摘要:

L e t t e r Synthesis and crystal structure of [Au2(N-Ts[9]aneNS2)Cl2 ] 2 {N-Ts [9]aneNS2= 7-(toluenesulfonyl)-7-aza-1,4-dithiacyclononane} incorporating AuÆ Æ ÆAu and pñp interactions Angelo J. Amoroso,a Alexander J. Blake,a Jonathan P. Danks,a Dieter Fenskeb and Martin Schroé der*a a School of Chemistry, University of Nottingham, University Park, Nottingham, UK NG7 2RD. E-mail : m.schroder=nottingham.ac.uk b Institut Anorganische Chemie, Karlsruhe, Engesserstr. Geb-Nr. 30.45, fué r Universitaé t 76128 Karlsruhe, Germany Received (in Cambridge) 31st October 1998, Revised m/s received 3rd February 1999, Accepted 23rd February 1999 Reaction of AuCl(tht) (tht = tetrahydrothiophene) with NTs[ 9]aneNS2 {N-Ts[9]aneNS2 = 7-(toluenesulfonyl)-7-aza- 1,4-dithiacyclononane} aÜords the tetranuclear Au(I) species which exhibits AuÆ Æ ÆAu and pñp [Au2(N-Ts [9]aneNS2)Cl2 ] 2 interactions leading to an overall in–nite chain structure.The co-ordination of Au(I) by thioether crowns is well documented. 1h3 The Au(I) centre usually adopts a distorted tetrahedral geometry as in and [Au([9]aneS3)2]`1 whereas linear co-ordination is observed [Au([18]aneS6)]`,2 in where only one thioether donor is bound Au([9]aneS3)Cl3 to Au(I).Ligation of amines to Au(I) fragments is known,4 but is much less common than ligation with sulfur due to Au(I) being a typical class B/soft metal ion.5 Although coordination of to Au(I) leads to an unstable [18]aneN2S4 product which decomposes to form elemental gold, its methylated analogue, gives a stable species Me2[18]aneN2S4 , in which only the thioether donors [Au(Me2[18]aneN2S4)]` are co-ordinated to the metal.2 Little has been reported6 on the interaction of Au centres with mixed N/S donor macrocycles such as and and we therefore [9]aneN2S [9]aneNS2 , undertook a study of the interaction of Au(I) with [9]aneNS2 .Initial reactions of AuCl(tht) (tht\tetrahydrothiophene) with in 2 : 1 and 1 : 1 stoichiometries resulted in [9]aneNS2 the formation of insoluble products, probably polymeric species, which could not be characterised fully.We have found previously that thioether crowns can form extended polymers with d10 metal ions.7 In light of this, the macrocyclic precursor, N-Ts[9] [7-(toluenesulfonyl)-7-aza-1,4-dithia aneNS2 cyclononane], was used in the hope that the tosyl group would make the resulting product more soluble.Also, the protection aÜorded to the amino group might be expected to inhibit reduction of Au(I) to the metal. Reaction of AuCl(tht) with N-Ts[9] in a 2 : 1 molar aneNS2 ratio in resulted in a colourless solution. Addition of CH2Cl2 yielded a white solid which was collected by –ltration Et2O and dried. Microanalytical and spectroscopic data for the complex are consistent with the stoichiometry Au2(NCrystals of the dichloromethane hemi- Ts[9]aneNS2)Cl2 .§ solvate suitable for X-ray diÜractionî were obtained by the slow diÜusion of vapour into a solution of the complex Et2O in CH2Cl2 .The structure contains two units Au2(N-Ts[9]aneNS2)Cl2 linked across a crystallographic inversion centre by two equivalent Au… … …Au interactions [Au… … …Au 3.1155(5) (Fig. 1). Aé ] The macrocyclic ring has a strained [225] conformation8 similar to that observed by Parker and co-workers in While the Au… … …Au distance is longer than Au([9]aneS3)Cl.3 that found in metallic gold (2.884 it is shorter than twice Aé ),9 the van der Waals radius for gold (3.32 and comparable Aé )10 to those of known Au… … …Au interactions, some of which have been calculated as having bond strengths lying between those of the strongest hydrogen-bonds and those of the weakest covalent bonds (42»46 kJ mol~1).11 The FAB mass spectrum of the complex shows a fragmentation peak corresponding to at m/z 1064.Each Au(I) centre is bound to a Cl~ [Au2L2Cl]` anion and one thioether donor in an almost linear fashion [S(1)»Au(1)»Cl(1) 171.27(5), S(4i)»Au(2)»Cl(2) 178.86(5)°], the Au… … …Au interactions being orthogonal to these bonds [Cl(1)» Au(1)»Au(2) 89.47(4), Cl(2)»Au(2)»Au(1) 96.13(4)°] (Fig. 1). Fig. 1 Two views (a) and (b) of the tetranuclear fragment of [Au2(NSelected bond lengths and angles (°) : Ts[9]aneNS2)Cl2]2. (Aé ) Au(1)… … …Au(2) 3.1155(5), Au(1)»S(1) 2.2581(14), Au(2)»S(4i) 2.2711(13), Au(1)»Cl(1) 2.2664(15), Au(2)»Cl(2) 2.2770(14), S(1)»Au(1)»Cl(1) 171.27(5), S(1)»Au(1)»Au(2) 98.73(3), Cl(1)»Au(1)»Au(2) 89.47(4), S(4i)» Au(2)»Cl(2) 178.86(5), S(4i)»Au(2)»Au(1) 84.72(3), Cl(2)»Au(2)»Au(1) 96.13(4) (i\[x]2, [y, [z).New J. Chem., 1999, 23, 345»346 345Fig. 2 Crystal packing of The p»p interactions (dashed lines) give a centroid»centroid distance of 4.14 and an MAu2(N-Ts[9]aneNS2)Cl2N= .Aé interplanar distance of 3.48 Aé . Additionally, the crystal structure determination also shows the presence of p»p interactions in the solid state (Fig. 2), with the aromatic rings of neighbouring dimeric units lying parallel to each other and separated by a perpendicular distance of 3.48 The ring centroids lie 4.14 apart and are oÜset by Aé .Aé 2.24 relative to one another, as is typically observed in aro- Aé matic p»p stacking.12 While these interactions may not persist in solution, they obviously have a signi–cant role to play in the overall geometry of the complex in the solid state. Such interactions, together with hydrogen-bonding, has allowed Munakata and co-workers to form13 in–nite linear chains of Cu(I) quinoline-2-thione complexes.The same author has also observed p»p stacking playing a considerable role in the structure of many polymeric complexes.14 Additionally, p»p stacking, together with Ag… … …Ag interactions, has played a vital role in the construction of a polymeric Ag(I) complex of 1,2-trans- (4-pyridyl)ethene.15 In conclusion, the complex [Au2(N-Ts[9]aneNS2)Cl2]2 shows interesting p»p and Au… … …Au interactions which combine to give an overall polymeric motif.With a great variety of tosylated mixed N/S macrocycles available and the replacement of the tosyl group for larger aromatic systems possible, an interesting and varied chemistry is anticipated from this system. Acknowledgements thank the EPSRC for –nancial support, the EPSRC We National Service at the University of Swansea (for mass spectrometry), and Dr N.R. Champness for helpful discussions. Notes and references § Experimental procedure: a solution of 7-(toluenesulfonyl)-7-aza-1,4- dithiacyclononane (0.016 g, 0.05 mmol) in the minimum amount of (1 cm3) was added to a stirred solution of AuCl(tht) (0.032 g, CH2Cl2 0.10 mmol) in the minimum amount of (1 cm3).After stirring CH2Cl2 for 30 min, (10 cm3) was added to the clear solution to aÜord a Et2O white precipitate which was collected by –ltration. Yield 0.034 g, 86% (Calc. for C, 19.96 ; H, 2.45 ; N, 1.79. Found: C13H19Au2Cl2NO2S3 : C, 19.77 ; H, 2.46 ; N, 1.53%). FAB mass spectrum (nitrobenzyl alcohol matrix) : m/z 1064 10%), 977 18%), (Au2L2Cl`, (Au3LCl2 `, 831 42%), 746 27%), 514 (AuL`, 100%). 1H (AuL2 `, (Au2LCl`, NMR 300 MHz, 293 K): d 7.60 (d, J\8.3, 2H, Ar), 7.37 (d, (CDCl3 , J\8.4 Hz, 2H, Ar), 3.51 (br s, 4H), 3.45, (br s, 8H), 2.46 (s, 3H, ArCH3). î Crystal data for 1: M\824.77, C13H19Au2Cl2NO2S3 Æ 0.5CH2Cl2 , triclinic, space group a\7.977(1), b\11.053(1), c\12.455(1) P1, Aé , a\107.37(1), b\95.23(1) c\90.33(1)°, U\1043.08(18) Z\2, Aé 3, k\14.741 mm~1, T \203(2) K, 9778 re—ections collected, 5240 independent re—ections [I[2p(I)]\0.0361, [Rint\0.048], R1 wR2(all data)\0.0987.All signi–cant diÜerence electron density features lie within 1.07 of Au(1) or Au(2). The solvent molecule is dis- Aé CH2Cl2 ordered about a crystallographic inversion centre. CCDC reference number 440/104. See http ://www.rsc.org/suppdata/nj/1999/345/ for crystallographic –les in …cif format. 1 A. J. Blake, R. O. Gould, J. A. Greig, A. J. Holder, T. I. Hyde and M. Schroé der, J. Chem. Soc., Chem. Commun., 1989, 876; A. J. Blake, J. A. Greig, A. J. Holder, T. I. Hyde, A. Taylor and M. Schroé der, Angew. Chem., Int. Ed. Engl., 1990, 29, 197. 2 A. Taylor, PhD Thesis, University of Edinburgh, 1991; A. J. Blake, A. Taylor and M. Schroé der, J. Chem.Soc., Chem. Commun., 1993, 1097; A. J. Blake, D. Collison, R. O. Gould, A. J. Holder, T. I. Hyde, G. Reid, A. Taylor and M. Schroé der, Molecular Electrochemistry of Inorganic, Bioinorganic and Organometallic Compounds, eds. A. J. L. Pombeiro and J. A. McCleverty, Kluwer Academic Publishers, 1993, pp. 121»129; A. J. Blake, R. O. Gould, C. Radek, G. Reid, A. Taylor and M. Schroé der, Proceedings of the 1st International Conference on the Chemistry of the Copper and Zinc T riads, Edinburgh 1992, Royal Society of Chemistry, Cambridge, 1993, pp. 95»102. 3 D. Parker, P. S. Roy, G. Ferguson and M. M. Hunt, Inorg. Chim. Acta, 1989, 155, 227. 4 See for example: D. M. P. Mingos, J. Yau, S. Menzer and D. J. Williams, J. Chem. Soc., Dalton T rans., 1995, 319. 5 J. J. Guy, P.G. Jones, M. J. Mays and G. M. Sheldrick, J. Chem. Soc., Dalton T rans., 1977, 8. 6 U. Heinzel and R. Mattes, Polyhedron, 1991, 10, 19; J. P. Danks, N. R. Champness and M. Schroé der, Coord. Chem. Rev., 1998, 174, 417. 7 A. J. Blake, W.-S. Li, V. Lippolis and M. Schroé der, Chem. Commun., 1997, 1943. 8 J. Dale, T etrahedron, 1974, 30, 1683. 9 D. N. Batchelder and R. O. Simons, J. Appl. Phys., 1965, 36, 2864. 10 A. Bondi, J. Phys. Chem., 1964, 68, 441. 11 P. Pyykkoé , Chem. Rev., 1997, 97, 597. 12 C. A. Hunter and J. K. M. Sanders, J. Am. Chem. Soc., 1990, 112, 5525. 13 S. Kitagawa, S. Kawata, Y. Nozaka and M. Munakata, J. Chem. Soc., Dalton T rans., 1993, 1399. 14 M. Munakata, T. Kuroda-Sowa, M. Maekawa, A. Honda and S. Kitagawa, J. Chem. Soc., Dalton T rans., 1994, 2771; T. Kuroda- Sowa, M. Munakata, H. Matsuda, S. Akiyama and M. Maekawa, J. Chem. Soc., Dalton T rans., 1995, 2201. 15 A. J. Blake, N. R. Champness, S. S. M. Chung, W.-S. Li and M. Schroé der, Chem. Commun., 1997, 1675. L etter 8/08494G 346 New J. Chem., 1999, 23, 345»346

ISSN:1144-0546    

DOI:10.1039/a808494g

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

L e t t e r Phosphate anion binding and luminescent sensing in aqueous solution by ruthenium(II) bipyridyl polyaza receptors Paul D. Beer* and James Cadman Inorganic Chemistry L aboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QR Received 28th January 1999, Accepted 25th February 1999 New ruthenium(II) bipyridyl polyaza receptors have been prepared and shown to bind and sense, via MLCT luminescent emission quenching, phosphate and ATP anions in aqueous solution.Anionic guest species play a pivotal role in biological processes and their eÜects as environmental pollutants are now being realised.1 In consequence of this there is a growing need for the construction of speci–c receptors containing signalling or reporter groups that have the capability of sensing anions in solution.2 Water in particular is a challenging medium in which to detect the anion recognition event as anion hydration energies are large.3 Rare examples of luminescent anion responsive systems have combined the anthracene —uorophore with polyammonium,4 guanidinium5 and zinc(II) amine6 anion recognition sites.In addition cyclen appended phenanthridinium europium and terbium complexes are showing promise as luminescent sensors for halide and hydroxide ions in water.7 We report here a new class of anion luminescent sensor consisting of a series of new ruthenium(II) bipyridyl polyaza receptors which bind and detect phosphate anions in water via MLCT luminescent emission quenching. The condensation of 4,4@-diformyl-2,2@-bipyridine8 with two equivalents of the appropriate mono-BOC protected diamine and N,N@-dimethylethylenediamine followed by in situ sodium tetrahydroborate reduction gave the new bipyridyl substituted amine compounds in 70»75% yield (Scheme 1).Re—uxing these compounds with in (bipy)2RuCl2 … 2H2O aqueous ethanol for 18 hours and subsequent addition of hydrochloric acid aÜorded the new receptors L1»L4 isolated as hydrochloride salts in near quantitative yield (Scheme 2).Potentiometric methods in water using the SUPERQUAD program9 (0.1 M 25 °C) enabled speciation diagrams, KNO3 , Scheme 1 protonation and receptor»phosphate complex stability constant values to be determined for L1»L4. Table 1 shows, as expected, values increase as the diamine alkyl spacer pKa increases from L1 to L3.Taking into account speciation diagrams and the respective values of the receptors and pKa phosphate and ATP, Tables 2 and 3 report the various receptor»anion complex species stability constant values. All the receptors bind both phosphate and ATP with the strength of binding dependent upon levels of protonation of receptor and guest anion.§ As expected there is a general selectivity preference for ATP over phosphate and obviously chloride and nitrate.The luminescent properties of the centre are Ru(bipy)32` well documented10 and we have recently shown that perturbation of the emission spectra of ruthenium(II) bipyridyl amide based receptors in polar DMSO and acetonitrile organic solvents occurs upon halide and dihydrogen phosphate binding.11 It was of interest therefore to investigate whether Scheme 2 Table 1 Protonation constants for L1»L4 receptors Receptor pKa1 pKa2 pKa3 pKa4 L1 10.12(2)a 8.97(2) 5.00(3) 3.53(4) L2 10.37(1) 9.68(1) 6.71(1) 5.65(2) L3 10.77(1) 10.09(1) 7.57(1) 6.65(2) L4 8.37(2) 4.35(3) 2.78(3) ca. 1.50b a Figures in parentheses are the standard deviations of the last digit. was difficult to re–ne successfully with the program SUPER- b pKa4 QUAD as it is low.New J. Chem., 1999, 23, 347»349 347Table 2 Stability constants for receptor»phosphate complexesa LH6PO4 LH5PO4 LH4PO4 LH3PO4 LH2PO4 Receptor as LH44` …H2PO4~ as LH33`…H2PO4~ as LH22` …H2PO4~ as LH22` …HPO42~ as LH` …HPO42~ L1 » 4.97(1) 3.78(2) 3.62(1) 3.09(1) L2 4.09(9) 3.22(7) 3.65(4) 3.39(3) 3.34(4) L3 » 2.03(9) 2.69(3)b 2.87(4) 3.41(4) L4 » 9.31(9) 7.91(6) 5.77(4)c 4.67(3) a log bœs for in at 25 °C, 0.1 M are 12.03, 7.03, 2.64.b As c As KH2PO4 H2O KNO3 LH33` …HPO42~. LH` …H2PO4~. these water soluble receptors were capable of sensing anions via MLCT emission perturbation in the aqueous environment. The luminescent behaviour of L1»L4 in the pH range 2»10 was investigated via titration of the respective tetrahydrochloride compound with base, using to lower the HNO3 pH.The emission intensity versus pH and MLCT pro–le jmax for L1 is shown in Fig. 1. As the solution becomes more acidic an increase in the emission intensity is observed which reaches a maximum at pH ca. 5.5 before a gradual quenching eÜect occurs towards lower pH values. Concomitant with this emission intensity variation there is a red shift of the MLCT jmax Fig. 1 pH dependence of luminescence intensity and wavelength for L1 (10~5 M). Colour version available at : http ://www.rsc.org/ suppdata/nj/1999/347/. from 605 nm at high pH values to 625 nm at pHO4. Similar observations were also obtained with L2 and L3 receptors. At high pH where the amine groups remain largely unprotonated the luminescence is quenched via photoinduced electron transfer (PET) from the unprotonated amine groups to the moiety.Retrieval of quenched luminescence takes Ru(bipy)32` place as the pH is lowered and the amine groups become protonated. Why luminescence quenching occurs at pH valuesO5.5 is difficult to rationalise. Grigg et al.12 reported similar quenching behaviour at low pH with 5,5@- diaminomethyl-2,2@-bipyridyl ruthenium(II) complexes.They attributed the quenching to the location of proximal protonated amine positive charges in the vicinity of the Ru(bipy)32` moiety which may lead to Ru»N bond –ssion in the excited state. A similar mechanism could be in operation here, at lower pH the values of L1»L3 (Table 1) suggest proto- pKa nation of the groups occurs generating bipyridyl»CH2»amino positive charges in close proximity to the ruthenium(II) bipyridyl centre.The luminescence pH titration curve for L4 (Fig. 2) displays similar quenching characteristics at lower pH, however, at higher pH values (pH[7) there is a gradual increase in emission intensity. The addition of and salts to non- KH2PO4 Na2H2ATP deaeratedî buÜered aqueous solutions of L1»L4 at pH 6 resulted in signi–cant quenching of luminescence (Fig. 3) by up to 15%. Analogous anion titration experiments with produced no eÜect. These phosphate anion Ru(bipy)32` induced quenching eÜects contrast with ruthenium(II) bipyridyl amide systems exhibiting an intensity enhancement of luminescence emission response to chloride and dihydrogen phosphate anions in acetonitrile and DMSO solutions.11 The Fig. 2 pH dependence of luminescence intensity and wavelength for L4 (10~5 M) in water, T \25 °C. Table 3 Stability constants for receptor»ATP complexesa LH7ATP LH6ATP LH5ATP LH4ATP LH3ATP LH2ATP Receptor as LH44` …H3ATP~ as LH44` …H2ATP2~ as LH33`…H2ATP2~ as LH33` …HATP3~ as LH22` …HATP3~ as LH22` …ATP4~ L1 » 5.02(8) 5.13(7) 5.22(5) 4.27(5) 3.74(3) L2 4.83(5) 3.79(6) 4.37(4)b 5.21(2) 4.62(2) 3.28(2) L3 4.12(4) 2.75(9) 3.41(4)b 4.99(1) 3.66(2)d 1.89(9) L4 » » 4.28(9) 3.53(9)c 4.14(9) » a log bœs for in at 25 °C, 0.1 M are 6.78, 4.01, 2.03.b As c As d As Na2H2ATP H2O KNO3 LH44` …HATP3~. LH2 2` …H2ATP2~. LH3 3` …ATP4~. 348 New J. Chem., 1999, 23, 347»349Fig. 3 Emission spectral titration of with L1 (10~5 M) in KH2PO4 water at pH\6.02 buÜered with N-morpholinoethylsulfonic acid (MES) (10~2 M), T \25 °C.Colour version available at : http :// www.rsc.org/suppdata/nj/1999/347/. formation of a receptor»anion complex that increases the rigidity of the receptor, thereby disfavouring the non-radiative decay processes can tentatively explain the latter eÜect in organic solvents. However, further photophysical investigations are needed to determine the nature of this anion» receptor quenching mechanism in water. This preliminary study has demonstrated that these new ruthenium(II) bipyridyl polyaza receptors represent a new class of anion luminescent sensor capable of binding and sensing phosphate and ATP anions in the aqueous environment. Acknowledgements thank the EPSRC for a studentship and for use of the We mass spectrometry service at University College, Swansea.Notes and references § 31P NMR titration experiments in gave evidence for 1 : 1 stoi- D2O chiometric complexes with ATP. î Analogous anion titration experiments in argon aerated aqueous solutions gave very similar quenching observations. 1 F. P. Schmidtchen and M. Berger, Chem. Rev., 1997, 97, 1609; P.D. Beer and D. K. Smith, Prog. Inorg. Chem., 1997, 46, 1; J. L. Atwood, K. T. Holman and J. W. Steed, Chem. Commun., 1996, 1401. 2 A. P. de Silva, H. Q. N. Guarantee, T. Gunnlaugsson, A. J. Huxley, C. P. McCoy, J. T. Rademacher and T. R. Rice, Chem. Rev., 1997, 97, 1515; P. D. Beer, Acc. Chem. Res., 1998, 31, 71; P. D. Beer, Chem. Commun., 1996, 689; R. V. Slone, D. I.Yeon, R. M. Calhoun and J. T. Hupp, J. Am. Chem. Soc., 1995, 117, 11813. 3 A. Bianchi, K. Bowman-James and E. Garcïç a-Espan8 a (Editors), Supramolecular Chemistry of Anions, Wiley-VCH, New York, 1997. 4 A. W. Czarnik, Acc. Chem. Res., 1994, 27, 302. 5 A. P. de Silva, H. Q. N. Guarantee, C. McVeigh, M. G. E. Maguire, P. R. S. Maxwell and E. OœHanlon, Chem. Commun., 1996, 2191. 6 L.Fabbrizzi, G. Francese, M. Licchelli, A. Perotti and A. Taglietti, Chem. Commun., 1997, 581. 7 D. Parker, P. K. Senanayake and J. A. G. Williams, J. Chem. Soc., Perkin T rans. 2, 1998, 2129. 8 P. Dupau, T. Renouard and H. Le Bozec, T etrahedron L ett., 1996, 37, 7503. 9 P. Gans, A. Sabatina and A. Vacca, J. Chem. Soc., Dalton T rans., 1985, 1195. 10 J.-P. Sauvage, J.-P. Collin, J.-C. Chambron, S. Guillerez, C. Coudret, V. Balzani, F. Barigelletti, L. De Cola and L. Flamigni, Chem. Rev., 1994, 94, 993. 11 F. Szemes, D. Hesek, Z. Chen, S. W. Dent, M. G. B. Drew, A. J. Goulden, A. R. Graydon, A. Grieve, R. J. Mortimer, T. Wear, J. S. Weightman and P. D. Beer, Inorg. Chem., 1996, 35, 5868; P. D. Beer, F. Szemes, V. Balzani, C. M. Salaç , M. G. B. Drew, S. W. Dent and M. Maestri, J. Am. Chem. Soc., 1997, 119, 11864. 12 R. Grigg and W. D. J. A. Norbert, J. Chem. Soc., Chem. Commun., 1992, 1300. L etter 9/01534E New J. Chem., 1999, 23, 347»349 349

