Electronic Spectroscopy and the Photochemistry of Compounds which Contain Quadruple Metal-Metal Bonds Citation Trogler, William Charles (1978) Electronic Spectroscopy and the Photochemistry of Compounds which Contain Quadruple Metal-Metal Bonds. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/nr7a-jz27. https://resolver.caltech.edu/CaltechTHESIS:03092026-183448574 Abstract The electronic spectra and structures of binuclear transition metal compounds which contain a quadruple metal-metal bond are discussed. Experimental results are presented for the electric-dipole forbidden transitions of (n-Bu 4 ) 2 [Re 2 Cl 8 ], (n-Bu 4 ) 2 [Re 2 Br 8 ], Mo 2 (O 2 CCH 3 ) 4 , Mo 2 (O 2 CCD 3 ) 4 , Mo 2 (O 2 CH) 4 , and Mo 2 (O 2 CCF 3 ) 4 . The near infrared absorption system in K 3 Mo 2 (SO 4 ) 4 · 3.5 H 2 O has also been analyzed. Aqueous acid solutions of Mo 2 (SO 4 ) 4 4- , Mo 2 (aq) 4+ , and Mo 2 Cl 8 4- yield molecular hydrogen and oxidized molybdenum products upon ultraviolet irradiation. The compounds Mo 2 Cl 4 P 4 (P = P(n-Bu) 3 , P(Et)3 and P(Et)(Ph) 2 ) undergo photooxidation to chloride bridged binuclear Mo(III) compounds subsequent to 254 nm irradiation in chlorocarbon solvents. The nature of these photooxidation reactions are examined. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Gray, Harry B. Thesis Committee: Unknown, Unknown Defense Date: 31 May 1977 Record Number: CaltechTHESIS:03092026-183448574 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:03092026-183448574 DOI: 10.7907/nr7a-jz27 ORCID: Author ORCID Trogler, William Charles 0000-0001-6098-7685 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 18407 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 10 Mar 2026 20:33 Last Modified: 10 Mar 2026 20:44 Thesis Files PDF - Final Version See Usage Policy. 46MB Repository Staff Only: item control page

ELECTRONIC SPECTROSCOPY AND THE PHOTOCHEMISTRY OF COMPOUNDS WHICH CONTAIN QUADRUPLE METAL-METAL BONDS Thesis by William Charles Trogler In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 91125 1978 (Submitted May 31, 1977) ii ACKNOWLEDGMENTS Research at Caltech has been enjoyable. I thank Professor Harry B. Gray for introducing me to the fascinating field of metal clusterso Most of all, though, I thank him for teaching me the importance of clear and concise logic and for reinforcing my sense of scientific independence. I am proud to have been part of the "Gray Group". Professor Edward Solomon has had a profound influence on my training in inorganic spectroscopy for which I am incredibly grateful. Professor George Hammond has furthered my photochemical education by sharing his insight into photochemical reaction mechanisms. Professor William Goddard and Larry Harding are acknow- ledged for many instructive discussions about the electronic structure of metal cluster compounds. I also thank Professor John Bercaw, Dave Dooley, Dave Erwin, Dro Jeff Hare, Professor Mitsuru Kubota, Dro Kent Mann, Dr. Robert Richman, Professor George Rossman, and Dave Tyler for teaching me new experimental techniques. I sincerely thank Professor Luigi Marzilli for introducing me to the field of inorganic chemistry and for his continued support. I gratefully acknowledge the expert skills of Laura Dagen in the typing of this thesiso Also, I thank her and Sue Brittenham for their aid on many occasions during my tenure at Caltech. iii The other members of the third floor: Bob, Cathy, Dao, D. T. - K. T., Duke, Flip, Nate, Pam, Penny, Porn, and Yo are saluted for those less serious moments. I especially thank Pat for her friendship. Finally, I dedicate this work to my parents. iv ABSTRACT The electronic spectra and structures of binuclear transition metal compounds which contain a quadruple metalmetal bond are discussed. Experimental results are presented for the electric-dipole forbidden transitions of (n-Bu4 )J Re 2Cl8 ] , (n-Bu )J Re Br 2 4 ] 8 and Mo (O2CCF 2 3 ) Mo 2(O2CCH , 4 • 3 ) 4 , MoiO2CCD 3 \, Mo (O CH) 2 2 4 , The near infrared absorption system in KsM0 2 (SO4 ) 4 ° 3. 5 H2O has also been analyzed. Aqueous acid solutions of Mo2(SO4 ) 4 Mo2Cls 4- 4, Mo 2 (aq) '1+ , and yield molecular hydrogen and oxidized molybdenum products upon ultraviolet irradiation. (P = P(Q_-Bu) 3 , The compounds Mo 2 Cl4 P 4 P(Et) and P(Et) (Ph) 2) undergo photooxidation to 3 chloride bridged binuclear Mo(III) compounds subsequent to 254 nm irradiation in chlorocarbon solvents. The nature of these photooxidation reactions are examined. V TABLE OF CONTENTS PAGE CHAPTER 1. ELECTRONIC SPECTRA OF COMPOUNDS WHICH 1 CONTAIN QUADRUPLE METAL-METAL BONDS. I. Electronic Spectra and Structures of Compounds 2 Containing Quadruple Metal-Metal Bonds. An Overview of the Field. IT. 32 Further Studies of the Electronic Spectra of 2- Re2Cla 2- and Re~r 8 Assignment of the Weak • Bands in the 600-350 nm Region. Estimation of the Dissociation Energies of Metal-Metal Quadruple Bonds. III. Emission Spectra and Lifetimes of Re 2Cl 8 Re2Bra 2- , and Mo 2Cl8 of the 6 IV. 4- 2, 52 at 1. 3 K Upon Excitation o* Transition. Studies of the Polarization Behavior, Temperature 65 Dependence, and Vibronic Structure of the 23,000 cm -1 Absorption System in the Electronic Spectra of Mo 2 (O2CCH 3 ) 4 and Related Compounds. Emission Spectrum of Mo 2(O2CCF 3 ) 4 at 1. 3 K. V. Electronic Structure and Spectrum of MoiO2 CH) 4 and Related Compounds. 99 vi PAGE CHAPTER 2. THE PHOTOCHEMISTRY OF COMPOUNDS 112 WHICH CONTAIN A QUADRUPLE MOLYBDENUMMOLYBDENUM BOND. I. Production of Hydrogen by Ultraviolet Irradiation of Mo 2 (S04 ) 4 4- 113 in Aqueous Sulfuric Acid. Electronic Absorption Spectrum of KsMo 2 (S04 ) 4 • 3. 5 H2 0 at 15 K. II. Photochemical Oxidation Reactions of Binuclear Molybdenum(Il) Complexes in Aqueous Acids and in Chlorocarbons. 124 1 CHAPTER 1 ELECTRONIC SPECTRA OF COMPOUNDS WHICH CONTAIN QUADRUPLE METAL-METAL BONDS 2 Electronic s· ectra and Structures of Com ~ . An Overview of the Field Parallel spectroscopic and theoretical studies of diatomic molecules 1 have provided a basis for molecular bonding theory. The electronic structure of many polyatomic molecules 2 may be adequately described by postulating diatomic-like a and 1T bonds betweens and p orbitals on different atomic centers. An attempt to extendthis model to explain metal-metal bonding in transition metal cluster compounds must recognize several new fm tures. The most important is the presence of metal d orbitals in the valence shell. Now the spectroscopic manifestation of strong metal-metal bonding will be explored. Binuclear compounds M2 L8 (M = Mo(II), Re(III) or Tc(III); L = ligand donor atom)of second and third row Group VI and VII B 4 transition metal ions with the d electronic configuration are systems well-suited for examining the consequences of metal-metal bonding. Chemical and structural properties of these compounds are consistent with the presence of an unusually strong metal-metal bond. 3 Both the net positive charge and high atomic numbers of these ions are expected to effect large (n-1) d - (n)s and (n)p orbital energy separations. 4 Therefore, the valence shell of the metal ion approximately consists of partially occupied ct-like orbitals, and consequently the bonding between two metal ions primarily involves the d orbitals. 3 From a more practical viewpoint, the high symmetry (D4 h) of the M2L8 compounds is particularly favorable for spectroscopic investigations. We will now consider the electronic structure of these compounds with a special emphasis on the near infrared-visibleultraviolet absorption bands associated with the M2 chromophore. Simple MO Diagram for d Orbital Splittings The qualitative splitting pattern expected for the d orbitals of a homonuclear diatomic molecule is depicted in Figure lao If the internuclear axis is chosen as the z axis of right and left-handed coordinate systems, respectively for the two atoms, then a, 1T and 6 bonds are generated from the appropriate linear combinations of the (dz 2), (dxz, dyz) and (dxy' dx 2 -y 2) orbitals in that order. Intro2- duction of the eight ligands as for example in Re 2 Cl8 along the x and y a.xis will dramatically raise the energy of the o and 6* orbitals composed of the dxz-yz components (Figure lb). This pair of orbitals is primarily metal-ligand a-antibonding in charactero A similar but less pronounced destabilization would also occur for the a and a* orbitals in a crystal field approach. According to the model Re2Cla 2- would contain a quadruple metal-metal bond and so it was proposed by Cotton 5 to explain the diamagnetism, the eclipsed geometry and the unusually short Re Re distance which is obtained. This qualitative MO diagram suggests that the lowest energy electric-dipole-allowed l transition in such compounds is A1 g - l A 2u (o - o*). Our discussion of electronic spectra will be centered around Re 2Cl8 2- as it has been the most extensively studied of those compounds which contain a quadruple metal-metal bond. 5 For Mo 2Cl 8 4- a z polarized transition is observed at 19,128 cm -1 -1 and at low temperatures a vibrational progression in a 336 cm 1 vibration (a g (MoMo) in ground state is 346 cm - )1 1 is observed. 14 1 4- The compound Mo 2 (SO4 ) 4 exhibits a structureless absorption at -1 15 15 1 1 19,400 cm . From the polarization data a A1 g - A2 u (o o*) assignment is indicated; however, the intensity of this absorp16 tion is rather low with an Emax 170. ~ A confusing characteristic of the o - o* absorption band has been its relatively low intensity. Generally, one-electron N - V excitations of the bonding to antibonding kind are among the most intense molecular transitions. 17 Recently it was suggested 18 that the diminished intensity of the o - o* transition is a consequence 17 of weak orbital overlap. Mulliken has shown that the oscillator strength (f) of a one electron N - V diatomic-like transition in a weak coupling or valence bond formalism is approximated by 11 (f) = 1. 096 X 10 2 2 VS r -1 where v is the transition energy in cm , s is the orbital overlap, and r the bond length in cm. As the o overlap is on the order of O. 1 or less, a valence l>ond description is more appropriate than the MO modeL Employing the valence bond formula, oscillator strengths in the neighborhood of 10 -2 are predicted for typical values or rand v. One further point of interest is the location of the triplet state 3 A2 u or more correctly, the A1 u and Eu spin orbit components of the o - o* excitation. It was originally proposed 19 that a weak structured 6 -1 absorption at 6250 cm not all Mo 2Cl8 4- in the diffuse reflectance spectra of some, but 1 salts was A1 g - 3 Azu. As we shall discuss shortly, oxidized Mo species may exhibit a structured absorption in the 6000-1 8000 cm region. More recently, a single crystal polarized spectral 15 study of ~Mo 2Cla • 2H 2O revealed an (x 7 y) polarized shoulder (ca. -1 1 1 17,500 cm )on the lower energy shoulder of A1 g - A2u (o - o*)o It was suggested that this resulted from a transition to the Eu spin-orbit 3 component of A2u. Subsequently, the emission spectra of Re 2Cl 8 Re2Brs 1 A1 g - 21 , and Mo 2Cl8 2, 4- were recorded at 1. 3 K upon excitation of the 20 A2 u absorption. Lack of overlap between the emission spectrum and the 0-0 line of the o - o* absorption was observed for all three compounds. The emission band shape was similar to that of the 6 - o* ab- sorption and the luminescent lifetimes were all on the order of 100 ns. It was concluded 20 that the emission probably originated from one of 3 1 3 the spin-orbit components of A2 uo Therefore, a A 2 u - A2u separation of 1000 - 3000 cm -1 is indicated. Upon one-electron oxidation or reduction of binuclear group VI or VIIB transition metal compounds, one would expect an electron to be removed from a o bonding orbital or added to a o antibonding orbital, respectively. A series of epr studies by Cotton and Pedersen 21 have shown that this is indeed the case for oxidized species such as Mo2(SO4)4 3- + and Mo2(O2CC 3 I:I., )4 and for the reduced species Tc 2Cl 8 3- Re 2 (O2 CC 6 H5 ) 4 + and Re 2 Cl4 (P(C 2H5 ) 3 ) 4 +. It is now becoming apparent that a near infrared electronic absorption band is characteristic of these single electron reduction and oxidation products. Ebner and , 7 Walton 22 have noted the presence of an absorption at 7100 cm - l (1400 nm) with a molar extinction coefficient greater than 2000 22 for Re~ 5 (PRPh 2)3 ( X = Cl or Br). Similar behavior obtains for [ReC1 (Ph PCH CH2PPH )] and [ReCl (Ph2PCH PPh )]n. 23 3 2 2 2 2 Another reduced species Tc 2 Cl8 3 2 2 3- has been studied and an absorp-1 24 tion at 6200 cm (Emax 630) was observed. At low temperatures, a well resolved 3 20 cm -1 vibrational progression is detected. An absorption at 7100 cm -1 (Emax 143) was also observed in the electronic spectrum of the oxidized species Mo 2(SO4 ) 4 low temperatures the band 1 progressions of 350 cm - • of a 1 g (MoMo) for Mo2(SO4 ) 4 3• 16 At displays well resolved vibrational For comparison, the ground state value 25 3- -1 is ca. 3 80 cm . On the basis of the sharp molecular vibronic structure observed for Moz(SO4 ) 4 3- and the lack of any observable splitting of Mo signals in the X-ray photoelectron spectrum, a 6 - 6* assignment for the near infrared absorption system rather than an in.tervalance charge transfer assignment was 16 proposed. A 6 - 6* assignment was also suggested 24 for Tc2Cl8 3- with the aid of an Xa calculation. An analogous assignment therefore appears likely for the compounds examined by Walton, et al. 22 , 23 It is interesting to note that the Emax of the near infrared absorption is similar to that of the 6 - 5* absorption in corresponding compounds which possess a full metal-metal quadruple bond. a o - 6* assignment for the 13,000 cm -1 These results suggest (Emax 2600) absorption of Mo 2 (O2 CCF 3 ) 4 - rather than the metal to ligand charge transfer (MLCT) assignment which was originally suggested. 26 Although this absorption 8 -1 is at higher energy (ca 6000 cm ) than in the other examples cited, the position of the o - 6* absorption of the unreduced 1 compound is also (ca 15,000 cm- ) at an increased energy. 27 Electronic structure simulations have been attempted employing a variety of theoretical techniques. The earliest 2- studies of Re 2 Cl8 and Mo 2 (O2 CCH 3 ) 4 were of the Extended Ruckel type. 28 , 29 Although these calculations qualitatively accounted for the bonding features, they we re inadequate for a description of the excited states. In particular, the lowest unoccupied virtual orbitals were incorrectly predicted to be a- nonbonding orbitals which were localized along the metal-metal axis. New impetus on the theoretical front was provided when Mortola, Moskowitz, and Rosch 30 applied the multiple scattering Xa method to Re 2 Clg 1 2- . Although underestimating -1 the energy of the A 2 u state by some 10,000 cm , .the qualitative ordering of the excited states was consistent with experiment. 6 ' 1 O 431 Calculations of the SCF-Xa SW type for Mo 2 Cl8 supported the 10 -1 earlier o - 5* assignment for the 19, 128 cm absorptiono Once again, the energy 0£ the transition was underestimated by 10,000 cm -1 0 For MoiO2 CH) 4 SCF-Xa-SW calculations incorrectly predict the 6 32 6* excitation to be the lowest excited singlet stateo Also, the energy 1 1 of the A2u state again was underestimated by some 15,000 cm - • An accurate prediction of a 6 - 6* excitation energy was recently accomplished, 24 3 with an Xa calculation of Tc 2 Cl~ - . The success in this instance was ascribed to the fact that calculation of doublet to doublet excitations 9 for the odd electron system are free of the ambiguities associated with estimating triplet and singlet energies for the closed shell system. This cannot be the entire explanation as experimental evidence 20 suggests a small singlet-triplet splitting of o - o* (vide supra). Another possible explanation would be the failure of the Xa method to adequately account for electron repulsion contributions to the B - B* excitation energy of the closed shell systems. With regard to this point it should be recalled that the o - 5* transition energies are much less in both the singly oxidized and reduced species. For these cases the ground and excited states contain equal numbers of singly occupied orbitals so that electron repulsion effects will be similar for both the ground and excited states. In the closed shell systems, the o - o* excited state contains two singly occupied 6-like orbitals whereas the ground state contains one doubly occupied 15 orbital. A considerable amount of ionic 1 character in the A2u state would effectively place both o electrons on the same atom with a large electron repulsion term between them. It seems likely that unfavorable electron repulsion effects contribute 1 to the high energy of the A2u state for these systemso Since the metalmetal bond length only changes slightly upon one-electron oxidation or reduction, the magnitude of the unfavorable electron repulsion terms can be estimated from the energy separation between the 15 o* transition in the closed shell case and either the oxidized or reduced case. Values of 12,000, 8000, 17,000 cm Mo2(S04)4 4-/3- , Tc 2Cl 3-/2- o/1- , 8 -1 and Mo (0 CCF 4_ ) 2 2 3 are obtained for , respectively. The 10 exact origin of these electron repulsion energies remains as a problem for future theoretical studies. The Metal Localized 6 - 1r* Excited State Four regions of weak absorption are observed between -1 the more intense 6 - 6* (14, 700 cm ) and 1T (Cl) -1 32, 600 cm ) bands of Re 2 Cl8 2- (Figure 2). The first weak band of (n-Bu4 N) 2 [ Re 2 Cia] maximized at 17, 675 cm • o* (vide infra; -1 and at 15 K consists -1 of three progressions in a 225 cm excited state a 1 g (ReRe) stretching vibration built on three origins. 7 Two of the progressions are polarized parallel and one perpendicular with respect to the ReRe -1 molecular axis. The band red shifts by only 200 cm in (n-Bu4 N) 2 7 [ Re 2 Br8 ] . These data are consistent with a metal localized state 1 which possesses greater antibonding character than A 2 u (o - o*) as reflected by the diminished value of a 1 g (ReRe). vibration was 248 cm -1 Recall that this 1 in the A 2 u state. The presence of several vibronic origins and the weakness of this band suggest an electricdipole-forbidden, vibronically-enabled transition. The o - 1T* 1 7 tA1 g - Eg) assignment was adopted as it would be expected to lie lower than 1T - 1r* (vide infra; was preferred over 1T - -1 39, 215 cm ). A o - 1r* assignment o* since the 7T orbital possesses substantial halogen character(% Cl, 1r, 44; 1r*, 19). The o - 1r* absorption system plays a particularly prominent role iri the electronic spectra of Mo2 (02 CR) 4 (R = H, CH 3 , CD 3 , CF 3 ) compounds. These complexes, which are isostructural with Cu 2 (O 2CCH3 ) 4 , contain four bridging carboxylate ligands. Unlike the other compounds which contain a quadruple metal-metal bond, there is no absorption 11 -1 in the 14,000 - 21,000 cm range. Furthermore, no electronic absorption bands of appreciable intensity were detected at lower energy. 3 2 The first feature in the solution electronic absorption spectra of these species 2 9 , 32 , 33 is a weak band centered at -1 32 22, 700 cm (Ema,--x: 100) and the frozen solution spectrum of 29 Mo 2 (O2 CH) 4 is shown in Figure 3. Dubicki and Martin observed ~ a single vibrational progression of 350 cm -1 for the 22, 700 cm -1 band in the 77 K diffuse reflectance spectrum of Mo 2 (O2CCH 3 ) 4 • They were not able to identify whether the progressional mode was MoMo or MoO stretching and no firm assignment for the 22, 700 cm -1 absorption system was reached; however, it was concluded that this absorption was not the 6 - 6* transition, which therefqre must be at higher energy. More direct evidence that this weak absQrption system was not o - o* came from the polarized crystal spectra of Mo 2 (O2 CCH 2 NH 3 ) 4 (SO4 ) 2 • 4H 2 O and Mo 2 (O2 CH) 4 • 34 , 15 A study 18 of the polarized spectra of a crystal of Mo (O CCH ) 2 2 3 4 as well as films of Mo 2 (O2 CCD3 ) 4 , Moz(O2 CCF 3 ) 4 , and Mo 2 (O2 CH) 4 at 15 K provided a definitive assignment for the 22, 700 cm -1 absorption band. The energy of this transition was found to be essentially independent of the carboxylate ligand (Figure 4) implying a metallocalized excitation. From a comparison of the shifts of the pro- gressional frequency and ground state infrared and raman fundamentals, it was established 18 that the vibration responsible for the excited state progression was a 1 g (MoMo). The direct observation of the 190 cm -1 12 hot band related to the first intense vibronic component conclusively showed that the transition was allowed in the Herzberg-Teller 18 sense and therefore was electric-dipole forbidden. Comparison with the infrared spectra of these compounds identified the 190 cm 18 vibration as an a 2 u (MoO) bending vibration. As this vibration -1 lowers the symmetry of the Mo2 L8 chromophore, it is reasonable that such a vibration would be important in a vibronic coupling mechanism. The presence of two kinds of perpendicular polarization 18 was interpreted as a low symmetry site splitting of a degenerate electronic state. For Mo 2 (O2 CCF 3 ) 4 , a structured emission was observect 18 at 1. 