摘要:

J. Chem. Soc., Dalton Trans., 1997, Pages 313–315 313 DALTON Kinetics and mechanism of the oxidation of sulfite by trans- [Ru(tmc)O2]2+ (tmc = 1,4,8,11-tetramethyl-1,4,8,11- tetraazacyclotetradecane) Tai-Chu Lau,* Kwok-Ho Chow, Kent W. C. Lau and Wenny Y. K. Tsang Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon Tong, Hong Kong The kinetics of the oxidation of SO3 22 by trans-[Ru(tmc)O2]2+ (tmc = 1,4,8,11-tetramethyl-1,4,8,11- tetraazacyclotetradecane) has been studied in aqueous solution at 25.0 8C.The reaction has the following stoichiometry: trans-[RuVI(tmc)O2]2+ + SO3 22 + H2O æÆ trans-[RuIV(tmc)O(OH2)]2+ + SO4 22. Sulfite and sulfur-containing products were analysed by ion chromatography. No S2O6 22 could be detected. The rate law is 2d[RuVI]/dt = k/{1 + ([H+]/K)}[RuVI][SIV] with k = (7.0 ± 1.4) × 104 dm3 mol21 s21 and K = (3.4 ± 1.0) × 1027 mol dm23 at I = 1.0 mol dm23. The value of k is more than two orders of magnitude greater than predicted by Marcus theory, and an inner-sphere mechanism involving the intermediate [O]] RuVI]] O]SO3] is proposed.This may decompose by one- or two-electron pathways. Ruthenium forms an extensive series of oxo complexes with oxidation states ranging from IV to VIII.1 These are in general potent oxidants and there have been many reports on their reactions with organic substrates.1,2 We have been studying the reactions of oxoruthenium complexes, especially trans-dioxoruthenium( VI) complexes, with simple inorganic substrates since much less is known in this area.We are particularly interested in trans-[RuVI(tmc)O2]2+ (tmc = 1,4,8,11-tetramethyl- 1,4,8,11-tetraazacyclotetradecane) for the following reasons. The presence of the macrocyclic tertiary amine tmc makes this complex particularly stable, hence complications arising from ligand exchange or decomposition can be eliminated. It is also a mild oxidant (Eo = 0.56 V),3 thus reactions would be likely to occur at convenient rates for most substrates.The kinetics of the reduction of this complex with PPh3,4 Fe2+(aq),5 and I2 6 have been reported. We describe here kinetic studies of its reduction with sulfite. Although the oxidation of sulfite by transition-metal complexes has been extensively studied,7 most of these reactions are either outer sphere involving substitutioninert oxidants, or inner sphere involving labile oxidants. In the present study an additional pathway such as oxygen-atom transfer is possible.Experimental Materials The complex trans-[RuVI(tmc)O2][PF6]2 was prepared by a literature method.8 Sodium sulfite (AR grade) was obtained from BDH and used as received. Ionic strength was maintained with sodium trifluoroacetate. Water for kinetic experiments was distilled twice from alkaline permanganate, and was deaerated with argon for 30 min before use for preparation of solutions. Sodium sulfite solutions were standardized by iodometry.Kinetics Kinetic experiments were performed under pseudo-first-order conditions using either a Hewlett-Packard 8452A diode-array spectrophotometer or an Applied Photophysics DX-17MV stopped-flow spectrophotometer. The progress of the reaction was monitored by measuring absorbance changes at 260 nm (lmax of RuVI). Pseudo-first-order rate constants, kobs, were obtained by non-linear least-squares fits of At vs. time t according to the equation At = A• + (A0 2 A•)exp(2kobst), where A0 and A• are the initial and final absorbances, respectively.Products The SO3 22 and sulfur-containing products were analysed by ion chromatography (IC) with a Wescan ICM 300 ion chromatograph equipped with a 335 suppressor module and an Anion/R column. The mobile phase was 1.8 mmol dm23 Na2CO3 and 1.7 mmol dm23 NaHCO3. In a typical experiment a 1.5 × 1024 mol dm23 solution (2.5 cm3) of [Ru(tmc)O2]2+ in water was added to a 2.4 × 1023 mol dm23 solution (2.5 cm3) of Na2SO3 in acetate buffer (pH 4.52 and I = 0.01 mol dm23) at 25 8C.After a reaction time of 5 min the solution was diluted to 10.0 cm3 with water and analysis by ion chromatography indicated that the final solution contained 5.6 × 1024 mol dm23 SO3 22 and 3.5 × 1025 mol dm23 SO4 22. Thus, 1 mol of SO3 22 reacted with 1 mol of RuVI to produce 1 mol of SO4 22. No S2O6 22 was detected. Results and Discussion Spectral changes and stoichiometry Preliminary repetitive scanning indicated rapid spectral changes when a solution of RuVI was mixed with a solution of SIV- (HSO3 2 + SO3 22).A well defined isosbestic point at 290 nm was maintained throughout the reaction (Fig. 1). The final spectrum showed quantitative formation of trans-[RuIV(tmc)O(OH2)]2+ [lmax/nm(e/dm3 mol21 cm21): 420 (150), 280 (1600) and 210 (9800)].9 Analysis by ion chromatography indicated that 1 mol of RuVI reacted with 1 mol of SO3 22 to produce 1 mol of SO4 22.Thus the stoichiometry of the reaction is as in equation (1). [RuVI(tmc)O2]2+ + SO3 22 + H2O [RuIV(tmc)O(OH2)]2+ + SO4 22 (1) N N N N Me Me Me Me tmc D6/03655D/A1314 J. Chem. Soc., Dalton Trans., 1997, Pages 313–315 Kinetics In the presence of at least 10-fold excess of SIV clean pseudo- first-order kinetics was observed for over three half-lives. The kinetics was followed at 260 nm (lmax of RuVI). The pseudo- first-order rate constants, kobs, were independent of [RuVI] from 2.5 × 1025 to 1 × 1024 mol dm23.Most kinetic runs were done under anaerobic conditions using syringe techniques to transfer degassed solution to the stopped-flow spectrophotometer, since oxidation of SO3 22 by certain metal complexes is known to be highly sensitive to contamination by atmospheric dioxygen.7b,d A few experiments were done by mixing ruthenium(VI) solutions saturated with air with degassed sulfur(IV) solutions in the stopped-flow apparatus. The rate constants were found to be identical to those obtained when both solutions were degassed.The effect of CuII on the reaction was investigated at pH 4.52. When reactions were carried out with solutions containing up to 2.6 × 1024 mol dm23 CuSO4?5H2O the rate constants were similar to those obtained when there was no added CuII. The reaction is first order in [SIV] up to 2000 [RuVI]. Thus the experimentally determined rate law is as in equation (2). 2d[RuVI]/dt = kobs[RuVI] = k2[RuVI][SIV] (2) Representative results are summarized in Table 1; k2 increases with decreasing ionic strength, as expected for a bimolecular reaction between ions of opposite charges; a plot of log k2 versus I� �� /(1 + I� �� ) was linear with slope = 23.6 ± 0.2 and intercept = 4.5 ± 0.1. The kinetics of the oxidation of SIV by RuVI were studied over the range pH 3.26–6.28.A plot of 1/k2 against [H+] gave a straight line (Fig. 2) consistent with the relationship (3).A nonk2 = k/{1 + ([H+]/K)} (3) linear least-squares fit of the data using equation (3) yielded k = (7.0 ± 1.4) × 104 dm3 mol21 s21 and K = (3.4 ± 1.0) × 1027 mol dm23 at 25 8C and I = 1.0 mol dm23. Mechanism The mechanism for the outer-sphere oxidation of sulfite by metal complexes is well established.7a,b A similar outer-sphere Fig. 1 Spectral changes during the reduction of [RuVI(tmc)O2]2+ (ª5 × 1025 mol dm23) by SIV (ª5 × 1024 mol dm23) at pH 3.8, I = 0.5 mol dm23 and T = 298.0 K.Spectra were recorded at 4 s intervals Fig. 2 Plot of 1/k2 vs. [H+] for the reduction of [RuVI(tmc)O2]2+ by SIV at 298.0 K and I = 1.0 mol dm23 Table 1 Representative rate constants at 25.0 8C for the oxidation of sulfite by [Ru(tmc)O2]2+ * I/mol [SIV]/ k2/dm3 dm23 pH mol dm23 kobs/s21 mol21 s21 1.0 3.26 4.60 × 1024 3.19 × 1022 (5.25 ± 0.01) × 10 1.0 3.26 1.59 × 1023 5.32 × 1022 1.0 3.26 2.30 × 1023 1.33 × 1021 1.0 3.26 3.40 × 1023 1.67 × 1021 1.0 3.26 4.60 × 1023 2.65 × 1021 1.0 3.26 9.20 × 1023 4.16 × 1021 1.0 3.26 4.69 × 1022 2.46 1.0 4.08 1.59 × 1023 2.63 × 1021 (2.72 ± 0.01) × 102 1.0 4.08 2.30 × 1023 5.85 × 1021 1.0 4.08 3.40 × 1023 8.12 × 1021 1.× 1023 1.08 1.0 4.08 9.20 × 1022 2.49 × 10 1.0 4.30 5.00 × 1024 1.69 × 1021 (6.40 ± 0.05) × 102 1.0 4.30 1.20 × 1023 5.78 × 1021 1.0 4.30 2.30 × 1023 1.32 1.0 4.30 3.50 × 1023 2.06 1.0 4.30 5.00 × 1023 3.10 1.0 4.30 9.20 × 1023 5.71 1.0 4.52 4.60 × 1024 2.16 × 1021 (8.03 ± 0.10) × 102 1.0 4.52 1.16 × 1023 6.08 × 1021 1.0 4.52 2.30 × 1023 1.29 1.0 4.52 3.40 × 1023 1.94 1.0 4.52 4.60 × 1023 2.93 1.0 4.52 9.20 × 1023 5.44 1.0 4.52 2.76 × 1022 2.14 × 10 1.0 4.52 3.68 × 1022 2.87 × 10 1.0 4.52 4.66 × 1022 3.69 × 10 1.0 4.80 1.16 × 1023 3.02 (1.77 ± 0.06) × 103 1.0 4.80 2.30 × 1023 3.61 1.0 4.80 3.40 × 1023 4.38 1.0 4.80 4.60 × 1023 1.00 × 10 1.0 4.80 9.20 × 1023 1.35 × 10 1.0 4.80 1.84 × 1022 3.07 × 10 1.0 4.80 2.76 × 1022 4.79 × 10 1.0 4.80 3.68 × 1022 6.94 × 10 1.0 4.80 4.60 × 1022 7.66 × 10 1.0 4.88 4.60 × 1024 2.47 × 1021 (1.88 ± 0.04) × 103 1.0 4.88 1.16 × 1023 1.15 1.0 4.88 2.30 × 1023 2.86 1.0 4.88 3.40 × 1023 4.48 1.0 4.88 4.60 × 1023 6.33 1.0 4.88 9.20 × 1023 1.22 × 10 1.0 4.88 4.64 × 1022 8.60 × 10 1.0 5.93 4.60 × 1024 4.59 (1.37 ± 0.03) × 104 1.0 5.93 1.16 × 1023 1.48 × 10 1.0 5.93 2.30 × 1023 3.21 × 10 1.0 5.93 3.40 × 1023 4.43 × 10 1.0 5.93 4.60 × 1023 6.78 × 10 1.0 5.93 9.20 × 1023 1.13 × 102 1.0 5.93 1.84 × 1022 2.59 × 102 1.0 5.93 2.76 × 1022 3.73 × 102 1.0 6.28 4.60 × 1024 7.95 (2.78 ± 0.08) × 104 1.0 6.28 1.16 × 1023 2.89 × 10 1.0 6.28 2.30 × 1023 6.53 × 10 1.0 6.28 3.40 × 1023 1.02 × 102 1.0 6.28 4.60 × 1023 1.31 × 102 1.0 6.28 9.20 × 1023 2.52 × 102 0.10 4.40 4.60 × 1024 2.39 (4.96 ± 0.10) × 103 0.10 4.40 1.15 × 1023 5.96 0.10 4.40 3.45 × 1023 1.77 × 10 0.10 4.40 4.60 × 1023 2.28 × 10 0.08 4.40 4.60 × 1024 2.51 (5.60 ± 0.10) × 103 0.08 4.40 1.15 × 1023 5.76 0.08 4.40 3.45 × 1023 1.87 × 10 0.08 4.40 4.60 × 1023 2.56 × 10 0.036 4.40 4.60 × 1024 3.21 (6.45 ± 0.10) × 103 0.036 4.40 1.15 × 1023 6.42 0.036 4.40 3.45 × 1023 2.11 × 10 0.036 4.40 4.60 × 1023 2.98 × 10 * The pH was maintained with acetate buffer (below 5.5) and phosphate buffer (above 5.5).Ionic strength was adjusted with NaO2CCF3. Each kobs value is the average of at least two determinations. Results were reproducible to within 5%.J. Chem. Soc., Dalton Trans., 1997, Pages 313–315 315 mechanism for the present reaction that is consistent with the experimental results is shown in equations (4)–(9).This HSO3 2 Ka SO3 22 + H+ (4) [O]] RuVI]] O]2+ + SO3 22 k1 k21 [O]] RuV]] O]+ + SO3 2 (5) [O]] RuVI]] O]2+ + SO3 2 k2 [O]] RuV]] O]+ + SO3 (6) [O]] RuV]] O]++SO3 2+2H+ k3 [O]] RuIV]OH2]2++SO3 (7) 2[O]] RuV]] O]+ + 2H+ k4 [O]] RuVI]] O]2+ + [O]] RuIV]OH2]2+ (8) SO3 + H2O k5 2H+ + SO4 22 (9) mechanism leads to the observed rate law with k = k1 and K = Ka under the conditions that the equilibrium Ka is established rapidly and that the k2, k3 and k4 steps compete efficiently with the back electron-transfer step k21.In an acidic medium dioxoruthenium( V) is a stronger oxidant than dioxoruthenium(VI), and would rapidly oxidize SO3 2 (k3 step) or undergo disproportionation 3 (k4 step). The measured value of Ka (3.4 × 1027 mol dm23) is in reasonable agreement with a literature value 10 (1.2 × 1027 mol dm23). If the oxidation of SO3 22 by RuVI is indeed a simple outersphere reaction as shown above then the rate constant k1 should be comparable to the theoretical value k12 obtained by the Marcus cross-relation 11 (neglecting the work term) (10) and the relationship of f12 is given in equation (11).The k12 = (k11k22K12 f12)� �� (10) log f12 = (log K12)2/4log(k11k22/1022) (11) equilibrium constant K12 is calculated from the reduction potentials for the RuVI–RuV (0.56 V12) and the SO3 2–SO3 22 (0.72 V7a) couples. A value of 1 × 105 dm3 mol21 s21 is used for k11, the self-exchange rate for RuVI–RuV;3 k22, the self-exchange rate for SO3 2–SO3 22, is 4 dm3 mol21 s21.7a The value of k12 obtained in this manner is 6.3 × 102 dm3 mol21 s21, lower than the observed one by a factor of more than 102.This suggests that the oxidation of SO3 22 probably goes through an inner-sphere pathway. A one-electron inner-sphere pathway involving co-ordination of SO3 22 to the metal centre is common for substitution-labile metal complexes as oxidants.Since the ruthenium(VI) complex is substitution inert and the bulky macrocyclic ligand would prevent formation of a seven-co-ordinate complex, this pathway is highly unlikely. Moreover, S2O6 22, which is usually produced in such pathways, is not detected in our system. A more reasonable inner-sphere pathway involves the intermediate [O]] RuVI]] O]SO3], in which the S atom in SO3 22 is bonded to an oxygen atom of a ruthenium–oxo bond.This intermediate may undergo subsequent decomposition in two ways, namely by one-electron transfer (12) or by oxygen-atom transfer, which is [O]] RuVI]] O]SO3] [O]] RuV]] O]+ + SO3 2 (12) a two-electron process (13). The inner-sphere one-electron [O]] RuVI]] O]SO3] [O]] RuIV]OSO3] (13) pathway would have a rate law similar to that of the outersphere pathway. The mechanism for the proposed oxygen-atom transfer pathway is shown in Scheme 1. By applying the steadystate approximation to [O]] RuIV]OSO3], the rate law (16) is 2d[RuVI]/dt = Kbk6k7[RuVI][SIV]/(k26 + k7) × {1 + ([H+]/Ka)} (16) HSO3 2 Ka H+ + SO3 22 (4) [O]] RuVI]] O]2+ + SO3 22 Kb [O]] RuVI]] O]SO3] (14) [O]] RuVI]] O]SO3] k6 k26 [O]] RuIV]OSO3] (13) [O]] RuIV]OSO3] + H2O k7 [O]] RuIV]OH2]2+ + SO4 22 (15) Scheme 1 obtained.This agrees with the observed rate law with k = Kbk6k7/(k26 + k7) and K = Ka. Although our kinetic results cannot distinguish between the two inner-sphere pathways, we tend to favour the oxygen-atomtransfer mechanism for the following reasons.First the kinetics is not very sensitive to the presence of O2, suggesting that the SO3 2 radical may not be an intermediate. Secondly the oxygenatom- transfer pathway is rather common for oxoruthenium complexes, e.g oxidation of PPh3 4 and olefins.13 It has also been proposed for the oxidation of I2.6 Similar intermediates have also been proposed for the oxidations of SIV by other oxometal species. In the oxidation by CrVI in acidic solution [O3CrOSO3]22 was proposed to be an intermediate.7g In a detailed study of the oxidation of sulfite by MnO4 2 in alkaline solution the initial step was suggested to be outer-sphere reduction of MnVII to MnVI. However, in a more recent and similar study, the intermediate [O3MnOSO3]2 was proposed, which may decompose by one- or two-electron mechanisms.7e The rate constants obtained in the two studies were comparable, but much higher than predicted by Marcus theory,7a so the inner-sphere mechanism seems to be more reasonable. Acknowledgements Financial support from the Hong Kong Research Grants Council and the City University of Hong Kong is gratefully acknowledged.References 1 C. M. Che and V. W. W. Yam, Adv. Inorg. Chem. Radiochem., 1992, 39, 233. 2 W. P. Griffith, Transition Met. Chem., 1990, 15, 251; Chem. Soc. Rev., 1992, 21, 179. 3 C. M. Che, K. Lau and T. C. Lau, J. Am. Chem. Soc., 1990, 112, 5176. 4 K. Y. Wong and C. M. Che, J. Chem. Soc., Dalton Trans., 1989, 2056. 5 T. C. Lau, K. W. C. Lau and C. K. Lo, Inorg. Chim. Acta, 1993, 209, 89. 6 T. C. Lau, K. W. C. Lau and K. Lau, J. Chem. Soc., Dalton Trans., 1994, 3091. 7 (a) R. Sarala, M. A. Islam, S. B. Rabin and D. M. Stanbury, Inorg. Chem., 1990, 29, 1133; (b) R. Sarala and D. M. Stanbury, Inorg. Chem., 1990, 29, 3456; (c) L. I. Simandi, M. Jaky, C. R. Savage and Z. A. Schelly, J.em. Soc., 1985, 107, 4220; (d) J. M. Anast and D. W. Margerum, Inorg. Chem., 1981, 20, 2319; (e) T. Ernst, M. Cyfert and M. Wilgocki, Int. J. Chem. Kinet., 1992, 24, 903; ( f ) E. S. Gould, Acc. Chem. Res., 1986, 19, 66; (g) G. P. Haight, jun., E. Perchonock, F. Emmenegger and G. Gordon, J. Am. Chem. Soc., 1965, 87, 3835. 8 C. M. Che, K. Y. Wong and C. K. Poon, Inorg. Chem., 1985, 24, 1797. 9 C. M. Che, T. F. Lai and K. Y. Wong, Inorg. Chem., 1987, 26, 2289. 10 R. M. Smith and A. E. Martell, Critical Stability Constants, Plenum, New York, 1989, vol. 6. 11 R. A. Marcus, Annu. Rev. Phys. Chem., 1964, 15, 155. 12 C. M. Che, K. Y. Wong and F. C. Anson, J. Electroanal. Chem. Interfacial Electrochem., 1987, 226, 211. 13 C. M. Che, W. H. Leung, C. K. Li and C. K. Poon, J. Chem. Soc., Dalton Trans., 1991, 379. Received 28th May 1996; Paper 6/03655D

ISSN:1477-9226    

DOI:10.1039/a603655d

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 317–322 317 Formation constants for complexes of 1,4,7,10-tetraazacyclododecane-1,7- diacetic acid and the crystal structure of its nickel(II) complex† Jennifer M. Weeks, Max R. Taylor and Kevin P. Wainwright * Department of Chemistry, The Flinders University of South Australia, GPO Box 2100, Adelaide, South Australia 5001, Australia The crystal structure of the nickel(II) complex of 1,4,7,10-tetraazacyclododecane-1,7-diacetic acid (H2doda) has been determined.The ligand binds in a cis-octahedral fashion through the four nitrogen atoms and two carboxylate groups and dimerizes via an apparently symmetrical hydrogen bond from the non-bonding oxygen of a carboxylate group to a lone hydrogen ion originating from a perchloric acid molecule of crystallization. Carbon- 13 NMR studies of the corresponding zinc(II) complex indicated that it has a similar structure. pKa Measurements for doda, at 298.2 K in 0.1 mol dm23 NEt4ClO4, suggested that the two secondary amines are the first to protonate followed by the two carboxylates and then the two tertiary nitrogen atoms.A marked disparity in the magnitude of the pKa values for the two carboxylic groups [4.00(1) and 2.36(3)] suggests that internal hydrogen bonding between the two stabilizes one of the protons. Formation constants for some divalent metal complexes (Co, Ni, Cu, Zn, Cd or Pb) of H2doda have been determined. The results from this work are compared to data for analogous macrocycles having three or four carboxymethyl groups.A convenient approach to synthesizing the highly stable, neutral metal complexes that are frequently sought for applications in diagnostic and therapeutic medicine has been to use the high stability conferred by a polyaza macrocycle, in conjunction with the metal-ion charge neutralization provided by a group of pendant carboxylates equal in number to the charge on the metal ion.1–3 Thus, neutral complexes of divalent metal ions with 1,4,8,11-tetraazacyclotetradecane (cyclam) have been formed by appending two carboxymethyl (acetate) groups to two of the nitrogen atoms.2,3 By selectively appending these carboxymethyl groups on the diagonally related 1,8-nitrogen atoms, advantage can be taken of the fact that when cyclam binds in its most stable trans-III configuration one carboxylate will project below the macrocyclic plane and the other above.This allows the complex to assume octahedral stereochemistry without inducing undue ring strain in any of the chelate rings, and in so doing probably maximizes the stability that can be achieved for many metal ions with this particular macrocycle– pendant arm combination.This approach to generating neutral complexes also has the advantage that the lipophilicity of the complex can be further enhanced by alkylation of any secondary amines that are not carboxymethylated.3 Synthetic strategies for selectively difunctionalizing the smaller tetraazamacrocycle 1,4,7,10-tetraazacyclododecane (cyclen) have been somewhat limited and thus the possibility of preparing neutral complexes of divalent metal ions, through the attachment of two carboxymethyl groups to two of the nitrogen atoms of the macrocycle, has remained largely unexplored.This is significant as it frequently turns out that complexes originating from cyclen-derived ligands are more stable than their cyclam-derived counterparts, especially when the divalent metal ion is one of large ionic radius.4,5 Recently, however, van Westrenen and Sherry 6 have demonstrated that selective sulfomethylation at the 1,7 positions of cyclen can be accomplished in quantitative yield, opening a viable pathway to these compounds.Nitrogen atoms at positions 4 and 10 are prevented from reacting by allowing the synthesis to proceed under conditions in which they are protonated, namely in aqueous solution at pH 7.The sulfonate moieties can be replaced by cyano moi- † Non-SI unit employed: mB ª 9.27 × 10224 J T21. eties which are then hydrolysed in situ giving the amino acid dihydrate in ca. 60% isolated yield over the three steps summarized in Scheme 1. In the work reported here we have used this ligand synthesis to prepare 1,4,7,10-tetraazacyclododecane-1,7-diacetic acid (H2doda) and have made a study of its complexing properties. Its protonation constants, formation constants for a range of divalent metal complexes and the crystal structure of the nickel(II) complex are reported, and compared with the corresponding data for 1,4,7,10-tetraazacyclododecane-1,4,7,10- tetraacetic acid (H4dota), 1,4,7,10-tetraazacyclododecane- 1,4,7-triacetic acid (H3dotra) and their complexes.We also investigated the consequences of exposing H2doda to conditions which favour conversion into the diacid chloride (concentrated HCl or thionyl chloride). With the cyclam analogue this leads to spontaneous internal lactamization to form a tricyclic compound,2,3 however in the case of H2doda no corresponding reaction resulted, probably due to the greater steric rigidity of the smaller macrocycle.Scheme 1 (i) Water, pH 7; (ii) 2NaCN; (iii) concentrated HCl, 65 h318 J. Chem. Soc., Dalton Trans., 1997, Pages 317–322 Results and Discussion Macrocycle protonation constants The stepwise protonation constants for doda were measured by potentiometric titration of H4doda2+ with tetraethylammonium hydroxide using 0.1 mol dm23 tetraethylammonium perchlorate as the background electrolyte.Values for the protonation constants, Kx (x = 1–6), relate to the equilibria shown in equations (1)–(6), where log10 Kx = pKax. doda22 + H+ K1 Hdoda2 (1) Hdoda2 + H+ K2 H2doda (2) H2doda + H+ K3 H3doda+ (3) H3doda+ + H+ K4 H4doda2+ (4) H4doda2+ + H+ K5 H5doda3+ (5) H5doda3+ + H+ K6 H6doda4+ (6) Previous work has shown that polyaminocarboxylates have the ability to bind sodium and, to a lesser extent, potassium ions, resulting in a reduction of the perceived protonation constants when either of these ions is present during the potentiometric titration.7 It is for this reason that tetraethylammonium was used as the background cation in this work.For K1 to K4 the logarithms of the protonation constants were found to be 11.45(2), 9.54(3), 4.00(1) and 2.36(3). The log K values for the final two protonations were not measured and are presumably below 2.3 since the species H5doda3+ and H6doda4+ were not detected within the pH range of the potentiometric titration (2.3–11.2).These values are shown in Table 1 where they are compared to the corresponding values for dota, dotra and cyclen which have been measured by the same method, at the same ionic strength and in a non-co-ordinating electrolyte. Comparison of these sets of protonation constants shows that there is a progressive trend towards higher basicity at the first protonation site as the number of carboxylate groups increases.This is understandable, as increasing the number of carboxylates increases the overall charge on the macrocyclic anion undergoing protonation. With doda and dotra the question arises as to whether the first protonation occurs on a secondary or a tertiary nitrogen. The crystal structure of H2dotra has been determined previously 8 and shows, as has been postulated on numerous previous occasions, that it is a trans-related pair of nitrogen atoms that accepts the first two protons. Of the two possible trans pairs it is the secondary–tertiary rather than the tertiary–tertiary pair that protonate, probably indicating that even though the higher than normal basicity of the molecule arises from the carboxylate groups that are attached to the tertiary nitrogen atoms, it is still the secondary amine that is the more basic.Originally it has been suggested that in dota the high basicity originated from an intrapendant-arm hydrogenbond stabilization of a proton that had added to a tertiary nitrogen,12 however the crystal structure of H2dotra shows no evidence for this and it now seems more likely that the high basicity is simply a function of overall charge and that, where the situation arises, the normal sequence of basicity observed in macrocyclic amines (with secondary amines being more basic than tertiary amines) is preserved. Thus, we suggest here that the first protonation on doda22 occurs at one of the secondary amines, and, since the second protonation will almost certainly occur trans to the first, the second protonation will also occur at a secondary nitrogen.This is supported by log K2 for doda being higher than the value for either dotra or dota, where it can only be ascribed to protonation at a tertiary nitrogen. It is not clear to us, however, why log K2 for doda should be lower than that for cyclen where the proton would be adding to a cation rather than an anion.The value of log K3 for doda must correspond to protonation of the first of the carboxylate groups and is similar to the corresponding dota and dotra values. It is interesting that in doda, dotra and dota the first carboxylate site (first and second in dota) is markedly more basic than is normally the case for a-amino acids, where this protonation is typically characterized by a log K value of between about 2.0 and 2.5 (2.35 for glycine 16), or for a single carboxymethyl group pendant to a tetraazamacrocycle, where a log K of 3.01 has been measured.17 A possible explanation may be that the first protonated carboxylate group in these polycarboxymethyl macrocyclic molecules undergoes stabilization through intramolecular hydrogen bonding to the trans-related carboxylate, as shown in Scheme 2.Alternative hydrogen-bonding interactions involving one of the amine moieties do not account for the fact that the second carboxylate in doda does not show a similar high basicity, but instead has a more normal log K of 2.36, similar to log K5 and log K6 of dota which probably also describe protonation of the two carboxylates initially involved in the stabilization of their trans-carboxylic partners.The speciation of doda at different pH is shown in Fig. 1. Metal complex formation constants Glass-electrode potentiometric titration of H4doda2+ with NEt4OH in the presence of an M2+ (M = Co, Ni, Cu, Zn, Cd or Pb) cation revealed the existence of two complex species, [M(Hdoda)]+ and [M(doda)], for all M.These results indicate that doda can co-ordinate, giving complexes of significant stability, either as a penta- or as a hexa-dentate ligand, although there is no way of being certain, from these data, whether it is an amine or a carboxylate that is protonated in pentadentate Hdoda2. This observation parallels findings with dota where [M(dota)]22, [M(Hdota)]2 and [M(H2dota)] complexes were observed for all of the metal ions used in this study (except Pb2+, where only the first two species were observed) indicating that complete ligand deprotonation is not essential for stable complex formation.9 In contrast to dota,9,18 no 2 : 1 metal : ligand species were observed, which is consistent with the lower number of donor atoms available in doda.The formation constants for the observed species are given in Table 2, together with data for the cyclen and dota complexes for comparison.The values of the formation constants tend to be intermediate between those measured for cyclen and dota and no particular metal-ion selectivity is seen within the group of metal ions studied here. The cadmium(II) complex shows the greatest stability enhancement (103.9) compared to the corresponding cyclen complex, which is consistent with the observation made by Hancock and co-workers 22 that metal ions with an ionic radius of about 1 Å derive the greatest stability enhancement from attaching pendant acetates to cyclen.Scheme 2 Proposed stabilization of a carboxylate proton in macrocycles having more than one pendant acetate, which may account for the pairwise disparity in the pKa values for the carboxylic groups. The pKa values shown are those for dodaJ. Chem. Soc., Dalton Trans., 1997, Pages 317–322 319 Table 1 Protonation constants a at 298.2 K for doda, determined in this study, together with literature values for dota, dotra and cyclen Ligand log10 K1 log10 K2 log10 K3 log10 K4 log10 K5 log10 K6 dodab dotra c dota c dota d cyclen e cyclen f 11.45(2) 11.59 11.73 11.08 11.32 11.04 9.54(3) 9.24 9.40 9.23 9.72 9.86 4.00(1) 4.43 4.50 4.24 <2.3 <2 2.36(3) 3.48 4.19 4.18 <2.3 <2 <2.3 1.88 <2.3 1.71 a K1, K2, etc.refer to the equations Lx2 + H+ HL(x21)2, HL(x21)2 + H+ H2L(x22)2, etc. where x = 2, 3, 4 and 0 for doda, dotra, dota and cyclen, respectively.b I = 0.1 mol dm23 NEt4ClO4. c From ref. 8, I = 0.1 mol dm23 NMe4Cl, earlier determinations for dota may be found in refs. 7 and 9–12. d From ref. 12, I = 1 mol dm23 NaCl, corrected for sodium complexation. e From ref. 5, I = 0.1 mol dm23 NaNO3. f From ref. 13, I = 0.1 mol dm23 NaClO4; earlier values may be found in refs. 14 and 15. Table 2 Formation constants at 298.2 K for doda,a dota b and cyclen complexes of some divalent metal ions log10 b Overall reaction Ligand Co2+ Ni2+ Cu2+ Zn2+ Cd2+ Pb2+ M + L + 2H+ M(H2L)2+ M + L + H+ M(HL)+ M + L ML doda dota doda dota cyclen doda dota c 27.73 21.9(1) 24.35 13.8 e 16.9(1) 20.27 c 28.26 >19.3 d 23.54 16.4 f >13.3 d 20.03 c 29.80 24.1(1) 26.03 23.3 g 21.1(2) 22.25 c 28.79 22.2(1) 25.28 16.2 h 18.2(1) 21.10 c 28.73 21.7(1) 25.70 14.3 h 18.2(1) 21.31 cc 21.9(1) 26.55 15.9 h 18.3(1) 22.69 Derived reactions M + H2L2+ M(H2L)2+ M + HL+ M(HL)+ doda dota doda dota c 5.88 10.5(1) 12.26 c 6.41 >7.9 d 11.45 c 7.95 12.7(1) 13.94 c 6.94 10.8(1) 13.19 c 6.88 10.3(1) 13.61 cc 10.5(1) 14.46 a This work, I = 0.1 mol dm23 NEt4ClO4.A value of 13.8 was obtained for the apparent pKw under these conditions. b From ref. 9, I = 0.1 mol dm23 NMe4NO3; earlier values may be found in refs. 7 and 10. c Reaction not observed. d Only the lower limit is given owing to very slow and hence doubtful equilibration during the titration. e From ref. 19. f From ref. 20. g From ref. 21. h From ref. 14. Fig. 1 Speciation diagram for doda produced using data from Table 1320 J. Chem. Soc., Dalton Trans., 1997, Pages 317–322 Equilibration of the titration solution following the addition of each aliquot of base was judged to be complete when the rate of change of potential fell below 0.1 mV min21. This occurred within a period of 5 min for all metal ions studied with doda except nickel(II) and is similar to the rate of equilibrium noted previously for dota.9 Structure of the doda complexes To provide some information about the co-ordination mode of H2doda in the solid state the nickel(II) complex was isolated and the crystal structure solved.The complex was synthesized by adding an ethanolic solution of hydrated nickel(II) perchlorate to a refluxing aqueous solution of H2doda?2H2O, in a 1 : 1 molar ratio. The resulting solution precipitated purple crystals on cooling. These were recrystallized, from a slowly evaporating aqueous solution, giving crystals, suitable for X-ray diffraction studies, that analysed as [Ni(doda)]? 0.5HClO4?1.5H2O.The X-ray diffraction data revealed that the complex is dimeric, as shown in Fig. 2, with each half-unit symmetrically hydrogen bonded to the other through the interaction of a nonco- ordinating carboxylate oxygen with a lone hydrogen ion. The hydrogen ion is located on an approximate centre of symmetry within the dimer. The remainder of the asymmetric unit is composed of a perchlorate anion and three water molecules.The water molecule oxygen atoms are separated by distances that indicate possible hydrogen bonding between them [range 2.742(6)–2.880(7) Å] and they appear to be hydrogen bonded to the perchlorate ion. The only possible hydrogen bond to the dimer is from water O(2) to O(15b) [2.815(4) Å]. The individual half-units have approximate C2v symmetry. Each consists of an octahedrally co-ordinated nickel(II) ion about which the macrocycle has folded, along the N(4)]N(10) axis, so that it binds through the four nitrogen atoms and two carboxylate groups, which occupy cis-related co-ordination sites, as shown in Fig. 3. The equatorial N]Ni]N angles average 109.48, while the axial N]Ni]N angles average 155.48. There are no significant distortions in the Ni]N bond lengths which range from 2.113 to 2.144 Å. A selected listing of bond lengths and angles is provided in Fig. 2 Structure of the [Ni(doda)] dimer Fig. 3 An ORTEP23 drawing of one half-unit (a) of the [Ni(doda)] dimer showing displacement ellipsoids of the non-hydrogen atoms at the 50% probability level Table 3. Both NH groups and both carboxymethyl groups project in the same direction, thus, using the terminology of Bosnich et al.,24 the macrocycle has the cis-I configuration. Each half-unit has a similar arrangement of atoms to those in [Ni- (H2dota)] and [Cu(H2dota)], which have been analysed previously, 25 and in which the two carboxylate groups not present in doda are protonated and directed away from the inner coordination sphere of the metal ion.The hydrogen atom, H(1), lies on the line of centres between O(19a) and O(19b) (Fig. 2) at distances of 1.21(4) and 1.24(4) Å from the O atoms, respectively. The apparently symmetrical hydrogen bond O(19a) ? ? ? H(1) ? ? ? O(19b), responsible for the dimeric [Ni- (doda)] structure, has an O? ? ? O separation of 2.448(4) Å which is at the short end of the known spectrum of hydrogen bond lengths. A search of the Cambridge Crystallographic Data Base 26 showed that this pattern of hydrogen bonding between co-ordinated carboxylates is not uncommon, but revealed only ten structures containing an O? ? ? O hydrogen-bonded distance (where the two oxygen atoms were specified as the non-metalco- ordinating carboxylate oxygen atoms of a co-ordinated carboxylate) of less than 2.5 Å.27 The zinc(II) complex of H2doda was prepared and purified in the same way as the nickel(II) complex, described above, but analysed as [Zn(doda)]?1.5HClO4?2H2O, suggesting that this complex also hydrogen bonds in the solid state through hydrogen ion incorporation into the crystal lattice, but in a different manner to that seen with [Ni(doda)].The 13C NMR spectrum of [Zn(doda)] in D2O is superficially similar to the spectrum of uncomplexed H2doda, with a pattern of resonances (detailed in the Experimental section) consistent with C2v symmetry. Thus it appears that the zinc(II) complex is cisoctahedral in solution, at least on a time-averaged basis, with doda bound in the same fashion as seen for [Ni(doda)] in the solid state.Experimental General Carbon-13 NMR spectra were recorded at 75.46 MHz, using a Varian Gemini 300 spectrometer, in D2O at ambient temperature. Chemical shifts are quoted with respect to internal 1,4- dioxane, for which the resonance position was taken as d 67.00.Elemental analyses were performed by Chemical and Microanalytical Services Pty Ltd. The compound cyclen was prepared according to the procedure described by Richman and Atkins.28 Dowex 50W-X8 (H) was supplied by BDH. Syntheses 1,4,7,10-Tetraazacyclododecane-1,7-diacetic acid dihydrate, H2doda?2H2O. The crude compound was formed according to the procedure used by van Westrenen and Sherry,6 summarized in Scheme 1. Final purification of the hydrolysed nitrile was effected in the following way.(At the conclusion of the hydrolysis reaction the product is contaminated with sodium Table 3 Selected bond lengths (Å) and angles (8) for the [Ni(doda)] dimer Half-unit a Half-unit b Ni(1)]N(1) Ni(1)]N(4) Ni(1)]N(7) Ni(1)]N(10) Ni(1)]O(16) Ni(1)]O(20) 2.113(3) 2.144(3) 2.135(3) 2.137(3) 2.019(2) 2.068(2) 2.115(3) 2.132(3) 2.144(3) 2.138(3) 2.031(2) 2.061(2) N(4)]Ni(1)]N(10) N(1)]Ni(1)]N(7) O(16)]Ni(1)]O(20) 155.0(1) 109.3(1) 86.4(1) 155.8(1) 109.5(1) 86.6(1)J. Chem.Soc., Dalton Trans., 1997, Pages 317–322 321 chloride, which arises from the sodium sulfite produced during the formation of the dinitrile. The sodium sulfite is not removed as the combination of stages two and three of the synthesis gives for a better overall yield of the dicarboxylic acid.) The crude hydrolysis product, obtained originally from 3 mmol of cyclen, was dissolved in water (9 cm3) and the resulting solution divided into three aliquots. Each aliquot was separately applied to a column of Dowex 50W-X8 (particle size 0.39–1.00 mm) cationexchange resin (bed volume 20 cm3) in its H+ form.The column was washed with 0.5 mol dm23 HCl (200 cm3) to displace sodium cations and then with water (250 cm3) until the eluent was neutral. Elution of the product was then commenced using 0.5 mol dm23 aqueous ammonia solution. The product began to leave the column following a throughput of 150 cm3 of ammonia solution, at which stage the pH of the eluent had risen to 10.Collection of the product was complete after the passage of more ammonia solution (175 cm3) by which stage the pH had risen to 11.5. The fraction containing the product was rotary evaporated to dryness and the residue triturated with ethanol (5 cm3) giving an off-white powder. This was filtered off, washed with diethyl ether (5 cm3) and dried in vacuo. The combined yield from the three aliquots was 0.56 g, 57% based on cyclen (Found: C, 44.45; H, 8.6; N, 17.05.C12H24N4O4?2H2O requires C, 44.45; H, 8.7; N, 17.25%); 13C NMR (D2O) d 179.5 (2C), 57.8 (2C), 50.6 (4C) and 43.9 (4C). [Ni(doda)]?0.5HClO4?1.5H2O. CAUTION: perchlorate salts of metal complexes are potentially explosive. Although we have had no incidents with this or the following perchlorate salt suitable precautions should be taken. A solution of H2doda?2H2O (0.1 g, 0.31 mmol) in water (3 cm3) was brought to reflux and a solution of hexaaquanickel(II) perchlorate (0.11 g, 0.312 mmol) in ethanol (5 cm3) was added dropwise over 10 min.Upon cooling the resulting solution to room temperature fine purple crystals precipitated. These were filtered off and recrystallized from water giving the pure product (0.87 g, 68%) (Found: C, 33.85; H, 5.75; N, 12.95. C24H51ClN8Ni2O15 requires C, 34.15; H, 6.1; N, 13.25%); meff = 3.1 mB. [Zn(doda)]?1.5HClO4?2H2O. This complex was prepared in 71% yield using the same method as described above for the nickel(II) complex (Found: C, 26.8; H, 4.8; N, 10.25.C12H27.5Cl1.5N4O12Zn requires C, 26.75; H, 5.15; N, 10.4%); 13C NMR (D2O) d 179.4 (2C), 57.9 (2C), 53.8 (4C) and 44.7 (4C). Potentiometric titrations The potentiometric titrations were carried out under an inert atmosphere of water-saturated nitrogen in a water-jacketed vessel maintained at 25 8C. Data were obtained from aliquots (10 cm3) of solution containing 0.005 mol dm23 HClO4, 0.100 mol dm23 NEt4ClO4 and approximately 1.0 × 1023 mol dm23 macrocycle titrated with 0.10 mol dm23 NEt4OH.A Metrohm E665 Dosimat autoburette equipped with a 5 cm3 burette was used to deliver the titrant and the potential measured by an Orion Ross Sure Flow 81-65BN combination electrode connected to an Orion 290A pH meter. The autoburette and pH meter were interfaced to an IBM-compatible personal computer which controlled the addition of titrant using a program written by Drs. A. P. Arnold and P. A. Duckworth so that successive additions of titrant caused a decrease of ca. 4 mV in the potential reading. The electrode was calibrated by a titration in the absence of macrocycle and fitting the resulting data from this strong acid– strong base titration by use of the Nernst equation to find correct values for E0 and pKw. The pKa and stability constants were determined using the program SUPERQUAD.29 Stability constant data were gathered from solutions to which 0.1 mol dm23 metal perchlorate solution was added so as to give a metal-toligand ratio in the range 0.5 : 1 to 2 : 1.At least three titrations, with different ratios, were performed for each metal ion. Crystallography Unit-cell and intensity data for [Ni(doda)]?0.5HClO4?1.5H2O were measured on a Rigaku AFC7/R diffractometer using graphite-monochromated Cu-Ka X-radiation. Parameters associated with unit-cell dimensions, intensity data collection, structure solution and refinement are given in Table 4. Absorption corrections were calculated by Gaussian integration.Computer programs of the XTAL system 30 were used throughout the structure solution and refinement. The structure was solved by direct methods. Non-hydrogen atomic coordinates and anisotropic displacement parameters for all atoms were refined by full-matrix least squares on F2 minimizing ow(|Fo|2 2 |Fc|2)2, where w = 1/s2(Fo 2). Values of 1/s2(Fo 2) were obtained from counting statistics. Neutral atom scattering factors with anomalous dispersion corrections were used.An extinction correction was refined [g = 635(67)].31 The perchlorate anion showed high thermal motion and disorder. It was modelled by placing two differently oriented rigid groups on the site and refining individual anisotropic displacement parameters for all atoms and a group population parameter. The highest shift/error values at the end of the refinement were associated with the eulerian angles of the rigid groups. With the exception of the hydrogen atoms on the water molecules and the bridging hydrogen atom, hydrogen atoms were placed at calculated positions and although their coordinates and isotropic displacement parameters were not refined the coordinates were readjusted several times during the refinement.The water molecule hydrogen atoms were not included in the calculations. The position of the bridging hydrogen atom was located in a difference map and its positional and isotropic displacement parameters refined.There is non-crystallographic symmetry present in the structure as revealed by the program BUNYIP.32 The half-units a and b are related by an approximate inversion centre at 0.239(4), 0.749(5), 0.2522(4) and the Cl atoms of the perchlorate groups appear to be similarly related about this point and symmetrically equivalent points. The water molecule O atoms, however, are not. The assignment of the crystal system and Table 4 Crystal data and refinement summary for the [Ni(doda)] dimer Empirical formula M Crystal system Space group a/Å b/Å c/Å b/8 U/Å3 l/Å Z Dm a/g cm23 Dc/g cm23 F(000) T/K m/mm21 Tmin, Tmax qmax/8 h, k, l ranges No.reflections sampled No. reflections measured No. used in refinement No. parameters refined Rb wR(F2) Goodness of fit Final shift/error (maximum, average) Minimum, maximum Dr/e Å23 (C12H22N4NiO4)2?HClO4?3H2O 844.50 Monoclinic P21/c 16.815(8) 8.494(1) 24.333(2) 98.81(2) 3434(2) 1.541 80 4 1.61(2) 1.633 1776 293(2) 0.278 0.475, 0.896 60 0–18, 0–9, 227 to 26 8646 (w–2q scans) 5117 4904 (F2> 0) 477 0.051 0.113 4.18 0.041, 0.004 20.56, 0.52 a By flotation in a mixture of chloroform and dibromoethane. b R = o(||Fo| 2 |Fc||)/o|Fo| based on 4904 reflections.322 J.Chem. Soc., Dalton Trans., 1997, Pages 317–322 space group was initially made by inspection of precession photographs for Laue symmetry and systematic absences. The program CREDUC33 did not reveal a unit cell other than the metrically monoclinic one reported here.It is also noted that if there were an exact inversion centre at 0.25, 0.75, 0.25 the cell would be A-centred. Although the reflections with k + l odd are weak their average intensity is substantial and some are quite intense. We conclude that the extra symmetry is indeed noncrystallographic. Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC).See Instructions for Authors, J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/313. Acknowledgements The authors thank Dr. T. W. Hambley of the University of Sydney who collected the X-ray data and Ms. C. L. Price for technical assistance. References 1 D. Dischino, E. J. Delaney, J. E. Emswiler, G. T. Gaughan, J. S. Prasad, S. K. Srivistava and M. F. Tweedle, Inorg. Chem., 1991, 30, 1265. 2 I. M. Helps, D. Parker, J. Chapman and G. Ferguson, J. Chem. Soc., Chem. Commun., 1988, 1094. 3 J. Chapman, G. Ferguson, J. F. Gallagher, M. C. Jennings and D. Parker, J. Chem. Soc., Dalton Trans., 1992, 345. 4 R. D. Hancock, Acc. Chem. Res., 1990, 23, 253. 5 M. L. Turonek, P. A. Duckworth, G. S. Laurence, S. F. Lincoln and K. P. Wainwright, Inorg. Chim. Acta, 1995, 230, 51. 6 J. van Westrenen and A. D. Sherry, Bioconjugate Chem., 1992, 3, 524. 7 R. Delgado and J. J. R. Fraústo da Silva, Talanta, 1982, 29, 815. 8 K. Kumar, C. A. Chang, L. C. Francesconi, D. D. Dischino, M. F. Malley, J. Z. Gougoutas and M. F. Tweedle, Inorg. Chem., 1994, 33, 3567. 9 S. Chaves, R. Delgado and J. J. R. Fraústo da Silva, Talanta, 1992, 39, 249. 10 E. T. Clarke and A. E. Martell, Inorg. Chim. Acta, 1991, 190, 27. 11 H. Stetter and W. Frank, Angew. Chem., Int. Ed. Engl., 1976, 15, 686. 12 J. F. Desreux, E. Merciny and M. F. Loncin, Inorg. Chem., 1981, 20, 987. 13 T. Koike, S. Kajitani, I. Nakamura, E. Kimura and M. Shiro, J. Am. Chem. Soc., 1995, 117, 1210. 14 M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., 1977, 2269. 15 A. P. Leugger, L. Hertli and T. A. Kaden, Helv. Chim. Acta, 1978, 61, 2296. 16 B. B. Owen, J. Am. Chem. Soc., 1934, 56, 24. 17 M. Studer and T. A. Kaden, Helv. Chim. Acta, 1986, 69, 2081. 18 A. Riessen, M. Zehnder and T. A. Kaden, Helv. Chim. Acta, 1986, 69, 2074. 19 M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., 1980, 327. 20 V. J. Thöm and R. D. Hancock, J. Chem. Soc., Dalton Trans., 1985, 1877. 21 V. J. Thöm, G. D. Hosken and R. D. Hancock, Inorg. Chem., 1985, 24, 3378. 22 H. Maumela, R. D. Hancock, L. Carlton, J. H. Reibenspies and K. P. Wainwright, J. Am. Chem. Soc., 1995, 117, 6698. 23 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 24 B. Bosnich, C. K. Poon and M. L. Tobe, Inorg. Chem., 1965, 4, 1102. 25 A. Riessen, M. Zehnder and T. A. Kaden, Helv. Chim. Acta, 1986, 69, 2067. 26 F. H. Allen, J. E. Davies, J. J. Galloy, O. Johnson, O. Kennard, C. F. McCrae, E. M. Mitchell, G. F. Mitchell, J. M. Smith and D. G. Watson, J. Chem. Inf. Comput. Sci., 1991, 31, 187. 27 See, for example, J. Roziere and C. Belin, Acta Crystallogr., Sect. B, 1979, 35, 2037. 28 J. E. Richman and T. J. Atkins, J. Am. Chem. Soc., 1974, 96, 2264. 29 P. Gans, A. Sabatini and A. Vacca, J. Chem. Soc., Dalton Trans., 1985, 1195. 30 S. R. Hall, G. S. D. King and J. M. Stewart (Editors), XTAL3.4 User’s Manual, University of Western Australia, Lamb, Perth, 1995. 31 W. H. Zachariasen, Acta Crystallogr., Sect. A, 1967, 23, 558. 32 J. R. Hester and S. R. Hall, BUNYIP XTAL3.4 User’s Manual, eds. S. R. Hall, G. S. D. King and J. M. Stewart, University of Western Australia, Lamb, Perth, 1995. 33 Y. Le Page and H. D. Flack, CREDUC XTAL3.4 User’s Manual, eds. S. R. Hall, G. S. D. King and J. M. Stewart, University of Western Australia, Lamb, Perth, 1995. Received 25th July 1996; Paper 6/05208H