ISSN:1144-0546    

DOI:10.1039/a901534e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

L e t t e r Synthesis and characterization of a new kind of l-oxo Mn2 III, IV complex: [Mn2O2(terpy)2(H2O)2 ](NO3)3 Æ6 H2O, terpy= 2,2º : 6º,2/-terpyridine Marie-Noeé lle Collomb,*a Alain Deronzier,*a Aureç lien Richardota and Jacques Pe ç cautb a L aboratoire dœElectrochimie Organique et de Photochimie (CNRS UMR 5630), Reç dox Joseph Fourier, BP 53, 38041 Grenoble cedex 9, France Universiteç de Recherche Fondamentale sur la Service de Chimie b Deç partement Matie` re Condenseç e, Inorganique et Biologique, L aboratoire de Chimie de Coordination (CNRS URA 1194), CEA/Grenoble, 38054 Grenoble 9, France ceç dex Received (in Montpellier, France) 29th January 1999, Accepted 15th February 1999 A new kind of mixed-valence binuclear l-oxo complex of manganese having two aqua ligands has been synthesized and structurally characterized.Although the exact nature of the oxygen-evolving center in photosystem II remains ill-characterized, it is believed to contain a tetranuclear oxo-manganese cluster.1 In order to mimic this natural active center much eÜort has been devoted during the last decade to designing molecules containing multinuclear l-oxo-bridged manganese complexes.Among the numerous molecules synthesized for this purpose,2 only the mixed-valence binuclear complex [Mn2III, IVO2(bpy)4]3` (bpy\2,2@-bipyridine) and the analogous 1,10-phenanthroline have been reported to be active for evolution from water O2 oxidation under heterogeneous conditions.3 However, this result remains questionable4 and we reported recently5 that the assumed active species is unstable [Mn2IV, IVO2(bpy)4]4` in water, leading to the stable tetranuclear species which is inactive as a homogeneous [Mn4IVO6(bpy)6]4`, O2 catalyst.Very recent work by Limburg et al.6 reported that oxidation of some mononuclear complexes with [Mn(L)2]n planar tridentate ligands L, like dpa (dpa\dipicolinate, n\[1) or terpy (terpy\2,2@:6@,2A-terpyridine, n\]2), by oxone leads to evolution ; the complex containing terpy O2 appears to be the more stable. Although the mechanism involved is unknown, these authors observed the formation of a green intermediate dimer having a MnIIIMnIV feature, as characterized by UV-visible and EPR spectroscopies.Furthermore, they suggested that the MnIIIMnIV complex has a bis l-oxo structure, the open coordination sites being occupied by the solvent The better activity of these kinds of (H2O).complex versus bpy derivatives could result from the possible formation of higher oxidation states containing MnxO species able to oxidize water by an inner-sphere mechanism. In this context, we report here the synthesis and the structural characterization of this new kind of mixed-valence complex 1, containing terpy as [Mn2O2(terpy)2(H2O)2](NO3)3 … 6H2O, a tridentate ligand, and a terminally bound water molecule on each Mn.Reaction of 1.44 equiv. of terpyridine ligand with 1 equiv. of followed by subsequent addition of 0.42 equiv. of Mn(NO3)2 , aÜords a green solution that yields a green precipi- KMnO4 , tate of 1 upon cooling.§ Elemental analysis of 1 shows that it is the mixed-valence complex with the above formulation, which is con–rmed by X-ray structure analysis.î The unit cell contains two distinct but very similar dimeric cations, six nitrate anions and twelve water molecules.The ORTEP diagram of the two cations, is [Mn2III, IVO2(terpy)2(H2O)2]3`, represented in Fig. 1. Each cation possesses a crystallographically imposed center of symmetry implying that the two Mn ions (MnIII and MnIV) for each dimer are crystallographically equivalent.To our knowledge, this unusual equivalence of the two Mn centers in a MnIIIMnIV(l-O) dimeric 2 structure has only been reported by Stebler et al.7 in the case of the complex.8 [Mn2III, IVO2(phen)4](PF6)3 …CH3CN Two possibilities, not distinguishable, have been suggested in order to explain the equivalence of the two Mn centers :8,9 (1) the observed structure corresponds to a superposition of MnIIIMnIV and MnIVMnIII complexes (class I in the classi- –cation of Robin and Day,9 static disorder) and (2) the rate of the electron transfer MnIIIMnIV]MnIVMnIII, is faster than the measurement time scale, thus the two Mn ions appear to be identical (class II,9 dynamic disorder).A third possiblity, in which the two manganese ions are chemically equivalent since the d electrons are totally delocalized over the two Mn centers (class III,9 ordered structure), has been excluded after analysis of the thermal displacement parameters.7 In our case, a 16-line EPR spectrum is observed for 1 (see further in the text), indicating that this dimer is probably a d electron localized species ; thus only static disorder could explain the equivalence of the Mn ions.Bond angles suggest (Fig. 1) that the geometry at each manganese center is approximately octahedral, the coordination spheres consisting of the three N atoms from a single terpy ligand, the two cis-bridging oxo ligands and the O atom of the coordinated water (noted The coordination in the Oaqua).equatorial plane is provided by the three O atoms and an N pyridine atom while the two axial positions are coordinated by the two other pyridine N atoms. The orientation of the two water molecules is trans. A trans geometry is also observed in the few manganese complexes, as yet isolated, containing a bond such as (L\bpy10 MnwOaqua [Mn3IVO4(L)4(H2O)2]4` and phen11), and [Mn2IIIO(O2CCH3)2(bpy)2(H2O)2]2`,12 complexes.[Mn2IVO2(O2CCH3)(bpy)2(H2O)2]3` 13 In the following discussion, the average of the atom bond distances of the two dimeric cations present in the structure of 1 are used. The MnwMn bond distance of 2.7321(6) in 1 is Aé at the high end of the range (2.588»2.741 observed for Aé ) other structurally characterized MnIIIMnIV di-oxo-bridged complexes14 and slightly longer than the values of 2.716 and 2.700 in the and Aé [Mn2III, IVO2(bpy)4]3` 15 [Mn2III, IVO2- complexes.The MnwO distances of 1.820(2) and (phen)4]3` 7 1.809(4) for the oxo bridges are consistent with the values of Aé 1.820 and 1.808 observed by Stebler et al.7 in the phen Aé New J. Chem., 1999, 23, 351»353 351Fig. 1 X-ray structure of (ORTEP [Mn2III, IVO2(terpy)2(H2O)2]3` representation of the two similar dimeric cations present in an unit cell). Selected bond distances and angles (°) : (Aé ) Mn(1)wMn(1)j1 2.7315(12), Mn(1)wO(1) 1.818(3), 1.812(3), Mn(1)w Mn(1)wO(1)j1 O(2) 2.009(3), Mn(1)wN(1) 2.132(4), Mn(l)wN(2) 2.044(3), Mn(l)wN(3) 2.131(3), 97.61(12), Mn(1)wO(1)wMn(1)j1 O(1)wMn(1)wO(1)j1 82.39(12), N(1)wMn(1)wO(1) 94.09(12), O(2)wMn(1)wO(1) 175.40(11), N(2)wMn(1)wO(1) 90.24(12), N(3)wMn(1)wO(1) 96.75(12).Mn(2)w 2.7327(12), Mn(2)wO(11) 1.821(2), 1.805(3), Mn(2)j2 Mn(2)wO(11)j2 Mn(2)wO(12) 2.012(3), Mn(2)wN(11) 2.125(4), Mn(2)wN(12) 2.037(3), Mn(2)wN(13) 2.125(3), 97.83(12), O(11)w Mn(2)wO(11)wMn(2)j2 82.17(12), N(11)wMn(2)wO(11) 95.22(12), O(12)w Mn(2)wO(11)j2 Mn(2)wO(11) 176.54(11), N(12)wMn(2)wO(11) 90.37(12), N(13)w Mn(2)wO(11) 95.64 (12).complex and are in good agreement with those of other similar MnIIIMnIV(l-O) complexes cited in the literature.14 2 Moreover, the MnwO distance is comparable to the RuwO distance in the homogeneous water oxidation catalyst (1.869 [Ru2IIIO(bpy)4(H2O)2]3` Aé ).16 The bond length [2.011(2) distance compares MnwOaqua Aé ] well with those observed in the trinuclear bpy10 and phen11 complexes (average 2.045 and 2.009 respectively) and in Aé , the complex13 (average [Mn2IVO2(O2CCH3)(bpy)2(H2O)2]3` 1.991 On the other hand, it is shorter than the Aé ).RuIIIwOaqua bond in (average 2.136 and the [Ru2IIIO(bpy)4(H2O)2]3` Aé )16 bond in MnIIIwOaqua [Mn2IIIO(O2CCH3)2(bpy)2(H2O)2]2` (average 2.312 The distance in 1 is also sig- Aé ).12 MnwOaqua ni–cantly longer than the MnIVwOH bond distance of 1.881 reported by Wieghardt et al.,17 con–rming that the bound Aé species is an aqua group.The IR spectrum of 1 exhibits in the 670»770 cm~1 region only an intense band at 706 cm~1 assigned to a stretching mode of the di-l-oxo bridge,18 while vibrations at 1384(vs) and 830 cm~1 con–rm the presence of as counter ions.NO3~ Several spectroscopic measurements carried out in stabilized° aqueous solutions con–rm the mixed-valence nature of complex 1. In contrast to the bpy and phen di-l-oxo complexes, 4,5,18 1 is insoluble in usual organic solvents like CH3CN and but soluble and stable for several hours in CH2Cl2 unbuÜered pure water solution at pH 4.5. Obviously, aqueous solutions of 1 are also stable in the presence of terpyridyl buÜer at pH 4 (for pH[4, the terpy ligand precipitates in solution). The addition of electrolyte like 0.1 M or NaNO3 does not aÜect this stability.The UV-visible absorp- Na2SO4 tion spectrum of a green solution of 1 measured in distilled water at pH 4.5 is reminiscent of that of in [Mn2O2(bpy)4]3` a bpy/bpyH` buÜer (Fig. 2).5 The band at 553 nm (e\678 l mol~1 cm~1) is possibly ascribable to a d»d transition band, while that at 654 nm (e\585 l mol~1 cm~1) is due to a ligand (O)-to-metal charge transfer.18 As expected, EPR spectroscopic measurements of a solution of 1 in (pH 4.5) containing 20% at 100 K give a H2O CH3 CN 16-line signal, centered at g\2.18 1 can be also characterized by electrospray ionization mass spectroscopy (ES).As previously observed for some dinuclear oxo-iron complexes containing aqua ligands,19 these ligands are too labile to remain coordinated to metal atoms under ES conditions. The most abundant peak at m/z 732 in the spectrum is due to the cation. [Mn2O2(terpy)2](NO3)2 ` The cyclic voltammogram (CV) at a vitreous carbon electrode of 1 in water containing 0.1 M (pH 4.5) has NaNO3 been compared to that obtained in an aqueous bpy buÜer for the complex.It should be recalled that the [Mn2O2(bpy)4]3` latter complex is stable in aqueous solution only in this speci –c medium.4,18a For the bpy complex, the quasi-reversible oxidation wave leading to the MnIVMnIV species appears at V vs. Ag/AgCl while the irreversible reduction E1@2\1.16 wave producing the mononuclear complex as [Mn(bpy)3]2` –nal product is seen [Fig. 3(A)] at V.5 For the Epc\0.40 terpy complex, potentials are strongly shifted to more positive values with regard to those of the bpy one, in accordance with the smaller electron-donating eÜect of the aqua ligand than that of pyridine. As a consequence, the irreversible reduction peak, leading to the mononuclear complex is [Mn(terpy)3]2`, now located at V [Fig. 3(B)]. Upon oxidation, no Epc\0.56 reversible redox system corresponding to formation of the MnIVMnIV species is observed. On the CV curve a large continuous increase of the anodic current with a shoulder at 1.00 V is observed by scanning the potential up to 1.40 V. On the reverse scan only a small irreversible peak at 0.86 V is seen.In order to explain the shape of the oxidation system of this complex, two possibilities can be proposed: (1) the oxidation potential of the MnIIIMIV/MnIVMnIV system is located outside the electroactivity range of the solvent or (2) the irreversibility of the system is due to a catalytic eÜect toward the oxidation of water. We are currently undertaking a careful bulk electrolysis study in order to try to solve these questions.In conclusion, this structural study validates the proposal of Limburg et al.6 about the formation of a binuclear l-oxo Fig. 2 Visible spectra in an aqueous solution containing 0.1 M at pH 4.5 of (a) 1 (0.9 mM) and (b) (1 mM) NaNO3 [Mn2O2(bpy)4]3` in 0.05 M bpy/bpyH` buÜer. 352 New J. Chem., 1999, 23, 351»353Fig. 3 Cyclic voltammograms in aqueous solutions (pH\4.5) at a carbon working electrode (diam. 0.5 cm) (scan rate 20 mV s~1) of (A) (2 mM) in 0.05 M bpy/bpyH` buÜer containing [Mn2O2(bpy)4]3` 0.1 M and (B) 1 (1 mM) containing 0.1 M Dashed NaBF4 NaNO3 . line curve is of the pure electrolyte. Potentials are referenced to an Ag/AgCl reference electrode. complex of manganese having two aqua ligands, during the operation of their catalytic water oxidation system.Acknowledgements authors thank Dr. J.-M. Latour for laboratory facilities The and for helpful discussions. Notes and references § Selected data for elemen- [Mn2O2(terpy)2(H2O)2](NO3)3 … 6H2O: tal analysis calculated for C 38.39, H 4.08, C30H38Mn2N9O19: N 13.43, Mn 11.71 ; found: C 38.17, H 3.87, N 13.71, Mn 11.63 ; ES-MS (4.5 eV) m/z : 732.0 687.0 [Mn2III, IVO2(terpy)2](NO3)2 `, 670.0 [Mn2IV, IVO2(terpy)2(OH)](NO3)`, [Mn2III, IVO2(terpy)2]- 625.0 454.0 (NO3)`, [Mn2III, IVO2(terpy)2(OH)]`, [Mn2IV, IVO2- 437.0 IR (KBr) (terpy)(OH)](NO3)`, [Mn2 III, IVO2(terpy)](NO3)`; cm~1: 3390(m, br), 1600(s), 1573(m), 1477(s), 1451(s), 1384(vs), 1247(w), 1166(w), 1025(m), 1015(m), 830(w), 777(s), 706(m).î Green-black, air-stable crystals of 1 suitable for X-ray diÜraction analysis can be obtained by evaporating the green solution of 1 at 0 °C for 6 h.Crystal data for 1. M\938.57, tri- C30H38Mn2N9O19 , clinic, space group (no 2) ; R\0.0617, wR\0.1673 ; a\7.7480(9), PY b\13.775(2), c\19.596(2) a\74.470(2), b\88.066(2), Aé , c\79.844(2)° ; U\1983.3(4) k\7.26 cm~1 T \293 K; Z\2; Aé 3; 12930 measured re—ections, 9085 were independent and 4292 were observed with I[2p(I).CCDC reference number 440/102. ° The stability of 1 in aqueous solutions was evaluated by the persistence of the green color of solutions and by the fact that no change in the absorption spectrum appears after several hours. 1 V. K. Yachandra, K. Sauer and M. P. Klein, Chem. Rev., 1996, 96, 2927 and references therein. 2 See, for instance : R. Manchanda, G. W. Brudvig and R. H. Crabtree, Coord. Chem. Rev., 1995, 144, 1 and references therein. 3 (a) R. Ramaraj, A. Kira and M. Kaneko, Angew. Chem., Int. Ed. Engl., 1986, 25, 825 and Chem. L ett., 1987, 261. (b) G. J. Yao, A. Kira and M. Kaneko, J. Chem. Soc., Faraday T rans. 1, 1988, 84, 4451. 4 M. G. Ghost, J. W. Reed, R. N.Bose and E. S. Gould, Inorg. Chem., 1994, 33, 73. 5 (a) M.-N. Collomb Dunand-Sauthier, A. Deronzier, X. Pradon, S. Meç nage and C. Philouze, J. Am. Chem. Soc., 1997, 119, 3173. (b) M.-N. Collomb Dunand-Sauthier, A. Deronzier, A. Piron, X. Pradon and S. Meç nage, J. Am. Chem. Soc., 1998, 120, 5373. 6 J. Limburg, G. W. Brudvig and R. H. Crabtree, J. Am. Chem. Soc., 1997, 119, 2761. 7 M. Stebler, A. Ludi and H.-B. Bué rgi, Inorg. Chem., 1986, 25, 4743. 8 R. Manchanda, G. W. Brudvig, S. De Gala and R. H. Crabtree, Inorg. Chem., 1994, 33, 5157. 9 M. B. Robin and P. Day, Adv. Inorg. Chem. Radiochem., 1967, 10, 247. 10 J. E. Sarneski, H. H. Thorp, G. W. Brudvig, R. H. Crabtree and G. K. Schulte, J. Am. Chem. Soc., 1990, 112, 7255. 11 K. R. Reddy, M. V. Rajasekharan, N. Arulsamy and D. J. Hodgson, Inorg. Chem., 1996, 35, 2283. 12 S. Meç nage, J.-C. Girerd and A. Gleizes, J. Chem. Soc., Chem. Commun., 1988, 431. 13 B. C. Dave, R. S. Czernuszewicz, M. R. Bond and C. J. Carrano, Inorg. Chem., 1993, 32, 3593. 14 S. Pal, M. M. Olmstead and W. H. Armstrong, Inorg. Chem., 1995, 34, 4708 and references therein. 15 P. M. Plaskin, R. C. Stoufer, M. Mathew and G. J. Palenik, J. Am. Chem. Soc., 1972, 94, 2121. 16 J. A. Gilbert, D. S. Eggleston, W. R. Murphy, Jr., D. A. Geselowitz, S. W. Gersten, D. J. Hodgson and T. J. Meyer, J. Am. Chem. Soc., 1985, 107, 3855. 17 K. Wieghardt, U. Bossek, B. Nuber, J. Weiss, J. Bonvoisin, M. Corbella, S. E. Vitols and J.-J. Girerd, J. Am. Chem. Soc., 1988, 110, 7398. 18 (a) S. R. Cooper and M. Calvin, J. Am. Chem. Soc., 1977, 99, 6623. (b) S. R. Cooper, G. C. Dismukes, M. P. Klein and M. Calvin, J. Am. Chem. Soc., 1978, 100, 7248. 19 U. N. Anderson, C. T. McKenzie and G. Bojesen, Inorg. Chem., 1995, 34, 1435. L etter 9/00814D New J. Chem., 1999, 23, 351»353 353