3 K. As for the weak absorption system, a progression in a 1 g (MoMo) is observed. Because of the 1800 cm -1 gap between the 0-0 band of the absorption and emission spectra, and the long luminescent lifetime (2 ms) it ·was suggested that triplet emission from the state related to the weak absorption was observed. It 18 1 1 was concluded that only a A1 g - Eg transition can account for the 22,700 cm -1 absorption system. Of the available low-lying 1 excitations, only o-- 11* or 1T - 6* yield Eg states. The latter assignment was rejected, as Xa calculations 32 on MoiO CH) have 2 indicated that 6 - r;* falls below 7T - 4 6* and more importantly, only the former transition is metal localized (% Mo: 6 89; rr*, 96; 1T, 65; 6*, 86). Furthermore, the calculations suggest that the shoulder at 30,600 cm -1 (Figure 3) with an oscillator strength of approximately 0. 03 should be assigned to the o . . . 6* excitation. 13 The anomolou.s spectral behavior of the Mo 2 (O2 CR) 4 compounds may be a manifestation of strong covalent Mo-O 7T bonding. Calculations 32 suggest that strong metal ligand bonding can indirectly 35 strengthen the metal-metal bond. Vibrational studies of binuclear Mo(II) compounds with various ligand systems tend to support this 36 37 conclusion. Photoelectron spectra ' have also been measured for Mo 2 (O2 CR) 4 compounds and the energy ordering of the bonding orbitals MoMo o > MoMo 1T > MoO ~MoMo a was consistent with the Xa description. Other Metal Localized Excited States The next absorption band, in order of increasing energy, for Re 2 Cl8 2- -1 is another weak (£max 65) band centered at 20, 940 cm (Figure 2). 7 Only perpendicularly polarized progressions of ca. 400 cm - l are observed. 7 This high frequency clearly is not attributable to a 1 g (ReRe), and was assigned to an excited state a 1 g 11 -1 (ReCl) stretch (ground state 359 cm ). This suggests a transition to an unoccupied, metal-chlorine orbital. The vibronic nature of the transition is supported by the low intensity, the presence of several vibronic origins and the red shift and intensity increase between 15 7 -1 and 298 K. A corresponding absorption was observed at 19, 150 cm in (n-Bu4 N)J Re 2 Br8 ] . It was pointed out that this 1790 cm -1 red -1 shift upon halide substitution parallels the 1400 - 1500 cm red shift of the ~y - ~ 2 -Y 2 transition in (n-Bu4 N)J OsNX4 ] (X = Cl, 7 1 1 . Br). Therefore, a A10o. - A1 u (6 - dX 2 -y 2 (b 2 u)) assignment was adopted. 7 Weak absorptions observed in the 22,000 - 24,000 cm 1 region of K.iMo 2 Clg • 2H 2O have been in part ascribed to the A1 o-t, - -1 14 1 )) t rans1·t·10n. 14,31 A?ff (u - rl-X 2-y~ (b 1 cr ""b b Proceeding to higher energy in the electronic spectrum of ? Re2Cls 2- (Figure 2), an extremely weak progression is seen at about 23,650 cm -1 superimposed on more intense absorption. The progressional frequency of 185 cm -r was attributed 7 to the symmetrical bending mode a 1 g (ReReCl). This band was tentatively 7 assigned to the electric-dipole forbidden transition rr - o* tA1 g 1 B u). Excitation of a (ReReCl) was pointed out 7 as consistent with 1 (J" 1 b the considerable halogen character of the rr orbital shown by the Xa . 6 . -1 calculat10n. Part of a weak absorption system at 22,000 - 24,000 cm 4- 1 in Mo 2Cl8 1 has been assigned to A1 g 31 of an Xa calc ulation Eg (1r - o*) on the basis 0 Before moving to the subject of charge transfer excitations, the second lowest "allowed" metal-localized excitation will be considered. A transition centered at 39,215 cm -1 in the 5 K polarized crystal spectrum of ((n-C 4 Hg )4 N)J Re 2Cl8 ] (Figure 2) has been 1 1 6 attributed to the A1g - A 2u (1r - 1r*) transition. The modest red shift of this band to 37,735 cm -1 in the crystal spectrum of ((n-C 4 H9 ) 4 N) 2- Re 2Br8 is consistent with the 44. 4% Cl character of the rr orbital predicted 6 by an Xa calculation for Re 2Cl8 z-. Intense absorption -1 in KqMo 2Cl8 • 2H 20 was in part assigned to the 1T - rr* excitation with a calculated energy 31 of 39,400 cm -r. greater than 34, 000 cm A well resolved band with an oscillator strength of 0. 02 occurs at about 44,000 cm -1 in the solution spectra of Mo2(0 2CR) 4 compounds. 27 29 32 ' ' 15 On the basis of an Xa calculation 1 1 A1 g _ 32 this was assigned to the A 2 u (7T - 1T*) transition. The relative order of the rr - 1r* energies for the compounds discussed parallels that of the 6 - o* transition energies. A metal localized a - a* type transition has never been observed experimentally and Xa calculations consistently predict this excitation to lie in the as yet unexplored vacuum ultraviolet region of the spectrum. Ligand - o* and o - Ligand Charge Transfer Excitations Next the charge transfer states which are possible for the M2 L8 compounds will be discussed. The o and o* orbitals are respectively the highest occupied and lowest unoccupied molecular orbitals in these compounds. Therefore, it is not surprising that metal to ligand charge transfer (MLCT) and ligand to metal charge transfer (LMCT) transitions which originate or terminate in the o or o* orbitals have been observed. First, the LMCT excitations will be considered. If the symmetry adapted ligand 7T and a-like orbitals are considered, one obtains: = 1T II where aL denotes ligand a orbitals and 1r11 denotes ligand 7f orbitals which lie in a plane containing the metal-metal axis and 7Tl. represents ligand 1r orbitals which lie in a plane perpendicular to the metal-metal axiso Only LMCT transitions from the b 2 or one of the three e g g 16 ligand orbitals to the lowest unoccupied b 1 u (o*) orbital are electricdipole allowed. Intense (x, y) polarized absorption at 30, 870 (E of 6000 - 9000 in solution) and 23, 630 cm -1 max in the 5 K polarized crystal spectra of ((~-C 4 Hg )4 N)J Re2 Cls] (Figure 2) and .((~.... C4 ~ )4 N):r [ Re 2 Br8 ] respectively, were assigned to the LMCT transitions l l 6 A1 g --- Eu (Cl(eg 7T11 and 1TJ - 6*). Consistently, an MCD A term is associated with these bands. 13 The halide dependence of the transition, of course, substantiates the LMCT assignment. An Xa cakulation of the electronic structure of Mo 2 Cl8 l 1 predicted the A1 g 37,500 and 3 8, 600 cm 4- Eu (7r(Cl) - o*) transitions to fall at -1 , respectively O In accord with this pre-1 diction, an A term was observed at 37,000 cm in 6 N HCl solutions 4 containing the Mo Cl - ion. 13 To lower energy of the intense LMCT 2 8 absorptions of Re 2Cl 8 2, there are two sharp parallel polarized peaks at 27,000 and 28, 100 cm -1 (Emax of 400 in CH 3 CN solution). 7 -1 These absorptions shift to 21, 290 and 22,080 cm , respectively, 7 in the analogous bromide derivative Because of the small separation O between the two transitions, their LMCT behavior and the polarization 1 3 along the metal axis, the eg rr11 and 1T1- (halide) - 6* ( A1 g - A2u (a, b , Eu)) assignments were proposedo 7 Similar absorptions have not been observed for Mo 2 Cl8 4, perhaps in part because spin~rbit coupling should be smaller for Mo(II) than for Re(III). Complex absorption systems at 35, 700 cm Re2Cla 2- and 26,300 cm -1 in Re 2 Br8 2- -1 (Figure 2) in also appear to be electric- 17 dipole allowed LMCT transitions. 13 Because of the mixed z and (x,y) polarization and the complex MCD signal observed, it was suggestect 13 that the absorption system was due to the overlapping 1 1 1 transitions A1g - A2u (Cl(1r.L) - 6*) and A1 g - 1 Eu (Cl(a) - 6*). A LMCT transition to a metal orbital other than 6* has 414 been proposed for Mo 2Cl8 • The relatively weak absorption band at 28, 800 cm -1 was polarized along the MoMo axis; its 1 intensity was independent of temperatureo Therefore, a A1g 1 14 A 2 u (Cl(1r) - dx 2-y 2) assignment was adopted. Examination 38 of the crystal spectrum of (NH,J 4 Mo 2 Br8 should provide useful additional information about the charge transfer states in the binuclear Mo(II) halides. Transitions of the MLCT type have so far only been identified for binuclear Mo(II) carboxylate compounds o The most intense feature (Emax ~ 7000, f = 0. 19) in the spectrum of Mo2 (O2 CH) 4 1 (Figure 3) is an intense peak at 41,900 cm- . 32 AnXacalculation 32 1 1 has supported a o - (C-O)rr* assignment, specifically A1g - Eu(b2g - eu). Similar bands are observed for Mo 2(O2CCH 3 ) 4 , 29 29 As the ionization potentia139 33 Mo (O CC H Mo2(O2CCF 2 ) 3 , 4 ) 2 5 5 • 4 of the corresponding carboxylic acid increases, the energy of this band decreases which is consistent with the MIC T assignment. Implications of the Spectroscopic Studies with Respect to Metal-Metal Bond Strengths Resonance enhancement of a 1 g(ReRe) and a 1g (MoMo) have been 4 -19 22- 12 observed for Re 2 Cl 8 , Re 2 Br3 , and Mo 2Cl8 saltso As many 18 as 10 overtones of the aw metal-metal stretching vibration were b 4 19 observed for Mo Cl - . This allowed Clark and Franks 12 , 19 to evaluate 2 8 the anharmonicity constant, that is, how the spacing between adjacent vibrational levels decreases with increasing vibrational quantum numbers. The lack of sensitivity of a 1 g(ReRe) to halide substitution suggested to us that this vibration corresponds to almost pure ReRe stretching with little contribution from the other vibrational modes of a 1 symmetry. Because of this and the success of the diatomic metal chro7 mophore description of the electronic spectrum, we suggested that the Birge-Sponer extrapolation technique, well known in the vibrational spectroscopy of diatomic molecules, 40 might apply in this instance. Simply stated, the decrease of the vibrational level separation with increasing vibrational quantum number is used to extrapolate to a dissociation limit. Employing this technique, values of 115 - 130 and 7 110 - 160 kcal/mole were estimated for ReRe and MoMo quadruple bond energieso As pointed out earlier, the electron repulsion contribution to the o - o* excitation may be factored out by examining the single electron oxidized and reduced specieso It is this value which more nearly corresponds to a simple bonding-antibonding orbital splitting. . . 6 21 . 6 30 31 32 41 42 43 Recall that the spectroscopic studies ' and calculations ' ' ' ' ' ' consistently indicate that the o and o* orbitals are almost pure linear combinations of metal dxy orbitals. Because of these facts, an upper limit of the o-5 bond energy should be equal to the electron repulsion 19 corrected bonding-antibonding separation. With the 0-0 energies -1 for the o - 6* transition in the odd electron 3 -15 3 -15 compounds Mo2 (SO4 ) 4 and Tc 2Cl8 estimates of 18 and 17 of 6314 and 5900 cm kcal/mole for the maximum strength of the o-o bond are obtained. 20 References 1. G. Herzberg, Molecular Spectra and Molecular Structure I. Spectra of Diatomic Molecules, Van Nostrand Reinhold, New York, N. Y., 1950, and references therein. 2. G. Herzberg, Molecular Spectra and Molecular Structure III.Electronic Spectra of Polyatomic Molecules, Van Nostrand Reinhold, New York, N. Y., 1966, and references therein. 3. F. A. Cotton, Chem. Soc . Rev., 4, .,.... 27 (1975), and references therein. 4. I. N. Levine, Quantum Chemistry, Volume I: Quantum Mechanics and Molecular Electronic Structure, Allyn and Bacon, Boston, Mass., 1970, pp. 287-289. 5. F. A. Cotton, et al. , Science, 145, 1305 (1964); F. A. Cotton, ~ Inorg. Chem. , 4, 334 (1965) . .r-. 6. A. P. Mortola, J. W. Moskowitz, N. Rosch, C . D. Cowman, and H. B. Gray, Chem. Phys. Lett., 32, 283 (1975) . .,.....,..__ 7. W. C. Trogler, C. D. Cowman, H. B. Gray, and F. A. Cotton, J. Am. Chem. Soc. , .r-..r-. 99 , 2993 (1977) . 8. F. A. Cotton, N. F. Curtis, B. F. G. Johnson, and W. R. Robinson, Inorg. Chem., 4.r-., 326 (1965) . 9. W. C. Trogler and H. B. Gray, unpublished observations. 10. C. D. Cowman and H. B. Gray, J. Am. Chem. Soc., 95, ,..__,.._ 8177 (1973) . 11. W. K. Bratton, F. A. Cotton, M. Debeau, and R. A. Walton, . J. Coord. Chem., .,.... 1,121 (1971) . 21 R J. H. Clark and Mo Lo Franks, J. Am. Chem. Soc. , .,...__,..._, 98 0 2763 (1976). 13. C. Do Cowman, W. Co Trogler, and H. Bo Gray, Israel J. Chem., in press. 14. P. E. Fanwick, D. So Martin, F o A. Cotton, and T. R. Webb, Inorg. Chem. , in press 15. 0 F. A. Cotton, D. S. Martin, P. E. Fanwick, TO JO Peters, and T. R. Webb, J. Am. Chem. Soc., -"-"98, 4681 (1976). 16. Do K. Erwin, G. L. Geoffroy, H. B. Gray, G. S. Hammond, E. I. Solomon, W. C. Trogler, and A. Ao Zagars, J. Am. Chem. Soc., .,.-..,-,. 99, 3620 (1977) . 170 R. S. Mulliken, J. Chem. Phys., .,.... 7, 20(1939) . 18. W. C. Trogler, E. I. Solomon, Io Trajberg, C. J. BalThausen, and H. B. Gray, Inorg. Chem. , 16, 82 8 (1977). ~ 19. R. J. H. Clark and M. L. Franks, J. Am. Chem. Soc. , 97, r,.,-,. 2691 (1975)0 200 Wo C. Trogler, Eo I. Solomon, and H. B. Gray, Inorg. Chem., submitted. 21. F A. Cotton and E. Pedersen, J. Am. Chem. Soc. , 97, 303 (1975), O /'-"- and references therein. 22. J. R. Ebner and Ro A. Walton, Inorg. Chim. Acta, -"-"14, L45 (1975)o 23. JO 15' .,.._,.._ 24. R. Ebner, D. R. Tyler, and R. A. ·walton, Inorg. Chem., 833 (1976)0 F. Ao Cotton, P. E. Fanwick, Lo Do Gage, B. Kalbacher, and D. S. Martin, J·. Am. Chem. Soc., submitted. 22 25. A.. Loewenschuss, J. Shamir, and Mo Ardon, Inorg. Chem., 15, 238 (1976) 0 -"-" 26. J. H. Baxendale, C. D. Garner, Ro G. Senior, and P. Sharpe • J ~ Am. Chem. Soc., 98, 637 (1976) . .r-.r,. 27. W. Co Trogler, PhoDo Thesis, The California Institute of Technology, 1977. 28. F. A. Cotton and C. B. Harris, Inorg. Chem., .,..._ 6, 924 (1967) . 29. L. Dubicki and R. L. Martin, Aust. J. Chem. , -"-" 22, 1571 (1969). 30. Ao P. Mortola, J. W O Moskowitz, and N. Rosch, Int. J . Quant. Chem. Symp., No. 8, 161 (1974). 31. J. G. Norman and H. J. Kolari, J. Chem. Soc., Chem. Commun. , 303 (1974); J . Am. Chem. Soc., 97, 33 (1975). r,.r,.. 32. J. Go Norman, H. J. Kolari, H. B. Gray, and W. C. Trogler, Inorg . Chem., .,..._.,..._ 16, 987 (1977) . 33. F. A. Cotton and J . G. Norman, J. Coord. Chem., .,..._ 1, 161 (1971) 0 34. F. A. Cotton, D. S. Martin, T. Ro Webb, and T. J. Peters, Inorg. Chem., 15, 1199 (1976). r,.r,.. 35. A. P. Ketteringham, C. Oldham , and C. J. Peacock, J. Chem. Soc. Dalton, 1640 (1976) . 36 0 J. Co Green and A. J. Hayes, Chem. Phys. Lett., 31, 306 (1975). 37. F. A. Cotton, J. G. Norman, B. Ro Stults, and T. R. Webb, r,.r,.. J. Coord. Chem. , .,.... 5, 217 (1976) . . 38. J. V. Brencic, I. Leban, and P. Segedin, Z. Anorg. Allg. Chem. , 427 , 85 (1976). ~ 23 39. D. A. Sweigart and D. W. Turner, J o Am. Chem. Soc., 94, /'-", 5592 (1972). 400 Reference 1, p. 43 8. 41. M. B. Cingi and E. Tondello, Inorg. Chim. Acta, 11, L3 (1974). ""' 42. M. Benard and A. Veillard, Nouveau J. DeChimie, .,.._ 1, 97 (1977) . 43. F. A. Cotton and B. JO Kalbacher, Inorg. Chem. , in press 0 'Sj'i N Table I -- 2- 2- Compound Re 2 Cl 8 Re 2 Br8 '1- Ema_'( 16,955 170 3015 2 KBri 18,600 17,390 :s; 700 :s; 1500 -3 Et 2 O11 22,000 :s; 1500 4+ 4- Et 2O11 l9,500 f a g(M-M) cm 1 lJ-\.10- -1 o. 015f 346g 277c 274':: 289d 33 68 255d 248a 1A2lt 278i 320j - - - -::..-- 342i 1530b 0.023 19,400 33 7 1050 19, 840 483 l Energy max . A1 0' 1 A2u(o - 0*) in cm- 1 14,654a Polarization za 14,104d Sample condition zd • x-ta la/Methanol-HClb x-tala 19,128e 10- M HSO3 CF 3 m 20,920 14,363d 14,290 ze x-talh za x-tale /6M HClf Tc 2Cl 0 x-tala, d M0zCl Re~Cl 0 (PEt 3) 2 KBr/CHaCl) 0 1- M0Cl1 (PEt M HSO3 CF 3 m 1100 zk 10 17,090 2200 x-talk/5M H2SO4 Mo2(en) 1 CH 3 Cl/ 18, 870 3) 4 4- Mo 2Cl,i (SMe 2 ) 4 CHC13 i 1300 Mo 2(SO1 ) 1 ,i+ Mo 2 Cl,i (bipy) 2 16,610 Mo 2 (aq) Mo2Cl,i(CH3 CN) 4 CH 3 CNi 4- Et 2O11 Mo 2Br1 (bipy)2 Re 2 (CH 3) 8 Cr 2 (CH 3) 8 1- Moz(CH 3) 8 25 Table I Footnotes a. c. D. Cowman and H. B. Gray, J. Am. Chem. Soc., 95, 8177 r.r.. (1973). b. F. A. Cotton, N. F. Curtis, B. F. G. Johnson, and WO R. Robinson, Inorg. Chem., 4, 326 (1965) . .r,,. c. W K. Bratton, F o A. Cotton, M. Debeau, and R. A. Walton, 0 J. Coord. Chem., 1, 121 (1971) . .r,,. d 0 C D. Cowman, Ph. Do Thesis, California Institute of Technology, 0 1974. e. p. E. Fanwick, D. S. Martin, F. A. Cotton, and T. R. Webb, Inorg. Chem., in press. f. W. C. Trogler, E. I. Solomon, I. Trajberg, C. J. Ballhausen, and H. B. Gray, Inorg. Chem., .,.._,.,_ 16, 828 (1977)0 g. R. J. H. Clark and M. L. Franks, J. Am. Cher.n. Soc., 97, r..r,. 2691 (1975). ho K. Schwochau, K. Hedwig, H. J. Schenk, and 0. Greis, Inorg. Nucl. Chem. Lett., 13, 77 (1977) . .r,,..r,,. L Value reported for Re 2 Cl6 (P(n-C 3 H7 ) 3 ) 2 and Mo 2 Cl4 (PBu3 ) 4 , J. San Filippo and Ho J. Sniadoch, Inorg. Chem., .,.._,,.,_ 12, 2326 (1973); Jo San Fillippo, H. J. Sniadoch, and R. L. Grayson, Inorg. Chem., 13, .,..._,.._ 2121 (1974) . j. vV. C. Trogler and Ho B. Gray, unpublished results. k. F. A. Cotton, D. S. Martin, P. E. Fanwick, T . J. Peters, and To R. Webb, J. Am. Chem. Soc., 98, 4681 (1976) . .r,,..r,,. 1. D. K. Erwin, G. L . Geoffroy, H. B. Gray, G. S. Hammond, E. I. Solomon, W. C. Trogler, and A. Ao Zagars, J. Am. Chem. Soc. , 99, 3620 (1977) -.r,,..r,,. 26 m. A. R. Bowen and H. Taube, Inorg. Chem., £, 2245 (1974). n. A. P. Sattelberger and J. P. Fackler, J. Am. Chem. Soc., 99, 1258 (1977). """ 27 Figure Captions Figure 1. Simple MO diagram for the d orbital splittings in a homonuclear diatomic molecule (A) and in a M L compound with D h symmetry (B). 2 Figure 2. 4 8 Single crystal polarized electronic absorption spectra of (n-Bu4 N) 2[ Re 2Cla] at 5 K (outer spectra) and 15 K (central portion). Three different crystals were used to obtain these spectra. Figure 3. Electronic absorption spectrum of Mo 2 (O2CH) 1 in a 4:1 ethanol:methanol glass at 80 K. Figure 4. Electronic absorption spectra of the o - 1r* band in Mo 2(O2CCH ) 3 , 4 Moz{O2CH) and Mo2(O CCF 4 2 ) 3 • 4 Spectra were obtained at 15 K for samples which had been sublimed onto a quartz window. 28 a* --- - - - - -- o* 0 lj * 1k o* ----- - - 11 - --- ---- ·--.------- A - · -· 0 1J -- -- --· - - -- - - - ---'-.>'.--11 8 a 1I 29 ABSORBANCE ~ 0 0 0 0 f\) ()7 0 (>J (Jl 0 0 0 (>J 0 0 (>J 0 0 0 0 (>J ()7 0 f\) (Jl 0 0 0 ~ 0 0 , ..f·.r··· _fr,· ~ (Jl 0 f\) z $ 0 0 n 0 ~ 0 J_ (Jl 0 0 i i l. \ "~; (Jl (Jl ..,. ·7 ··2 0 ---;.., ..--· 1. ..... ., ~~ \.. () '? \ \ I t- (]) (Jl 0 en -."'L_:, -=--.l...~ c:.:·· "' 0 0 r---~:-...... ..~:> -· 0 0 0 30 40000_ ---~3~0~0~0_0_ _ _ _ _ _ _ _2~0~000 ---1-- r 7000 BOK .6000 5000 4000_ 3000 ~ 2000 X 10 1000 0 200 300 400 WAVELENGTH (nm) 500 31 "'u C 0 D .,,0 D <{ 410 420 430 440 A (nm) 450 460 470 32 Further Studies of the Electronic Spectra of Re 2 Cl8 2 - and R~Br 8 20 Assignment of the Weak Bands in the 600-350 nm Region. Estimation of the Dissociation Energies of Metal-Metal Quadruple Bonds. Contribution No. 5433 from the Arthur Amos Noyes Laboratory of Chemical Physics 2 Califcrnia Institute of Technology! Pasadena, California 91125 and F. Albert Cotton ~ Department of Chemistry, Texas A & M University, College Station, Texas 77843 Abstract: ~ Previously published evidence bearing on the polarizations and 2- assignments for the strong bands in Re2 Cl 8 and related species is synthesized and it is shown that work from our two laboratories gives a consistent picture. New polarized spectra of thick crystals of fu-Bu 4 N)2 LR~ Cl 8 J and fu-Bu 4 N)2 [R~Br 8], measured at 15 K, are reported and discussed. Four new regions 1 of weak absorption between the more intense 6 - o* (14, 700 c m- ) and 7T ( Cl) - 1 2- o* (32,600 cm - ) bands of Re2 Cl 8 have been characterized. The first weak band of fu-Bu 4 N)z [R~ Cl 8 ] maximizes at 17,675 cm -l and consists of three progressions in a 225 cm -l excited state a 1 g (ReRe) stretching vibration built on three origins. Two of the progressions are polarized parallel and one perpendicular with respect to the ReRe molecular axis. 1 The band red shifts by only 200 cm- in (n-Bu4 N)z[ R~Br 8 ] . The results indicate that the band is attributable to the electric-dipole-forbidden, metal- 33 localized transition 6 - 1T* ( A1 g - 1 E g). A second absorption at 1 20 ,940 cm- exhibits only perpendicularly polarized progressions on 1 several vibronic origins with a common vibrational spacing of ca. 400 cm- (a1g (ReCl) stretch). (15 to 300 K). This band also intensifies and red shifts upon warming The 1 790 cm -i red shift of this band in @-Bu 4 N)2 [Re2 Br 8 J pinpoints it as a o - dx2 _ y2 (b2 u) ( A1 g - 1 A1 u) transition. The third 1 band, which is at 23,645 cm- , is extremely weak and only one perpendicularly polarized progression in the a 1 g (ReReCl) skeletal bending mode was detected. A 1r - o* ( A1 g - 1 B1 u) assignment is suggested. The fourth absorption system is comprised of two sharp peaks at 27,000 and 28,100 cm - i , which are parallel polarized. These peaks shift to 21,290 and 22,080 cm -i in fu.