond.[16] A detailed theoretical analysis of the nature of the chemical bond in 1 will certainly be undertaken in the future and it is likely that there will be a renewed debate about metalmetal multiple bonding. To fully understand the bonding in this complex is of crucial significance and many fundamental questions remain to be answered. Whatever the conclusion turns out to be regarding the bond multiplicity in the newly synthesized compound [ArCrCrAr], this work undoubtedly inspires scientists from a theoretical and experimental point of view. The greatest achievement here is the preparative work, which has opened a new area previously considered nonexistent. This work will certainly encourage others to investigate transitionmetal complexes bearing ultralarge ligands in more detail. The combination of both the isolation of further compounds of the type [RMMR] on the one hand, and their precise characterization on the other, may facilitate an accurate experimental assessment of the bonding situation and the bond strength in these www.angewandte.org

Chemie

Figure 5. The molecular orbitals forming the chemical bond between two uranium atoms in U2. Orbital labels are given below each orbital, together with the number of electrons occupying this orbital or pair of orbitals in the case of degeneracy[14a] (reprinted by permission from Macmillan Publishers Ltd: Nature 2005, 433, 848, copyright 2005).

7 1114), and is more complex than any other known diatomic bond. If two of the twelve electrons of U2 were removed, some simplifications of the electronic structure are predicted.[14b] Quantum-chemical calculations, based on multiconfigurational wave functions and including relativistic effects, show that the U22+ system has a large number of low-lying electronic states with S = 02 and L ranging from zero to ten. A bond length of approximately 2.30 (cf. 2.43 for neutral U2) is common for these states. The lowest electronic state corresponds to an electron configuration (sg)2(pu)4 and suggests a triple bond. The s orbital is a hybrid comprising 7s, 6ds, and 5fs atomic orbitals, and the pu orbitals are mainly 6d in character. The next four electrons, two localized on each of the uranium atoms, occupy 5fd and 5ff orbitals, which are essentially nonbonding. Recently, a related theoretical paper presented model calculations on linear singlet [HThThH], which should be a likely candidate for the so far unknown multiple, in this particular case triple, bond between f elements.[14c] The orbital picture and the bonding analyses suggest substantial f character in the ThTh

bond in linear 1Sg-[HThThH] bonding orbitals. The molecular orbitals of this compound correspond to s(ThH) bonds, one s(ThTh) bond (HOMO1), and a double p(ThTh) bond (HOMO) featuring up to 22 % f character in the bond. According to the calculations the f-orbital participation stabilizes the linear geometry of the molecule. The model of the two-electrontwocenter bond, as introduced by G. N. Lewis in 1916, which features a single bond to be formed by one pair of electrons, is one of the most important concepts in chemistry. This also covers multiple bonds since they are regarded as composed of two, three or four twoelectron components. The last few years have witnessed an in depth discussion on multiple bonding between main group elements, which has revealed that simple concepts mainly emerging from the peculiarities of carbon (or the second period of elements) chemistry are not valid for heavier elements. Apparently, the models, which are consistent and clear for the lighter main group elements, become considerably more complicated than anticipated when applied to their heavier counterparts. The same holds true for the term bond order,