ISSN:1144-0546    

DOI:10.1039/a900814d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

L e t t e r The synthesis and structure of a macrobicyclic hexahalogenide trisdioximate as a promising precursor of functionalized clathrochelates Yan Z. Voloshin,*a Oleg A. Varzatskii,b Aleksei V. Palchik,b Adam I. Stasha and Vitaly K. Belskya a Karpov Institute of Physical Chemistry, 10, ul. V orontsovo Pole, Moscow 103064, Russia. Fax: ]7 095 975 24 50; E-mail : voloshin=cc.nifhi.ac.ru b Institute of General and Inorganic Chemistry, pr.Palladina 32/34, Kiev 252142, Ukraine Received (in Montpellier, France) 18th December 1998, Accepted 27th January 1999 The reactive hexachloride precursor of functionalized adioximate clathrochelates and the product of their reaction with thiophenol have been prepared and structurally characterized by single-crystal X-ray diÜraction. Compounds containing an encapsulated metal ion in the three-dimensional cavity of a macropolycyclic ligand have recently become of great interest to scientists working in different –elds of chemistry and biochemistry.This is mainly due to the unique properties displayed by ions of a metal that is encapsulated into the ììcageœœ of a macropolycyclic ligand, to a great extent isolated from external factors.1h5 High selectivity in the formation of such complexes, the screening of an encapsulated metal ion, and the low reactivity exhibited by the majority of the synthesized compounds limit studies of reactions of coordinated metal ions and the reactivity of macrobicyclic ligands.These factors can also explain difficulties in the synthesis of such compounds with a variety of metal ions.The –rst stage of the preparation of functionalized macrobicyclic oximates, oximehydrazonates, and their modi–ed complexes has involved the development of synthetic procedures for halide-containing precursors that permit the subsequent modi–cation and functionalization to be carried out by exploring the reactivity of halideoxime fragments. Two strategies for the synthesis of such precursors can be proposed (Scheme).The –rst one is based on halogenation of presynthesized clathrochelates. High yields and the availability of these clathrochelates are, undoubtedly, the advantages of this approach. However, the proposed reaction pathway cannot commonly be used because of a partial substitution of hydrogen atoms for the halides.In addition, side reactions of the macrobicyclic framework eÜect by halogenation agents may also take place. Furthermore, with some substituents at the boron atoms, halogenation can also occur under the same conditions. The second strategy, which is based on the use of preformed dihalodioximes, does not have the disadvantages of the –rst one. Although conventional synthetic routes were unsuccessful, we have managed to select conditions to give yields of 90%.Stepwise addition of the complex to a Fe(CH3CN)4Cl2 boiling solution/suspension of dichloroglyoxime (H2Cl2Gm) and phenylboronic acid in nitromethane with a 1 : 3.5 : 2 ratio for 1 h and dehydration of the solvent with 4 molecular ” sieves followed by distillation of a proportion of the solvent led to the formation of a solid clathro- Fe(Cl2Gm)3(BC6H5)2 chelate.After the suspension was cooled, the precipitate was –ltered oÜ, washed with nitromethane, a small amount of chloroform, diethyl ether, hexane and recrystallized from hot methylene dichloride. The hexachloride compound 1 readily reacted with active thiol and amino groups to produce functionalized clathrochelates that were capable of coordinating metal ions using peripheral fragments.The intensive stirring of a solution/suspension in hot 1,4- Fe(Cl2Gm)3(BC6H5)2 dioxane with an excess of thiophenol in the presence of led to the formation of a red-brown reaction mixture. K2CO3 The latter was evaporated to a small volume, and the product was precipitated by adding a –vefold volume of ethanol. The resulting crystalline precipitate Fe[(C6H5S)2Gm]3(BC6H5)2 (2) was washed with ethanol, diethyl ether and reprecipitated from chloroform with hexame.§ Our attempts to remove the Fe2` ion from the cavity and isolate free macrobicyclic ligands were unsuccessful due to decomposition of the latter. Scheme New J.Chem., 1999, 23, 355»357 355Fig. 1 Molecular structure of 1. Selected bond lengths and bond (”) angles (°) are as follows : FewN1, 1.894(8) ; FewN2, 1.905(8) ; FewN3, 1.913(8) ; O1wB, 1.51(3) ; N1wC1, 1.27(1) ; Cl1wC1, 1.712(9) ; 1.42(2) ; 78.7(6) ; N1wFewN2, 85.4(4) ; C1wC1j1, N1j1wFewN1, O1wN1wFe, 121(1) ; C1wN1wFe, 117.8(8) ; N1wO1wB, 110(2) ; 112.8(6) ; O2wBwO1, 106(1). Symmetry transform- N1wC1wC1j1, ations used to generate equivalent atoms: [x]1, y, [z]1/2.j1 The 1H NMR spectrum of precursor 1 contained two multiplets at approximately 7.4 and 7.85 ppm, whereas that of complex 2 revealed only one multiplet at approximately 7.1 ppm. The 13C NMR spectra (with and without 13C»1H interaction decoupling) permitted the identi–cation of a signal of Fig. 2 Molecular structure of 2. H atoms are omitted and phenyl substituents are denoted as Ph for clarity.Selected bond lengths (”) and bond angles (°) are as follows : FewN1, 1.920(4) ; FewN2, 1.901(4) ; FewN3, 1.922(4) ; O1wN1, 1.361(4) ; O1wB1, 1.495(7) ; N1wC1, 1.297(6) ; C1wC6, 1.447(6), N6wFewN1, 78.6(2) ; C1wN1wFe, 117.7(3) ; O2wB1wO1, 109.5(4) ; C1wS1wC19, 96.1(2) ; N1wO1wB1, 113.2(4). Fig. 3 The iron(II) coordination polyhedra in precursor 1 (top) and in the resulting clathrochelate 2 (bottom).the azomethine fragment in the precursor (131.6 ppm) practically coinciding with one of the signals of a phenyl substituent at the boron atom (127.7, 128.9 and 131.7 ppm). The signal of the carbon atom bound to the boron atom was not detected because of quadrupole broadening. The 13C NMR spectrum of the starting dichloroglyoxime contained a signal due to the azomethine groups at 130.9 ppm, which is lower than those of alicyclic and acyclic dioximes.The spectrum of clathrochelate 2 indicated that four more intense signals of thiophenyl groups appeared in the same spectral range (128»132 ppm), and the signal of azomethine groups was observed in the range characteristic of the majority of alicyclic, acyclic and aromatic clathrochelate trisdioximates (148.5 ppm).The parameters of the Moé ssbauer 57Fe spectra of the complexes obtained characterize the s-electron density on the iron nucleus and the ligand –eld force (isomer shift, IS) as well as the electric –eld gradient on it (quadrupole splitting, QS). The QS magnitude is determined by the geometry of a coordination polyhedron and can be used for its prediction.6 A comparison of the parameters of complexes 1 (IS\0.39 mm s~1, QS\0.68 mm s~1) and 2 (IS\0.34 mm s~1,QS\0.25 mm s~1) revealed that the clathrochelate ligands have close ligand –eld force values and appreciably diÜerent geometries.All compounds of this type have a geometry intermediate between that of a trigonal prism (TP, twist angle u\0°) and that of a trigonal antiprism (TAP, u\60°). Judging from the correlation relationships6 and equations,7 the observed discrepancy in the QS values corresponds to an increase in the twist angle value by about 15 to 20° in passing from precursor 1 to func- 356 New J.Chem., 1999, 23, 355»357tionalized complex 2. The X-ray analyses of complexes 1 and 2 from crystals obtained by slow evaporation of their solutions in (1 : 1) and benzene, respectively, has con–rmed CHCl3»CCl4 this conclusion (Fig. 1»3).î The mean Fe»N distances (1.90 and 1.91 and the bite angles (half of the chelate angles) are ”) practically identical (39.0 and 39.5°) and typical for clathrochelate trisdioximate iron(II) complexes, whereas the twist angles (5.4 and 25.6°) and the distances between the trigonalprismatic coordination polyhedron bases are signi–cantly different (2.39 and 2.33 respectively), which is in agreement ”, with the Moé ssbauer 57Fe data.Hexachloride complex 1 has the lowest value among the known compounds of this type.9 The relatively short (approximately 3.5 S… … …S contacts ”) between the adjacent molecules in the crystals of complex 2 should also be noted.Essential diÜerences in the geometry of coordination polyhedra and in the behavior of the substituents in dioxime fragments are responsible for appreciable diÜerences in the UV-vis spectra : in the spectrum of complex 2 the intense Md]Lp* charge transfer band in the visible region is shifted to the long-wavelength region by approximately 40 nm and has twice the intensity compared to the spectrum of complex 1.Acknowledgements of the Russian Fund of Fundamental Researches Support (Grants N96-03-33512a and N97-03-33776) is gratefully acknowledged. Notes and references § Analytical, MS(FAB, PD), IR(CsI) and UV-vis (solvent : CHCl3) spectral data: 1. Anal. calcd. for C, 31.02 ; H, C18H10N6O6B2Cl6Fe: 1.44 ; N, 12.06 ; Cl, 30.59 ; Fe, 8.01. Found: C, 30.95 ; H, 1.44 ; N, 11.99 ; Cl, 30.78 ; Fe, 7.95%.MS(PD): m/z (%) 696 (20) 573 (100) [M]`~, [M 541 (35) IR: 1533 (CxN); 905, [Cl2C2N2]`~, [M[Cl2C2N2O2]`~. 980 (NwO); 1225 (BwO) cm~1; e/mol L~1 cm~1)\266 j'(10~3 (12), 290 (7.9), 308 (4.9), 346 (2.8), 425 (5.9), 454 (14) nm. 2. Anal. calcd. for C, 56.90 ; H, 3.51 ; N, 7.38 ; Fe, 4.90. Found, C54H40N6O6B2FeS6 : C, 56.76 ; H, 3.56 ; N, 7.30 ; Fe, 4.86%.MS(FAB): m/z (%) 1139 (100) [M]H]`, 1062 (40) IR: 1582 (CxN); 893, 972 [M]H[C6H5]`. (NwO); 1233 (BwO) cm~1; (10~3 e/mol L~1 cm~1)\276 (23), j' 320 (18.5), 388 (4.5), 495 (25) nm. î Crystal data for 1: crystal size C18H10B2Cl6FeN6O6 , 0.60]0.20]0.02 mm, M\696.49, monoclinic, space group C2/c, a\24.290(5), b\8.169(2), c\15.807(3) b\127.42(3)°, ”, U\2491.0(9) Z\4, g cm~3, k\1.300 mm~1, ”3, Dc\1.857 F(000)\1384. 864 independent re—ections were collected at 293 K on a Syntex diÜractometer using radiation (j\0.71073 P16 Mo-K a ”) with h/2h scans (2.11\h\22.49°). The structure was solved by the heavy-atom method, re–nement was done by a full-matrix least squares on F2 for all data with anisotropic thermal parameters for non-hydrogen atoms; phenyl substituents were constrained in accordance with a riding model Mgoodness-of-–t\1.108, –nal R indices [I[2p(I)R1\0.0480, wR2\0.1098, w~1\p2(Fo2)](0.046P)2 All oxygen atoms are statistically dis- ]20.19P, P\(Fo2]2Fc2)/3]N.ordered in two equivalent positions each. Crystal data for 2: crystal size C54H40B2FeN6O6S6 , 0.70]0.35]0.35 mm, M\1138.75, monoclinic, space group C2/c, a\16.874(2), b\20.477(3), c\31.705(3) b\104.62(3)°, ”, U\10600(7) Z\8, g cm~3, k\0.578 mm~1, ”3, Dc\1.427 F(000)\4688. 7437 independent re—ections were collected at 293 K on a CAD4 diÜractometer using radiation (j\0.71073 Mo-K a ”) with h/2h scans (1.60\h\22.47°). The structure was solved by the heavy-atom method, re–nement was by a full-matrix least squares on F2 for all data with anisotropic displacement parameters for nonhydrogen atoms; phenyl substituents were constrained in accordance with a riding model Mgoodness-of-–t\1.034, –nal R indices [I[ 2p(I)]R1\0.0342, wR2\0.0947, w~1\p2(Fo2)](0.063P)2]26.56P, P\(Fo2]2Fc2)/3]N.All calculations were made using the SHELXTL-93 program package.8 CCDC reference number 440/098. See http ://www.rsc.org/ suppdata/nj/1999/355/ for crystallographic –les in .cif format. 1 J.-M. Lehn, Supramolecular Chemistry, VCH, Weinheim, 1995. 2 A. M. Sargeson, Pure Appl. Chem., 1984, 561, 1603. 3 A. M. Sargeson, ibid., 1986, 58, 1511. 4 N. A. Kostromina, Y. Z. Voloshin and A. Y. Nazarenko, Clathrochelates : Synthesis, Structure and Properties, Naukova Dumka, Kiev, 1992. 5 A. M. Sargeson, Coord. Chem. Rev., 1996, 151, 89 and references therein. 6 Y. Z. Voloshin, N. A. Kostromina and A. Y. Nazarenko, Inorg. Chim. Acta, 1990, 170, 181. 7 A. Y. Nazarenko, E. V. Polshin and Y. Z. Voloshin, Mendeleev Commun., 1993, 45. 8 G. M. Sheldrick, SHEL XT L -93 Programs for Structure Re–nement, University of Goé ttingen, Germany, 1993. 9 Y. Z. Voloshin, T. E. Kron, V. K. Belsky, V. E. Zavodnik, Y. A. Maletin and S. G. Kozachkov, J. Organomet. Chem., 1997, 536ñ537, 207 and references therein. L etter 8/09881F New J. Chem., 1999, 23, 355»357 357

ISSN:1144-0546    

DOI:10.1039/a809881f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

L e t t e r A combined covalent and coordination approach to dendritic multiporphyrin arrays based on ruthenium(II) porphyrins Scott L. Darling,a Chi Ching Mak,a Nick Bampos,a Neil Feeder,a Simon J. Teatb and Jeremy K. M. Sanders*a a Cambridge Centre for Molecular Recognition, University Chemical L aboratory, L ens–eld Road, Cambridge, UK CB2 1EW . E-mail : jkms=cam.ac.uk b CL RC Daresbury L aboratory, Daresbury, W arrington, UK WA4 4AD Received (in Montpellier, France) 11th January 1999, Accepted 11th February 1999 Two strategies have been employed in the construction of multiporphyrin arrays ; the principles are illustrated by a sevenporphyrin array consisting of two three-porphyrin dendrons coordinated to a single ruthenium(II) porphyrin, and by a sixporphyrin array containing three ruthenium(II) porphyrin monomers coordinated to a tripyridyl porphyrin trimer.These systems possess both redox and photophysical functionality built-in to the overall structure. We present here two approaches leading to large arrays of porphyrins. In essence, our approaches employ oligomeric porphyrins as donor ligands that coordinate to ruthenium(II) porphyrin monomers.1,2 The principles are illustrated here by a seven-porphyrin array containing two three-porphyrin dendrons coordinated to a single ruthenium(II) porphyrin and by a six-porphyrin array containing three ruthenium(II) porphyrin monomers coordinated to a tripyridyl porphyrin trimer.This strategy can be extended to dendritic arrays, and is complementary to a related approach we have described elsewhere. 3 Our –rst approach is based on porphyrin oligomers (dendrons) that contain pyridine units for coordination to ruthenium(II) porphyrin monomer building blocks (Scheme 1). Porphyrin monomers 1 and 2 were synthesised according to previously reported procedures.4 A palladium-mediated coupling reaction of 1 and 2, followed by removal of the pyridine N-oxide protecting group by aÜords the pyridyl PCl3 , porphyrin dendron 3 in 55% yield.The N-oxide-protected pyridyl porphyrin 2 gave higher yields than the unprotected analogue and allowed easier chromatographic separation of the N-oxide-protected dendron 3 following the palladium coupling. The ruthenium(II) porphyrin 4, used as the core of the large arrays, was synthesised essentially as described previously.5 Complexation of 4 with a stoichiometric amount of dendron 3 at room temperature (Scheme 2), produced the monocoordinated complex 5 by displacement of the labile solvent ligand, leaving the axially bound CO ligand intact.Characterisation of these solution-state arrays by 1H NMR spectroscopy allowed assignment of the resonances of the array, in particular the NH resonances associated with the coordinated dendron 3, which appear as four sharp singlets in a 2 : 2 : 1 : 1 ratio for the outer and inner NHs (Fig. 1), whilst three sharp resonances (1 : 1 : 2) were evident for the meso protons. Characteristic up–eld shifts for the doublets (d 5.86 for the bpyridyl protons) of the ruthenium-bound pyridyl protons were also identi–ed. Addition of a further equivalent of 3 to 5 under photolysis conditions6 for 24 h leads to displacement of the Scheme 1 (i) 25 °C, 14 h.(ii) 25 °C, 16 h. Pd2(dba)3 , AsPh3 , NEt3, CH2Cl2 , PCl3, CH2Cl2 , New J. Chem., 1999, 23, 359»364 359Scheme 2 (i) 1 equiv. of 3, 25 °C. (ii) 1 equiv. of 3, hv, 24 h. C6D6, C6D6 , axial CO ligand of 4 (Scheme 2) aÜording mixtures of both mono- 5 and bis-coordinated 6 species as judged by 1H NMR. The bis-coordinated complex 6 can be characterised by two new singlets corresponding to the inner NH protons, in addition to up–eld-shifted resonances for the meso signals and the b-pyridyl doublet (d 5.42) of the metal-bound dendron 3.Our second approach was to construct porphyrin oligomers containing tethered pyridine units to act as multidentate ligands for the coordination of ruthenium(II) monomer building blocks (Scheme 3).Porphyrin monomer 7 was synthesised according to standard procedures,4 while palladium coupling of 7 and 2, with subsequent deprotection of the N-oxide as before, aÜords the tridentate trimer ligand 8 in 61% yield. Titration of three equivalents of 4 in solution (mM) to CDCl3 a solution of 8 provides the hexamer array 9 (Scheme 4), the integrity of which was con–rmed by 1H NMR spectroscopy (Fig. 2). The two meso resonances of 8 are observed in a 1 : 2 ratio corresponding to the two types of pyridine-appended porphyrins of the ligand (P1 and P2). In the hexamer array 9 four meso resonances are observed, two for the P1 and P2 components of the central trimer and two for the ruthenium porphyrins bound to P1 and P2 of the ligand (Ru-P1 and Ru-P2).In addition, two new doublets in a 1 : 2 ratio are observed at d 5.84 and 5.81 assigned to P1 and P2 rutheniumbound b-pyridine protons of 8. The neighbouring a1 and a2 protons are hidden in the congested aliphatic region, but could be identi–ed at d 1.66 and 1.64 with the aid of a COSY spectrum. In the NOESY spectrum two crosspeaks are detected between the a- and b-pyridine protons of 8 and the Ru-P1 and Ru-P2 meso resonances, allowing assignment of the P1, P2, Ru-P1 and Ru-P2 meso protons and consequently the aromatic and aliphatic resonances.Fig. 1 High-–eld region of the 1H NMR spectrum (400 MHz, 300 K) after photolysis of ruthenium(II) porphyrin 4 in the presence of the CDCl3 , pyridine porphyrin dendron.The NH resonances of the dendron 3 in the 1H NMR showing (a) free dendron 3, (b) mono-coordinated complex 5 and (c) mixture of both mono- 5 and bis- 6 coordinated species. 360 New J. Chem., 1999, 23, 359»364Scheme 3 (i) 25 °C, 14 h. (ii) 25 °C, 16 h. Pd2(dba)3 , AsPh3 , NEt3, CH2Cl2 , PCl3, CH2Cl2 , As yet, crystals of 9 suitable for X-ray diÜraction analysis have not been obtained.However, we have prepared smaller coordination arrays ; suitable crystals of 10 7 were prepared by treatment of one equivalent of the tridentate ligand tripyridyltriazine with three equivalents of 4 in (Py3T) CH2Cl2 (Scheme 5). After many attempts only very small single crystals of 10 could be grown. The best crystals were obtained by slow solvent diÜusion of methanol into a toluene solution.These crystals proved to be too small to determine the structure by normal laboratory X-ray methods, so it was necessary to exploit the high intensity of a synchrotron radiation source to obtain the structure. The structure (Fig. 3) consists of three ruthenium porphyrins each axially bound to the pyridine nitrogens of the ligand, with the aromatic rings of Py3T Py3 T coplanar.All the porphyrin aryl substituents lie perpendicular to the central ligand, presumably due to steric con- Py3T straints. The angles of the nitrogens with respect to the Py3T porphyrin plane are in the range of 85»90°, and the distances [2.19(3), 2.22(8) and 2.15(6) are similar Ru»N(Py3T) ”] to those reported previously.8 No mass spectral characterisation was possible for any of these coordination arrays : even when neutral matrices were used we observed only ions due to the ligand and ruthenium porphyrin core.The two strategies employed demonstrate the ease with which functionality can be incorporated into large multiporphyrin arrays. Variation in the structural framework of the porphyrin building blocks and investigations into the photophysical properties are currently being undertaken.Experimental 1H NMR spectra (400, 500 or 800 MHz) were recorded on Bruker DRX-400, DRX-500 or DRX-800 spectrometers, respectively. 13C NMR spectra were obtained on a DRX-400 operating at 100.6 MHz. All NMR measurements were carried out at room temperature in deuterochloroform unless otherwise speci–ed.Routine UV/visible spectra were obtained on a Uvikon 810 spectrometer in 10 mm oven-dried cuvettes. Distilled solvents were used throughout and when used dry, were freshly obtained from the solvent stills. Triethylamine and Scheme 4 (i) 1 equiv. of 8, 25 °C. CDCl3 , New J. Chem., 1999, 23, 359»364 361Fig. 2 Selected down–eld regions of the 1H NMR spectrum (400 MHz, 300 K) demonstrating assembly of ruthenium building CDCl3 , blocks around tridentate trimer ligand 8; (a) free tridentate trimer ligand 8 and (b) hexamer complex 9.dichloromethane were distilled from under (CH2Cl2) CaH2 argon while toluene and tetrahydrofuran (THF) were distilled from or sodium, also under argon. MALDI-TOF mass CaH2 spectra were recorded on a Kratos Analytical Ltd Kompact MALDI IV mass spectrometer.A nitrogen laser (337 nm, 85 Scheme 5 (i) 1 equiv. of 25 °C. Py3T, CH2Cl2 , Fig. 3 A ball-and-stick representation of the X-ray structure of 10. kW peak laser power, 3 ns pulse width) was used to desorb the sample ions, and the instrument was operated in linear time-of-—ight mode with an accelerating potential of 20 kV. Results from 50 laser shots were signal-averaged to give one spectrum.An aliquot (1 ll) of a saturated solution of the matrix (sinapinic acid) was deposited on the sample plate surface. Before the matrix completely dried, a small volume (1 ll) of analytes (dissolved in dichloromethane»chloroform at 1 mg ml~1) was layered on the top of the matrix and allowed to air-dry. Syntheses Synthesis of dendron 3: Complex 1 (137.7 mg, 130.8 lmol), 2 (70.0 mg, 62.3 lmol), (8.5 mg, 9.3 lmol), and Pd2(dba)3 AsPh3 (33.2 mg, 74.7 lmol) were dissolved in freshly distilled CH2Cl2 (8 ml) and (8 ml).The solution was saturated with argon NEt3 (three freeze»thaw cycles) and stirred at room temperature for 14 h, after which the solvent was removed in vacuo and the residue chromatographed through eluting –rst with SiO2 , (5 : 1 : 1) until a –rst band clearly C6H14 : CHCl3 : EtOAc separated.The elution was continued with (4 : 1 : 1) giving the N-oxide analogue C6H14 : CHCl3 : EtOAc of 3 (135 mg, 73%) as a –ne dark red powder. The N-oxide analogue of 3 (35.0 mg, 11.8 lmol) was dissolved in freshly distilled (20 ml) and cooled to 0 °C. (5.1 ll, 58.8 CH2Cl2 PCl3 lmol) was added dropwise and the solution was stirred under dry air for 16 h, after which the solution was washed with (2]100 ml) and (3]100 ml), and the solvent NaHCO3 H2O was removed in vacuo.The compound was recrystallized by layered addition of methanol to a chloroform solution of 3, –ltered, and dried in vacuo to aÜord a red powder of 3 (19.2 mg, 55%). 1H NMR (400 MHz, 10.36 (s, 2H, inner CDCl3) : d meso), 10.24 (s, 4H, outer meso), 9.05 (d, J\4.8 Hz, 2H, a-py), 8.54 (t, J\1.5 Hz, 1H, inner ArH), 8.53 (t, J\1.5 Hz, 2H, inner ArH), 8.14 (d, J\8.1 Hz, 4H, outer ArH), 8.12 (d, J\4.8 Hz, 2H, b-py), 8.05 (d, J\8.1 Hz, 4H, outer ArH), 7.92 (d, J\1.8 Hz, 4H, outer ArH), 7.81 (t, J\1.8 Hz, 2H, outer ArH), 4.12 (m, 8H, inner 4.00 (m, 16H, outer Por-CH2), 2.88 (s, 6H, inner pyrrolic 2.58 (s, 12H, outer Por-CH2), CH3), pyrrolic 2.56 (s, 6H, inner pyrrolic 2.47 (s, 12H, CH3), CH3), 362 New J.Chem., 1999, 23, 359»364outer pyrrolic 2.33 (m, 8H, inner 2.20 CH3), Por-CH2-CH2), (m, 16H, outer 1.86 [m, 8H, inner Por-CH2-CH2), Por- 1.74 [m, 16H, outer 1.58 [m, (CH2)2-CH2], Por-(CH2)2-CH2], 8H, inner 1.51 (s, 36H, But), 1.48 [m, 16H, Por-(CH2)3-CH2], outer 1.38 [m, 24H, inner/outer Por-(CH2)3-CH2], Por- 0.99 [m, 12H, inner 0.90 [m, (CH2)4-CH2], Por-(CH2)5-CH3], 24H, [2.28 (s, 1H, inner NH), [2.33 (s, Por-(CH2)5-CH3], 1H, inner NH), [2.37 (s, 4H, outer NH) ; 13C NMR (100.6 MHz, 152.3, 150.8, 149.9, 149.0, 145.3, 144.7, 144.0, CDCl3) : d 2 ]143.5, 143.1, 141.9, 141.6, 141.5, 141.4, 140.9, 136.7, 136.2, 135.7, 135.4, 133.2, 131.0, 129.7, 127.6, 123.4, 122.8, 121.1, 119.6, 116.6, 116.2, 114.6, 97.5, 97.0, 90.9, 89.6, 35.1, 33.5, 33.3, 32.1, 32.0, 31.7, 30.1, 30.0, 26.8, 22.8, 15.6, 2]15.0, 14.3, 14.2, 14.1 ; UV/vis 411, 507, 541, 547, (log[e/M~l jmax (CH2Cl2)/nm: cm~l]) 5.8, 4.8, 4.2, 4.3 ; MALDI-MS: requires C199H261N13 2835.37, found 2834.91 (MH`).General procedure for photolysis reaction : Complex 3 (13.6 mg, 4.8 lmol) and 4 (2.9 mg, 2.3 lmol) contained in a quartz NMR tube (Aldrich) were dissolved in (1 ml).The tube C6D6 was degassed by evacuating the area above the sample solution at room temperature and exposed to a 60 W broad-band mercury lamp for 24 h. Preparation of 5: Complex 5 was prepared from the stoichiometric addition of 3 to 4. 1H NMR (400 MHz, CDCl3) : d 10.23 (s, 4H, outer meso), 10.07 (s, 2H, Ru meso), 10.02 (s, 2H, inner meso), 8.47 (t, J\1.7 Hz, 1H, inner ArH), 8.38 (d, J\1.7 Hz, 2H, inner ArH), 8.11 (d, J\8.0 Hz, 4H, outer ArH), 8.05 (t, J\1.7 Hz, 2H, Ru ArH), 7.99 (d, J\8.0 Hz, 4H, outer ArH), 7.91 (d, J\1.7 Hz, 6H, Ru/outer ArH), 7.82 (t, J\1.7 Hz, 2H, outer ArH), 7.81 (t, J\1.7 Hz, 2H, outer ArH), 5.86 (d, J\6.5 Hz, 2H, b-py), 3.98 (t, J\7.2 Hz, 16H, outer 3.89 (br t, 8H, Ru 3.70 (br t, 8H, Por-CH2), Por-CH2), inner 2.75 (s, 6H, inner pyrrolic 2.55 (s, 12 H, Por-CH2), CH3), Ru pyrrolic 2.46 (s, 12H, outer pyrrolic 2.45 (s, CH3), CH3), 12H, outer pyrrolic 2.27»2.14 (m, 32H, Ru/inner/outer CH3), 2.17 (s, 6H, inner pyrrolic 1.94 [m, 8H, Por-CH2-CH2), CH3), Ru 1.84»1.63 [m, 24H, inner/outer Por-(CH2)2-CH2], Por- 1.57 (s, 18H, Ru But), 1.53 (s, 18H, Ru But), 1.51 (CH2)2-CH2], (s, 36H, outer But), 1.48»1.27 [m, 66H, Ru/inner/outer Pora- py], 0.95»0.87 [m, 48H, Ru/ (CH2)3-CH2 , Por-(CH2)4-CH2 , inner/outer [2.39 (s, 2H, outer NH), Por-(CH2)5-CH3], [2.40 (s, 2H, outer NH), [2.66 (s, 1H, inner NH), [2.97 (s, 1H, inner NH) ; 13C NMR (100.6 MHz, d 149.9, CDCl3) : 149.7, 149.1, 148.3, 145.3, 145.0, 144.7, 143.8, 143.5, 2]143.4, 143.2, 143.1, 142.9, 142.7, 142.5, 141.6, 141.5, 141.4, 141.3, 141.1, 141.0, 140.9, 137.6, 136.6, 136.0, 2]135.7, 134.3, 133.2, 130.9, 128.3, 127.8, 127.5, 125.1, 123.3, 122.7, 121.1, 2]120.7, 119.6, 116.5, 116.1, 112.7, 98.7, 97.2, 97.0, 90.8, 89.5, 58.4, 35.2, 35.1, 35.0, 33.3, 33.1, 32.0, 2]31.9, 2]31.7, 31.6, 2]30.0, 2]29.9, 26.9, 26.8, 22.8, 3]22.7, 18.4, 15.5, 15.0, 14.9, 14.2, 14.1, 13.4 ; UV/vis 407, 509, 574, (log[e/ jmax (CH2Cl2)/nm: Mv1cmv1]) 5.8, 4.8, 4.3.Synthesis of tridentate trimer 8: Complex 8 was synthesised by a similar procedure to that reported above for 3, except that 7 was used instead of 1 (yield : 61%). 1H NMR (400 MHz, 10.33 (s, 2H, P1 meso), 10.26 (s, 4H, P2 meso), 9.03 CDCl3) : d (d, J\4.7 Hz, 2H, P1 a-py), 8.98 (d, J\4.7 Hz, 4H, P2 a-py), 8.55 (s, 1H, P1 ArH), 8.54 (s, 2H, P1 ArH), 8.10 (d, J\8.0 Hz, 4H, P2 ArH), 8.06 (d, J\5.1 Hz, 2H, P1 b-py), 8.03 (d, J\8.0 Hz, 4H, P2 ArH), 8.01 (d, J\5.1 Hz, 4H, P2 b-py), 4.10 (m, 8H, P1 3.99 (m, 16H, P2 2.89 Por-CH2), Por-CH2), (s, 6H, P1 pyrrolic 2.57 (s, 12H, P2 pyrrolic 2.53 CH3), CH3), (s, 6H, P1 pyrrolic 2.48 (s, 12H, P2 pyrrolic 2.31 CH3), CH3), (m, 8H, P1 2.20 (m, 16H, P2 Por-CH2-CH2), Por-CH2-CH2), 1.85 [m, 8H, P1 1.74 [m, 16H, P2 Por- Por-(CH2)2-CH2], 1.58»1.38 [m, 48H, P1/P2 (CH2)2-CH2], Por-(CH2)3-CH2 , 0.99 [m, 12H, P1 0.91 Por-(CH2)4-CH2], Por-(CH2)5-CH3], [m, 24H, P2 [2.38 (br s, 6H, P1/P2 NH) ; Por-(CH2)5-CH3], 13C NMR (100.6 MHz, d 150.7, 149.0, 145.0, 144.0, CDCl3) : 143.9, 143.7, 142.8, 141.8, 2]141.7, 141.4, 2]136.1, 135.8, 135.4, 135.2, 133.1, 131.0, 128.4, 123.3, 122.9, 117.4, 114.3, 2]97.3, 90.8, 89.7, 33.4, 33.3, 32.0, 31.9, 30.1, 29.9, 29.7, 26.7, 2]22.7, 15.6, 14.9, 2]14.1 ; UV/vis 412, jmax (CH2Cl2)/nm: 508, 540, 574, (log[e/Mv1 cmv1]), 5.6, 4.5, 3.9, 4.0 ; MALDI-MS: requires 2612.92, found 2612.54 (MH`). C181H227N15 Preparation of 9: Complex 9 was prepared in solution by stiochiometric addition of 4 to 8. 1H NMR (800 MHz, d 10.03 (s, 2H, P1 meso), 10.01 (s, 2H, Ru-P1 meso), CDCl3) : 10.00 (s, 4H, Ru-P2 meso), 9.95 (s, 4H, P2 meso), 8.39 (s, 1H, P1 ArH), 8.32 (s, 2H, P1 ArH), 8.04 (br t, 6H, Ru-P1 ArH), 7.94 (d, J\8.6 Hz, 4H, P2 ArH), 7.90 (br d, 4H, P2 ArH), 7.89 (s, 6H, Ru-P2 ArH), 7.81 (s, 6H, Ru-P2 ArH), 5.84 (d, J\6.3 Hz, 2H, P1 b-py), 5.81 (d, J\6.3 Hz, 4H, P2 b-py), 3.96 (m, 28H, 3.82 (m, 8H, 3.68 (m, 4H, Por-CH2), Por-CH2), Por- 3.63 (m, 8H, 2.69 (s, 12H, pyrrolic 2.44 CH2), Por-CH2), CH3), (s, 24H, pyrrolic 2.53 (s, 24H, pyrrolic 2.41 (s, CH3), CH3), 12H, pyrrolic 2.26 (m, 24H, 2.13 (m, CH3), Por-CH2-CH2), 4H, 2.02 (m, 8H, 1.91 (m, 4H, Por-CH2-CH2), Por-CH2-CH2), 1.87 (m, 8H, 1.79 [m, 36H, Por-CH2-CH2), Por-CH2-CH2), 1.70 [m, 4H, 1.66»1.25 Por-(CH2)2-CH2], Por-(CH2)2-CH2], [m, 218H, P1/P2 a-py, Por-(CH2)2-CH2 , Por-(CH2)3-CH2 , But], 0.92»87 [m, 60H, Por-(CH2)4-CH2 , Por-(CH2)5-CH3], 0.82 [t, J\7.5 Hz, 12H, [2.72 (s, 1H, P1 Por-(CH2)5-CH3], NH), [2.80 (s, 2H, P2 NH), [3.02 (s, 1H, P1, NH), [3.06 (s, 2H, P2 NH) ; 13C NMR (100.6 MHz, d 199.0, 151.0, CDCl3) : 149.7, 144.9, 144.7, 2]143.9, 143.8, 143.5, 143.3, 142.9, 142.6, 142.4, 141.4, 140.9, 137.5, 136.0, 134.3, 134.1, 2]132.9, 130.9, 128.3, 127.8, 125.0, 120.7, 98.7, 97.0, 35.2, 35.0, 33.2, 33.1, 32.0, 31.9, 31.8, 31.7, 30.0, 29.8, 29.7, 29.3, 26.9, 26.7, 26.5, 22.8, 22.7, 15.0, l4.0 ; UV/vis 406, 514, (log[e/Mvl jmax (CH2Cl2)/nm: cmv1]) 5.7, 4.6.Acknowledgements thank the EPSRC, the BBSRC (funding the purchase of We the 800 MHz NMR spectrometer), the Croucher Foundation and Zeneca for generous –nancial support.Notes and references 1 For multiporphyrin arrays with donor ligands tethered to the porphyrin periphery see : (a) W. T. S. Huck, A. Rohrer, A. T. Anilkumar, R. H. Fokkens, N. M. M. Nibbering, F. C. J. M. van Veggel and D. N. Reinhoudt, New J. Chem., 1998, 22, 165; (b) C. M. Drain, F. Ni–atis, A. Vasenko and J.D. Batteas, Angew. Chem., Int. Ed., 1998, 37, 2344. 2 For multiporphyrin arrays with porphyrins as the donor ligands themselves see : (a) K. Funatsu, T. Imamura and A. Ichimura, Inorg. Chem., 1998, 37, 4986; (b) N. Kariya, T. Imamura and Y. Sasaki, Inorg. Chem., 1998, 37, 1658; (c) K. Funatsu, T. Imamura, A. Ichimura and Y. Sasaki, Inorg. Chem., 1998, 37, 1798; (d) E. Alessio, M.Macchi, S. Heath and L. G. Marzilli, Chem. Commun., 1996, 1411; (e) M. R. Johnston, M. I. Cunter and R. N. Warrener, Chem. Commun., 1998, 2739. 3 C. C. Mak, N. Bampos and J. K. M. Sanders, Chem. Commun., submitted. 4 C. C. Mak, N. Bampos and J. K. M. Sanders, Angew. Chem., Int. Ed., 1998, 37, 3020. New J. Chem., 1999, 23, 359»364 3635 V. Marvaud, A. Vidal-Ferran, S. J.Webb and J. K. M. Sanders, J. Chem. Soc., Dalton T rans., 1997, 985. 6 F. R. Hopf, T. P. OœBrien, W. P. Scheidt and D. G. Whitten, J. Am. Chem. Soc., 1975, 97, 277. 7 Crystal data for 10: M\4208.97, triclinic, C270H360N18O3Ru3 , space group a\19.417(2), b\22.545(2), c\33.326(3) P1 6 , Aé , a\84.930(10), b\73.730(10), c\65.400(10)o, U\12727(2) ”3, Z\2, Mg mv3, k\0.69170 l\0.231 mmv1, Dc\1.098 ”, F(000)\4524, T \150(2) K. Crystal of size 0.04]0.04]0.02 mm. Data were collected on a Siemens SMART CCD diÜractometer on the single-crystal diÜraction station (No. 9.8) at the Daresbury Laboratory Synchrotron Radiation Source (UK). The structure was solved by direct methods and re–ned by full-matrixblock least-squares on F2; 52789 re—ections measured of which 34245 were observed as unique. The structure was re–ned to R\0.1184, Rw\0.3124, including disordered solvent molecules. CCDC reference number 440/101. See http ://www.rsc.org/ suppdata/nj/1999/359/ for crystallographic –les in .cif format. 8 (a) K. Funatsu, A. Kimura, T. Imamura, A. Ichimura and Y. Sasaki, Inorg. Chem., 1997, 36, 1625; (b) R. G. Little and J. A. Ibers, J. Am. Chem. Soc., 1973, 95, 8583; (c) S. L. Darling and J. K. M. Sanders, unpublished results. L etter 9/00344D 364 New J. Chem., 1999, 23, 359»364