-Bu 4 N)z [ Re2 Br 8 ] , consisb:mt with a ligand to metal charge transfer 1 1 (LMCT) assignment. As the LMCT transitions to a E and b E states u u 2are not split appreciably in RE:z Cl 8 , it is suggested that the fourth weak band system is derived from the corresponding eg1i(Cl) . . . o*( A1 g 3 4 ..A.>zu (a, b E u )) e.."'<citations. Dissociation energies of Re1~Re and Mo °:i.vio bonds have been estimated from a Birge-Sponer extrapolation, giving values of115-130 kcal/mol for the former and 110-160 kcal/mol for the latter. 34 ~ 1 2 The earliest, tentative suggestivns ' as to how to assign the 2- electronic spectrum of R~Cl 8 the ~pecies in which the existence of 3 quadruple bonds was discovered, l, as well as the extension of such assign, ments to the spectra of closely related complexes, have not withstood recent experimental 4-1 0 and th eore t.1ca.16 ' 1 0-12 scru t.my. wor k ' d one wit hin the past 12 few years 4 - ,has provided new spectral assignments that seem certain to be I correct in their essentials and bring us to a point of departure for delving more deeply into the details of the spectra and bonding in the entire class of 13 compounds containing quadruple M-M bonds, and other multiple bonds between metal atoms. One purpose of this paper is to present some new ?- and pertinent facts concerning the spectra of the Re2 X8 - (X = Cl, Br) ions. In the past it has appeared that there have been disagreements between publications from our two laboratories as to observations and 2- interpretations regarding the electronic structural properties of Re2 Cl 8 and related species containing quadruple bonds. In actual fact, we are in essentially complete agreement as to the observations and their interpretation. In view of the importance of establishing a clear picture of the situation 1 3 regarding the prototypal ' quadruple bond, _i.e. , that in R~ c1:-, we also present in this report a synthesis of some isolated observations recorded in earlier publications, as well as a needed critique of our previous estimates of the bond dissociation energies (BDE's) of quadruple bonds. 5 Contrary to the impression created in a recent publication from one of our laboratories, there was in fact a sound basis for the claim, which was 1 advanced 4 ' 6 first by Cowman and Gray , that the 14,000 cm- absorption band .,_ in R~Cl8 - is polarized parallel to the Re-Re a."'<:is. The crystallographic 35 results, showing as they do the existence of a disorder in the crystals, 4 7 taken together with the earlier crystal spectra, , confirmed in the rnost direct way the aforementioned assignment of polarization. The misunderstanding between our laboratories was in large part attributable to the fact that the logic of the Cowman-Gray argument . was not presented 4 6 in full form in either of the two published documents. ' To set the record straight, we shall now outline the main parts of the argument. It should be recalled that each of the three complexes, (g-Bu 4 N\lRezC1 8 J, fu.-Bu 4 N)2 [R~Br 8 J, and Re2 Cl6 [P(C2 H5 ) 3 ] 2 , exhibits a moderately intense band at about 14,000 cm-1; in each case the integrated intensity of the band is independent of temperature foetween 5 and 300 K) and a vibrational progression in a 1 a- (ReRe) is exhibited at 5 K. Studies on single crystals of b 1 Re2 Cl 6 [ P(C2 H5 ) 3 ] 2 showed conclusively that the 14,000 cm- band is 4 ReRe-axis (parallel or z) polarized. Starting with this incontestable result, three important points should be emphasized: (1) There is an obvious 2- analogy of the 14,000 cm -i bands in R~Cl 8 R~Cl 6 [P(C2 H5 ) 3 L. 2- and Re Br 2 with that in 8 (2) The fact that the solution MCD spectra of fu.-Bu 4 N)2 show that the intense bands at 32 ~ 600 and 23,630 cm - i , respectively, exhibit A terms.6 ' 7 ' 14 These results, which lR6z Cl ] 8 and (n-Bu N) [R~Br 4 2 ] 8 show that the excited state is degenerate, establish that the intense bands 1 are A1 g - 1 Eu, and therefore that they must be polarized perpendicular to the molecular axis. As the crystal spectra show, 4 , 7 the 14,000 cm- 1 l band displays polarization character opposite to that of A1 g must therefore be parallel polarized. 1 Eu. It (3) The previously unpublished observation that a few crystals of (:g_-Bu 4 N)2 lR~Br 8 ] possess a nondichroic -1 7 orange-yellow face, with the 14,000 cm band absenL No such face 36 would be possible if the band were perpendicularly polarized. 2 1 The early reports ' ~ of the electronic absorption spectrum of fu.-Bu 4 N\ lRe2 Cl8 ] did not concern themselves with the region of weak absorption between the 14, 700 and 32,600 cm -i bands. By examining concentrated 2- solutions containing R~Cl8 , we have now shown that there are four weak bands in this region. We shall first describe here the single crystal spectra and assignments of these new band systems for (g_-Bu 4 N)2 LRezCl8 ] , and also report related spectroscopic results for (g_-Bu4 N)2 [RezBr 8 ] . Ex erimental Section Crystals of (Q.-Bu 4 N\ [R8zCl8 ] and (n-Bu 4 N) 2 lR~Br 8 J were grown 4 and oriented as previously described. Low temperature polarized spectra were recorded on a Cary 17 spectrometer using a Cary liquid helium dewar and tandem Glan-Taylor air-spaced calcite polarizers. The 0. 2 and O. 5 absorbance slidewires were used to record the weak bands. Temperature was measured with a calibrated carbon resistor. Polarizations are denoted II and l along the extinction directions of the (100) crystal face, corresponding 4 to projections predominantly 11 and 1 to the ReRe molecular axis. ' 7 Results and Discussion The electronic spectrum of (n-Bu4 N)2 LRe.2 Cl8 J in ac etonitrile solution at 298 K (Figure 1) shows fom.' absorption bands (I - IV) between the electricdipole-allowed 6 - 6* ( A1 g - 1 ~u) and 7T(Cl) - o*( A1 g - 1 Eu) systems at 682 and 314 nm, respectively. In a crystal at 15 K, band I, which is 1 centered at 17,675 cm- , consists of three vibrational progressions (Figure 2 and Table 1 )" Owing to overlap with the 6 - 6* band at lower energy, it is not possible to locate the true progressional origins with certainty. A strong 37 1 parallel polarized progression in a 225 cm- vibration originates at 17 ,062 cm-1, and a less intense one begins at 17,452 cm -l O The weaker o-ression could either represent a new vibronic origin or a single pr Ob 1 quantum of a 1 g (ReCl) built on the one at 17,062 cm- • A perpendicularly polarized progression of the same frequency is built on a vibronic origin 1 at 17,117 cm- 1 • The 225 cm- progression is assigned to the excited state a 1 g metal-metal stretching modeo For comparison, the ground and 16 4 6 7 -1 . and 248 ' ' cm , respectively·. A 0 _ o* state values are 2 74 maximum in the progression is reached at the four th band, or third vibration, 4 in contrast to the o - o* system, which peaks at the first vibration. ' 7 This indicates a greater distortion of the R~ bond in the excited state associated with band I than that accompanying o -► o* excitation. The reduced value of a 1 g (ReRe) for band I is also consistent with a transition to an orbital that possesses a large amount of metal-metal antibonding character. The mixed polarization, the presence of at least two vibronic origins, and the weakness of band I suggest an electric-dipole-forbidden, vibronically-enabled transition. A similar band, with unresolved, overlapping vibronic structure, is observed at 17,475 cm -l for fu-Bu 4 N\[R~Br8 ]. Such insensitivity to halide substitution confirms the metal1 localized nature of the excitation. The transition o . . . 1r* ( A1 g . . . . E g) is consistent with the experimental data, and would be expected to lie lower than 1r 5 K. 6 1 1T*, which is at 39,215 cm- in a crystal of (n-Bu4 N\[R~Cl 8 ] at We prefer o -- 1r* to 7f - B* for the 17,675 cm -i band, as the 7T orbital possesses a substantial halogen character. 6 38 -1 Band II, which is centered at 20,940 cm (Figure 2 and Table 1), exhibits only perpendicularly polarized components with several vibronic origins, complex overlapping progressions, and a common vibrational -1 spacing of ca. 400 cm . Thts high frequency progression clearly is not attributable to a1 g(ReRe), and is assigned to an excited state a 1 g (ReCl) 1 6 stretch (ground state1 359 cm - ), which suggests a transition to an unoccupied, metal-chlorine orbital. The vibronic nature of the transition is supported by the low band intensity, the presence of several vibronic origins, and the red shift and approximately twofold increase in intensity upon increasing the temperature from 15 to 298 K. (It was not possible to assess the temperature dependence of bands I or III, owing to concomitant thermal broadening of the more intense systems with which they overlap.) A corresponding band was observed at 19, 150 cm - l in (n-Bu,1N)J Re 2 Br 8 ] . This modest (1790 cm -i) red shift upon halide substitution parallels the 1400 - 1500 cm -1 red shift 8 The above of the dXY - dX 2 -y-?. transition in (n-Bu 4 N)[ OsNX 4 ] (X = Cl, Br)! data support c1 5 -- dx 2 -y 2 assignment for band IL In a D4 h complex such 2- as Re 2 Cl8 1 , there are b1 g and b 2 u linear combinations of the dx 2 -y 2 orbitals, 1 yielding A:zg and A 1 u excited states, respectively, for promotion of a o electron. A transition to either of these states is orbitally forbidden. 1 In 1 the Herzberg-Teller coupling scheme,1 9 A 1 g -- A 2g can gain vibronic intensity 1 both parallel and perpendicular but A 1 g - 1 A 1 u can 2nly_ be allowed in perpen- dicular polarization, as a D4 h MzX:8 unit does not possess an a 2 g vibrational mode. As band II exhibits only perpendicular polarization, the assignment 0 1 - dx2-y2 (b 2 u) ( A1 g - 1 A1 u) is adopted. Band III, which is centered at 23, 645 cm -1 , is barely detectable on the low energy tail of band rv (Figure 2 and Table 1). clear progression in a 185 cm -1 There is only one vibration, and it is perpendicularly polarized; 39 however, the presence of parallel components cannot be ruled out, as -1 the intensities are very low. The 185 cm symmetrical bending mode, a 1 g (ReReCl). vibration is attributed to the One striking aspect is the length of the progression, which, despite its weakness, is detectable for 12 to 13 vibrations. Band III is not observable in (n-Bu,1N) 2 [ Re 2 Br 8 ] , owing to the presence of interfering charge transfer absorption. We tentatively assign band III to the electric-dipole-forbidden transition 1 1 6 1T _ 6* ( A1 0" B1 u). The fact that an X a calculation has shown that b the 1f orbital of Re 2 Cl8 2- possesses considerable halogen character is consistent with the assignment, as excitation of a 1g (ReReCl) is reasonable 1 for a transition to B1 u. Band system IV, which consists of two sharp, parallel polarized peaks at 27,000 and 28, 100 cm -1 , shifts to 21, 290 and 22,080 cm respectively, in (n-Bu4 1'DJ Re 2 Br8 ] . for Re 2 Cl 8 2- and Mo 2 Cl 8 4- , This constitutes conclusive evidence for a ligand to metal charge transfer (LMCT) assignment. spaced, allowed eg1r(Cl) - <'>* (A1 g - -1 1 Two closely- Eu) transitions have been predictect 6 , lO by Xa calculations, and the intense band -1 observed in a crystal of (!:!_-Bu4 N)J Re 2 Cl 8 ] at 30,870 cm is split into two 6 components. We suggest that band IV comprises the transitions to the 3 two A2 u spin-orbit components derived from the LMCT states a E and 3 u b Eu. 40 2- Previous estimates made by us of the BDE for Re-Re in R~Cl 8 20 now seem likely either to be too low (ca. 80 kcal/mol), because salvation 20 21 contributions to excited state reactivity could be substantial, , or too 2 high ~- 360 kcal/mol) because of accumulated errors in the MO calculation. 22 It is interesting to note that for the Re-Re double bonds in Re 2 Cl 9 and Re 3 Br9 the BDE's have been estimated from thermochemical data to be 85 and 103 40 kcal/mol, respectively. 23 This suggests that for the quadruple bond a value somewhat over 100 kcal/mol might be sensible. As another approach to this problem, we have utilized available 2- 2- 4- vibrational data for Re2 Cl 8 , R8zBr3 , and M°-2Cl 8 • Data first . 16 ' 17a ' 24 ' 25 h ave rec entl y b een improved . reported for these species and 2 extended by Clark and Franks for Re2 X8 - 26 a (X = Cl, Br) and for 4 M0zCls - . 26 b In each case the Raman band that is due principally to totally symmetric metal-metal stretching was observed not only as the fundamental, but through several overtones. first three overtones and for M°-2Cl8 4- 2- For the R8zX8 ions, the the first ten overtones were observed. From these overtone progressions Clark and Franks were able to evaluate w and x , the harmonic frequency and anharmonicity constant, respectively, e e for each of the MzX 8 n- species. It is well known in the vibrational spectro27 2 scopy of diatomic molecules ' 8 that such information may be used to estimate BDE 1 s by a method known as the Birge-Sponer extrapolation. Specifically, the BDE is estimated as (w 2 /4x. )-w /2. This is only an e e e approximation, even for diatomics; it can give results either too high or 27 too low, but usually too high. Its reliability in a polyatomic system is uncertain; the main difficulty arises because the "bond stretching mode", so-called, for a given bond will in general be coupled to other modes of the same symmetry and it is not clear whether the degree of coupling will vary With the state of excitation of the vibration. In the MzX 8n- ions, with D4h symmetry, there are two other totally symmetric ( a 1 g) vibrations besides " vMM , " namely "v:MX " and "vMMX·" Thus we apply the Birge-Sponer extrapolation in these cases with considerable circumspection. 41 Table 2 gives the input data from Clark and Franks 26 and the BDE's estimated. The uncertainty intervals on the latter derive solely from those listed for the~ values. It can be seen that all of the BDE's obtained are in a range we consider reasonable. If we assume that the values obtained are more likely to be too high than too low, we might say, allowing for an error in this direction of 15-20o/c,that the BDE for Re~Re 4 may be 115-130 kcal/mol, whereas that for Mo Mo is probably in the 0 range 110-160 kcal/mol. It should be noted that these are very substantial bond energies. The only homonuclear units with BDE's that are clearly higher are C=C (196) and N-N (22 6), and beyond the first short period only 29 P==P (125 kcal/mol) has a BDE in the same range. Conclusions ~ (1) There exists excellent experimental evidence of several kinds for the entire assignment of band polarizations for RezCl/- given previously. 4 , 5 This leads, in turn, to reliable transition assignments for these bands, including four weak features, as has been detailed above. No significant disag-reements exist between our two laboratories in these respects. 4 (2) Ear lier estimates of the Re Re bond energy based on photo2 20 chemical observations and a crude Iv10 calculation are recognized to be, 0 respectively, too low and too high. We suggest, very tentatively and with full awareness of the dangers inherent in the method, that by a Birge-Sponer n- extrapolation using Clark1 s Raman data on overtones of the ~X8 species, 4 4 the Mo Mo and Re Re bond dissociation energies most likely lie in the 0 0 ranges 110-160 and 115--130 kcal/mol, respectively. 42 - We thank E. I. Solomon for helpful discussions. W.C.T. acknowledges an IBM Fellowship for 1976-1977. This research was supported by the National Science Foundation. 43 References and Notes 1. F. A. Cotton, Inorg. Chem., i_, 334 (1965). 2 . F. A. Cotton and C. B. Harris, Inorg. Chem. , ,§, 924 (1967). 3. F. A. Cotton, et al., Science, W, 1305 (1964). 4 _ c. D. Cowman and H.B. Gray, J. Am. Chem. Soc.,~' 8177 (1973). 5. F. A. Cotton, B. A. Frenz, B. R. Stults, and T. R. Webb, J. Am. Chem. Soc., Q§_, 2768 (1976). 6. A. P. Mortola, J. W. Moskowitz, N. Rosch, C. D. Cowman, and H. B. Gray, Chem. Phys. Lett., ..,..__,.._ 32, 283 (1975) . 7. c. D. Cowman, Ph.D. Thesis, California Institute of Technology, 1974. 8. D. S. Martin, P. E. Fanwick, and F. A. Cotton, unpublished works. 9. F. A. Cotton, D. S. Martin, P. E. Fanwick, T. J. Peters, and T. R. Webb, J. Am. Chem. Soc., .,..__,..,._ 98, 4681 (1976) . 10. F. A. Cotton, B. W. Kalbacher, and L. Gage, SW-Xa calculations and spectra of Tc 2 Cl 8 11. 3, to be published. J. G. Norman and H. J. Kolari, J. C. S. Chem. Comm., 303 (1974); J. Am. Chem. Soc., 97, 33 (1975 ) . .r-./'. 12. F. A. Cotton, R. F. Fenske, T. Block, and D. Lichtenberger, calculations on MOz (SO4\ 1-/3- . , to be published. 13. F. A. Cotton, Chem. Soc. Rav., ,.._ 4, 27 (1975). 14. F. A. Cotton and E. Pedersen, J. Am. Chem. Soc. , .,..,_,,._ 97, 303 (1975), cf. ref. 17, citing a personal communication from P. N. Schatz. 15. F. A. Cctton, N. F. Curtis 1 B. F. G. <Johnson, and \V. R. Robinson, Inorg. Chem. , ,.._ 6, 326 (1965 )o 16. W. K. Bratton, F. A. Cotton, M. Debeau, and R. A . Walton, J. Coard. Chem .., 1, ,.._ 121 (1971). , 44 l7. (a) A. P. Ketteringham and C. Oldham, J.C.S. Dalton, 2304 (1973-); (b) A. P. Ketteringham, C. Oldham, and C. J. Peacock, ibid., 1640 (1976). 18. C. D. Cowman, W. C. Tr ogler, K. R. Mann, C. K. Poon, and H B. Gray, Inorg. Chem., 15, .,.._,,... 1747 (1976) . 0 19. G. Herzberg, Molecular Spectra and Molecular Structure, Vol. III, Van Nostrand Reinhold Company, New York, 1966, p. 140. 20. G. L. Geoffroy, H. B. Gray, and G. S. Hammond, J. Am. Chem. Soc., 96, 5565 (1974). ~ 21. R. H. Fleming, G. L. Geoffroy, H. B. Gray, A. Gupta, G. S. Hammond, D. S. Kliger, and V. M. Miskowski, J. Am. Chem. Soc . , ~' 48 (1976). 22. Since the results in reference (2) appear to give separations between the ground state and the excited states that are too large by factors of 1. 5 to 2. 0, it could be argued , crudely, that the estimate of total 40 Re-Re bond ener gy ou ght to be revised down by a factor of O. 5-0. 7, .Le., down to something in the range of 180-240 kcal/ mol , but even this range is likely to be t oo high. 23. D. V. KoroPkov, Kh. Missner, and K. v. Ovchinnikov, Zh. Strukt. Khim., 14, .,.,_,.,_ 717 (1972) . 24. C. L. Angell J F. A. Cott on : B. L . Frenz , and T. R. \Vebb, J.C. S. Chem. Comm., 399 (1973). 25. J. San Filippo and H. J. Sniad och , Inorg. Chem. , U, 2 326 (1973). 26, (a) R o J. H. Clark and M. L. Franks r J. Am. Chem. Soc., 1t3, 2763 (1976); ro) idem., ibid., 17, 2691 (1975) . 27 • G. Herzberg, Molecular Spectr a and Molecular Str ucture. I. Spectra of Diatomic Molecules, Van Nostrand Company , New York , 1950, P 438. 2 8, A. G. Gaydon , Dissociation Energies, Dover P ub lications, Inc., New York, 1950, Chapter 5. 45 29. D. A. Johnson, Some Thermodynamic Aspects of Inorganic Chemistrv, Cambridge Univ. Press, 1968. 46 Table 1. l i,,11(-j:l•.:l) t"!l l ~I~{;_ I 17(\';2 f1f. ·1. /. l ·;117 5'i f.. 2 ttf:l 1'1:'0j S~C.. G 1-;-:qJ 573.0 17-\ :"•~ [170 . 1 l'i 5-l l f,G8. 9 l 'i57f. !•!':•~• j 11 :: (<· I\\ - ') p ~1l;1 ri :· :•l ic•tl 'A, ,.~ -'-I I\: l. ~:~:1_ nG .I. 2~!l ~'lL ~22 .I. 231 !.iGS. C 17G 'M !" G:!. 7 17771 5f.l . 5 17800 55S. 4 1'1908 556. 0 17%6 555. 3 1 SOI 0 551. 3 18139 519. 1 18202 547.7 18258 51~. 9 16352 • 5·13. 6 I 83So 541. 8 4 97. 3 18157 20109 .I. 493 . 8 20251 .I. 492. 3 20313 1 II - ·---;.- ,- - · ~ 2'..iO_ 23-1 .I. 201 215 231 .I. 218 216 490. 4 2039 1 4 88. 'I 204 62 4 67. 4 20517 4 83. 9 20 GG5 482. 5 20725 480. 6 20 807 47ti. 9 20881 477. 6 20938 474 . 4 21079 473. 1 21137 471. 2 21222 ~G9.9 21231 468. 6 21340 456 . 7 21427 465. 4 21187 461, 4 21533 213 .I. 199 194 .L ~03 . .l .l J. 414 421 l-"·.l J. .l .l .I. 414" -.1; .l -.:-a .l .l 402 J. .l 408 .l 40.G I .l .l 39•1 .l 397 + 462, 2 21636 461. l 21G87 460, 1 21731 J. 458. 2 2182-1 .l ••. 456 . 9 218 87 455, 6 219·19 451 . 2 2:: 01·1 J. 452.8 220% J. 419. 2 22262 J. 415. 1 2~167 J. -1 ·1 '.~. 2 2:•,; I 1 .j. .l ,_ .l 438 l. t A,., 47 III 43 9. 