ISSN:1477-9226    

DOI:10.1039/a605208h

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 321–322 321 Cluster growth reactions with selenido–carbonyl clusters. Synthesis and structural characterization of [M2Ru2(Ï4-Se)2(Ï-CO)4(CO)6(PPh3)2] (M 5 Mo or W)‡ Daniele Cauzzi, Claudia GraiV, Chiara Massera, Giovanni Mori, Giovanni Predieri *,† and Antonio Tiripicchio Dipartimento di Chimica Generale ed Inorganica, Chimica Analitica, Chimica Fisica, Centro di Studio per la Strutturistica Diffrattometrica del CNR, Università di Parma, Viale delle Scienze, I-43100 Parma, Italy The open-triangular, nido cluster [Ru3(m3-Se)2(CO)7(PPh3)2] (50 electrons, seven skeletal electron pairs, s.e.p.s) reacted at room temperature with [M(CO)3(MeCN)3] (M = Mo or W) to give unexpectedly the bicapped, square-planar clusters [M2Ru2(m4-Se)2(m-CO)4(CO)6(PPh3)2] as unique products; they exhibit an electron-deficient closo geometry with 60 electrons and six s.e.p.s.The chemistry of chalcogenido–carbonyl metal compounds constitutes an active area of research leading to a variety of molecular architectures which span from dinuclear metal species to giant molecular clusters.1,2 Work by several groups3 has demonstrated that small chalcogenido clusters, such as dinuclear and trinuclear Group 8 metal complexes, are useful synthons for cluster-growing processes by reaction with suitable transition-metal species.Amongst the possible synthetic routes available to prepare small chalcogenido clusters, that involving the oxidative addition of phosphine chalcogenides to zerovalent metal carbonyl species appears advantageous, if a phosphine-substituted cluster is the target compound.4,5 In this regard, we recently found that the reaction of [Ru3(CO)12] with SePPh3 is quite selective, affording the disubstituted open-triangular cluster [Ru3(m3-Se)2(CO)7(PPh3)2] 1 in high yield.6 The availability of this complex in workable amounts prompted us to investigate its behaviour in cluster-growing processes by reaction with [M(CO)3(MeCN)3] (M = Mo or W).The preliminary results of this study, including the molecular structure of the new bimetallic cluster [M2Ru2(m4-Se)2(m-CO)4- (CO)6(PPh3)2] (M = Mo 2; M = W 3), are the subject of this communication. Cluster 1, shown in Scheme 1, is an open-triangular, nido cluster with 50 electrons and seven skeletal electron pairs, in accord with the requirements of the 18-electron rule and of the skeletal electron pair (s.e.p.) theory respectively.As a consequence, cluster 1 could be prone to add a zero-s.e.p. fragment, Scheme 1 (i) [M(CO)3(MeCN)3], CH2Cl2, N2, room temperature PPh3 Ru Se Ru Se Ru M Se Ru Se Ru M PPh3 Ph3P Ph3P ( i) 1 M = Mo 2 M = W 3 † E-Mail: predieri@ipruniv.cce.unipr.it ‡ Dedicated to Professor Pascual Royo on the occasion of his 60th birthday. such as M(CO)3 (M = Mo or W), to give the hypothetical closo clusters [MRu3(m4-Se)2(CO)10(PPh3)2]. This would have been expected considering the existence of the analogous monophosphine sulfur derivative [WRu3(m4-S)2(CO)11(PMe2Ph)], obtained by irradiating a solution of [Ru3(m3-S)2(CO)9] and [W(CO)5(PMe2Ph)].7 For this purpose, we have treated the labile intermediates [M(CO)3(MeCN)3] with complex 1 (1 : 1 molar ratio) in dichloromethane.The reactions, monitored by IR spectroscopy, take place in a few hours, at room temperature, giving the clusters 2 and 3 (Scheme 1, carbonyls omitted) as unique isolable products.§ The IR spectra of 2 and 3 are superimposable, suggesting the same molecular structure for both as confirmed by the Fig. 1 View of the molecular structure of [Mo2Ru2(m4-Se)2- (m-CO)4(CO)6(PPh3)2] 2 showing the atom numbering scheme. Selected bond distances (Å) and angles (8) [the values in square brackets refer to cluster 3]: Mo(1)]Se(1) 2.644(2) [2.595(2)], Mo(1)]Se(1A) 2.646(2) [2.620(2)], Mo(1)]Ru(1) 2.844(2) [2.803(2)], Mo(1)]Ru(1A) 2.860(2) [2.845(2)], Mo(1)]C(1) 2.39(2) [2.44(2)], Mo(1)]C(5) 2.02(1) [2.00(2)], Ru(1)]Se(1) 2.572(2) [2.579(2)], Ru(1)]Se(1A) 2.589(2) [2.598(2)], Ru(1)]P(1) 2.372(3) [2.373(3)], Ru(1)]C(1) 2.02(2) [2.06(2)], Ru(1A)]C(5) 2.50(2) [2.59(3)]; Ru(1)]Mo(1)]Ru(1A) 87.62(4) [89.3(1)], Mo(1)]Ru(1)]Mo(1A) 92.38(4) [90.7(1)] § The compound [Ru3Se2(CO)7(PPh3)2] 1 (200 mg, 0.17 mmol) and [Mo(CO)3(MeCN)3] (50 mg, 0.16 mmol) were stirred in dry CH2Cl2 (50 cm3) for 12 h at room temperature under N2.The resulting dark solution was evaporated to dryness and the residue was dissolved in CH2Cl2 (10 cm3).The brown product (cluster 2, 55 mg) was separated and purified by TLC on silica, using CH2Cl2–light petroleum (b.p. 40– 60 8C) (1: 1) as eluent. Crystals suitable for X-ray analysis were obtained by layering methanol on a dichloromethane solution. IR (CH2Cl2), nCO(cm21): 2046ms, 2009vs, 1982s, 1910ms, 1840m, 1805m. 31P-{1H} NMR, d 42.0 (s). The same reaction, under the same conditions, took place between 1 and [W(CO)3(MeCN)3], affording cluster 3; IR (CH2Cl2), nCO(cm21): 2044m, 2009vs, 1981s, 1893m, 1834m, 1804w. 31P-{1H} NMR, d 42.8 (s).322 J. Chem. Soc., Dalton Trans., 1998, Pages 321–322 X-ray analyses.¶ In fact they are isostructural, and the structure of 2 is shown in Fig. 1. Both of these bimetallic clusters exhibit electron-deficient (60 electrons), planar, centrosymmetrical arrays of two Group 6 metals and two ruthenium atoms, bicapped above and below by two quadruply bridging selenium atoms.As a result, the six atoms of the M2Ru2Se2 core from distorted octahedrons, in which four carbonyls asymmetrically bridge the M]Ru edges. The two phosphorus ligands remain attached to the ruthenium atoms. Apart from the unusual mixed-metal core, clusters 2 and 3 are unique, as they contain only six skeletal electron pairs instead of seven, as observed in all other clusters of this family3c–e,6–9 and as required by the s.e.p.theory.10 In particular, [Mo2Fe2S2(CO)12]22 is very similar to 2 and 3, having the same number of two-electron ligands, but it is dianionic, attaining the predicted electron-pair count.8 On the contrary, any anionic nature of clusters 2 and 3 should be ruled out, as they are insoluble in polar solvents (water and light alcohols) and their 1H NMR spectra do not exhibit any other peak (possibly due to hypothetical cations or SeH groups) besides those of the phenyl rings. Furthermore, the final electron density maps for both show a number of residual peaks only around the heavy metals; in the case of 2 the maximum residual peak (1.46 e Å23) is located at 0.95 Å from Mo(1).The surprising stability of the unsaturated clusters 2 and 3 could be attributed to the presence Fig. 2 Cyclic voltammogram at 50 mV s21 of a dichloromethane solution of cluster 2 (1023 M) and [NBun 4][PF6] (0.1 M) on a glassy-carbon electrode ¶ Crystal data for complex 2: C46H30Mo2O10P2Ru2Se2, M = 1356.58, triclinic, space group P1� , a = 10.129(3), b = 11.147(4), c = 11.151(5) Å, a = 95.56(2), b = 97.55(2), g = 110.29(3)8, U = 1156.8(7) Å3, Z = 1, m = 124.61 cm21.Intensity data collected at room temperature. 4395 Unique reflections measured and used in refinement. Final R1 factor [for 2993 reflections with I > 2s(I)] 0.0663. Final R factor for all data 0.0965. Crystal data for complex 3: C46H30O10P2Ru2Se2W2, M = 1532.40, triclinic, space group P1� , a = 10.102(5), b = 11.156(5), c = 11.205(5) Å, a = 95.86(2), b = 97.33(2), g = 110.16(2)8, U = 1161.2(9) Å3, Z = 1, m = 72.68 cm21.Intensity data collected at room temperature. 4087 Unique reflections measured and used in refinement. Final R1 factor [for 3412 reflections with I > 2s(I)] 0.0783. Final R factor for all data 0.0914. CCDC reference number 186/824. of two strongly electron-donating phosphine ligands, which could contribute to the increased electron density on the metal atoms. The electrochemical behaviour has been investigated by recording cyclic voltammograms|| for clusters 2 (Fig. 2), 3 and for the homometallic, seven s.e.p.s cluster [Ru4Se2(CO)9(PPh3)2] 4.6 Clusters 2 and 3 behave similarly showing reduction peaks at 21080 and 21027 mV respectively both on platinum and glassy-carelectrodes. On platinum only, after some scans, an adsorption peak appears at ca. 2830 mV for both mixedmetal clusters and for 4, which does not show any diffusive reduction peak until 21500 mV, where the solvent reduction takes place.The small, irreversible anodic peaks appearing at ca. 2300 mV in the voltammograms of 2 and 3 could be diffusion peaks, as suggested by experiments performed on a rotating disc electrode. Moreover the ratio ipc/ipa does not change either by holding the potential at more negative value than Epc, or by increasing the scan rate from 50 to 1000 mV s21, indicating the presence of oxidizable species which possibly form from the reduction products under steady-state conditions.References 1 L. C. Roof and J. W. Kolis, Chem. Rev., 1993, 93, 1037. 2 D. Fenske, in Clusters and Colloids, ed. G. Schmid, VCH, Weinheim, 1994, pp. 212–297. 3 (a) V. Day, D. A. Lesch and T. B. Rauchfuss, J. Am. Chem. Soc., 1982, 104, 1290; (b) R. D. Adams, J. E. Babin, J.-G. Wang and W. Wu, Inorg. Chem., 1989, 28, 703; (c) P. Mathur, B. H. S. Thimmappa and A. L. Rheingold, Inorg. Chem., 1990, 29, 4658; (d ) P.Mathur, I. J. Mavunkal, V. Rugmini and M. F. Mahon, Inorg. Chem., 1990, 29, 4838; (e) P. Mathur, D. Chakrabarty and Md. Munkir Hossain, J. Organomet. Chem., 1991, 418, 415; ( f ) P. Mathur, D. Chakrabarty, Md. Munkir Hossain and R. S. Rashid, J. Organomet. Chem., 1991, 420, 79; (g) P. Mathur, Md. Munkir Hossain and A. L. Rheingold, Organometallics, 1994, 13, 3909; (h) M. Shieh, T.-F. Tang, S.-M. Peng and G.-H. Lee, Inorg. Chem., 1995, 34, 2797; (i) P. Mathur and P. Sekar, Chem. Commun., 1996, 727; (l) S. N. Konchenko, A. V. Virovets and N. V. Podberezskaya, Polyhedron, 1997, 16, 1689. 4 D. Cauzzi, C. Graiff, M. Lanfranchi, G. Predieri and A. Tiripicchio, J. Chem. Soc., Dalton Trans., 1995, 2321. 5 A. M. Z. Slawin, M. B. Smith and J. D. Woolins, J. Chem. Soc., Dalton Trans., 1997, 1877. 6 P. Baistrocchi, D. Cauzzi, M. Lanfranchi, G. Predieri, A. Tiripicchio and M. Tiripicchio Camellini, Inorg. Chim. Acta, 1995, 235, 173. 7 R. D. Adams, T. A. Wolfe and W. Wu, Polyhedron, 1991, 10, 447. 8 P. A. Eldredge, K. S. Bose, D. E. Barber, R. F. Bryan, E. Sinn, A. Rheingold and B. A. Averill, Inorg. Chem., 1991, 30, 2365. 9 B. F. G. Johnson, T. M. Layer, J. Lewis, A. Martin and P. R. Raithby, J. Organomet. Chem., 1992, 429, C41. 10 D. M. P. Mingos, Acc. Chem. Res., 1984, 17, 311. Received 19th November 1997; Communication 7/08355F || Dichloromethane solutions (1023 M) containing [NBun 4][PF6] (0.1 M); platinum and glassy carbon were used as working electrodes; the reference was Ag–AgCl–KCl (saturated); scan rate 50 mV s21.