ISSN:1144-0546    

DOI:10.1039/a900344d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Degradation of palmitic (hexadecanoic) acid deposited on self-cleaning glass : kinetics of disappearance, TiO2-coated intermediate products and degradation pathways Virginie Romeç as,a Pierre Pichat,*a Chantal Guillard,a Thierry Chopinb and Corinne Lehautb a URA au CNRS ììPhotocatalyse, Catalyse et Environnementœœ Ecole Centrale de L yon, BP 163, 69131 Ecully CEDEX, France. E-mail : Pierre.Pichat=ec-lyon.fr Centre de Recherches, 93308 Aubervilliers CEDEX, France b Rho� ne-Poulenc, Received (in Montpellier, France) 13th July 1998, Revised manuscript received 17th February 1999, Accepted 18th February 1999 Self-cleaning glass was prepared by depositing anatase nanoparticles as a transparent –lm onto glass previously coated by a barrier layer.A photoreactor was built to evaluate the efficiency.Palmitic acid, chosen as a compound representing stains from various sources and sprayed over these plates to form an uniform layer of ca. 580 nm thickness, was shown to disappear at a rate of ca. 60 nm h~1 under UV irradiation corresponding to the average solar radiant power at midlatitude. That clearly demonstrated that this glass is efficient enough, at least with this type of grease stain, to implement its use.Our main eÜorts explored the identi–cation of palmitic acid degradation products under these conditions by use of chromatographic methods, various solvents for recovering products on the glass, and various adsorbents (cartridges or SPME) for gas-phase analysis. The 39 products identi–ed revealed the gradual splitting of the palmitic acid chain yielding the whole series of linear aldehydes and carboxylic acids, some alkanes and two alcohols.In the closed photoreactor, complete mineralization was achieved. When the photoreactor atmosphere was renewed every hour, formaldehyde, acetaldehyde and acetone reached the highest concentrations in the gas phase. Mechanisms involving the initial attack of palmitic acid either by photogenerated holes or by hydroxyl radicals with subsequent formation of alkylperoxy radicals, hydroperoxides and tetroxides are discussed to account for the products. 1 Introduction UV-irradiated is able to destroy organic molecules TiO2 present on its surface, which represents a further potential application of heterogeneous photocatalysis.1 Therefore deposition of as a –lm on glass or ceramics panes has been TiO2 proposed for obtaining self-cleaning surfaces. For that purpose, methods to deposit a transparent, abrasion resistant and photocatalytically efficient thin –lm onto glass sur- TiO2 faces have been recently reported.2h4 Some processes have been patented.5,6 The Japanese –rm Toto has developed both anti-bacterial ceramic tiles and glass plates coated with In Fujishimaœs group8 and Hellerœs group9 glass plates TiO2 .7 have been prepared and shown to clean themselves from grease and tobacco stains or from coÜee stains, respectively.The objectives of our research were (i) to further assess the efficiency of transparent glass plates, coated by a thin TiO2- –lm, in making a sprayed layer dis- containing C15H31CO2H appear under UV-A irradiation ; and (ii) to identify and, if possible, quantify the intermediate products formed during the destruction of the palmitic acid layer in order both to complete the data concerning the self-cleaning properties of the glass plates and to give an insight into the photocatalytic degradation pathways.Palmitic acid was selected for several reasons.First, this compound is the most abundant fatty acid in human sebum10 and is thus present in –nger marks. A further clue of the presence of palmitic acid on articles handled manually was found in a study of chromatographic ìghost peaksœ11 which indicated that they are due to –nger fatty acids, especially palmitic acid, deposited on the syringe needle by the experimenter. Second, palmitic acid is emitted during the cooking of some foodstuÜs, particularly meat;12 it can therefore be regarded as a good representative of grease stains that can be found on the inner side of window panes.Third, it is also found in tobacco smoke particulate phase (10»155 lg per cigarette).13 Fourth, analyses of tropospheric solid particles in the Los Angeles area14 showed that palmitic acid is the most abundant compound (100»250 ng m~3) among the identi–ed organics (ca. 1000 ng m~3). Sources include the combustion of fossil fuels, wood and organic detritus ; particles containing nalkanoic acids, among which palmitic acid, also result from leaf surface waxes which are blown up by winds. Therefore, dry or wet deposition of these atmospheric particles onto the outer side of window panes can contribute to the presence of palmitic acid in the solid layers that blur the sight through the glass.Finally, as it contains only a saturated hydrocarbon linear chain and a group that are both quite stable, palmitic CO2H acid is a good candidate to test the capabilities of the selfcleaning glass plates. The photocatalytic degradation of octadecanoic (stearic) acid on thin –lms attached on glass has been studied by TiO2 Sitkiewitz and Heller.15 The formation of was monitored CO2 but no degradation intermediate products were searched for.These –lms were found to be eÜective in the complete TiO2 removal of 200 nm-thick layers of stearic acid. In this study, we have shown the efficiency of glass plates whose photocatalytic coating was prepared from nanoparticles synthesized by Rho� ne-Poulenc.We have identi–ed numerous products of the degradation of palmitic acid and also evaluated to what extent some of these products can be rel- New J. Chem., 1999, 23, 365»373 365eased in the atmosphere. To our knowledge, these essential aspects regarding the fate of an organic layer deposited on a coating submitted to solar light are explored for the –rst TiO2 time.Photocatalytic degradation pathways are tentatively proposed and discussed. 2 Experimental 2.1 Materials Palmitic acid (99%) was purchased from Sigma and its degradation intermediate products from Aldrich, Fluka or Merck. All these products were used as received. The glass coating was prepared from nanoparticles TiO2 ([250 m2 g~1) synthesized by Rho� ne-Poulenc and a photostable binder.The solution was deposited via the dip-coating method. The coated plates were then annealed at a temperature higher than 623 K. The mass ratio TiO2/photostable binder was 80/20. The layer was roughly 45 nm thick, had an anatase crystalline structure, and was hydrophilic. SEM photographs of the thin layers are shown in Fig. 1. TiO2 2.2 Palmitic acid deposition Palmitic acid was deposited (Fig. 2) by spraying (system : Jet Pack Spray Gun, Bioblock) a 8 g L~1 chloroformic solution of this compound. Approximately 1.5 mg of palmitic acid was sprayed on each plate, i.e. ca. 0.2 lmol cm~2 or 50 lg cm~2 which corresponded to a layer of ca. 580 nm thick as calculated from the density of palmitic acid at 335 K.Reproducibility in the deposited amount was evaluated to be ^15% under somewhat optimized conditions (nature of the solvent ; palmitic acid concentration ; distance between the gas container and the glass plate ; angle a, Fig. 2). These deposits exhibited a regular and uniform visual aspect ; no attempt was made to determine the uniformity of the layer thickness along Fig. 1 SEM photographs of nanoparticles deposited on glass, TiO2 (a) top view, (b) side view.Fig. 2 Illustrations showing the method by which palmitic acid is deposited on glass plates. the glass plate. The palmitic acid disappearance rates given in the results section correspond to a calculated mean value of the deposit thickness. 2.3 Photoreactor and light sources A photoreactor of ca. 2.2 L (Fig. 3) was designed and constructed to study the photodegradation of the thin palmitic acid layers covering the glass substrates. Five TiO2-coated glass plates (10 cm long, 3 cm wide and 4 mm thick) were vertically disposed at equal distance (ca. 35 mm) from the axis Fig. 3 Photoreactor used in this work. 366 New J. Chem., 1999,65»373of the cylindrical reactor. They were –xed in slits drilled in a Te—on disk, which was magnetically rotated (ca. 20 rpm) to ensure a regular mean irradiation of the glass plates. In some cases, the reactor was –lled with dioxygen (purity [99.998%, HP 48 Carboxyque, France) and closed before starting UV irradiation ; in other cases, a 50 mL min~1 dioxygen —ow was circulated through the reactor during the photocatalytic degradation.The relative humidity of the inner atmosphere was kept between 50 and 80% by adding a few drops of pure water in a hole drilled in the Te—on disk. Six Philips TL 29D/16 09N UV tubular lamps (Fig. 4) were vertically and regularly placed in a circle outside the reactor at about 2 mm from the external wall. They emitted in the 315»400 nm range with a maximum at 365 nm. A circulating water jacket kept the temperature between 291 and 295 K inside the reactor.The average radiant —ux reaching the glass plates, measured by use of a UDT 21 A powermeter, was found to be between 7 and 10 W m~2, i.e. of the same order of magnitude as the average solar radiant power on a horizontal plane at midlatitudes for wavelengths O380 nm.16,17 2.4 Analyses 2.4.1 Measurements of the palmitic acid amounts.To measure the amount of palmitic acid deposited on the glass plates, these plates were washed with 5 mL of chloroform containing 200 mg L~1 of dodecanoic acid (used as internal standard). The resulting solution was analyzed with a Varian 3400 gas chromatograph equipped with a —ame ionization detector and a CP Sil 5 CB column (25 m long, 0.32 mm i.d., 1.2 lm –lm thickness). 2.4.2 Detection of the organic intermediate products on the glass plates. In a –rst series of experiments, the TiO2-coated glass plates, initially covered with ca. 1.5 mg palmitic acid and UV irradiated for 3 h in the photoreactor under an —ow, O2 were immersed in 200 mL of a 2 g L~1 NaOH aqueous solution. After 15 min sonication (40 Hz), the solution was concentrated in a rotary evaporator down to 5 mL and then acidi–ed (pH\2) with phosphoric acid before being analyzed by HPLC.The HPLC apparatus comprised a LDC Constametric 3000 isocratic pump, a Waters 486 UV Detector set at 210 nm and a 20 lL injection loop. A Spherisorb ODS 2 column (25 cm long, 4.6 mm i.d., 5 lm particles diameter) was used. The mobile phase (—ow rate : 1 mL min~1) was deionized, doublydistilled water (Milli-Q system) containing 10 mmol L~1 phosphoric acid.Fig. 4 Spectral emission of the Philips TL 29 D 16/09 N lamp (according to the manufacturer). In another series of experiments, the glass TiO2-coated plates, also initially covered with ca. 1.5 mg palmitic acid, were irradiated for durations comprised between 30 min and 3 h in the closed photoreactor.After each irradiation duration, the side of each of the –ve plates was washed TiO2-coated with 5 mL (Pestinorm grade) which was not enough CH2Cl2 to collect all the organic matter remaining on the glass but enabled the subsequent GC-MS analysis to be carried out without further concentration. This solution was analyzed with a HP 5890 series II gas chromatograph equipped with a CP Sil 5 CB column (25 m long, 0.25 mm i.d., 1.2 lm –lm thickness) ; the temperature was programmed from 348 to 513 K at 15 K min~1, hold time: 13 min.Detection and identi–cation were performed with a HP 5971A MS detector using the electron ionization mode. The major organic compounds in the sample were initially identi–ed by computer matches to standard reference mass fragmentograms in the NBS 49K library and the identi–cations were checked later by comparicson with standards.In both cases, control analyses were performed in which the same procedures were used but without irradiating the glass plates. 2.4.3 Detection of the organic intermediate products in the gas phase. Case of the photoreactor through which a 50 mL min~1 —ow was circulated.An ORBO 32 (Supelco) activat- O2 ed charcoal cartridge was placed in the gaseous output of the photoreactor during the irradiation. At the end of the irradiation period, the activated charcoal was transferred into a vial containing 1 mL Pestinorm grade dichloromethane and manually shook during 1 min. No –ltration was carried out in order to avoid the loss of analytes. This solution was analyzed with the GC-MS apparatus equipped with the same CP Sil 5 CB column; the column temperature was programmed from 311 to 403 K at 3 K min~1 and from 403 to 493 K at 20 K min~1.To detect carbonyl compounds, two Supelco Lp DNPH10 cartridges were connected in series in the gaseous output of the photoreactor during the irradiation. These cartridges contained silica coated with 2,4-dinitrophenylhydrazine (2,4- DNPH) which converts carbonyl compounds into their 2,4- dinitrophenylhydrazone derivatives.Only one glass plate coated with a low amount of palmitic acid was irradiated in this case in order to avoid sample breakthrough. At the end of the irradiation, the cartridges were each eluted with 4 mL HPLC grade acetonitrile. The resulting solution was then analyzed by reversed phase HPLC under the following conditions : solvent delivery system: Varian 9010; column: Spherisorb ODS2 (25 cm]4.6 mm i.d., 5 lm particles) ; mobile phase: A: acetonitrile (30%)» tetrahydrofuran (10%)»water (60%), B: acetonitrile (60%)» water (40%), 100% A for 1 min, linear gradient to 100% B over 10 min, hold time 30 min; —ow rate : 1.5 mL min~1; injected volume: 20 lL; detector : UV Varian 9065 Polychrom set at j\360 nm.A standard mixture of derivatized carbonyl compounds (Supelco TO11/IP6A Carbonyl DNPH Mix) was injected under the same analytical conditions in order to assign the chromatographic peaks and to estimate the amounts of acetone and aldehydes collected in the cartridges. Case of the closed photoreactor –lled with Direct O2 .analysis of the enclosed atmosphere was carried out with an Intersmat IGC 120FL gas chromatograph equipped with a —ame ionization detector and a Vocol column (105 m long, 0.53 mm i.d., 3 lm –lm thickness) heated at 313 K. Identi–cations were con–rmed by the use of the HP GC-MS apparatus equipped with a CP Wax 57 CB column (50 m long, 0.25 mm i.d., 1.2 lm –lm thickness) at 303 K, and operated in the electron ionization mode.Solid phase micro-extraction (SPME) was used to adsorb the organic products contained in the enclosed atmosphere of New J. Chem., 1999, 23, 365»373 367the photoreactor in order to detect analytes at lower concentrations. This technique18,19 is an alternative to other concentration methods. It consists of two processes : partitioning of analytes between the silica –ber coating and the sample and desorption of concentrated analytes into an analytical instrument.Before each use, the –bers, purchased from Supelco, were thermally desorbed according to the supplierœs instructions. The –ber holder assembly containing the retracted SPME –ber was introduced into the reactor through the septum.The –ber was exposed to the gaseous phase for various durations with the UV lamps on or oÜ (see Results section). The –ber was then retracted into the holding assembly which was removed from the reactor and transferred to the injection port, set at 493 K, of the HP GC-MS apparatus equipped with a SPME inlet liner. The temperature of the same CP Sil 5 CB column was programmed from 313 K (hold time: 5 min) to 493 K at 5 K min~1, and the electron ionization mode was used.measurements. Carbon dioxide was analyzed in 2.4.4 CO2 the closed photoreactor by use of a catharometer gaschromatograph (Intersmat, model IGC 20MB), equipped with a Porapak Q column (80»100 mesh, 3 m]1/4 in). In the case of the photoreactor through which a 50 mL min~1 —ow was circulated, the amount of was too O2 CO2 low to be measured in the exiting gas —ow by gas chromatography.Consequently, we employed a chemical method: the gas —ow was bubbled into three successive bottles containing each 200 mL of barium dihydroxide solution (1 mol L~1). Carbon dioxide was precipitated as barium carbonate. The excess barium dihydroxide was titrated with potassium hydrogenphthalate (1.5 mol L~1).However, as volatile carboxylic acids produced by palmitic acid degradation also react with we thought it Ba(OH)2 , useful to also employ the following sequential procedure which allows one to measure, by gas chromatography, the amounts produced. Palmitic acid-coated plates were CO2 placed into the photoreactor; the photoreactor was –lled with dioxygen and closed ; the plates were UV-irradiated during 1 h; concentration in the photoreactor was then measured CO2 by GC; the photoreactor was –lled again with dioxygen; the absence of was checked.One-hour irradiation sequences CO2 were repeated until was no longer emitted, i.e. the degra- CO2 dation of palmitic acid and its organic products were complete. The total amount of corresponded to the amount CO2 of carbon dioxide produced in a periodically renewed atmosphere during the photocatalytic degradation of palmitic acid.Note that if a 50 mL min~1 dioxygen —ow was circulated through the reactor, the inner atmosphere was renewed about every 45 min. Therefore the sequential procedure we used can be considered as a satisfactory approximation to the situation in which dioxygen is permanently circulated. 3 Results 3.1 Photocatalytic disappearance of palmitic acid When palmitic acid was deposited onto the non-TiO2-coated side of the glass plates and UV irradiated under the conditions described above in the closed photoreactor, no decrease in the palmitic acid amount was observed after 15 h irradiation. Disappearance of palmitic acid sprayed onto the TiO2- side is shown in Fig. 5. Under our conditions, an coated apparent zero-order decay was observed with a rate constant k\0.60^0.12 lmol h~1, which means that an approximately 60 nm thick layer of palmitic acid was destroyed per hour. That removal was accompanied by a gradual improvement in the transparency of the glass plates according to visual estimations. Fig. 5 Photocatalytic disappearance of palmitic acid deposited on the self-cleaning glass plate ; the two experiments give an idea of the reproducibility. 3.2 Identi–cation of degradation intermediate products Determining the nature and amount of the photooxidation products generated is required (i) to know whether this cleaning process may release into the atmosphere organic compounds which can be toxic or noxious, and (ii) to give an insight into the photocatalytic degradation pathways. 3.2.1 Intermediate products detected on the glass plates. A sodium hydroxide aqueous solution or dichloromethane were used to dissolve the products remaining on the glass plates. HPLC analysis of the NaOH solution indicated the presence of formic and acetic acids, whereas GC-MS analysis of the solution showed the presence of the –ve linear car- CH2Cl2 boxylic acids containing 8 to 12 carbon atoms.The use of propan-1-ol instead of dichloromethane allowed the detection of tetradecanoic acid. Identi–cations were con–rmed by use of standards. 3.2.2 Intermediate products detected in the gas phase. The various analytical techniques and procedures we employed allowed us to detect 39 products of palmitic acid degradation (Table 1) as is detailed below.Under our laboratory conditions, experiments carried out in the closed photoreactor have had the advantage of allowing the identi–cation of numerous intermediate products whose detection was impossible otherwise. In the real world, the atmosphere in the vicinity of a window glass plate is constantly changed whether we consider the outdoor or the indoor side, since renewing the atmosphere every hour is recommended in houses or buildings.20 Consequently, we also thought it was necessary to determine the amount of volatile Table 1 Organic intermediate products (all linear) detected on the glass plates (P) or in the gas phase (G) during the photodegradation of palmitic acid deposited on self-cleaning glass plates Primary Alkanes alcohols Aldehydes Ketones Acids C1 G / P, G C2 G / P, G C3 G G G C4 G G C5 G G C6 G G C7 G G G G C8 G G G G C9 G G P, G C10 G G P, G C11 G G P, G C12 G G P, G C13G G G C14G G P C15 G 368 New J.Chem., 1999, 23, 365»373organic compounds released during palmitic acid degradation under the conditions where —ow through the photoreactor O2 was at a velocity corresponding to this renewal rate.Acetone, pentanal, hexanal, heptane and octane produced during the photodegradation of palmitic acid sprayed on the glass plates were trapped into cartridges containing activated carbon and identi–ed by GC-MS analysis. Adsorption/reaction of carbonyl products in cartridges containing 2,4-DNPH-coated allowed the detection of SiO2 acetone and linear aldehydes.Only trace amounts were C1»C6 adsorbed in the second cartridge (see Experimental section), which showed that the levels of carbonyl products in the –rst cartridge did not reach the saturation limit and therefore that the expected carbonyl products were quantitatively collected in the –rst cartridge. The estimated amounts formed from 0.57^0.08 lmol palmitic acid are indicated in Table 2.Formaldehyde and acetaldehyde were formed in higher amounts than the other linear aldehydes, in particular linear C4»C6 aldehydes were found in trace amounts. The sum of the amounts of acetone and linear aldehydes emitted C1»C3 under these conditions represented 22^4% of the amount of organic carbon initially present in the photoreactor. Note that acrolein (propenal) and crotonaldehyde (buten-2-al) would have been detected if they had been produced in amounts [0.5 lg from 0.15 mg palmitic acid.The amounts of acetone and acetaldehyde were sufficient to monitor their concentration by GC-FID (Fig. 6) in the enclosed atmosphere of the photoreactor (note that formaldehyde cannot be detected by FID). The maximum detected amounts of acetone (ca. 4.5 lmol) and acetaldehyde (ca. 7 lmol) were obtained when ca. 70% of palmitic acid was eliminated (Fig. 6). The photoreactor initially contained about 5]1.5 mg of palmitic acid, i.e. ca. 30 lmol. The lower ratio of the amount of acetaldehyde to that of actone in the closed photoreactor than when oxygen was circulated (Table 2) is consistent with the better stability of acetone toward oxidation.Analysis of the enclosed atmosphere carried out with SPME-GC-MS led to the detection and the identi–cation of numerous volatile intermediate products. Every linear carbox- Table 2 Estimated amounts of acetone and linear aldehydes C1»C3 released from 0.57^0.08 lmol of palmitic acid under a 50 mL min~1 —ow O2 Product Released amount/lmol Acetone 0.13^0.01 Formaldehyde 0.34^0.05 Acetaldehyde 0.59^0.09 Propanal B0.05 Fig. 6 Temporal variations in the amounts of acetone and acetaldehyde in the closed photoreactor during palmitic acid (initial amount: 30 lmol) degradation. ylic acid containing from 1 to 12 carbon atoms was detected by use of a 85 lm polyacrylate –ber. A 65 lm polydimethylsiloxane»divinylbenzene (PDMS»DVB) –ber enabled the detection of 29 products (Fig. 7) : four linear alkanes (from two linear primary alcohols and C12»C15), (C7 acetone, ten linear aldehydes (from to and twelve C8), C5 C14) linear carboxylic acids and to (C1, C2 C4 C13). A control experiment was carried out with palmitic acid deposited onto the side. Analysis of the non-TiO2-coated enclosed atmosphere with the PDMS»DVB –ber showed that some unidenti–ed peaks in the chromatogram obtained under the usual operating conditions were in fact due to compounds released from the –ber under UV irradiation ; but the aforementioned products were not found in this control experiment.To ensure that all the identi–ed products were not generated by interactions between the –ber phase and some analytes when irradiating during 15 h, another experiment was carried out.Instead of leaving the –ber in the photoreactor during the whole photodegradation, the –ber was exposed to the enclosed atmosphere of the photoreactor during only 15 min, in the dark, after having irradiated the glass plates for various durations. Under these conditions, the previously identi–ed products were also detected, which con–rmed they were intermediate products of palmitic acid degradation.However, the chromatographic peaks were less numerous and less intense (compare Fig. 8 and 7). On one hand, C8»C13 acids and aldehydes were not detected possibly C9»C14 because it took a longer time for these less volatile compounds to equilibrate with the –ber coating. On the other hand butanal, nonane, decane and undecane were detected ; these products are, within their chemical category, more volatile than those observed when the –ber was continuously exposed to the photoreactor atmosphere.These short exposure-time SPME-analyses brought some information on the chronological order of formation of the intermediate products. For example, butanoic acid was detected in the chromatogram obtained after 1 h irradiation, whereas propanoic acid and acetic acid were detected in the chromatograms obtained after 2.5 and 5.5 h irradiation, respectively.Among the carboxylic acids, the shorter ones were not surprisingly emitted later than the longer ones, which revealed the gradual splitting of the palmitic acid chain (e.g. compare the chromatograms shown in Fig. 8). Fig. 7 GC chromatogram, obtained after 15 h adsorption on a polydimethylsiloxane»divinylbenzene 65 lm –ber, of the products emitted into the gas phase during palmitic acid degradation.Identi–- cation performed by mass spectrometry. Irradiation duration: 15 h. 1, acetone; 2, formic acid ; 3, acetic acid ; 4, pentanal; 5, hexanal; 6, butanoic acid ; 7, heptanal ; 8, pentanoic acid ; 9, heptan-1-ol ; 10, hexanoic acid ; 11, octanal ; 12, octan-1-ol ; 13, heptanoic acid ; 14, nonanal; 15, octanoic acid ; 16, decanal; 17, dodecane; 18, nonanoic acid ; 19, undecanal ; 20, tridecane ; 21, decanoic acid ; 22, dodecanal; 23, tetradecane ; 24, undecanoic acid ; 25, tridecanal ; 26, pentadecane; 27, dodecanoic acid ; 28, tetradecanal ; 29, tridecanoic acid.New J. Chem., 1999, 23, 365»373 369Fig. 8 GC chromatogram, obtained after 15 min adsorption on a PDMS»DVB 65 lm –ber, of the products emitted into the gas phase during palmitic acid degradation. Product identi–cation by mass spectrometry. 