2 22'769 435.5 22962 432.7 23111 -, 193 .L 1 E J_ g 150 J_ 177 429. 4 232 88 426.4 23452 • 422. 9 23646 ,.l 164 J_ 194 , J_ 169 419.9 23815 416.4 24015 413.1 24 2or7 J_ 200 J_ 192 J_ 183 IV 410.0 24390 40G.7 24588 .l.. 370. 27000 356. 28100 II ·II .l.. 198 3 A2u (a Ell) 3 A 2 u (b E u·) 48 Table 2. Birge-Sponer Extrapolated Re2 and Ma.a Bond Dissociation Energies in ~X8 n- Complexes a -1 w (cm ) e X (cm-1) e BDE (kcal/mol) K4 [M~Cl8 ] 347 0. 50 ± o. 05 172 ± 20 Rb 4 l Mo..aC1 8 ] 339 190 ± 25 Cs 4 [MD.,aC1 8 ] 342 fu_-Bu 4 N\ [ R~Cl 8 ] 273 fu.-Bu 4 N\[Re2 Br 8 ] 276 o. 43 ± o. 05 o. 66 ± o. 07 o. 35 ± 0. 05 o. 39 ± 0. 06 Compound a Data from ref. 26. - 127 ± 15 152 ± 20 139 ± 25 49 ~ Figure 1. Electronic absorption spectrum of (n-Bu 4 N) 2 f Re 2 Cl 8 ] in acetonitrile solution at 298 K in the region 600-350 nm. Figure 2. Single crystal polarized spectra of bands I - III in fu-Bu 4 N) 2 [Re2 Cl 8 ] at 15 K. Bands II and III were recorded at a vertical scale expansion 2. 5 times that used for band I. The measurements were made at a spectral bandwidth less than 0. 75 nm. 50 0 o · (.C) 0 0 ...-Lf) E -c ....._.. ,< 0 0 c;;t o · o 0 0 r<) (\J (,_W:) 1-V'J).3 0 0 51 0 r------~--~g H 0 I.[) I.[) 0 --- ~ E C ---.-< 8 <:j" aJUOQ..JOsq\]' 52 2 Emission Spectra and Lifetimes of Re2 Cl 8 - , at 1. 3 K U on Excitation of the o - o* Transition William C. Trogler, Edward I. Solomon, and Harry B. Gray* Contribution No. 5609 from the Arthur Amos Noyes Laboratory of Chemical Physics, California Institute of Technology, Pasadena, California 91125 ~: l Re Br 2 ] 8 Excitation (650 nm) of (n-Bu 4 N}z[Re2 Cl 8 } and fu-Bu 4 N)zin KBr disks at 1. 3 K gives broad emission centered at _1 13,020 and 12,530 cm , respectively. Emission lifetimes are as follows: Re Cl 2 2, 8 150 ns; Re Br 2 also recorded for the Mo2 Cl 8 4- 2, 8 110 ns. Luminescence was ion doped into a crystal of (enH2 )C½. Again a broad, structureless emission band was observed (14,950 cmT 75 ns, 1. 3 K). Owing to the lack of overlap 0f the emission with the 0-0 transition in absorption, it is proposed that the 3 luminescent state is one of the spin-orbit components of Azu· This places the o - o* singlet-triplet splitting in the range 1000-3000 cm- 1 • 1 53 Absorption spectroscopy has been employed to characterize the lowest spin-allowed transitions in compounds that contain a metal1 metal quadruple bond. With the exception of dimolybdenum(II) 2 carboxylates, the lowest singlet excited state in dinuclear complexes 1 of Mo(II) and Re(III) is A u (in D h microsymmetry) derived from the 2 4 3 one-electron excitation 5 --- 5*. The location of A u (o - o*) (or, 2 more properly, its A1 u and Eu spin-orbit components) in these complexes, however, has been the subject of some speculation. And, although scattered wave Xa calculations have been performed for 2- 3 4- 4 5 Re Cl Mo Cl and Mo (0 CH) it is generally agreed that , 2 , 2 8 , 2 8 2 4 the level of sophistication is not sufficient to predict singlet-triplet splittings with any confidence. • . 6 • Clark and Franks noticed a weak, structured absorption at 6250 cm -1 4- Mo2Cla in the diffuse reflectance spectra of some but not a ll of the salts they examined, and the suggestion was made that this 1 3 near IR band might be attributable to the A1 g - A 2 u (o - 5*) transition. 1 However, it has been pointed out 7 that the 12,000 cm - singlet-triplet splitting that follows from such an interpretation is unexpectedly large , 7 8 and the possibility of near IR impurity absorption has been raised. ' 1 4- 3 -1 An alternative placement of A1 g - A 2 u in Mo 2Cl 8 could be 17, 500 cm , where Fanwick et al. 7 found a weak shoulder in one polarized absorption spectrum of a single crystal of ~Mo Cl 2 • 8 2H 0. 2 54 In an attempt to gather more evidence concerning the position 3 of the A2u state in quadruply-bonded dinuclear complexes, we have measured the emission spectra and lifetimes at 1. 3 K of (:Q_-Bu4 N) 24 [ Re2Cla] , (!!_-Bu4 N)J Re 2Br8 ] , and 0. 2% Mo 2Cl 8 - in a single crystal of (enH 2)Cl2. The results reported herein support a relatively small singlet-triplet splitting of the o - 6* excited states. Experimental Section The compounds (n-Bu 4 N)z[Re2 Cl8 j and (n-Bu 4 N)z[Re2Br 8 ] were prepared by the literature method. 9 Crystals of (enH2)C~ doped with Mo2 Cl 8 4- were prepared by the followirg modification of the 10 1 literature synthesis of (enH2 )z[Mo2 Cl 8 ] . Mo2 (02 CCH3 ) / was added to a 6N HCl solution saturated with (enH 2)Cl 2 • The solution was filtered before slowly cooling to 5 °C, yielding a crop of translucent purple crystals. Calculated for C2 H10N2 Clz: C, 17.93 %; H, 8.22 %; N, 20. 91 %; Cl , 52. 93 %. Found: C, 17. 82 %; H, 7. 48 %; N, 20. 47 %; Cl , 54.13%; Mo, 0. 35 %. The Mo analysis corresponds to a mole fraction of Mo2 Cl 8 4- of ca. 0. 2% (assuming Mo2Cl8 4- . is the only Mo species present). The C, H, N, and Cl analyses were obtained from Chemalytics , Inc., and the Mo analysis was performed by Schwarzkopf Microanalytical Laboratory. Emission spectra were obtained at the A point of liquid He in an optical dewar. A pulsed nitrogen laser was used to pump a dye laser (output pulse width of 32 ns) and excitation wavelengths of 650 and 540 nm were used for the Re(III) and Mo(II) compounds, respectively. Emission lifetimes we r e recorded at the wavelength of maximum emission intensity . Boxcar integration techniques were applied to the signal from a GaAs phototube for data collection. A resolution of 55 0. 6 nm was maintained for the emission spectra. Emission spectra were corrected for phototube and monochromator response before plotting. Samples of fu.-Bu 4 N) 2 [Re2 Cl 8 J and fu.-Bu 4 N)z [Re2 Br8 ] consisted of crystals suspended in 13 mm diameter KBr pressed disks. The low-temperature single crystal polarized absorption spectra of the crystal doped with Mo 2 Cl 8 4- (vide supra) were recorded 1 employing experimental conditions as previously described. Parallel 12 polarization is for the long axis (c), w.ai.ch was also an extinction direction of the (enH 2)Cl 2 crystal. The crystal was mounted over a pinhole on a copper disk with heat-conducting copper grease3 Results 1 Upon 650-nm excitation of the A1 g - 1 A2 u (6 - o*) transition 1 of (n-Bu4 N) 2 [ Re 2 Cl 8 ] , a broad emission centered at 13,020 cmis observed (Figure 1) with a lifetime of 150 ns at 1. 3 K. Two features of the emission spectrum are extremely important. Firstly, it is evident that a mirror symmetry relationship does not exist between 1 1 the emission band and the A1 g - A 2 u absorption system. Notably, the sharp vibronic structure observed in absorption is not present in emission. Secondly, and more importantly, there is no overlap between 1 the emission band and the 0-0 transition of A1g - 1 A 2 u that has been positioned from single crystal polarized absorption measurements. 13 Similar 650-nm excitation of (n-Bu4 N)J: Re 2Br8 ] results in broad 1 • emission centered at 12,530 cm- with a lifetime of 110 ns at 1. 3 K. The 490-cm -1 red shift of the emission maximum from that in Re 2 Cl 8 2- is reasonable for a o - 6* type transition, considering the fact that 1 A1 g - 1 A 2 u (o - o*) in Re 2 Br8 - is red-shifted in absorption by 586 cm- • 14 2 1 56 4- The related species Mo 2Cl8 behaves analogously to the Re(III) compounds, as excitation at 540 nm gives a broad, structureless emission centered at 14950 cm -1 (Figure 2) with a 75 ns lifetime at 1 3 K; however, the separation between the emission and absorption o systems is nearly twice that in the Re(III) compounds. It is also evident in this case that the emission band does not overlap the 0-0 transition reported 7 at 180 83 cm -i for a single crystal of K4 Mo 2Cl 8 • 2H 0. As the polartzation of this band has been established 7 to be 2 •parallel to Mo-Mo, the results for the doped crystal prove that the Mo2Cla 4- ion orients with Mo-Mo directed primarily along the needle axis of the crystal. A broad absorption centered at 405 nm, which is not attributable to Mo2 Cl 8 4, is present in the doped crystal. This absorption does not occur in a pure (enH 2)Cl 2 crystal. When a 6N HCl solution containing K.iMo 2Cl 8 is allowed to stand overnight under anaerobic conditions, the 530 nm absorption decreases at the same time a new band grows in at 418 nm. The latter absorption is present in solutions of Rb 3 Mo 2Cl8 H 15 4prepared from oxidation of the Mo 2 Cl8 ion in concentrated HCl. 3- Therefore it is likely that Mo 2Cl 8 H or a related species is trapped during growth of the doped crystals. Because the 405 nm band was found not to be polarized, it appears that this species is randomly incorporated into the (enH 2)Cl 2 lattice. Emission related to this second component was not observed, as only the lowest absorption was pumped using 540 nm radiation. Observation of emission in a doped crystal, where the rate of resonance energy transfer should be small, shows 57 that impurity emission from an exciton trap probably does not occur. Discussion For Re 2 Cl 8 2- and Mo 2 Cl 8 4- the lack of overlap between the emission band and the 0-0 peak in absorption strongly rules against l an assignment of simple fluorescence from the A 2 u state. Unfortunately, the emission lifetimes are not definitive in this respect, as they are consistent with either fluorescence and little radiationless decay 16 or "phosphorescence II with a large contribution from nonradiative deactivation pathways. However, two additional observations are l inconsistent with a A2 u emitting state, namely, the absence of vibronic structure for the Re 2 Cl8 separation in Mo 2 Cl 8 4• 2- emission and the increased emission/absorption Ligand substitution effects in the Re(ill) complexes, as well as the similarity of the absorption and emission half-bandwidths and the overall band profiles, suggest that luminescence occurs from a l state that is bonding to the same extent as is A2 u· For the reasons just stated, then, we believe the emitting state is one of the spin-orbit com3 l ponents of A 2 u (A1 u or Eu). Assignment of the emission to A2 u l A1 g, however, is not entirely out of the question, as a large geometrical 1 distortion in the A2 u state of Mz)C 8 n- complexes could lead to nonover17 lapping emission and absorption systems. The Duschinsky effect could then operate to give an emission band shape that is considerably distorted from that of the corresponding absorption. 18 For the M X n- (M = Re, 2 8 Mo; X = Cl, Br; n = 2 or 4) compounds, the excited state distortion most likely would involve an eclipsed (D4 h) to noneclipsed (D4 d) conformational change. Even if this type of distortion occurs, it apparently does not affect the a 1 vibrational modes to a large extent, as the ground-to-excited 58 state frequency shifts of a 1 (ReRe) and a 1 (ReReCl) are small. 13 4 4 Additionally, the calculations of Norman and Kolari ' 5 for Mo Cl 2 8 suggest that chloride-chloride attractions contribute to the stabilization l of the eclipsed geometry. Thus there is doubt whether the A2 u state would distort at all from the eclipsed configuration. The evidence favors assignment of the emission to one of the 3 3 transitions A1 u( A2 u) or Eu( A 2 u) l 1 A1 g. This interpretation places 3 -1 the singlet-triplet ( A 2 u- A 2 u) separation in the range 1000-3000 cm . This range is quite in line with the only documented singlet-triplet splitting in a dinuclear Re(III) or Mo(II) compound, which is 1800 cm 3 1 2 for the ' Eg(o - 1r*) excited states of Mo 2 (O2 CCF 3 ) 4 • Further elucidation of the nature and position of the low-lying triplet excited states in M;½: 8 n- complexes will require much additional theoretical and experimental effort. Acknowledgments We thank Ib Trajberg for experimental assistance and Professors F. A. Cotton and D. So Martin, Jr. for communication of their results 7 prior to publication. Professor W. A. Goddard III is acknowledged for helpful discussions. VI. C. T. held an IBM Fellowship for 1976-1977. This research was supported by the National Science Foundation (CHE-7519086). -1 59 References 1. W. C. Trogler, C. Do Cowman, H. B. Gray, and F. Ao Cotton, J . Am. Chem. Soc., ,,..__,,.__ 99, 2993 (1977), and references therein. 2. W. C. Trogler, E. I. Solomon, Ib Trajberg , Co J. Ballhausen, and H. B. Gray, Inorg. Chem. ,}i, 828 (1977); F. Ao Cotton, D. S. Martin, T. R. Webb, and To J. Peters, Inorg. Chem., 15, ~ 1199 (1976); F. A. Cotton, D. S. Martin, P. E. Fanwick, T. J. Peters, and T. R. Webb, J. Am. Chemo Soc., .,..._,..,_ 98, 4681 (1976) . 3. A. P. Mortola, J. W. Moskowitz, N. Rosch, C. P. Cowman, and H. B. Gray, Chem. Phys. Lett., .,..._,..,_ 32, 283 (1975) . 4~ J. G. Norman and H.. J. Kolari, J. Am. Chem. Soc., 97, 33 (1975). 5. J. G. Norman, H.J. Kolari , H. B. Gray, and W. C. Trogler, ~ lnorg. Chem., .,.,_,.,_ 16, 987 (1977) . 6, R. J. H. Clark and M. L. Franks, J. Am. Chem~ Soc., "'-'"'97, 2691 (1975) . 7. P. E. Fanwick, D. S. Martin, F. A. Cotton, and T. R. Webb , Inorg. Chem. , in press. 8. In this connection, it should be mentioned that the spectrum of the oneelectron oxidation product of Mo 2 (SO'ot) 4 absorption band at 7100 cm -1 4- exhibits a structured 6 - 6* (D. K. Erwin, G. L. Geoffroy, H. B. Gray, G. S. Hammond, E .. I. Solomon, W. C. Trogler, and A. A. Zagars, J. Am . Chem. Soc.) .,..._,..,_ 99,3620 (1977)) . 9. F. A. Cotton, N. F. Curtis, B. F. G. Johnson, and W. R. Robinson, Inorg. Chem., 4, .,.... 326 (1965) . 10. J. V. Brencic and F. A. Cotton, Inorg. Chem. , ~, 2698 (1969) . 11. T. A. Stephenson, E. Bannister, and G. Wilkinson, J. Chem. Soc., 2538 (1964). 60 12. T. Ashida and S. Hirokawa, Bull. Chem. Soc. Jap., 36, .,...._,.._ 704 (1963) . 13. C. D. Cowman and H. B. Gray, J. Am. Chem. Soc . , "-"95, 8177 (1973) . 14. C. D. Cowman, Ph.D. Thesis, California Institute of Technology, 1974. 15. M. J . Bennett , J. V. Brencic, and F. A. Cotton, Inorg. Chem., 8, 1060 (1969); F . A. Cotton and B. J. Kalbacher, Inorg. Chem. , ,-,. 15, ,.,_,.,_ 16., 522 (1976). J . B. Birks, "Photophysics of Aromatic Molecules 11 , Wiley- Interscience, New York, N. Y., 1970, pp. 87-88. 17. F. Duschinsky, Acta . Physiochim. U. R.S.S., 1,551 (1937). 18. C. J. Ballhausen and A. E. Hansen, Ann. Rev. Phys. Chem., .rv-. 23, 15 (1972). 61 Figure Captions Figure 1 Corrected emission (1. 3 K) and absorption (15 K) of (n-Bu N) 4 Figure 2 [ 2 Re Cl 2 ] 8 in a KBr disk. Corrected emission (1. 3 K) and absorption (15 K) spectra of a single crystal of (enH2 )C½ doped with 4- Mo2Cla • (0. 2 mol%). Figure 3 Single crystal polarized absorption spectra at 15 K of a single crystal of (enliz)C½ doped with Mo Cl 2 48 (0. 2 mol%). Polarizations are denoted parallel and perpendicular to the crystal needle axis. 62 EMISSION INTE1\JSITY I'\) 0 0 0 (>J 0 0 0 0 ~ I ~ 0 0 0 0 I 0 0, 0 0 0 (}) 0 0 0 ABSORBANCE 63 EM;SS!ON INTENSITY +> 0 0 (J) 0 0 0 0 s I m 0 0 ru 0 0 0 ABSORBANCE 64 II POL w 0 z <t co a: 0 <f) co <t J_ POL 400 500 NM 600 65 Studies of the Polarization Behavior, Temperature Dependence, and Vibronic Structure of the 23, 000 cm -1 Absorption System in the Electronic Spectra of Mo2 (02 CCH3 ) 4 and Related Compoundso Emission Spectrum of Mo 2 (02 CCF 3 ) 4 at 1. 3 K Contribution No. 5447 from the Arthur Amos Noyes Laboratory of Chemical Physics, California Institute of Technology, Pasadena, California 91125 Abstract: The 23, 000 cm -1 absorption band in the electronic spectrum ~ of Mo (0 CCH 2 2 ) 3 4 is primarily polarized perpendicular to the metal- metal axis. Rich vibronic structure is observed for this absorption system in the spectra of Mo (0 CCH 2 2 ) 3 Mo (0 CCD , 2 4 2 ) 3 Mo (0 CCF , 2 4 2 ) 3 , 4 and Mo 2 (0 2CH) 4 at 15 K. Analysis of the temperature dependence of the hot bands in the spectrum of Mo (0 CCF 2 2 ) 3 4 has established that the transition is vibronic, being allowed primarily by an a u (MoMoO) 2 -1 bending vibration (~ 190 cm ). Interpretation of the vibronic structure in the spectrum of Mo2 (02 CCH 3 ) 4 suggests that the 0-0 transition is split by 275 cm -1 into x- and y-polarized components. This splitting, which is much larger in Mo 2 (02 CCF 3 ) 4 , correlates with departures from ideal D h symmetry that are evident from crystal structure 4 determinations. A dominant progression in the MoMo totally symmetric 66 stretch is observed on each of several vibronic origins in the spectrum of Mo 2 (O2 CCH 3 ) 4 o Franck-Condon analysis demonstrates 0 that a 0. 1 A elongation of the MoMo bond occurs in the excited state. From the lowest energy vibronic origin in Mo 2 (O2 CCF 3 ) 4 , an estimate of 90 cm -1 for the metal-metal torsional frequency is obtained. Structured emission is observed for Mo 2 (O2 CCF 3 ) 4 at 1. 3 K ( T ~ 2 ms). The emission origin lies 1800 cm -1 lower than that of the absorption systemo The results are consistent with the assignment of the absorption band to the orbitally-forbidden, metal-localized transi1 tion A1 g - 1 Eg (o - 1f*), and the emission to the corresponding triplet -► singlet. 67 Introduction ~ It is now well established that the low-energy absorption 3 system of moderate intensity (E ~ 10) in the electronic spectra of Re 2Cl8 2, Re 2 Br8 2, Re 2Cle( P(CJI 5 ) 3 ] 2, and Mo 2Cl 8 4- is attributable to the electric-dipole-allowed transition o - o* 1 1 2-4 ( A1 g - A2 u). By contrast, spectra of :Mo2(O2CR\ (R / 2 aromatic) exhibit only a weak low-energy band (E ~ 10) which -1 peaks at about 440 nm (~ 23,000 cm ). 5 Recently, Cotton, Martin, and coworkers have examined the polarized crystal spectra of Mo 2 (0 2CCH 2NH 3 ) 4 (S04 ) 2 • 4H 20 6 and Mo 2(02 CH) 4 • 7 Although a firm assignment of the weak system was not made, it was shown 6 conclusively that the transition in question in the Mo 2(0 2CR) 4 -type 1 compounds could not be o - 6* ( A100 • - 1 A 2u)o We have examined the polarized spectra of a crystal of Mo 2 (0 2CCH 3 ) 4 as well as films of Mo 2 (02CCD3 ) 4 , Mo 2 (02 CCF3 ) 4 , and Moz{O2CH) 4 at 15 K. Particular attention has been paid to the eluci- dation of the hot bands in the spectra of Mo 2 (02CCH3 ) 4 and Mo2(02CCF 3 ) 4 • The emission spectrum of Mo 2(O2CCF 3 ) 4 at 1. 3 K has also been measured o Our data suggest that the 23, 000 cm -1 system in molybdenum(II) acetate is derived from a metal-localized, electric-dipole-forbidden transition 1 to an excited state of Eg symmetry. Ex erimental Section The compounds Mo 2(02CCH3 ) 4 , 8 Mo 2(02CCF 3 ) 4 , 9 Mo2(O2CH) 4 , 10 and K.iMo 2Cl 8 11 were prepared and purified by standard literature methods. Satisfactory elemental analyses were obtained in all cases. 68 The deuterated complex Mo (O CCD 2 2 ) 3 4 was prepared by suspending 0o 6 g (0. 93 mmol) of Mo 2 (O2 CCF 3 ) 4 in 10 ml of deuterated acetic acid. After warming and stirring for 2 h, the pale yellow powder was filtered and washed with ethanol, pentane, and dried in vacuo. Calcd for Mo (O CCD 2 2 ) 3 • 4 C, 21. 83; D, 5. 49. Found. C, 21. 4; D, 5. 2. -1 Low energy (100-500 cm ) infrared spectra were recorded on a Perkin Elmer 180 spectrometer. Sample temperatures of 40 K were attained in a Cryogenics Technology Model 20 Cryocooler, which was equipped with polyethylene windows. Samples of molybdenum(II) carboxylates for the low energy IR spectra were Vaseline mulls spread on a thin polyethylene disk. Infrared -1 spectra in the range 300-4000 cm were recorded on samples of KBr pellets using a Perkin Elmer 225 instrument. All samples were prepared under a nitrogen atmosphere. Electronic spectra were recorded on a Cary 17 spectrometer. Low temperatures (15 or 80 K) were obtained with a Cary liquid helium dewaro Temperature was monitored with a calibrated carbon resistor next to the sample on the cooling block. The temperature dependence study was performed by allowing controlled increases from 80 to 300 Kover a 5 h period. Spectra of Mo (O CCH 2 and Mo (O CCD 2 2 ) 3 4 2 ) 3 , 4 Mo (O CCF 2 2 ) 3 , 4 Mo (O CH) 2 2 , 4 were obtained from sublimed films on optasil quartz disks (1/2 in. diameter; 1/32 in. thick) that were mounted in the dewar using heat-conducting copper grease for good thermal 69 contact. Pellet spectra (KBr) were also recorded and resembled the film spectra but were of lower quality. Slitwidth and lamp settings were such that a spectral bandwidth of less than 0. 07 nm was maintained for all low temperature optical spectra. The emission spectrum was measured at 1. 3 K, obtained by pumping on liquid helftum in an optical dewar. A pulsed nitrogen laser (337 nm) was used to excite the sample whose emission was recorded at 0. 