Spectroscopic Investigation of the Structure of a Novel Zerovalent Cobalt Nitrosyl in Zeolite Matrices BY H ~ B N E PRALIAUD, GIS~LE F. COUDURIER AND YOUNES BEN TAARIT* C.N.R.S., Institut de Recherches sur la Catalyse, 79, Boulevard du 11 Novembre 1918, 69626 Villeurbanne Cedex, France Received 12th April, 1978 Cobalt (II) hexahydrate introduced in synthetic zeolites by conventional ion-exchange was shown to give up readily five of its water molecules upon evacuation at room temperature. Subsequently the cobalt ion was coordinated by three lattice oxide ions from the zeolite lattice and by a single water molecule so as to acquire a slightly distorted tetrahedral symmetry. This tetrahedral CoII species readily reacted at room temperature with nitric oxide to yield a zerovalent paramagnetic cobalt dinitrosyl within the zeolite framework.The complex was identified on the basis of its i.r., U.V. and e.p.r. spectra as a C2h distorted octahedral complex with a dX2-,,2 ground state with a co- ordination sphere made up of three lattice oxide ions [possibly O(4) ions], a water molecule and two NO ligands which transferred their odd electrons to the cobalt d orbitals. In recent years, transition metal exchanged zeolites attracted much attention and were studied using a myriad of techniques. The prominent feature of these solids resides in the ease with which they form within their framework a variety of complexes upon interaction of transition metal zeolites with various organic and/or inorganic reagents. In fact, even when initially located in remote sites, transition metal ions were conclusively shown to migrate to accessible sites in order to co- ordinate adsorbed molecuIes.1-7 These phenomena have been observed by both direct and indirect methods. E.s.r. spectroscopy appeared to be one of the most efficient and informative means to investigate such cation behaviour within the zeolite lattice, though powder patterns could provide valuable information on the symmetry and electronic structure of the complexed ion and generally on the location of the complex. As to their structure, most transition metal complexes were shown to behave in a similar fashion to their solution analogues, although specific interaction with the zeolite framework could also modify this behaviour. In particular, the zeolite lattice appeared to be a stabilising matrix for a variety of cations in unusual coordination or oxidation states.2* 4* 8*12 We report in this study the formation of a novel zerovalent cobalt nitrosyl complex generated within the zeolite framework. EXPERIMENTAL AND MATERIALS The starting zeolite material was a synthetic faujasite type zeolite (both X and Y) with respective unit cell formulae (NaA102)86(Si02)1 6xH20 and (NaA102)5 6(Si02)1 6yH20.The cobalt zeolites were obtained by conventional ion-exchange of the parent materials with dilute cobalt (11) sulphate solutions. Several exchange levels (2.9, 6, 8.5, cobalt %) were reached upon repetitive procedures. The samples were then thoroughly washed with deionised water. E.s.r. spectra were recorded on a Varian E9 spectrometer equipped with a dual cavity operating in the X band mode and DPPH was used as a standard for g value determination.Q band spectra were obtained using a variable temperature accessory, Mn2+ impurities in the zeolite were used as a convenient internal reference. 3000H. PRALIAUD, G. F. COUDURIER AND Y. BEN TAARIT 3001 Diffuse reflectance spectra were recorded, at room or liquid nitrogen temperature in the range 220 to 2500 nm, on an Optica Milano CF4N1 spectrometer using a differential method, the reference being MgO or the starting zeolite. Infrared measurements were performed on a Perkin Elmer spectrometer model 125. The zeolite sample was compressed into a thin wafer at a pressure of 1 ton cm-2 and inserted into a sample holder. A conventional i.r.cell, as described elsewhere,13 was used. Nitric oxide was purified by the freeze-pump-thaw technique and finally distilled from a liquid nitrogen+pentane slurry (138 K) to remove NOz impurities. RESULTS Fully hydrated cobalt zeolites were light pink in colour and exhibited only a broad e.s.r. signal at g x 3.3 at whatever temperature, The U.V. spectrum showed two absorption maxima at 525 