1, butanal; 2, acetic acid ; 3, pentanal; 4, propanoic acid ; 5, hexanal; 6, butanoic acid ; 7, heptanal; 8, pentanoic acid ; 9, nonane; 10, hexanoic acid ; 11, octanal ; 12, decane; 13, heptanoic acid ; 14, nonanal; 15, undecane; 16, decanal; 17, dodecane; 18, tridecane ; 19, tetradecane ; 20, pentadecane; B, this peak was present in the blank analysis.None of the intermediate products were found in two types of control experiments, i.e. without UV irradiation or when palmitic acid was deposited on the side of non-TiO2-coated the plates. 3.3 Mass balance Quanti–cation in the enclosed atmosphere showed that, after a regular increase within at least the –rst 10 h of irradiation, concentration reached a plateau after ca. 22 h (Fig. 9). CO2 The amount of palmitic acid initially introduced was 29^4 lmol, which potentially corresponded to 0.47^0.07 mmol whereas 0.38^0.04 mmol were found to be CO2, CO2 emitted according to our measurements. Considering these accuracies, the mineralization of palmitic acid can be regarded as being complete when the photocatalytic degradation took place in the closed photoreactor.When was quanti–ed in the 50 mL min~1 —ow CO2 O2 coming out from the reactor after trapping in barium dihy- Fig. 9 Production of during the photocatalytic degradation of CO2 palmitic acid in the closed photoreactor. Horizontal lines : maximum and minimum (considering the accuracy in the amount of deposited palmitic acid) of expected from complete mineralization. CO2 droxide solutions, the conversion rate from palmitic acid into was estimated to be 80^15%.CO2 When was quanti–ed in the periodically renewed CO2 atmosphere, because of an increase in the total uncertainty (ca. 25%) resulting from the addition of amounts mea- CO2 sured separately, the conversion of palmitic acid into was CO2 evaluated to be comprised between 72 and 100%.Note that, under these conditions, volatile organic products were eliminated when the atmosphere was renewed and thereby could not be mineralized. The two results obtained for the conversion of palmitic acid into are consistent. We can conclude that the conversion CO2 into lies between 72 and 95% if we consider the overlap CO2 between the ranges evaluated by the two types of measurements.We also tried to evaluate the total amount of the products that were detected with the PDMS»DVB SPME –ber. For that purpose, a solution in dichloromethane of linear C9»C15 alkanes, aldehydes (which were the aldehydes not C7»C14 quanti–ed by use of the 2,4-DNPH-coated silica cartridges, cf.Table 2) and acids was analyzed under the same con- C4»C13 ditions as those used to analyze the products collected on the –ber. From this analysis we deduced that the amounts, collected on the –ber, of all these products corresponded to only ca. 0.01% of the organic carbon initially present in the photoreactor. Even if we assume that only a small percentage of these volatile organics (say 1%, to be very pessimistic) were collected on the –ber, we can infer that less than 1% of the palmitic acid carbon was emitted as linear alkanes, C9»C15 aldehydes and acids.C7»C14 C4»C13 In short, under the conditions corresponding to the renewal of the photoreactor atmosphere every hour, the mass balance from palmitic acid photocatalytic degradation on the TiO2- glass plates is the following (expressed in percentages coated of the palmitic acid carbon): 72»95%; (a) CO2 : (b) acetone, acetaldehyde and formaldehyde: 22^4%; (c) linear alkanes aldehydes and acids (C9»C15), (C7»C14) maximum 1%.(C4»C13) : Amounts of volatile carboxylic acids and alkanes were not estimated. However, the sum of the amounts of acetone CO2 , and linear aldehydes represented between 91 and 100% C1»C6 of the carbon amount initially contained in palmitic acid.This means the amounts of products which were not detected or quanti–ed corresponded to \9% of the carbon contained in palmitic acid. 4 Discussion 4.1 Palmitic acid elimination. Glass self-cleaning efficiency The zero-order degradation kinetics for palmitic acid disappearance was consistent with the constant saturation of the surface by the palmitic acid molecules which were TiO2 sprayed on the top layer of nanoparticles (Fig. 1).TiO2 We have shown that, despite the fact that the self-cleaning glass was not optimized and despite the chemical stability of palmitic acid, layers of several tens of nm of this compound could be destroyed under solar-like irradiation. This corresponded to a quantum yield of the order of 0.002 (with respect to UV-A photons incident on the plates) as usual in photocatalysis.Palmitic acid removal was accompanied by a gradual improvement in the transparency of the glass plates according to visual estimations (Fig. 10). This result illustrates the efficiency of the glass to clean itself from a TiO2-coated compound representative of grease stains.This efficiency was further demonstrated by the mass balance. In the closed photoreactor, palmitic acid could be totally mineralized within experimental accuracy. When an oxygen —ow was circulated through the photoreactor to simulate more realistic conditions, palmitic acid and its interme- 370 New J.Chem., 1999, 23, 365»373Fig. 10 Visual aspect of glass plates initially covered with ca. 6 mg palmitic acid, after increasing irradiation durations (from left to right : 0, 8, 16, 24, 32, 40 h UV irradiation in the photoreactor). diate products of higher molecular weight were transformed principally into and also into a few VOCs (Table 2) CO2 swept out from the reactor and corresponding to about 22% of palmitic acid carbon under our conditions. 4.2 In—uence on indoor air quality Because of the huge dilution capacity of the troposphere, VOCs formed from palmitic acid and related compounds contained in grease stains on glass panes, will obviously not aÜect the outdoor air quality. However, the question may be relevant for indoor air. Assuming that the amounts of acetone, formaldehyde and acetaldehyde listed in Table 2 are suddenly released in a 24 m3-room at ca. 293 K , the concentrations in ppbv of these products would be 0.13 (acetone), 0.34 (formaldehyde) and 0.59 (acetaldehyde). These values are 35»125 times lower than the average concentrations in ppbv reported for indoor air in buildings : for instance 8.6 (acetone),21 42 (formaldehyde)21 and 21 (acetaldehyde).22 In other words, the perturbation in air quality would become signi–cant only for a quantity of palmitic acid about 35 times greater than that corresponding to Table 2 (ca. 0.57 lmol). This latter quantity represented a palmitic acid mean thickness of ca. 60 nm on a 30 cm2-plate and gave the visual impression of a quite dirty glass. That means that the air quality might be signi–cantly altered if a 60 nm-thick layer of palmitic acid covered in the order of 35]30 cm2\0.105 m2 of glass pane, i.e. ca. 7% of the window panes (ca. 1.5 m2) in a 24 m3-room. Unfortunately, to our knowledge, data about the amounts of organic material present on indoor glass surfaces are lacking. Accordingly, we cannot indicate whether the dirt level we calculated might be reached.In general, window panes gradually become dirty, and moreover the cleaning process of the glasses being continuous, the resulting oxi- TiO2-coated dized VOCs will be released progressively and not suddenly, as assumed in our calculations, so that these calculations correspond to upper limit concentrations. In addition, the clean part of the glass panes TiO2-coated can adsorb the VOCs and mineralize them as illustrated by our experiments in the closed photoreactor.However, on the other hand, adsorption on the glass of hydrocar- TiO2-coated bons contained in indoor air and subsequent photocatalytic transformations of these substances will both decrease the hydrocarbon concentrations and contribute to increase the concentrations in oxygenated VOCs produced from the hydrocarbons if the formation rate of these oxidized products dominated their mineralization rate.This is another debate. The only purpose for the comparison between the amounts of carbonyl products formed from palmitic acid and the concentrations in these products usually present in indoor air was to provide some data illustrating under what conditions an unfavorable in—uence of the self-cleaning glasses upon air quality might appear. However, in the absence of knowledge about dirt levels formed on glass panes from both continuous and irregular sources, it is impossible to tell whether these conditions might be met. 4.3 Remarks on the detection of intermediate products As is detailed in the Results section, numerous intermediate products of the photodegradation of palmitic acid on the glass plates have been identi–ed (Table 1).TiO2-coated However, it would be naive to claim that the list of these products is complete. Indeed, detection of analytes is strongly dependant on the analytical technique. For instance, the whole range of linear carboxylic acids were detected C1»C14 in the atmosphere of the photoreactor.Some of these carboxylic acids were detected on the glass plates as well : formic and acetic acids dissolved in a NaOH solution, acids dis- C8»C12 solved in and tetradecanoic acid dissolved in CH2Cl2 , propan-1-ol. It is obvious that the carboxylic acids containing from three to seven carbon atoms were also present on the glass plates as a result of an equilibrium between the adsorbed/deposited phase and the gas phase for each product.However, only the products that were present in a sufficient amount on the plates and that were soluble enough in the solvents we used were detected on the plates. Also, when a plate was taken out of the photoreactor, the equilibrium between the adsorbed and the gaseous phase was modi–ed and only strongly adsorbed and/or weakly volatile products remained on the plates.Therefore we consider that palmitic acid yielded the whole series of linear aldehydes, acids and alkanes (no C15»C1 attempt was made to detect alkanes because they were C6»C1 quantitatively negligible given the mass balance). We also think that primary alcohols were formed in very low amounts, so that only heptan-1-ol and octan-1-ol could be detected. 4.4 Suggested degradation pathways Analyzing the intermediate products at diÜerent stages of the degradation showed the gradual splitting of palmitic acid. Degradation pathways are considered hereafter to explain the formation of these intermediate products. Reaction of carboxylic acids with a photogenerated hole leads to decarboxylation.23h25 This reaction could be favored by the interaction of these acids with the surface via the TiO2 carboxyl group [eqn.(1), for palmitic acid]. R\CH3(CH2)14 RCH2CO2H]h`]RCH2~]CO2]H` (1) Easy addition of dioxygen to the alkyl radical would yield the corresponding alkylperoxyl radical [eqn. (2)] RCH2~]O2 ]RCH2OO~ (2) According to literature, can (i) abstract an H atom RCH2OO~ from a molecule26 e.g. palmitic acid (or possibly from another molecule depending on the stage the cleaning process has reached) to produce a hydroperoxide and (ii) RCH2OOH, combine either with an identical or non-identical alkylperoxyl radical27h30 or with a hydroperoxy radical formed either from conduction band electrons, dioxygen and protons [eqn.(3) or via eqn. (7) (below)] e~]O2 ]O2~~ O2~~]H`HHO2~ (3) Tetroxides or are (RCH2)2O4 , RCH2»O4»CH2R@ RCH2O4H believed to be generated in that latter case.31,32 However, in the condensed phase the production of hydroperoxides should predominate over radical recombination as a result of the high pool of nontransformed molecules, provided they contain labile H atoms.Hydroperoxides are photochemically stable in the spectral range used, so they can decompose only thermally [eqn.(4) and (5)].26,29 RCH2OOH]RCH2O~]OH~ (4) 2RCH2OOH]RCH2O~]RCH2OO~]H2O (5) New J. Chem., 1999, 23, 365»373 371The resulting alkoxy radical could undergo several reactions [eqn. (6)»(9)].26,29,30 RCH2O~]R@H]RCH2OH]R@~ (6) RCH2O~]O2 ]RCHO]HO2~ (7) RCH2O~]HCHO]R~ (8) 2RCH2O~]RCH2O»OCH2R (9) The peroxides supposed to be produced via the latter reaction could not be detected under our analytical conditions. The same types of products could be formed from the tetroxide either directly [eqn.(10)»(12)]32 or (RCH2)2O4 , through the radical generated by reaction (13). RCH2O~ (RCH2)2O4 ]RCHO]RCH2OH]O2 (10) (RCH2)2O4 ]2RCHO]H2O2 (11) (RCH2)2O4 ]2HCHO]2R~]O2 (12) (RCH2)2O4 ]2RCH2O~]O2 (13) In the case of the tetroxide reactions equivalent RCH2O4H, to reactions (10) and (13) could occur [eqn.(14) and (15)]. RCH2O4H]RCHO]H2O]O2 (14) RCH2O4H]RCH2O~]OH~]O2 (15) Eqn. (4)»(15) show that aldehydes, in particular formaldehyde, and primary alcohols can be formed from the alkylperoxy radicals obtained through reaction (2). These pathways assume the formation of either hydroperoxides and/or tetroxides and the subsequent thermal decomposition of these species.Aldehydes can easily be oxidized to the corresponding acids. Alkyl radicals [eqn. (6), (8) and (12)] can recombine to yield alkanes. Therefore these pathways allows one to account for all the products of palmitic acid degradation we detected except acetone. Concerning the formation of hydroperoxides, palmitic acid, at least initially, should be the main source of H atoms in the reaction [eqn.(16)]. RCH2O2~]R@H]RCH2OOH]R@~ (16) Considering the bond strengths, H abstraction should involve one of the groups in palmitic acid, and the resulting alkyl CH2 radicals would lead to carbonylated acids, and subsequently diacids, through pathways equivalent to those presented above. Carbonylated acids and diacids were not detected at least under our analytical conditions.We therefore hypothesize that the carboxylic H atom could be predominantly abstracted possibly because of its proximity to the TiO2 surface and because of a weakening of the COO»H bond strength due to adsorption. The resulting radical would readily decarboxylate [eqn. (17)] R@~\RCH2COO~]RCH2~]CO2 (17) giving an alkyl radical equivalent to that obtained by reaction of a photogenerated hole with palmitic acid or with the acids subsequently produced [eqn.(1)]. The same hypothesis can be evoked regarding the role of hydroxyl radicals. These radicals can be formed during the photocatalytic process when a photogenerated hole is trapped by a water molecule. h`]H2O]OH~]H` (18) They could also be produced from hydroperoxides [eqn.(4)] and tetroxides [eqn. (15)]. If they abstracted H RCH2O4H atoms from the groups in palmitic acid, carbonylated CH2 acids and diacids would be expected. Because these products were not observed, we again suggest that H abstraction mainly occurred at the group as in the case where the CO2H H abstracting radical is [eqn. (19)] RCH2OO~ RCH2CO2H]OH~]RCH2COO~]H2O (19) which would subsequently give rise to radical RCH2~ [eqn.(17)].Consequently, product analysis does not allow one to discriminate between an initial attack of palmitic acid by a hole [eqn. (1)] or a hydroxyl radical [eqn. (19)]. In addition, and radicals can abstract H atoms from the OH~ RCH2OO~ aldehydic groups and less easily from the groups in CH2 alkanes. Formation of acetone (Figs. 6 and 7, Tables 1 and 2) is not directly accounted for by the reactions suggested above for the fate of alkylperoxy radicals. If acetone was formed only via the combination of an acetyl radical, derived from acetaldehyde, and a methyl radical, other ketones would be reasonably expected to be produced from equivalent radicals with a higher number of C atoms; no ketone but acetone was found though our analytical methods were appropriate for detecting higher ketones.This difficulty in explaining the formation of acetone indicates that the mechanisms suggested are not comprehensive even if they account for the formation of all the other products. 5 Conclusion Our results show that layers of some tens of nm of fatty acids (i.e. a few tens of nmol cm~2), such as palmitic acid, can be removed within hours under solar-like irradiation of the anatase-coated glass plates we used.As analyses carried out in our laboratory have indicated that a thumb print on glass contains between 0.5 and 1.5 nmol cm~2 of palmitic acid, selfcleaning from this type of stain will indeed be very rapid even on a cloudy day. Furthermore, progress is expected in the efficiency of the which was not optimized.TiO2-coating Two main conclusions can be drawn from our analyses of the products of palmitic acid. From the practical viewpoint, our study showed that more than 72% of palmitic acid was converted into when the photoreactor atmosphere was CO2 , renewed every hour. The majority of the organic products identi–ed were present at negligible concentrations in the gas phase.Only formaldehyde, acetaldehyde and acetone were found at higher levels ; ca. 22% of palmitic acid carbon was converted into these three products. From the fundamental viewpoint, the absence of rami–ed compounds and of carbonylated or hydroxylated acids among the products has led us to suggest that the reacting site of palmitic acid (and other acids formed from it) is the carboxylic group.To complete these data, an on-going study addresses the case of a still more stable compound which represents organic deposits emanating from other sources. Acknowledgements R. is grateful to the CNRS and to Rho� ne-Poulenc for her V. Ph.D. scholarship. We thank Rho� ne-Poulenc for partial –nancial support and Saint-Gobain for preparing the TiO2-coated glass plates.References 1 See for example: P. Pichat, in Handbook of Heterogeneous Catalysis, ed. G. Ertl, H. Knoé zinger and J. Weitkamp, Wiley-VCH, Weinheim, 1997, vol. 4, pp. 2111»2122. 2 H. Cui, H. S. Shen, Y. M. Gao, K. Dwight and A. Wold, Mater. Res. Bull., 1993, 28, 195. 3 Y. Paz, Z. Luo, L. Rabenberg and A. Heller, J. Mater. Res., 1995, 10, 2842. 4 N.Neigishi, T. Iyoda, K. Hashimoto and A. Fujishima, Chem. L ett., 1995, 841. 5 A. Fujishima, K. Hashimoto, T. Iyoda and S. Fukayama, Eur. Pat. Appl., 0 737 513 A1, 1996. 6 Saint-Gobain Vitrage, Fr. Pat. 95 10839, published in BOPI no. 12, March 1997. 372 New J. Chem., 1999, 23, 365»3737 Super-hydrophilic photocatalyst and its applications, TOTO Ltd; Photocatalyst project team, Toto web site : http ://www.toto.co.jp/ Hydro-E/hydro1-E.htm 8 A.Fujishima, L ook Jpn., 1995, 41, 471. 9 A. Heller, Acc. Chem. Res., 1995, 28, 503. 10 P. Boulanger, F. Tayeau, P. Mandel and G. Biserte, Biochimie I, Masson, Paris, 1962, p. 534. Meç dicale 11 S. B. Hawthorne, D. J. Miller, J. Pawliszyn and C. L. Arthur, J. Chromatogr., 1992, 603, 185. 12 W. F. Rogge, L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit, Environ. Sci. T echnol., 1991, 25, 1112. 13 M. R. Guerin, R. A. Jenkins and B. A. Tomkins, T he Chemistry of Environmental T obacco Smoke, Composition and Measurements, L. Clar, Chelsea, MI, 1992, pp. 43»61. 14 W. F. Rogge, M. A. Mazurek, L. M. Hildemann, G. R. Cass and B. R. T. Simoneit, Atmos. Environ., 1993, 27, 1309. 15 S. Sitkiewitz and A. Heller, New J. Chem., 1996, 20, 233. 16 L. R. Koller, Ultraviolet Radiation, Wiley & Sons, New York, 1965, ch. 4. 17 J.-M. Herrmann, J. Disdier, P. Pichat, S. Malato and J. Blanco, Appl. Catal. B, 1998, 17, 15. 18 Z. Zhang, M. J. Yang and J. Pawliszyn, Anal. Chem., 1994, 66, 844. 19 J. Pawliszyn, Solid Phase Microextraction, T heory and Practice, Wiley-VCH, New York, 1997. 20 L. Molhave, B. Bach and O. F. Pedersen, Environ. Int., 1986, 12, 167. 21 J. J. Shah and H. B. Singh, Environ. Sci. T echnol., 1988, 22, 1381. 22 A. H. Miguel, F. R. de Aquino Neto, J. N. Cardoso, P. de C. Vasconcellos, A. S. Pereira and K. S. G. Marquez, Environ. Sci. T echnol., 1995, 29, 338. 23 B. Kraeutler and A. J. Bard, J. Am. Chem. Soc., 1977, 99, 7729. 24 H. L. Chum, M. RatcliÜ, F. L. Posey, J. A. Turner and A. J. Nozik, J. Phys. Chem., 1983, 87, 3089. 25 Y. Mao, C. Schoé neich and K. D. Asmus, J. Phys. Chem., 1991, 95, 10080. 26 R. A. Larson and E. J. Weber, Reactions Mechanisms in Environmental Organic Chemistry, Lewis, Boca Raton, FL, 1994, ch. 4. 27 A. F. Marchaj, D. G. Kelley, A. Bakac and J. H. Espenson, J. Phys. Chem., 1991, 95, 4441. 28 A. F. Wagner, I. R. Slagle, D. Sarzynski and D. Gutman, J. Phys. Chem., 1990, 94, 1853. 29 P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D. Hayman, M. E. Jenkin, G. K. Moortgat and F. Zabel, Organic Peroxy Radicals, Commission of the European Communities, Brussels, 1993. 30 R. Atkinson, Int. J. Chem. Kinet., 1997, 29, 99. 31 J. E. Bennett and J. A. Howard, J. Am. Chem. Soc., 1973, 95, 4008. 32 C. von Sonntag and H.-P. Schuchmann, Angew. Chem., Int. Ed. Engl., 1991, 30, 1229. Paper 9/01342C New J. Chem., 1999, 23, 365»373