6 nm resolution using boxcar integration techniques. The spectrum was corrected for phototube response before plotting o Both Mo (O CCH ) / 2 and Mo (O CCF ) 9 crystallize in the 2 2 3 2 2 3 4 triclinic space group Pl. Crystals of reasonable size were obtained only for MoiO2 CCH3 ) 4 • The needlelike crystals were observed to grow with two kinds of well-formed faces . These types were readily distinguished , as one face exhibited extinction directions nearly parallel (within 2°) and perpendicular to the needle axis, whereas the extinction directions for the other face made an approximately 15 ° angle with the needle axis. Polarized spectra were recorded for the former type of crystal face. Subsequent X-ray oscillation, Weissenberg, and precession photographs showed conclusively that the needle axis was the crystallographic b axis, and that polarized spectra were taken II and J. to b within 10 ° of the be crystal face. Polarization ratios were calculated for transitions assumed to be polarized along the MoMo(z) and the mutually perpendicular 70 MoO(x, y) axes. These calculations were performed by projecting the molecular axes along the extinction directions. Adopting the same notation used in the crystal structure analysis, 12 we report the following polarization ratios 11b/J. b for the be crystal face: z = 3. 61, x = 0. 004 (MoO(l), 0. 008; MoO(3), 0. 000), y = 0. 55 (MoO(2), 0. 618; MoO(4), 0. 476). The calculations show that the assumption that the principal absorption directions lie along the MoMo and MoO axes does not hold rigorously for the molecule in the low-symmetry crystal environment, and we emphasize that uncertainties are to be associated with the reported polarization ratios; nevertheless, the calculations are entirely consistent with the observation of strong parallel- and both strong and weak pe rpendj_c ularly-polarized features. The oscillator strength of the 23,000 cm -1 band in Mo2 (O2CCF 3 ) 4 was determined from measurements on an oxygen-free acetonitrile solution at room temperature. From a plot of the spectrum on an energy scale, the oscillator strength was found by cut-and-weight integration of the peak, allowing for the overlapping tails of intense bands at 297 and 335 nm. For KiMo 2 Cl8 , a freshly-prepared, de- gassed (20° C, 6 N HCl) solution was employed to obtain an oscillator strength. Results and Discussion The oscillator strength of the 22,700 cm -1 band in the absorp- tion spectrum of Mo 2 (O2 CCF 3 ) 4 in acetonitrile solution at 298 K is 0. 0011. This value is an order of magnitude less than that (0. 015) 13 71 obtained for the o - 5* band at 19, 300 cm -1 in K4 Mo 2 C lg . Exam- ination of the spectra of Mo 2 (O2 CCF 3 ) 4 and Mo 2 (O2 CH) 4 at 40 K between 4000 and 22, 000 cm -1 revealed no lower energy electronic bands of appreciable intensity (E > 10). Upon cooling, the 23,000 cm band in each of four complexes (Mo (O CCH Mo (O CCD Mo (O CCF and Mo (O CH) 7 shows remarkable vibronic structure ) 2 ) 2 2 3 , 4 2 3 , 4 ) 2 2 3 , 4 ) 2 2 4 (Figures 1, 2; Tables 1-4). The integrated intensity of this band increases slightly (10-15%) and a red shift is exhibited between 15 and 300 K. It is unlikely that the weak band represents a spinforbidden transition, as a similarly-structured emission system -1 (T 2 ms) was observed at lower energy (0-0 gap of 1800 cm ) in a film of Mo2 (O2 CCF 3 ) 4 at 1. 3 K (Figure 3; Table 5). From the oscillator strength of the absorption band, a radiative lifetime of 14 2 µs is estimated. Therefore, the long-lived emission is logically assigned to a transition from the triplet state that is related to the -1 singlet excited in absorption at 23,000 cm . The Mo (O CCH 2 2 ) 3 4 complex does not emit, which may reflect the more efficient radiationless processes expected for a molecule possessing high frequency (CH) vibrations in the weak coupling limit. 15 Upon examination of the low-energy side of the first intense -1 vibronic component in the spectrum of Mo 2 (O 2CCF 3 ) 4 (22,064 cm ; A, Figure 4), using a thick sample, additional weak features were observed at 21,950 and 21,862 cm -1 (B, C, Figure 4) . At 15 K these weak features cannot be hot bands on the 22,064 cm -1 "origin", as their intensities are several orders of magnitude too large. This 1 72 observation implies that there must be a lower energy forbidden origin. When the temperature is increased to 80 K, hot bands 1 (D, E, Figure 4) are exhibited at 21, 773 and 21,671 cm - • The temperature dependence of the la tter hot band is clearly resolved over a large temperature range (80 - 260 K), whereas the 21,773 cm hot band was obscured by broadening of the more intense 22,064 cm -1 -1 component at higher temperatures . As a hot band arises from thermal population of ground state vibrations, its temperature dependent intensity should follow a Boltzmann distribution. In Figure 5 the Boltzmann plot for the 21,671 cm -1 hot band is shown. The slope obtained corresponds to a ground state vibrational spacing -1 of 206 ± 20 cm . This means that the hot band is built on a weak electronic origin (i.e. , the true 0-0 energy for the transition) approximately midway between 21, 671 cm - 1 and the first intense -1 vibronic component (22, 064 cm ) of the transition. In fact, a weak feature is observed at 21,862 cm -1 (C, Figure 4), where the hot band analysis predicts the 0-0 transition to occur. Moreover, the intensity of this weak peak is temperature independent (observable between 15 - 80 K), ruling out its assignment to a low frequency, vibronicallyallowed transition. Therefore, we conclude that 21, 862 cm -1 is the 0-0 transition energy, and that its intensity is derived from magneticdipole allowedness . 16 The 0-0 feature is about 100 times less intense than the vibronic portion of the system, which places its oscillator 5 strength at about 10- (a reasonable order of magnit ude for a magnetic1 dipole t r ansition). 17 As will be shown later, the excited s tate is Ecr b 73 l • l in D4 h, and A1 g - Eg is allowed as a magnetic-dipole transition. 2 Because the crystallographic symmetry is and the 0-0 c/ transition will be shown to be even - even, the electric-dipole forbiddeness and magnetic-dipole allowedness are rigorous. Similar hot bands are observed in the spectrum of Mo2 (O2 CCH 3 ) 4 , and their polarizations are shown in Figure 6. Unfortunately, resolution of the vibronic structure was not sufficient to observe the 0-0 band directly. The 21,302 cm -1 hot band (A, Figure 6) and the first intense vibronic component at 21,696 cm -1 (A, Figure 1) are polarized similarly. This provides additional evidence that this hot band and first intense component derive their intensities by coupling the same non-totally symmetric vibration (ground state 191 cm -1 for the hot band and excited state energy of 202 cm -1 for the first intense component) with an electric-dipole-forbidden 0-0 transition approximately midway between them. The spectrum of Mo 2 (O2 CH) 4 also exhibits a similarly-behaved hot band ~ 375 cm -1 to lower energy of the first intense peak of the system (Table 4). Therefore, the allowedness of the first vibronic component of the 23, 000 cm -1 . band in Mo 2 (O2 CCH3 ) 4 , Mo 2 (O2 CCF 3 ) 4 , and Mo2 (O2 CH) 4 is attributed to coupling with a ~ 190 cm vibration. -1 An important feature of the polarized spectra (Table 1) of Mo 2 (O2 CCH 3 ) 4 is the presence of vibronic components with three different polarization ratios. For example, the first intense com- ponent (A, Figure 1) is strong]y polarized perpendicular, the weak 74 feature at 21,870 cm -1 (C, Figure 7) is polarized parallel, and -1 the intense peaks at 21, 969 and 22, 244 cm (C, E, Figure 1) are weakly polarized perpendicular to the metal-metal axis. The presence of two perpendicularly polarized bands (strong and weak) suggests a low symmetry splitting into x and y components with different polarization ratios, as expected (see Experimental Section), owing to the alignment of the molecule with respect to the crystallographic axes. There are two possible explanations for the low symmetry splitting. If the excited state involved in the forbidden 1 transition were orbitally degenerate (i.e., Eg) in D4 h symmetry, then in the Ci site of the crystal a splitting of the electronic origin would occur, producing two different components with x and y polarization ratios (a Davydov splitting can be ruled out, as there is only one molecule in the unit cen 12). Alternatively, the split transition could arise as a result of a nondegenerate electronic excited state 1 1 1 1 ( A1 g(u), A2g, B1 g(u), or B;,-g(u)) coupling with a low-symmetrysplit eu (g) vibration of the molecule. To investigate the latter possibility, we have examined the infrared spectra of mulls of Mo 2 (O2 CCH3 ) 4 , Mo2 (O2 CCF 3 ) 4 , and Mo2 (O2 CH) 4 in the 100-500 cm intense band system centered at 360 cm -1 ponent at 372 cm region at 40 K. An is exhibited in the spectrum of Mo2 (O2 CCH 3 ) 4 • This system is split by 25 cm -1 -1 -1 into a sharp com-1 and a somewhat broader one at 347 cm . Force constant calculations 18 have shown that this intense system arises from the IR active a 2 u and eu MoO stretching vibrations . The broader band, which presumably represents the eu vibration, has a linewidth 75 -1 of ca. 30 cm . Bands were also observed in the MoMoO bending region at 175 and 228 cm -1 for Mo 2 (O2 CCH 3 ) 4 , at 170 and 232 cm for Mo2 (O2 CH) 4 and 190 and 216 cm -1 for Mo 2 (O2 CCF 3 ) 4 • -1 Force constant calculations for Mo 2 (O2 CCH 3 ) 4 have suggested that these bands are attributable to a u and eu MoMoO bending vibrations. 18 2 As none of the eu vibrations is split sufficiently to explain the separation of the x and y polarized vibronic components, the elec1 tronic transition must be to a split Eg state. Of interest is the fact that all three compounds show a vibrational band in the energy -I range (170-190 cm ) where the hot band analysis placed the vibronically active fundamental for the first intense origin. Our analysis suggests an a 2 u assignment for this band, as a 2u would induce x and y polarized I components in a transition to a split Eg state, according to the 19 Herzberg-Teller coupling scheme. -1 The vibrational f:ine structure of the 23, 000 cm band may be assigned after proper account is taken of the following: As the pure electronic transition is forbidden, all of the strong features at 15 K must arise from excitation of a quantum of one of several possible antisymmetric vibrations in the excited state (Table 6). A progression of spacing 355-370 cm -1 is built on each vibronic origin (three strong ones for Mo 2 (O2 CCH3 ) 4 and only one for Mo 2 (O2 CCF 3 ) 4 ). • Weak features pose a problem, as they can be either new vibronic origins or progressions on one of the intense vibronic bands with a small Franck-Condon factor. For example, if the Franck-Condon factor is 0.4, a progression with relative intensities of 1:0.4:0. 08:0. 01. .. 76 is expected. Therefore, in a complex spectrum, only the first two bands would be observable. Only one dominant progression, which must represent quanta of a totally symmetric vibration, is exhibited by Mo2 (O2 CCF 3 ) 4 -1 -1 (ca. 355 cm ) and Mo2 (O2 CCH 3) 4 (ca. 370 cm ). There are only two a 1 g modes of the Mo 2 0 8 unit that fall in this energy region. Bands attributable to the awt, (MoMo) stretch were observed in the Raman - 1 18 -l 9 spectra of Mo2 (02 CCH3 ) 4 (406 cm ) and Mo 2 (02 CCF 3 ) 4 (397 cm ). 1 A Raman peak also was observed 20 at 322 cm - for Mo (0 CCH ) , 2 2 3 4 where force cons~ant calculations place a 1 g (MoO). 18 As the average :rvioO bond length in Mo 2 (02 CCF 3 ) 4 is shorter than that in Mo2 (02 CCH3 ) 4 (vide infra), we would expect the a 1g (MoO) stretch to occur at higher energy in the former complex; however, the progressional energies display the opposite trend. 370 cm -1 For this reason, the spacings of 355 and are logically attributable to an excited state a 1 g (MoMo) stretch that is reduced somewhat from the ground state value. The emission system of Mo 2 (O2 CCF 3 ) 4 exhibits one major progression with -1 a spacing of ca. 390 cm , which is in good agreement with the Raman -1 9 band at 397 cm assigned to a 1 g (MoMo). A progression in the metal-metal stretching frequency reflects a displacement of the excited state geometry from the ground state along this normal coordinate. The intensity 111 of then 'th member of 21 a vibrational progression is given by the Poisson distribution I n = I oSn/n'• 77 The Franck-Condon factor S is most simply obtained from the ratio 11 /Io. This is related to the distortion, Q 0 a, from equilibrium along an a 1 g mode of frequency wa by: S = (1/2)k(Q 0 at/hw 01 • Force constant calculations on Mo2 (O2 CCH 3 ) 4 have given kMoMo ~ 1 3. 73 mdyne/Afor the 406 cm- a 1 g mode (~ 75% MoMo stretching). 18 Using these parameters and a value S = 2. 2 obtained from the electronic spectrum, we calculate Q MoMo ~ O. 1 A. 0 As mentioned above, the most intense vibronic component is allowed by coupling the 0-0 transition with an a 2 u vibration. In the ground state, the a 2 u(MoMoO) vibrational quantum for Mo.:/O2 CCF 3 ) 4 -1 is 191 cm , and similar values are found for Mo2 (O2 CCH 3 ) 4 and Mo2 (O2 CH) 4 • According to vibronic selection rules (Table 6), either an a 1 u, a 2 u, b1 u, or b 2 u mode can induce perpendicularly polarized 1 1 intensity into A1 g .... Eg. If we assume that the molecular symmetry is reduced to D2h, we note that a 1 u and b1 u (a2 u and b 2 u) correlate with 1 an au (b1 u) vibration, which induces x (y) polarization in the Ag 1 1 B2g component of the parent A1 g 1 tion to Ag - 1 1 Eg transition, and y (x) polariza- B3g, thereby accounting for the observed behavior. The a 1 u metal-metal torsional vibration is expected to be of lw fre-1 quency. In fact, the lowest energy weak feature (21,950 cm ; B, Figure 4) in Mo 2 (O2 CCF 3 ),c1 might be the a 1 u vibronic origin, which -1 would place the metal- metal torsional frequency at about 90 cm . The related hot band is observed at 21, 773 cm -1 (D, Figure 4). -1 Polarization of the corresponding feature (21,428 cm ; B, Figure 6) 78 in Mo 2 (O2 CCH3 ) 4 is consistent with the proposed assignment. The a 2 u, b1 u, and b2 u modes are all MoMoO deformations, -1 18 calculated to lie near 200 cm in Mo 2 (O 2CCH 3 ) 4 • We favor a 2 u as the dominant, vibronically active mode, particularly in view of the close agreement between the vibrational quantum derived -1 from the hot band (191 cm ) and the 190 cm -1 peak in the low temperature IR spectrum of MoiO2 CCF 3 ) 4 • The parallel polarized component of low intensity at 21, 870 cm -l (or ca. 370 cm -1 from the estimated position of the 0-0 band) in the spectrum of Mo 2 (O2CCH 3 ) 4 (C, Figure 7) must represent another vibronic origin (rather than a -1 quantum of an a 1 g vibration built on 21,696 cm ). It is likely that coupling with the eu (MoO) stretching vibration induces the z -1 polarization (ground state 347 cm ). A weak band at 21, 818 cm -1 that is strongly xy polarized in the spectrum of Mo2 (O2 CCH3 ) 4 is assigned to a transition involving -1 a quantum of a 2 u(MoO) built upon the 0-0 component at about 21,500 cm . -1 The two intense features at 21,969 and 22,244 cm , which are polarized weakly perpendicular, are attributed to vibronic origins on the second 1 component of the 0-0 A1 g - 1 Eg (vide supra). As the second 0-0 component, unlike the first, could not be located with certainty, definitive assignments are not possibleo If we assume that the 21,969 cm -1 peak represents a quantum of a 2 u(MoMoO) built upon 1 the second 0-0 component, then the low symmetry splitting of Eg is about 275 cm -1 -1 (i.e. , second 0-0 at ~ 21, 775 cm ) . 79 The 22, 244 cm -1 peak (E, Figure 1) is not a member of an a 1 g progression on either of the intense vibronic origins, because the third member of such a progression is not observed. Therefore, it is a vibronic origin on one of the two 0-0 components, corresponding either to a 469 or 744 cm -1 vibration. Such high energies cannot result from MoO vibrations, but must involve a carboxylate ligand deformation that lowers the symmetry of the Mo 2 O8 unit, thereby allowing the transition. Support for this suggestion comes from a comparison with the spectrum of the deuterated complex, Mo 2 (O2 CCD 3 ) 4 (Figure 2). Relative to the positions of the other peaks in the spectrum, only the 22,244 cm -1 band (E, Figure 2) displays a noticeable red -1 shift (~ 40 cm ) in the deuterated complex. Band D (Figure 1) is absent in Mo 2 (O2 CCD3 \ , which implies either a low or high energy shift to a position at which it is obscured by more intense peaks -1 at 21, 981 (C, Figure 2) or 22,095 cm . In IR spectra measured in KBr pellets at ambient temperature, a band attributable to OCO bending shows a similar red shift (675 to •·l 645 cm ). Although an absorption assignable to the pr(OCO) deformation was not observed in the IR spectrum of Mo2 (O2 CCH3 ) 4 , one 1 22 such feature occurs at 460 cm- in acetate ion. Therefore, we assign the 22, 244 cm -I peak to a vibronic origin involving a 469 cm OCO deformation built on the second 0-0 component. Band D is attributed to a transition involving a similar OCO deformation built on the first 0-0 component. We assume that the apparent absence -1 80 of band Din the spectrum of Mo 2 (O2 CCD3 ) 4 is due to a 40 cm --1 red shift, which results in the weak absorption being buried beneath band C (Figure 2). Bands F and G (Figure 1), which occur 318 and 320 cm -1 above C and E, respectively, are assigned to transitions involving a quantum of the a 1 g(MoO) stretching vibration built on the two strong vibronic origins. The polarizations of F and G match C and E, as expected, and force constant calculations predict a1 g -1 18 (MoO) to ba about 320 cm in the ground state. This a 1 g progression has a small Franck-Condon factor (ca. O. 3), and therefore diminishes according to 1:0. 3:0. 045. Hence, the third member in the progression would not be observable. The small Franck-Condon factor for the symmetric MoO stretch indicates that the excited state distortion along the normal coordinate is minimal. Because no polarization data or force constant calculations for Mo 2 (O2 CCF 3 ) 4 are available, the vibrational assignments given in Table 2 are less certain. Assignments are made by analogy with the interpretation of the spectrum of Mo 2 (O2CCH3 ) 4 • Discussion of the differences in the spectra of Mo 2 (O2 CCH3 ) 4 and Mo 2 (O2 CCF 3 ) 4 is facilitated by recalling the reported 9 ' 12 bond distances in the two complexes, as follows: 81 2. 121(4) o"- /o 2. 110(4) Mo 2. 107(5) 0 / " 0 2. 14(2) 20 137(4) o"- /o 2. 01(2) "short axis" Mo 1. 97(2) o/ "-o 2. 11(2) ''long axis'' The relative simplicity of the spectrum of Mo 2 (O2 CCF 3 ) 4 may be associated with the substantial distortion of the Mo 2 O8 unit from D4 h symmetry. Only one strong component is exhibited in the spectrum, and the medium intensity component (A, Figure 2) actually is comprised of two overlapping weak progressions (Table 2). The spectrum of Mo 2 (O2 CCH 3 ) 4 exhibits, by contrast, three strong progressional origins of comparable intensity, and band C (Figure 2), which does not red shift appreciably in Mo 2 (O2 CCD3 ) 4 (10 cm -1 or less), and which has no counterpart in Mo 2 (O2 CCF3 ) 4 , is one of the more intense vibronic features. It is logical to propose that, in the distorted Mo 2 O8 unit of Mo 2 (O2 CCF 3 ) 4 , one of the com1 ponents of A1 g - 1 Eg is moved to higher energy and is buried under more intense peaks, whereas the other component, which is associated with the "long-axis" (OMoO) 2 plane, occurs at approximately the same energy as in Mo 2 (O2CCH 3 ) 4 • The lack of sensitivity of this 0-0 component to ligand substituents (CH 3 or CF 3 ) rules out any interpretation based on charge transfer to or from the ligand 7T or 11* orbital. Considering the energies of the metal 6 and 6* orbitals, 23 82 the transition must be 6 - 1r* or 1r - o*. The latter assignment may be rejected, as Xa calculations 23 on Mo (O CH) have indicated 2 2 4 that o - 1r* falls below rr - o *, and, more importantly, that only the former transition is metal-localized(% Mo: 5, 89, 1r*, 96; rr, 65, o*, 86). We conclude, therefore, that the 23,000 cm 1 in Mo 2 (O2CCH 3 ) 4 is attributable to the transition A1 g - 1 -1 system Eg (o - rr*), -1 which is low-symmetry-split by about 275 cm 24 It has been shown that the o - 1r* band in the electronic 0 spectrum of Re 2 Cl 8 2- lies only 3700 cm -1 above that attributable to o - o*. Apparently, in the molybdenum(II) carboxylates, splitting of the o levels has increased to the point where o - 11* is now slightly lower in energy than o - o*, and it has been suggested 23 that the latter transition in Mo 2 (O2 CH) 4 be associated with an in-1 tense shoulder at 30, 800 cm O The difference in the spectral be423 havior of Mo 2 (O2 CH).1 and Mo 2Clg may be related to the fact that 0 the metal-metal bond distance is 0. 