ISSN:1477-9226    

DOI:10.1039/a708355f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 323–329 323 Dimethylthiophosphoryl hydrazone ligands and their copper complexes: crystal structures and analysis of their solution complexation behaviour by electrospray mass spectrometry Andrei S. Batsanov, Andrei V. Churakov, Morag A. M. Easson, Linda J. Govenlock, Judith A. K. Howard, Janet M. Moloney and David Parker * Department of Chemistry, University of Durham, South Road, Durham, UK DH1 3LE Received 14th September 1998, Accepted 3rd December 1998 Four new thiophosphoryl ligands containing NNPS or CNPS connectivities have been prepared and three of them characterised by single-crystal X-ray diVraction.The complexation of these ligands with salts of CuI, CuII, NiII and ZnII has been studied by electrospray mass spectrometry and uv-visible spectrophotometry in acetonitrile and methanol solution. Evidence for 1 : 1 and 2 : 1 complexation is presented. Two copper(I) complexes were isolated including a mixed chloride–bromide copper(I) complex in which a dimeric structure is adopted with a 22-membered metallacycle.The incorporation of phosphorus-containing donors within hydrazine and hydrazide ligands has been reported in the development of new ligand systems for the co-ordination of transition metals.1 Many of these ligands possess P–N–N–P or N– N–P–N–N connectivities and some of these systems with P]] S donors have been explored 2 for the complexation of 99mTc or used as ligands in complexes with PdII, CuI and CoI.3 Thiophosphorylamide ligands themselves have been studied for well over 30 years, and sterically demanding systems with an SPNPS connectivity have attracted considerable attention,4 in view of the tendency of nickel(II) and zinc(II) complexes to form well defined ML2 complexes, involving tetrahedral S4 ligation.5 Given our recent interest in developing the co-ordination chemistry of azaphosphinate 6 and tetradentate azathiaphosphinate ligands,7 we were attracted by the structures of the azathiaphosphoryl ligands L1–L4 with a view to exploring their coordination chemistry in ML and ML2 complexes.An added impetus for this exploratory study was the structural resemblance of L1 and L2 to the 1,2-dicarbonyl-derived thiosemicarbazone ligands ‘PTSM’ and ‘ATSM’, which are being used experimentally as perfusion tracers in positron emission tomography as their readily reduced complexes of 62CuII (62Cu: b1, t2� 1 9.8 min).8 N N HN P HN P R R S S Me Me Me Me N Me NH P S Me Me N HN NH P S Me Me P S Me Me L1 R = Me L2 R = Ph L3 L4 N N N N R R' S S Cu NHMe NHMe R = Me, R' = H (CuII-PTSM) R = Me = R' (CuII-ATSM) Ligand synthesis and structure The general approach to ligands of the type L1 to L4 that was considered first involved an attack of a dialkylthiophosphoryl bromide on the appropriate hydrazine or primary amine.The precursor Me2P(S)Br, may be prepared by cleavage of tetramethyldiphosphine disulfide with molecular bromine at 215 8C.9,10 Reaction of Me2P(S)Br in CCl4 with the appropriate dihydrazone, prepared following Cope’s general procedure,11 in the presence of Et3N gave the desired ligands L1 and L2 in good yield (Scheme 1).In the absence of a base, the reaction with benzil dihydrazone unexpectedly led to formation of a mixture of products including an eight-membered tetrazocine ring (Scheme 1), the formation of which is well defined from the more obvious and established condensation reaction of benzil dihydrazone with benzil itself.12 Reaction of Me2P(S)Br with 3 equivalents of 2-amino-6- methylpyridine gave L3 in 82% yield after crystallisation: reducing the amount of the aminopyridine gave rise to concomitant formation of the N,N-disubstituted ligand 6- MeC5H3N{N[PMe2(S)]2-2}, which was identified by ESMS, 1H and 31P NMR but was not isolated.The tridentate ligand L4 was prepared in the same way. Crystal structures of L1, L2 and L3 Crystals suitable for analysis by X-ray diVraction of the ligands L1, L2 and L3 were obtained from diethyl ether by slow evaporation.Compounds L1 and L2 contain the same S]] P(Me2)- NHN]] C–C]] NNHP(Me2)]] S backbone and diVer only in the substituents at the central carbon atoms. In each case the amino N atoms have planar (sp2) bonding geometry within experimental error. However, the molecular conformations are quite diVerent. The molecule L1 possesses a crystallographic inversion centre (Fig. 1). All non-H atoms (save the phosphorusbound methyl carbons) are coplanar, with the maximum deviation from the plane being 0.07 Å and the mean one 0.03 Å. In L2 the backbone adopts a ‘curled’ conformation (Fig. 2) being twisted by ca. 728 around the central C(1)–C(2) bond and by a smaller degree around other single bonds (see Table 1). The phenyl rings at C(1) and C(2) form dihedral angles of 11 and 198 with the planes of adjacent C]] N bonds.Comparison of the bond lengths in L1 and L2 (Table 1) reveals small but significant diVerences indicative of greater p conjugation along the chain in L1, particularly the shortening324 J. Chem. Soc., Dalton Trans., 1999, 323–329 Scheme 1 N N HN P HN P R R S S Me Me Me Me N N NH2 NH2 R R P S Me Me Br N N N N Ph Ph Ph Ph P P S S Me Me Me Me N NH2 H2N N NH HN P P S S Me Me Me Me N NH2 Me N NH Me P S Me Me N N Me P S Me Me P Me Me S Et3N + CCl4 Br2 CCl4 L1, R = Me L2 R = Ph L4 2 equivalents L3 3 equivalents of the central C–C bond by ca. 0.03 Å. The bulkiness of the phenyl substituents is not suYcient to explain the change of conformation, as a trans disposition of the phenyl groups around the C(1)–C(2) bond (as with the methyls in L1) would not produce substantial (intramolecular) steric strain. The absence of significant hydrogen bonding in both L1 and L2 is noteworthy. In L2 the intramolecular contacts N(2)– Fig. 1 Molecular structure of L1 showing 50% probability ellipsoids.Primed atoms are symmetrically related via the inversion centre. Fig. 2 Molecular structure of L2 showing 50% probability ellipsoids and intramolecular H ? ? ? N contacts. H? ? ? N(3) and N(4)–H ? ? ? N(1) exhibit H ? ? ? N distances of 2.55(2) and 2.53(2) Å (somewhat shorter than the sum of van der Waals radii,13 2.74 Å) and N–H–N angles of 116(2) and 119(2)8. They cannot be regarded as hydrogen bonds, since the directions of the contacts do not correspond to the lone pair sites of N(1) and N(3). In L1 the only intermolecular N(1)– H? ? ? S contact of 2.90 Å is equal to the sum of van der Waals radii (H, 1.10 1 S, 1.81 Å).13 Molecule L3 (Fig. 3) contains a roughly planar S]] P–NH– pyridyl system. Inversion-related molecules are linked by N(2)– H? ? ?S9 hydrogen bonds, lying practically in the same plane. The H ? ? ? S distance of 2.54(2) Å (for the normalised H position) is common for N–H ? ? ?S]] C bonds (2.28–2.72, mean 2.46 Å),14 while the PS ? ? ? H angle of 120(1)8 is wider than CS ? ? ?H (angles usually are 98–1148), although it corresponds well to the lone pair direction of the sulfur atom.The P–S bond lengths in R3P]] S molecules average 1.954(5) Å, but are contracted to an average of 1.922(14) Å if one of the carbon substituents (R) at the phosphorus atom is replaced by a more electronegative O or N group.15 Particularly, in the Me2(RNH)P]] S moieties, the P–S distance is 1.940 Å in the absence of hydrogen bonds, but can increase to 1.949–1.962 Å when the sulfur atom is engaged in such bonds.16 Our results are in good agreement with the earlier data: P–S distances average 1.941 Å in L1 where sulfur atoms accept no hydrogen bonds compared to 1.966 Å in L3 with a strong hydrogen bond.Preliminary assessment of solution complexation by electrospray mass spectrometry The use of electrospray mass spectroscopy in the qualitative investigation of the solution speciation of metal complexes has attracted considerable recent attention.17–20 Samples were prepared in dry acetonitrile using equimolar mixtures of [CuI- (MeC4]BF4 or Cu(ClO4)2, and were examined in positive ion mode in the presence of dry methanol.With L1 and copper(II) perchlorate a single species was observed at m/z 359.94 corresponding very closely to the theoretical isotope model forJ. Chem. Soc., Dalton Trans., 1999, 323–329 325 [Cu(L1 2 H)]1 (C8H19CuN4P2S2 requires m/z 359.98).Using the copper(I) salt, as the source of copper, the only species observed (100 V core voltage) was at m/z 360.78, corresponding well to C8H20CuN4P2, i.e. [CuL1]1 (Calc. m/z 360.99). With Ni(ClO4)2 a strong signal was observed for [NiL1 2 H]1, but with Zn(ClO4)2 several species were identified including [ZnL1(L1 2 H)]1. Under the experimental conditions, copper may form monopositive 1 : 1 complexes with L1 in both the monovalent and divalent state, in the latter case by loss of one of the ligand NH protons.With the C-phenyl analogue, L2, addition of Cu(ClO4)2 gave rise to species at m/z 484.74 and 906.27, corresponding to [ML]1 and [ML2]1 formation. The isotope pattern at lower mass showed peaks separated by one mass unit and corresponded to formation of a 1: 1 copper(I) complex (C18H24CuN4P2S2 1 requires m/z 485.02). The higher mass pattern corresponds to the loss of one ligand proton, suggesting that it may be due to a copper(II) complex.Complexation with zinc produced a similar result and [Zn(L2 2 H)]1 and [ZnL2(L2 2 H)]1 species were observed, but in addition a [ZnL2(ClO)4]1 species was found and a doubly charged [ZnL2]21 species at half-mass (i.e. with non-integral isotope separations). Complexes of nickel produced similar spectra to those of zinc, with [NiL2]1, [NiL2(L2 2 H)]1 and [NiL2]21 species dominant. With the bidentate ligand L3 separate sets of experiments were carried out with 1 : 1 and 2 : 1 ligand to metal ion ratios. With Cu(ClO4)2, at both 2 : 1 and 1 : 1 stoichiometry, an [ML2] species dominated the spectrum, in addition to some [ML] and the protonated ligand.The [CuL3]1 and [Cu(L3)2]1 species at m/z 262.98 and 462.95 were consistent with the loss of ligand protons in neither species, suggesting that copper(I) complexes had formed. With both nickel and zinc the major species Table 1 Selected bond distances (Å), bond and torsion angles (8) for L1, L2, L3 and complex 1 L1 P–S N(1)–N(2) C(1)–C(19) 1.955(1) 1.382(1) 1.478(2) P–N(1) C(1)–N(2) P–C mean 1.676(1) 1.296(2) 1.801(2) L2 P(1)–S(1) P(1)–N(2) N(1)–N(2) C(1)–N(1) C(1)–C(2) N(4) ? ? ? N(1) N(1)–C(1)–C(2)–N(3) S(1)–P(1)–N(2)–N(1) P(1)–N(2)–N(1)–C(1) 1.935(1) 1.680(2) 1.376(2) 1.292(2) 1.506(2) 3.156(2) 71.9(2) 167.4(1) 2168.7(1) P(2)–S(2) P(2)–N(4) N(3)–N(4) C(2)–N(3) P–C mean N(2) ? ? ? N(3) N(3)–N(4)–P(2)–S(2) C(2)–N(3)–N(4)–P(2) 1.947(1) 1.689(2) 1.380(2) 1.291(2) 1.786(5) 3.139(2) 2177.1(1) 2165.6(1) L3 P–S C(2)–N(2) N(2) ? ? ?S9 N(2)–H(2) ? ? ?S9 S–P–N(2)–C(2) 1.966(1) 1.400(2) 3.544(2) 166(1) 176.2(1) P–N(2) P–C mean P–N(2)–C(2)–N(1) 1.677(1) 1.798(2) 7.2(2) 1 Cu–S(1) Cu–Cl/Br P(1)–S(1) P(1)–N(1) N(1)–N(2) C(1)–N(2) S(1)–Cu–S(29) S(1)–Cu–Cl/Br S(1)–P(1)–N(1)–N(2) P(1)–N(1)–N(2)–C(1) 2.270(1) 2.293(2) 1.988(1) 1.678(2) 1.386(3) 1.296(3) 117.45(3) 117.75(3) 2151.5(1) 2168.1(1) Cu–S(29) C(1)–C(2) P(2)–S(2) P(2)–N(4) N(3)–N(4) C(2)–N(3) S(29)–Cu–Cl/Br N(3)–N(4)–P(2)–S(2) P(2)–N(4)–N(3)–C(2) 2.265(1) 1.485(4) 1.997(1) 1.685(2) 1.398(3) 1.296(3) 124.33(3) 270.1(1) 153.9(1) observed at 2 : 1 stoichiometry was the protonated ligand, and at a 1 : 1 ratio this species was also dominant.A much less intense (<20%) peak of an [ML3(L3 2 H)]1 species was conspicuous in each case, with the nickel complex twice as intense as the corresponding zinc one. Using the tridentate NS2 ligand L4, addition of copper(II) perchlorate gave a simple mass spectrum (m/z 355.8 and 357.75 in 2 : 1 ratio) showing only formation of a 1 : 1 complex of copper(I).Analysis of the spectra of the corresponding nickel and zinc perchlorate complexes revealed that several species were present including [M(L4 2 H)]1, [M(L4 2 H)(ClO)4]1, [ML4(L4 2 H)]1 and [M(L4)2(ClO)4]1. In this case, as with copper complexes of L1–L3, it is unclear if reduction to a copper( I) species occurred prior to mass spectral study or during the ionisation process.Accordingly, the copper complexes were examined further by uv-visible spectral analysis. Spectrophotometric studies of copper complexes Ultraviolet and visible spectra of ligands L1 to L4 were recorded in acetonitrile in the absence and presence of [Cu(MeCN)4]BF4 and anhydrous Cu(CF3SO3)2. Comparison of the results in the presence of added copper-(II) and -(I) salt (Table 2) reveals that only L1 forms a well defined copper(II) complex (ld–d max = 636 nm; e = 210 dm3 mol21 cm21) while in the presence of L2, L3 and L4 the metal ion is reduced and a copper(I) complex is formed.Direct addition of the copper(I) tetraacetonitrile complex to L2, L3 and L4 gives rise to a species with identical spectral characteristics. In contrast, addition of CuI and CuII to L1 yields the corresponding ligated complexes of CuI and CuII. In the latter case there is a strong LMCT band at 392 nm, and in the former a relatively intense MLCT band at 430 nm responsible for the observed yellow colour in solution.The copper(I) complex was isolated as an analytically pure orange solid and characterised as a neutral complex, [Cu(L1 2 H)], formed by monodeprotonation of the ligand. It gave identical ESMS and spectrophotometric data to those of the same complex prepared in situ and discussed above. The complexes formed in solution between L1–L4 and copper( II) triflate were also examined in MeOH. With L1 a dark green solution was observed again (lmax 638, e 150 dm3 mol21 cm21) while with L2, L3 and L4 there was no sign of this d–d band, consistent with reduction to a copper(I) complex.The only significant diVerence was with L2, for which an extra charge transfer band at 404 nm (e = 180 dm3 mol21 cm21) was identified which may be related to the enhanced stabilisation of the charge-transfer state by the protic solvent. Isolation of [Cu2(L1)2Cl2 2 xBrx] 1 and X-ray analysis Repeated attempts to isolate crystals of the copper complexes of L1, L2 and L3 with triflate and tetrafluoroborate counter ions Fig. 3 Molecular structure of L3 showing 50% probability ellipsoids and hydrogen bonds with an inversion related molecule (primed atoms).326 J. Chem. Soc., Dalton Trans., 1999, 323–329 were unsuccessful. However when an acetonitrile solution of [Cu(L1 2 H)] was inadvertently exposed to HCl/HBr vapour it slowly lost its very pale yellow colour and deposited colourless crystals [labs (MeCN) = 274 nm with no 430 nm band; dp 66.0 (CD3CN); dH 1.91 (24 H, d, J = 10.7 Hz, PMe), 1.97 (6 H, s, CMe overlapping with solvent), 2.05 (6 H, s, CMe)].A solution of this solid complex in MeCN was stable for at least 2 months in dry acetonitrile, but decomposed fairly rapidly in the presence of water or methanol. Analysis of an aged (24 h) solution containing MeOH or water by 31P NMR revealed a loss of the original peak (P]] S) and appearance of 2 peaks at d 59.0 and 32.1 consistent with hydrolysis of one of the phosphorus– sulfur bonds.Such a reaction may be mediated by the metal acting as a charge sink, facilitating attack at phosphorus by a water molecule, leading to the formation of a P=O bond. At the same time the band at 274 nm disappeared to be replaced by a new band at 230 nm. An electrospray mass spectrum of a fresh sample in CD3CN revealed a relatively intense peak of a singly charged species clustered around m/z 862 {corresponding to [Cu2Cl4(L1 2 H)]1} with a daughter peak at m/z 800.The peaks both disappeared slowly following addition of water or MeOH. Complex 1 was characterised by a single-crystal structure analysis. The 2 : 2 dimer molecule (Fig. 4, Tables 1 and 3) possesses crystallographic Ci symmetry. Each copper(I) centre is coordinated in a trigonal-planar fashion by two sulfur atoms from diVerent L1 ligands and a chloride ligand, which is in part isomorphously substituted by bromide. Thus the molecule contains an unusual 22-membered ring, stabilised by two transannular N(1)–H ? ? ? Cl(Br) hydrogen bonds.The central P HN S Me Me N Me Me N NH P S Cu Cu S S X X Me Me P N N Me Me N HN P Me Me Me Me H 1 X = Cl, Br Table 2 Spectral details of ligands L1 to L4 in MeCN and in the presence of 1 equivalent of [Cu(MeCN)4]BF4 or Cu(CF3SO3)2 Ligand or complex L1 L2 L3 L4 L1 1 CuII L2 1 CuII L3 1 CuII L4 1 CuII L1 1 CuI L2 1 CuI L3 1 CuI L4 1 CuI lmax/nm 274 276 232 286 244 300 206 302 392 636 208 280 232 286 240 302 204 274 430 280 232 286 240 302 e/dm3 mol21 cm21 3.1 × 104 3.1 × 104 1.1 × 104 6.3 × 103 7.5 × 103 6.0 × 103 1.96 × 104 1.25 × 104 4.7 × 103 2.1 × 102 shoulder 7.8 × 104 4.9 × 104 1.5 × 104 shoulder 8.8 × 103 4 × 104 3.5 × 104 2.0 × 103 7.8 × 104 1.85 × 104 6.25 × 104 1.84 × 104 1.2 × 104 N(1)N(2)]] C(1)C(2)]] N(3)N(4) system of the L1 ligand adopts an essentially planar, all-trans, conformation, as in uncoordinated L1 (see above).However, the P]] S bonds are no longer coplanar with this moiety, but are tilted out of its plane to the same side.The methyl substituents at C(1) and C(2) are no longer equivalent, the former pointing into the molecular cavity and the latter outwards. Co-ordination with CuI results in a lengthening of the P]] S bonds by ca. 0.04 Å, but other bond distances change insignificantly. There is only one other structurally characterised metal complex of a ligand that incorporates the NNPS sub-unit,21 which involves a sulfur bridged copper dimer.There are, however, several examples of heterocyclic compounds with such a constitution.22 The N(4)–H groups, pointing outward, form hydrogen bonds with the halogenide ligands of adjacent molecules, symmetrically related via a 21 axis, giving rise to a layered molecular packing. The H ? ? ? Cl(Br) distances (2.37 Å for the intramolecular and 2.48 Å for the intermolecular bonds) are comparable with the average H ? ? ? Cl (2.22 Å) and H ? ? ? Br (2.39 Å) hydrogen bond lengths.23 In solution, this isolated halogenocopper(I) dimer behaved quite diVerently to the complexes prepared in the presence of the non-co-ordinating triflate or perchlorate anions: such differences are often encountered in the co-ordination chemistry of relatively weakly binding ligands, such as those described herein.Experimental General procedures and instrumentation have been reported recently.6,20 Dimethylthiophosphinic bromide was prepared according to the literature method,9,10 mp 33–34 8C (lit.,10 34 8C), and the hydrazones of butane-2,3-dione 24 and benzil 11 were prepared following Cope’s original procedure 11 as analytically pure samples.Electrospray mass spectrometry Electrospray mass spectra were obtained using a VG-Platform II (Fisons Instruments), with a capillary voltage of 4 kV and a source temperature of 60 8C. Cone voltages and analyte concentrations were varied according to the nature of the experiment, but were typically 40 to 60 V and 1025 to 1026 M respectively. The solvent flow was maintained using a Hewlett Packard HPLC instrument that was directly linked to the mass spectrometer.The sample was inserted into the flow using an injection valve with a 10 mL steel loop and transported to the electrospray capillary through a silica tube. Fig. 4 Molecular structure of complex 1, showing 50% probability ellipsoids and hydrogen bonds. Primed atoms are related via an inversion centre, double primed via a 21 axis.J. Chem.Soc., Dalton Trans., 1999, 323–329 327 Major ions are quoted as a percentage of the base peak intensity and isotope patterns were modelled using Mass lynx software. Typically, a stock solution of the ligand was prepared (ca. 0.3 mM) in freshly distilled dry methanol (Aristar grade, BDH) and stock solutions of the dried metal trifluoromethanesulfonate or perchlorate (CAUTION: hazard) salts (ca. 1.0 mM) also prepared in dry methanol. Samples were prepared in polypropylene Eppendorf vials and transferred using Gilson Pipetman micropipettes. To a sample of the ligand solution (1 ml) and methanol (1 ml) was added the appropriate volume of the metal triflate/perchlorate salt, to make a 1 : 1 or 1 : 2 metal : ligand ratio. A 10 mL sample of this solution was injected and mass spectra recorded in positive or negative ion mode, as stated. Preparations Butane-2,3-dione bis[(dimethylthiophosphoryl)hydrazone] L1.To a solution of Me2P(S)Br (27.6 mmol) in carbon tetrachloride (10 ml), butane-2,3-dione dihydrazone (1.5 g, 13.2 mmol) and triethylamine (3 g, 30 mmol) were added. The reaction mixture was stirred under an inert atmosphere for 15 h. A yellow precipitate formed which was isolated by filtration and washed thoroughly with water to remove the triethylamine salts. The pale yellow solid was dried in vacuo (3.5 g, 89%); mp > 250 8C; dP (CDCl3) 65.43; dH (CDCl3) 1.92 [12 H, d, 2JP = 13.5 Hz, P(CH3)2], 1.98 (6 H, s, CH3C) and 5.99 (2 H, d, 2J = 21 Hz, PNHN); dC (CDCl3) 9.93 (C–CH3), 22.64 [d, 1J = 68 Hz, P(S)(CH3)2] and 148.30 (s, CH3]C]] NNP); m/z (EI1) 298 (20, M1) and 205 [100%, M 2 P(S)(CH3)2]; nmax 3270 br (NH), 1580, 1410–1400, 1375, 1350, 1290, 1280, 1060s, 940, 900, 850, 730, 630 and 580 cm21 (Found: C, 31.8; H, 6.67; N, 18.3%.C4H10N2PS requires C, 32.2; H, 6.76; N, 18.8%); lmax(MeCN) 274 nm (e = 3.1 × 104 dm3 mol21 cm21). 1,2-Diphenylethane-1,2-dione bis[(dimethylthiophosphoryl)- hydrazone] L2. To a solution of dimethylthiophosphinic bromide, (2.9 mmol) in carbon tetrachloride (10 ml), benzil dihydrazone (328 mg, 1.4 mmol) and triethylamine (0.42 ml, 3.03 mmol) were added. The reaction mixture was stirred under an inert atmosphere for 15 h and yielded a white precipitate. The solid was removed by filtration and the solvent evaporated from the mother-liquor under reduced pressure to leave an orange oil.This was triturated with diethyl ether and dark yellow crystals formed (4.0 g, 70%). If crystals did not form easily, the ether solution was, instead, washed with water (3 × 30 ml), dried (Na2SO4) and the solvent removed under reduced pressure to give the pale orange solid product, mp 164–167 8C; dP (CDCl3) 66.71; dH (CDCl3) 2.05 [12 H, d1d, 2JP = 12, P(S)(CH3)2], 6.15 (2 H, d, 2JP = 21.5 Hz, PNHN), 7.26–7.37 (6 H, m, m-, p-H of Ph) and 7.51–7.55 (4 H, m, o-H of Ph); dC (CDCl3) 22.71 [d, 1JP = 68, P(S)(CH3)2], 23.20 [d, 1JP = 68, P(S)(CH3)2], 126.52 (m-C of Ph), 129.74 (o-C of Ph), 130.88 (p-C of Ph), 133.53 (PPh) and 143.00 (d, 3J = 14.5 Hz, Ph-C]] NNH); m/z (ESMS1, MeOH) 423.07 (100%, M 1 H1); nmax (KBr) 3290, 1440, 1380, 1290, 1125, 1060, 1035, 1010, 990, 950, 910, 860, 740s, 670s and 600s cm21 (Found: C, 51.3; H, 5.78; N, 13.0. C9H12N2PS requires C, 51.2; H, 5.73; N, 13.3%); lmax (MeCN) 276 nm (e = 3.1 × 104 dm3 mol21 cm21). 2-(Dimethylthiophosphorylamino)-6-methylpyridine L3.To a solution of Me2P(S)Br (5.46 mmol) in carbon tetrachloride (50 ml) was added 2-amino-6-methylpyridine (1.78 g, 16.4 mmol) and the resultant solution stirred for 5 h at room temperature under an inert atmosphere. A white precipitate formed, was isolated by filtration and then washed thoroughly with cold diethyl ether. White crystals formed in the ether layer which were isolated by filtration and dried in vacuo (0.9 g, 82%); mp 138–140 8C; dP (CDCl3) 59.96; dH (CDCl3) 2.17 [6 H, d, 2JP = 14, PS(CH3)2], 2.41 (3 H, s, CH3 of py), 5.14 (1 H, br s, NH), 6.41 [1 H, d, 3J = 8, H(3) of py], 6.68 [1 H, d, 3J = 7.5, H(4) of py] and 7.41 (1 H, t, 3J = 8 Hz); dC (CDCl3) 23.53 (d, 1JP = 68 Hz, PCH3), 24.74 (s, CH3), 130.81, 131.31, 131.82, 162.17 and 169.65.nmax 3280 (aryl H), 1600 (aryl H), 1570, 1450–1370, 1280s, 950s, 910–860, 780, 740s, 700, 580 and 530 cm21; m/z (DCI) 201 (100, M 1 H1) and 185 (5%, M 2 CH3) (Found: C, 47.6; H, 6.50; N, 13.80.C8H13N2PS requires C, 48.0; H, 6.54; N, 14.0%). lmax (MeCN) 232 (e = 1.6 × 104) and 286 nm (e = 6.3 × 103 dm3 mol21 cm21). 2,6-Bis(dimethylthiophosphorylamino)pyridine L4. To a stirred solution of Me2P(S)Br (4.28 mmol) in carbon tetrachloride (15 ml) 2,6-diaminopyridine (110 mg, 1.02 mmol) and triethylamine (0.31 ml, 2.24 mmol) were added and the reaction mixture was stirred for 16 h at room temperature under an inert atmosphere. A white precipitate formed which was isolated by filtration, washed with water and dried in vacuo (0.75 g, 60%); mp 201–203 8C (decomp.); dP (CDCl3) 56.99; dH (CDCl3) 2.14 [12 H, d, 2J = 13.7, P(CH3)2], 5.19 (2 H, d, 2JP = 5.5, NH), 6.31 [2 H, d, 3J = 8, H(3), H(5) of py], 7.43 [1 H, t, 3J = 8 Hz, H(4) of py]; dC (CDCl3) 24.06 [d, 1JP = 68, P(CH3)2], 104.42 [d, 2JP = 4.5 Hz, C(3) and C(5) of py], 140.92 [C(4) of py] and 153.9 [C(2) and C(6) of py]; nmax 3150 (aryl H), 1600s, 1580s, 1450s, 1400, 1290, 1210, 1160, 1040, 950–930, 890, 790s, 720s, 650 and 580 cm21; m/z (DCI) 294 (100, M 1 H1) and 202 [20%, M 2 P(S)Me2 1 2H1] (Found: C, 36.1; H, 5.81; N, 14.17.C9H17N3P2S2? 0.5H2O requires C, 35.8; H, 6.00; N, 13.90%); lmax (MeCN) 244 (e 7.5 × 103) and 300 nm (e 6 × 103 dm3 mol21 cm21). {Butane-2,3-dione bis[(dimethylthiophosphoryl)hydrazonato]} copper(I) [Cu(L1 2 H)]. To a solution of tetrakis(acetonitrile) copper(I) tetrafluoroborate (295 mg, 0.94 mmol) in acetonitrile (35 ml) a solution of L1 (280 mg, 0.94 mmol) in dichloromethane (30 ml) was added.After stirring the deep orange solution a precipitate formed which was isolated by centrifugation, washed with acetonitrile (2 × 10 ml) and dried in vacuo to give an orange solid (350 g, 85%); dP (CD3CN) 66.74; m/z (ESMS1, MeOH) 360.82 (CuL1) (Found: C, 26.5, H, 5.47; N, 15.1. C8H19CuN4P2S2 requires C, 26.6; H, 5.30; N, 15.5%); lmax (MeCN) 274 (e 3.5 × 104) and 430 mn (2.0 × 103 dm3 mol21 cm21).A solution of this complex in acetonitrile slowly lost its colour on exposure to an atmosphere of HCl/HBr vapour, and colourless crystals of the dimer 1 were deposited over 18 h: l (MeCN) 274 nm (no band at 430 nm); dP (CD3CN) 166.0; dH (CD3CN, 300 MHz, 293 K) 1.91 (24 H, d, J = 10.5 Hz, PMe), 1.97 (6 H, s, CMe) and 2.05 (6 H, s, CMe) (Found: C, 46.0; H, 9.31; N, 26.7. C16H40Br0.5Cl1.5Cu2N8P4S4 requires C, 46.2; H, 9.65; N, 27.0%). {Butane-2,3-dione bis[(dimethylthiophosphoryl)hydrazone]- copper(II) bis(trifluoromethanesulfonate), [CuL1][CF3SO3]2.To a stirred solution of copper(II) triflate (120 mg, 0.34 mmol) in dry acetonitrile (7 ml) was added solid L1 (100 mg, 0.34 mmol). The reaction mixture immediately turned deep green and was heated at 80 8C for 4 h then left to cool. Diethyl ether (15 ml) was added and the solution left to stand at 25 8C, however crystallisation did not occur over 48 h. The solvents were removed under reduced pressure to give a green oil. Crystallisation was attempted using tetrahydrofuran but no solid was formed although visible spectroscopy and mass spectral analysis con- firmed the formation of the complex; m/z (ESMS1, MeOH) 359.98 (100, M1) and 362.02 (50%) corresponding to calculated copper isotope pattern; lmax (MeCN) 206 (e 1.96 × 104), 302 (1.25 × 104), (392 4.7 × 103) and 636 nm (210 dm3 mol21 cm21).{1,2-Diphenylethane-1,2-dione bis[(dimethylthiophosphoryl)- hydrazone]}copper(I) tetrafluoroborate, [CuL2]BF4.To a solution of tetrakis(acetonitrile)copper(I) tetrafluoroborate (37 mg, 0.12 mmol) in acetonitrile (5 ml) was added a solution of L2328 J. Chem. Soc., Dalton Trans., 1999, 323–329 Table 3 Crystal data for compounds L1–L3 and complex 1 Formula Formula weight Colour Crystal size/mm T/K Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 Z Dc/g cm23 m/cm21 Maximum 2q/8 Data total, unique R(int) Transmission, minimum, maximum Data observed, I > 2s(I) Number of variables R(F, observed data) wR(F 2, all data) Goodness of fit Maximum, minimum Dr/e Å23 L2 C18H24N4P2S 422.5 Yellow 0.3 × 0.5 × 0.5 150 Triclinic P1� (no. 2) 9.588(1) 11.402(1) 11.620(1) 69.43(1) 87.11(1) 69.74(1) 1112.1(2) 2 1.26 3.9 51 4911, 3567 0.027 0.828, 0.906 3319 332 0.032 0.094 1.05 0.26, 20.30 L1 C8H20N4P2S2 298.3 Orange 0.15 × 0.2 × 0.4 120 Monoclinic P21/n (no. 14) 6.220(1) 12.146(1) 10.141(1) 97.94(1) 758.8(2) 4 1.31 5.5 55 5381, 1718 0.024 0.726, 0.843 1595 114 0.025 0.073 1.06 0.38, 20.27 L3 C8H13N2PS 200.2 Colourless 0.2 × 0.3 × 0.4 150 Monoclinic P21/c (no. 14) 11.415(1) 8.348(1) 11.983(1) 114.54(1) 1038.7(2) 4 1.28 4.2 61 8012, 2806 0.028 0.857, 0.952 2425 161 0.033 0.082 1.16 0.35, 20.26 1 C16H40Br0.54Cl1.46Cu2N8P4S4 415.1 Colourless 0.1 × 0.35 × 0.35 150 Monoclinic P21/c (no. 14) 11.649(1) 15.064(1) 9.682(1) 94.11(1) 1694.7(2) 4 1.63 27.5 60 13814, 4588 0.041 0.442, 0.787 4087 241 0.039 0.086 1.31 0.46, 20.44 (50 mg, 0.12 mmol) in acetonitrile (5 ml).After mixing, the yellow solution was cooled to 25 8C but crystallisation attempts were unsuccessful, even following the slow addition of diethyl ether. Visible spectroscopy and mass spectral analysis confirmed the formation of a copper(I) complex; m/z (ESMS1, MeOH) 485.02 (Calc. for [C18H24CuN4P2S2]1 m/z 485.02 corresponding to the calculated isotope pattern); lmax (MeCN) 280 nm (e 7.8 × 104 dm3 mol21 cm21). Crystal structure analyses X-Ray diVraction experiments were carried out on a Siemens SMART 3-circle diVractometer with a CCD area detector, using graphite-monochromated Mo-Ka radiation (l � = 0.71073 Å) and a Cryostream open-flow cryostat.Crystal data and experimental parameters are listed in Table 3. Data were collected by scanning over a full hemisphere of reciprocal space in frames of 0.38 w. Structures were solved by direct methods and refined by full matrix least squares against F 2 of all data, using SHELXTL software.25 Absorption corrections were performed for L1, L2 and L3 by a semiempirical method based on Laue equivalents and multiple measurements of strong reflections, using the SADABS program.26 For complex 1, crystal faces were indexed and the absorption correction was made by numerical integration. All non-H atoms were refined with anisotropic displacement parameters; all H atoms were located from the Fourier diVerence map and refined in isotropic approximation. Statistically disordered Cl and Br atoms were refined as a single atom with a mixed scattering factor; the contributions were refined to 27.2(3)% Br and 72.8(3)% Cl.CCDC reference number 186/1271. See http://www.rsc.org/suppdata/dt/1999/323 for crystallographic files in .cif format. Acknowledgements We thank EPSRC for studentship support, the Commissioners of the 1851 Exhibition for a Fellowship (to L. J. G.), the Leverhulme Trust for a visiting fellowship (to A.S. B.) and the Royal Society for support (to A. V. C.). References 1 K. V. Katti, V. S. Reddy and P. R. Singh, Chem. Soc. Rev., 1995, 97. 2 W. A. Volkert, P. R. Singh, A. R. Ketring, K. V. Katti and K. K. Katti, J. Labelled Compd. Radiopharm., 1993, 32, 15. 3 K. V. Katti, P. R. Singh and C. L. Barnes, Inorg. Chem., 1992, 31, 4588. 4 A. Schmidtpeter and J. Ebeling, Chem. Ber., 1968, 101, 815; R. Bohm, H. Groeger and H. Schmidtpeter, Angew. Chem., 1964, 76, 860; E. Fluck and F.L. Goldman, Chem. Ber., 1963, 96, 3091. 5 M. R. Churchill and J. Wormald, Chem. Commun., 1970, 703; A. Davison and E. S. Switkes, Inorg. Chem., 1971, 10, 837; M. R. Churchill, J. Cooke, J. Wormald, A. Davison and E. S. Switkes, J. Am. Chem. Soc., 1969, 91, 6518. 6 S. Aime, M. Botta, A. S. Batsanov, R. S. Dickins, S. Faulkner, C. E. Foster, A. Harrison, J. A. K. Howard, J. M. Moloney, T. J. Norman, D. Parker, L. Royle and J. A. G. Williams, J. Chem. Soc., Dalton Trans., 1997, 3623; C.D. Edlin and D. Parker, Tetrahedron Lett., 1998, 2797; G. B. Bates, E. Cole, R. Kataky and D. Parker, J. Chem. Soc., Dalton Trans., 1996, 2693; T. Kiss, M. Jezowrska-Bojezuk, H. Kozlowski, P. Kosfarski and A. Antozak, J. Chem. Soc., Dalton Trans., 1991, 2275. 7 M. A. M. Easson D. Parker, Tetrahedron Lett., 1997, 6091. 8 J. K. Lion, C. J. Mathias and M. A. Green, J. Med. Chem., 1997, 40, 132; Y. Fujibayashi, H. Taniuchi, Y. Yanekura, H. Ohtani, J. Konishi and A.Yokayama, J. Nucl. Med., 1997, 38, 1155. 9 H. Reinhardt, D. Bianchi and D. Molle, Chem. Ber., 1957, 1656. 10 W. Kuchen and H. Buchwald, Angew. Chem., 1959, 4, 162. 11 A. C. Cope, D. S. Smith and R. J. Cotter, Org. Synth., 1963, Coll. Vol. IV, 377. 12 H. Schlesinger, Angew. Chem., 1960, 5, 563. 13 R. S. Rowland and R. Taylor, J. Phys. Chem., 1996, 100, 7384. 14 F. H. Allen, C. M. Bird, R. S. Rowland and P. R. Raithby, Acta Crystallogr., Sect. B, 1997, 53, 680. 15 F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen and R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1987, S1. 16 C. Silvestru, R. Röster, I. Haiduc, R. Cea-Olivares and G. Espinosa- Perez, Inorg. Chem., 1995, 34, 3352; D. Cupertino, R. Keyte, A. M. Z. Slawin, D. J. Williams and J. D. Woollins, Inorg. Chem., 1996, 35, 2695; N. S. Hosmane, A. M. Arif and A. H. Cowley, Acta Crystallogr., Sect. C, 1987, 43, 2013. 17 J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong and C. M. Whitehouse, Science, 1989, 246, 64. 18 G. Hopfgartner, C. Piquet, J. D. Henion and A. F. Williams, Helv. Chim. Acta, 1993, 76, 1759. 19 E. Leize, A. JaVrezic and A. Van Dorsellaer, J. Mass Spectrom., 1996, 31, 537.J. Chem. Soc., Dalton Trans., 1999, 323–329 329 20 M. Goodall, P. M. Kelly, D. Parker, K. Gloe and H. Stephan, J. Chem. Soc., Perkin Trans. 2, 1997, 59. 21 B. Delavaux-Nicot, N. Lugan, R. Mathieu and J.-P. Majoral, Inorg. Chem., 1992, 31, 334. 22 A. Schmidtpeter, J. Gross, E. Schrenk and W. S. Sheldrick, Phosphorus Sulphur Relat. Elem., 1982, 14, 49; B. Wallis, C. Donath, M. Meisel and J. Fuchs, Acta Crystallogr., Sect. C, 1991, 47, 2423; G. L’Abbe, J. Flemal, J. P. Declercq, G. Germain and M. van Meerssche, Bull. Soc. Chim. Belg., 1979, 88, 737. 23 T. Steiner, Acta Crystallogr., Sect. B, 1998, 54, 456. 24 S. Wolfe, Can. J. Chem., 1971, 49, 1099. 25 G. M. Sheldrick, SHELXTL, Version 5/VMS, Bruker axs, Analytical X-ray systems, Madison, WI, 1995. 26 G. M. Sheldrick, SADABS, Program for scaling and correction of area detector data, University of Göttingen, 1996. Paper 8/07142J