ISSN:1144-0546    

DOI:10.1039/a901342c

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Bi-ionic system: theoretical investigation on the ionic —uxes through an ion-exchange membrane Sami Mokrani, Lasa� ad Dammak, Christian Larchet and Bernard Auclair* L aboratoire des Echangeurs dœIons, de Paris-V al de Marne, 94010 Mateç riaux Universiteç Creç teil, France. E-mail : lmei=univ-paris12.fr Received (in Montpellier, France) 21st December 1998, Accepted 12th February 1999 We have established a theoretical description of the co-ion and counter-ion —uxes through an ion-exchange membrane (IEM) used in a bi-ionic system.This study is based on the modi–ed Nernst»Planck equation applied to three parallel homogenous phases: two diÜusion boundary layers (DBLs) and an IEM. The role of the IEM and of the DBLs in the control of the interdiÜusion process is discussed. The in—uence of the common concentration, the co-ion and counter-ion diÜusion coefficients in the membrane, the membrane selectivity and affinity, the concentration of functional sites and the convection velocity are evaluated.Syste` me bi-onique : e ç tude theç orique des —ux ioniques a ` travers une membrane e ç changeuse dœions. Nous avons eç tabli une eç tude theç orique des —ux du co-ion et des contre-ions a` travers une membrane eç changeuse dœions (MEI) utiliseç e dans un syste` me bi-ionique.Cette eç tude est baseç e sur lœeç quation modi–eç e de Nernst»Planck appliqueç e a` trois phases paralle` les : deux couches limites de diÜusion (CLE) et une MEI. La part de la MEI et des CLE dans le contro� le du processus dœinterdiÜusion est discuteç e.Lœin—uence de la concentration commune, des coefficients de diÜusion du co-ion et des contre-ions dans la membrane, de la seç lectiviteç et de lœaffiniteç membranaires, de la concentration des sites fonctionnels et de la vitesse de convection sont eç valueç s. A bi-ionic system (BIS) is constituted of an ion-exchange membrane (IEM) separating two electrolyte solutions having the same co-ion, Y, at the same concentration but diÜer- C0 , ent counter-ions : A and B.It will be denoted as AY (C0) In this system, we observe an interdif- /IEM/BY(C0 ). fusion process (IDP) between the two counter-ions, under the in—uence of their concentration gradients. Quickly, the system will become multi-ionic because the two counter-ions will exist in both compartments.In most papers dealing with biionic systems,1h9 attention has been focused on potential measurements and/or computations. But the analysis of transmembrane ionic —ux has been neglected because of the difficulty of its study.10 For example Dammak et al.11 have proposed an experimental setup to maintain the system as near as possible to the real state of a BIS and the corresponding measurement protocol for determining the bi-ionic potential (BIP).They have elaborated a theoretical treatment, based on the modi–ed Nernst»Planck equation12 to interpret their BIP experimental results. They have shown also the suitability of these equations for describing the transport phenomenon through an IEM and the utility of studying this phenomenon over the entire concentration range (10~5»1 M) For the ion —uxes, we remark that until now, all analytical solutions, giving these through an IEM have been established by several authors9,10,13 for the two limiting concentration domains only.V ery low concentrations. One of the concentrations is less than 10~2 M; the authors9h13 have shown that in this case the diÜusion boundary layers (DBLs) control completely the IDP.V ery high concentrations. Both concentrations are higher than 1 M. It has been shown9 that in this case the IEM controls completely the IDP. However, for the intermediate concentrations, the most used experimentally, there are neither signi–cant theoretical nor experimental studies. In this paper, we propose to solve the same equations used by Dammak et al.11 in the case of a bi-ionic system in order to study theoretically the transmembrane ionic —ux variations over all concentration domains and with diÜerent in—uencing parameters.This approach will be used in the future to develop the study on multi-ionic systems (MIS), considered as the real representation of crossed ionic dialysis (Donnan dialysis). We have organized this paper as follows.First, we will present a short theoretical development, then we will give the general variation curves of the counter-ion and co-ion —uxes and –nally we will evaluate the in—uence of each parameter taking part in the interdiÜusion process. Theoretical development The system under study is shown in Fig. 1. It consists of a cation-exchange membrane (CEM), with univalent –xed anions at a total concentration of X, separating two solutions a and b whose compositions are maintained constant in time.In this system we distinguish four diÜerent interfaces : (i) bulk solution a»DBLa, at (ii) DBLa»CEM, at x\0, (iii) x\[da , CEM»DBLb, at x\d, and (iv) DBLb»bulk solution b, at Thermodynamic equilibrium is presumed between x\d]db . the diÜerent phases at each interface.9 The kinetics of ion exchange at the level of the DBL»IEM interfaces is assumed to be instantaneous,14 thus only the diÜusion process will be studied.Fig. 1 The study system of the cation-exchange membrane. New J. Chem., 1999, 23, 375»380 375Recent studies on the BIP8,15 lead to the following assumptions, in order to be as close as possible to the real conditions of a BIS: (i) mixed control of the interdiÜusion process by both the charged membrane and the diÜusion boundary layers, (ii) the co-ion —ux is diÜerent from zero, (iii) the solvent —ux is not nil, (iv) the affinity and selectivity coefficients are neither constant nor equal to one, (v) the electroneutrality is respected in each phase, and (vi) a zero current density through the studied system is considered. The Nernst»Planck equation is the most frequently used to describe the transport phenomenon of matter in solution and through IEMs.However, this equation does not take into account the solvent —ux. We use then the modi–ed Nernst» Planck equation that includes the correction term to Ci VM calculate the —ux of ion i : Ji Ji\[DiG+(Ci)]Ci+[Ln(ci)]]ziCi F RT +(t)H]Ci VM (1) where denotes the diÜusion coefficient of the ionic species i, Di its valence, its molar activity coefficient, F is Faradayœs zi ci constant, R is the ideal gas constant and T is the absolute temperature.The ion concentration and the electric poten- Ci tial t are assumed to vary only in the x direction. is the VM solution center-of-mass velocity in the membrane. In the usual concentration range mol L~1) and for (C0B1 almost all electrolytes (diÜerent from HCl and NaOH) we can suppose is constant, but it can be diÜerent from unity.In ci fact, the electrolyte quantity absorbed by the membrane remains negligible compared to X (functional site concentration) in recently developed membranes. Thus, the electrolyte strength in the membrane will be independant of the external concentration. Eqn.(1) for a 1 : 1 electrolyte becomes Ji\[Di AdCi dx ]Ci F RT dt dx B]Ci VM (2) where the bars denotes the membrane phase. In this part we will describe the mathematical treatment only for the membrane phase. The same treatments may be achieved for the two DBLs, by considering x\0, x\d and removing the bar on the physico-chemical magnitudes.For the three ionic species within the membrane, which is considered as homogeneous,16 we can write : JA\JA\[DA AdCA dx ]CA F RT dt dx B]CA VM (3a) JB\JB\[DB AdCB dx ]CB F RT dt dx B]CB VM (3b) JY\JY\[DY AdCY dx [CY F RT dt dxB]CY VM (3c) The electroneutrality condition and the zero current density condition lead, respectively, to two equations : CA]CB\CY]X (4a) JA]JB\JY (4b) Removing the potential gradient from eqn.(3), we obtain : [ JA DA \ dCA dx ]CA VMX][(DY[DA) dCA/dx ](DY[DB) dCB/dx] [(DY]DA)CA ](DY]DB)CB[DYX] [ CA DA VM (5) For the counter-ion B, we obtain a similar equation by permuting A and B in the last equation. After rearrangements we obtain : [ JA DA [(DY]DA )CA](DY]DB )CB[DYX] ] CA DA VM [(DY]DA )(CA[X)](DY]DB )CB] \[(2CA]CB[X)DY]CBDB ] dCA dx ]CA(DY[DB ) dCB dx (6a) [ JB DB [(DY]DB )CB](DY]DA )CA[DYX] ] CB DB VM[(DY]DB )(CB[X)](DY]DA )CA ] \[(2CB]CA[X)DY]CADA ] dCB dx ]CB(DY[DA ) dCA dx (6b) If we de–ne: aA0\[ JA DA [(DY]DA )CA](DY]DB )CB[DYX] ] CA DA VM[(DY]DA )(CA[X)](DY]DB )CB ] aA1\[(2CA]CB[X)DY]CBDB ] aA2\CA(DY[DB ) aB0\[ JB DB [(DY]DB)CB](DY]DA )CA[DYX] ] CB DB VM[(DY]DB )(CB[X)](DY]DA )CA ] aB1\CB(DY[DA ) aB2\[(2DB]CA[X)DY]CADA ] Eqn.(6a) and (6b) may be written as : aA0\aA1 dCA dx ]aA2 dCB dx (7a) aB0\aB1 dCA dx ]aB2 dCB dx (7b) We can re-write the system (7a, 7b) in order to give dCA/dx and in the form of two diÜerential equations as a func- dCB/dx tion of the parameters X, d and DA , DB , DY , CA , CB , VM: dCA dx \ aA0 aB2[aB0 aA2 aA1 aB2[aB1 aA2 \f (DA , DB , CA , CB , X, d, VM) (8a) dCB dx \ aB0 aA2[aA0 aB2 aB1 aA2[aA1 aB2 \g(DA , DB , DY , CA , CB , X, d, VM) (8b) In these equations, the convection velocity will be considered as a function of the common concentration, the counter-ion pair and the membrane, but constant through the DBLs and the IEM.The values used in the calculations are those VM obtained experimentally for a typical system (NaCl/CM2/ LiCl).17 In our procedure, we have to know the limiting conditions, that is the concentrations in the membrane at the interfaces.These are determined from the equilibrium condition between 376 New J. Chem., 1999, 23, 375»380the DBLs and the membrane using the Donnan equation, which leads to the two following equations : KAff\ CA CA CB CB (9a) Ksel\ CA CA CY CY (9b) where and are, respectively, the affinity coefficient KAff Ksel and the selectivity coefficient.9 These two coefficients present the ionic activity coefficient variations between the membrane and the solutions.In fact, if we introduce the coefficients KA , and de–ned by then eqn. (9a) and (9b) KB KY Ki\(ci/ ci ), become: KAff\ cA cB cA cB \ KA KB (10a) Ksel\AcAY cA B2 \ 1 KAKY (10b) By symmetry these equations can be written for the interfaces x\0 and x\d.Numerical solution of the two diÜerential equations written for each phase (DBLa, IEM or DBLb) needs a set of initial conditions. So, we choose two values of and obtained JA JB from Fickœs law by supposing that the three phases constitute only one continuous phase of thickness equal to d]2d separating the bulk solutions a and b.From these values of JA and we compute initial values of : (i) and by JB CA(0) CB(0) solving the two diÜerential equations (8a) and (8b) in the DBLa; (ii) and from and by using the CA(0) CB(0) CA(0) CB(0) affinity and selectivity coefficients ; (iii) and by CA(d) CB(d) solving the two diÜerential equations (8a) and (8b) in the membrane; (iv) and from and by using CA(d) CB(d) CA(d) CB(d) again the affinity and selectivity coefficients ; (v) and CA(d]d) by solving again eqns.(8a) and (8b) in the DBLb. CB(d]d) We compare then the computed values of and CA(d]d) to the initial imposed values, 0 and respectively. CB(d]d) C0 , If the diÜerence is large, we select other initial values of the —uxes and and repeat the procedure. This iteration can JA JB be made automatically (according to the Newton»Raphson algorithm), with respect to the diÜerence between the initial and the computed —ux values.For the present study, the diÜusion boundary layer thickness, d, the ionic diÜusion coefficients in the membrane, Di , the selectivity and the affinity coefficients, and and Ksel KAff , the –xed function sites concentration, X, are chosen arbitrarily but within the order of magnitude found in the literature.We have investigated the system KCl for (C0)/IEM/LiCl (C0) which we have collected a maximum of data and the values of the parameters. In addition, the 1:1 electrolyte pair, KCl and LiCl, has a great diÜerence in their cation mobilities, leading to high values of the ion —uxes and bi-ionic potential.Results and discussion General shape of the curve Ji= f(C0) On Figs. 2a and 2b, we report the variations of the counterion and co-ion —uxes vs. the common concentration, which varies from 0.01 to 10 mol L~1. The values of the parameters used in this computation are given in Table 1. The A`, B` and Y~ diÜusion coefficients are respectively 1.95]10~5, 1.05]10~5 and 2.05]10~5 cm2 s~1 (corresponding to the particular case K`, Li` and Cl~).From Fig. 2a, using real coordinates, we remark, for high concentrations mol (C0P1 L~1), that the ionic —uxes vary linearly with However, C0 . from Fig. 2b corresponding to the curves J\f (log it C0), appears that for low concentrations mol L~1) : (i) the (C0O1 Fig. 2 Variation of A`, (]) B` and (]) co-ion —uxes vs.the (L) common concentration : (a) linear coordinates, (b) semi-logarithmic coordinates. —ux variations are asymptotic, decreasing to zero as tends C0 to zero and (ii) the co-ion —ux must be con- (JY\JA[JB), sidered as negligible, and not of the same order of magnitude as the counter-ion —ux. However, the co-ion —ux is of the same order of magnitude as the counter-ion —uxes for C0P1 mol L~1.We also observe over the entire concentration range that the higher the mobility of the counter-ion, the higher its —ux. The set of experimental results, obtained recently by Dieye and colleagues,13,18 concerning the application of Donnan dialysis for de—uorinating drinking water using anionexchange membranes (AEM), con–rm our previsions on the general shape of the curves In fact, even if Jcountervion\f (C0).Dieyeœs experimental conditions are diÜerent from those corresponding to the proposed theoretical treatment (existence of a concentration gradient through the membrane, diÜerent counter-ions and co-ion nature, diÜerent membrane nature and characteristics, etc.) we observe, for the three AEM used (Fig. 3) small variations of the counter-ion —uxes for low concentrations followed by a very sharp increase for high concentrations. The single diÜerence rests in the limits of the concentration domains, which are lower for Dieyeœs results, and are a function of the factors enumerated previously.In addition, the veri–cation of the analytical equations, established in the absence of water —ow and only for the two concentration domains (low and high), leads to the same conclusions concerning the phase that controls the interdiffusion process :19 the DBL for low concentrations and the IEM for high ones.Study of each parameter In this section, we will study the in—uence of one parameter if we take all the others constant and equal to their values listed Table 1 Parameter values used in the compilation of the ionic —uxes Parameter Value used DK/cm2 s~1 1.95]10~5 DLi/cm2 s~1 1.05]10~5 DCl/cm2 s~1 2.05]10~5 d/lm 135 d/lm 60 X/mol mL~1 1.66 VM/cm s~1 0 KAff 1 Ksel 1 New J. Chem., 1999, 23, 375»380 377Fig. 3 Fluoride —ux vs. the upstream compartment concentration for diÜerent membranes.18 in Table 1. The tested parameter will vary arbitrarily within a range –xed from the literature data.DiÜusion boundary layer thickness and the selectivity coefficient. To evaluate the DBL thickness in—uence on the diÜerent —ux values, we have computed these —uxes for diÜerent values of d at diÜerent common concentrations. The results obtained (Fig. 4) show the existence of two concentration domains: mol L~1 and mol L~1. For mol C0O0.1 C0P1 C0O0.1 L~1, the asymptotic evolution of the A` —ux vs.depends C0 sharply on the d values (Fig. 4a). The higher the DBL thickness, the faster the rate of decrease of with concentration. JA For mol L~1, we observe (Fig. 4b) that the evolution C0P1 of the A` —ux vs. is linear. C0 To study thoroughly the in—uence of the DBL thickness on the counter-ion —ux, we present in Fig. 5 (curve a) the variation of the diÜerence between the —uxes corresponding to *JA the two limiting values of the diÜusion boundary layer thickness (d\60 lm and d\200 lm) vs.the common concentration We note that for low concentrations, up to C0 . C0\0.1 Fig. 4 Counter-ion —ux vs. common concentration for diÜerent diÜusion boundary layer thicknesses d (in lm). (a) low values of (b) all C0; values of C0 . Fig. 5 The variation of the diÜerence in the A` —ux at d\60 and 200 lm, vs.the common concentration [with equal to 1 (curve a) Ksel or variable (curve b), see Table 2]. mol L~1, the diÜerence increases ; it then decreases *JA between 0.1 and 1 mol L~1, and then again increases above mol L~1. For weak concentrations, the value of the C0\1 DBL thickness in—uences directly the evolution of the —ux.The diÜerence increases from 0 to 10~8 mol cm~2 s~1. *JA For intermediate concentrations, this diÜerence diminishes to 0.4]10~9 mol cm~2 s~1, which corresponds to a nonnegligible role of the IEM in the control of the interdiÜusion process. It is well-known9 that the higher the common concentration, the greater the role of the membrane in this control. However, for mol L~1, *J increases sharply.C0P1 This is related to the decreasing in—uence of the DBL on the ionic —uxes. To explain the shape of curve a presented in Fig. 5, we have to take into account the sorbed electrolyte, which can be quanti–ed by the selectivity coefficient (eqn. 9b). In our computation, we have considered the experimental values of Ksel obtained by Belaid20 using an atypical homogenous IEM membrane (CM2 from Tokuyama Soda) and NaCl electrolyte (Table 2).The results are given on the same –gure (Fig. 5, curve b). In this case for mol L~1, we remark a con- C0P0.1 tinuous decrease of the diÜerence between the counter-ion —uxes corresponding to d\60 lm and d\200 lm. It appears that (i) the in—uence of DBL thickness decreases with the common concentration and (ii) the selectivity coeffi- C0 cient plays an important part at high concentrations, but not at low ones.Co-ion diÜusion coefficient in the membrane. On Fig. 6 we report the counter-ion —ux variation vs. the common concentration for diÜerent co-ion diÜusion coefficients in the membrane, varying from 10~7 to 5]10~7 cm~2 s~1. For low concentrations, there is no in—uence of the co-ion diÜusion coefficient on the counter-ion —ux values.However, for high concentrations, this in—uence is very important and increases with concentration. In all cases, the increase of the co-ion diffusion coefficient implies an increase in the counter-ion —ux diÜerences, but the K` —ux always remains higher than the Li` —ux. We can explain these results by the fact that at low concentrations, the co-ion is almost completely excluded from the membrane, thus its does not contribute to the ion —ux.Di Table 2 Experimental values of the sorbed electrolyte quantity (in microequivalents) and the selectivity coefficients obtained with a CM2 membrane and NaCl electrolyte20 C0/M nCl Ksel 1.0 11.80 6.53 0.8 9.20 5.41 0.2 2.04 1.56 0.1 0.96 0.83 0.05 0.46 0.43 0.02 0.18 0.18 0.01 0.09 0.09 378 New J.Chem., 1999, 23, 375»380Fig. 6 Variation of the A` —ux vs. the common concentration for diÜerent co-ion diÜusion coefficients in the membrane. The Aë counter-ion diÜusion coefficient. In this paragraph, we discuss the in—uence of a counter-ion diÜusion coefficient (for example on the ionic —uxes (see Table 1). has DA ) DA been varied from 10~7 to 10~6 cm2 s~1 and the obtained results are reported in Fig. 7. Two concentration ranges are observed: at low concentrations, the co-ion —ux is negligible and the interdiÜusion process is entirely controlled by the diffusion boundary layers, while at very high concentrations, the in—uence of is very important; the higher the higher the DA DA —ux of the counter-ion A and the higher the co-ion —ux.Fixed site concentration. In this study of the in—uence of the –xed site concentration, X, we have computed the ionic —uxes for diÜerent values of X: 1.66, 2.50 and 5.00 mole per liter of wet membrane in H` form. The results are reported in Fig. 8. We observe the existence of three diÜerent ranges of the common concentration For very low or very high values C0 . of the –xed site concentration in—uence on the ion —uxes C0 , is not very important.At low concentrations, the DBL controls the interdiÜusion process. Then the membrane plays no part. At high concentrations, the electrolytes AY and BY penetrate considerably into the membrane and their concentrations become very high compared to X. Then, the –xed site concentration does not in—uence the diÜusion of ions.At intermediate concentrations, this in—uence on all the ion —uxes Fig. 7 Variation of the co-ion —ux vs. the common concentration for diÜerent values of the diÜusion coefficient of the counter-ion A` in the membrane. Fig. 8 Variation of the A` —ux vs. the common concentration for diÜerent –xed site concentrations. is obvious, because the interdiÜusion process is controlled by both the ion-exchange membrane and the diÜusion boundary layers.For all curves, we note that the higher the value of X the higher the ionic —ux, because the A` and B` concentration gradients in the membrane and in the DBLs increase with X. Affinity coefficient. To study the affinity coefficient KAff in—uence on the ionic —ux, we have used our treatment to compute these —uxes for diÜerent values of (1, 1.5 and 2).KAff These values are chosen to be as near as possible to the usual range of affinity coefficients corresponding to alkaline chloride electrolyte pairs.21 We show on Fig. 9, the ionic —ux variations vs. the common concentration and the affinity coefficient. It appears that for all concentration domains, the counter-ion —ux variations are negligible.Convection velocity. In Table 3 we give the variation of the A` counter-ion —ux vs. the common concentration and the convection velocity The second column corresponds to a VM . zero value of Column 4 corresponds to the computed A` VM . —uxes taking into account the experimental values of VM (column 3) given by Dammak et al.22 for the bi-ionic system NaCl/CM2/LiCl, and only for less than 1 mol L~1.The C0 Fig. 9 Variation of the A` —ux vs. the common concentration for diÜerent values of the affinity coefficient. Table 3 Variation of the A` —ux vs. the common concentration and the convection velocity VM . C0/M JA/10~8 mol cm~2 s~1 VM/10~6 cm s~1 JA/10~8 mol cm~2 s~1 JA/10~8 mol cm~2 s~1 a 0.01 0.8743 0 0.8743 0.8743 0.025 1.5953 0 1.5953 1.5953 0.05 2.2077 0 2.2077 2.2077 0.1 2.7471 [0.30 2.7555 2.8312 0.25 3.3002 [0.75 3.3209 3.5091 0.5 3.7974 [1.37 3.8262 4.0896 1 4.8797 [1.31 4.88122 4.8983 a For 10 … VM .New J. Chem., 1999, 23, 375»380 379comparison of the two data sets leads us to assume that to the water transport contribution in the cross-ionic —uxes is negligible (less than 1%). However, for the bi-ionic potential, the convection velocity has a contribution of 6»7% of the BIP values.These same proportions can be obtained for convection velocities ten times higher than those obtained experimentally (column 5). Conclusion We have proven the important in—uence of the selectivity coefficient for mean and high concentrations, but the (Ksel) affinity coefficient plays a negligible part.We have also discussed the role of the membrane and the diÜusion boundary layers, and presented a useful abacus for to deter- J\f (DY ) mine the experimental values of under diÜerent experimen- DY tal conditions (membranes, electrolytes, etc.). We con–rm that the DBLs control the interdiÜusion process at low concentrations, but at higher ones the process is controlled by the membrane. In the intermediate concentration domain, control of the interdiÜusion process is mixed.All these theoretical previsions will be veri–ed experimentally and will be the subject of a future paper. References 1 N. Lakshminaraynaiah, T ransport Phenomena in Membranes, Academic Press, New York, 1969. 2 G. Scatchard and F. HelÜerich, Disc. Farad. Soc., 1956, 21, 70. 3 K. Inenaga and N. Yochida, J. Membrane Sci., 1980, 6, 271. 4 K. Kaibara and H. Kimizuka, Bull. Chem. Soc. Jpn., 1982, 55, 1743. 5 D.Mackay and P. Meares, Kolloidn Zh., 1959, 167, 31. 6 M. Tasaka, H. Sugioka, M. Kamaya, T. Tanaka, S. Suzuki and Y. Ogawa, J. Membrane Sci., 1988, 38, 27. 7 M. Tasaka, S. Iwaoka, K. Yamagishi and Y. Ikeda, J. Membrane Sci., 1985, 24, 29. 8 A. Guirao, S. Mafeç , J. A. Manzanares and J. A. Ibane` z, J. Phys. Chem., 1995, 99, 3387. 9 F. HelÜerich, Ion Exchange, McGraw-Hill, New York, 1962. 10 T. Ktari and B. Auclair, J. Membrane Sci., 1987, 32, 251. 11 L. Dammak, C. Larchet, V.V. Nikonenko, V.I. Zabolotsky and B. Auclair, Eur. Polym. J., 1996, 32, 1199. 12 R. Schloé gl and U. Schodel, Z. Phys. Chem., 1955, 5, 372. 13 A. Dieye, PhD Thesis, Universiteç de Paris XII, 1995. 14 K. Bunzl, J. Chem. Soc., Faraday T rans. 1, 1993, 89, 107. 15 L. Dammak, C. Larchet and B. Auclair, J. Membrane Sci., in press. 16 F. G. Donnan, Chem. Rev, 1925, 1, 73. 17 L. Dammak, PhD Thesis, Universiteç de Paris XII, 1996. 18 A. Dieye, C. Larchet, B. Auclair and C. Mar-Diop, Eur. Polym. J., 1998, 34, 67. 19 A. Dieye, C. Larchet, B. Auclair and C. Mar-Diop, Eur. Polym. J., 1999, 35, 461. 20 N. N. Belaid, PhD Thesis, Universiteç de Paris XII, 1993. 21 H. Miyoshi, M. Yamagami and T. Kataoka, Chem. Express, 1990, 10, 717. 22 L. Dammak, R. Lteif, C. Larchet and B. Auclair, New J. Chem., 1998, 22, 605. Paper 8/09931F 380 New J. Chem., 1999, 23, 375»380