05 A shorter in the former complex, but further studies will be needed to clarify why the o o* splitting is affected more than o - 11T*. - We thank George Rossman for assistance with certain experimentso Professors F. A. Cotton, D. S. Martin, and J. G. Norman, Jr., are acknowledged for helpful discussions. W. Co T. held an IBM Fellowship during 1976-77. was supported by the National Science Foundationo This research 83 References and Notes ~ ; ~~ 1. (a) California Institute of Technology; (b) Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139; (c) Chemical Laboratory IV, University of Copenhagen, Denmark. 2. C. D. CowmanandH. B. Gray, J. Am. Chem. Soc., 95, 8177 (1973). 3. A. P. Mortola, J. W. Moskowitz, N. Rosch, C. D. Cowman, and -"-" H.B. Gray, Chem. Phys. Lett., 32, .,.._,._ 283 (1975) . 4. J. G. Norman and H. J. Kolari, J. Am. Chem. Soc., -"-"'97, 33 (1975). 5. L. Dubicki and R. L. Martin, Aust. J. Chem., -"-'"' 22, 1571 (1969). 6. F. A. Cotton, D. S. Martin, T. R. ·webb, and T. J. Peters, Inorg. Chem., 15, 1199 (1976). "'""' 7. F. A. Cotton, D. S. Martin, P. E. Fanwick, T. J. Peters, and T. R. Webb, J. Am. Chem. Soc., .......,._ 98, 4681 (1976) . 8. T. A. Stephenson, E. Bannister, and G. Wilkinson, J. Chem. Soc., 2538 (1964) . 9. 10. F. A. Cotton and J. G. Norman, J. Coard. Chem. , 1, 161 (1971). "' J. G. Norman and H. J. Kolari, Chem. Commun., 649 (1975). -"""""' 11. J. V. Brencic and F. A. Cotton, Inorg. Chem. , Q_, 351 (1970). 12. F. A. Cotton, A. C. Mester, and T. R. Webb, Acta Cryst., ~ ' 2768 (1974). 13. The relatively small o - o* oscillator strength is in line with theoretical expectations for such a one-electron N - V transition (R. S. Mulliken, J. Chem. Phys., 7, 20 (1939)). The effective dipole "' moment, Q, is given by Q ~ Sr in the model (valence bond) that is ' particularly appropriate for small orbital overlaps. The oscillator N 11 strength, fy, is found to be 1. 096 x 10 2 2 vS r , where Sis the orbital overlap, r the internuclear distance in cm, and v the transition energy in cm -i. Using \Vavefunctions for Mo+ (H. Basch and H. B. Gray, 84 Theoret. Chim. Acta (Berl.), 4, 367 (1966)), we find a o overlap -'"'of 0.10 for r = 2.14 x 10- 8 cm, which in turn predicts an oscillator strength of 0. 01 for a o - o* transition at 19,300 cm -i. 14. J. B. Birks, Photophysics of Aromatic Molecules, Wiley-Interscience, New York, 1970, pp. 87-88. 15. G. W. Robinson and R. P. Frosch, J. Chem. Phys., ~' 1187 (1963). 16. G. Herzberg, Molecular Spectra and Molecular Structure, Vol. III, Van Nostrand Reinhold Co. , New York, 1966, pp. 134-136. 17. C. J. Ballhausen, Introduction to Ligand Field Theory, McGraw-Hill, New York, 1962, pp. 185-186. 18. W. K. Bratton, F. A. Cotton, M. Debeau, and R. A. Walton, J. Coard. Chem., .,..,. 1, 121 (1971) . 19. Ref. 16, p. 140. 20. J. San Filippo and H. J. Sniadoch~ Inorg. Chem., ll_, 2326 (1973). 21. E. I. Solomon and C. J. Ballhausen, Mol. Phys. , ~' 279 (1975). 22. K. Nakamoto, Infrared Spectra of Inorganic and Coordination Compounds, 2nd ed., Wiley, New York , 1970, p. 223. 23. J. G. Norman, H.J. Kolari, H.B. Gray, and V.1. C. Trogler, Inorg. Chem., in press. 24. W. C. Trogler, C. D. Cowman, H. B. Gray, and F. A. Cotton, J. Am. Chem. Soc., in press. 85 Table 1. Vibronic Structure of I.he Lowest Energy Visible Band in the Spectrum of Mo,(02 CCH 3), at 15 K. ~ Polarizationd Assignm ente 469.44 21302a lS "• - a,u(g) 466.68 21428a lW "• - a,u(g) v. + --E.!!!_ 21500b 460.91 21696 1S 459. 66 21755c lW 458.78 21797C lW 458.34 21818c lS 457.25 21870 455.19 21969 lW II~+ 454.19 22017 1S v 0 + l~(OCO) 453.19 22066 lS "• 370 451.16 22165 lW 449.56 22244 lW 443.69 22287 lW 447.51 22346 lW 446.61 22391 lS 445 . 71 22436 lS 443 . 18 22564 lW 442.13 22618 lW 441. 00 22676 lW 440.18 22718 lW 439.23 22767 lS 438. 29 22816 lS 435. 81 22946 lW 434 . 84 22997 lW 433. 05 23092 lW 431.35 23183 lS 429.00 23310 lW 427.88 23371 lW 426 . 2-1 23461 lW 424. 59 23552 1s 422.42 23673 lW 421.25 23739 lW 419 .46 23840 lW 418. 01 23923 lS 377 a,u + ru VO + ru "• + a,~ Vo + eu ~U v~ + ru(OCO) v~ + ~u + ai g 374 370 374 372 389 376 380 379 382 374 367 374 364 369 369 363 368 379 371 aHot bands were not detected at 15 K, but were evident in a thick crystal at 80 K. b0-0 energy estimated from the hot band (but not observed in the spectrum). cWeak features observed in a thick crystal at 15 K. d ll denotes polarization along the MoMo axis. 1w signifies weak polarization perpendicular to the MoMo axis. ls signifies strong polarization perpendicular to the MoMo axis. eanly the origins for prog ressions in the totally symmetric MoMo stretch are shown since higher members of the progr ession are indicat_ed by the hash marks. The grou nd state MoMoO bending mode is ":,u(r,;) , and a,u(g) is the corre.~ponding metal-metal torsional vibration. The two µire electronic origins ar~ denoted by v and v '. Excited st ate MoO stretchin g modes are a,u', eu, and a, , and a, i s an e~cit ed' state MoMoO bending mode. An excited state antis y mmetrice;carboxyYate d ef ormation is r~(OCO) , and v 0 + ru denotes an unassigned vibronic origin. 86 T.ab le 2, Vihronic Structure of the Lowe,;t Ener gy Vis ible Band in the Spectrnm of 111o,{O,CCFJ., at 15 K. , _1 - nm ·- 4Gl.45 --~'-- Assi gnment 459.28 21773a 457.41 218G2b,c 455.58 21950h l'u + a,u 455. 10 2HJ73b Vo + ru 453 . 23 22064 Vo + a2u 451. 22 221G2b Vo + 450.11 22217b Vo 449. 4G 22249b C - a,u(µ;} Vo - alU (r:) 21(;71;\ v~ Vo + ru Vo + 361 447.89 22327 445.93 22425 443.16 22565d 440.70 22691 438.77 22791 436.17 22927 433.86 23049 432.10 231~3 429. 50 23283 427.30 23403 425.68 23492 423.08 23636 420.86 23761 419.46 23840 416.91 23986 415.00 24096 413.36 24192 410.83 24341 408.76 24464 407.35 24549 358 404. 87 21699 b Vo 3G4 ru ru + ru(OCO) Vo + ~u + alg 366 362 358 352 3JG 354 349 353 358 348 350 335 352 355 68 357 aHot bands were not detected at 15 K, but were evident at 80 Kin a thick sample. bweak bands observed in thick path length sample . c0-0 ba11d obs e rved in the spectrum and confirmed as such by the hot band temper:i.ture dependence. dThis hand is broacl with a distinct high energy shoulder and remains as such throu v;l10ut the progression, indicating two overbpping weak pro;;ressions. eNotation as in Table l ; a,u is an e:,cited state metal-metal torsional vibration. Table 3. 87 Vibronic Structure of the Lowest Energy Visible Band in the Spectrum of Mo 2 (O 2 CCD3 \ at 15 K. -1 nm cm 460. 29 21725 456.76 21893 454.94 21981 452.59 22095 449.76 22234 447.29 22357 Assignmenta 370 376 v~ + ru(OCO) 377 [2 445.00 22472 442.35 22606 440.00 22727 370 368 1375 437.82 22840 435.14 22981 432.94 23098 ;371 I I 368 369 430.88 23208 428. 26 23350 370 354 426.11 23468 424.41 23562 421. 59 23720 419. 2 9 23850 418.12 23917 415.17 24087 a Notation as in Table 1. 370 382 355 367 88 Table 4. cm -1 nm 465.27 Vibronic Fine Structure of Mo2 (O2 CH).; al 15 K a 21 493 457.29 21 868 456.79 21 892 456. 50 b 21 906 454.65 21 995 451. 94 22 127 451. 21 b 22 163 449. 82 b Assignment e 22 231 449. 26 22 259 44 ~ b 22 305 447. 53 b 22 345 444. 80 b 22 482 444. 08 b 22 518 442. 88 b 22 579 442.16 b 22 616 441.45 22 653 440.53 22 700 436. 29 d 22 920 429. 59 d 23 278 423. 06 d 23 637 416. 71 d 23 997 410. 71 d 24 350 405.18 d 24 680 373 363 355 355 367 348 373 357 358 359 360 353 330 a Broad hot band observed at 80 K. b Shoulder. c Not resolved. d Maximum of a broad peak that contains many i.lnrcsolved components. e Notation as in Table 1. 89 Tohk._5. Vibronic Structure in the Em ission Spectrum of Mo,(O,CCF 3 ) ., ~t I. 3 K. _, ____!!!!!_ ~ 498. 80 20048 499 . 73 20011 501. 61 1 9936 503.88 19846 506. 95 1 9726 508 . 57 19663 509. 92 1 9611 512. 11 19527 513.82 1 9462 517.41 19327 519. 61 19262 520. 51 19212 522.55 191 37 524.27 1 9074 528. 26 1 8930 529. 80 18875 531. 26 18823 533. 28 1 8752 535. 62 18670 r 385 I ! 400 : 409 I 399 401 l,s 390 538.94 540.98 18485 542 . 62 18429 544. 25 18374 546.78 18289 547.41 18268 550. 60 18162 I 1389 :i I>- 18048 555 . 56 18000 558.47 17 906 559. 79 17 864 562.46 1777 9 564.02 17730 566.70 17646 568.41 17593 572.18 17477 574.38 17 410 576.27 17 353 578 . 87 17275 580.89 17215 585.58 17077 587.54 17 020 590. 07 16947 594. 32 16 826 599. 88 16670 I j 404 L j 375 I : 390 394 378 j 381 I ,- 393 '-- : 378 obscured 554. 08 397 I 387 ~85 18555 ~. 381 ·374 I : 383 L 4"4 378 402 407 387 377 I_ 371 378 400 389 L. 38~ L 407 90 Table 6. Vibronic Intensity Inducing Vibrations ( A1g Ground State) Excited States Polarization Direction X D 4h Symmetry 1 Eg State D2 h Symmetry 1 1 B 3 g State Bzg State a a1u ~u blU b1u y au b2U blU alU au blU ~u blU au blU b2u z eu u b3U b2u 91 Figure Legends Figure 1. Single crystal polarized spectra of Mo 2 (O2 CCH 3 ) 4 at 15 K. Polarizations are denoted II and ..L to the b crystallographic a..,""{iS on the be crystal face. Spectr_a l ba,ndwidth was 0. 07 nm. Figure 2. Electronic spectra of Mo 2 (O2 CCH3 ) 4 , Mo 2 (O 2 CCD3 ) 4 , and Mo 2 (O2 CCF 3 \ at 15 K. Spectra of sublimed films on quartz disks at 0. 07 nm bandwidth were recorded. Figure 3. Corrected emission spectrum of a thick sublimed film of Mo 2 (O 2CCF 3 ) 4 on quartz at 1. 3 K. The sample was irradiated at 337 nm with a pulsed nitrogen laser and the spectrum was recorded at 0. 6 nm bandwidth. Figure 4. The band origin in the electronic spectru-m of Mo2 (O2 CCF 3 ) 4 at 15 K. Hot bands (80 and 150 K) are shown at a vertical • scale expansion of 5. The spectra were recorded at 0. 07 nm bandwidt'h . Figure 5. Boltzmann plot of the temperature dependence of the hot band intensity (Ir) for M~ (O2 CCF 3 ) 4 • Figure 6. Hot bands in the single crystal polarized spectra of M~ (~CCH 3 \ at 80 K. Polarizations are denoted II and 1 to the b crystallographic axis on the be face. Figure 7. Weak features in the single crystal polarized spectra of Mo2 (4 CCH 3 ) 4 • Polarizations are denoted II and l to the b crystallographic axis on the be face. The spectra were recor ded at 15 K with 0. 07 nm bandwidth. N Cl) (l) 0 C 0 .0 ~ V> 0 ..0 <r 410 420 :": /,' •: ... : 430 A (nm) ' .=: \! ) . ' . ....... ,"\ ;\j':.f-: I 440 :=J ... ·: G :_J : ! ·. ,. .. .. ·.: ::11: F E 450 , (: nC ;: : D: : :!: .' ·_: : B II :\___.._. .................L ~ A I\ j 460 93 ., g 0 .D :;: .0 <1 410 420 I 430 440 A (nm) 450 460 470 94 Intensity ()) N 0 ()) 8 <.n 8 ::J 3 c.n ~ 01 0 <.n ~ 0 <.n N 0 - 95 A (l) u C o . .a ~ 0 Cf) ..0 <[ 450 460 \ (nm) 96 i-----· - - - - - - - - - - - - 1 • -1.0- -2.0 5.0 10.0 97 J_ (1) u C 0 B _o L 0 Cf) _Q <[ 470 465 A (nm) 98 . • .• Q) 0 C 0 • ..D .. ..•• • L , 0 (/) ..D .. <[ .• .• • •. • C ... ............. _L 11 • ..... 455 460 A (nm) 465 99 Electronic Structure and Spectrum of Mo2 (O2 CH) 4 and Related ~ * Introduction The quadruple-bond formulation 1 of the strong metal-metal interactions in binuclear Mo(II) complexes has been confirmed by 42 Xa calculations on Mo 2 Cl8 and Mo 2 (O2CH) 4 o There is some evidence that the lowest excited states in the two complexes are 2 different, however. In this paper, we present an analysis of the electronic spectrum of Mo2 (O2 CH) 4 , and make comparisons with Mo 2 Cl 8 4- as well as with other binuclear Mo(II) carboxylate complexes. Experimental The compound Mo 2 (O2 CH) 4 vms prepared from Mo 2 (O2 CCH 3 ) 4 2 and formic acide Calcd. for C4 H4 Mo 2 O8 : C, 12. 92; H, 1. 08. Found: C, 12. 99; H, 1.10. ~Mo 2Cl8 was prepared by a published method. 3 Far infrared spectra were obtained for vaseline mulls on polyethylene disks using a Perkin-Elmer model 180 spectrometero A Cryogenics Technology model 21 cryocooler equipped with polyethylene windows was employed for measurements at 40 K. All electronic spectra were recorded on a Cary 17 double beam instrument. Low temperature spectra were obtained for a *W. C. Trogler's contribution to the joint paper by J. G. Norman, H. J . Kolari, H. B. Gray, and W. C . Trogler, Inorg . Chem., 1§_, 987 (1977). 100 thin film of Mo 2 (O2 CH) 4 sublimed onto a quartz disk. Temperatures of 80 or 15 K were obtained for the sample mounted in a Cary liquid helium dewar through use of either liquid nitrogen or helium as coolant. Sample temperature was measured with a calibrated carbon resistor secured to the cooling block. For the 15 K spectrum slits were controlled so that a resolution of 0. 05 nm was achieved. Oscillator strengths were measured at room temperature by preparing oxygen-free methanol solutions so that the band to be measured had an absorbance of 0. 4 - 0. 8. The spectrum was replotted on an energy scale and cut and weigh-integrated to obtain an oscillator strength. Results and Discussion The electronic spectra of Mo 2 (O2 CH) 4 (Figure 1) and ~Mo 2 Cl8 (Figure 2) are clearly different. For example, the low temperature spectrum of ~Mo 2 Cl8 shows bands at 18, 200 -1 1 and 41,670 cm , which may be attributed to A1 g 1 and A1 g - 1 A2 u ( • 1T - 1r*) transitions. 2 1 A 2 u (o - o*) The former band contains -1 a progression in the a 1 g Mo-Mo stretch at 40 K (360 cm ) . These assignments are consistent with those proposed for analogous bands 4 2in the spectrum of Re 2 Cl8 • The UV-VIS spectrum of lv102 (O2 CH) 4 (Figure 1) displays four bands, all of which are at considerably higher energy than o -- 5* in Mo 2 Cl8 4• -4 Band I, at 22, 990 cm -1 an oscillator strength of only 8 x 10 . Bands III and II overlap has 101 but are resolved into a peak at 33,780 cm -1 and a shoulder at -1 30,580 cm at 80 Kin a 4:1 ethanol-methanol glass (Figure 3). The oscillator strengths of bands II and III are estimated to be 0. 03 and 0. 19, respectively. Finally, band IV with an oscillator -1 strength of 0. 021 occurs at 44,250 cm . At J5 K band I (Figure 3) shows considerable vibronic structure. Analogous bands occur in Mo 2 (O2 CCH 3 ),1 and Mo2 (O2 CCF 3 ) 4 • 5 Polarized spectral studies including full vibronic analysis 5 have I suggested that band I be assigned to the vibronically-allowed transi1 1 tion A1 g - Eg (o - 11*). The 40 K low energy ir spectrum of Mo 2 (O2 CH) 4 was obtained (Figure 4) and peak positions and assignments are listed in Table 1. Identification of Mo-O modes were possible by comparison with Mo (O CCF ) 6 and Mo (O CCH ) , 2 2 3 4 2 2 3 4 with fo r ce constant calculations 7 available fo r the latter complex. With this information detailed assignment of band I is possible. The distortion of the M2 O8 moiety in Mo 2 ( O2 CCH3 ) 4 and 5 Mo 2 (O2 CCF 3 ) 4 from ideal D4 h to D2h symmetry has been noted. This is manifested in low symmetry splitting of the electronic (and 1 -1 hence vibronic) origins for the Eg state. The small (24 cm ) splitting of band A in Mo2 (O2 CH) 4 implies that distortion from D4 h symmetry is not significant. In fact, such a small splitting could be a factor group effect as there are four molecules per unit cell. 6 A temperature dependent hot band similar to that observed -1 for Mo 2 (O2 CCH3 ) 4 and Mo2 (O2 CCF 3 t,1 occurs 385 cm to lower energy of band A above 80 K. This is expected since the Mo- O a 2 u vibration 102 (which vibronically enables the transition) is essentially the -1 same energy (170, 183, and 190 cm ) in all three complexes. Therefore the true (24 cm -1 split) 0 - 0 band for Mo 2 (O2 CH) 4 lies -1 at about 21,685 cm . The sole vibrational progression is attributed -1 to the Mo-Mo a 1 g stretch (360 cm ). This excited state distortion 1 in the Mo-Mo bond is appropriate for the antibonding Eg (o - 1r*) state. Other weaker features are assigned by analogy with 5 Mo (O CCH ) • 2 2 3 4 We have carefully examined the spectrum of Mo 2 (O2 CH) 4 from 80 to 20, 000 cm -1 at 40 K and find no electronic absorption bands of appreciable intensity. Therefore we must look to higher energy for a band attributable to the 5-5* transition. In fact, band II with an oscillator strength of 0. 03 is a reasonable candidate, as it is of comparable intensity to the 5-5* bands in Re 2 Cl8 45 and Mo2 Cl8 (0. 015). 2- (0. 023) 9 Band II, as well as bands I and IV, do not appear to depend on the nature of the coordinated carboxylate (Table 2). This suggests a metal-localized transition assignment for all three bandso Band III, on the other hand, depends drastically on the nature of the carboxylate. The huge shift observed for the benzoate compound hints at a transition involving carboxylate 7f or iT* orbitals. As the IP 10 of the corresponding carboxylic acid increases the energy of .band III decreases, suggesting assignment to a charge transfer transition o - L(1r*). 1 Since the transition apparently is allowed it must be either A1 b0' 1 1 A2U or Aw b 1 Eu. It is likely that band I derives its vibronic 103 intensity (which is chiefly polarized perpendicular to the Mo-Mo 1 1 5 axis ) from band Ill, suggesting a A1 g - Eu assignment for the latter. Band IV, considering its oscillator strength and slight 1 carboxylate dependence, is reasonably assigned to A1 g - 1 A2 u (7T - 7T*). The o-6'!' and 1T-1T* transitions are proposed to lie 4700 and 2200 cm -1 to higher energy in the carboxylate complexes (as compared to Mo 2 Cl8 4)• vVe take this to be an indication of the I stronger metal-metal interactions in the carboxylate complexes, which are reflected in the considerably shorter Mo-Mo bond length 1 of 2. 09 Ain Mo 2 (02 CR) 4 compounds. The Mo-Mo bond length in other non-bridged complexes is ca. 2. 15 A. The o-o* transition is affec ted most by this change of bond length. We infer that it is primarily the 6* level which is affected. This he lps rationalize the relatively low energy position of the o - 1r* transition. Electron repulsion factors will also need to be considered, as they could easily be larger in Mo2 0 8 than in Mo 2 Cl8 4• 104 References 1. F. A. Cotton, Chem. Soc. Rev., 4, .,..._ 27(1975), and references therein. 2. J. G. Norman and H. J . Kolari, ,J. Am. Chem. Soc., 97, 33 ~ (1975) and Chem. Commun., 649 (1976). 3. J. V. Brencic and Fo A. Cotton, Inorg. Chem., .,..._ 9, 351 (1970) . 4. C. D. Cowman and Ho B. Gray, J. Am. Chem. Soc., 95, 8177 ~ (1973). 5. W. Co Trogler, E. L Solomon, lb. Trajberg, C. Jo Ballhausen, and H. B. Gray, Inorg. Chem., 16, .,..._.,..._ 828 (1977) . 6. F. A. Cotton and J. G. Norman, J. Coord. Chem., .,..._ 1, 161 (1971) . 7. Wo K. Bratton, Fo A. Cotton, M. Debeau, and R. A. Walton, J. Coord. Chem., .,..._ 1, 121 (1971) . 8. F. A. Cotton, J. G. Norman, B. R. Stults, and T. R. Webb, J. Coord. Chem. , .,..._ 5, 217 (1976) . 9. F. Ao Cotton and C. G. Harris, Inorg. Chem., .,..._ 6, 924 (1967) . 10. D. A. Sweigart and D. \V. Turner, J. Am. Chem. Soc., .,..._.,..._ 94, 5592 (1972). 105 Far Infrared Spectrum of Moi02CH) 1 at 40 K. Table 1. a cm -1 Assignment 471 s a2 u (Mo-0 stretch) 445 s eu (Mo-0 stretch) 231 m eu (Mo-0 bend) 170 w a2U (Mo-0 bend) 128 w eu 410 m 384 m 371 s 345 w (Mo-0 bend) 78 w a s = strong, m :-:: medium, and w = weak intensity. 106 Comparison of the electronic absorption spectra of Mo 2(0 2CR) 1 Table 2. complexes for R = CH , H, CF 3 , 3 and C H 6 • 5 Band Ia Band II Band III Band IV 22 ,730 n. o. 34,010 42,900 22,990 30,580 33,780 44,250 Mo 2 (0 2CCF 3 ) 4 23,0QQC 30,210c 32, 360c 44, 200d Mo 2 (0 2CC 6 H5 ) 4b n. o. n. o. 23,700 44,300 0.0008 0.03 0.19 0. 021 Assignment 6 --- 1r* o --- o* o - L(7r*) 7T - 7T* a -1 Mo 2 (0 2CCH 3 ) 4 Mo 2 (0 2CH) 4 b C Oscillator strength e Band energies in cm . b L. Dubicki and R. L. Martin, Aust. J. Chem., .,,._,.._ 22, 1571 (1969) . c This work. d Reference 6. e Our values for Mo 2 (0 2CH) 4 107 F io-ure Cautions ~ Figure 1. • Electronic absorption spectra of Mo 2 (O 2CH)" in methanol solution at room temperature. Figure 2. Electronic absorption spectrum of K.1Mo 2 Clg at 4 0 K in a mineral oil mull. Figure 3. Electronic absorption spectrum of Mo2 (O2 CH) 4 at 15 K for a sample sublimed onto a quartz window. Figure 4. Low energy infrared spectrum of MoiO2CH) 4 at 40 K. The sample consisted of Mo 2 (02CH) 4 powder suspended in vaseline on a polyethylene disk. 108 l I T ERS /,.,.O LE 500 1000 0 0 0 0 0 s N ""Tl s 0 :::r :3 ::J r-1- (D r-1- OJ 0 z 0 ~ OJ +::,,(D 3: ::) 50 . 0 0 "'0 0 I 00 LI TE R 5/MO LE 109 absorbance (:nbitrary units) ~ 0 0 ::s s Ul 0 0 ::s s O') 0 0 ::s 8 110 absorbance (arbitrary units) I I I ,p.. ~ 0 ,p.. 0) 0 111 TRANSMITTANCE (Jl 0 0 +:> 0 0 n 0J :::::;-o =o N 0 0 0 0 112 CHAPTER 2. THE PHOTOCHEMISTRY OF COMPOUNDS WHICH CONTAIN A QUADRUPLE MOLYBDENUM - MOLYBDENUM BOND 113 production of Hydrogen by Ultraviolet Irradiation of J\/[Oz (SO4 ) / - in Aqueous Sulfuric Acid. Electronic Absorption Spectrum of K3 Mo2 (SO4 \ • 3. 