and 1200 nm [fig. l(a)J. Upon evacuation at room temperature (1-2 h) to a final pressure of Torr, the cobalt zeolite turned purple and the e.s.r. spectrum was not significantly affected. By contrast the U.V. spectrum was modified drastically : the absorption maximum at 525 nm decreased in intensity and an absorption occurred in two ranges : 1200-1600 nm on the one hand and 520-640 nm with maxima at 540, 575 and 620 nm on the other hand, as shown in fig.l(b). The infrared spectrum still showed the 6H,O strong band at 1640 cm-l characteristic of molecular water and at 1605 cm-l due to H20 specifically co- ordinated to Co2+ ions.14 I I I 1 I 1 1 . 200 400 600 800 A/nm FIG. l.-(a) U.V. diffuse reflectance spectrum of COII(H~O)~ in faujasite type zeolites. (b) U.v spectrum of CoII zeolite degassed at room temperature for 2 h. (c) U.V. spectrum of the nitric oxide adduct. Adsorption of nitric oxide resulted in the growth of two strong vNO bands at 1890 and 1806 cm-l (fig. 2). Both bands grew at the same rate upon further contact with NO. Their growth did not affect the band at 1605 cm-l. The sample now turned grey at room temperature and greenish-grey at 77 K.Under these con- ditions a new e.s.r. spectrum could be recorded at 77 K (fig. 3). On progressively warming the sample, the signal intensity decreased and the features broadened. At around 153 K, the signal was broadened beyond detection. The X band pattern shown in fig. 3 is characteristic of an intense hyperfine interaction of the electron with a spin 712 cobalt nucleus. The g and hyperfine principal components which3002 I . R . , U.V., E.P.R. OF COO NITROSYL IN ZEOLITES could be extracted from the X band spectrum are listed in table 1. As the con- siderable line overlap set a lower limit for the accuracy of the g and A values, Q band spectra were sought and they confirmed the departure of the g tensor from axial symmetry (fig.4). However, in spite of the fairly low temperature achieved in the variable temperature device, the Q band spectrum was obscured by appreciable line broadening thus precluding any improvement in the accuracy of the gx and gr components. Again the g and A values are also listed in table 1. They are in fair agreement with those deduced from the X band spectrum. In the U.V. spectrum, absorptions at 240, 360, 460-70, 750 and 1200 nm could be seen [fig. l(c)] at all temperatures. - * 1806 1 I 9 Y FIG. 3.-X band spectrum of the nitric oxide adduct recorded at 77 K.€I. PRALIAUD, G. F. COUDURIER AND Y. BEN TAARIT 3003 t * I I l I I J gu FIG. 4.-Q band spectrum of cobalt nitrosyl recorded at 108 K. TABLE 1 . C . s . ~ . PARAMETERS OF COBALT (0) ,NITROSYL X band 2.13 (0.05) 2.11 (0.05) 2.36 (0.05) +118 (5) +118 (5) -75 (5) mode gx 0, Bz AxIG-1 AxlG-1 A.IG-1 Q band 2.14 (0.05) 2.10 (0.05) 2.29 (0.05) +120 (5) +12O (5) -75 (10) DISCUSSION The magnitude of theihyperfine coupling to the cobalt nucleus together with the strong g shift from the free electron g value, clearly indicate that such species as NO, NO2, NO;-, etc., should be ruled out and that the odd electron is most certainly a cobalt d electron. However, the low spin d7C0" complexes are usually octahedral or square- pyramidal with a d,Z ground state.6* I 5 - l 7 The gz value for such complexes was typically very close to the free spin g value while the other two g components were usually larger than both gz and ge.This is obviously not what is observed for this particular nitrosyl complex.Furthermore, the formation of a d7 nitrosyl compound would require the presence of two NO molecules and a formal dismutation of the NO ligand into NO- and NO+ species to ensure the d7 configuration : this is definitely inconsistent with the experimental infrared spectrum. A single NO ligand could also secure a d7 cobalt configuration, but the odd electron of the NO molecule would still be unpaired and we would have to deal with a triplet state situation. COORDINATION OF COBALT SPECIES The hydrated cobalt zeolite showed a U.V. spectrum characteristic of octahedral COII(H~O)~ in the large cavities of the zeolite, the 525 and 1200 nm maxima being respectively due to the 4T,,(F) + 4T1,(P) and 4T2,(F) transitions. Following partial dehydration and prior to NO adsorption, the U.V.spectrum