ISSN:1477-9226    

DOI:10.1039/a807142j

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 325–326 325 The first direct observation of N]O bond cleavage in the oxidative addition of an oxime to a metal centre. Synthesis and crystal structure of the methyleneamide complex trans- [Re(OH)(N] CMe2)(Ph2PCH2CH2PPh2)2][HSO4] Cristina M. P. Ferreira, M. Fátima C. Guedes da Silva, Vadim Yu. Kukushkin, João J. R. Fraústo da Silva and Armando J. L. Pombeiro *,† Centro de Química Estrutural, Complexo I, Instituto Superior Técnico, Av.Rovisco Pais, 1096 Lisboa codex, Portugal The first direct observation of oxidative addition of an oxime upon N]O bond cleavage has been reported in the reaction of Me2C]] NOH with trans-[ReCl(N2)(Ph2PCH2CH2PPh2)2] in the presence of Tl[BF4]–Tl[HSO4], to form, in a single-pot experiment, the methyleneamide complexes trans-[Re(OH)- (N]] CMe2)(dppe)2][A] (A = BF4 1a or HSO4 1b) which undergo ready replacement of hydroxide by fluoride upon reaction with HBF4; X-ray crystallography (1b) showed that the linearly bound methyleneamide behaves as an effective p acceptor and exerts a significant trans influence on the hydroxide ligand.The co-ordination chemistry of oximes, RR9C]] NOH, is rich, extensively investigated and has been recently reviewed by two of us.1 However, at electron-rich metal centres, in particular those which can bind dinitrogen and other substrates of nitrogenase that have been the object of our interest,2 the coordination chemistry is still unknown.Moreover, oxidative addition of oximes to metal centres is an essentially unexplored area. To the best of our knowledge, only one paper has been published, by Deeming et al.,3 on the O]H bond splitting in the oxidative addition of Me2C]] NOH to the osmium cluster [Os3(CO)10(MeCN)2] giving [Os3(m-H)(m-Me2C]] NO)(CO)10] whose thermal isomerization leads to the hydroxo isomer [Os3- (m-OH)(m-Me2C]] N)(CO)10] along with some other unidentified products.The mechanism of the conversion has not been studied, but if it includes the intermediate formation of [Os3- (m-Me2C]] NOH)(CO)10], as was suggested by the authors,3 the reaction can be considered as oxidative addition of the oxime due to N]O bond splitting. We herein report the first direct observation of such a reaction by treatment of a THF solution of trans-[ReCl(N2)(dppe)2] (dppe = Ph2PCH2CH2PPh2) with 2-propanone oxime, Me2C]] NOH, in the presence of a chloride abstractor, Tl[BF4]– Tl[HSO4], and in sunlight (to promote N2 loss), to give the hydroxo–methyleneamide complexes trans-[Re(OH)(N]] CMe2)- (dppe)2][A] (A = BF4 1a or HSO4 1b)‡ [equation (1)].This trans-[ReCl(N2)(dppe)2] 1 Me2C]] NOH 1 T1[A] æÆ trans-[Re(OH)(N]] CMe2)(dppe)2][A] 1 N2 1 TlCl (1) reaction also provides a novel single-pot synthesis of a methyleneamide (azavinylidene, alkylideneamide or ketimide) complex from a convenient and commercially available precursor, an oxime (see also below).The X-ray crystal structure analysis § of compound 1b shows (Fig. 1) that the methyleneamide ligand is linearly co-ordinated thus behaving as a formal three-electron donor, Re]� N]] C(27)Me2, and allowing the complex to attain the 18-electron † E-Mail: pombeiro@alfa.ist.utl.pt configuration. The significant double-bond character of the methyleneamide co-ordination bond is indicated by the Re]N distance, 1.901(5) Å, which is shorter than the average value, 2.107 Å, quoted 4 for nitrile complexes of Re, in particular those 5 with an identical co-ordination metal centre.Moreover, the methyleneamide ligand behaves as an effective p-electron acceptor, competing with the diphosphines for the available metal dp electrons, as indicated by the average Re]P distance, 2.461(2) Å, which is identical to that 6 of the related aminocarbyne complex trans-[ReCl(CNHMe)(dppe)2][BF4] (in which the CNHMe ligand is a strong p-electron acceptor) and longer Fig. 1 Molecular structure of the complex cation of trans- [Re(OH)(N]] CMe2)(dppe)2][HSO4] 1b. Selected bond distance (Å) and angles (8): Re]N 1.901(5), Re]O 2.015(4), Re]Pave 2.461(2), N]C(27) 1.251(9); Re]N]C(27) 178.9(5) ‡ The oxime (15.3 mg, 0.209 mmol) and the Tl1 salts {61 mg, ca. 0.21 mmol, mainly Tl[BF4] with a much smaller amount of Tl[HSO4] which co-precipitated with the former salt in its synthesis by reaction of Tl2SO4 with Ba(OH)2 followed by treatment with [NH4][BF4]} were added to a THF solution (150 cm3) of trans-[ReCl(N2)(dppe)2] (0.10 g, 0.095 mmol), and the products 1 were isolated, after ca. 4 h, as brick red solids which were recrystallized from CH2Cl2–Et2O (ca. 60% yield). Compound 1a is the dominant isolated product, but it can contain some co-precipitated 1b (Found: C, 58.0; H, 4.8; N, 1.1. C55H55BF4- NOP4Re 1a requires C, 57.7, H, 4.8; N, 1.2%). Selected spectroscopic data: IR (KBr, cm21): n(OH) ca. 3420s (br); n(C]] N), d(OH) ca. 1640w (br); BF4 2 ca. 1090vs and ca. 1050vs (1a). 13C NMR (CD2Cl2): d 147.11 (s, N]] CMe2), 4.90 [q, JCH = 129.4 Hz, N]] C(CH3)2]. 1H NMR (300 MHz, CD2Cl2): d 2.70 [s, 6 H, N]] C(CH3)2]. 31P-{1H} NMR (CD2Cl2): d 2128.24 [relative to P(OMe)3]. 19F NMR (CD2Cl2): d 2151.46 (s) (relative to CFCl3) (1a). § Crystal data for 1b: C55H55NO5P4ReS, M = 1152.14, triclinic, space group P1� , a = 11.998(2), b = 16.724(2), c = 13.970(2) Å, a = 86.59(1), b = 67.84(1), g = 75.46(1)8, U = 2511.0(6) Å3, Z = 2, T = 293(2) K, m(Mo-Ka) = 2.638 mm21, 6776 reflections collected, 6427 independent reflections (Rint = 0.017), final R indices (all data) R1 = 0.0339, wR2 = 0.087.CCDC reference number 186/832.326 J. Chem. Soc., Dalton Trans., 1998, Pages 325–326 than the average value, 2.428 Å, reported 4 for Re]dppe bonds, in particular those of comparable nitrile 5 or isocyanide 7 complexes. Interestingly, the Re]O distance in 1b, 2.015(4) Å, is signifi- cantly longer than the average value, 1.795 Å, quoted 4 for the Re]OH bond length, thus suggesting an appreciable structural trans influence of the methyleneamide on the hydroxide ligand, in contrast to the common behaviour observed 8 for linear methyleneamide ligands which do not exhibit an obvious lengthening effect on the trans metal–ligand bond.Possibly related to that observation is the ready replacement of the hydroxide ligand by fluoride upon treatment of a CH2Cl2 solution of compound 1a with HBF4 to give trans- [ReF(N]] CMe2)(dppe)2][BF4] 2¶ [equation (2)], although the trans-[Re(OH)(N]] CMe2)(dppe)2][BF4] 1 HBF4 æÆ trans-[ReF(N]] CMe2)(dppe)2][BF4] 1 H2O?BF3 (2) displacement promoting effect owing to the conceivable protonation of the former ligand to give a labile aqua complex should also play a relevant role.The formation of compounds 1a and 1b can be related to the interesting synthesis of complexes of the type [(h6-C6R6)- M(N]] CR9R0)(L)][PF6] [M = Os or Ru; R = H or Me; CR9R0 = CPh2, CMe(Ph), CMe2 or C(CH2)4CH2; L = organophosphine] which were obtained by Werner and co-workers 9 by reaction of the corresponding oximes HON]] CR9R0 with [(h6-C6R6)MHX(L)] (X = Cl or I) in the presence of Ag[PF6].It proceeds via the hydride–oxime intermediates [(h6-C6R6)- MH(HON]] CR9R0)(L)][PF6] which, upon subsequent dehydration when chromatographed over Al2O3, yield the final products, without changing the initial metal oxidation state. In our ReI system, the lability of two ligands (rather than a single one as in the Os or Ru complexes described above), the greater electron richness of the metal centre relative to osmiumand ruthenium-(II) sites, and the ability of the rhenium(I) centre (with a high p-electron releasing character) to form multiple bonds to unsaturated ligands promotes oxidative addition of the oxime to this centre and the preferential cleavage of the N]O bond (to give the p acceptor N]] CMe2 ligand) rather than the split of the O]H bond (which would generate the oximate group without such a p-accepting ability).In addition, the steric hindrance at our centre, with the bulky diphosphines, conceivably also plays a role, favouring the stabilization of a product with end-on cdination of a linear group (such as the methyleneamide but not the oxime nor oximate species). ¶ Complex 2 precipitated on addition of Et2O to a CH2Cl2 solution (5 cm3) of 1a (61 mg, 0.054 mmol) with [Et2OH][BF4] (0.18 mmol, 0.84 cm3 of a 1 : 25 Et2O diluted solution of commercial 85% HBF4 in Et2O), as a brick red solid (ca. 65% yield) (Found: C, 56.6; H, 4.6; N, 1.0. C55H55BF5NOP4Re requires C, 52.6; H, 4.7; N, 1.2%). Selected spectroscopic data: IR (KBr, cm21): n(N]] C) 1640w. 31P-{1H} NMR (CDCl3): d 2127.51 [relative to P(OMe)3] (d, 2JPF ª 39 Hz). 19F NMR (CDCl3): d 2196.73 (relative to CFCl3) (qt, 2JPF ª39 Hz). We have previously established 10 a different route for methyleneamide complexes based on the activation to b-protonation of a nitrile ligand by a rhenium centre, i.e.at [ReCl(NCR)- (dppe)2] to give [ReCl(N]] CHR)(dppe)2]1. In the present work the inability of the oxime (which does not have the p-accepting character possessed by nitriles) to stabilize such an electron-rich site by simple co-ordination prevents the isolation of any oxime intermediate and the reaction proceeds further to give a p-acceptor derivative, the methyleneamide ligand.This work thus extends the rare application of electron-rich metal sites to the synthesis of methyleneamide complexes, which contrasts with their common and quite different preparative procedures 8 involving medium or high oxidation state metal sites. Acknowledgements We thank Professor Vitaly Belsky (L. Ya. Karpov Physico- Chemical Institute, Moscow) for the X-ray diffraction analysis, and the financial support from JNICT (National Board for Scientific and Technological Research), the PRAXIS XXI Programme (Portugal), INVOTAN, RFBR (Russian Fund for Basic Research) and the EC Network ERBCHRXCT 940501.References 1 V. Yu. Kukushkin, D. Tudela and A. J. L. Pombeiro, Coord. Chem. Rev., 1996, 156, 333. 2 For reviews, see for example, A. J. L. Pombeiro, New J. Chem., 1997, 21, 649; 1994, 18, 16; A. J. L. Pombeiro, in Transition Metal Carbyne Complexes, ed. F. R. Kreissl, Kluwer Academic Publishers, Dordrecht, 1993, pp. 105–121. 3 A. J. Deeming, D.W. Owen and N. I. Powell, J. Organomet. Chem., 1990, 398, 299. 4 A. G. Orpen, L. Brammer, F. H. Allen, O. Kennard, D. G. Watson and R. Taylor, J. Chem. Soc., Dalton Trans., 1989, S1. 5 M. F. C. G. Silva, A. J. L. Pombeiro, A. Hills, D. L. Hughes and R. L. Richards, J. Organomet. Chem., 1991, 403, C1; A. J. L. Pombeiro, M. F. C. G. Silva, D. L. Hughes and R. L. Richards, Polyhedron, 1989, 8, 1872. 6 A. J. L. Pombeiro, M. F. N. N. Carvalho, P. B. Hitchcock and R. L. Richards, J. Chem. Soc., Dalton Trans., 1981, 1629. 7 M. F. N. N. Carvalho, M. T. Duarte, A. M. Galvão and A. J. L. Pombeiro, J. Organomet. Chem., 1994, 469, 79. 8 B. F. G. Johnson, B. L. Haymore and J. R. Dilworth, in Comprehensive Coordination Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon Press, Oxford, 1987, ch. 13.3, p. 99 and refs. therein. 9 T. Daniel, W. Knaup, M. Dziallas and H. Werner, Chem. Ber., 1993, 126, 1981; H. Werner, T. Daniel, W. Knaup and O. Nürnberg, J. Organomet. Chem., 1993, 462, 309. 10 A. J. L. Pombeiro, D. L. Hughes and R. L. Richards, J. Chem. Soc., Chem. Commun., 1988, 1052; J. J. R. Fraústo da Silva, M. F. C. Guedes da Silva, R. A. Henderson, A. J. L. Pombeiro and R. L. Richards, J. Organomet. Chem., 1993, 461, 141. Received 6th October 1997; Communication 7/07213I

ISSN:1477-9226    

DOI:10.1039/a707213i

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, Pages 327–328 327 A stacking spin-crossover iron(II) compound with a large hysteresis † Zhuang Jin Zhong,* Jian-Qing Tao, Zhi Yu, Chun-Ying Dun, Yong-Jian Liu and Xiao-Zeng You * Coordination Chemistry Institute, State Key Laboratory of Coordination Chemistry, Nanjing University, Nanjing 210093, China An inclusion compound [Fe(dpp)2(NCS)2]?py has been prepared and characterized by magnetic and Mössbauer spectral measurements (dpp = dipyrido[3,2-a:2939-c]phenazine and py = pyridine); it shows a spin-crossover behavior with a large hysteresis (DTc = 40 K).The study of the spin-crossover phenomenon is an important field of magnetochemistry. A variety of iron(II) complexes are known to show a transition from the high-spin state 5T2 (S = 2) to the low-spin state 1A1 (S = 0) on cooling, upon increasing pressure or by light irradiation.1–8 Such compounds have attracted much attention since their bistable nature could allow their use as molecular switches in new electronic devices.9,10 We are interested in investigating the main factors in spin-crossover systems that control abruptness, hysteresis and critical temperature, which are very important for designing molecular switches. Of the iron(II) systems, [Fe(phen)2(NCS)2] (phen = 1,10- phenanthroline) is one of the most throughly studied spincrossover complexes.1,11 The influence of a modification of the hydrogen atoms with electron-donating, electron-withdrawing or bulky groups,1 and by replacing the CH groups with nitrogen atoms has been investigated.12 However, to our knowledge, no study has been reported so far on the effect of adding an extended aromatic ring.Ligands with extended aromatic rings may be expected to show strong intermolecular p–p interactions to enhance the cooperativity, which has a close relevance to abruptness and thermal hysteresis.2,3,10 Along this line, we began to use ligands containing a more extended aromatic ring system than phen to study the spin-transition behavior of iron(II) complexes.Here we report the structure and magnetic properties of such a new spin-crossover system [Fe(dpp)2(NCS)2]?py (dpp = dipyrido[3,2-a:2939-c]phenazine, py = pyridine). The complex [Fe(dpp)2(NCS)2]?py was prepared by addition of a hot pyridine solution (20 cm3) of dpp13 (0.5 mmol) to a hot pyridine solution (80 cm3) of equimolar [Fe(py)4(NCS)2].14 After filtration, the filtrate was allowed to stand for about 3 months at room temperature to produce the complex as dark violet crystals.All operations were conducted under a nitrogen atmosphere (Found: C, 63.37; H, 3.32; N, 18.83. Calc. for C43H25FeN11S2: C, 63.31; H, 3.09; N, 18.89%).‡ † Non-SI unit employed: mB ª 9.274 × 10224 J T21. ‡ Crystal data: C43H25FeN11S2, M = 815.71, dark violet cubic crystal of dimension 0.23 × 0.28 × 0.34 mm, monoclinic, space group Pn, Z = 2, a = 13.381(2), b = 8.708(2), c = 16.301(6) Å, b = 104.16(3)8, U = 1841.6(9) Å3, F(000) = 836, Dc = 1.471 g cm21, m(Mo-Ka) = 0.573 mm21, T = 294 K. 4202 Reflections collected in the range 1.77 < q < 25, 3686 independent data (Rint = 0.06). The structure was solved by direct methods and refined on F2 by full-matrix least-squares methods using SHELXTL version 5.0:15 R = 0.0632, R9 = 0.0845, goodness of fit = 1.028 for 1704 observed reflections [I > 2s(I)]. CCDC reference number 186/837.The compound [Fe(dpp)2(NCS)2]?py crystallizes in the monoclinic Pn space group at room temperature. The crystal structure shows that there is one guest pyridine molecule per iron complex (Fig. 1). Each iron atom is octahedrally coordinated to four nitrogen atoms from two dpp bidentate ligands and nitrogen atoms of two NCS ions in a cis arrangement. The octahedron geometry is distorted, as Fe]N (CS) bond lengths (2.10 Å on average) are shorter than Fe]N (dpp) (Fig. 1). The Fe]N (dpp) lengths also differ considerably for different dpp ligands (2.17 and 2.23 Å on average).Moreover, the geometric constraints of the bidentate ligands cause signifi- cant reduction of the N (dpp)]Fe]N (dpp) bond angles from the ideal 908 value [N(2)]Fe(1)]N(1) 74.18, N(7)]Fe(1)]N(8) 75.58]. This distortion of the FeN6 core from Oh symmetry is also seen in the high-spin phase structure of similar [FeL2- (NCS)2] (L = phen, bipyridine or 1,4,5,8-tetraazaphenanthrene) complexes.11,12,16 The NCS groups are almost linear [N(5)]C(19)]S(1) 173.78], N(6)]C(20)]S(2) 178.08], whereas the Fe]NC(S) linkages are bent.The two dpp ligands in Fig. 1 are nearly planar, and the largest deviations from the mean planes [C(1)]C(12), N(1)]N(4)] and [C(21)]C(32), N(7)]N(10)] are 0.0378 Å and 0.0292 Å respectively. The dihedral angle between these two planes is 73.08. Between the nearest [Fe(dpp)2(NCS)2] molecules there is a p–p stacking of their dpp ligands with an interplanar distance of 3.50 Å and a dihedral angle of 0.68, thus forming a column structure along the diagonal of the unit cell (Fig. 2). The solvent pyridine molecule inserts between these columns, with two sets of overlapping dpp ligands around it. The shortest distances from the center of py to the surrounding atoms are 3.452 Å [to C(89), symmetry transformation 2��� 1 x, 1 2 y, 2��� 1 z] and 3.502 Å [to C(289), symmetry transformation 2��� 1 x, 1 2 y, ��� 1 z], suggesting a close contact.The dihedral angles between py and the dpp ligands are about 768. The temperature dependence of the effective magnetic moment meff is illustrated in Fig. 3, showing a S = 2 (HS) S = 0 (LS) spin-crossover behavior. Fairly abrupt transitions are observed around 123 K in the cooling curve and 163 K in the heating curve, leading to a well shaped hysteresis loop of about 40 K (DTc). This hysteresis feature can be repeated for more than one cycle. The moment at 275 K (5.2 mB) is close to the quintet state, while the value at 86 K (1.6 mB) is considerably higher than the pure singlet state.It could be due to the presence of some FeIII impurities, which were detected by an EPR signal at liquid-nitrogen temperature. The Mössbauer spectra agree with the magnetic curve. The results show that FeII ions in [Fe(dpp)2(NCS)2]?py at room temperature are in the high-spin state (DEq = 2.655 mm s21, IS = 0.828 mm s21, relative to 57Co– Pd) whereas at 80 K they are in the low-spin state (DEq = 0.423 mm s21, IS = 0.268 mm s21, relative to 57Co–Pd).A small amount of FeIII (high spin) impurity is present at ca. 6 mol% as deduced from analysis of the Mössbauer spectra, an amount which is entirely consistent with the observed effective magnetic moment at 80 K. Note that no [FeL2(NCS)2]-type spincrossover compounds have yet shown DTc values larger than 30 K, making [Fe(dpp)2(NCS)2]?py a novel example. Until now,328 J. Chem.Soc., Dalton Trans., 1998, Pages 327–328 Fig. 1 Molecular structure of [Fe(dpp)2(NCS)2]?py. Selected bond lengths (Å) and angles (8): Fe(1)]N(6) 2.090(4), Fe(1)]N(5) 2.104(3), Fe(1)]N(2) 2.165(3), Fe(1)]N(1) 2.186(3), Fe(1)]N(7) 2.227(4), Fe(1)]N(8) 2.237(3), S(1)]C(19) 1.636(4), S(2)]C(20) 1.581(4), N(5)]C(19) 1.114(5), N(6)]C(20) 1.175(5); N(6)]Fe(1)]N(5) 96.4(1), N(6)]Fe(1)]N(2) 99.0(1), N(5)]Fe(1)]N(2) 93.1(1), N(6)]Fe(1)]N(1) 86.1(2), N(5)]Fe(1)]N(1) 167.2(1), N(2)]Fe(1)]N(1) 74.1(1), N(6)]Fe(1)]N(7) 165.8(1), N(5)]Fe(1)]N(7) 87.6(1), N(2)]Fe(1)]N(7) 94.5(1), N(1)]Fe(1)]N(7) 93.1(1), N(6)]Fe(1)]N(8) 90.3(1), N(5)]Fe(1)]N(8) 100.9(1), N(2)]Fe(1)]N(8) 162.3(1), N(1)]Fe(1)]N(8) 91.6(1), N(7)]Fe(1)]N(8) 75.5(1), C(19)]N(5)]Fe(1) 154.3(3), C(20)]N(6)]Fe(1) 149.1(4), N(5)]C(19)]S(1) 173.7(4), N(6)]C(20)]S(2) 178.0(5) the spin-crossover systems with a large hysteresis have been found to contain communication networks of hydrogen bonds or covalent bridges.1,2,9,17 In our system, only intracolumn p–p stacking and intercolumn van der Waals interactions exist, suggesting that these kinds of supramolecular interactions can also be effective in constructing cooperative spin-crossover systems. Further studies are in progress.Fig. 2 Crystal packingFig. 3 Temperature dependence of meff versus T for [Fe(dpp)2(NCS)2]?py Acknowledgements This work was supported by grants from the State Science and Technology Commission and the National Nature Science Foundation of China.References 1 P. Gütlich, Struct. Bonding (Berlin), 1981, 44, 83. 2 E. König, G. Ritter and S. K. Kulshreshtha, Chem. Rev., 1985, 85, 219. 3 P. Gütlich, A. Hauser and H. Spiering, Angew. Chem., Int. Ed. Engl., 1994, 33, 2024. 4 D. Boinnard, A. Bousseksou, A. Dworkin, J. M. Savariault, F. Varret and J. P. Tuchagues, Inorg. Chem., 1994, 33, 271. 5 Z. Yu, N. Boris, G. Schmitt, H. Spiering and P. Gütlich, J. Phys.: Condens.Matter, 1995, 7, 777. 6 J. K. McCusker, A. L. Rheingold and D. N. Hendrickson, Inorg. Chem., 1996, 35, 2100. 7 M.-L. Boillot, C. Roux, J.-P. Audiere, A. Dausse and J. Zarembowitch, Inorg. Chem., 1996, 35, 3975. 8 P. J. Kunkeler, P. J. van Koningsbruggen, J. P. Cornelissen, A. N. van der Horst, A. M. van der Kraan, A. L. Spek, J. G. Haasnoot and J. Reedijk, J. Am. Chem. Soc., 1996, 118, 2190. 9 J. Krüber, E. Codjovi, O. Kahn, F. Groliere, J. Charlotte and C. Jay, J. Am. Chem. Soc., 1993, 115, 9810. 10 J. A. Real, E. Andréz, M. C. Muñoz, M. Julve, T. Granier, A. Bousseksou and F. Varret, Science, 1995, 268, 265. 11 B. Gallois, J. A. Real, C. Hauw and J. Zarembowitch, Inorg. Chem., 1990, 29, 1152. 12 J. A. Real, M. C. Muñoz, E. Andréz, T. Granier and B. Gallois, Inorg. Chem., 1994, 33, 3587. 13 R. D. Gillard, R. E. E. Hill and R. Maskill, J. Chem. Soc. A, 1970, 1447. 14 N. E. Erickson and N. Sutin, Inorg. Chem., 1966, 5, 1834. 15 Siemens, SHELXTL Version 5.0, Siemens Industrial Automation Inc., Analytical Instrumentation, Madison, WI, 1995. 16 M. Konno and M. Mikami-Kido, Bull. Chem. Soc. Jpn., 1991, 64, 339. 17 W. Vreugdenhil, J. H. van Diemen, R. A. G. de Graaff, J. G. Haasnoot, J. Reedijk, A. M. van der Kraan, O. Kahn and J. Zarembowitch, Polyhedron, 1990, 9, 2971. Received 22nd September 1997; Communication 7/06841G