5 1-IzO at 15 K. la and Arnis A. Zagars lb Contribution No. 5462 from the Arthur Amos Noyes Laboratory of Chemical Physics, California Institute of Technology J Pasadena. California 91125; the Department of Chemistry. Pennsylvania State University, University Park, Pennsylvania 16802; the Department of Chemistry: 2 Division of Natural Sciences, University of California at Santa Cruz, California 95060; and the Denartment of Chemistry. Massachusetts Institute of 'JechnologL_ Cambridge, Massachusetts 02139 Ultraviolet irradiation (7\. 254 nm) of Mo2 (SO4 \ 4- in 5 M ~SO4 prcduces H2 and the one-electron oxidation product Mo~ (SO4 ) 4 disappearance quantum yield of MOz (SO4 \ 1- 3• The is 00 06 at 254 nm. The spectrum of K3 M0z(SO4 ) 4 • 3. 5 H2 O exhibits an absorption band at 1405 nm (E 143); this band shows a vibrational progression in a 350 cm- 1 fundamental even in 5 M D2 SO4 solution at room temperature. At 15 K additional vibronic structure is resolved. This band is assigned to the N - E type transition 114 Introduction ~ It has been shown previously that ultraviolet irradiation of aqueous 2 acidic solutions containing one-electron reductants such as F e + yield 2 hydrogen. We are presently explormg the photoredox behavior of low-valent molybdenum complexes containing strong metal-metal bonds, and in particular are directing our attention to systems that are capable of reducing hydronium ions. Herein we report the results of an investigation 4 of the photoreactiv ity of M0z (SOJ 4 - in aqueous sulfuric acid. The near infrared spectrum of K3 Mo2 (SO4 \ · 3. 5 H 2 O at 15 K has also been studied. erimental Section The compounds K M9a(SO 4 4 ) } , 4 and K MaJSO 3 \ · 4 3. 5 I1zcPwere prepared according to standard procedures. All manipulations with the molybdenum complexes were conducted under nitrogen or in degassed solution. For the irradiations , the appropriate solution was placed in a special vacuumtight, 1-cm quartz spectrophotometer cell and degassed through three freeze-pump-thaw cycles. Low pressure Hg lamps were used for the 254 nm irradiations. The electronic absorption spectra of the reaction solutions were monitored using a Cary 17 spectrophotometer. Quantum yields at 254 nm were determined using ferrioxalate actinometry such that the sample absorbed 5. 77 x 10- 7 einsteins/min. Mass spectral analyses of the gas above the irradiated solutions were obtained with an AEI-MS-902 high resolution mass spectrometer. The low temperature near IR spectrum of K3 Mo2 (SO4 \ · 3. 5 l1z O was obtained by positioning a sample in a Cary Liquid Helium Dewar. X-ray photoelectron spectra were obtained with a HP5950A spectrometer from powdered samples on double stick tape. Evolution of ~ was measured for exhaustively photolyzed samples 4- of M~ (SO4 \ in vacuum-tight 1 cm spectrophotometer cells equipped with 115 a glass side-chamber and needle valve. The amount of evolved }\ was determined by Toepler pumping the stirred photolyzed solution through three liquid nitrogen traps into a calibrated volume and manometrically measuring the pressure . To insure that only ~ was collected, the gas was then passed over hot copper oxide, yielding water that was condensed in a liquid nitrogen trap, leaving no detectable gas. Photochemistry of M~ (804\ 4- Irradiation (254 nm) of M~ (SO4 \ 4- in 5 M D2 8O4 results in the spectral changes shown in Figure 1. As irradiation proceeds, small gas bubbles 3 4 4are formed. The band at 515 nm (E ~170) characteristic ' of M~(SO4 ) 4 decreases in intensity, whereas a structured absorption system diagnostic (vide infra) of MqJS04 ) 4 3- grows in at 1200-1600 nm (Emax of 143). Isosbestic points are initially maintained at 578 and 420 nm and shoulders develop at 570 and 405 nm, A quantum yield of 0. 06 \Vas obtained for disappearance of Mo2 (SO4 \ 4• Mass spectral analysis of the gas formed by irradiating M~ (SO4 ) 4 4- in 5 M ~S04 solution confirmed that ~ was produced. Thus the principal photoreaction must correspond to Eq. 1: (1) An exhaustively-photolyzed solution yielded 0. 31 moles of Hz per mole of K4 Mo2 (SO4 \ . The fact that this is less than the theoretical amount of ~ (0. 5 moles) is not unexpected, as isosbestic points are not maintained (Fig. 1); that is , the r eaction observed upon 254 nm irradiation is not 6 stoichiometric at hi gh conversion. 116 3- Near Infrared Spectrum o f ~ We have found that MOz (SO4 ) / - exhibits a previously unreported 5 absorption band in the near IR at 1405 nm (Emax 143 for a 5 M D2 SO4 solution). Remarkably, a vibrational progression with a spacing of 350 cm- 1 attributable to a 1 g (MoMo) was observed even in solution at room temperature. In contrast, the lowest energy electronic absorption band of K4 MOz (SO4 ) 4 , 7 which occurs at 519 nm (E 170), has been shown to be due to the transition 1 o - o* ( A1 g - 1 Azu). There are two plausible explanations for the near IR band of K3 Mo2 (SO4 ) 4 • 3. 5 ~ 0. As the compound is formally of mixed valence, the transition could be assigned intervalence charge transfer in 8 9 the weak interaction description. ' Alternatively, the transition could 10 2 2 be viewed as the N - E type transition o - 6* ( Bzg - B1 u). The former description seems inappropriate for several reasons. Firstly, the MoMo 2 11 bond lengths and stretching frequencies1 are approximately equal in K4 MOz(SO4 ) 4 and K3 Maz{SO4 ) 4 • 3. 5 ~0. Therefore, the o bonding model applicable to K4 Mo2 (SO4 ) 4 would be expected to obtain for K3 Mo2 (SOJ 4 • 3Hz 0 as well. EPR measurements for K3MOz(SO4 ) 4 • 3. 5 1\0 suggest that the odd 13 electron occupies a o orbital delocalized over both metal centers. V.l e also have examined the X-ray photoelectron spectrum of K3 M0z(SO4 ) 4 ° 3. 5 ~O and have found no splitting of the Mo 3p signal. One final point, and most i 2 convincing, is the fact that the 1405 nm band in the spectrum of K3 Mo2 (SO4 ) 4 • 3. 5 ~O shows sharp vibronic structure at 15 K (Figure 2, Table 1 ). An intervalence charge transfer band, because the transition is a multi phonon 14 excitation, 9 , would not be expected to exhibit well resolved molecular 117 vibronic structure. Both the vibrational progression in quanta of 1 350 cm- and the Franck-Condon factor of about 1. 5 are more consistent with the o - o* assignment. The vibrational structure of the 1405 nm band in the spectrum of K3 Mo2 (SO4 \·3.5 f½O is assigned as follows: There are five (A-E, Figure 2, Table 1) origins for progressions in a 1 g(MoMo). Conspicuous is the presence of the two major components A and C with slightly different Franck-Condon factors. The progression that begins at C has a slightly higher progressional frequency than that beginning at A. These split components are hterpreted to result from fue o - o* transition of two 11 crystallographically nonequivalent Mo2(S0.1)/- ions in the unit cell. 12 Previously, it has been noted that this nonequivalence results in slightly 1 different ground-state values of a 1 g (MoMo) (373 and 386 cm- ). The two weak peaks, B and E, might possibly represent transitions involving quanta of aw modes built on A and C , respectively . Alternatively, these weak b peaks could result from Davydov splitting of the pure origins (A and C). The other weak peak, D, may be attributed to a transition involving a quantum of an a 1 g mode built on either A or C. rt is apparent, however, that definitive assignments of the three weak peaks cannot be made with the limited data available. ~ We thank Robert Gay for experimental assistance. W. C. T. acknowledges an IBM Fellowship during 1976-77. This research was supported by the National Science Foundation. 118 References and Notes 1. a) California Institute of Technology. b) Pennsylvania State University. c) University of California at Santa Cruz. d) Massachusetts Institute of Technology. 20 V. Balzani and V. Carassiti, "Photochemistry of Coordination Compounds, 11 Academic Press, New York, N. Y., 1970, pp. 159-161, and references therein. 30 A. Ro Bowen and H. Taube, J. Am. Chem. Soco, .,.__,,.._ 93, 3287 (1971) . 4. A. R. Bowen and H. Taube, lnorgo Chem., -"'-"' 13, 2245 (1974). 5. A. Pernick and M. Ardon, J. Am. Chem. Soc., .,.__,.,_ 97, 1254. (1975) . 6. Irradiation of Mo 2(SO1 ) 4 3- 4- in 5 M H~O4 with visible irradiation also produces Mo 2(SO,i) 4 and hydrogen (W. C. Trogler and H. B. Gray, unpublished results). Preliminary experiments have shown that the photoreaction under these conditions is stoichiometric at high conversion (0. 49 moles of H2 were liberated per mole of Mo2(SO4)4 4- irradiated). Experiments directed toward the eluci- dation of the mechanistic features of the photoreaction are in progress. 7. F. A. Cotton, D. S. Martin, P. E. Fanwick, T. J. Peters, and T. R. Webb, J. Am. Chem. Soc., .,.__,,.._ 98, 4681 (1976) . 8. G. C. Allen and N. S. Hush, Prog. Inorg. Chem., .,.... 8, 357 (1967) . 9. N. S. Hush, Prog. Inorg. Chem., 5!_, 391 (1967). 10. R. S. Mulliken, J. Chem. Phys., ....-.. 7, 20 (1939) . 119 lL F. Ao Cotton, B. A. Frenz, E. Pedersen, and T o R. Webb, Inorg. Chem., 14, 391 (1975) . .r-..r-. 120 A o Loewenschuss, J o Shamir, and Mo Ardon, Inorg o Chem., 15, 238 (1976) . .rv--. 13. F. A. Cotton and E. Pedersen, J. Am. Chem. Soc., 97, 303 ,r..,-... (1975). 140 J o Ko Beattie, N. S. Hush, and P. R. Taylor, Inorg. Chemo, 15, 991 (1976). ""' 120 Table 1. Vibronic Structure of the Near IR Band in tl-Je Spectrum of K3Mo2 (SO4 ) 4 ·3.5 HzO in a KBr pellet at 15 K -1 nm cm 1583. 7 6314 1575.6 6347 1558.2 6418 1520.0 6579 1500.3 6665 1492.3 6701 1482. 4 6746 1475.9 6776 1442.9 6930 1425.0 7018 1417.3 7055 1406. 7 7109 1402.3 7131 1374.0 7278 1357.2 7368 1339.0 7468 1312.9 7616 1296.2 7715 A B C -D E 351 354 358 351 353 354 363 355 348 350 359 347 121 ~ Figure 1. Electronic absorption spectral changes during 254 nm irradiation of K4 Mo2 (SO4 ) 4 in 5 M D2 SO4 solution. Figure 2. Electronic absorption spectrum of K3 M0.;i(SO4 ) 4 • 3. 5 H2 0 in a KBr pellet at 15 K. 122 Ab,o,to-ce r - - - --·-;;;.;;:=-..::::-----::=.-=-:-· 8 .. ! 8~ I I 8 I 0 8 I i I -- _________! 123 Absorbance N (J1 0 vJ 0 0 .Is (Jl 0 0i - 8 124 Photochemical Oxidation Reactions of Binuclear Molybdenum(II) Complexes in Aqueous Acids and in Chlorocarbons. William C. Trogler, David K. Erwin, Gregory L. Geoffroy, and Harry B. Gray* from the Arthur Amos Noyes Laboratory Contribution No. 5598 of Chemical Physics, California Institute of Technology, Pasadena, California 91125 Abstract: ~ The disappearance quantum yield at 254 nm of Mo 2 (SO4 ) 4 in aqueous H2SO4 to yield one-half mole of H2 and Mo 2 (SO4 ) 4 independent of acid concentration (0. 001 - 5 M H2SO4 ) . irradiation of Mo 2 (aq) 4+ 4+ • Eventually Mo 2 Cl8 H is irradiated at thermally decomposes to yield one mole of hydrogen gas and Mo2 (µ-OH) 2 (aq) 4+ 4- is initially formed with a quantum yield of 0. 13. 3- Mo 2 (O2 CH2 NH 3 ) 4 Ultraviolet . The quantum yield for this process is 0. 035 upon 254 nm excitation. Vlhen Mo 2 Cl8 254 nm, Mo 2 Cl8 H is in 1 M HO3 SCF 3 produces one mole of hydrogen and Mo2 (µ-OH) 2 (aq) 3- 3- 4- 4+ . No reaction was observed when in 0. 25 M HCl was subjected to uv irradiation. Irradiation (254 nm) of Mo 2 Cl4 P 4 (P = P(Et) 3 , P(~-Bu) 3 , and P(Et)(Ph) 2 ) in chlorocarbon solvents yield triply chloride bridged molybdenum(III) compounds. Quantum yields for these reactions are reported. Consideration of the spectra of these compounds suggests that the photoactive excited state has considerable LMCT character. 125 Introduction ~ Photooxidation reactions have been observed for a variety of metal ions in aqueous media. l- 3 Frequently, molecular hydrogen is the reduction product of the redox process. excitation of Cr 2+ (aq), Fe 2+ (aq) , Eu 2+ For example, ultraviolet (aq), and Ce s+ (aq) results in a one-electron oxidation of the metal center concomitant with dihydrogen 1 4 formah on. Recently we reported the first example of an analogous reaction for a metal cluster compound. It was shown that 254 nm irradiation of aqueous sulfuric acid solutions of K4 Mo2 (SOJ 4 yielded H2 3- and Mo2(S04 ) 4 • In order to test the generality of this process for bi- nuclear Mo(II) comm unds we have examined the photochemistry of Mo2(aq) 4+ , Mo2 Cl8 4, and Mo2 (02CCH 2NH 3 ) 4 4+ in aqueous acid solutions. Although photooxidation reactions in aqueous media are often loosely described as the result of charge transfer to solvent (CTTS) excitation, solvated electron production has onl y been established 2+ 444- 1-3 (e.g., Cr (aq), Fe(CN) 6 , W(CN) 8 , and Mo(CN) 8 ) for those compounds where independent spectroscopic studies also indicate a low-lying CTTS absorption. This is to be compared with the fact that a solvated electron intermediate has not been observed subsequent to 5 3 flash photolysis of aqueous ferrous solutions. Mechanistic studies • 2+ have implicated the presence of H atoms during photolysis of Fe(aq) . 6 An investigation of the reaction of H atoms with Fe2\aq) had suggested 2+ the formation 0f an unstable hydride complex FeH . Kinetic evidence 7 for the photooxidation reaction was consistent with a mechanism involving hydrogen evolution from a FeH 2+ species. Herein we present evidence that photochemical production of 126 hydrogen from binuclear Mo(II) ions occurs via an intermediate hydride complex. It is proposed that the photoactive state results from excitation of a LMCT transition. Organometallic compounds have also been observed to undergo photooxidation reactions in halocarbon solvents. 8 Irradiation of ferrocene in the presence of CC14 and CHC13 and of 9 ruthenium in CC14 and CH2 C1Cf4Cl yield the ferricenium and ruthenicenium ions. A net photooxidation also obtains for the photodeclusterification of metal-metal bonded compounds in halocarbon solvents. lO In order to further elucidate the photochemistry of binuclear Mo(II) compounds the reactions of Mo2 Cl4 P 4 (P = P Et 3 , Pfu-Bu)3 and P(Et)(Ph)z) have been examined in chlorocarbon solvents upon 254 nm irradiation. ~ 12Rb 13, 14 The compounds K4 Mo 2C lg · 2H 2 0, 11 K.iM0 2 (SO) . sM0 2 Cl8 H, 4 4, Mo 2 (0 2 CCH 2 NH 3 ) 4 (S04 )2 . 4H 2 0, 15 Mo 2Cl4 [ P (Et) 3 ] 4 , 16 Mo2 Cl4 [ P (n-Bu) 3 ] 4 , 17 17 and Mo 2 C1J P(Et)(Ph) 2 ] were prepared by published methods. The phosphine derivatives were recrystallized from CH 2 Cl 2 by the addition of methanol to filtered solutions and cooling to 0° C and satisfactory C, H 4+ analyses were obtained. Solutions of Mo2 (aq) in 1 M H03 SC F' 3 were prepared as described by Bowen and Taube. 18 Although binuclear Mo(II) compounds are air stable for short (3 h) to long (1 yr) periods of time in the solid state, they are exceedingly oxygen sensitive in solution. For the Mo 2 Cl4 P 4 compounds we noticed that traces of moisture are also detrimental. Therefore the following procedures were adopted. Solids were stored under N2 prior to use. Solution spectra were measured in special evacuable 127 cells. The apparatus consisted of a 1. 0 or 0. 1 cm quartz cuvette attached to a pyrex tube which in turn was connected to a teflon needle valve and either a 10 ml pyrex bulb or a 5 ml pyrex calibrated volume tube. A 24/ 40 ground glass joint after the needle valve allowed attachment to a vacuum line. Solid samples weighed in room atmosphere were transferred to the cuvette sidearm. For aqueous solutions 5 ml of the appropriate acid mixture was pipeted into the 10 ml bulb. Solutions of Mo2 (aq) 4+ were transferred in a nitrogen filled glove bag. The cell was then sealed, attached to the vacuum line and freeze-pump-thaw degassed for three to four cycles and only then was the solution mixed with the solid in the sidearm. Extinction coefficients were determined in 0.1 cm cells so that a sufficiently large amount (0. 02 g) of material could be used for an accurate weighing. No correction for sample volume was made as it is an ins ignificant source of error. Molar extinction coefficients were reproducible to± 5 %and are believed precise to± 10%. A similar procedure was employed with chlorocarbon solvents except that excess solvent was vacuum distilled from Ca}\ into the cooled (77 K) pyrex sidearm of the solution cell. The sidearm was calibrated to either 3 or 7 ml. After warming, excess solvent was slowly removed until the volume was reduced to a calibration line. The solvent was then subjected to three freeze-pump-thaw cycles before mixing with the solid in the quartz sidearm. Failure to take these precautions may 19 account for the anomalously low Emax value reported for the 588 nm absorption band of Mo2 ClJ P(n-Bu\] 4 ~ 128 Irradiations at 546, 505 and 436 nm were obtained from the output of a 1000 W Hg-Xe arc lamp which was coUimated before passing through the appropriate interference filter. For some 546 nm irradiations light from a 450 W Hanovia medium pressure Hg lamp was passed through a Corning 4-102 glass filter combination. The focused beam from a 200 W Hg-Xe arc was used for some "broad band" UV photolysis. A 30 W low pressure Hg (Hanovia Germicidal) lamp with a Vycor envelope was used as a source of 254 nm radiation. Because of competing thermal side reactions it often was vital to have an intense source of 254 nm radiation. Therefore the output of the low pressure Hg lamp was not filtered. The only other spectral line of appreciable intensUy is at 57 8 nm. 20 In each instance we have shown that the compounds examined were not photoactive in this spectral region. Quantum yields were determined at 254 nm employing the 20 ferrioxalate actinometer . The procedure was modified to adopt the precautions recently suggested by Bowman and Demas. 21 Disappearance quantum yields of the binuclear Mo(II) compounds were obtained by monitoring the decrease of the visible o - o* absorption band. Aqueous solutions of binuclear Mo(II) ions decompose 13 14 18 thermally. ' ' Correction for these thermal background reactions was necessary for the quantum yield determinations. This was accomplished by alternating periods of time with and without irradiation for each sample. In most instances this correction was less than 510%. vVe have-also observed an initial slow decomposition of the Mo 2 Cl 4 P 4 compounds in CH 2 Cl 2 , even with the aforementioned precautions . Presumably traces of oxygen or the presence of adventitious moisture 129 account for this observation. Chloroform reacted with the Mo 2 Cl4 p 4 compounds thermally and amounted to a 10-30% correction in the quantum yields in this instance. No corrections for inner filter effects were applied as the quantum yields were obtained for less than 10% photolysis. Typically three measurements were obtained with an internal consistency of± 5%. Except in those cases where the thermal correction was large (i.e., the photochemical reactions in chloroform, 0. 001 M ~SO4 and 6 M HCl) reproducibility of the quantum yields was ± 15%. Hydrogen evolution was measured for photolyzed samples in the vacuum-tight 1 cm spectrophotometer cells . The amount of evolved ~ was determined by Toepler pumping the stirred photolyzed solution through three liquid nitrogen traps into a 12. 2 ml volume and manometrically measuring the pressure. To insure that only~ was collected, the gas was then passed aver hot copper oxide, yielding water that was condensed in a liquid nitrogen trap, leaving only traces ( < 5%) of residual gas. Typically 0. 03-0. 07 mmoles of hydrogen gas were collected and the measurements were reproducible The compound Mo 2 Cl3 (µ-Cl)s[ P(Et)(Ph) 2 chemically by the following method. [ P(Et)(Ph) 2 to ± 5-10%. L was prepared photo- One quarter gram of Mo 2 Cl4 - L and a teflon stir bar were placed in a quartz tube (3. 5 diameter, 14 cm height) which could be sealed with a teflon needle valve and attached to a vacuum line via a 24/ 40 ground glass joint. The cell was evacuated and ca 50 ml of CH 2 Cl2 (which was stored in vacuo over CaH 2 ) was distilled into the tube. The solution was then irradiated with the unfiltered light .from a 200 W Hg-Xe lamp (ca 5 hr) 130 until the blue colored solution changed to a bright maroon. After reattaching to the vacuum line the volume was reduced to 1 0 ml and ~ 30 ml of diethylether (stored in vacuo over LiAlH4 ) was distilled into the tube. A maroon oil formed and the mixture was stirred for one day until a maroon powder was obtained. The tendency of the compound to form an oil made purification very difficult. This powder was collected, redissolved in CH2 C12 , and precipitated again to yield 0. 