Angew. Chem. Int. Ed. 2006, 45, 3006 3010

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

3009

Highlights
types of complexes. The main objective for experimental as well as theoretical chemists is undoubtedly to work together to develop an understanding of this unusual bonding situation between dblock metalsand similar compounds comprising f-block metals are not out of reach!
bond with a high bond order, which is comparable to that of RSiSiR: g) S. Brownridge, T. S. Cameron, H. Du, C. Knapp, R. Kppe, J. Passmore, J. M. Rautiainen, H. Schnckel, Inorg. Chem. 2005, 44, 1660; h) S. K. Ritter, Chem. Eng. News 2005, 83, 49. [7] Ar: -C6H3-2,6-dipp2 (dipp = C6H3-2,6iPr2); Ar*: -C6H3-2,6-trip2 (trip = C6H22,4,6-iPr3). [8] a) M. Stender, A. D. Phillips, R. J. Wright, P. P. Power, Angew. Chem. 2002, 114, 1863; Angew. Chem. Int. Ed. 2002, 41, 1785; b) A. D. Phillips, R. J. Wright, M. M. Olmstead, P. P. Power, J. Am. Chem. Soc. 2002, 124, 5930; c) L. Pu, B. Twamley, P. P. Power, J. Am. Chem. Soc. 2000, 122, 3524; d) P. P. Power, Appl. Organomet. Chem. 2005, 19, 488; e) C. Cui, M. M. Olmstead, J. C. Fettinger, G. H. Spikes, P. P. Power, J. Am. Chem. Soc. 2005, 127, 17530; for selected reviews see: f) P. P. Power, Chem. Rev. 1999, 99, 3463; g) M. Weidenbruch, Organometallics 2003, 22, 4348; h) P. P. Power, Chem. Commun. 2003, 2091. [9] a) Y. Xie, R. S. Grev, J. Gu, H. F. Schaefer, P. v. R. Schleyer, J. Su, X.-W. Li, G. H. Robinson, J. Am. Chem. Soc. 1998, 120, 3773; b) G. H. Robinson, Chem. Commun. 2002, 2175; c) F. A. Cotton, A. H. Cowley, X. Feng, J. Am. Chem. Soc. 1998, 120, 1795; d) M. M. Olmstead, R. S. Simons, P. P. Power, J. Am. Chem. Soc. 1997, 119, 11705; e) T. L. Allen, W. H. Fink, P. P. Power, J. Chem. Soc. Dalton Trans. 2000, 407; f) N. J. Hardman, R. J. Wright, A. D. Phillips, P. P. Power, J. Am. Chem. Soc. 2003, 125, 2667; g) R. Ponec, G. Yuzhakov, X. Girons, G. Frenking, Organometallics 2004, 23, 1790; h) J. Grunenberg, N. Goldberg, J. Am. Chem. Soc. 2000, 122, 6045; i) N. Takagi, M. W. Schmidt, S. Nagase, Organometallics 2001, 20, 1646. [10] See for example: a) R. Kppe, H. Schnckel, Z. Anorg. Allg. Chem. 2000, 626, 1095; b) see also ref. [9h]. [11] a) T. Nguyen, A. D. Sutton, M. Brynda, J. C. Fettinger, G. J. Long, P. P. Power, Science 2005, 310, 844; b) G. Frenking, Science 2005, 310, 796. [12] F. A. Cotton, S. A. Koch, M. Millar, Inorg. Chem. 1978, 17, 2084. [13] For the potential energy curve of Cr2 see: a) S. M. Casey, D. G. Leopold, J. Phys. Chem. 1993, 97, 816; for theoretical calculations on Cr2 see for example: b) N. E. Schultz, Y. Zhao, D. G. Truhlar, J. Phys. Chem. A 2005, 109, 4388; c) E. A. Baudreaux, E. Baxter, Int. J. Quantum Chem. 2004, 100, 1170; d) B. O. Roos, Collect. Czech. Chem. Commun. 2003, 99, 265; e) G. L. Gutsev, C. W. Bauschlicher, Jr., J. Phys. Chem. A 2003, 107, 4755; f) B. O. Roos, K. Anderson, Chem. Phys. Lett. 1995, 245, 215; g) K. Anderson, Chem. Phys. Lett. 1995, 237, 212; h) K. Anderson, B. O. Roos, P.. Malmqvist, P.-O. Widmark, Chem. Phys. Lett. 1994, 230, 391; i) M. M. Goodgame, W. A. Goddard III, Phys. Rev. Lett. 1985, 54, 661. [14] a) L. Gagliardi, B. Roos, Nature 2005, 433, 848; b) L. Gagliardi, P. Pyykk, B. O. Roos, Phys. Chem. Chem. Phys. 2005, 7, 2415; c) M. Stratka, P. Pyykk, J. Am. Chem. Soc. 2005, 127, 13090. [15] Note the difference to Cr2 featuring antiferromagnetic coupling, which causes spin-pairing of the electrons and usually provides some additional small bonding contributions and some extra stabilization. In the particular case of U2, this ferromagnetic coupling can be attributed to favorable exchange stabilization, that is interaction between nonbonding 5f electrons and one-electron bonds. [16] Bonding pictures might also be totally unexpected, as experienced in the case of U2 ; see also: M.-M. Rohmer, M. Bnard, Chem. Soc. Rev. 2001, 30, 340

[1] E. Peligot, C. R. Hebd. Seances Acad. Sci. 1844, 19, 609. [2] a) F. A. Cotton, N. F. Curtis, C. B. Harris, B. F. G. Johnson, S. J. Lippard, J. T. Mague, W. R. Robinson, J. S. Wood, Science 1964, 145, 1305; b) F. A. Cotton, N. F. Curtis, B. F. G. Johnson, W. R. Robinson, Inorg. Chem. 1965, 4, 326; c) F. A. Cotton, C. B. Harris, Inorg. Chem. 1965, 4, 330; d) F. A. Cotton, Inorg. Chem. 1965, 4, 334. [3] F. A. Cotton, L. A. Murillo, R. A. Walton, Multiple Bonds Between Metal Atoms, 3rd ed., Springer, Berlin, 2005. [4] F. A. Cotton, D. G. Nocera, Acc. Chem. Res. 2000, 33, 483. [5] a) L. Gagliardi, B. O. Roos, Inorg. Chem. 2003, 42, 1599; b) F. Ferrante, L. Gagliardi, B. E. Bursten, A. P. Sattlberger, Inorg. Chem. 2005, 44, 8476. [6] a) A. Sekiguchi, R. Kinjo, M. Ichinohe, Science 2004, 305, 1755; b) N. Wiberg, W. Niedermayer, G. Fischer, H. Nth, M. Suter, Eur. J. Inorg. Chem. 2002, 1066; c) N. Wiberg, S. K. Vasisht, G. Fischer, P. Mayer, Z. Anorg. Allg. Chem. 2004, 630, 1823; d) M. Weidenbruch, Angew. Chem. 2005, 117, 518; Angew. Chem. Int. Ed. 2005, 44, 514 (Highlight); e) M. Weidenbruch in The Chemistry of Organic Silicon Compounds, Vol. 3 (Ed.: Z. Rappoport, Y. Apeloig), Wiley, Chichester, 2001; f) N. Takagi, S. Nagase, Eur. J. Inorg. Chem. 2002, 2775; Recently, Passmore and co-workers impressively succeeded in isolating the S2I42+ cation featuring a sulfur sulfur

3010

www.angewandte.org

 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2006, 45, 3006 3010

Вам также может понравиться