showed the existence of cobalt (11) species in tetrahedral symmetry, as could be deduced from the appearance and development of the two ranges of absorption between 520-640 and 1200-1600 nm at the expense of the absorption due to [Co(H20),l2+. The existence of the two absorption ranges is due to transitions from the non-degenerate ground state 4A,(P) to the triply degenerate excited 4T1(P) and 4T1(F) states. The splitting of the bands corresponding to the excitation to the 4T1(P) level with maxima at 540, 575 and 620 nm could be due to dynamic Jahn-Teller effect or to a low symmetry3004 I.R., U . V . , E . P . R . OF COO NITROSYL ILN ZEOLITES perturbation which lifts the excited state degeneracy. The tetrahedral Co" cation should therefore be located in an SII position at the hexagonal windows within the supercage.Upon adsorption of nitric oxide the decrease in the 520-640 nm absorption and the simultaneous occurrence of bands at 360, 460-70 and 750-60 nm are indicative of changes in the coordination and electronic configuration of the cobalt complex. The infrared spectrum of the nitric oxide adduct is consistent with a dinitrosyl species. In fact the two vNO bands are similar to those observed in the case of well defined nitrosyl compounds where the bands at 1890 and 1806 cm-l are due to the symmetric and antisymmetric NO vibration modes. The presence of a residual water molecule in the coordination sphere of the cobalt (11) ions as witnessed by the presence of the 1605 cm-1 band in addition to three oxide ions prior to NO adsorption is consistent with the tetrahedral symmetry inferred from the U.V.spectrum. Addition of two NO molecules as deduced from the i.r. spectrum leads to an octahedral type symmetry. It is highly probable that both the zeolite lattice and the specific ON-Co NO bond angle contribute to impose drastic alteration of the octahedron. The occurrence of the vNO vibrations at 1890 and 1806 cm-l indicative of a dinitrosyl species and the following analysis of the U.V. spectrum strongly favour a dgCoO dinitrosyl structure arising from the transfer of both odd electrons of the two NO ligands to the initial Co" tetrahedrally coordinated ion thus forming a Coo octahedral dinitrosyl. Indeed the 460-70 and 750-60 nm transitions characteristic of the NO adduct are in agreement with a distorted algCo0 complex : the 460-70 nm absorption is assigned to an internal Coo transition from the ground 2D(3d9) state to the excited 2F(3d84s1) configuration.l* On the other hand the 750-60 nm transition is more difficult to assign but also more informative.Copper" d9 distorted Oh complexes are known to exhibit d-d transitions in the 620-930 nm range.19 In particular [CU(H~O)~]~+ in zeolite absorbs at 800 nm.20 Substitution of water by N-ligands such as NH, shifts the absorption maximum to around 620 to 570 nm.21 Similarly dgN? complexes have been identified in zeolite matrices following either Hz reduction of Ni" ions 22 or addition of nitric oxide to Ni" exchanged In the former instance Nil species gave rise to an absorp- tion maximum around 740 nm ascribed to an internal transition of the Nif ion.In the latter case octahedrally distorted Ni' nitrosyl identified on the g basis of its e.p.r. spectrum (911 = 2.365 and gl = 2.193) and its i.r. spectrum (vNO at 1892 cm-l) gave rise to transitions at 360, 650,740 and 1650 nm.23 As previously 22 the 740 nm band is characteristic of the internal (d9-d84s1) transition of Nil while the 650 nm absorption is due to d-d transition of 3d9Ni1 distorted octahedral complex. As the charge on the central atom of the octahedrally distorted complex having N ligands de- creased from 2 to 1 Cu" and Ni', the energy of the corresponding d-d transition was also expected to decrease, as was actually observed, from 570 to 650 nm.Therefore, the transition energy of the isoelectronic Coo octahedral nitrosyl complex, con- sidering the zero charge on the central atom, should be even lower ; hence, the 750 nm maximum seems to be a likely candidate for such a d-d transition. In fact Coo complexes have already attracted considerable interest in recent years. In particular, paramagnetic cobalt tetracarbonyl has been reported to form upon sublimation of CO,(CO)~ onto a cold finger held at 77 K in the microwave cavity of an e.p.r. spectr~meter.