ISSN:1144-0546    

DOI:10.1039/a809931f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

About the TATB hypothesis : solvation of the and ions As/4 ë B/4 ó and their tetrahedral and spherical analogues in aqueous/nonaqueous solvents and at a waterñchloroform interface§ Rachel Schurhammer and Georges WipÜ* L aboratoire de Modelisation et Simulations (CNRS UMR 7551), L ouis Moleç culaire Universiteç Pasteur, Institut de Chimie, 4 rue B. Pascal, 67 000 Strasbourg, France. E-mail : wipÜ=chimie.u-strasbg.fr Received (in Montpellier, France) 15th January 1999, Accepted 11th February 1999 Based on molecular dynamics (MD) and free energy (FEP) simulations, we investigate the eÜect of ]/[ charge on the solvation properties of large ììsymmetricalœœ ions in water, acetonitrile and chloroform solutions.The nearly isostructural and tetrahedral ions, which have been assumed to display identical energies of transfer Asz4 ` Bz4~ from water to any solvent (ììTATB hypothesisœœ), are found to display marked diÜerences in solution. The anion interacts more than the cation with water, chloroform and acetonitrile, due to short-range electrostatic interactions, in relation to the solvent granularity and shape of the ion.The importance of charge distribution is demonstrated by the simulations on four diÜerent models of and on –ctitious and ions, and on neutral Asz4 ` Bz4~, Bz4 ` Asz4~ and analogues.The role of ion shape is demonstrated by MD and FEP simulations on isovolumic Bz4 0 Asz4 0 spherical S` and S~ ions, which also display marked diÜerences in solvation properties, but opposite to those of and In water, S` is much better hydrated than S~, due to clathrate-type hydration around S`, while Asz4 ` Bz4~.in acetonitrile, S` and S~ display similar solvation energies. The question of ion pairing is also examined in the three solvents. At a liquid»liquid water»chloroform interface represented explicitly, the anion is found to be Bz4~ more surface active than A number of methodological issues are addressed in the paper.These results are Asz4 `. important in the context of the TATB hypothesis and for our understanding of solvation of large hydrophobic ions in pure liquids or in heterogeneous environments like aqueous interfaces. The precise determination of solvation properties of electrolytes represents a problem of utmost importance in physical chemistry. An inherent difficulty in this problem is our inability to measure individual solvation energies of ions,1 corresponding to their transfer from a –xed point in the gas phase to a –xed point in solution.2 Experiments provide thermodynamic solvation functions of salts, which are generally assumed to be the sum of the cation and anion contributions, but assignment of individual ion contributions rests on ììextrathermodynamic assumptionsœœ, pioneered by Bjerrum and Larsson in 1927.3 As reviewed by Kim4 and Marcus,5,6 the most widely practiced method is ììbased on the assumption that the medium eÜect for a large reference cation can be equated to that of a reference anion of molecular similarity œœ.After the introduction of as a reference electrolyte Pz4 `Bz4~ by Grunwald et al.,7 other pairs of reference ions were suggested, involving the TAB` quaternary ammonium (TAB`\triisoamyl-n-butylammonium8) or TA` (TA\tetraphenylarsonium cations and the Asz4 ` 9) Bz4~ anion.It is indeed generally assumed that bulk cations and anions of the same charge display similar solvation properties in pure aqueous or in mixed water»liquid solvents.7 According to the TATB (TATB\TetraphenylArsonium TetraphenylBorate) hypothesis5,10 and the anion Asz4 ` Bz4~ have the same free energy of transfer from water to any solvent s : *Gwat?s(Asz4 `)\*Gwat?s(Bz4~)\1/2*Gwat?s(Asz4Bz4) (1) § Non-SI units employed: 1 atmB101 kPa; 1 a.u.B2.63]106 J mol~1; 1 kcalB4.184 kJ.Based on this hypothesis, the free energies of transfer of single anions and single cations from water to all kinds of solvents (polar/apolar, protic/aprotic, etc.. . .) have been put on the same scale.5 According to Marcus, this reference electrolyte concept ììdepends on the expectations that the interactions of its cation and its anion with their solvent environments should be equal and independent of the sign of the charge, provided that these ions meet certain criteria.These are that the ions should (i) have unit charge, (ii) be similar in most respects, (iii) have the same size, (iv) be very large, (v) be as nearly spherical as possible and (vi) have an inert peripheryœœ.10 Free energies of solvation involve solute»solvent and solvent»solvent interactions, including entropy and enthalpy components, which cannot be assessed from experiment. However, most of the arguments in favour of the TATB hypothesis deal with one of these components, that is speci–c electrostatic solute»solvent interactions. They are inspired from the continuum Born model, according to which the excess free energy of solvation of a sphere of radius r and ionic charge Q, imbedded in a continuum of dielectric constant e, is and therefore independent of the sign *GBorn\(Q2/2r)(1/e[1), of Q.According to other electrostatic models where a spherical ion interacts with a neutral solvent molecule, a charge reversal of the ion does not change the ion»dipole interaction energy, but reverses the sign of the ion»quadrupole interactions, 11 which are generally assumed to be negligeable.4,5 Based on the fact that the cation is somewhat larger Asz4 ` than the anion, it has been argued, based on the Born Bz4~ model, that the anion might be better solvated than the cation (see discussions in refs. 4 and 5). Another aspect and consequence of the TATB hypothesis concerns the interfacial behaviour of two large ions of opposite charge at a liquid» New J. Chem., 1999, 23, 381»391 381liquid interface : if they are solvated in a similar fashion in the two adjacent liquid phases, they also should display similar properties at the interface between these two liquids, be they ììaloneœœ or in interaction.Here we report a series of molecular dynamics (MD) simulations in explicit solvents, in relation with the TATB hypothesis. They deal with two series of model ions that are likely to meet the above criteria (i) to (vi). The –rst ions are tetrahedral and and the second ones are spherical (S` and (Asz4 ` Bz4~) S~).For comparison, we also consider their neutral analogues of tetrahedral and and of spherical (S0) type, where (Asz4 0 Bz4 0 ) all atomic charges have been set to zero. As solvents, we consider water, chloroform and acetonitrile. Water corresponds to the reference state for free energies of transfer If *Gwat?s .the latter are the same for two ions of opposite charge, the hydration properties of these two ions should be very similar, if not identical. Chloroform and acetonitrile are ììreceiving phasesœœ, of the weakly polar and polar aprotic type, respectively. Two series of analyses are performed. The –rst one deals with the solvation properties of individual species.Structural features are characterized by the ion»solvent radial distribution functions (RDFs) and typical snapshots extracted from the MD trajectories. Energy component analysis is also reported with a particular focus on the electrostatic contributions of the solute»solvent interactions, and on short-range (ìì –rst shell œœ)/long-range contributions. Although the corresponding numbers are not physically observable quantities, they are important for our understanding of diÜerences in solvent interactions when the sign of the ionic charge is reversed. Beyond single ion behaviour, we also investigate, in relation with the TATB hypothesis, the question of ion pairing between the and ions in the three solvents, and Asz4 ` Bz4~ the interfacial behaviour of these ions at a water»chloroform interface represented explicitly.The second series of computations addresses the diÜerences in solvation free energies between a given cation and the related anion in a given solvent. These simulations start with the neutral tetrahedral or spherical S0 species in a given solvent, that is Asz4 0 , Bz4 0 where the cavitation energy12,13 contribution of the solvation process is already paid for.Based on statistical perturbation calculations, we calculate the changes in Gibbs free energy corresponding to the electrostatic charging processes to the anion (*G0~) and to the cation (*G0`). According to the TATB hypothesis, these two quantities, which involve enthalpic and entropic components, should be equal and independent of the solvent.We –nd that this is far from being the case. The calculations also allow to predict diÜerences in free energies of transfer of a given cation/anion from water to a given solvent. From a computational point-of-view, a number of important issues are also addressed and four models of and Asz4 ` Bz4~ are compared. Methods We used the modi–ed AMBER4.1 software14 with the following representation of the potential energy: U\ ; bonds Kr(r[req)2] ; angles Kh(h[heq)2 ] ; dihedrals ;n Vn(1]cos n/) ] ; i:j Aqi qj Rij [2eijARij * RijB6 ]eijARij * RijB12D The geometries of the and ions (see Fig. 1) were Asz4 ` Bz4~ taken from a statistical analysis of 479 and 1073 Asz4 ` Bz4~ ions in the Cambridge crystallographic data base. The average BwC distance (1. 66 is somewhat shorter than the AswC Aé ) distance (1.91 The parameters used to calculate U come Aé ). Table 1 Charge distributions in and See Fig. 1 for Asz4 ` Bz4~. atom de–nitions Asz4 ` Bz4~ ESP a set-1 b set-2 b set-3 b ESP a As/B 0.76 0.5 1.0 0.024 [0.48 C1/C5 [0.12 0.0113 0.0 0.022 [0.29 C2/C4 [0.16 0.0113 0.0 0.022 [0.15 C3 [0.07 0.0114 0.0 0.022 [0.24 C6 [0.12 0.0114 0.0 0.022 0.30 H1/H5 0.15 0.0114 0.0 0.022 0.14 H2/H4 0.18 0.0114 0.0 0.023 0.13 H3 0.15 0.0114 0.0 0.022 0.15 a Charges obtained from a 3-21G basis set.b In these models, the atomic charges of and are identical but of opposite sign. Asz4 ` Bz4~ from the AMBER force –eld,15 allowing for internal ion —exibility and dynamics. The atomic charges used throughout the whole study were derived from ab initio 3-21G electrostatic potentials (ììESP chargesœœ ; Table 1).Additional tests were performed with three other electrostatic models (set-1 to set-3 ; Table 1). In set-1, the charge of the central As/B atom is ]0.5/ [0.5 e and the remaining charge is equally spread over all other atoms. In set-2, the total ]/[1 charge sits on the central atom, while all others are neutral.In set-3, the charge is equally diluted on all atoms, including the central one. In the neutral analogues and all atoms have a zero Asz4 0 Bz4 0 charge. The torsion potentials for the AswC bond of Vi Asz4 ` and the BwC bond of were set to zero. The water, ace- Bz4~ tonitrile and chloroform solvents were represented explicitly with the TIP3P 16, OPLS and OPLS 17 models, respectively. Other models of chloroform were also tested.In the ììstandard calculationsœœ the non-bonded interactions were calculated with a residue-based cutoÜ of 11 in water, 13 in acetoni- Aé Aé trile and 15 in chloroform. Other test calculations were per- Aé formed using larger cutoÜ distances. The solutes were immersed at the center of cubic boxes of pure solvents or at the interface between two adjacent boxes of pure chloroform and water (Fig. 2 and Table 2), represented with periodic boundary conditions in the three directions. The MD simulations were performed at 300 K, at P\1 atm. All CwH, OwH, H… … …H, CwCl and Cl… … …Cl ììbondsœœ Fig. 1 De–nition of atom labels (left) and atom types (right) used in both and ions. Asz4 ` Bz4~ Fig. 2 Representation of an ion pair at a water» Asz4 `Bz4~ chloroform interface. 382 New J. Chem., 1999, 23, 381»391Table 2 Simulation conditions in water, chloroform, acetonitrile and at the water»chloroform interface Number of solvent Time/ Cut oÜ/ Solute Solvent Box size/Aé 3 molecules ps Aé Asz4 `, Bz4~ Acetonitrile 35]36]36 505 200 13 Asz4, Bz4 a Chloroform 37]38]38 390 200 15 Asz4 `, Bz4~ Polar chloro. 37]38]38 390 200 15 Asz4, Bz4 a Water 29]30]30 873 200 11 Asz4Bz4 pairb Acetonitrile 48]37]36 709 600 13 Chloroform 50]39]38 543 600 15 Water 42]30]31 1279 600 11 Asz4 `Asz4 ` pair Chloroform 43]38]37 450 200 15 Water 38]38]38 1860 200 11 Bz4~Bz4~ pair Chloroform 44]38]38 476 200 15 Water 38]37]37 1723 200 11 Asz4, Bz4 b Interface 260 chloro 12 water»chloro. 962 water 322](34]28) 1000 Asz4Bz4 pairb Interface 264 chloro 12 water»chloro. 1333 water 382](34]28) 1000 S`, S0, S~ Acetonitrile 36]36]35 501 200 13 Chloroform 38]38]37 389 200 15 Water 30]30]29 871 200 11 a The same conditions have been used for and neutral species. b The same conditions have been used for Asz4 `, Bz4~, Asz4~, Bz4 ` Asz4 0 , Bz4 0 and neutral species. Asz4 `, Bz4~ Asz4 0 , Bz4 0 were constrained with SHAKE, using a timestep of 1 fs.After 1000 steps of energy minimization, 5 ps of MD were –rst performed with the solute kept frozen. This was followed by 200 to 1000 ps of free MD (see Table 2). Free energy calculations The diÜerence in free energies of solvation between states A and B was calculated using the statistical perturbation FEP theory and the windowing technique,14 with *G\; *Go and *Go\RT logTexp (Uo[Uo`*o) RT Uo The potential energy was calculated using a linear com- Uo bination of charges and parameters of the initial state (k\1) and –nal state (k\0) : The number qo\k … q1](1[k) … q0 .of windows is given in Table 6 and in Section 7. In each window, 2 ps of equilibration were followed by 3 ps of data collection and the change of free energy *G was averaged from the forward and backward cumulated values.Analysis of results Average structures, radial distribution functions (RDFs), solute»solvent and solvent»solvent interaction ener- (Esx) (Ess) gies and their electrostatic/van der Waals components were calculated from the trajectories saved every 0.5 ps. For the simulations at the interface, the position of the interface was recalculated at each step, and de–ned as the intersection between the density curves of the water and chloroform liquids.Results The results are presented as follows. In Section 1, we describe the charge distribution used to model the and Asz4 ` Bz4~ ions. This is followed by the solvation characteristics of these ions, compared to tetrahedral analogues in water, chloroform and acetonitrile (Section 2), and by results concerning the question of ion pairing in solution (Section 3). In Section 4, the solvation of spherical S` and S~ ions is described in the three solvents.Section 5 deals with the interfacial properties of and at the water»chloroform interface. The ques- Asz4 ` Bz4~ tion of relative free energies of solvation of and Asz4 ` Bz4~, obtained from charging their neutral analogues and Asz4 0 is described in Section 6.These results are contrasted Bz4 0 , with those corresponding to the electrostatic charging of a neutral sphere S0 to S` and to S~. We –nally consider the question of diÜerences in free energies of transfer from water to the organic solvents (Section 7). 1 Charge distribution in the and ions As/4 ë B/4 ó In principle, in the context of the TATB hypothesis, the precise charge distribution should be of little importance, provided that the ions have ììa similar and inert peripheryœœ.There is no unique atomic charge distribution for a given molecular solute, as the atomic charges have no physical meaning and depend on the method of derivation. There is also no –rm criteria to provide a balanced description of –rst row (B, C, H)/third row (As) atoms.Our choice was to –t ESP potentials, following a methodology widely used for polar solutes,18 using a 3-21G basis –tted consistently for the As, B, C and H atoms. Table 1 shows that the central atom bears most of the total charge (As`0.76; B~0.48), while the remaining charge is spread on the aromatic groups.In both ions, the aromatic CwH bonds display a clear Cd~Hd` polarity, where the protons are positively charged. Thus, the charges in Asz4 ` and are not simply reversed. It can be noticed that the Bz4~ HOMO is of negative energy ([0.17 a.u.), that is, the anion is correctly described by a bound state. In addition, with these ESP charges, the and ions have a similar Asz4 ` Bz4~ Cd~Hd` periphery.All simulations on and were Asz4 ` Bz4~ performed with this set of charges. In addition, three other models (set-1 to set-3) were tested for the free energy calculations. 2 The and ions and their (hypothetical) As/4 ë B/4 ó tetrahedral analogues and in pure (B/4 ë, As/4 ó, As/4 0 B/4 0) water, chloroform and acetonitrile solutions In solution, the and ions are not rigid, but Asz4 ` Bz4~ undergo coupled rotations of the their phenyl rings (ììgearing eÜectœœ) that lead to the average time equivalence of all ortho and meta protons, in agreement with NMR data.The two ions display very diÜerent interaction energies with water, Esx with acetonitrile, as well as with chloroform (Table 3). The New J. Chem., 1999, 23, 381»391 383Table 3 The and ions and their analogues simulated in pure water, chloroform and acetonitrile liquids. Average solute»solvent Asz4 ` Bz4~ Esx and solvent»solvent interaction energies (kcal mol~1) calculated between 50»200 ps Ess Solute Asz4 ` Bz4~ Asz4 0 Bz4 0 Asz4~ Bz4 ` Asz4 ` e Bz4~ f Water Esx total a [85 [121 [33 [32 [79 [115 [84 [118 Esx elec a [53 [93 0 0 [49 [83 [53 [88 Ess total b [6934 [6886 [7023 [6939 [6899 [6906 [6936 [6910 Ess elec b [8131 [8074 [8250 [8133 [8097 [8097 [8130 [8106 Chloroform Esx total a [65 [87 [46 [45 [76 [65 ]62 [97 Esx elec a [15 [38 0 0 [28 [18 [16 [38 Ess total c [1979 [1977 [1997 [1979 [1985 [1972 [1981 [1985 Ess elec a [49 [45 [53 [51 [47 [47 [48 [47 Acetonitrile Esx total a [94 [101 [47 [45 » » » » Esx elec a [46 [58 0 0 » » » » Ess total d [2557 [2555 [2595 [2578 » » » » Ess elec c [950 [947 [976 [970 » » » » ahd Fluctuations are about (a) 80, (b) 30, (c) 20, (d) 6»7 kcal mol~1.e Calculations performed on with an AswC distance of 1.66 instead Asz4 ` Aé , of 1.91 f Calculations performed on with a BwC distace of 1.91 instead of 1.66 Aé . Bz4~ Aé , Aé . anion interacts more than the cation with the three solvents 7, and 22 kcal mol~1, respectively), mostly due to (*Esx\36, the diÜerences in the electrostatic components (*Esx elec\40, 12 and 23 kcal mol~1, respectively).We notice that the latter do not follow the order of solvent polarities (water[acetonitrile[chloroform). Some speci–c solvation features are analyzed below, with a particular focus on –rstshell solvent molecules, which clearly determine the observed trends.We –rst consider the aqueous solution. Table 3 shows that the better hydration of is not due to the somewhat Bz4~ smaller size of the anion. Indeed, a lengthening of the BwC bond from 1.66 to 1.91 decreases the interaction of Aé Bz4~ with water by 3 kcal mol~1, while a shortening the AswC bond of to 1.66 decreases this interaction by 1 kcal Asz4 ` Aé mol~1 only.As a result, the ììstretchedœœ anion still Bz4~ interacts much more with water than the ììshortenedœœ Asz4 ` cation kcal mol~1), again due to electrostatics. (*Esx\34 The radial distribution functions (RDFs) and snapshots in water (Fig. 3) provide a structural basis for the better hydration of The –rst-shell water molecules of display Bz4~. Asz 4 ` similar average and distances, as their Hw… … …As Ow… … …As Fig. 3 The (top) and (bottom) ions in water. Typical Bz4~ Asz4 ` snapshots and RDFs around the central atom (B or As) of (dotted Ow line) and (full line). Hw OwH dipoles point to the centre of the carbon rings, forming OwH… … …p interactions. Thus, these water molecules do not display an optimal orientation with respect to Asd`.Such an orientation would disrupt the OwH… … …p ììbondsœœ. The –rst hydration shell of is quite diÜerent, as some water mol- Bz4~ ecules can achieve both OwH… … …p interactions with the phenyl rings and charge/dipole interactions with Bd~. As a result, their water protons are closer to Bd~ than the Hw Ow atoms (Fig. 3). Some water molecules achieve bridging p… … …HOH… … …p interactions over two phenyl rings, while others achieve one