ISSN:1477-9226    

DOI:10.1039/a706841g

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 329–344 329 Crystal supramolecularity: hexagonal arrays of sextuple phenyl embraces amongst chemically diverse compounds Marcia Scudder and Ian Dance * School of Chemistry, University of New South Wales, Sydney, NSW 2052, Australia The sextuple phenyl embrace (SPE) occurs widely between molecules with XR3 (R = aryl) moieties, and has at least quasi-three-fold symmetry. The SPE can have exact 3� symmetry, and therefore can be aligned with the principal axis of trigonal crystal lattices. Using the Cambridge Structural Database we report that this does occur in the crystals of a chemically diverse set of compounds, ranging from P(C6H4Me-4)3 and Ph3CCO2H, through salts such as [Ph3PH1]2[Ga2Cl6]22 and [Ph3PMe1]2[Cu4I6]22, small molecules such as Ph3PAlMe3 and Ph3POs(CO)3PPh3, to larger molecules such as Ph3SiOTiN(CH2CH2O)3 and Ph3PCu(m-SPh)3U(m-SPh)3- CuPPh3(thf)6 (thf = tetrahydrofuran).The XR3 moieties and the SPEs occur in hexagonal nets, and this generic crystal structure type (in space groups R3� , P3� , R3� c, P3� c1) is named the hexagonal array of sextuple phenyl embraces, HASPE.The hexagonal nets can be planar, or puckered by expansion or compression along the trigonal axis. The linkages around these hexagonal nets can be further multiple phenyl embraces, or may be elongated substantially. The non-embracing sections of the molecules can occupy the centres of the hexagons.The HASPE lattice type shows considerable flexibility in order to accommodate diverse components, but the integrity of the SPEs is maintained, attesting to their strong contribution to the lattice energy. Lower symmetry portions of molecular structure, and lower symmetry supramolecular motifs such as the CO2H? ? ?HO2C dimer, are forced to disorder by the dominant hexagonal array of SPEs. Quasi HASPE lattices with lower crystal symmetry have been recognised. In previous papers 1–4 we have identified and described multiple phenyl embraces, which are concerted supramolecular motifs maintained by phenyl–phenyl attractive interactions.These multiple phenyl embraces are frequently dominant factors in crystal packing, and are recognisable by investigation through the Cambridge Structural Database (CSD).5 The dominant multiple phenyl embrace (MPE) is the sextuple phenyl embrace (SPE), structure I, which occurs frequently in crystals of compounds with terminal Ph3P ligands, and in crystals containing the Ph4P1 cation.In the supramolecular domain of the SPE, three phenyl rings on one molecule are arrayed between three phenyl rings on the other molecule, such that each ring has edge-to-face (ef) interactions with two rings of the other molecule. Each phenyl ring projects two H atoms towards C atoms of a ring across the domain, and the coulombic attraction of Hd1 ? ? ?Cd2 provides the directionality and some of the attractive energy of the embrace.The SPE comprises six such ef interactions in a cyclic sequence, as shown by the arrows in structure I. This SPE can have 3� symmetry.4 Another multiple phenyl embrace engaged by Ph3P and Ph4P1 groups is the quadruple phenyl embrace (QPE) involving two phenyl groups on each partner, with two geometrical subclasses depending on whether the CipsoPCipso planes on each molecule are approximately parallel (the PQPE) or orthogonal (the OQPE).3 The PQPE comprises one offset-face-to-face (off) interaction between a pair of phenyl rings, and two vertex-toface (vf) interactions, while the OQPE has four ef interactions.The double phenyl embrace (DPE) involves an off interaction between one phenyl group on each molecule. We recently reported 4 that the crystal structure of [MePh3P]2- [CdBr4] contains pairs of MePh3P1 cations in sextuple phenyl embrace, and that these SPE are further organised in puckered hexagonal layers such that the trigonal lattice is effectively a pseudo-diamondoid array of SPE.The anions are contained in cavities surrounded by twelve cations. We named this lattice type the hexagonal array of sextuple phenyl embraces (HASPE) * E-Mail: I.Dance@unsw.edu.au and noted that it could vary its dimensions in order to accommodate other anions [CdI4]22 and [Cu4I6]22, and also accommodate solvent CH2Cl2 with [CdBr4]22. The three-fold symmetry of MePh3P1 cations allows the HASPE lattice to have exact three-fold symmetry (which can enforce disorder of the associated anion or solvent 4).We wondered whether other molecules containing XPh3 (or indeed XR3, R = aryl) moieties and with potential three-fold symmetry could also crystallise with the HASPE lattice, and accordingly we interrogated the Cambridge Structural Database. We find that the HASPE lattice is prevalent amongst a variety of chemical identities and molecular structures. The HASPE lattice structure can contain other supramolecular interactions in330 J.Chem. Soc., Dalton Trans., 1998, Pages 329–344 addition to the SPEs, and occurs with neutral molecules as well as with cation–anion combinations. Further, the HASPE lattice can be accommodated in a variety of trigonal space groups, with a range of lattice dimensions.† It is evident that the HASPE lattice is a significant motif in crystal supramolecularity, and in this paper we describe and analyse the instances we have dis- Fig. 1 The array of X atoms in a typical HASPE lattice.Each of the pink vertical rods represents an SPE, and the blue rods show the hexagonal lattice of X atoms. Every X atom is involved in one SPE Fig. 2 Hexagonal lattices of SPEs (pink vertical rods) which are expanded (a) or compressed (b) along the trigonal axis, causing puckering of the hexagonal network of X atoms † This variability precludes recognition of crystal supramolecular motifs when crystallographic isomorphism alone is considered.covered. We are concerned also with exceptions to the HASPE lattice, for their revelations about its potency. Generic Description of the HASPE Lattice We first outline the principal characteristics of the HASPE lattice of XR3 moieties. The predominant feature is the alignment of all SPE with the trigonal axis of a hexagonal or rhombohedral crystal lattice. Fig. 1 demonstrates the hexagonal array of X atoms in a typical HASPE and the SPEs which link them.Each X atom is involved in only one SPE, and as a consequence repetition along the hexagonal axis can occur only at the fourth layer, and stacking of the layers is . . . ABCABC . . . and cannot be . . . ABAB . . . The hexagonal layers of X atoms need not be planar as in Fig. 1, but can be puckered by elongation or compression along the hexagonal axis, as illustrated in Fig. 2. This variability of placement of the SPEs along the trigonal axis is related to the volumes and shapes of the molecules comprising the HASPE lattice, and the volumes of any associated species such as counter ions.The HASPE lattices can also be described in terms of the array of centroids of the SPEs: we use the symbol Ç for the centroid of an SPE, which in all cases except one is a centre of inversion in the lattice because the SPEs have exact 3� symmetry. The space group of the array of centroids is included in our tabulations of HASPE structures. The space groups encompassed by the HASPE lattices are both hexagonal and rhombohedral, and in Fig. 3 we remind the reader of the relationships between a primitive hexagonal lattice, a primitive rhombohedral lattice, and the description of a rhombohedral lattice with hexagonal axes as is common practice. Occurrence of the HASPE Lattice Our source of data is the graphical version of the Cambridge Structural Database (April 1997), which was screened first by selection of rhombohedral or hexagonal structures, then by occurrence of the phrase ‘triphenyl’ in the compound name.The lattice packing of each the 144 hits was then displayed and assessed visually for the occurrence of SPE interactions. We Fig. 3 Unit-cell relationships for HASPE lattices. (a) The primitive cell of space group P3� . (b) The rhombohedral cell and the hexagonal cell of space group R3� : the hexagonal setting contains three times as many lattice poindral settingJ. Chem. Soc., Dalton Trans., 1998, Pages 329–344 331 recognise that this is not an exhaustive search protocol, and we have discovered other instances of HASPE lattices in general exploration of the structural database.Also, we describe below some de facto HASPE lattices which have lower crystallographic symmetry, and it is possible that there are many others. In the following, crystal structures are identified by their CSD reference codes, and the literature citations for these refcodes are listed at the end of the paper.The molecular structures of the compounds revealed by this search include XR3 moieties more elaborate than XPh3, and can be grouped in four classes. The simplest molecular structure A is just XR3; a larger class has a single atom or small group Y with three-fold symmetry attached to X (molecular structure B); a third class has a molecular structure C in which a linear spine connects XR3 to another more distant possibly larger group with three-fold symmetry; the fourth class (D) has two XR3 ends linked through a collection of atoms which also has three-fold symmetry.We describe the HASPE lattices by tabulating crystallographic details and presenting diagrammatic pictures of representative structures. HASPE Lattices with Simple Molecules of Type A or B Small simple molecules which crystallise with the HASPE lattice in space group R3� are described in Table 1. There are four instances of a type A molecule, As(C6H4OCH3-4)3 and E(C6H4CH3-4)3 (E = P, As or Sb), which crystallise with virtually flat hexagonal nets, with good quadruple aryl embraces along all connections in the hexagonal net as well as the SPE between the nets.The 4-substituents do not interfere. The other four compounds in Table 1 have a different HASPE in space group P3� , which is illustrated in Fig. 4. There are two types of hexagonal net, one relatively flat and one puckered, with two flat nets and one puckered net in the cell repeat.There are two different SPEs, one between flatter nets and the other between puckered and flatter nets: the former SPE is shorter (C ? ? ?C 5.91 Å in Ph3CBr, TPHMBR02) and has crystallographic 3� symmetry, while the longer SPE at 6.24 Å has crystallographic symmetry 3 with a quasi inversion centre. The C ? ? ? C distances within the flatter net are 8.13 Å, and in the puckered net are 8.44 Å. The Br atoms which are distal to each SPE are directed towards the centre of the contiguous hexagon [see Fig. 4(b)], and along the c axis point towards another Br atom, at a distance of 3.19 or 3.48 Å, within the sum of the van der Waals radii. As is partly evident from Fig. 4(b) there are QPE interactions along the 8.13 and 8.44 Å vectors of the hexagonal nets: the geometry of the longer interaction is actually a better QPE. The compound Ph3CCO2H [ZZZVTS01] is noteworthy, because its carboxyl group cannot adhere to the three-fold symmetry of the molecular site, but this does not disrupt the HASPE which imposes rotational disordering of the carboxyl group about the three-fold axis: further, the normal hydrogenbonded dimerisation of carboxyl groups occurs along the crystallographic c axis, through the centre of a hexagon.HASPE Lattices with Molecules of Type B A set of compounds with triphenylphosphonium cations (molecular structure type B) and various dianions also adopt the HASPE lattice. These compounds are described in Table 2.The HASPE lattice of [Ph3PH1]2 [Ga2Cl6]22 [FUPSEO] is portrayed in Fig. 5; the [Ga2Br6]22 and [Ga2I6]22 homologs are isostructural. There are short and tight SPEs which project alternately from either side of a virtually planar regular hexagonal network of P atoms. The anions, with a Ga]Ga bond and in staggered conformation, are positioned on three-fold axes at the centres of the hexagons, and therefore the H atoms of the phosphonium cations which are directed along three-fold axes are close to a triangle of Cl atoms at the end of an anion.Along each crystallographic three-fold axis the sequence is . . . 1HPPh3? SPE?Ph3PH1 ? ?2[Cl3GaGaCl3]2 ? ?1HPPh3?SPE?Ph3PH1 ? ? 2[Cl3GaGaCl3]2. Along the edges of the hexagons (7.99 Å) there are well developed double phenyl embraces [see Fig. 5(b)]. The data in Table 2 show that as the size of the anion increases from [Ga2Cl6]22 to [Ga2I6]22 the length of the SPE is unchanged but the DPE expands to enlarge the hexagon.Two compounds with Ph3PCl1 cations and [MCl6]22 anions (M = Sn or Mo; see Table 2) demonstrate HASPE lattices very similar to that shown in Fig. 5. The SPE (P ? ? ? P 5.6 Å) and DPE (P ? ? ? P 7.6 Å) are short and tight. Note that in this crystal structure a Cl atom of the phosphonium cation is adjacent to a triangle of Cl atoms from the anion: this Cl ? ? ? Cl distance is 3.8 Å in both compounds. There are four compounds with Ph3PMe1 cations and dianions listed in Table 2, all in space group R3� c.One of these, with [Cu4I6]22, has virtually planar hexagonal nets like the HASPE structures just described, while the other three have strongly puckered hexagonal nets, elongated along the hexagonal axis. We have described and analysed the details of these crystal structures previously:4 in the present context there are two important points to make. One is that the HASPE is elongated (in the [CdBr4]22 and [CdI4]22 compounds) or not (in the [Cu4I6]22 compound) in order to allow for the volume of the anion, which is prolate along the three-fold axis for the Cd anions and oblate for the Cu anion (all the anions being orientationally disordered).The second point is that the change in space group from R3� to R3� c is a consequence of the intimate relationship between the anions and the surrounding six cations in each hexagon, restricting rotational conformation of the Ph3PMe1 cations, and doubling the c repeat distance (see Table 2).Also included in Table 2 are three compounds which form the HASPE lattice and contain two neutral species, JACWAL (JACWEP) contains Ph3PO (Ph3AsO) and toluene-psulfonamide in the ratio of 2 : 3. The SPEs formed here have P ? ? ? P 6.47 Å but around the nearly planar hexagon there are only weak interactions consisting of single edge-to-face interactions. Distal to the SPE, where two O atoms of Ph3PO are directed towards each other, each of three toluene-psulfonamide molecules bridges the two O atoms with hydrogen bonds from NH2 (see below).The complex with refcode BTZANI contains Ph3PO and hexakis(benzotriazolyl)hexakis(vinylamine)trinickel(II). The SPEs are slightly longer than usual, at P ? ? ? P 7.08 Å and the hexagonal net, which has P ? ? ? P distances of 10.12 Å is slightly N H H O O P P 3332 J. Chem. Soc., Dalton Trans., 1998, Pages 329–344 Table 1 HASPE lattices formed by small simple molecules Refcode a BADBIRg CANTEQ TOLARS01 TPTOSB JICWUN TPHMBR02h ZZZVTY12 ZZZVTS01i Compound As(C6H4OCH3-4)3 P(C6H4CH3-4)3 As(C6H4CH3-4)3 Sb(C6H4CH3-4)3 (C6F5)3BPH3 Ph3CBr Ph3CCl Ph3CCO2H Molecular structure class A A A A B B B B Space group R3� R3� R3� R3� P3� P3� P3� P3� Cell b axes a,c/Å 13.3, 18.9 12.6, 19.7 12.7, 19.8 12.8, 20.1 15.1, 14.6 13.9, 13.4 14.0, 13.2 14.1, 13.1 Crystallographic locations of X atomsc As, c P, c As, c Sb, c B, c,d,d C, c,d,d C, c,d,d C, c,d,d Stacking of hexagonal nets d All layers identical All layers identical All layers identical All layers identical Repeat: f,p,f Repeat: f,p,f Repeat: f,p,f Repeat: f,p,f SPE Length X? ? ? X/Å 7.47 6.97 7.22 7.58 6.19, 6.40 5.91, 6.24 5.87, 6.30 6.16, 6.24 Hexagonal X? ? ?Xe/Å 7.76 7.31 7.36 7.45 8.76 8.99 8.13 8.44 8.18 8.52 8.28 8.54 Angle at Xd,f/8 117.8 f 119.7 f 119.3 f 118.6 f 118.6 f 113.9 p 117.9 f 111.3 p 117.6 f 110.5 p 116.9 f 111.3 p Aryl embraces in hexagonal net Good QPE Good QPE Good QPE Poor QPE QPE a Structure identifier used by the Cambridge Crystallographic Database.b Rhombohedral unit cells are described in the hexagonal axial setting. c Small italic letters identify the crystallographic special positions with three-fold local symmetry occupied by atoms X, according to ref. 6. d f = Flat hexagonal net; p = puckered hexagonalithin the hexagonal net. f X? ? ?X? ? ? X Angles within the hexagonal net. g Space group for Ç R3� m; OCH3 occupies space between backs of SPEs.h Space group for Ç P3� m1. i CO2H Group disordered, H bonding.J. Chem. Soc., Dalton Trans., 1998, Pages 329–344 333 Fig. 4 Representations of the HASPE lattice in the crystal structure of Ph3CBr [TPHMBR02], with the unit cell outlined (c vertical). (a) The net of supramolecular connections between the central C atoms, with all Br atoms included (orange). Each of the pink vertical rods is an SPE. The relatively flat and the puckered hexagonal nets are evident, as is the occurrence of two relatively flat nets and one puckered net in the repeat.Carbon atoms (green) for rings which form the two crystallographically independent SPE interactions are included. Note that all Br atoms (orange) point towards the centre of an adjacent hexagon. The blue rods signify the QPE interactions around the hexagons. (b) Top (c axis projection) and side views of Ph3CBr molecules in a puckered hexagon (C green, Br orange) with the Br atoms of two adjacent molecules directed towards the centre of the hexagon: the Br ? ? ? Br distance is 3.48 Å puckered.The large size of the Ni cluster pushes the Ph3PO molecules apart, so that there are no interactions around the hexagonal net. Instead, however, there are interactions between the Ph3PO molecules and the aromatic rings of the Ni cluster. The phenyl ring of Ph3PO which is directed towards the cluster takes part in one good ef interaction with one benzo334 J.Chem. Soc., Dalton Trans., 1998, Pages 329–344 Fig. 5 (a) The HASPE lattice of [Ph3PH1]2 [Ga2Cl6]22 [FUPSEO]. The SPE are represented as pink rods, and the almost planar hexagonal nets by blue rods. The space group is R3� , with the hexagonal unit cell (with three times the contents of the primitive rhombohedral cell) outlined. Phenyl rings are marked for six cations in three SPE: the cation H atoms are omitted. All anions [Ga2Cl6]22 (with a Ga]Ga bond and Cl atoms in staggered conformation) are included (Ga fawn, Cl orange).(b) Three-fold and side views of one hexagon of Ph3PH1 cations centered by [Ga2Cl6]22. The DPE between adjacent cations are evident in the three-fold view: the distance between the centroids of the two phenyl rings of the DPE is 3.86 Å HN N N Ni Ni N H2 N CH H2C N NH Ni NH 2 HC CH2 3 3 3 3 ring and a second somewhat distorted off interaction with another. HASPE Lattices with Molecules of Type C Table 3 provides details for HASPE lattices adopted by compounds in which XR3 is at one end of a linear molecule, in some cases with a relatively large group at the other end.Fig. 6 shows the HASPE lattice in a representative structure in this class, namely Ph3POGaCl3. The space group is R3� and the hexagonal net is almost planar. In this regard this structure is very similarJ. Chem. Soc., Dalton Trans., 1998, Pages 329–344 335 Table 2 HASPE lattices formed by triphenylphosphonium salts and triphenylphosphine oxide Refcode a FUPSEO FUPSIS FUPSOY MPPICU MTPHCI Ref. 4 Ref. 4 FUTPIT PINSOU JACWALf JACWEPf BTZANI Compound [Ph3PH1]2[Ga2Cl6]22 [Ph3PH1]2[Ga2Br6]22 [Ph3PH1]2[Ga2I6]22 [Ph3PMe1]2[Cu4I6]22 [Ph3PMe1]2[CdI4]22 [Ph3PMe1]2[CdBr4]22 [Ph3PMe1]2[CdBr4]22 1 CH2Cl2 [Ph3PCl1]2[MoCl6]22 [Ph3PCl1]2[SnCl6]22 [Ph3PO]2[H2NSO2C6H4CH3-4]3 [Ph3AsO]2[H2NSO2C6H4CH3-4)3 [Ph3PO][Ni cluster] g Space group R3� R3� R3� R3� c R3� c R3� c R3� c R3� R3� R3� R3� R3� c Cell b axes a,c/Å 13.8, 17.6 14.1, 17.6 14.5, 18.3 14.0, 40.1 11.0, 64.0 10.9, 59.2 10.8, 62.4 13.2, 18.9 13.2, 18.9 18.7, 27.5 18.8, 27.4 17.4, 50.5 SPE length X? ? ? X/Å 6.23 6.20 6.24 6.22 6.74 6.39 6.78 5.62 5.59 6.47 7.08 Hexagonal X? ? ?Xc/Å 7.99 8.13 8.39 8.07 7.45 7.18 7.22 7.63 7.63 10.93 10.12 Angle at Xd,e/8 119.8 f 119.9 f 120.0 f 119.7 f 94.9 p 98.4 p 97.1 p 119.2 f 119.2 f 117.1 f 118.3 sl p Aryl embraces in hexagonal net Good DPE Good DPE Good DPE ef DPE Space group for X or Ç only X R3� m Ç R3� m X R3� m Ç R3� m, c9=c/2 Ç R3� m a Structure identifier used by the Cambridge Crystallographic Database.b Rhombohedral unit cells are described in the hexagonal axial setting. c X? ? ? X Distances within the hexagonal net. d X? ? ?X? ? ? X Angles within the hexagonal net. e f = Flat hexagonal net; p = puckered hexagonal net; sl = slightly. f These two structures are included in the data base with the incorrect space group R3, instead of R3� . There are no atomic coordinates for JACWEP but it is isomorphous and isostructural with JACWAL.g See text for structural formula of cluster. Table 3 HASPE lattices formed by molecules of type C, with XR3 at one end of a linear molecule Refcode a FUFVAD KIFSOH KIFSUN PIHGIW JEJLEP JEJLAL JEJLIT WAWKAG WEKBIX BAYFUC CINSIB CIWZUD KOMMOO WEWLEP FUTTAP Compound (C6F5)3Ge]GeEt3 Ph3P]AlMe3 (2-CH3C6H4)3P]AlMe3 Ph3As]GaI3 Ph3P]] O]GaCl3 Ph3P]] O]AlCl3 Ph3P]] O]AlBr3 Ph3P]] O]AlMe3 Ph3P]] N]] TiCl3 Ph3P]Au]V(CO)6 Ph3P]Au]C(CN)Cl2 Ph3Si]O]Si(CH]] CH2)3 Ph3Si]O]TiN(CH2CH2O)3 Ph3P]Au]Tc(NC6H3Pri 2-2,6)3 Ph3As]I]I 1 1.5 C6H5Meg Space group R3� R3� f R3� c f R3� R3� R3� R3� R3� R3� R3� R3� R3� R3� R3� R3� c Cell b axes a,c/Å 12.2, 30.4 14.3, 16.5 15.0, 34.2 14.8, 16.6 13.8, 18.4 13.7, 18.3 14.0, 18.4 13.8, 18.2 13.7, 18.5 13.6, 23.6 12.8, 21.1 14.8, 18.0 9.6, 45.4 14.9, 39.6 13.3, 53.9 SPE length X? ? ? X/Å 7.34 6.89 6.48 6.62 6.42 6.36 6.36 6.45 6.38 7.35 6.91 6.40 6.91 6.49 6.59 Hexagonal X? ? ?Xc/Å 7.59 8.40 8.68 8.64 7.95 7.89 8.10 8.00 7.92 7.85 7.38 8.53 9.57 10.93 8.05 Angle at Xd,e/8 107.2 p 117.3 f 119.2 f 118.4 f 119.9 f 119.9 f 119.9 f 119.8 f 119.9 f 119.6 f 120.0 f 119.8 f 120.0 f 86.2 p 111.5 p Aryl embraces in hexagonal net None ODPE ODPE DPE DPE DPE DPE Good DPE Good DPE Good DPE ODPE Network of 1/2 interactions (see text) Poor DPE Space group for X or Ç only X R3� m Ç R3� m X R3� m Ç R3� m Ç R3� m c9 = c/2 a Structure identifier used by the Cambridge Crystallographic Database.b Rhombohedral unit cells are described in the hexagonal axial setting. c X? ? ? X Distances within the hexagonal net. d X? ? ?X? ? ? X Angles within the hexagonal net. e f = Flat hexagonal net; p = puckered hexagonal net. f Structures reported in both the publication and in the CSD in non-centrosymmetric space group, which gives half empty lattice. g Toluene disordered. to that shown in Fig. 5, but the chemical difference is that the trihalogeno unit is part of the embracing molecule in the lattice rather than a separate entity as in Fig. 5, and the trihalogeno unit points towards the centre of a hexagon rather than away from it. Adjacent GaCl3 ends of the molecule are directed towards each other, in staggered intermolecular conformation, as shown in Fig. 6(b). Other similar molecules have the same structure (Table 3), which occurs also with the AlMe3 at the end of the molecule. The linearity at O in these molecules is unusual, and the original paper 7 alludes to possible disordering at this atom, in which case the HASPE lattice would be the overriding influence.There have been reports of cell dimensions for polymorphs of Ph3POGaCl3 in both orthorhombic and monoclinic space groups.7 Very similar HASPE lattices with almost planar hexagons occur for the chemically different compounds Ph3P]Au] V(CO)6 (in which the V co-ordination is octahedral distorted by the seventh Au ligand along the three-fold axis), and in Ph3As]GaI3 and Ph3Si]O]Si(CH]] CH2)3 where the DPE along the hexagon edge is longer and more offset.The compound Ph3P]Au]Tc(NC6H3Pri 2-2,6)3 has a much larger but three-fold entity at the end of the molecule, which requires that the hexagonal layers be severely puckered. The compound Ph3Si]O]TiN(CH2CH2O)3 [KOMMOO] is terminated with a more compact three-fold entity, and adopts a HASPE lattice which is unique because all of the SPEs project from the same side of the hexagonal net (which is planar).This is illustrated in Fig. 7. All phenyl groups occur in the same plane (i.e. their centroids are coplanar) and within this plane in this plane has intermolecular interactions with four others around it. In this lattice there is complete segregation of phenyl–phenyl and alkyl–alkyl intermolecular interactions, and we regard this crystal lattice as being particularly informative about crystal supramolecularity.336 J.Chem. Soc., Dalton Trans., 1998, Pages 329–344 Fig. 6 The HASPE lattice structure of Ph3POGaCl3 [JEJLEP]. (a) The overall lattice, space group R3� , including details (green C atoms) of one SPE (pink rods) at a 3� site. O atoms are red, Ga fawn, Cl orange. (b) Three-fold and side views of the hexagon of molecules and two adjacent molecules with their GaCl3 ends directed into the hexagon.The six DPE interactions around the edges of the hexagon are evident HASPE Lattices with Molecules of Type D, with XR3 at Both Ends There is a class of molecules (Table 4) with three-fold symmetry and XR3 groups at both ends. The lattice structures have similarities with those already described, but now the molecular sections distal to the XR3 and the SPE are connected rather than simply pointing towards each other. Fig. 8 shows this structure for a relatively small molecule, Ph3P]Os(CO)3]PPh3 [COSPHP], in which there is no symmetry operation relatingJ.Chem. Soc., Dalton Trans., 1998, Pages 329–344 337 Fig. 7 Three-fold and side views of a centred hexagon on the HASPE lattice structure of Ph3Si]O]TiN(CH2CH2O)3 [KOMMOO]. Note that all of the Si atoms are in the same plane, and that the SPEs (not shown in full) project from the same side of this plane. Note also that all phenyl groups are in the one plane, within which each phenyl group has intermolecular attractive interactions with four other phenyl rings from three different molecules Fig. 8 Three-fold and side views of part of the crystal structure of Ph3P]Os(CO)3]PPh3 the two ends of the molecule. Both ends, however, take part in SPE. The space group now has a c glide and an elongated c axis. The larger complex Ph3P]Cu(m-SC6H4CH3-4)3]Mo]Cu- (m-SC6H4CH3-4)3]PPh3 [SUDYUL] with octahedrally coordinated Mo at the centre forms a similar HASPE lattice. The 4-substituent leads to separation of molecules in the ab plane to leave triangular cavities which are occupied by disordered tetrahydrofuran (thf) molecules.This is shown in Fig. 9. The complex Ph3P]Cu(m-SPh)3]U]Cu(m-SPh)3]PPh3 [YIRHOW] is similar again, except that the molecules are pushed apart around the hexagonal net to accommodate disordered solvent in between. Aryl–aryl interactions around the hexagonal net are consequently precluded. The complex Ph3Si]N]] C]] N]SiPh3 [TPSICI] is a significant molecule because the central stem is narrow (see Fig. 10). This crystal structure has many phenyl embraces in addition to the338 J. Chem. Soc., Dalton Trans., 1998, Pages 329–344 Table 4 HASPE lattices formed by molecules of type D, with XR3 at both ends of a linear molecule Refcode a COSPHP DNTPIR SUDYUL YIRHOW TPSICI LIBHIN SOMMOW SOMMUC SOMNAJ Compound Ph3P]Os(CO)3]Ph3P Ph3P]IrH(NO3)2]Ph3P Ph3P]Cu(m-SC6H4CH3-4)3]Mo] Cu(m-SC6H4CH3-4)3]PPh3 1 9thf Ph3P]Cu](m-SPh)3]U](m-SPh)3] Cu]PPh3 1 6thf Ph3Si]N]] C]] N]SiPh3 (4-CH3C6H4)3Pb]Ge(C6H4CH3-4)3 (4-CH3C6H4)3Sn]Sn(C6H4CH3-4)3 (4-CH3C6H4)3Pb]Pb(C6H4CH3-4)3 (4-CH3C6H4)3Pb]Sn(C6H4CH3-4)3 Space group P3� c1 P3� c1 R3� P3� R3� R3� R3� R3� R3� Cell b axes a,c/Å 15.8, 23.2 16.3, 22.8 24.1, 16.6 20.7, 17.6 15.0, 48.0 13.3, 36.6 13.3, 36.8 13.2, 37.4 13.3, 37.2 Crystallographic locations of X atomsc P on c,d,d 3 P on a,b,b 3 P on 3 P on c,d,d 3 Si on 3,3,3,3 Pb/Geg on 3 Sn on 3 Pb on 3 Pb/Sng on 3 SPE length X? ? ? X/Å 6.83, 6.97 6.74 6.82 7.03 5.81, 6.74 6.40 6.37 6.54 6.50 Hexagonal X? ? ?Xd/Å 9.14 9.47 13.99 11.99 8.78 9.63 9.67 9.66 9.69 Angle at Xe, f/8 119.6 f 119.1 119.2 119.4 118.0 sl p 87.7 v p 86.9 v p 86.3 v p 86.6 v p Aryl embraces in hexagonal net Ph ? ? Ph2 Weak Ph ? ? Ph2 None None PQPE h h h h Space group for Ç only Ç P3� m1 c9 = c/2 Ç R3� m Ç R3� m a Structure identifier used by the Cambridge Crystallographic Database.b Rhombohedral unit cells are described in the hexagonal axial setting.c Small italic letters identify the crystallographic special positions with three-fold local symmetry occupied by atoms X, according to the International Tables for Crystallography. d X? ? ? X Distances within the hexagonal net. e X? ? ?X? ? ? X Angles within the hexagonal net. f f = flat hexagonal net; p = puckered hexagonal net; sl = slightly, v = very. g Metal atoms are disordered. h Embrace involving six rings. For each of a pair of molecules, one ring is from R3X which is involved in the SPE and two are from the other XR3.J.Chem. Soc., Dalton Trans., 1998, Pages 329–344 339 Fig. 9 The HASPE lattice in Ph3P]Cu(m-SC6H4Me-4)3]Mo]Cu(m-SC6H4Me-4)3]PPh3 [SUDYUL]. Disordered solvent molecules occupy the ‘empty’ spaces Fig. 10 Three-fold and side views of crystal structure of Ph3Si]N]] C]] N]SiPh3 [TPSICI]: N blue, Si yellow SPE. Around the hexagonal ring, there are PQPE, while each ring of the molecule occupying the centre of the hexagon takes part in three ef interactions, two with one molecule of the hexagonal net, and one with a second.This is a classic wheel and axle compound, which as a consequence of the well-packed SPEs does not form an inclusion lattice. All the type D molecules just described form HASPE lattices in which both ends of the molecule form SPEs. In contrast to this, there are four crystal structures listed in Table 4 which behave as though only one end of the molecule contained the XR3 moiety.These are the four structures of the type (4- MeC6H4)3X]Y(C6H4Me-4)3, where X and Y are Ge/Pb [LIBHIN], Sn/Sn [SOMMOW], Pb/Pb [SOMMUC] and Pb/Sn [SOMNAJ]. In these structures, the molecules form only one SPE, and the HASPE lattice is similar to those described above for type A, B or C molecules. The space between noninteracting aromatic rings is occupied by the methyl groups protruding from surrounding molecules. For the mixed-metal340 J. Chem.Soc., Dalton Trans., 1998, Pages 329–344 Fig. 11 Three-fold and side views of part of the crystal packing of the rhombohedral polymorph of Ph3SnCl, illustrating the outer hexagonal net of molecules aligned with the three-fold axis (coloured green) and the inner hexagon of 12 molecules in six SPEs (coloured blue) compounds, (4-MeC6H4)3Ge]Pb(C6H4Me-4)3 [LIBHIN] and (4-MeC6H4)3Sn]Pb(C6H4Me-4)3 [SOMNAJ], the metals are disordered over the two sites, so that for packing purposes, it is irrelevant which metal is at the ‘end’ of the molecule which forms the SPE.The cation [PN(PPh3)2]1 could be considered as suitable for the formation of this class of HASPE if the central PNP geometry is linear (which is often not the case). There are three hexagonal/rhombohedral crystal structure types including [PN(PPh3)2]1 in the CSD, but none forms the HASPE lattice. In [PN(PPh3)2]1[M(CO)6]2, M = Nb [BILNEP10] or Ta [BUVGII], the anion is situated between linear cations along the hexagonal axis.In the [PN(PPh3)2]1 salt of the cluster [(m-CO)3- (m3-C]] C]] O){Ru(CO)2}3(m3-CuI)]22[PAJWOM] there are two different columns in the c direction: in one, the anion is between cations, and in the second, there are only cations, but they are too far apart to embrace and are disordered. In [PN- (PPh3)2]1[Fe4(CO)13]2[PIMNFE01] the cations are not linear and do not align with the hexagonal axis. Unusual HASPE Lattices The compound Ph3SnCl [TPSNCL03] has a rhombohedral crystal lattice ‡ which has an additional unusual feature in which SPE interactions occur within the hexagons as well as between the hexagons.The hexagonal net of molecules with SPE occurring on opposite sides is similar to that in Ph3CX (X = Cl or Br: see Fig. 4). However in the distal region of each of these SPE, where two Cl atoms point towards each other, there is another set of twelve Ph3SnCl molecules not aligned with the crystallo- ‡ There is also a monoclinic dimorph in which the interactions between the phenyl rings are not multiple embraces.The related compounds ClSiPh3, BrSnPh3 and BrGePh3 have sgraphic three-fold axis and which form a local hexagon of six SPEs. This is illustrated in Fig. 11. There are ef interactions between the two sets of molecules. Both types of SPE are remarkably short, 5.29 Å along the three-fold axis and 5.57 Å oblique to the three-fold axis. The unconventional compound [Ph3As]I]I]I]AsPh3]1 [Ph3- AsCoI3]2 [PELZAH], prepared in the solid state by reaction of Ph3AsI3 and Co2(CO)8, contains a mixture of double ended embracers and single ended embracers.The long c axis in space group R3c is comprised of alternating [Ph3As]I]I]I]AsPh3]1 and [Ph3AsCoI3]2 species, which also alternate around the hexagonal lattice as shown in Fig. 12. The embrace sequence along the c axis is ? ? ? Ph3AsCoI3 ? ? Ph3AsIIIAsPh3 ? ? ? Ph3As- CoI3 ? ? .The linearity of the central As]I]I]I]As segment is unusual, and is presumably a consequence of the HASPE lattice structure. A chemically different and distinctive compound forming a HASPE lattice is HOSi{h6-C6H5]Cr(CO)3}3, WAWXUN10, in space group R3� (see Fig. 13). The SPE (Si ? ? ? Si 6.45 Å) is formed by the SiPh3 group, aligned with the three-fold axis, but the obverse face of each phenyl ring is h6 bound to Cr(CO)3. One of these CO ligands is directed almost parallel to the three-fold axis.Where the OH groups align along the threefold axis, the O ? ? ? O distance is 2.77 Å and there is presumably hydrogen bonding, although the hydrogen atoms are necessarily disordered. Summary Account of the Symmetry and Dimensions of HASPE Lattices All structures have SPEs aligned with the three-fold axis of the hexagonal or rhombohedral space groups. Analysis of the array of centroids Ç of the SPEs shows that all structures with only one X atom in the asymmetric unit have an array of centroids inJ.Chem. Soc., Dalton Trans., 1998, Pages 329–344 341 Fig. 12 Trigonal and side views of the HASPE lattice in crystalline [Ph3As]I]I]I]AsPh3]1[Ph3AsCoI3]2 [PELZAH]: Co blue, As fawn Fig. 13 Three-fold view of the crystal structure of HOSi{h6-C6H5]Cr(CO)3}3, WAWXUN10: Si yellow, Cr white, space group R3� . Note that the O atoms of three CO groups from three different columns of SPE overlap along the three-fold axes between the columns of SPE space group R3� m, irrespective of whether the space group of the complete structure was R3� or R3� c, but for those in R3� c the c repeat for the centroid array is half of c for the complete lattice.For those complete structures in space group P3� or P3� c1, the asymmetric unit includes three different X atoms, and there are two independent SPEs (for example see Fig. 4): the arrays of SPE centroids can adhere to space groups R3� m or P3� m1. In all cases except one [KOMMOO], the SPEs around a hexagon of X atoms are directed alternately along 1c and 2c.There is variety in the molecular sizes and shapes of YXR3, with Y ranging from nothing [BADBIR] through single atoms, bulky three-fold units such as Au]Tc(NC6H3Pri 2-2,6)3 [WEWLEP], short double ended connectors (N]] C]] N [TPSICI]), to bulky double ended connectors like Cu(m-342 J. Chem. Soc., Dalton Trans., 1998, Pages 329–344 Table 5 Refcode FUPSEO PINSOU FUFVAD WEWLEP BADBIR PIHGIW X]Y? ? ?Y]X P]H? ? ? Cl3GaGaCl3 ? ? ?H]P P]Cl ? ? ? SnCl6 ? ? ? Cl]P Ge]GeEt3 ? ? ? Et3Ge]Ge P]Au]Tc(NC6H3Pri 2-2,6)3 ? ? ? (NC6H3Pri 2-2,6)3Tc]Au]P As ? ? ? As As]GaI3 ? ? ? I3Ga]As R/Å 11.3 13.3 23.1 33.1 11.4 9.9 2 × SPE/Å 12.5 11.2 14.7 13.0 14.9 13.2 Angle at X/8 119.8 119.2 107.2 86.2 117.8 118.4 Ring geometry Planar Planar Elongated Elongated Compressed Compressed Fig. 14 The hexagonal lattice of Ph3Ge]GeMe3 [HAYSEF]: Ge red. The CH ? ? ? Ph interactions are shorter around the hexagonal net than along the three-fold axis SC6H4CH3-4)3]Mo]Cu(m-SC6H4CH3-4)3 [SUDYUL].This variety is accommodated by variation in the dimensions and puckering of the HASPE array of X atoms. The X ? ? ? X lengths of the SPEs range from 5.59 to 7.58 Å (and 5.29 Å in the unusual lattice of TPSNCL03), but the X ? ? ? X distances in the hexagons cover a much larger range, 7.18–13.99 Å. Where this X? ? ?Xhexagon distance is relatively short it signifies additional multiple phenyl embraces (QPE, DPE).Compression or elongation of the lattice can be understood in the following way. Defining R as the length X]Y? ? ?Y]X, when R = 2 × SPE, the hexagonal net is planar (Fig. 1). If R is greater than twice the SPE the hexagonal net must pucker by elongation [Fig. 2(a)], while when R is less than twice the SPE the hexagonal is compressed [Fig. 2(b)]. This is illustrated in Table 5. HASPE Lattices without Trigonal Symmetry Triphenylmethyl chloride, Ph3CCl, is reported to crystallise with a number of polymorphs, although full details are not available.The compound with refcode ZZZVTY12, in space group P3� , has been described above. Two other polymorphs [ZZZVTY03, ZZZVTY04], both in space group P1� , do in fact contain columns of SPE in a pseudo HASPE lattice. In a recent report, the structure of Ph3Si]C]] ] C]H with O]] PPh3,8 which crystallises in space group P1� , has been shown to have four crystallographically different columns of molecules taking part in SPEs. The overall packing, however, is a pseudo HASPE array.It is more difficult to uncover low symmetry HASPE lattices from the CSD. Exceptions While two mixed triaryl, trialkyl derivatives Ph3P–AlMe3 [KIFSOH] and (2-CH3C6H4)3P]AlMe3 [KIFSUN] form the HASPE lattice (Table 3), the related compounds Ph3X–YMe3, where X/Y = Si/Si [BIMSUL], Ge/Si [GAFGID], Ge/Ge [HAYSEF] or Si/Ge [SENYIT] do not. Instead, a hexagonal array is formed in which the triplet of methyl groups is directed towards the three phenyl rings of the adjacent molecule along the threefold axis, as illustrated in Fig. 14. The difference between these two groups of related compounds is that the first set has molecules alternating in direction along three-fold axis, while the second set has all molecules stacked in the same orientation. These two structure types are informative about subtle differences of interaction energy between segregated aryl–aryl and alkyl–alkyl contacts, versus desegregated aryl–alkyl contacts.The shortest contacts between the methyl groups and phenyl rings, however, is around the hexagonal net, rather than along the three-fold direction. Also forming the latter type of hexagonal lattice is (p-ClC6H4)3AsO [DATDIL01] where the O atom interacts with the phenyl rings of the adjacent molecule, forming three C]H? ? ? O hydrogen bonds with C ? ? ? O distances 3.22 Å and H ? ? ? O 2.35 Å. The structures SUDYUL and YIRHOW in Table 4 form the HASPE lattice despite the inclusion of solvent, thf.However, for structures such as Ph3Ge]GePh3 with benzene [HPGEBZ10], Ph3Sn]SnPh3 with benzene [VINXOF], Ph3Si]O]SiPh3 with benzene [CECXAJ01], and Ph3Si]O]SiPh3 with piperidine [DOHDOT], the solvent disrupts the SPEs and interacts itself with the three phenyl rings at one end of the molecule. The alternative interaction of benzene with Ph3Sn]SnPh3 inJ. Chem. Soc., Dalton Trans., 1998, Pages 329–344 343 Fig. 15 Part of the structure of Ph3SnSnPh3 crystallised with benzene, showing how the benzene (blue) blocks the SnPh3 face, obstructing the normal SPE formation VINXOF is shown in Fig. 15. We believe that this alternative to the HASPE lattice occurs because the two XPh3 ends of these short molecules have insufficient intramolecular separation to allow the intercolumn interactions of the HASPE lattice (see, for example, Fig. 10). In CECXAJ01 and DOHDOT (Ph3Si]O] SiPh3 with benzene and piperidine respectively) the geometry is linear at the O atom like the structures of Ph3P]O]GeCl3 [JEJLEP] and others in Table 3.Conclusion We have established (1) the generality of the HASPE lattice as a crystal supramolecular motif; (2) the variety of molecular structures which adopt this crystal supramolecularity; (3) the variability of the lattice to accommodate variations of molecular structure; (4) the dominance of the SPE and its contribution to lattice cohesion.The salts in which Ph3PX1 cations (X = Me,ASPE lattice are like inclusion structures with the anions located in cavities of the host lattice of sextuply embracing cations. Acknowledgements This research is supported by the Australian Research Council. References 1 I. G. Dance and M. L. Scudder, J. Chem. Soc., Chem. Commun., 1995, 1039. 2 I. G. Dance and M. L. Scudder, J. Chem. Soc., Dalton Trans., 1996, 3755. 3 I. G. Dance and M.L. Scudder, Chem. Eur. J., 1996, 2, 481. 4 C. Hasselgren, P. A. W. Dean, M. L. Scudder, D. C. Craig and I. G. Dance, J. Chem. Soc., Dalton Trans., 1997, 2019. 5 F. H. Allen, J. E. Davies, J. J. Galloy, O. Johnson, O. Kennard, C. F. Macrae and D. G. Watson, Chem. Inf. Comput. Sci., 1991, 31, 204. 6 International Tables for X-Ray Crystallography, Kynoch Press, Birmingham, 1952. 7 N. Burford, B. W. Royan, R. E. v. H. Spence, T. S. Cameron, A. Linden and R. D. Rogers, J. Chem. Soc., Dalton Trans., 1990, 1521. 8 T. Steiner, J. van der Maas and B. Lutz, J. Chem. Soc., Perkin Trans. 2, 1997, 1287. BADBIR, TOLARS01, A. N. Sobolev and V. K. Belsky, J. Organomet. Chem., 1981, 214, 41. BAYFUC, M. G. B. Drew, Acta Crystallogr., Sect. B, 1982, 38, 254. BILNEP10, BUVGII, F. Calderazzo, U. Englert, G. Pampaloni, G. Pellizi and R. Zamboni, Inorg. Chem., 1983, 22, 1865. BIMSUL, L. Parkanyi and E. Hengge, J. Organomet. Chem., 1982, 235, 273. BTZANI, J. Meunier-Piret, P.Piret, J.-P. Putzeys and M. van Meerssche, Acta Crystallogr., Sect. B, 1976, 32, 714. CANTEQ, A. N. Sobolev, V. K. Bel’skii, I. P. Romm and E. N. Gur’yanova, Zh. Strukt. Khim., 1983, 24, 123. CECXAJ01, DOHDOT, K. Suwinska, G. J. Palenik and R. Gerdil, Acta Crystallogr., Sect. C, 1986, 42, 615. CINSIB, E. G. Perevalova, Yu. T. Struchkov, V. P. Dyadchenko, E. I. Smyslova, Yu. L. Slovokhotov and K. I. Grandberg, Izv. Akad. Nauk SSSR, Ser. Khim., 1983, 2818. CIWZUD, A. I. Gusev, M.G. Los, S. D. Vlasenko, V. I. Zhun and V. D. Sheludyakov, Zh. Strukt. Khim., 1984, 25, 172. COSPHP, J. K. Stalick and J. A. Ibers, Inorg. Chem., 1969, 8, 419. DATDIL01, V. K. Bel’skii and V. E. Zavodnik, J. Organomet. Chem., 1984, 265, 159. DNTPIR, D. N. Cash, R. O. Harris, S. C. Nyburg and F. H. Pickard, J. Cryst. Mol. Struct., 1975, 5, 377. FUFVAD, R. I. Bochkova, Yu. N. Drozdov, E. A. Kuz’min, L. N. Bochkarev and M. N. Bochkarev, Koord. Khim., 1987, 13, 1126. FUPSEO, FUPSIS, FUPSOY, M.A. Khan, D. G. Tuck, M. J. Taylor and D. A. Rogers, J. Cryst. Spectrosc., 1986, 16, 895. FUTPIT, R. L. Richards, C. Shortman, D. C. Povey and G. W. Smith, Acta Crystallogr., Sect. C, 1987, 43, 2309. FUTTAP, B. Beagley, C. B. Colburn, O. El-Sayrafi, G. A. Gott, D. G. Kelly, A. G. Mackie, C. A. McAuliffe, P. P. MacRory and R. G. Pritchard, Acta Crystallogr., Sect. C, 1988, 44, 38. GAFGID, L. Parkanyi, C. Hernandez and K. H. Pannell, J. Organomet. Chem., 1986, 301, 145.HAYSEF, L. Parkanyi, A. Kalman, S. Sharma, D. M. Nolen and K. H. Pannell, Inorg. Chem., 1994, 33, 180. HPGEBZ10, M. Drager and L. Ross, Z. Anorg. Allg. Chem., 1980, 469, 115. JACWAL, JACWEP, G. Ferguson and C. Glidewell, J. Chem. Soc., Perkin Trans. 2, 1988, 2129. JEJLAL, JEJLEP, JEJLIT, N. Burford, B. W. Royan, R. E. v. H. Spence, T. S. Cameron, A. Linden and R. D. Rogers, J. Chem. Soc., Dalton Trans., 1990, 1521.344 J. Chem. Soc., Dalton Trans., 1998, Pages 329–344 JICWUN, D.C. Bradley, M. B. Hursthouse, M. Motevalli and Z. Dao-Hong, J. Chem. Soc., Chem. Commun., 1991, 7. KIFSOH, KIFSUN, D. A. Wierda and A. R. Barron, Polyhedron, 1989, 8, 831. KOMMOO, W. M. P. B. Menge and J. G. Verkade, Inorg. Chem., 1991, 30, 4628. LIBHIN, H.-J. Koglin, K. Behrends and M. Drager, Organometallics, 1994, 13, 2733. MPPICU, G. A. Bowmaker, G. R. Clark and D. K. P. Yuen, J. Chem. Soc., Dalton Trans., 1976, 2329. MTPHCI, C. Couldwell and K. Prout, Acta Crystallogr., Sect. B, 1978, 34, 2312. PAJWOM, A. S. Gunale, M. P. Jensen, D. A. Phillips, C. L. Stern and D. F. Shriver, Inorg. Chem., 1992, 31, 2622. PELZAH, S. M. Godfrey, H. P. Lane, A. G. Mackie, C. A. McAuliffe and R. G. Pritchard, J. Chem. Soc., Chem. Commun., 1993, 1190. PIHGIW, L.-J. Baker, C. E. F. Rickard and M. J. Taylor, J. Organomet. Chem., 1994, 464, C4. PIMNFE01, G. van Buskirk, C. B. Knobler and H. D. Kaesz, Organometallics, 1985, 4, 149. PINSOU, E. Rentschler and K. Dehnicke, Z. Kristallogr., 1994, 209, 89. SENYIT, K. H. Pannell, R. N. Kapoor, R. Raptis, L. Parkanyi and V. Fulop, J. Organomet. Chem., 1990, 384, 41. SOMMOW, SOMMUC, SOMNAJ, C. Schneider and M. Drager, J. Organomet. Chem., 1991, 415, 349. SUDYUL, P. M. Boorman, H.-B. Kraatz, M. Parvez and T. Ziegler, J. Chem. Soc., Dalton Trans., 1993, 433. TPHMBR02, A. Dunand and R. Gerdil, Acta Crystallogr., Sect. B, 1984, 40, 59. TPSICI, G. M. Sheldrick and R. Taylor, J. Organomet. Chem., 1975, 101, 19. TPSNCL03, S. W. Ng, Acta Crystallogr., Sect. C, 1995, 51, 2292. TPTOSB, A. N. Sobolev, I. P. Romm, V. K. Belskii and E. N. Gur’yanova, J. Organomet. Chem., 1979, 179, 153. VINXOF, H. Piana, U. Kirchgassner and U. Schubert, Chem. Ber., 1991, 124, 743. WAWKAG, F. J. Feher, T. A. Budzichowski and K. J. Weller, Polyhedron, 1993, 12, 591. WAWXUN10, K. L. Malisza, L. C. F. Chao, J. F. Britten, B. G. Sayer, G. Jaouen, S. Top, A. Decken and M. J. McGlinchey, Organometallics, 1993, 12, 2462. WEKBIX, T. Rubenstahl, D. W. von Gudenberg, F. Weller, K. Dehnicke and H. Goesmann, Z. Naturforsch., Teil B, 1994, 49, 15. WEWLEP, A. K. Burrell, D. L. Clark, P. L. Gordon, A. P. Sattelberger and J. C. Bryan, J. Am. Chem. Soc., 1994, 116, 3813. YIRHOW, P. C. Leverd, M. Lance, M. Nierlich, J. Vigner and M. Ephritikhine, J. Chem. Soc., Dalton Trans., 1994, 3563. ZZZVTY12, A. Dunand and R. Gerdil, Acta Crystallogr., Sect. B, 1982, 38, 570. ZZZVTS01, ZZZVTY03, ZZZVTY04, B. Kahr and R. L. Carter, Mol. Cryst. Liq. Cryst. A, 1992, 219, 79. Received 27th August 1997; Paper 7/06252D