15 g of product. The solution spectrum of the compound in CH2 Cl 2 was similar but not identical to that of the photoproduct. A molecular weight of 950 g/mole was measured in chloroform solution. Anal: Found C, 46. 88 ; H, 40 42; P, 9. 7. Calcd for Mo2 C1J P(C 2 H5 )(C 6 H5 ) 2] 3 : C, 48. 17; H, 4. 33; P, 8. 87; MW = 1047 g/mole. The infrared spectrum measured in a K.Br pellet showed two strong Mo-Cl stretches at -1 315 and 332 cm . Ultraviolet-visible absorption spectra and spectral changes were recorded utilizing a Cary 17 spectrometer. Infrared spectra were recorded with a PE 225 instrument. Molecular weights were determined using an HP 301A vap-or pressure osmometer and benzil as a calibranto Elemental analyses were obtained from Chemalytics, Inc. and the Caltech microanalytical laboratory. 131 Results ~ Photooxidation Reactions in Ag ueous Media - Previously we demonstrated 4 that ultraviolet irradiation of Mo (SO ) / - in 5 M H SO 2 yielded H2 and Mo 2 (SO4 ) 4 3• 4 2 4 At high conversion this reaction was not stoichiometric. Therefore we have measured the amount of H2 evolved in partially photolyzed (ca 80%) solutions and find 0. 49 moles 4- of H2 literated per mole of Mo(SO4 ) 4 consumed. When considered together with the earlier 4 results, the initial stoichiometry is unambiguously defined as aqueous H~O4 4>254 = 0. 18 The quantum yield of this process is essentially independent of acid concentration from 0. 001 M to 5. 0 1\1 H~O4 (Table 1). Attempts to extend the measurements to lower acid concentrations were hampered by increasingly rapid hydrolysis. It should also be noted that a regular increase in Emax of the 6 - B* transition was observed for decreasing acid concentrations. This may result from either partial substitution of the sulfate ligand by water or a change in the solvation structure of the Mo 2 (SO4 ) 4 4- ion. 4 Excitation (254 nm) of Mo2 (aq) + in 1 M trifluoromethanesulfonic acid results in the electronic absorption spectral changes shown in Figure 1. As the irradiation proceeds, the color of the solution changes from red to green, and the band at 505 nm, characteristic of Mo (aq) 4+, 4 decreases in intensity. Absorption bands attributable 2 to product occur at 620, 570, 490, and 362 nm, and isosbestic points at 568 and 424 nm are maintained throughout the conversion. No 132 reaction was observed under similar conditions in the dark or when the wavelength of irradiation was 366 nm. As 254-nm irradiation proceeds, small gas bubbles are formed and mass spectral analysis confirmed the presence of H2 • Measurement of the hydrogen evolved yielded 1. 06 ± 0. 1 moles per mole of Mo 2 (aq) 10 -3 to 1. 4 x 10 -2 Min Mo 2 (aq) 4+ 4+ for solutions 5 x . Furthermore, the absorption spectrum of the green photoproduct is identical to that of a Mo(ll) species first prepared by Hartmann and Schmidt 23 and formulated as a hydroxybridged dimer by Ardon and Pernick. 24 The overall photochemical reaction may be formulated as in Eq. 1: (1) The disappearance quantum yield for Mo 2 (aq) 4+ at 254 nm is 0. 036. 4- Concentrated HCl solutions of the Mo 2 Cl8 ion are readily 3 13 14 oxidized to Mo2 Cl 8 H -. ' The rate of this reaction is dramatically decreased in dilute HCl solutions. For example, in 3 M HCl at ambient temperature only 15% reaction occurs over a period of 18 hr. We shall refer to the acid solutions of ~Mo2 Cl 8 • 2H 2 0 as containing the Mo 2 Cl 8 4- ion although it should be realized that in solution an equilibrium among partially aquated species obtains. Irradiation (254 nm) of Mo2 Cla 4- in 3 M HCl induces the electronic absorption spectral changes shown in Figure 2. As the reaction proceeds the color of the solution changes from red to green; however, there are no isosbestic points. The first absorption band attributable to products 133 appears at 415 nm. Upon further irradiation this peak grows slightly, a new band appears at 369 nm and bubbles of gas are evolvedo The 369 nm band is also found if a photolyzed solution containing the 415 nm absorption is allowed to stand in the dark. If an authentic sample of RbM0 2 Cl 8 H is dissolved in 3 M HCl, an absorption band is initially observed at 415 nm (Figure 3). Subsequently, the intensity of this band decreases concomitant with the growth of a 14 new band at 369 nm. Cotton and Kalbacher have shown that the reaction of Mo 2 Cl 8 H with D 2 0 yields HD. We find that 0. 92 moles of H2 are evolved per mole of Rb3 Mo 2 Cl 8 H in 3 M HCl. solution of Mo 2 Cl 8 4- When a in 3 M HCl is photolyzed to completion 1. 16 moles of H2 were evolved per mole of Mo2Cl 8 4- consumedo The electronic spectrum of the green product derived from the decomposition of Rb 3 Mo 2 Cl 8 H in 3 M HCl exhibits bands at 369 (E ~ 820) and 655 (E ~ 85) nm and a shoulder at 600 (E ~ 75) nm. A dimeric formulation of the products was assumed to calculate the extinction coefficientso This spectrum appears identical to that reported 23 for Mo(III)(H 2 O) 6 to be Mo 2 (µ-OH) 2 (aq) 4+ 3 +. 1n . 2 M HCl which was later shown . As for Mo 2 (aq) 4+ , photolysis of Mo 2 Cl 8 24 4- ultimately yields one equivalent of H2 and the hydroxy-bridged molybdenum(III) dimer. Consistent with the observations of Cotton and Kalbacher 3- we find that Mo 2 Cl 8 H3 14 is stable in more concentrated acid solutions. Photolysis (254 nm) of Mo 2 Cl8 4- in 6 M HCl initially (< 5% reaction) 134 3- quantitatively produces Mo 2Cl 8 H with a disappearance quantum yield of 0. 092. At greater than 10% conversion, the production 3- of Mo2 Cl8 H levels off and then decreases upon further irradiation. This is most likely a result of secondary photolysis. In fact, 3- Mo2Cla H in 6 M HCl disappears with a quantum yield of 0. 040 when excited at 254 nm. The photoproduct in this instance posesses an intense absorption at 305 nm and a weaker band at 512 nm. An isosbestic point is maintained at 490 nm throughout the photolysiso 3- Because Mo 2Clg H absorbs more strongly than Mo 2Cl 8 4- at 254 nm the secondary photolysis rationale is vindicated. Similarly, secondary photolysis can account for the fact that conversion to Mo 2(µ-OH) 2(aq) 4+ is not quantitative at high conversions for the photolysis carried out in 3 M HCL In fact, the development of a shoulder at 300-305 nm at long irradiation times (Figure 2) is evidence for the product of secondary photolysis. No spectral changes were observed when a solution of Mo2(O2CCH 2NH3 ) 4 4+ in 0. 25 M HCl was irradiated for 12 hr with the unfiltered radiation of a 200 W Hg-Xe lampo Photooxidation Reactions in Chlorocarbon Solvents - When a solution of Mo 2C¼[ P(Et) 3 ] 4 is photolyzed at 254 nm, the spectral changes shown in Figure 4 occur. A band attributable to the 6 ... o* absorption decreases concomitant with the growth of an intense absorption at 314 nm and a shoulder develops ca 400 nm. In more concentrated solutions at long irradiation times two weak absorptions 135 are found at 515 and 570 nm. An isosbestic point is initially (< 50% photolysis) maintained at 525 nm. Similar spectral changes are observed when Mo 2 C1,J P(Et) 3 ] 4 reacts with CC1 3 and CHC13 by a thermal pathway. In both cases an intense absorption appears at 314 nm, a shoulder develops ca 400 nm, and broad weak absorptions peaking at 510 and 570 nm are detected. Initially, an isosbestic point obtains between 525 and 530 nm. Clearly the thermal and photochemical processes yield very similar or identical products. 25 Walton and coworkers have shown that workup of the reaction mixture of Mo2 Cl4 [ P(Et) 3 ] 4 with CC14 in CH 2 Cl2 yields 29% of (Et3 PC1) 3 Mo 2 Clg; however, this product will not redissolve in CH 2Cl 2 • We have observed the formation of small amounts of this compound when the product of the photochemical re action (see Experimental Section) is recrystallized from CH 2 Cl 2 -CC14 • Addition of ether to the photolyzed solutions of Mo 2 Cl4 [ P(Et) 3 ] 4 yielded an oil which eventually could be induced to form a powder with irreproducible elemental analysis. The infrared spectrum of this compound (KBr pellet) in the Mo-Cl stretching region showed bands at 330 (s), 309 (s), and 280 (m) very similar to those at 334 (s), 307 (s), 275 (m), and 250 (w) in (Et 3 PC1) 3 Mo 2 CJ.g. Photolysis (254 nm) of Mo 2 C14 [ P(n-Bu) 3 ] 4 and Mo 2ClJ P(Et)(Ph) 2] 4 in CH 2Cl2 results in spectral changes similar to those observed for the photochemical reaction of Mo 2CliP(Et) 3 ] 4 with CH 2Cl2. The spectra of the photoproducts from the reaction of Mo2CliP(n-Bu) 3 ] 4 with CH 2 C12, EtCl, and CHC13 were identical. It was possible to isolate a CH 2 C1 2 soluble product from the photochemical reaction of Mo 2 C14 [P(Et)(Ph) 2] 4 with CH 2Cl 2 136 whose elemental analysis and molecular weight were consistent with an Mo 2 C4,[ P(Et)(Ph) 2]s formulation; however, its electronic spectrum was not identical to that of the photolyzed solution. It therefore appears likely that photolysis of Mo 2Cl4 P 4 compounds initially yields an Mo 2Cl 5 P 4 or a similar species which subsequently can loose phosphine ligands ultimately producing the M(½Clg s- ion. Quantum yields at 254 nm for the photochemical reactions of Mo2 Cl4 P 4 (P = P(Et) 3 , P(n-Bu) 3 , and P(Et) (Ph) 2 ) with various chlorocarbon solvents are presented in Table 1. For all photooxidation reactions both in aqueous (vide supra) and chlorocarbon solvents, no appreciable photoactivity (Le. , <P > ~ 0. 0005) was observed when • samples were excited at 546, 505, and 436 nm. As for the photo2 chemistry of Re Cl - 26 no reactivity occurs when the lowest energy 2 8 o - o* transition is excited. Discussion In summary, the photochemical processes of binuclear Mo(II) compounds upon 254 nm excitation are Mo2(SO4)4 3- + 1/2 H2 H /0"'-. (aq)Mo "'-./ 0 H (aq)HCl Mo(aq) 4+ + 1 H2 137 RX "' / -Mo(III) [ / ..___ Cl ~ C~ l - Mo(III)~ Cl ----- l Spectroscopic studies 27 - 3o of compounds which contain quadruple metalmetal bonds have demonstrated that the lowest energy visible absorption band arises from the metal localized one-electron N- V excitation o l>* 0 Irradiation of this absorption band in Re 2 Cl 8 2- and in those compounds studied in this work does not lead to metal-metal bond cleavage. When Re 2 Cl 8 2- in CH 3 CN is subjected to ultraviolet 26 irradiation (A < 366 nm) rupture of the quadruple bond is observed. 31 This is now thought to proceed by a solvent assisted pathway rather than by homolytic cleavage of the ReRe bond. As shown herein, UV irradiation of binuclear molybdenum (II) compounds yields oxidized metal species and a binuclear structure is maintained in the products. These results are in sharp contrast to the photochemical reactions of other compounds which contain metal-metal bonds where, in general, homolytic bond cleavage obtains. lO, 32 For example, excitation of the a - a* trai:1sition 33 in M2 (CO) 10 (M = Mn, Tc, Re) 32 compounds leads to efficient metal-metal bond cleavageo lO, 138 A recent spectroscopic analysis 34 suggested values of 115-130 and 110-160 kcal/mole for the metal-metal bond energies in Re 2Cl8 2- and Mo 2Cl8 4, respectively. Also, the low lying (700 - 350 nm) metal-localized electronic absorption bands of these compounds often exhibit well-resolved vibronic structure 27 - 30 , 34 , 35 and the excited state value of a 1 g (MM) is consistent with strong, residual metal-metal bonding in these excited states. The large metal-metal bond energy and the absence of an appreciable weakening of this bond in low lying metal-localized excited states explains the absence of photolytic fragmentation in compounds which contain a quadruple metal-metal bond. The magnetic circular dichroism (MCD) spectrum of Mo2Cl3 4- in 6 M HCl shows an A term at 270 nm which has been attributed to an allowed Cl(1r) - o* ligand to metal charge transfer (LMCT) transition. Therefore it is likely that a LMCT state is initially produced when Mo 2Cl 8 4- and Mo2Cl4 P 4 are photolyzed at 254 nm. Because intersystem crossing between charge transfer and metal localized states may be slow 36 , 37 it is possible that a LMCT state is responsible for the photochemistry. absorption bands of Mo 2(aq) identified yet. 4+ and Mo2(SO4 ) 4 4- The LMCT have not been Using the difference in optical electronegativities 38 - - between Cl and H2 O and between Cl and SO4 2- the LMCT bands are estimated at 190 and 230 nm, respectively. The quantum yields (Table I) for Mo 2Cl8 4- and MoiSO4 ) 4 4- are similar while that of 139 Mo 2 (aq) 4+ is much less. This is consistent with the proximity of the LMCT band with respect to 254 nm excit.ation. It is interesting to note that Mo 2(O2 CCH 2NH3 ) 4 the ultraviolet (>.. ;,,. 200 nm)o 4+ did not react when irradiated in For Mo 2(O2CH) 4 the lowest energy charge transfer transition centered at 300 nm has been assigned 39 to a metal to ligand charge transfer (MLCT) transition. No LMCT transitions were found for Mo 2(O2CH) 4 for X ~ 200 nm.. The solution spectrum reportect 15 for Mo2(O2 CCH 2NH 3)/+ is virtually identical to that of Mo 2 (O2 CH) 4 • Therefore, there is strong circumstantial evidence which suggests that the photoactive state possesses substantial LMCT charactero A particularly significant observation is the intermediacy of a metal hydride for the photochemical production of hydrogen from Mo2Cla 4- . One reasonable mechanism for the formation of the hydride is as follows 140 An analogous scheme could be drawn for Moiaq) 4+ and Mo 2 (SO4 ) 4 except that for the latter compound attack of MoiSO4 ) 4 MoiSO4)4 4- 3- to yield Mo2 (SO4)4 2- 4- on must obtain. The oxidation state of molybdenum in the LMCT state is formally 1. 5. Binuclear Mo (1. 5) compounds have never been isolated. An extremely unstable Mo 2 (O2 CCF 3 ) 4- species has been generated by pulsed radiolysis of methanol solutions containing Mo 2 (O2 CCF 3 ) 4. 40 Mechanisms for the reactions of Mo 2 C¼P4 with chlorocarbons can be imagined similar to the above or involving electron transfer to solvent. There is no reason to favor any mechanism at this point. Flash photolysis studies would be of help in the further elucidation of these reactions. - We thank Professor George Hammond for helpful discussions. W. C. T. held an IBM Fellowship for 1976-1977. This research was supported by the National Science Foundation Grant No. CHE 75l9086. 141 References ~ 1. V. Balzani and V. Carassiti, Compounds, 2. Photochemistry of Coordination Academic Press, 1970, and references therein. P. C. Ford, R. E. Hintze, and J. D. Petersen in Concepts of Inorganic Photochemistry, ed. by A. W. Adamson and P. D. Fleischauer, John Wiley and Sons, 1975, and references therein. 3. M. Fox in Concepts of Inorganic Photochemistry, ed. by A. W. Adamson and P. D . Fleischauer, John Wiley and Sons, 1975, and references therein. 4. D. K. Erwin, G. L. Geoffroy, H. B. Gray, G. S. Hammond, E. I. Solomon, W. C. Trogler, and A. A. Zagars, J. Am. Chem. (1977). 5. J. Jortner and G. Stein, J . Phys. Chem., ~ ' 1258 (1962). 6. J. Halpern, G. Czapski, J. Jortner, and G. Stein, Nature, -"'-'""-" 186, 629 (1960) and references therein. 7. J . Jortner and G. Stein, J . Phys. Chem.,~' 1264 (1962). 8. J. C. D. Brand and W. Snedder, Trans. :!:i'araday Soc., .,.,,,,.,_ 53, 894 (1957); 0. Traverso and F. Scandola, Inorg. Chim. Acta, 4, "' 493 (1970). 9. O. Traverso, S. Sostero, and G. A. Mazzocchin, Inorg. Chim. Ac ta, 11, 2 3 7 (1 974). 10. M. S. Wrighton, Chem. Rev., 74, 401 (1974). """ lL J . V. Brencic and F. A. Cotton, Inorg. Chem., 9, 351 (1970). "' 12. F. A. Cotton, B. A. Frenz, E. Pedersen, and T. R. Webb, Inorg. Chem., .,.,,,... 14 , 391 (1975) . 142 13. M. J. Bennett, J. V. Brencic, and F. A. Cotton, Inorg. Chem. , !!_, 1060 (1969). 14. F. A. Cotton and B. J. Kalbacher, Inorg. Chem. , 15, 511 (1976). -""""' It was shown that the formulation given in ref. 13 was incorrect. 15. F. A. Cotton and T. R. Webb, Inorg. Chem., 15, 68 (1976). """ 16. Jo San Filippo, Inorg. Chem., ]1, 3140 (1972). 17. H. D. Glicksman, A. D. Hamer, T. J. Smith, and R. A. Walton, Inorg. Chem., 15, 2205 (1976). """ 18. A. R. Bowen and H. Taube, Inorg. Chem., 13, 2245 (1974). """" 19. J. San Filipo, H. J. Sniadoch, and R. L. Grayson, Inorg. Chem., 13, 2121 (1974). -""""' 20. J. G. Clavert and J. N. Pitts, "Photochemistry," John Wiley and Sons, Inc., New York, 1966. 21. W. D. Bowman and J. N. Demas, J. Phys. Chem., .,._ 80, 2434 (1976) . 22. F. A. Cotton, D. S. Martin, P. E. Fanwick, T. J. Peters, and T. R. Webb, J. Am. Chem. Soc., .,.._,.., 98, 4681 (1976) . 23. H. Hartmann and H. J. Schmidt, Z. Phys. Chem., (Frankfurt am 24. Main), 11, 234 (1957). """ M. Ardon and A. Pernick, Inorg. Chem., 13, 2275 (1974). """ 25. H. D. Glicksman, A. D. Hamer, T. J. Smith, and R. A. Walton, Inorg. Chem., 15, 2205 (1976) . .rv-,. 26. G. L. Geoffroy, H. B. Gray, and G. S. Hammond, J. Am. Chem. Soc., 96, 5565 (1974)n - - """" 27. C. D. Cowman and H. B. Gray, J. Am. Chem. Soc., ~' 8177 (1973). 143 28. C. Do Cowman, W. Co Trogler, and Ho B. Gray, Israel J. Chem., in press. 29. P. E. Fanwick, D. S. Martin, F. A. Cotton, and T. R. Webb, Inorg. Chem., in press. 30. F. A. Cotton, D. S. Martin, P. E. Fanwick, T. J. Peters, and T. R. Webb, J. Am. Chem. Soc. , ""'98, 4681 (1976). 31. R. H. Fleming, G. L. Geoffroy, H. B. Gray, A. Gupta, G. S. Hammond, D. S. Kliger, and V. M. Miskowski, J. Am. Chem. Soc., .,._.,..,_ 98, 48 (1976) . 320 Co Ro Bock and E. Ao .Koerner Von Gustorf, Adv. in Photochem., 10, 221 (1977). ""'- 330 R. A. LevensonandHo B. Gray, J. Am. Chem. Soc., .,--.,... 97, 6042 (1975). 34. W. C. Trogler, C. D. Cowman, Ho B. Gray , and F . Ao Cotton, J. Am. Chem. Soc. , "-"99, 2993 (1977). 35. W. C. Trogler, E. I. Solomon, I. Trajberg, C. J. Ballhausen, and H. B. Gray, Inorg. Chem., .,.,_.,,_ 16, 828 (1977) . 36. Y. Ohashi, K. Yoshihara, and S. Nagakura, J. MoL Spectrosc., 38, 43 (1971). ""'- 37. G. A. Crosby, K. W. Hipps, and WO Ho Elfring, J. Am. Chem. Soc. 96, 629 (1974). ""'- 38. C. K. J<t>rgensen, "Essays in Coordination Chemistry", Birkhauser Verlag, Basel, 1964. 39. J . G. Norman, H. J. Kolari, Ho B. Gray, and W. Co Trogler, Inorg . Chem. , 16, 987 (1977). "'"' 144 40. J. H. Baxendale, C. D. Garner, R. G. Senior, and P. G. Sharpe, J. Am. Chemo Soc. , 98, 637 (1976). ~ 145 Table 1. Disappearance Quantum Yields at 254 run for Photooxidation Reactions of Binuclear Mo{II} Compounds Compound [CompoundJ Moles / ! K4 Mo2 (SO4 ) 4 4. 41 X 10- K4 Mo2 (SO4 ) 4 ~ax Solvent (o-o*) q,254 5.0 M ~SO4 150 0.19 3. 90 X 10- 3 2.5 M ~SO4 207 0.17 K4 M~(SO4 ) 4 3.90x10- 3 0.5 M ~SO4 230 0.16 K4 Mo2 (SO4 ) 4 3. 72 X 10- 3 0. 05 M ~SO4 237 0.15 K4 M~(SO4 ) 4 3. 84 X 10- 3 0. 001 M ~SO4 250 0.18 3 1 M HO3 SCF3 337a 0.035 1. 93 X 10- 3 1 M HO3 SCF3 / 337 0.031 Mo2 (aq) M~(aq) 4+ 4+ 3 2.15 X 10- 0015 M CflsOH K4 M~Cl8 ·2~0 1. 71 X 10- 3 3M HCl 960 0.14 K4 Mo2 Cl 8 •2H2 O 1. 06 X 10- 3 3M HCl 960 0.13 K4 M~Cl 8 • 2~0 9. 33 X 10- 4 6 M HCl 1050 0.092 Rb 3 Mo2 Cl 8 H 5.37x10 6 MHCl 1400b 0.040 Mo2 Cl 4 [ P(Et\] 4 3. 06 X 10- Cf½C½ 3065 0.055 Mo2 Cl 4 [ P(Et\j 4 4.54x10 C1¾C½/ 3065 0.055 CHC1s 3065c 0.060 CHC1 3 3065c 0.063 -4 4 -4 ~10 -5 M in galvinoxyl -4 Mo2 Cl 4 [ P(Et\J 4 3. 04 X 10 Mo2 Cl4 [ P(Et) 3 ] 4 1. 70 X 10- 4 146 Table 1 (Continued) Compound [Compound] Moles/£ Mo2 Cl 4 [P(i:!__-Bu) 3 L 2. 47 x 10- Solvent 4 Emax (o-o*) 4)254 EtCl 3150c 0.016 4 M~Cl 4 [P(g-Bu\] 4 4. 20 X 10- C~C½ 3150 0.058 M~CliP(n-Bu) 3 L 4.29X10- CHC13 3150c 0.054 4 Mo2 Cl 4[P(Et)(Ph\L 5. 31 x 10- 5 c~c½ 3125 0.019 M°-2Cl4[P(Et)(.Ph\ L 1. 62 X 10 - 4 CHC1 3 3125c 0.035 aValue of Emax taken from ref. 18. bThe Emax value is for the band centered at 419 nm which is not o - o* in this compound. cThese values assumed identical to those obtained in CH.zC½. 147 Figure Captions Figure 1. Electronic absorption spectral changes during 254 nm irradiation of M~ (aq) 4+ in 1 M HSO3 CF 3 solution. Spectral changes were observed to be more complex for concentrations of HO3 SCF 3 less than 0.1 M. Figure 2. Electronic absorption spectral changes during 254 nm 4- irradiation of Mo 2Cl8 in 3 M HCl solution. For the 300-500 nm region spectra for a more concentrated sample are also shown. Figure 3. Electronic absorption spectral changes of a 5. 84 x 10 -3 solution of Rb 3 Mo 2Cl 8 H in 3M HCl subsequent to mixing. The cell path length was 0. 1 cm. Figure 4. Electronic absorption spectral changes during 254 nm irradiation of Mo 2 C1J P(Et) 3 ] 4 in CH 2 Cl2 • M 149 ABSORBANCE (jJ 0 0 (Jl 0 0 CJ) 0 0 150 0 0 (JJ 0 0 ~ 0 0 z ~ en 0 0 CD 0 0 ----J 0 0 ABSORBANCE 0 (Jl b 151 ABSORBANCE 0J 0 0 ~ 0 0 en 8 0 s: 0) 8 0 """ 0 z