~~ This radical species was also studied by Symons et aZ.25 and its reactivity towards O2 demonstrated. More recently Ozin and coworkers26 reported an i.r., e.p.r. and U.V. study of Co(CO), and CoCO),. In a more recentH . PRALIAUD, G. F .COUDURIER AND Y. BEN TAARIT 3005 study, Symons and coworkers 2 7 succeeded in isolating a substituted cobalt (0) carbonyl [Co(CO), pb PhJ- which is thought to retain the structure of the parent carbonyl, i.e., trigonal bipyramidal. Coo paramagnetic complexes were also pre- pared by chemical or electrochemical reduction of Co' organometalh compounds and again a distorted tetrahedral structure was favoured.28 Lastly a mixed nitrosyl dicobalt carbonyl [Cp,(Co),NOCO] and the Cp2(Co),(CO), anion were shown to be paramagnetic and relevant g parameters to be close to those reported for Co(CO), etc. Simultaneous X-ray structural determinations showed both com- pounds to have a distroted tetrahedral stru~ture.~ The overriding conclusion emanating from all these studies was the general agreement on the magnitude and ordering of the g values as can be seen from table 2: g1 > gz N ge. In all cases this was interpreted as due to a dgCoo complex in a C,, distorted tetrahedral sym- metry consistent with a d,t ground state.TABLE 2 . C . S . R . PARAMETERS OF DISTORTED TJ3TRAHEDRAL coo COMPLEXES compound reference g, gY gz AJG-1 Ay/G-l Az/G-I Co(CO), in C O ~ ( C O ) ~ matrix 24 2.134 2.134 2.02 56 56 67 Co(CO), in CO matrix 26 2.128 2.128 2.007 55 55 58 Co(CO), Pb Ph; in Co(CO), Pb Ph3 27 2.020 2.017 1.996 -48 -48 f 3 7 As one can see from table 1 our g values differ from those reported in table 2 in the ordering, with an additional splitting of the perpendicular component into two distinct values both of them larger than ge, and gz far larger than gx and gr.On the basis of the relative magnitude of the g tensor components st distorted tetrahedral symmetry with a dz2 ground state should be definitely ruled out, which is reasonable in view of the ligand count. By contrast, this particular ordering of the g values is similar to that observed for d9Cu2+ complexes in distorted octahedral symmetry and to that reported for Ni' nitrosyl complexes formed within the zeolite cavities 4* and also to NO2 adduct also forming Ni' specie^.^ All these complexes were shown to have distorted octahedral symmetry with a dxz-y2 ground state. Additional evidence for a CjX2-,,2 ground state could be provided upon resolving the hyperfine tensor. Within experimental error, the hyperfine tensor appeared as nearly axial, and could be resolved into its isotropic and dipolar parts in four possible ways depending on the choice of sign for Al and All, i.e., All = +75 and AL = 11 8 G.Taking the signs as all positive or all negative results in too small an aniso- tropy to account for the 3d character of the electron. Taking A , , = +75 and Al = - 118G yields Aiso = - 53.7 and 2B = + 128.7. This is consistent with a dz2 ground state, which had to be ruled out on the basis of the ordering and magnitude of theg values with respect tog,. The alternative choice, which we favour, gives Aiso= 59.7 and 2B = - 128.7. This result is consistent with a high population in thedx2-,2 orbital and is also consistent with the observed g tensor and the U.V. spectrum. The residence time in this orbital, considering that for unit occupancy 2B0 = - 152.3 G,27 is quite reasonable = 84.5 %.On the other hand the usual value for Aiso for cobalt com- plexes is around - 100 G. To achieve a positive value would imply an appreciable admixture of the dX2-,,2 orbital with the outer 4s shell. This is quite reasonable in the case of D2,, or CZh distortion of the octahedral symmetry. Such a distortion3006 I . R . , U . V . , E . P . R . OF COO NITROSYL I N ZEOLITES within the zeolite framework could scarcely be avoided. Thus the approximate 4s population that could be deduced from this positive value is within 11.5 %. Hence 96 % of the odd electron is exclusively delocalized on the metal atom. This is consistent with the absence of any detectable nitrogen superhyperfine coupling.The departure of the g tensor from axiality is also consistent with a C,, or DZh deviation