ISSN:1477-9226    

DOI:10.1039/a706252d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 335–339 335 Crystal structure, luminescence and other properties of some lanthanide complexes of the polypyridine ligand 6,69-bis[bis(2- pyridylmethyl)aminomethyl]-2,29-bipyridine Anders Døssing,*,a Hans Toftlund,b Alan Hazell,c James Bourassa d and Peter C. Ford d a Department of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen Ø, Denmark b Department of Chemistry, Odense University, DK-5230 Odense M, Denmark c Department of Chemistry, Aarhus University, DK-8000 Århus C, Denmark d Department of Chemistry, University of California, Santa Barbara, CA 93106, USA Europium(III) and gadolinium(III) form 1 : 1 complexes with the octadentate polypyridine ligand 6,69-bis[bis(2- pyridylmethyl)aminomethyl]-2,29-bipyridine L2.The crystal structure of the complex [GdL2Cl2]Cl?2H2O?0.5EtOH has been determined. The gadolinium atom is co-ordinated to seven nitrogen atoms of the L2 ligand and two chloride ligands.One of the pyridine groups of L2 is dangling unco-ordinated. Irradiation of water or MeOH solutions of the [EuL2]Cl3 complex with UV light led to a metal-centred luminescence. The luminescence lifetimes at 296 K in water, D2O, MeOH and CD3OD were 0.78, 1.28, 0.98 and 1.25 ms, respectively. From these data the average number of co-ordinated solvent molecules (water or MeOH) was calculated to be 0.5 according to earlier empirical correlations. The luminescence consists of a series of narrow lines corresponding to the envelopes of 5D0 æÆ 7FJ transitions, the most intense occurring at 613 nm (Æ7F2).The integrated intensities of these emission lines give the respective quantum yields of 0.046 and 0.089 in water and D2O. The equilibrium constant for formation of the europium complex in MeOH, I = 0.100 mol dm23 NBu4ClO4, was determined to be 107.09 ± 0.09 dm3 mol21. Recent years have witnessed considerable interest in the preparation and characterization of new materials based on the luminophores europium(III) and terbium(III) that display luminescence in aqueous solution.1 Such europium(III) or terbium( III) complexes are luminescent in aqueous solution if three conditions are fulfilled: (i) the ligand includes a highly absorbing chromophore (since the f–f transitions in the metal have very low molar absorption coefficients), (ii) energy transfer from the ligand-centred excited states to the metal centre is fast and efficient, (iii) water is largely excluded from the first coordination sphere, since the non-radiative deactivation pathway for the excited metal atom is promoted by the high-frequency vibrational modes of H2O.The antenna effect represented by a combination of requirements (i) and (ii) 2 makes the ligand design crucial to the photophysical properties of the complexes. Branched macrocyclic ligands containing bipyridine or phenanthroline chromophores have proved especially useful in this regard.Among the first systems to be investigated in this context were the europium(III) and terbium(III) complexes of the cage ligand L1.1a,1b Comparisons of the luminescence lifetimes in water and D2O solutions using the empirical correlation described below suggested that in aqueous solutions the complexes [EuL1]3+ and [TbL1]3+ each had an average of two to three water molecules present in the first co-ordination spheres. A later crystal structure determination confirmed that (in the solid) two water molecules were indeed co-ordinated.3 The present investigation describes the preparation and characterization of gadolinium(III) and europium(III) complexes of the potentially octadentate ligand 6,69-bis[bis(2- pyridylmethyl)aminomethyl]-2,29-bipyridine L2.Comparison of the drawings of L1 and L2 illustrates the structural relationship between the two; L2 can be imagined to be the result of hydrogenolysis of the 2,29 C]C bonds in two of the bipyridines of L1 to give a more flexible podand-type compound.Experimental Reagents The compounds EuCl3?6H2O (99.99%), GdCl3?6H2O and D2O (99.9%) were obtained from Aldrich. Solvents for spectroscopy were ‘spectroscopic grade’ and water was doubly distilled. Tetra-n-butylammonium perchlorate was obtained from Fluka and dried at 363 K under reduced pressure prior to use. Preparation L2. This compound was prepared as described elsewhere.4 UV/VIS (MeOH, I = 0.100 mol dm23 NBu4ClO4): lmax/nm336 J.Chem. Soc., Dalton Trans., 1997, Pages 335–339 245, 261, 267 and 288 (1023 e/dm3 mol21 cm21 15.0, 20.0, 19.1 and 16.9). [EuL2]Cl3?1.5H2O 1. A mixture of EuCl3?6H2O (195 mg, 0.532 mmol), MeCN (30 cm3) and trimethyl orthoformate (3.5 cm3, 32 mmol) was refluxed for 4 h with stirring. Then L2 (308 mg, 0.532 mmol) was added, and the mixture was refluxed for an additional hour. The solvent was then removed, and the slightly yellow residue dissolved in ethanol (5 cm3).The solution was filtered, and the complex reprecipitated by addition of diethyl ether (50 cm3). It was washed with diethyl ether then dried in vacuum (400 mg). The crude product was redissolved in ethanol (3 cm3), the solution was filtered, and the complex reprecipitated by vapour diffusion of diethyl ether into the ethanol solution. The final step was repeated several times yielding a white product (ca. 200 mg) (Found: C, 49.8; H, 4.4; N, 12.8.C36H37Cl3EuN8O1.5 requires C, 50.0; H, 4.3; N, 13.0%). [GdL2]Cl3?5H2O 2. This complex was made as for 1 (Found: C, 46.4; H, 4.3; N, 12.1. C36H44Cl3GdN8O5 requires C, 46.6; H, 4.8; N, 12.1%). Absorption spectra The UV/VIS absorption spectra were measured on a Perkin- Elmer Lambda 17 spectrophotometer. Crystallography X-Ray-quality crystals of [GdL2Cl2]Cl?2H2O?0.5EtOH were obtained by vapour diffusion of diethyl ether into an ethanol solution of complex 2. A colourless crystal of dimensions 0.45 × 0.50 × 0.58 mm was mounted on a Huber diffractometer and the cell dimensions determined from setting angles of 30 reflections measured at ±2q with graphite-monochromatized Mo-Ka radiation (l = 0.710 73 Å).Intensity data were measured at 294 K using w–2q scans; two standard reflections were measured every 50. Crystal data are given in Table 1. Intensity data are corrected for background, Lorentz-polarization effects, decay and absorption.5 The structure was determined by direct methods using SIR 92 6 followed by Fourier-difference syntheses.In the full-matrix least-squares refinement (on F) the pyridylmethyl groups were constrained 7 to be identical as were the pyridylmethyl halves of the dimethylbipyridines. Gadolinium and chlorine atoms were refined anisotropically, nonhydrogen atoms isotropically and hydrogen atoms on the ligands were kept at calculated positions with Uiso 20% larger than that of the atom to which they were attached.A Rogers factor 8 refined to h = 0.04(10) indicating that the crystal was twinned containing almost equal numbers of domains of opposite polarity. The final R and R9 values were 0.057 and 0.071 for 4945 reflections with I > 3s(I). Atomic scattering factors, f 9 and f 0, were taken from ref. 9. Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Instructions for Authors, J.Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/316. Luminescence spectra and lifetimes For lifetime measurements, solutions of complex 1 were pipetted into Pyrex cells (2 cm diameter) fitted with Rotoflow stopcocks and deaerated by five freeze–pump–thaw cycles. Emission lifetimes were determined as described 10 but with some modifi- cations and updating of the instrumentation.The excitation source was the third harmonic (355 nm) of a Continuum NY- 61-20 Q-switched Nd/YAG pulsed laser (operating at 10 Hz with a pulse width of 10 ns). The emission was monitored at right angles to the excitation beam through a SPEX model 1680 Doublemate grating monochromator using an RCA 8852 PMT detector. The PMT output was recorded by an Tektronix TDS 540 digital oscilloscope and transferred to a dedicated Teltron 484 PC work station for data analysis and storage. The 100 digitized waveforms were signal averaged (500 points per waveform) and analysed by single-exponential curve fitting.Emission and excitation spectra were recorded on a SPEX Fluorolog11 spectrofluorimeter with a water-cooled Hamamatsu R928A PMT corrected for phototube response. For roomtemperature luminescence spectra, square Suprasil quartz cells were used for the solutions. For the solid and for the solutions at 77 K, cylindrical 0.5 cm diameter quartz cells were used to obtain emission spectra.Luminescence quantum yields were determined by comparing the integrated emission intensity of the europium(III) complex in solution to that of [Ru(bipy)3]Cl2 (bipy = 2,29-bipyridine) in water.12 Typical concentrations of 1 in these experiments were ª1026 mol dm23. Results and Discussion Crystal structure The asymmetric unit consists of two [GdL2Cl2]+ cations, two chloride ions, four waters of crystallization and a solvent molecule presumed to be ethanol.The gadolinium atoms are coordinated to seven nitrogen atoms of the L2 ligand and two chlorine atoms, and one of the pyridine groups is dangling unco-ordinated (see Fig. 1). The co-ordination polyhedron can be described as a monocapped cube; bond distances and angles about the gadolinium atoms are listed in Table 2. The Gd]Cl distances vary between 2.675(6) and 2.737(6) Å, with a mean of 2.71(1) Å, and the Gd]N distances vary between 2.59(1) and 2.77(2) Å with a mean value of 2.68(2) Å.The mean values in related compounds14 are 2.68(1) and 2.65(2) Å respectively. Carbon–carbon and –nitrogen distances have the expected values. The two halves of the bipyridyl groups are almost coplanar with torsion angles N(37)]C(36)]C(46)]N(47) 6(3) and N(379)]C(369)]C(469)]N(479) 2(2)8. The two cations are almost related by a glide plane at y = 0 and so have opposite chirality, a further difference is that the orientation of the nonco- ordinated pyridine groups differs by 1808.The pseudo-glide plane accounts for the ease with which zones of opposite polarity are achieved. The L2 ligand forms six five-membered chelate rings with N]Gd]N angles in the range 59.1–61.78. The mutual orientations of these chelate rings will in principle define the chirality of the molecule. However, the helical arrangement of the chain N(67), N(20), N(47), N(37), N(10), N(17) is a more Table 1 Crystal data for [GdL2Cl2]Cl?2H2O?0.5EtOH Formula C36H34Cl3GdN8?2H2O?0.5C2H6O M 901.43 Crystal symmetry Monoclinic Space group P21 a/ Å 12.652(4) b/ Å 14.670(7) c/ Å 20.860(9) b/8 90.022(23) U/Å3 3872(3) Z 4 Dc/g cm23 1.546 m(Mo-Ka)/cm21 19.73 No.of reflections 7083 No. observed [I > 3s(I)], No 4945 No. of variables, Nv 304 Ra 0.057 R9 b 0.071 Sc 1.40 a R = S||Fo| 2 |Fc||/S|Fo|. b R9 = [Sw(|Fo| 2 |Fc|)2/Sw|Fo|2]� �� , w = 1/ {[s(Fo 2) + 1.03Fo 2]� �� 2 |Fo|} with s(Fo 2) from counting statistics. c S = [Sw(|Fo| 2 |Fc|)2/(No]Nv)]� �� .J.Chem. Soc., Dalton Trans., 1997, Pages 335–339 337 obvious way to define the chirality in this case. This gives D for the complex cation in Fig. 1. Optical absorption spectra An absorption spectrum of L2 in MeOH is shown in Fig. 2. This consists of several bands, and the maxima at 246 and 288 nm are attributed to the bipyridine chromophore and those at 261 and 267 nm to the pyridine chromophores. Owing to low solubility, only a qualitative absorption spectrum of free L2 in water could be obtained,but it proved to be very similar to that taken in MeOH.The spectrum of 1 is very similar to that of free L2 in MeOH (Fig. 2), although a red shift is observed for the absorption band attributed to the bipyridine p æÆ p* transition at 288 nm. The pyridine bands are, however, relatively unaffected by metal complexation. Complex 2 showed similar behaviour. Concerning the stability of 1 in solution a spectrophotometric determination of the formation constant Kf [equation (1)] was carried out.By using MeOH, I = 0.100 mol dm23 M + L Kf ML (1) NBu4ClO4, as solvent instead of water the formation of hydroxo complexes and ligand protonation was avoided. The spectrum of solutions of 1 was measured over a wide range of Table 2 Selected bond distances (Å) and angles (8) for [GdL2Cl2]Cl? 2H2O?0.5EtOH Gd(1)]Cl(1) Gd(1)]Cl(2) Gd(1)]N(10) Gd(1)]N(20) Gd(1)]N(17) Gd(1)]N(37) Gd(1)]N(47) Gd(1)]N(57) Gd(1)]N(67) 2.712(6) 2.724(5) 2.770(21) 2.651(18) 2.770(10) 2.586(11) 2.606(10) 2.661(10) 2.652(11) Gd(19)]Cl(19) Gd(19)]Cl(29) Gd(19)]N(109) Gd(19)]N(209) Gd(19)]N(179) Gd(19)]N(379) Gd(19)]N(479) Gd(19)]N(579) Gd(19)]N(679) 2.675(6) 2.737(6) 2.769(20) 2.660(20) 2.740(10) 2.651(11) 2.669(11) 2.664(11) 2.644(12) Cl(1)]Gd(1)]Cl(2) Cl(1)]Gd(1)]N(10) Cl(1)]Gd(1)]N(20) Cl(1)]Gd(1)]N(17) Cl(1)]Gd(1)]N(37) Cl(1)]Gd(1)]N(47) Cl(1)]Gd(1)]N(57) Cl(1)]Gd(1)]N(67) Cl(2)]Gd(1)]N(10) Cl(2)]Gd(1)]N(20) Cl(2)]Gd(1)]N(17) Cl(2)]Gd(1)]N(37) Cl(2)]Gd(1)]N(47) Cl(2)]Gd(1)]N(57) Cl(2)]Gd(1)]N(67) N(10)]Gd(1)]N(20) N(10)]Gd(1)]N(17) N(10)]Gd(1)]N(37) N(10)]Gd(1)]N(47) N(10)]Gd(1)]N(57) N(10)]Gd(1)]N(67) N(20)]Gd(1)]N(17) N(20)]Gd(1)]N(37) N(20)]Gd(1)]N(47) N(20)]Gd(1)]N(57) N(20)]Gd(1)]N(67) N(17)]Gd(1)]N(37) N(17)]Gd(1)]N(47) N(17)]Gd(1)]N(57) N(17)]Gd(1)]N(67) N(37)]Gd(1)]N(47) N(37)]Gd(1)]N(57) N(37)]Gd(1)]N(67) N(47)]Gd(1)]N(57) N(47)]Gd(1)]N(67) N(57)]Gd(1)]N(67) 158.2(2) 88.2(4) 79.2(4) 70.2(3) 95.7(3) 73.0(3) 131.8(3) 101.4(3) 73.1(4) 117.8(4) 90.3(3) 85.1(3) 125.4(3) 69.9(3) 77.6(3) 166.0(5) 59.1(4) 61.2(4) 118.0(5) 129.3(4) 115.1(5) 110.0(4) 125.8(4) 64.2(5) 64.7(4) 62.2(4) 118.7(4) 143.1(4) 150.6(4) 64.6(4) 62.8(4) 82.1(4) 162.5(4) 63.3(4) 126.2(4) 89.3(4) Cl(19)]Gd(1)]Cl(29) Cl(19)]Gd(1)]N(109) Cl(19)]Gd(1)]N(209) Cl(19)]Gd(1)]N(179) Cl(19)]Gd(1)]N(379) Cl(19)]Gd(1)]N(479) Cl(19)]Gd(1)]N(579) Cl(19)]Gd(1)]N(679) Cl(29)]Gd(1)]N(109) Cl(29)]Gd(1)]N(209) Cl(29)]Gd(1)]N(179) Cl(29)]Gd(1)]N(379) Cl(29)]Gd(1)]N(479) Cl(29)]Gd(1)]N(579) Cl(29)]Gd(1)]N(679) N(109)]Gd(1)]N(209) N(109)]Gd(1)]N(179) N(109)]Gd(1)]N(379) N(109)]Gd(1)]N(479) N(109)]Gd(1)]N(579) N(109)]Gd(1)]N(679) N(209)]Gd(1)]N(179) N(209)]Gd(1)]N(379) N(209)]Gd(1)]N(479) N(209)]Gd(1)]N(579) N(209)]Gd(1)]N(679) N(179)]Gd(1)]N(379) N(179)]Gd(1)]N(479) N(179)]Gd(1)]N(579) N(179)]Gd(1)]N(679) N(379)]Gd(1)]N(479) N(379)]Gd(1)]N(579) N(379)]Gd(1)]N(679) N(479)]Gd(1)]N(579) N(479)]Gd(1)]N(679) N(579)]Gd(1)]N(679) 156.5(2) 87.8(4) 78.9(4) 71.1(3) 94.0(3) 72.6(3) 133.0(3) 100.9(3) 73.0(4) 118.7(4) 87.3(3) 88.5(3) 127.9(3) 70.4(3) 77.3(3) 165.8(5) 59.4(4) 61.4(4) 117.1(5) 129.2(4) 117.5(5) 110.9(5) 123.8(5) 63.5(5) 64.7(5) 61.4(5) 119.1(4) 143.6(4) 148.4(4) 65.5(4) 61.3(4) 83.5(4) 165.1(4) 65.2(4) 124.7(4) 87.3(4) concentrations.From the absorbance (A305) at the isosbestic point at 305 nm the complex concentration could be calculated. The absorbance at 288 nm (A288) can then be expressed as shown in equation (2).A288 = l(eL 288 2 eML 288) 2 Ê Ë÷1 K2f + 4A305 le305Kf 2 1 Kf � � + eML 288 e305 A305 (2) In the derivation of equation (2) eM 288 was neglected. Values for eL 288 and e305 were determined by independent experiments to be 16.9 × 103 and 7.70 × 103 dm3 mol21 cm21, respectively and the cell length, l, was 1.000 cm. The experimental values of A288 and A305 are plotted in Fig. 3. A least-squares fit refining eML 288 and Kf gave Kf = 107.09 ± 0.09 dm3 mol21 and eML 288 = 2.89(4) × 103 dm3 mol21 cm21.From the spectrum (Fig. 2, inset) of a 2.00 × 1025 mol dm23 solution of 1 in water the presence of free L2 is clearly seen. From the value of A288 (0.1546) a rough estimate of Kf in water (pH 7.0) can be obtained assuming that the values of eL 288 and eML 288 in water are similar to the values in MeOH. This gives Kf ª 105.4 dm3 mol21 showing that the H2O ligands are more strongly bound to the metal than are the MeOH ligands. From aqueous solutions of 1 precipitation of free L2 was observed at comple ª 5 × 1025 mol dm23.Fig. 1 View of one of the [GdL2 Cl2]+ cations.13 The thermal ellipsoids are drawn at the 50% probability level Fig. 2 Optical absorption spectra of complex 1 (——) and L2 ( ? ? ? ) in MeOH, I = 0.1 mol dm23 NBu4ClO4 both 2.00 × 1025 mol dm23 and of a 2.00 × 1025 mol dm23 aqueous solution of 1 (pH 7.0) at 296 K (inset).The cell length was 1.000 cm338 J. Chem. Soc., Dalton Trans., 1997, Pages 335–339 Luminescence Complex 1 is highly luminescent in the solid state and in water and MeOH solutions. The room-temperature luminescence spectrum of 1 in water is shown in Fig. 4(b). The spectrum, which is similar to that of 1 in MeOH, displays strong, sharp peaks which have been assigned as the 5D0 æÆ 7FJ transitions, the peaks at various J levels (0–6) of the 7FJ manifold.15 The emission spectra of 1 in D2O and CH3OD were identical to that recorded in aqueous solution.In frozen aqueous or MeOH solution at 77 K the emission decreased in intensity and showed only a broad featureless luminescence centred at around 640 nm. Upon thawing, the emission spectra and lifetimes recover completely. We have no simple explanation for these observations. The quantum yields for luminescence of the most intense band (l = 613 nm, assigned to the 5D0 æÆ 7F2 transition) in water and D2O at 296 K were found to be 0.022 ± 0.003 and 0.041 ± 0.003, respectively.The integrated intensity of all the 5D0 æÆ 7F2 transitions gives the respective quantum yields of 0.046 and 0.089. The emission lifetimes determined at 613 nm for solid 1 and for water, D2O, MeOH and CH3OD solutions of 1 are listed in Table 3. The presence of Eu3+(aq), the product of ligand dissociation from the EuL2 complex, does not interfere with the measurements since it is virtually nonluminescent under these conditions (luminescence is quenched Fig. 3 Absorbance at 288 nm vs. absorbance at 305 nm of methanol, I = 0.100 mol dm23 NBu4ClO4, solutions of [EuL2]Cl3?1.5H2O. Circles represent experimental data and the line a least-squares fit using equation (2) Fig. 4 (a) Excitation spectrum (– – – –, lem = 613 nm) of complex 1 in water at 296 K. For comparison the absorption spectrum (——) of 1 in MeOH is shown. (b) Emission spectrum (lex = 320 nm) of 1 in water at 296 K by the H2O ligands).Neither does the presence of free L2 interfere with the quantum-yield measurements since it does not absorb at the excitation wavelength (320 nm, see Fig. 2). Thus, only rigorously unexponential decays were observed. The average number of H2O ligands (nH2O) co-ordinated to the [EuL2]3+ emitter in aqueous solution can be estimated according to the empirical correlation (3) proposed by Horrocks.17a,b nH2O = k[(1/tH2O 296) 2 (1/tD2O 296 )] (3) Here tH2O 296 and tD2O 296 denote the emission lifetime at 296 K in water and D2O solution, respectively, k is a constant, 1.05, empirically determined for EuIII and the value of nH2O is considered to have an uncertainty of ±0.5.Use of equation (3) with the lifetimes listed in Table 3 gave nH2O = 0.5. The number of co-ordinated MeOH molecules in MeOH solution can be calculated analogously by using the lifetimes determined in MeOH and MeOD.17c In this case the number of OH oscillators in a coordinated MeOH molecule is half that of a co-ordinated H2O, thus a different k is used (2.1).In this manner it was shown that nMeOH also equals 0.5. However, the structure of the GdL2 complex predicts nMeOH and nH2O = 2 (after substitution of the chloride ligands). One possible explanation for this discrepancy is the occurrence of structural changes on going from the solid state to a solution, where the unco-ordinated pyridine group present in the solid state is co-ordinated to the metal in solutions (water or MeOH) of complexes EuIII and GdIII of L2.Comparison of the excitation spectrum and absorption spectrum of 1 [Fig. 4(a)] reveals that the relative intensity of the pyridine-centred absorption (265 nm) compared with the bipyridine-centred absorption (320 nm) is much lower than in the absorption spectrum. Thus, the energy transfer from the pyridine groups to the metal is apparently inefficient. In summary, the complex [EuL2]Cl3?1.5H2O is luminescent in the solid state and in MeOH and water solutions.In both solvents the average number of solvent molecules co-ordinated to the metal is 0.5. Compared to the EuL1 system 1a the present study shows that the more flexible L2 ligand more efficiently wraps the metal ion and thereby serves to shield it much more effectively from solvent molecules. Another more complicated ligand with four bipy moieties, two within a hexaazamacrocycle and two in pendant groups, is somewhat better in excluding water from the europium(III) co-ordination sphere, but is also somewhat unstable. 1f Major drawbacks in the EuL2 system are the instability to dissociation in water and the poor antenna properties of the pyridine groups, which results in smaller efficiency in converting incident ultraviolet photons into emitted visible photons. Acknowledgements P. C. F. is indebted to support by the U.S. National Science Foundation (Grant No. CHE-9400919). Laser flash photolysis experiments were carried out on a time-resolved optical system constructed using partial support from a US Department of Energy University Research Instrumentation Grant (No.DEFG05- 91ER79039) and in part from a National Science Foun- Table 3 Metal luminescence lifetimes (lex = 355 nm) in water, D2O, MeOH and MeOD at 296 K and in the solid state at 77 and 296 K [EuL2]Cl3?1.5H2O [EuL1]Cl3?6H2O t296 H2O/ms 0.780 0.34a t296 D2O/ms 1.28 1.70a t296 MeOH/ms 0.980 b t296 MeOD/ms 1.25 b t296 solid/ms 0.80 0.95c t77 solid/ms 1.15 b a Ref. 1(b). b Not reported. c At 300 K, ref. 16.J. Chem. Soc., Dalton Trans., 1997, Pages 335–339 339 dation Instrumentation Grant. A. D. is indebted to the Danish Natural Science Research Council (Grant No. 11-5962) for the UV/VIS spectrophotometer. A. H. is indebted to the Carlsberg Foundation and to the Danish Natural Science Research Council for the diffractometer. References 1 (a) B. Alpha, J.-M. Lehn and G. Mathis, Angew. Chem., Int.Ed. Engl., 1987, 26, 266; (b) B. Alpha, R. Ballardini, V. Balzani, J.-M. Lehn, S. Perathoner and N. Sabbatini, Photochem. Photobiol., 1990, 52, 299; (c) L. Prodi, M. Maestri, R. Ziessel and V. Balzani, Inorg. Chem., 1991, 30, 3798; (d) N. Sabbatini, M. Guardigli, I. Manet and F. Bolletta, Inorg. Chem., 1994, 33, 955; (e) V. Balzani, E. Berghmans, J.-M. Lehn, N. Sabbatini, R. Terörde and R. Ziessel, Helv. Chim. Acta, 1990, 73, 2083; (f) V. Balzani, J.-M. Lehn, N. Sabbatini and R.Ziessel, Angew. Chem., Int. Ed. Engl., 1991, 30, 190; (g) J. Bruno, B. R. Herr and W. DeW. Horrocks, jun., Inorg. Chem., 1993, 32, 756; (h) G. Bernardinelli, C. G. Bochet, J.-C. G. Bünzli, P. Froidevaux and C. Piguet, J. Chem. Soc., Dalton Trans., 1995, 83. 2 V. Balzani and F. Scandola, Supramolecular Photochemistry, Eltis Horwood, Chichester, 1991. 3 I. Bkouche-Waksman, J. Guilhem, C. Pascard, B. Alpha, R. Deschenaux and J.-M. Lehn, Helv. Chim. Acta, 1991, 74, 1163. 4 A. Døssing, A. Hazell and H. Toftlund, Acta Chem. Scand., 1996, 50, 95. 5 S. Parkin, B. Moezzi and H. Hope, J. Appl. Crystallogr., 1995, 28, 53. 6 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliard, M. C. Burla, G. Polidori and M. Camalli, J. Appl. Crystallogr., 1994, 27, 435. 7 R. G. Hazell, LINKON, a Program for Constrained Least-squares Refinement, Aarhus University, Århus, 1995. 8 D. Rogers, Acta Crystallogr., Sect. A, 1981, 37, 734. 9 International Tables for X-Ray Crystallography, Kynoch Press, Birmingham, 1974, vol. 4, pp. 72–98. 10 K. R. Kyle, C. K. Ryu, J. A. DiBenedetto and P. C. Ford, J. Am. Chem. Soc., 1991, 113, 2954; M. M. Mdleleni, J. S. Bridgewater, R. J. Watts and P. C. Ford, Inorg. Chem., 1995, 34, 2334. 11 J. van Houton and R. J. Watts, J. Am. Chem. Soc., 1976, 98, 4853. 12 Y. Haas and G. Stein, J. Phys. Chem., 1971, 75, 3668. 13 C. K. Johnson, ORTEP II, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 14 F. H. Allen, J. E. Davies, J. J. Galloy, O. Johnson, O. Kennard, C. F. Macrae, E. M. Michell, G. F. Michell, J. M. Smith and D. G. Watson, J. Chem. Inf. Comput. Sci., 1991, 31, 187. 15 J.-C. G. Bünzli, Lanthanide Probes in Life, Chemical and Earth Sciences, Elsevier, Amsterdam, 1989, ch. 7. 16 G. Blasse, G. J. Dirksen, D. van der Voort, N. Sabbatini, S. Perathoner, J.-M. Lehn and B. Alpha, Chem. Phys. Lett., 1988, 146, 346. 17 (a) W. DeW. Horrocks, jun. and D. R. Sudnick, J. Am. Chem. Soc., 1979, 101, 334; (b) W. DeW. Horrocks, jun. and D. R. Sudnick, Acc. Chem. Res., 1981, 14, 384; (c) W. DeW. Horrocks, jun., Inorg. Chem., 1991, 30, 3270. Received 13th May 1996; Paper 6/03336I

ISSN:1477-9226    

DOI:10.1039/a603336i

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 339–348 339 Tetranuclear grid-like copper(II) complexes with pyrazolate bridges: syntheses, structures, magnetic and EPR spectroscopic properties Karen L. V. Mann,a Elefteria Psillakis,a John C. JeVery,a Leigh H. Rees,a Nicholas M. Harden,a Jon A. McCleverty,*a Michael D. Ward,*a Dante Gatteschi,b Federico Totti,b Frank E. Mabbs,c Eric J. L. McInnes,c Peter C. Riedi d and Graham M. Smithd a School of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK BS8 1TS.E-mail: mike.ward@bristol.ac.uk b Department of Chemistry, University of Florence, Via Maragliano 75/77, 50144 Florence, Italy c EPSRC cwEPR Service Centre, Chemistry Department, University of Manchester, Manchester, UK M13 9PL d Department of Physics and Astronomy, University of St Andrews, North Haugh, St Andrews, Fife, UK KY16 9SS Received 30th September 1998, Accepted 30th November 1998 Reaction of 3-(2-pyridyl)pyrazole (HL1) and 6-(3-pyrazolyl)-2,29-bipyridine (HL2) with nickel(II) and zinc(II) salts aVorded the simple mononuclear pseudo-octahedral complexes [M(HL1)3][PF6]2 and [M(HL2)2][PF6]2 respectively (M = Ni or Zn) in which the ligands co-ordinate as neutral mononucleating chelates in the same manner as e.g. 2,29-bipyridine or 2,29:69,20-terpyridine respectively. However with CuII the complexes [Cu4(L1)6(solv)2][PF6]2 (solv = dmf or MeOH) and [Cu4(L2)4(dmf)4][PF6]4 were isolated and crystallographically characterised, in all cases containing four tetragonally elongated square-pyramidal copper(II) ions which are linked by pyrazolate bridges from the now deprotonated ligands L1 and L2.The approximate orthogonality of the diVerent ligands within each complex and the approximately square array of metal ions result in a grid-like structure. In [Cu4(L1)6(solv)2][PF6]2 there are successively two, one, two and one pyrazolate bridges between adjacent copper(II) ions around the Cu4 square resulting in two clearly diVerent magnetic coupling pathways; in [Cu4(L2)4(dmf)4][PF6]4 however, which has approximate S4 symmetry, each Cu ? ? ? Cu edge has a single pyrazolate bridge and the coupling pathways are all virtually equivalent. Prolonged drying of these compounds resulted in loss of the axial dmf ligands to give [Cu4(L1)6][PF6]2 and [Cu4(L2)4][PF6]4.Magnetic susceptibility studies on these showed the presence of two antiferromagnetic exchange pathways for [Cu4(L1)6][PF6]2 with J > 172 cm21 and J9 < 155 cm21 (strong correlation between the parameters precludes a more precise determination), but only one antiferromagnetic exchange pathway for [Cu4(L2)4][PF6]4 with J = 63.5 cm21, consistent with the crystal structures of the dmf adducts.The EPR spectra of [Cu4(L1)6][PF6]2 and [Cu4(L2)4][PF6]4 at a variety of frequencies and temperatures can be well simulated as arising from triplet species; however the spectrum of [Cu4(L1)6][PF6]2 also contains a feature which may be ascribed to the expected thermally populated quintet state.Introduction Self-assembly processes involving carefully designed multidentate ligands and metal ions with appropriate stereoelectronic preferences can lead to the eYcient and specific formation of architecturally highly sophisticated polynuclear complexes such as molecular helicates, grids, rings and boxes.1 Although in the first instances many of these structures arose by chance, recently a systematic attempt has been made to determine the relationship between the number, type and spatial disposition of binding sites on the ligand, and the stereoelectronic preferences of the metal ion, which would allow a considerable degree of control to be exerted over the nature of the structure.2 The major interest so far in these types of coordination complex has been in their remarkable structures and the specific self-assembly processes which lead to them.However there is also considerable scope for the study of metal– metal interactions in complexes where several metal ions are assembled in a well defined spatial array linked by suitable bridging ligands to mediate the interaction.3 Electrochemical properties 4 and magnetic 5 exchange interactions have been extensively studied between pairs of metal ions in dinuclear complexes, and the magnetic properties of one-dimensional chain-like complexes 5 and high-nuclearity clusters 6 have recently been of considerable interest for the development of new magnetic materials.In this paper we describe the syntheses, structural characterisation, and magnetic and EPR spectroscopic properties of some unusual grid-like 7 tetranuclear copper(II) complexes prepared from simple multidentate chelating ligands HL1 and HL2 containing pyridyl and pyrazolyl donors which act as bridges via deprotonated pyrazole groups. The assembly of these structures is driven by the stereoelectronic requirement of the copper(II) ion for an elongated tetragonal (as opposed to regular octahedral) geometry.The preferences of CuI and AgI for four-co-ordinate geometry, of most first-row transitionmetals for octahedral geometry, and of lanthanides for nineco- ordinate geometry have all been exploited in the assembly of multinuclear complexes by self-assembly methods.1 These new complexes represent a rare example8 of the exploitation of the unusual structural preference of CuII to direct self-assembly reactions and, to emphasise the point, complexes of the same340 J.Chem. Soc., Dalton Trans., 1999, 339–348 ligands with other first-row transition metal ions, having much simpler structures, are also described. The resultant grid-like copper(II) complexes contain arrays of four copper(II) ions linked by pyrazolate bridges which are known to be eVective mediators of magnetic exchange interactions; 9,10 accordingly, detailed magnetic susceptibility and EPR spectroscopic studies on these complexes are also described.A preliminary communication describing part of this work has been published.11 Experimental General details 3-(2-Pyridyl)pyrazole (HL 1) 12 and 6-(3-pyrazolyl)-2,29-bipyridine (HL2) 13 were prepared according to the published methods. Fast-atom bombardment (FAB) mass spectra were recorded on a VG-Autospec, using 3-nitrobenzyl alcohol as matrix, and electrospray (ES) mass spectra on a VG-Quattro instrument.Syntheses [M(HL1)3][PF6]2 (M 5 Ni or Zn). A mixture of the appropriate metal(II) acetate hydrate and HL1 in a 3 : 1 molar ratio in methanol was stirred to give a clear solution, from which a solid precipitated on addition of aqueous NH4PF6. The solids were filtered oV and dried in vacuo to give the desired products in 60–80% yield. Recrystallisation of both complexes from MeCN–diethyl ether aVorded X-ray quality crystals. Data for [Ni(HL1)3][PF6]2: ES MS [m/z (relative peak intensity, assignment)] 638 [5, {Ni(HL1)3(PF6)}1], 492 [15, {Ni(HL1)2(L1)}1], 347 [100, {Ni(HL1)(L1)}1] and 247 [1%, {Ni(L1)3}21] {Found: C, 35.5; H, 3.0; N, 15.3.Required for [Ni(HL1)3][PF6]2?H2O: C, 35.9; H, 2.9; N, 15.7%}. Data for [Zn(HL1)3][PF6]2: FAB MS [m/z (relative peak intensity, assignment)]: 580 [4, {Zn2(HL1)(L1)2F}], 560 [12, {Zn2- (L1)3}], 498 [6, {Zn(L1)(HL1)2}] and 353 [100%, {Zn(HL1)- (L1)}] {Found: C, 35.5; H, 2.8; N, 15.5. Required for [Zn(HL1)3][PF6]2?H2O: C, 35.6; H, 2.8; N, 15.6%}.[Cu4(L1)6(dmf)2][PF6]2. A mixture of HL1 and Cu(O2CCH3)2? H2O in a 3 : 2 molar ratio in MeOH was stirred for a few minutes at room temperature to aVord a dark green solution. Addition of an excess of aqueous KPF6 aVorded a green precipitate which was filtered oV and dried in vacuo. The complex was crystallised by slow diVusion of diethyl ether vapour into a concentrated solution of the crude material in dmf. X-Ray quality green crystals resulted in ca. 80% yield. FAB MS [m/z (relative peak intensity, assignment)]: 1260 {1, Cu4(L1)7}; 1135 {9, Cu4(L1)6F}, 993 {11, Cu4(L1)5F}, 972 {9, Cu4(L1)5}, 847 {15, Cu4(L1)4F}, 828 {8, Cu4(L1)4}, 784 {85, Cu3(L1)4F}, 765 {100, Cu3(L1)4}, 684 {30, Cu4(L1)3}, 640 {40, Cu3(L1)3F} and 621 {70%, Cu3(L1)3} and numerous smaller fragments (peak masses quoted for 63Cu; the isotopic patterns were consistent with the given formulations) {Found: C, 42.1; H, 3.1; N, 18.1. Required for [Cu4(L1)6(dmf)2][PF6]2: C, 41.7; H, 3.2; N, 18.0%}.After oven drying (70 8C) overnight the two dmf ligands were lost, as shown by IR spectroscopy (see Results and discussion) and N N NH N N NH N HL1 HL2 elemental analysis {Found: C, 41.2; H, 2.7; N, 18.0. Required for [Cu4(L1)6][PF6]2: C, 40.7; H, 3.0; N, 17.8%}. [Cu4(L1)6(MeOH)2][PF6]2. Crystals of this compound appeared when the crude tetranuclear complex, prepared as above, was crystallised from methanol by diVusion of ether vapour into the solution.The FAB mass spectrum was essentially identical to that of the dmf adduct above {Found: C, 40.4; H, 2.9; N, 16.7. Required for [Cu4(L1)6(MeOH)2][PF6]2: C, 40.7; H, 3.0; N, 17.1%}. [M(HL2)2][PF6]2 (M 5 Ni or Zn). A mixture of HL2 (0.060 g, 0.27 mmol) and the appropriate metal(II) acetate hydrate (0.14 mmol) in methanol (10 cm3) was stirred at room temperature to give a clear solution, from which a solid precipitated on addition of aqueous NH4PF6.The solids were filtered oV, dried and recrystallised from MeCN–ether to give microcrystalline powders in 70–80% yield. Data for [Ni(HL2)2][PF6]2: ES MS [m/z (relative peak intensity, assignment)] 501 [100, {Ni(HL2)(L2)}1], 279 [60, {Ni(L2)}1] and 251 [90%, {Ni(HL2)2}21] {Found: C, 38.6; H, 2.3; N, 14.2. Required for [Ni(HL2)2][PF6]2: C, 39.3; H, 2.5; N, 14.1%}. Data for [Zn(HL2)2][PF6]2: ES MS [m/z (relative peak intensity, assignment)] 508 [100, {Zn(HL2)(L2)}1], 286 [35, {Zn(L2)}1] and 254 [40%, {Zn(HL2)2}21] {Found: C, 38.5; H, 2.4; N, 13.7.Required for [Ni(HL2)2][PF6]2: C, 39.0; H, 2.5; N, 14.0%}. [Cu4(L2)4(dmf)4][PF6]4. A mixture of HL2 (0.060 g, 0.27 mmol) and Cu(O2CCH3)2?H2O (0.088 g, 0.44 mmol) in MeOH (20 cm3) was stirred at room temperature until a clear bluegreen solution was obtained. Addition of aqueous NH4PF6 aVorded a blue-green precipitate which was filtered oV, washed with water, and dried. Crystallisation from dmf–ether (as above) aVorded a crystalline precipitate of [Cu4(L2)4(dmf)4]- [PF6]4 in 50% yield.ES MS [m/z (relative peak intensity, assignment)]: 714 [8, {Cu4(L2)4(PF6)2}21] and 284 [100%, {Cu4L4}41] {Found: C, 36.6; H, 2.4; N, 13.2. Required for [Cu4(L2)4][PF6]4: C, 36.6; H, 2.1; N, 13.0% (i.e. the elemental analysis on the vacuum-dried crystals is consistent with loss of the co-ordinated dmf molecules, see Results and discussion)}. X-Ray crystallographic studies Suitable crystals were quickly transferred from the mother liquor to a stream of cold N2 on a Siemens SMART diVractometer fitted with a CCD-type area detector.In all cases data were collected at 2100 8C using graphite-monochromatised Mo-Ka radiation. A detailed experimental description of the methods used for data collection and integration using the SMART system has been published.14 Table 1 contains a summary of the crystal parameters, data collection and refinement. In all cases the structures were solved by conventional direct methods and refined by the full-matrix least-squares method on all F 2 data using the SHELXTL 5.03 package on a Silicon Graphics Indy computer.15 Non-hydrogen atoms were refined with anisotropic thermal parameters; hydrogen atoms were included in calculated positions and refined with isotropic thermal parameters riding on those of the parent atom.The complexes [M(HL1)3][PF6]2?H2O (M = Ni or Zn; Fig. 1) are isostructural and isomorphous, and have no imposed symmetry.Both [Cu4(L1)6(dmf)2][PF6]2?2dmf and [Cu4(L1)6- (MeOH)2][PF6]2?MeOH (Figs. 2–4) are centrosymmetric, so the asymmetric unit contains one half of the complex dication and one independent solvent molecule (dmf or MeOH, respectively); [Cu4(L2)4(dmf)4][PF6]4?6dmf (Fig. 5) has no imposed symmetry, so each asymmetric unit contains an entire complex unit and six independent molecules of dmf. CCDC reference number 186/1268.J. Chem. Soc., Dalton Trans., 1999, 339–348 341 See http://www.rsc.org/suppdata/dt/1999/339/ for crystallographic files in .cif format.Magnetic measurements Magnetic susceptibilities of [Cu4(L1)6(dmf)2][PF6]2 and [Cu4(L2)4(dmf)4][PF6]4 were measured using a Mètronique Ingènièrie MS-03 SQUID magnetometer in the temperature range 1.2–250 K. Both complexes possess an S = 0 ground state; the non-zero value of cT at low temperature is due to the temperature independent paramagnetism (TIP). Magnetic data were interpreted by using the exchange spin Hamiltonians in the forms (1) and (2) for [Cu4(L1)6(dmf)2][PF6]2 H1 = J(S1S2 1 S1aS2a) 1 J9(S1aS2 1 S2aS1) (1) H2 = J(S1S3 1 S1S4 1 S2S3 1 S2S4) = J[(S1 1 S2)(S3 1 S4)] = J?SASB (2) and [Cu4(L2)4(dmf)4][PF6]4, respectively.The eigenvalues for H1 were computed with the CLUMAG program16 while those for H2 were computed in the dimer scheme defining SA = (S1 1 S2), SB = (S3 1 S4) and ST = SA 1 SB. In both cases the computed energies were used in the Van Vleck equation. The magnetic susceptibility data for both complexes were fitted by the theoretical equations by means of an iterative least-squares minimisation routine, and the results of these fits are indicated by the solid lines in Fig. 6.17 In both cases a contribution from a paramagnetic impurity (r) was included in the calculated susceptibility. EPR spectroscopic measurements The EPR spectra on powdered solids were recorded at ca. 9.5, 24.0 and 34.0 GHz using a Bruker ESP300-E spectrometer between 295 and 4.2 K.Spectra of solutions in dmf were also recorded at 110 K using the same instrumentation. The 90 GHz spectra, in the temperature range 100 to 10 K, were recorded on an induction mode spectrometer designed and developed at the University of St Andrews.18 This spectrometer used an 8T Oxford Instruments superconducting magnet. The source was a frequency doubled Gunn diode, phase locked to an EIP frequency counter. The EPR spectrum simulations were performed on a Digital 200/4/233 Alpha Workstation using programs described previously.19 Results and discussion Complexes of HL1 Compound HL1 is a simple bidentate chelating ligand which also has the capacity, via deprotonation of the pyrazole NH group, to act as a dinucleating bridging ligand. We have recently observed both co-ordination modes in complexes with FeIII, and have seen how deprotonation and consequent bridging behaviour can aVord high nuclearity complexes.20 Recently oligomeric complexes of L1 with CuI and AgI have also been reported with the pyrazolate groups bridging.21 This prompted us to investigate the co-ordination behaviour of HL1 with other transition metal cations.Reaction of HL1 with zinc(II) or nickel(II) salts aVords the simple tris-chelates [M(HL1)3]21, here isolated as their hexa- fluorophosphate salts; this behaviour is in accord with that of various related pyridyl/pyrazole bidentate chelating ligands.22 The crystal structure of the complex cation [Zn(HL1)3][PF6]2? H2O is in Fig. 1; it is a simple mononuclear tris-chelate. Signifi- cant bond lengths and angles are collected in Table 2. The asymmetric bidentate ligands are arranged to give the sterically more favourable mer configuration in the pseudo-octahedral complex. One of the pyrazolyl NH groups [N(110)] is involved in a hydrogen bonding interaction to a water molecule, with the non-bonded O(1) ? ? ? N(110) separation being 2.784 Å.The nickel(II) complex [Ni(HL1)3][PF6]2 is isomorphous and isostructural with the zinc(II) complex; its bond lengths and angles are in Table 3. The electronic spectrum shows weak d–d transitions at 530 and 820 nm (e = 54 and 34 dm3 mol21 cm21 respectively) in CH2Cl2 whose position and intensity are entirely consistent with the complex having essentially octahedral geometry; for comparison, the two lower-energy d–d transitions of [Ni(bipy)3]21 are at 520 and 790 nm. The highestenergy d–d transition of [Ni(HL1)3][PF6]2 is hidden by the very strong ligand-centred transition at 286 nm.Thus, with metal ions that either have a stereoelectronic preference for octahedral geometry (e.g. NiII), or which have no particular aversion to it (e.g. ZnII), HL1 behaves as a simple bidentate chelating ligand and remains protonated. Since CuII has a marked stereoelectronic preference for elongated tetragonal geometries, we thought it unlikely that the same type of complex would form between HL1 and CuII.Reaction of HL1 with Cu(O2CCH3)2?2H2O (3 : 2) in methanol at room temperature aVorded a clear deep green solution from which a blue-green solid precipitated on addition of NH4PF6; this was crystallised from dmf–ether. FAB Mass spectrometry showed the presence of numerous peaks corresponding to polynuclear species [up to Cu4(L1)6] and the elemental analysis was consistent with the empirical formula [Cu2(L1)3(dmf)][PF6]. This product was formed using other ligand : metal stoichiometries, but the yield was subsequently optimised by using the required 3 : 2 ratio.The crystal structure (Fig. 2, Table 4) shows that the complex is in fact [Cu4(L1)6(dmf)2][PF6]2?2dmf. There are two approximately planar Cu2(m-L1)2 units, related by an inversion centre, in which each [L1]2 acts as a terdentate bridge linking the two metal centres; these units are stacked parallel and face-to-face (interplane separation 3.2–3.5 Å), with additional deprotonated ligands [L1]2 perpendicular to the two Cu2(m-L1)2 planes forming linking ‘cross-pieces’ between the Cu2(m-L1)2 units.Two of the metals [Cu(1) and Cu(1A)] have additional dmf ligands attached, and therefore have a squarepyramidal N4O environment in which the O ligands (dmf) are axial; Cu(2) and Cu(2A) have square-pyramidal N5 environments in which one of the pyrazole donor atoms is in the axial position. In every case the axial ligand is significantly further from the metal [Cu(1)–O(51), 2.339(2); Cu(2)–N(411A), 2.198(2) Å] than the four equatorial ligands (lengths in the range 1.97–2.07 Å), in keeping with the requirements of the Jahn–Teller eVect.A significant feature of this complex is the ease with which the axial dmf ligands are lost. An IR spectrum of the crystalline material shows strong peaks at 1670 and 1654 cm21, which we assign to the carbonyl stretching vibrations of non-coordinated (lattice) and co-ordinated dmf respectively, in agreement with the crystal structure.For free dmf, nCO can vary from about 1655 to 1695 cm21 depending on the extent of hydrogen Fig. 1 Structure of the complex cation of [Zn(HL1)3][PF6]2?H2O.342 J. Chem. Soc., Dalton Trans., 1999, 339–348 Table 1 Crystallographic data for the five crystal structures Formula M System, space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 Z Dc/g cm23 m/mm21 F(000) Crystal size/mm Reflections collected: total, independent, Rint 2q limits for data/8 Data, restraints, parameters Final R1, wR2 a,b Weighting factors a Largest peak, hole/e Å23 [Zn(HL1)3][PF6]2?H2O C24H23F12N9OP2Zn 808.82 Monoclinic, P21/n 11.5877(14) 12.651(2) 21.630(4) 96.208(13) 3152.3(8) 4 1.704 0.989 1624 0.5 × 0.4 × 0.1 14695, 5537, 0.0353 5–50 5532, 1, 469 0.0603, 0.1699 0.0783, 9.7755 11.405, 20.689 [Ni(HL1)3][PF6]2?H2O C24H23F12N9NiOP2 802.16 Monoclinic, P21/n 11.531(3) 12.642(3) 21.607(5) 96.26(2) 3130.8(13) 4 1.698 0.831 1608 0.2 × 0.1 × 0.1 19307, 7100, 0.0627 4–55 7100, 0, 442 0.0615, 0.1646 0.0685, 3.2831 10.990, 20.509 [Cu4(L1)6(dmf)2][PF6]2?2dmf C60H64Cu4F12N22O4P2 1701.43 Triclinic, P1� 10.410(2) 13.282(3) 14.170(2) 85.254(10) 70.087(12) 68.15(2) 1707.5(6) 1 1.655 1.373 864 0.5 × 0.4 × 0.1 8183, 5820, 0.0227 4–50 5820, 0, 473 0.0360, 0.1056 0.0686, 3.0096 10.703, 20.506 [Cu4(L1)6(MeOH)2][PF6]2?2MeOH C52H52Cu4F12N18O4P2 1537.22 Triclinic, P1� 10.888(2) 12.739(2) 13.050(2) 116.745(8) 101.313(11) 102.246(13) 1488.8(4) 1 1.712 1.560 776 0.3 × 0.2 × 0.1 15396, 6727, 0.0343 4–55 6727, 0, 480 0.0380, 0.0943 0.0492, 0 10.725, 20.903 [Cu4(L2)4(dmf)4][PF6]4?6dmf C82H106Cu4F24N26O10 2449.97 Monoclinic, P21/n 19.567(2) 26.313(3) 20.033(3) 91.441(10) 10311(3) 4 1.578 0.987 5008 0.7 × 0.15 × 0.1 63226, 23223, 0.0793 3–55 23223, 0, 1445 0.0792, 0.2000 0.0641, 21.66 11.005, 20.757 a Structure was refined on Fo 2 using all data; the value of R1 is given for comparison with older refinements based on Fo with a typical threshold of F > 4s(F).b wR2 = [Sw(Fo 2 2 Fc 2)2/Sw(Fo 2)2]� �� where w21 = s2(Fo 2) 1 (aP)2 1 bP and P = [max(Fo 2, 0) 1 2Fc 2]/3.J. Chem. Soc., Dalton Trans., 1999, 339–348 343 bonding to the carbonyl group; 23 the band generally moves to lower energy on co-ordination to a metal ion.24 On grinding the crystals and oven drying at about 70 8C, IR spectra taken at regular intervals showed that the 1670 cm21 band disappears first, followed more slowly by the 1654 cm21 band.After 24 h oven drying there was no trace of any lattice or co-ordinated dmf, and the elemental analysis of the dried material was consistent with this. The fact that crystallisation of this from dmf– ether aVorded again crystals of [Cu4(L1)6(dmf)2][PF6]2?2dmf confirms that loss of axial solvent molecules does not result in decomposition of the cluster core. The sample used for magnetic susceptibility measurements was dried in this way.Crystallisation of the above complex from methanol also resulted in a crystalline material, whose FAB mass spectrum was essentially identical to that of the dmf solvate and whose Fig. 2 Structure of the complex cation of [Cu4(L1)6(dmf)2][PF6]2? 2dmf. Table 2 Selected bond lengths (Å) and angles (8) for [Zn(L1)3][PF6]2? H2O Zn(1)–N(111) Zn(1)–N(311) Zn(1)–N(11) N(111)–Zn(1)–N(311) N(111)–Zn(1)–N(11) N(311)–Zn(1)–N(11) N(111)–Zn(1)–N(211) N(311)–Zn(1)–N(211) N(11)–Zn(1)–N(211) N(111)–Zn(1)–N(31) N(311)–Zn(1)–N(31) 2.129(4) 2.141(4) 2.149(4) 97.00(14) 76.50(14) 169.5(2) 163.72(14) 92.76(14) 95.50(14) 99.27(14) 75.8(2) Zn(1)–N(211) Zn(1)–N(31) Zn(1)–N(21) N(11)–Zn(1)–N(31) N(211)–Zn(1)–N(31) N(111)–Zn(1)–N(21) N(311)–Zn(1)–N(21) N(11)–Zn(1)–N(21) N(211)–Zn(1)–N(21) N(31)–Zn(1)–N(21) 2.160(4) 2.192(4) 2.201(4) 96.9(2) 95.73(14) 91.38(14) 92.08(14) 96.28(14) 75.22(14) 164.71(14) Table 3 Selected bond lengths (Å) and angles (8) for [Ni(HL1)3][PF6]2? H2O Ni(1)–N(31) Ni(1)–N(51) Ni(1)–N(11) N(31)–Ni(1)–N(51) N(31)–Ni(1)–N(11) N(51)–Ni(1)–N(11) N(31)–Ni(1)–N(41) N(51)–Ni(1)–N(41) N(11)–Ni(1)–N(41) N(31)–Ni(1)–N(61) N(51)–Ni(1)–N(61) 2.056(4) 2.074(4) 2.080(4) 95.66(14) 167.53(14) 91.82(14) 78.43(14) 170.72(14) 95.28(14) 96.36(14) 77.8(2) Ni(1)–N(41) Ni(1)–N(61) Ni(1)–N(21) N(11)–Ni(1)–N(61) N(41)–Ni(1)–N(61) N(31)–Ni(1)–N(21) N(51)–Ni(1)–N(21) N(11)–Ni(1)–N(21) N(41)–Ni(1)–N(21) N(61)–Ni(1)–N(21) 2.101(4) 2.127(4) 2.128(4) 94.96(14) 95.67(14) 92.15(14) 92.50(14) 77.54(14) 94.85(14) 167.60(14) elemental analysis also suggested a 3 : 2 ligand : metal ratio.Crystal structure analysis (Fig. 3, Table 5) showed this complex to be [Cu4(L1)6(MeOH)2][PF6]2?2MeOH, essentially identical to [Cu4(L1)6(dmf)2][PF6]2?2dmf but with the dmf molecules (both co-ordinated and free in the lattice) replaced by methanol molecules. The grid-like core of the structure is therefore suf- ficiently robust not to be aVected by changing the relatively labile monodentate solvent ligands.The Cu ? ? ? Cu separations in the methanol solvate (3.939 and 3.984 Å) are similar to those observed in the dmf solvate above (3.963 and 4.080 Å), so the {Cu4(L1)6}21 core is not substantially aVected by the change of co-ordinating solvent. Fig. 4(a) emphasises the grid-like structure of the molecules. We assume that the axial solvent molecules are also labile in this complex, but did not investigate this further.The dinuclear Cu2(m-L1)2 units of these complexes are reminiscent of the dinuclear complexes of the bridging ligand 3,5- bis(2-pyridyl)pyrazole, which forms planar [M2L2]n1 complexes having two deprotonated pyrazolate bridges with various metal ions.10e I from comparison of the structures of the copper(II) complexes of L1 with those of the mononuclear nickel(II) and zinc(II) analogues that the assembly of the gridlike architectures is driven principally by the stereoelectronic preference of the copper(II) ions for elongated tetragonal geometry: to satisfy this requirement necessitates deprotonation of the pyrazole rings and bridging behaviour of the ligand.We note that the particular stereoelectronic preferences of the copper(II) ion have also been exploited recently in directing the assembly of double helicates in which the two ligand strands are diVerent.25 Complexes of HL2 We described the preparation of HL2 recently, but used it as a Fig. 3 Structure of the complex cation of [Cu4(L1)6(MeOH)2][PF6]2? 2MeOH. Table 4 Selected bond lengths (Å) and angles (8) for [Cu4(L1)6- (dmf)2][PF6]2?2dmf Cu(1)–N(410) Cu(1)–N(210) Cu(1)–N(111) Cu(1)–N(11) Cu(1)–O(51) N(410)–Cu(1)–N(210) N(410)–Cu(1)–N(111) N(210)–Cu(1)–N(111) N(410)–Cu(1)–N(11) N(210)–Cu(1)–N(11) N(111)–Cu(1)–N(11) N(410)–Cu(1)–O(51) N(210)–Cu(1)–O(51) N(111)–Cu(1)–O(51) N(11)–Cu(1)–O(51) 1.977(2) 1.980(2) 1.982(2) 2.064(2) 2.339(2) 91.57(9) 170.61(9) 96.64(9) 90.51(9) 168.85(9) 80.57(9) 89.00(9) 98.09(9) 94.35(9) 92.90(9) Cu(2)–N(211) Cu(2)–N(110) Cu(2)–N(31) Cu(2)–N(21) Cu(2)–N(411A) N(211)–Cu(2)–N(110) N(211)–Cu(2)–N(31) N(110)–Cu(2)–N(31) N(211)–Cu(2)–N(21) N(110)–Cu(2)–N(21) N(31)–Cu(2)–N(21) N(211)–Cu(2)–N(411A) N(110)–Cu(2)–N(411A) N(31)–Cu(2)–N(411A) N(21)–Cu(2)–N(411A) 1.968(2) 1.996(2) 2.047(2) 2.073(2) 2.198(2) 96.60(9) 171.35(9) 90.31(9) 80.72(9) 165.66(9) 91.31(9) 105.84(9) 96.37(9) 78.44(9) 97.92(9)344 J.Chem.Soc., Dalton Trans., 1999, 339–348 precursor to a new hexadentate ligand and did not investigate its co-ordination properties.13 By analogy with HL1 it seemed likely that HL2 could behave as a simple terdentate chelating ligand to just one metal centre, or alternatively could act as a dinucleating bridging ligand via deprotonation of the pyrazole. Reaction of HL2 with nickel(II) acetate or zinc(II) acetate in methanol, followed by treatment of the solution with NH4PF6, aVorded complexes whose mass spectra and elemental analyses indicated that they were mononuclear complexes of formulation [M(HL2)2][PF6]2. We could not get X-ray quality crystals of them but they are likely to be unremarkable mononuclear octahedral complexes, and will not be discussed further here.In contrast, reaction with copper(II) acetate followed by precipitation of the complex as its hexafluorophosphate salt aVorded a Fig. 4 Edge-on views of the tetranuclear complex cations, emphasising the grid-like structure: (a) [Cu4(L1)6(MeOH)2]21; (b) [Cu4(L2)4- (dmf)4]41.Table 5 Selected bond lengths (Å) and angles (8) for [Cu4(L1)6- (MeOH)2][PF6]2?2MeOH Cu(1)–N(31) Cu(1)–N(52A) Cu(1)–N(21) Cu(1)–N(41) Cu(1)–N(11) N(31)–Cu(1)–N(52A) N(31)–Cu(1)–N(21) N(52A)–Cu(1)–N(21) N(31)–Cu(1)–N(41) N(52A)–Cu(1)–N(41) N(21)–Cu(1)–N(41) N(31)–Cu(1)–N(11) N(52A)–Cu(1)–N(11) N(21)–Cu(1)–N(11) N(41)–Cu(1)–N(11) 1.983(2) 1.999(2) 2.072(2) 2.076(2) 2.196(2) 97.23(9) 168.30(9) 89.67(9) 80.49(9) 170.12(9) 91.11(9) 110.40(9) 96.73(9) 77.96(8) 93.06(8) Cu(2)–N(51) Cu(2)–N(32A) Cu(2)–N(12) Cu(2)–N(61) Cu(2)–O(1) N(51)–Cu(2)–N(32A) N(51)–Cu(2)–N(12) N(32A)–Cu(2)–N(12) N(51)–Cu(2)–N(61) N(32A)–Cu(2)–N(61) N(12)–Cu(2)–N(61) N(51)–Cu(2)–O(1) N(32A)–Cu(2)–O(1) N(12)–Cu(2)–O(1) N(61)–Cu(2)–O(1) 1.982(2) 1.982(2) 1.991(2) 2.072(2) 2.349(2) 97.34(9) 168.89(9) 90.39(9) 80.60(9) 172.00(9) 90.65(9) 92.80(9) 100.28(9) 93.66(9) 87.57(9) blue-green solid which was crystallised from dmf–ether to give dark green X-ray quality crystals in high yield.Electrospray mass spectrometry indicated formation of a tetranuclear complex in solution, with the peak at the highest m/z value corresponding to {Cu4(L2)4(PF6)2}21. The elemental analysis indicated the empirical formula [Cu(L2)][PF6], i.e. a 1 : 1 metal : ligand ratio. The crystal structure (Fig. 5, Table 6) revealed the complex to be the tetramer [Cu4(L2)4(dmf)4][PF6]4?6dmf. This has many structural similarities to [Cu4(L1)6(solv)2][PF6]2 (solv = dmf or MeOH) (above). The overall structure is that of a 2 × 2 grid, with two pairs of parallel, stacked [L2]2 ligands mutually perpendicular to each other.The stacking distances between overlapping aromatic ligand fragments again are in the range 3.2–3.5 Å. Each metal ion is co-ordinated by the ter- Fig. 5 Structure of the complex cation of [Cu4(L2)4(dmf)4][PF6]4? 6dmf. Table 6 Selected bond lengths (Å) and angles (8) for [Cu4(L2)4- (dmf)4][PF6]4?6dmf Cu(1)–N(102) Cu(1)–N(21) Cu(1)–N(11) Cu(1)–N(31) Cu(1)–O(1) Cu(2)–N(72) Cu(2)–N(51) Cu(2)–N(41) Cu(2)–N(61) Cu(2)–O(2) N(102)–Cu(1)–N(21) N(102)–Cu(1)–N(11) N(21)–Cu(1)–N(11) N(102)–Cu(1)–N(31) N(21)–Cu(1)–N(31) N(11)–Cu(1)–N(31) N(102)–Cu(1)–O(1) N(21)–Cu(1)–O(1) N(11)–Cu(1)–O(1) N(31)–Cu(1)–O(1) N(72)–Cu(2)–N(51) N(72)–Cu(2)–N(41) N(51)–Cu(2)–N(41) N(72)–Cu(2)–N(61) N(51)–Cu(2)–N(61) N(41)–Cu(2)–N(61) N(72)–Cu(2)–O(2) N(51)–Cu(2)–O(2) N(41)–Cu(2)–O(2) N(61)–Cu(2)–O(2) 1.948(5) 1.956(5) 2.002(5) 2.049(5) 2.243(4) 1.952(4) 1.953(5) 2.012(5) 2.049(5) 2.270(4) 168.4(2) 102.5(2) 79.7(2) 96.0(2) 79.7(2) 157.6(2) 92.1(2) 98.8(2) 99.1(2) 92.7(2) 171.6(2) 103.8(2) 79.4(2) 95.6(2) 79.7(2) 156.8(2) 91.1(2) 96.1(2) 100.2(2) 92.0(2) Cu(3)–N(42) Cu(3)–N(111) Cu(3)–N(101) Cu(3)–N(121) Cu(3)–O(3) Cu(4)–N(12) Cu(4)–N(81) Cu(4)–N(71) Cu(4)–N(91) Cu(4)–O(4) N(42)–Cu(3)–N(111) N(42)–Cu(3)–N(101) N(111)–Cu(3)–N(101) N(42)–Cu(3)–N(121) N(111)–Cu(3)–N(121) N(101)–Cu(3)–N(121) N(42)–Cu(3)–O(3) N(111)–Cu(3)–O(3) N(101)–Cu(3)–O(3) N(121)–Cu(3)–O(3) N(12)–Cu(4)–N(81) N(12)–Cu(4)–N(71) N(81)–Cu(4)–N(71) N(12)–Cu(4)–N(91) N(81)–Cu(4)–N(91) N(71)–Cu(4)–N(91) N(12)–Cu(4)–O(4) N(81)–Cu(4)–O(4) N(71)–Cu(4)–O(4) N(91)–Cu(4)–O(4) 1.952(5) 1.955(5) 2.009(5) 2.056(5) 2.230(4) 1.949(5) 1.961(5) 2.011(4) 2.041(4) 2.226(4) 170.2(2) 105.2(2) 79.4(2) 94.3(2) 79.4(2) 156.4(2) 89.6(2) 98.1(2) 100.5(2) 92.8(2) 171.9(2) 103.6(2) 79.4(2) 96.5(2) 79.0(2) 156.1(2) 88.7(2) 98.3(2) 100.4(2) 92.8(2)J.Chem. Soc., Dalton Trans., 1999, 339–348 345 dentate pocket of one deprotonated ligand [L2]2, one pyrazole donor atom which is acting as a bridge to a ligand attached to another metal ion, and a dmf ligand. The result is an elongated N4O square-pyramidal geometry with the dmf ligand in the axial position; the axial bond lengths lie in the range 2.23– 2.27 Å, in contrast to the equatorial ones which lie in the range 1.95–2.06 Å.The four metal ions are crystallographically independent but chemically very similar, and the complex has approximate S4 symmetry {cf. the S2 axis of [Cu4(L1)6(dmf)2]- [PF6]2 implied by its inversion centre}. Fig. 4(b) shows an edgeon view of the complex cation. The four copper(II) ions are not now coplanar, but form a butterfly-like arrangement with the metal ions around the Cu4 ring alternately above and below their mean plane. As with the copper(II) complexes of L1, it is clear that the formation of this structure is driven by the stereoelectronic preference of copper(II) ions for an elongated tetragonal geometry.Again the dmf ligands are labile; the initially crystallised material has a broad, strong carbonyl signal at 1656 cm21 with a high-energy shoulder just discernible, which we assign to coordinated dmf and free (lattice) dmf molecules respectively, in agreement with the crystal structure. On heating these completely disappear, and the elemental analysis of the resulting material is consistent with the formulation [Cu4(L2)4][PF6]4; this is the material that was used for magnetic studies (below).Magnetic susceptibility studies The complexes [Cu4(L1)6][PF6]2 and [Cu4(L2)4][PF6]4, in which the axial solvent ligands have been removed, were subjected to magnetic susceptibility measurements in the temperature range 1.2–250 K; plots of c vs. T are given in Fig. 6. The room temperature value of c for [Cu4(L1)6][PF6]2 is much smaller than expected for uncoupled spins, suggesting dominant antiferromagnetic coupling within the cluster.This is confirmed by the broad maximum in the susceptibility observed at ca. 170 K Fig. 6 Plots of c vs. T for (a) [Cu4(L1)6][PF6]2 and (b) [Cu4(L2)4]- [PF6]4. The circles are measured data; the line is the calculated fit based on the parameters given in the text. [Fig. 6(a)]. At low temperature the compound is essentially diamagnetic.The increase in c below 20 K is presumably due to some paramagnetic impurity. A maximum is observed also in the susceptibility of [Cu4(L2)4][PF6]4 at ca. 50 K, again suggesting a dominant antiferromagnetic coupling, albeit weaker than in [Cu4(L1)6][PF6]2. The best fit values derived from the susceptibility data for [Cu4(L2)4][PF6]4 are g = 2.35, r = 6.5%, J = 63.5 cm21.† Although the complex has no crystallographically imposed symmetry and all four Cu ? ? ? Cu couplings could therefore be slightly diVerent, the susceptibility data can be satisfactorily accounted for by a single antiferromagnetic coupling constant of 63.5 cm21 along each edge of the Cu4 ring.The complex accordingly has an S = 0 ground state. The (approximate) S4 symmetry of the complex means that the magnetic d(x2 2 y2) orbital on each metal is eVectively at 908 to each of its neighbours, although each pyrazolate bridge does span two magnetic orbitals. This is illustrated in Fig. 7(b). At Cu(1) the plane of the magnetic orbital is defined by the equatorial donor atoms N(102), N(11), N(21) and N(31); i.e. it is being viewed ‘edge-on’ in the Figure. At Cu(3) the four equatorial donor atoms are N(101), N(42), N(121) and N(111), i.e. in the plane of the paper. Thus the two magnetic orbitals are spatially orthogonal to one another, Fig. 7 Excerpts from the crystal structures of (a) [Cu4(L1)6- (dmf)2][PF6]2 and (b) [Cu4(L2)4(dmf)4][PF6]4, emphasising the geometry of the bridging groups.† We note that the g values derived from the magnetic susceptibility data do not agree well with those derived from EPR spectra, which are far more reliable. This is quite common and arises because a g value derived from magnetic susceptibility data acts as a sink for all of the systematic errors in the curve fitting, and therefore has little signifi- cance. For [Cu4(L1)6][PF6]2 the strong correlation between J and J9 inevitably causes problems with the fitting.In [Cu4(L2)4][PF6]4 we have assumed that all four coupling pathways are equivalent, despite the fact that they are crystallographically slightly inequivalent. In addition it is entirely possible that the co-ordinatively unsaturated copper(II) centres could pick up axial water ligands from the air, at the sites vacated by the dmf ligands after drying: this would result in slight errors in the molecular weights used and in the diamagnetic correction, both of which would aVect the derived g values (but not the J values).346 J.Chem. Soc., Dalton Trans., 1999, 339–348 although they are of the same symmetry species and can therefore mix, and they are both linked by atoms N(101) and N(102) of the pyrazolate bridge. The result here is antiferromagnetic exchange, although it is rather weak. This is in agreement with the magnetic behaviour of a related tetranuclear copper(II) complex which has the same symmetry properties, viz. a diazene bridge linking mutually perpendicular magnetic orbitals in a tetranuclear complex of S4 symmetry: the antiferromagnetic coupling constant in this case was even weaker, at 12.2 cm21.26 Matters were slightly more complex in the fitting procedure of the susceptibility of [Cu4(L1)6(dmf)2][PF6]2 because the two parameters J and J9 which are needed are strongly correlated.The observed maximum at ca. 170 K suggests that there is at least one antiferromagnetic exchange constant J of ca. 170 cm21, but no direct information is available on the second one. In fact sample calculations showed that acceptable fits can be obtained, either with two similar values for J and J9, or with J > 170 cm21 and J9 < 170 cm21. The best fit values are g = 2.23, r = 5.9%, J = 172 cm21, J9= 155 cm21.† We interpret these values as a lower limit for J and as an upper limit for J9, and assume that the larger value is associated with the double pyrazolate bridge. The fact that both interactions are antiferromagnetic means that [Cu4(L1)6][PF6]2 also has an S = 0 ground state in which the four unpaired electrons will alternate in orientation around the Cu4 ring.The stronger coupling (J > 172 cm21) is within each doubly bridged {Cu2(L1)2} plane, i.e. between Cu(1) and Cu(2), and likewise between Cu(1A) and Cu(2A). It is to be expected that a strong antiferromagnetic coupling would occur in these cases as the magnetic d(x2 2 y2) orbitals are coplanar and overlap with the s orbitals of the coplanar bridging pyrazolate fragments.This is a common type of structure 9,10 and it is well understood how antiferromagnetism arises in such cases.9 The weaker antiferromagnetic coupling of J9 < 155 cm21 is between the singly bridged pair Cu(1) and Cu(2A), and likewise between Cu(2) and Cu(1A). The magnitude of this is less easy to understand because (i) there is only one bridging pyrazolate group, and (ii) it does not appear to interact with both magnetic orbitals on the two metal centres that it bridges.The relevant section of the crystal structure is shown in Fig. 7(a); whereas one of the pyrazolate donors [N(41A)] overlaps with the d(x2 2 y2) orbital of Cu(1A), the second [N(41B)] co-ordinates to Cu(2) along its z axis, orthogonal to the magnetic orbital. There are two possible answers to this problem. First, an additional coupling pathway exists via the pyridyl donor: the d(x2 2 y2) orbital on Cu(2) interacts with the pyridyl donor N(31) even though it does not interact with the pyrazolyl donor N(41B).Thus there is an additional Cu(1A)–pyrazolyl(N41A)– pyridyl(N31)–Cu(2) pathway which could provide a contribution to antiferromagnetic exchange. Secondly, the geometries about the copper centres are slightly distorted from regular square pyramidal. This provides a mechanism for some mixing of the d(x2 2 y2) and d(z2) orbitals, which are only orthogonal in high symmetries, such that the ‘axial’ pyrazolate donor N(41B) will interact with the unpaired electron on Cu(2) to some extent.However we emphasise that the derived value of J9 is only an upper limit: because of the strong correlation between J and J9, the actual value could be considerably lower. EPR spectroscopic studies The EPR spectra of [Cu4(L1)6][PF6]2 as a powder at room temperature consist mainly of a broad feature in the “g = 2” region at X-, K- and Q-band frequencies (Fig. 8). On cooling to 100 K these spectra showed some resolution, particularly at the higher frequencies. At 100 K a weak half-field feature was found in the spectra at all three frequencies, which is most apparent in Fig. 8(a) (X-band). Further cooling at K-band, and also spectra below 100 K at W-band, resulted in an overall decrease in the spectrum intensity accompanied by a relatively more rapid decrease in the feature marked * in the K-band spectrum at 100 K [Fig. 8(b)]. The main features in the spectra at 100 K at all four frequencies were well simulated as a spin-triplet spectrum, using the same set of spin-Hamiltonian parameters [gx = 2.220, gy = 2.060, gz = 2.050, D = 0.050 cm21, l (= E/D) = 0.22]; see Figs. 9 and 10 for two representative examples. The axis for gz and the principal axis of the zero-field splitting (Dzz) are assumed to be coincident. We have been unable to simulate the feature marked *, which we therefore tentatively attribute to the expected quintet state (see below).The good simulations at four diVerent frequencies indicates that the parameters for the spintriplet state are reliable. In addition, the temperature variation of the intensity of the spin-triplet spectrum confirms that this is not the ground state, in agreement with the magnetic susceptibility results. For [Cu4(L2)4][PF6]4 the spectra at X-band are less well resolved than those for [Cu4(L1)6][PF6]2. For this reason we concentrated on the spectra at both Q- and W-band (90 GHz).Representative spectra at both of these frequencies are in Figs. 11 and 12. These spectra are well simulated assuming that they arise from a spin-triplet state. At any given temperature the spin-Hamiltonian parameters required for the simulations at both frequencies are the same within experimental error (at 100 K, g|| = 2.033, g^ = 2.130, D = 0.017 cm21, l = 0). However, Fig. 8 Experimental spectra of a powder of [Cu4(L1)6(dmf)2][PF6]2 at (a) X-band, (b) K-band and (c) Q-band.In each case the upper spectrum is at 295 K and the lower is at 100 K. Fig. 9 Experimental (lower line) and simulated (upper line) EPR spectra of [Cu4(L1)6][PF6]2 as a powder at X-band and 100 K.J. Chem. Soc., Dalton Trans., 1999, 339–348 347 there appears to be a small increase in D as the temperature decreases (0.017 cm21 at 100 K to 0.030 cm21 at 10 K). These spectra are axial, with the rather unusual situation that g|| < g^. It is noticeable that the g|| value is very similar to that of the smallest g value of [Cu4(L1)6][PF6]2, whereas the g^ value is approximately the average of the other two g values for [Cu4(L1)6][PF6]2.This is understandable given the approximate S4 symmetry of the complex. If the principal axes of Cu(1) are taken to be along the metal–ligand bonds (assuming idealised symmetry with 908 bond angles), then we have Cu(1)–N(21), Cu(1)–N(31) and Cu(1)–O(1) as the axes. The first of these axes is co-parallel with the equivalent axis on the other three metal centres, but the other two axes are rotated by 908 on moving from one metal centre to the next and therefore interchange.It therefore is reasonable that in a coupled system the g value corresponding to the Cu(1)–N(21) axis at each site is unique, whereas the other two would be averaged and therefore equivalent. This accounts for the appearance of an axial spectrum and for the magnitudes of the g values. Similar g values, but associated with a much larger zero field splitting, were previously observed for a tetranuclear triazolato bridged copper(II) complex. 26 The present values seem to be closer to the dipolar contribution.Fig. 10 Experimental (lower line) and simulated (upper line) EPR spectra of [Cu4(L1)6][PF6]2 as a powder at K-band and 100 K. Fig. 11 Experimental (lower line) and simulated (upper line) EPR spectra of [Cu4(L2)4][PF6]4 under the following conditions: (a) powder spectrum, Q-band, 100 K; (b) powder spectrum, W-band, 100 K.The most surprising feature of the spectra of both complexes is that they are dominated by a triplet state; there is no unambiguous evidence of the expected quintet state which, given the values of J, we would expect to be thermally populated except at extremely low temperatures. However, there is the feature marked * in the spectra of [Cu4(L1)6][PF6]2 (Figs. 8 and 10) which is unaccounted for by the simulations.This feature is not due to solid-state eVects since it is present in the frozen solution spectra at 100 K; also, it is not due to monomeric impurities since its intensity decreases with decreasing temperature, the reverse of what would be expected from a monomeric centre. The temperature variation of the relative intensity of this feature suggests that it could belong to a spin state which is at a higher energy than the spin-triplet state, i.e. it could be from the expected spin quintet.This situation is analogous to that reported by Chaudhuri et al.27 wherein they inferred the presence of a spin-quintet state from an analysis of the linewidth and intensity variation with temperature of the X-band powder spectrum of [Cu4L4(Im)4][ClO4]4?2H2O, where L = 1,4,7-triazacyclononane and Im = the imidazolate anion. Conclusion The mixed pyridine–pyrazole ligands HL1 and HL2 form simple mononuclear octahedral complexes with NiII and ZnII in which the pyrazole remains protonated. With CuII however the requirement for an axially elongated geometry precludes this co-ordination mode, with the result that L1 and L2 act as anionic bridging ligands via deprotonation of the pyrazolate groups to give tetranuclear grid-like complexes in which adjacent copper(II) ions are linked by one or two pyrazolate bridges in various geometries.Variable-temperature magnetic susceptibility studies on these complexes show that (i) there is antiferromagnetic exchange between each pair of adjacent copper(II) ions in every case resulting in S = 0 ground states; (ii) the magnitude of the antiferromagnetic exchange depends on Fig. 12 Experimental (lower line) and simulated (upper line) EPR spectra of [Cu4(L2)4][PF6]4 under the following conditions: (a) powder spectrum, Q-band, 20 K; (b) powder spectrum, W-band, 25 K.348 J. Chem. Soc., Dalton Trans., 1999, 339–348 both the number of pyrazolate bridges and the relative orientations of the magnetic orbitals on the copper(II) ions concerned.The EPR spectroscopic measurements at a variety of frequencies and temperatures give spectra characteristic of triplet species, with (in one case) a feature ascribable to the thermally populated quintet state also being apparent. Acknowledgements We thank the EPSRC (UK), and Ministero dell’Università e della Ricerca Scientifica e Tecnologia and Consiglio Nazionale delle Ricerche (Italy), for financial support.Miss Claire White is also thanked for assistance with one of the crystal structures. References 1 D. Philp and J. F. Stoddart, Angew. Chem., Int. Ed. Engl., 1996, 35, 1155; J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim, 1995; C. Piguet, G. Bernardinelli and G. Hopfgartner, Chem. Rev., 1997, 97, 2005. 2 T. Beissel, R. E. Powers and K. N. Raymond, Angew. Chem., Int. Ed. Engl., 1996, 35, 1084. 3 A. J. Amoroso, A. M. W. Cargill Thompson, J. P. Maher, J. A. McCleverty and M.D. Ward, Inorg. Chem., 1995, 34, 4828; V. A. Ung, A. M. W. Cargill Thompson, D. A. Bardwell, D. Gatteschi, J. C. JeVery, F. Totti and M. D. Ward, Inorg. Chem., 1997, 36, 3447. 4 M. D. Ward, Chem. Soc. Rev., 1995, 121. 5 O. Kahn, Molecular Magnetism, VCH, Weinheim, 1993. 6 V. A. Grillo, Z. Sun, K. Folting, D. N. Hendrickson and G. Christou, Chem. Commun., 1996, 2233; V. A. Grillo, M. J. Knapp, J. C. Bollinger, D. N. Hendrickson and G. Christou, Angew. Chem., Int. Ed. Engl., 1996, 35, 1818; H.J. Eppley, H.-L. Tsai, N. de Vries, K. Folting, G. Christou and D. N. Hendrickson, J. Am. Chem. Soc., 1995, 117, 301; M. A. Bolcar, S. M. J. Aubin, K. Folting, D. N. Hendrickson and G. Christou, Chem. Commun., 1997, 1485. 7 Other examples of ‘molecular grids’: P. N. W. Baxter, J.-M. Lehn, J. Fischer and M.-T. Youinou, Angew. Chem., Int. Ed. Engl., 1994, 33, 2284; P. N. W. Baxter, J.-M. Lehn, B. O. Kneisel and D. Fenske, Chem. Commun., 1997, 2231; Angew. Chem., Int.Ed. Engl., 1997, 36, 1978; G. S. Hanan, D. Volkmer, U. S. Schubert, J.-M. Lehn, G. Baum and D. Fenske, Angew. Chem., Int. Ed. Engl., 1997, 36, 1842; C. Duan, Z. Liu, X. You, F. Xue and T. C. W. Mak, Chem. Commun., 1997, 381. 8 V. C. M. Smith and J.-M. Lehn, Chem. Commun., 1996, 2733. 9 V. P. Hanot, T. D. Robert, J. Kolnaar, J. G. Haasnoot, J. Reedijk, H. Kooijman and A. L. Spek, J. Chem. Soc., Dalton Trans., 1996, 4275 and refs. therein. 10 (a) T. Otieno, S. J. Rettig, R.C. Thompson and J. Trotter, Inorg. Chem., 1995, 34, 1718; (b) J. Casabó, J. Pons, K. S. Siddiqi, F. Teixidor, E. Molins and C. Miravitlles, J. Chem. Soc., Dalton Trans., 1989, 1401; (c) J. Pons, X. López, J. Casabó, F. Teixidor, A. Caubet, J. Ruis and C. Miravitlles, Inorg. Chim. Acta, 1992, 195, 61; (d ) J. Pons, F. J. Sánchez, A. Labarta, J. Casabó, F. Teixidor and A. Caubet, Inorg. Chim. Acta, 1993, 208, 167; (e) M. Munakata, L. P. Wu, M. Yamamoto, T. Kuroda-Sowa, M. Maekawa, S. Kawata and S. Kitagawa, J. Chem. Soc., Dalton Trans., 1995, 4099; ( f ) P. M. Slangen, P. J. van Koningsbruggen, K. Goubitz, J. G. Haasnoot and J. Reedijk, Inorg. Chem., 1994, 33, 1121. 11 J. C. JeVery, P. L. Jones, K. L. V. Mann, E. Psillakis, J. A. McCleverty, M. D. Ward and C. M. White, Chem. Commun., 1997, 175. 12 A. J. Amoroso, A. M. W. Cargill Thompson, J. C. JeVery, P. L. Jones, J. A. McCleverty and M. D. Ward, J. Chem. Soc., Chem. Commun., 1994, 2751; H. Brunner and T. Scheck, Chem. Ber., 1992, 125, 701. 13 J. S. Fleming, E. Psillakis, S. M. Couchman, J. C. JeVery, J. A. McCleverty and M. D. Ward, J. Chem. Soc., Dalton Trans., 1998, 537. 14 P. L. Jones, A. J. Amoroso, J. C. JeVery, J. A. McCleverty, E. Psillakis, L. H. Rees and M. D. Ward, Inorg. Chem., 1997, 36, 10. 15 SHELXTL 5.03 program system, Siemens Analytical X-Ray Instruments, Madison, WI, 1995. 16 D. Gatteschi and L. Pardi, Gazz. Chim. Ital., 1993, 123, 231. 17 W. H. Press, B. P. Flannery, S. A. Teukolsky and W. T. Vetterling, Numerical Recipes, Cambridge University Press, 1968. 18 G. M. Smith, J. C. G. LeSurf, R. H. Mitchell and P. C. Riedi, MTT Symposium Proceedings, Orlando, 1995, Rev. Sci. Instrum., 1998, 69, 3924. 19 F. E. Mabbs and D. Collison, Electron paramagnetic resonance of d-transition metal compounds, Elsevier, Amsterdam, 1992, Chs. 7 and 16. 20 P. L. Jones, J. C. JeVery, J. A. McCleverty and M. D. Ward, Polyhedron, 1997, 16, 1567. 21 K. Singh, J. R. Long and P. Stavropoulos, J. Am. Chem. Soc., 1997, 119, 2942. 22 W. R. Thiel and T. Priermeier, Angew. Chem., Int. Ed. Engl., 1995, 34, 1737; W. R. Thiel, M. Angstl and T. Priermeier, Chem. Ber., 1994, 127, 2373; J. Sieler and H. Hennig, Z. Anorg. Allg. Chem., 1971, 381, 219; Y. Luo, P. G. Potvin, Y.-H. Tse and A. B. P. Lever, Inorg. Chem., 1996, 35, 5445. 23 G. Eaton and M. C. R. Symons, J. Chem. Soc., Faraday Trans. 1, 1988, 84, 3459. 24 D. A. Bardwell, J. C. JeVery and M. D. Ward, Inorg. Chim. Acta, 1995, 236, 125 and refs. therein. 25 V. C. M. Smith and J.-M. Lehn, Chem. Commun., 1996, 2733. 26 A. Bencini, D. Gatteschi, C. Zanchini, J. G. Haasnoot, R. Prins and J. Reedijk, J. Am. Chem. Soc., 1987, 109, 2926. 27 P. Chaudhuri, I. Karpenstein, M. Winter, M. Lengen, C. ButzlaV, E. Bill, A. X. Trautwein, U. Flörke and H.-J. Haupt, Inorg. Chem., 1993, 32, 888. Paper 8/07599I