from octahedral symmetry. The rigid zeolite matrix can of course undergo an appreciable Si A1 alteration in order to accomodate various ions and complexes ; however, it is doubtful whether these alterations could be important enough to secure perfect symmetries in the case where oxide ions of the lattice are part of the coordination shell. Indeed X-ray studies 30 showed that 0-Co-0 angles in Co-Y zeolites are far larger than 90" : they are 97" to 117" depending on the nature of other attached ligands. Furthermore the vNO vibration bands relative intensities give a useful estimate of N-Co-N bond angle using the formula :31 4 = 2 arcotan JR(sym)/R(asym) where R is the intensity of the particular vibration mode, the N-Co-N bond angle deduced from the experimental R(sym)/R(asym) ratio of 0.37 is about 117" and clearly indicates a large departure from D2,, symmetry, which is expected considering the structural nature of the host.In conclusion the isolated complex is best described as a d9 Coo dinitrosyl complex with a ~,z-,J ground state in a C2h distorted octahedral symmetry. This is the first example of an octahedral zerovalent paramagnetic nitrosyl complex. The occurrence of such a complex is obviously favoured by the ability of the zeolite matrix to accommodate and stabilize such unusual valence and coordination states. Note that the complex was simply produced by NO addition and did not require any extreme treatment to be formed, thus emphasizing the suitability of the zeolite matrix for this particular type of chemical study.0 \ / \ / / \ / \ The authors gratefully acknowledge discussions with Prof. Che of the Pierre and Marie Curie University and with Dr. B. Imelik and Dr. C. Naccache. Thanks are due to G. Wicker and J. F. Dutel for technical assistance. C. Naccache and Y. Ben Taarit, Chem. Phys. Letters, 1971, 11, 11. C. Naccache and Y . Ben Taarit, J. Catalysis, 1971, 22, 171. P. Gallezot, Y . Ben Taarit and B. Imelik, J. Catalysis, 1972, 26, 295. C. Naccache and Y. Ben Taarit, J.C.S. Faraday I, 1973,69, 1475. P. Gallezot, Y . Ben Taarit and B. Imelik, J. Phys. Chem., 1973,77, 2364. E. F. Vansant and J. H. Lunsford, J.C.S. Chem. Comm., 1972, 830 and Trans. Faraday Soc., 1973,69,1028. ' K.A. Windhorst and J. H. Lunsford, J. Amer. Chem. SOC., 1975,97, 1407. * C. Naccache, Y. Ben Taarit and M. Boudart, A.C.S. Symposium series, Molecular Sieves 11, 1977, 40,156. P. H. Kasai and R. J. Bishop Jr., J. Amer. Chem. SOC., 1972,94,5960. lo C. Naccache, J. F. Dutel and M. Che, J. Catalysis, 1973, 29, 179. l1 M. Che, J. F. Dutel, P. Gallezot and M. Primet, J. Phys. Chem., 1976, 80, 2371. l2 E. Garbowski and J. C. Vedrine, Chem. Phys. Letters, 1977,48,550. l3 M. V. Mathieu and P. Pichat, in La Catalyse auLaboratoire et dansl'lndustrie (Masson, Paris, l4 G. F. Coudurier, to be published. l5 (a) J. P. Maher, J. Chem. SOC. A, 1968, 2918 ; (b) M. E. Kimball, D. W. Pratt and W. C. l6 M. C. R. Symons and J. G. Wilkinson, J. Chem. Soc. A, 1971,2069. 1967), p. 319. Kaska, Inorg. Chem., 1968, 7, 2006.H. PRALIAUD, G. F. COUDURIER AND Y. BEN TAARIT 3007 l7 F. D. Tsay, H. B. Gray and J. Danon, J. Chem. Phys., 1971,53,3760. l9 B. J. Hathaway and D. E. Billig, Coord. Chem. Reu., 1970,5,143. 2o J. D. Mikheikin, V. A. Shvets and V. B. Kazanskii, Kinetika i Kataliz, 1970, 11, 747. 'l J. H. Anderson, J. Catalysis, 1973, 28,76. 22 E. Garbowski, M. V. Mathieu and M. Primet, Chem. Phys. Letters, 1977, 49,247. 23 H. Praliaud, to be published. 24 H. J. Keller and H. Wawersik, 2. Naturforsch, 1965, 20b, 938. 2 5 S. A. Fieldhouse, B. W. Fullam, G. W. Neilson and M. C. R. Symons, J.C.S. Dalton, 1974,567. 26 L. A. Hanlan, H. Huber, E. P. Kiindig, B. R. McGarvey and G. A. Ozin, J. Amer. Chem. SOC., 27 0. P. Anderson, S. A. Fieldhouse, Ch. E. Forbes and M. C. R. Symons, J. Orgammetal. 28 H. Van Willigen, W. E. Geiger Jr. and M. D. Rausch, Inorg. Chem., 1977, 16, 581. '' I. Bernal, J. D. Korp, G. H. Reisner and W. A. Herrmann, J. Organometal. Chem., 1977,139, 30 P. Gallezot and B. Imelik, J. Chim. phys., 1974, 155. 31 W. Beck, A. Melnikoff and R. Stahl, Chem. Ber., 1966,99, 3721. C. E. Moore, Atomic Energy Leuels mat. Bur. Stand. Circ. 467, Washington D.C., 1952), vol. 2. 1975,97, 7054. Chem., 1976,110,247. 321. (PAPER 8/695)