ISSN:1477-9226    

DOI:10.1039/a807599i

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 341–346 341 Complexation of transition-metal ions, SnII, PbII and AlIII with nucleobase-substituted polyethers and dissociation of adduct ions studied by fast atom bombardment mass spectrometry † Mandapati Saraswathi and Jack M. Miller * Department of Chemistry, Brock University, St. Catharines, ONT L2S 3A1, Canada Metal-ion-complexed dinucleotide analogues were studied by FAB mass spectrometry using tri- and bi-valent transition-metal ions (Cr to Zn), SnII, PbII and AlIII.Pairs of thymine nucleobases linked by (CH2)2- [O(CH2)2]nO(CH2)2 chains (n = 1 1, 2 2, 3 3 or 4 4) and of adenine nucleobases (n = 1 5 or 4 6) co-ordinate to metal ions and form significant [M + M 2 H]+ and [M + MCl]+ ions. In addition, many transition-metal ions also give [M + M2X 2 2H]+ (X = Cl or NO3) ions but CrIII produces [M + CrCl2]+ ions. The effect of the spacer chain length is reflected in the ease of formation of [M + M 2 H]+ and [M + MX]+ ions.Fragmentation of [M + MX]+ and [M + M2X 2 2H]+ ions give [M + M 2 H]+ ions, suggesting that metal chelation through the nucleobase is more favoured than through the polyether. Complexes of compounds 1–6 with AlIII–glycerol give abundant [M + 117]+ complex ions (117 = Al + glycerol 2 2H). The intensities of these ions decrease with increasing number of ethylene oxide units from 1–4. The same trend is also observed in 5 and 6. Bimetallic ions with an unipositive charge are also produced from 1–4.Dissociation of adduct ions labelled with deuterium confirms the substitution of amide protons with aluminium(III) in the elimination of glycerol. Hydrogen-bonding interactions are crucial elements for catalysis and molecular recognition in biosystems. Many receptors have been synthesized to investigate the role of these interactions through the recognition of biologically important guests such as nucleobases.1 Model molecules derived from crown ethers and nucleotide bases can exist with Watson–Crick hydrogen-bonding interactions in a self-assembly process in solution.2 Gas-phase experiments provide a powerful tool for probing these interactions in the absence of a solvent.The FAB mass spectra of polyethers substituted with nucleobases showed the existence of intramolecular Watson–Crick interactions which are more effective between adenine and thymine than between thymine and thymine.3 Metal ions play an important role in biological systems and are often required as cofactors for enzymes.The metalcomplexing properties of crown ethers have biological as well as chemical significance.4 The chelation ability of alkali-metal ions with nucleobase-substituted acylic glymes falls between that of crown ethers and polyethylene glycols.3 Hydrogen bonds between two nucleobases keep the acylic system in a cyclic form, so as to encapsulate the metal ions more efficiently than their counterparts. Co-ordination of alkali-metal 3 and alkaline-earth-metal ions 5 with the nucleobases in the selvedge (the high pressure region just above the surface of the sputtered material in desorption experiments) was observed, in addition to complexation with ether oxygens.However, the chelation of alkali-metal ions through a carbonyl group of the nucleobase led to the hydrolysis of nucleobase.3 When we used alkali-earth-metal ions for complexation we observed intramolecular covalent bonding of metal between two nucleobases (N]M]N).5 In the present study, we discuss the interaction of FeIII, CrIII, CrII, MnII, FeII, CoII, NiII, CuII, ZnII, SnII, PbII and AlIII with these modified nucleotides and the dissociation of the metal complexes in the first field-free region of a conventional (EB geometry) mass spectrometer.Results and Discussion FeCl3–Thioglycerol Relative abundances of metal-containing ions obtained from dinucleotide analogues 1–6 using iron(III) chloride and a thioglycerol (HOCH2CHOHCH2SH) matrix are given in Table 1.The formation of metal-chelated ions such as [M + Fe 2 H]+, [M + FeCl]+ and [M + Fe2Cl 2 2H]+ indicates reduction of the metal from the +3 to the +2 state in the presence of thioglycerol. Thioglycerol also gives more abundant metal-chelated ions than does glycerol, as we have already observed for the complexation of alkaline-earth-metal ions.5 The [M + Fe 2 H]+ ion results from a covalent interaction between the amide nitrogen of the pyrimidine base and a metal ion.Similarly, [M + Fe2Cl 2 2H]+, a unipositive ion, may be formed by the covalent binding of one metal to two pyrimidine bases within the molecule. The other FeCl+ can co-ordinate with the ether oxygens. Polypyridyl and chiral bis(phenanthroline) complexes of ruthenium(II) have been shown to bind covalently to nitrogen † Non-S1 unit employed: Torr ª 133 Pa.342 J. Chem. Soc., Dalton Trans., 1997, Pages 341–346 Table 1 Relative abundances of organometallic ions from compounds 1–6 with FeCl3–thioglycerol * m/z (%) Compound [M + Fe 2 H]+ [M + FeCl]+ [M + Fe2Cl22H]+ Others 123456 14 25 34 57 22 79 88 40 12 6 37 4 20 22 15 21 43 582 (5), 583 (7), 637 (5), 787 (3), 841 (8) 749 (11), 839 (8) —— 556 (15) [M + Fe2Cl2 2 H]+ 688 (14) [M + Fe2Cl2 2 H]+ * m/z 153, [T?C2H4]+, is the base peak from compounds 1–4 and 6, [M + H]+ ion that from 5.Table 2 Relative abundances of organometallic ions from compounds 1–6 with CrCl3–thioglycerol * m/z (%) Compound [M + Cr 2 H]+ [M + CrCl 2 H]+ [M + CrCl]+ [M + CrCl2]+ Others 123456 22 21 22 24 12 17 10 10 12 345 26 11 10 443 18 788 —4 416 (10), 524 (26), 574 (20) 460 (5), 568 (14), 618 (12), 745 (6) 504 (4), 612 (10), 662 (10) 548 (8), 635 (9), 670 (6), 706 (9) — 665 (6), 715 (3), 930 (3) * m/z 153, [T?C2H4]+, is the base peak from compounds 1–4 and [M + H]+ ion that from 5 and 6.bases.6 Complexation of the alkali- and alkaline-earth-metal ions also demonstrated that metal ions can displace amide protons, 3,5 which is not a favourable reaction in solution.7 Similarly, in the gas phase, metal chelation of alkali-, alkaline-earth- and transition-metal ions by peptides is via covalent bonding to amide nitrogens.8 The abundance of the [M + Fe 2 H]+ ion increases with an increase in the length of the alkyl spacer chain between two nucleobases, while the abundance of [M + FeCl]+ ions declines.A similar trend is observed for 5 and 6 also.These observations suggest that in 1 and 2 there is more intramolecular hydrogenbonding. This keeps the system in a cyclic form for more effective chelation of the FeCl+ ion by the polyether in 1 and 2 compared to 3 and 4. These observations also suggest that the displacement of a hydrogen-bonded amide proton requires more energy for the substitution process than a free amide proton. The same phenomenon plays a role in solutions where alkali-metal ions cannot deprotonate amide nitrogens. Since the formation of the [M + Fe2Cl 2 2H]+ ion is a combination of metal chelation through the nucleobase and the polyether chain, there is no such variation in the relative abundances of this ion with an increase in the spacer chain length.However, in 5 and 6, the substitution of one proton by a metal ion is a major process and this results in the formation of significant amounts of the [M + Fe2Cl2 2 H]+ ion. The higher proton affinity of adenine 9 present in the complex results in the formation of abundant protonated molecules. In addition, the complexation of metal ions by the matrix is also observed in these spectra.Deuterium oxide exchange experiments with compound 2 show an increase of one and two units in the mass of the [M + M 2 H]+ and [M + MX]+ ions respectively, and no change in the mass of [M + M2X 2 2H]+ ions. These results indicate that during the substitution processes by metal ions the acidic protons have been eliminated from the nucleobase.From these observations we conclude that the bi- and tri-valent metal ions preferably bind covalently to nucleobases by displacing acidic protons when the system possesses more than one type of chelating site. CrCl3–Thioglycerol The interaction of CrCl3 with the dinucleotide analogues 1–6 produces ions by the complexation of both CrIII and CrII. The latter is due to partial reduction of chromium(III). The [M + Cr 2 H]+ and [M + CrCl]+ ions have chromium in a +2 state, while [M + CrCl 2 H]+ and [M + CrCl2]+ ions are nominally CrIII.The relative intensities of the metal-complexed ions obtained are listed in Table 2. Ions due to the complexation of analyte with metal-containing ions derived from matrix species are also observed for CrIII. Bimolecular metal-chelated ions are also formed, but are not as abundant as is found with alkali-metal ions.3 Metal-ion complexation of the dinucleotide analogues 1–6 has also been studied with CrCl2, MnCl2, FeCl2, CoCl2, NiCl2, Cu(NO3)2, Zn(NO3)2, SnCl2 and Pb(NO3)2 in a thioglycerol matrix.Transition-metal ions and potentially toxic metal ions (SnII and PbII) also produce significant metal-chelated species such as [M + M 2 H]+, [M + MX]+ and [M + M2X 2 2H]+, just as we observed for FeIII and CrIII. Copper(II) gives only [M + M]+ ions in addition to the fragment ions from the ligand. The relative abundances of these metal-chelated ions are given in Table 3.The site of co-ordination of the metal is analogous to that of chromium and iron which displace amide protons. The spectral data collected with equimolar solutions of com- Fig. 1 Relative intensities of the [M + M 2 H]+ (a) and [M + MCl]+ (b) ions vs. ligandJ. Chem. Soc., Dalton Trans., 1997, Pages 341–346 343 Table 3 Relative abundances of metal-containing ions from compounds 1–6 (a) Bivalent transition-metal ions m/z (%) MX2 Compound [M + M 2 H]+ [M + MX]+ [M + M2X 2 2H]+ Others CrCl2 MnCl2 FeCl2 CoCl2 NiCl2 Cu(NO3)2 Zn(NO3)2 12345612345612345612345612345612341234 15 14 16 20 23 28 13 16 23 41 28 88 9 18 27 68 17 60 13 10 21 46 16 45 10 15 20 33 25 25 26 16 12 13 26 29 21 25 22 863 25 — 100 36 20 8 64 — 91 32 11 7 41 — 72 19 74 25 — 38 19 11 6 26 ————— 34 11 —— ——56 —— 26 24 16 21 —— 24 14 13 32 —— 25 8 10 21 —————8 —————————— 524 (9), 574 (8), 783 (3) 745 (3), 871 (2) 662 (6) 706 (4), 757 (2) —— 581 (7), 634 (7), 839 (6), 928 (2) 748 (7), 837 (6) 669 (3), 722 (4) 713 (5), 766 (4), 891 (2) —— 547 (3), 583 (7), 637 (5), 841 (7) 627 (5), 749 (7), 839 (4) 671 (3), 725 (3) 715 (5), 769 (6) — 687 (18) 589 (11), 646 (5), 847 (6) 633 (4), 752 (4) 677 (2), 734 (2) 721 (5), 778 (5) — 694 (15) 587 (9), 789 (3) 751 (4) — 719 (3) 433 (20) — 499 (9), 561 (5) ———537 (26) 757 (10) —— (b) SnCl2 and Pb(NO3)2—thioglycerol m/z (%) Others from Compound [M + Sn 2 H]+ [M + SnCl]+ [M + Pb 2 H]+ [M + PbNO3]+ SnII PbII 123456 99 55 50 77 69 37 15 13 20 —— 65 45 38 49 20 57 11 11 74 —— 851 (6) 813 (12) ———— 840 (22) 884 (19), 910 (9), 930 (8) 930 (13) — 788 (6) 758 (12) pounds 1–4 and these transition-metal ions (with an excess of MX2) show the relative affinity of the metal ions towards the ligands as a function of the number of ethylene oxide units.The relative abundances of [M + M 2 H]+ ion [Fig. 1(a)] and [M + MX]+ ion [Fig. 1(b)] from each metal ion can be used as a probe of the affinity of metal ions towards the nitrogen base itself as the number of polyethylene oxide units between them is increased.This is an alternative to the affinity towards a polyether chain and the effect of hydrogen bonding between two nucleobases on the elimination of HX. Fragmentation of metal-complexed ions. The P+ = [M + M 2 H]+ ions, on collisionally activated decomposition (CAD) (M = Fe or Cr), in which the metal ion is covalently bonded to the nucleobase, undergo fragmentation to produce ions corresponding to [P 2 CH3]+, [P 2 43]+ and [P 2 152]+.Other metal-containing ions are observed due to consecutive losses of C2H4O units from [P 2 152]+ ions. These fragmentations suggest that the metal is in multiple co-ordination with the polyether, and on the elimination of neutral (Thy 2 H) + (C2H4) (152 mass units), the metal ion migrates to the polyether chain. Further fragmentation involves losses of C2H4O units. The difference in the m/z values of metalcontaining fragment ions obtained from [M + Cr 2 H]+ and [M + Fe 2 H]+ ions corresponds to the difference in the atomic weights of the metals.The dissociation of the [M + MCl]+ ion gives an abundant ion due to dehydrohalogenation. Further fragmentation of this ion is similar to the fragmentation of the [M + M 2 H]+ ion. The labile group of the metal complex is replaced in a nucleophilic substitution by an acidic proton of the nucleobase. The CAD of the [M + Fe2Cl 2 2H]+ ion also shows the formation344 J.Chem. Soc., Dalton Trans., 1997, Pages 341–346 Table 4 Relative intensities of ions obtained from compounds 1–4 using AlCl3–glycerol–CF3CO2H–water * m/z (%) Compound [M + H]+ [M + 117]+ [M + 139]+ [M + 141]+ [M + Al 2 2H]+ Others 1 23456 35 21 21 12 100 100 51 27 21 13 20 13 8 554 13 957 5 426 409 (5), 411 (9), 597 (12), 599 (7), 621 (6), 713 (5), 715 (3) 437 (9), 641 (4), 643 (3), 719 (6) —— 616 (12) 879 (10) * In compounds 1–4 m/z 153, [T?C2H4]+, is the base peak.of the [M + Fe 2 H]+ ion followed by loss of 152 mass units. This fragmentation process indicates that the metal-ion affinity is greater towards the nucleobase than towards the ether. The dissociation of complex ions between transition-metal ions and ribonucleoside monophosphates also gave [nucleobase + M2+ 2 H]+ ions in which the base was deprotonated and bound to the metal ion directly.10 Dissociation of the P+ = [M + M 2 H]+ ion of PbII and SnII from compounds 1 and 4 gives [P 2 NHCO]+, [P 2 152]+ followed by loss of C2H4O units.The same adduct ions, derived from CoII, MnII and FeII, produce [P 2 152]+ and [P 2 2(152)]+ ions in high abundance. However, the [M + Zn 2 H]+ ion of 4 produces the [P 2 152]+ ion as the major process. Fragmentation of [M + Cu]+ yields abundant [P 2 NHCO]+ and [P 2 152]+ ions. The dissociation of [M + MX]+ and [M + M2X 2 2H]+ is similar to that of the corresponding ions derived from FeIII and CrIII.The site of metal complexation is preferentially with the nucleobase and is analogous to the chelation of FeIII and CrIII. AlCl3–Glycerol–CF3CO2H–water As reported in our earlier studies 11,12 a metal–matrix mixture of AlCl3–glycerol (glyc)–CF3CO2H–water generates ions at m/z 117 ([Al + glyc 2 2H]+), 209 ([Al + 2 glyc 2 2H]+), 231 ([Al + CF3- CO2H + glyc 2 2H]+) and 233 ([2Al + 2 glyc 2 5H]+) along with species produced by substitution of the hydroxy proton of these ions by 116 mass units i.e.[Al + glyc 2 3H] and 114 mass units (CF3CO2H) in addition to the ions from glycerol matrix.The mass spectra of compounds 1–6 recorded using AlCl3, glycerol, CF3CO2H and water show ions due to [M + H]+, [M + 117]+, [M + 139]+, [M + 141]+, [M + Al 2 2H]+ and [M + 45]+, i.e. [(M 2 H) + AlF]+. Relative abundances of these ions are listed in Table 4. The [M + 117]+ ion shows an increase of three mass units during D2O-exchange experiments.An increase of two units is due to the nucleobase and one unit from the glycerol hydroxy group. For compound 2 the [M + 141]+ ion is observed even in experiments carried out with acetic acid in place of CF3CO2H. We thus confirm that the CF3CO2H is not involved in the formation of the [M + 141]+ ion. Therefore, the [M + 141]+ ion must be formed from aluminium and glycerol. The mass of the [M + 141]+ ion remains the same during the D2O exchange. This indicates that two exchangeable protons are involved in the substitution of AlIII at pyrimidine nitrogen (NH) to form the [M + 141]+ ion, i.e.[(M 2 2H) + (2Al + glyc 2 3H)]+. This bimetallic complex ion [M + 2Al + glyc 2 5H]+ is formed by the bonding of AlIII to two nitrogens (N3) of the nucleobases and one hydroxy of a glycerol unit. The other aluminium ion is also attached to the glycerol along with the remaining two hydroxy groups as proposed for structure a. The ab initio energy-minimized structure is in agreement with the proposed structure a.Similar substitution reactions were observed with alkali-metal and alkaline-earth-metal ions for these dinucleotide analogues.3,5 Reactions of AlIII–glycerol give ions due to metal-insertion reactions by substituting exchangeable protons with a-amino acids, peptides and glycols also.11,12 The highresolution mass measurements made on the [M + 141]+ ion at m/z 507 from 1 fit with the proposed elemental composition C19H25Al2N4O9 (calc. 507.12521, obs. 507.12310; D 4.1 ppm). The other adduct ion, corresponding to [M + 139]+, shows an increase of one mass unit during D2O experiments. This ion is not observed when CF3CO2H is changed to acetic acid. Hence, this ion must be formed by the substitution of AlIII with one amide proton of one nucleobase and CF3CO2H, [(M 2 H) + Al + (CF3CO2H 2 H)]+. The exact mass measurement on the [M + 139]+ ion at m/z 505 from compound 1 gives the elemental composition C18H21AlF3N4O8 with an error of 4.9 ppm (calc. 505.11268, obs. 505.11019). The characteristic loss of CF2CO2 (94 mass units) from the [M + 139]+ ion can yield [M + 45]+ ions in which AlF+ is bonded to a N3 of the thymine moiety. Finally, the formation of the [M + Al 2 2H]+ ion from [M + 117]+ can be accounted for by the elimination of glycerol in a substitution process. Compounds 5 and 6 also yield [M + 117]+ ions. Owing to their higher proton affinity, they did not form cluster ions as did 1–4 with aluminium ions.Instead the formation of [M + H]+ ions predominates in the spectrum. Loss of adenine (135 mass units) from the protonated bimolecule is consistent with previous reports.3,5 The accurate mass measurements on the [M + 117]+ ion at m/z 483 from 1 gives a deviation D 2.6 ppm from the elemental composition C19H28AlN4O9 (calc. 483.16716, obs. 483.16843). Mass spectra recorded with equimolar concentrations of compounds 1–4 yield the gradual decrease in the abundances of [M + 117]+ ions with an increase in the spacer between the two nucleobases.The competitive metal chelation of AlIII with 18- crown-6 (1,4,7,10,13,16-hexaoxacyclooctadecane) and 4 with six ethylene glycol units produces the [M + 117]+ ions in 5:1 ratio respectively. Especially in 4, the hydrogen-bonding interactions are not strong enough to keep the system in a cyclic form to make it more effective in metal chelation. Fragmentation of ionic complexes. The CAD of the [M + 117]+ ions of compounds 1–3 leads to similar fragment ions.Hence, only the fragmentation of ions obtained from 3 is discussed. The dissociation of [M + 117]+ ions gives ions cor-J. Chem. Soc., Dalton Trans., 1997, Pages 341–346 345 Scheme 1 The collisionally activated dissociation of the [M + 117]+ ion from compound 3 responding to [M + Al 2 2H]+, followed by the elimination of water and 152 mass units. Then, an abundant ion due to consecutive losses of C2H4O units, which is the characteristic elimination of neutral C2H4O from crown ethers 13 is observed from [M 2 244]+ ions.In addition, they form ions at m/z 153 and 110. The fragmentation pathway for this ion is established by the dissociation of the [M 2 D2 + 118]+ ion at m/z 574. All deuteriums are lost in the elimination of glycerol to give [M + Al 2 2H]+. The only ion at m/z 154 contains one D due to the thymines9 exchangeable protons. The remaining fragmentation of this labelled ion is similar to that of the unlabelled ion.These fragmentation pathways, by substituting the groups attached to the metal ion (Scheme 1), suggest covalent bonding of the metal ion through the nucleobase. The dissociation of the [M + 141]+ ion from compound 3 on CAD produces more abundant ions at m/z 500 and 479. The ion at m/z 479 is due to the formation of [M + Al 2 2H]+. Loss of 152 mass units from the ion at m/z 479 is also observed. The dissociation of [M + Al + CF3CO2H22H]+ from 3 at m/z 593 gives ions corresponding to loss of CF2CO2 and CF3CO2H followed by elimination of 152 mass units to form other fragment ions. Conclusion Transition-metal ions and potentially toxic metal ions (PbII and SnII) complex with nucleobase-derived polyether glymes.Fragmentation of these adduct ions showed that there is a greater affinity of the metals towards the nucleobase than towards the polyether chain. Hydrogen-bond interactions play a major role in the complexation of MX+ ions with these nucleotide-modified systems, which may be helpful to probe Watson–Crick interactions.Chelation of bi- or tri-valent metal ions with tautomeric forms of carbonyl oxygens may allow the metal ion to interact both with the polyether chain and/or only with the nucleobase. The formation of [M + 117]+ ions also showed that cyclic systems chelate with the metal ions more strongly, as observed in the competitive complexation reactions of 18-crown-6 and compound 4.The formation of abundant adduct ions with aluminium(III) decreases as the length of the spacer increases, which is another indication of the existence of intramolecular recognition through hydrogen bonding. The fragmentation of the [M + 117]+ ion produced a [M + Al 2 2H]+ ion, which involved the substitution of exchangeable protons. This is consistent with the dissociation of [M + 117]+ ions derived from a-amino acids, peptides and glycols.11,12 Our study demonstrates that metal chelation with modified dinucleotide analogues may be useful both to probe biologically relevant interactions and to mimic the mechanism by which far larger systems operate. Experimental Spectra were measured on a Kratos (Manchester, UK) Concept 1S EB geometry instrument equipped with a FAB gun346 J.Chem. Soc., Dalton Trans., 1997, Pages 341–346 (Ion Tech, Teddington, UK). Samples were prepared by dissolving the analyte and metal chlorides in a thioglycerol matrix for recording the spectra with transition-metal ions, SnII and PbII.For aluminium(III), compounds 1–6 were mixed in a matrix mixture containing AlCl3 in glycerol, CF3CO2H and water. For each sample 1 µl of prepared solution was loaded on to the 1 mm wide stainless-steel FAB probe tip and bombarded with fast xenon atoms generated at 6–8 kV using an ion current of 0.5–1 mA. The source housing pressure was 1–2 × 1025 Torr. The secondary-ion beam was accelerated to 8 kV.The system was calibrated with 2,4,6-tris(perfluoroheptyl)-1,3,5- triazine. The spectra recorded, 6–10 scans for each run, were averaged and the background was subtracted from spectra using a Kratos MACH 3/DART data system running on a SUN SPARC Station 10. Helium was used to reduce the main beam intensity by 45–50% in the first field-free region of the (EB geometry) mass spectrometer to obtain the CAD spectra. High-resolution mass measurements were made at 10 000 resolving power and 8 kV accelerating voltage.Spectra were obtained with equimolar concentrations of 1–4 (0.5 mg) using AlCl3 (2–3 mg), glycerol (25-30 mg), CF3CO2H (2 µl) and water (2 µl). Energy minimization of ion a was carried out using the Hartree-Fock method at the STO-3G level using the SPARTAN ab initio program running on a Silicon Graphics Power Indigo 2 Extreme computer. The initial geometry of the molecule was estimated using the Sybyl/x force field. The preparation of compounds 1–6 was described in our previous work.3 Matrices (thioglycerol and glycerol), metal chlorides and nitrates, CF3CO2H and D2O were obtained from Aldrich Chemical Co.(Milwaukee, WI) and used without further purification. Acknowledgements The authors are grateful to the Natural Sciences and Engineering Research Council (NSERC) of Canada for financial support in the form of operating and equipment grants (to J. M. M.) and to T. R. B. Jones for technical assistance. References 1 J. E. Kickham, S. J. Loeb and S. L. Murphy, J. Am. Chem. Soc., 1993, 115, 7031. 2 O. F. Schall and G. W. Gokel, J. Am. Chem. Soc., 1994, 116, 6089. 3 M. Saraswathi and J. M. Miller, J. Am. Soc. Mass Spectrom., 1996, 7, 42. 4 M. Moet-Ner (Mautner), J. Am. Chem. Soc., 1983, 105, 4912; R. B. Sharma and P. Kebarle, J. Am. Chem. Soc., 1984, 106, 3913; D. J. Cram and J. M. Cram, Acc. Chem. Res., 1978, 11, 8; D. J. Cram, Science, 1988, 240, 760. 5 M. Saraswathi and J. M. Miller, J. Mass Spectrom., 1996, 31, 1011. 6 N. Grover, N. Gupta and H. H. Thorp, J. Am. Chem. Soc., 1992, 114, 3390; N. Grover, T. W. Welch, T. A. Fairley, M. Cory and H. H. Thorp, Inorg. Chem., 1994, 33, 3544; J. K. Barton and E. Lolis, J. Am. Chem. Soc., 1985, 107, 708. 7 H. Sigel and R. B. Martin, Chem. Rev., 1982, 82, 385. 8 M. L. Gross, Acc. Chem. Res., 1994, 27, 361. 9 M. J. Moet-Ner, J. Am. Chem. Soc., 1979, 101, 2396. 10 J. H. Hamilton and J. Adams, 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Georgia, May 1995, P609. 11 M. Saraswathi and J. M. Miller, Can J. Chem., in the press. 12 M. Saraswathi and J. M. Miller, Rapid Commun. Mass Spectrom., 1996, 10, 1706. 13 Y. C. Lee, A. I. Popov and J. Allison, Int. J. Mass Spectrom. Ion Phys., 1983, 51, 267. 14 SPARTAN SGI version 4.0.3 GL, Wavefunction Inc., Levine, CA, 1995. Received 31st July 1996; Paper 6/05354H

ISSN:1477-9226    

DOI:10.1039/a605354h

出版商:RSC    

年代:1997

数据来源: RSC