OH… … …p interaction only (Fig. 3). Within 4.0 Aé from the central atom, there are 3.9 and 0.2 around Hw Ow whereas there are no water atoms around An Bz4~, Asz 4 `. energy component analysis, taking into account only those water molecules that are within 7 from the central atom Aé (Table 4) con–rms their larger interactions with com- Bz4~, pared to (by about 41^7 kcal mol~1).Both ions Asz4 ` display a peak in the RDFs at about 5 which corre- Ow Aé , sponds to the hydration of the ììcoreœœ, and another broad peak at 7.5»8.0 corresponding to the hydration of periph- Aé , erical protons. CwHpara This analysis makes clear that the sign of the ionic charge and the precise charge distribution in the ions determine their hydration properties.This conclusion is con–rmed by the calculations on the hypothetical and ions, where all Asz4~ Bz4 ` atomic ESP charges have been reversed (Table 3). Now, Bz4 ` becomes better hydrated than kcal mol~1), Asz4~ (*Esx\35 again due to the electrostatic component. The hydration schemes around are also quite diÜerent from Bz4 `and Asz4~ the ones around the corresponding (real) and Asz4 ` Bz4~ ions.Although chloroform is less polar than water, it also displays speci–c interactions with the ions. The –rst-shell CHCl3 molecules interact more strongly than the remaining ones by 26^4 kcal mol~1 (Table 4) because of their diÜerent orientations. Around they point their CwH ììprotonœœ toward Bz4~, the Bd~ atom in a arrangement leading to secondary C3v-type interactions between the Cl atoms of and CwH aro- CHCl3 matic protons (Fig. 4). As a result, the –rst peak is B… … …Cchlor found at a shorter distance than the peak. Around B… … …Clchlor the orientation of molecules is inverted : the Cl Asz4 `, CHCl3 atoms are closer to As than the H atoms, and the CwH dipole points outwards, as expected from the Asd`»CHCl3 charge»dipole interactions. This does not lead, however, to attractive interactions with the aromatic rings and, as a result, interacts less than with chloroform.19 Asz4 ` Bz4~ 384 New J.Chem., 1999, 23, 381»391Table 4 The and ions in pure water, chloroform, polar chloroform and acetonitrile liquids. Shell analysis of solute»solvent Asz4 ` Bz4~ interaction energies (kcal mol~1) performed between 50»200 ps.Group 1: First-shell solvent molecules [i.e., which have any atom within 7 (in Aé water and acetonitrile) or 8 (in chloroform) from the central As or B atom]. Group 2: All other solvent molecules within the cutoÜ distancea Aé Asz4 ` Bz4~ Polar Polar Water Chloro. chloro. Aceton. Water Chloro. chloro. Aceton. Esx1 elec [22 [5 [7 [15 [70 [26 [27 [32 Esx1 total [39 [34 [35 [40 [84 [60 [51 [53 Esx2 elec [31 [11 [15 [30 [25 [13 [27 [26 Esx2 total [46 [29 [34 [54 [39 [28 [51 [48 Esx elec [53 [16 [22 [46 [95 [39 [54 [58 Esx total [85 [63 [69 [94 [123 [88 [102 [101 a Fluctuations are about 3»7 kcal mol~1.In acetonitrile, the diÜerence between the ion»solvent interaction energies is smaller than in chloroform (*Esx\7^7 kcal mol~1). The acetonitrile molecules do not display any marked orientation preferences around the cation compared to the anion. The RDFs and typical snapshots (Fig. 5) show that, around a given ion, the two opposite orientations of the Fig. 4 The (top) and (bottom) ions in chloroform. Bz4~ Asz4 ` Typical snapshots and RDFs around the central atom (B or As) of (dotted line) and (full line). Clchlor Cchlor Fig. 5 The (top) and (bottom) ions in acetonitrile.Bz4~ Asz4 ` Typical snapshots and RDFs around the central atom (B or As) and (dotted line) of (full line). Nacet Meacet MeCN dipoles with respect to the central atom are simultaneously present. Generally speaking, when ion»solvent electrostatic attractions increase, the solvent»solvent interactions become less Ess attractive, as the orientation of the –rst-shell solvent molecules with respect to the solute prevents optimal solvent»solvent interactions (Table 3).For instance, the larger interactions of water with compared to kcal Bz4~, Asz 4 ` (*Esx\36^6 mol~1) are partly compensated by the change in water»water interactions kcal mol~1). Thus, relative sol- (*Ess\[48^80 vation energies cannot be obtained solely from the inter- Esx action energies. In addition, the statistical —uctuations on Ess are quite large.Free energy calculations are required to more quantitatively account for the diÜerences in thermodynamic aspects of the solvation (see Sections 6 and 7). 3 The ion pair and its analogue in pure As/4 ëB/4 ó As/4 0B/4 0 water, chloroform and acetonitrile solutions As in real transfer experiments the cations and anions are simultaneously present in solution, it is important to test the status of the ion pair as a function of the solvent.Asz4 `Bz4~ The TATB hypothesis implicitly assumes that the solvation properties of the salt can be described by those of the individual ions, which have therefore no direct contact. In this section, we report on MD simulations performed for 600 ps, starting from an intimate ion pair in water, chloroform and acetonitrile solutions, which show that the ion pair remains intimate in the three solvents (Fig. 6) : the As… … …B distance oscillates between 6.5 and 9.0 along the simulations. Both Aé ions rotate and display interionic electrostatic attractions, exchanging dynamically between p»p stacking (dipole»dipole) and/or CwH… … …p (quadrupole»quadrupole) interactions between their phenyl rings.In water, additional stabilization comes from ììhydrophobic forces œœ, as pointed out by simulations of the pair of the uncharged analogues (Fig. 6). Asz4 0Bz4 0 This pair remains intimate in water, at nearly the same As… … …B distance as in the charged pair (6.5 but dissociates rapidly Aé ), in chloroform and acetonitrile solutions.We thus conclude that the solvation of a given or ion may depend Asz4 ` Bz4~ on its counter ion. As it has been suggested that p-delocalized cations might display mutual attractions in solution20,21 or in the solid state,22,23 we also investigated the and Asz4 `Asz4 ` like ion pairs. They were found to dissociate in the Bz4~Bz4~ three solvents (Fig. 7). 4 Solvation of the spherical Së, Só and S0 species in water, chloroform and in acetonitrile solutions The importance of the positive/negative charge of the solute is further demonstrated by MD simulations performed on a large sphere in its neutral (S0) and charged states (S` and S~). These solutes are represented by van der Waals parameters R*\5.5 (close to the average radius of gyration of the Aé New J.Chem., 1999, 23, 381»391 385Fig. 6 The ion pair and its analogue From Asz4 `Bz4~ Asz4 0Bz4 0 . top to bottom: simulations in water, in chloroform and in acetonitrile. As… … …B distance as a function of time (ps). Bottom: snapshots of the ion pair in water at a As… … …B separation of 7.4 (at 460 ps) and of 9 Aé (at 370 ps). Aé Fig. 7 The and ion pairs in water (top) and Bz4~Bz4~ Asz4 `Asz4 ` in chloroform (bottom). B… … …B and As… … …As distances as a function of time.Fig. 8 Spherical S`, S0 and S~ species in water: (dotted line) and Ow (full line) RDFs around the centre of S. Hw and ions), and e\0.1 kcal mol~1 (close to the e Asz4 ` Bz4~ value of interacting solvent atoms). They meet the above mentioned criteria (i) to (vi) for the TATB hypothesis, but display marked diÜerences in solvation properties, especially in water.In water, the S` cation is found to be better hydrated than the S~ anion (Table 5) because of two cooperative features : S` interacts more than S~ with water kcal (*Esx\13^4 mol~1), and the water»water interactions are more attractive around S` than around S~ kcal mol~1, again (*Ess\13^30 mostly due to electrostatics).This large eÜect is related to the diÜerent orientations of water dipoles around the spheres. The RDFs show that solvation around S` or S0 is of the ììhydrophobic typeœœ, where the OwH bonds are more or less tangential to the solute (Fig. 8) : the and RDFs S… … …Ow S… … …Hw are similar below 7 As the charge is twice the Aé . Ow Hw charge, this structure clearly favours interactions with a positively charged solute, as observed in clathrate-type structures of water around quaternary ammonium ions.24h28 This contrasts with the hydration of S~, where some of the make Hws closer contacts with S~ than do the atoms.Additional Ow analysis of the –rst solvation shells of S0, S` and S~ con–rms the diÜerent orientations of water dipoles, as sketched in Fig. 9. The average SOd angle is about 110° in S`, 90° in S0 and 60° in S~. As a result, the –rst-shell water molecules display hydrogen bonding attractions around S` and repulsions around S~. We also notice that hydration of S0 is similar to that of S`, but diÜers from that of S~. This analysis makes clear why S` is better hydrated than S~, in contrast to Asz4 `, which interacts less with water than Bz4~.In chloroform, the trends are reversed compared to water: S` interacts less than S~ with the solvent kcal (*Esx\10 mol~1), due to the electrostatic component (Table 5). In acetonitrile solution, there is no marked diÜerence between the interaction energies of S` vs. S~ and the Esx solvent kcal mol~1) nor between the solvent» (*Esx\4^5 solvent energies, which are nearly identical (Table 5).The Ess RDFs do not reveal any clear structure of acetonitrile around S0, S` or S~. Thus, again, the role of positive/negative ionic charge markedly depends on the nature of the solvent. Discrimination between the model spherical S` and S~ species, which meet the criteria for the reference electrolyte hypothesis, is largest in water. 5 The and ions and their analogues at the As/4 ë B/4 ó waterñchloroform interface The simulations of the isolated ions, as well as of the Asz4 `- ion pairs, initially set at a water»chloroform interface, Bz4~ Table 5 Spherical S0, S` and S~ species in water, chloroform and acetonitrile solutions. Average solute»solvent and solvent»solute (Esx) (Ess) interaction energies (kcal mol~1) calculated between 50»200 ps Water Chloroform Acetonitrile S0 S` S~ S0 S` S~ S0 S` S~ Esx elec a 0 [51 [39 0 [9 [20 0 [37 [33 Esx total a [6 [55 [42 [8 [17 [27 [6 [40 [36 Ess elec b [8098 [8061 [8037 [44 [43 [40 [948 [932 [933 Ess total b [6911 [6882 [6869 [1248 [1240 [1240 [2478 [2464 [2465 a,b Fluctuations are around (a) 5 and (b) 30 kcal mol~1. 386 New J. Chem., 1999, 23, 381»391Fig. 9 Spherical S`, S0 and S~ in water. Angle SOd with the –rstshell molecules as a function of time. The selected water molecules have their atom within 6 (for S` and S~) or 8 (for S0) from Ow Aé Aé the centre of the sphere. also reveal that the two ions interact diÜerently with the two solvents. This is illustrated by the distances between the As and B atoms and the interface during the simulations and the snapshots of the –nal positions (Fig. 10). Both ions, whose central atom is initially right at the interface, rapidly move to the chloroform phase (Fig. 10). Despite their charge, they do not migrate into water because of the higher cavitation energy (energy cost for creating a cavity13) of water, compared to chloroform, in relation with the diÜerences of their surface tensions (72 and 27 mN m~1, respectively).29 oscillates a few Angstroé ms apart from Bz4~ the interface, while moves deeper into chloroform (up Asz4 ` to 6 As a result, the anion is more attracted by water than Aé ).the cation (by about 50^6 kcal mol~1 during the last 500 ps), while the diÜerence in interaction energies with chloroform is less (8^6 kcal mol~1). The importance of electrostatic forces on the interfacial behaviour is demonstrated by simulations on the neutral analogues and which both migrate into chloroform (at Asz4 0 Bz4 0 , Fig. 10 Snapshots of the and ions (top), of their Asz4 ` Bz4~ Asz4 0 and neutral analogues (middle) and the pair Bz4 0 Asz4 `Bz4~ (bottom) at the water»chloroform interface after 1 ns. Right: distances between the interface and the B (dotted line) or As (full line) atoms as a function of time (ns).Bottom centre : As… … …B distance as a function of time (ns). about 8 after 400 ps), without retaining any contact with the Aé water phase (Fig. 10). When the pair is simulated at the interface, Asz4 `Bz4~ starting with an ionic separation of about 11 the two ions Aé , collapse to form an intimate ion pair, although less tight than in pure water (the B… … …As distance ranges from 6.5 to 8 As Aé ).a result, both ions remain close to the interface, but sits Asz4 ` somewhat deeper than in chloroform (Fig. 10). Thus, the Bz4~ anion is more surface active than the cation. Concerning the S` and S~ ions, the interfacial behaviour is expected to be opposite that of the tetrahedral ions, as the S` cation interacts better with water than S~. 6 DiÜerences in free energies of solvation of tetrahedral and of spherical Së/Só species in water, As/4 ë/B/4 ó chloroform and acetonitrile solutions In this section, we address the question of relative free energies of solvation by FEP computations where the cation is mutated stepwise into the anion (or vice versa), via an ììalchemical routeœœ.30,31 In the case of the tetrahedral ions, we decomposed the to mutation via the non- Asz4 ` Bz4~ physical intermediate neutral states : Asz4 `]Asz4 0 (step 1: *Gs `0\[*Gs0`) Asz4 0 ]Bz4 0 (step 2: *Gs00) Bz40]Bz4~ (step 3: *Gs0~) In a given solvent s, *Gs `~\*Gs `0]*Gs00]*Gs0~. According to the TATB hypothesis, should be zero in *Gs `~ any solvent.Table 6 shows that this is not the case, whatever the model of ions and the simulation conditions are.We –rst notice (Table 6) that the contribution is negative, as *Gs00 expected from the Born or cavitation models,12,13,32 but quite small and negligible (from [0.1 to [0.5 kcal mol~1), compared to the two other terms. Thus, the diÜerence in solvation free energies mostly results from the and contri- *Gs0~ *Gs0` butions, which are negative and diÜerent from each other.In the following, we consider the values obtained with the largest cutoÜ distances. However, the observed trends are independent of the cutoÜ (Table 6). Among the tetrahedral ions, is better solvated than Bz4~ as *G0~ is more negative than *G0` in water ([41 Asz4 ` and [31 kcal mol~1, respectively), in chloroform ([22 and [8 kcal mol~1, respectively) and acetonitrile ([27 and [23 kcal mol~1, respectively).In the case of isovolumic spherical ions, the contribu- *Gs00 tion is zero. These ions behave opposite to the tetrahedral ones, and display nonzero values of *G`~ that are spectacularly solvent dependent. This value is largest and positive in water (15.9 kcal mol~1), negative in chloroform ([4.8 kcal mol~1), and nearly zero in acetonitrile.Indeed, in water, charging S0 to S` is far more favourable (more than (*Gwat0`) twice) than charging S0 to S~ As pointed out in (*Gwat0~).33 Section 4, this stems from the fact that the water structure around S0 is ììpreorganizedœœ for complexing a S` species, as it creates a negative potential ([9.5^2 kcal mol~1) at the centre of the sphere.In addition, charging S0 to S` can be performed without markedly disrupting the water structure, while charging S0 to S~ involves the rupture of –rst-shell hydrogen-bond networks, and inversion of water dipoles. We notice that the potential created by the water molecules around the neutral and species is also negative Bz4 0 Asz4 0 ([9.5^0.5 kcal mol~1), and thus should favour charging to compared to Thus, the fact that is better Asz4 ` Bz4~.Bz4~ hydrated than is due to speci–c interactions of water Asz4 ` with the anion as described above, rather than to a ììpreorganizationœœ of water around or The above Bz4 0 Asz4 0 . New J. Chem., 1999, 23, 381»391 387Table 6 Summary of consistent calculations of *Gs (kcal mol~1) and diÜerences in free energies of solvation of and S`/S~ *Gs `~ Asz4 `/Bz4~ in a given solvent s.The diÜerences between forward and backward calculated *Gs are about 0.2 kcal mol~1. Mutations were achieved in 21 windows in water and acetonitrile, and in 51 windows in chloroform. Unless otherwise speci–ed, all values for are obtained with the Asz4 `/Bz4~ ESP charges Solvent CutoÜ/Aé [*Gs0` **Gt a *Gs00 [*Gs0~ *Gs `~ Tetrahedral species Asz4 ` ] Asz4 0 ] Bz4 0 ] Bz4~ Asz4 ` ] Bz4~ Water 11 31.5 [0.4 [41.1 [10.0 » 11/15 29.7 [0.4 [37.8 [8.5 » 11b 37.1 [0.4 [20.4 16.3d » 11c 44.4 [0.4 [30.3 13.7e » 11d 33.2 [0.4 [17.0 15.8f » Chloroform 15e 7.6 [0.1 [22.1 [14.6 4.6 15f 7.2 [0.1 [20.3 [13.2 3.2 15b 9.1 [0.1 [17.4 [8.4d 24.7 15d 9.1 [0.1 [17.3 [8.3f 24.1 Acetonitrile 13 23.0 [0.4 [27.0 [4.4 [5.6 Spherical species S` ] S0 ] S~ S` ] S~ Water 11 30.7 [13.5 17.2 » 11/15 28.1 [12.2 15.9 » Chloroform 15 7.2 [12.0 [4.8 22.0 Acetonitrile 13 19.5 [18.8 0.7 16.5 is the diÜerence in free energies of transfer of the cation and the anion, computed consistently (see text).bhd The charges a **Gt\*Gt `[*Gt~ used for and are (b) set-1, (c) set-2 or (d) set-3 de–ned in Table 2.e OPLS model of chloroform. f All atom model of chloroform, Asz4 ` Bz4~ from ref. 53. analysis makes clear why ììwater cagesœœ (clathrates) are generally found around large cations,24h28 but never, to our knowledge, around large anions. This can also be considered as an indirect proof of the importance of sign reversal in ions. In chloroform, the S` and S~ spherical ions are not equally solvated, but the trend is inversed, compared to water: the anion is better solvated than the cation by about 5 kcal mol~1 kcal mol~1 and kcal (*Gchlor0~\[12.0 *Gchlor0`\[7.2 mol~1). In acetonitrile, both S0 to S~ and S0 to S` mutations lead to identical energy changes (*Gacet0`\*Gacet0~\[19 kcal mol~1), indicating that the solvation of these two ^0.5 ions is similar.This is fully consistent with the lack of signi–- cant diÜerences in interaction energies reported above. Esx Thus, a charge reversal from S` to S~ has minor energetic eÜects in acetonitrile. Comparison of tetrahedral vs. spherical species shows that the diÜerence in solvation energies of a large cation vs. an isostructural anion depends markedly on the solvent.EÜects are largest in water and smallest in acetonitrile. They thus do not simply follow the solvent polarity. These conclusions are strengthened by a number of computational tests, related to the treatment of long-range electrostatic interactions and to the sampling problem.34 For the tetrahedral species, the *Gs depend, as stressed above, on the precise charge distributions. This is –rst demonstrated by charging to and to Bz4 0 Bz4 ` Asz4 0 Asz4~ (ììinversed analoguesœœ with ESP charges).The corresponding *Gs ([47 and [18 kcal mol~1, respectively in water; [11 and [17 kcal mol~1 in chloroform) diÜer markedly from the values obtained for the ìì real ionsœœ. However, they lead to kcal mol~1, that is, again to a marked cation/ *Gwat `~\30 anion discrimination by the solvent.The second series of tests deals with the ììhand-madeœœ electrostatic models where the atomic charges of two ions are identical, but reversed (set-1 to set-3, de–ned in Table 2). In water, all three lead to a marked preference for cation hydration ranges from 13.7 to (*Gwat `~ 16.3 kcal mol~1; see Table 6). This can be explained by the negative potential created by water at the centre of the Asz4 0 and species and by the absence of speci–c solvation pat- Bz4 0 terns with these models.The important result is that in water, is far from being close to zero with any of these *Gwat `~ models. In chloroform, where only set-1 and set-3 models were compared, remains diÜerent from zero and negative *Gchlor `~ (about [8 kcal mol~1) ; is better solvated than Bz4~ Asz4 `, as with the ESP model. 7 DiÜerences in free energies of transfer of individual ions from water to the organic liquids Now comes the question to more quantitatively assess the differences in free energies of transfer from water to chloro- **Gt form for two ions of opposite charge. As illustrated in Scheme 1 for the pair, According Asz4 `/Bz4~ **Gt\*Gt `[*Gt~. to this thermodynamic cycle, **Gt\*Gwat `~[*Gchlor `~.Similarly, for the transfer to acetonitrile, **Gt\*Gwat `~ (\0 according to the ììTATB hypothesisœœ). In [*Gaceto `~ doing so, we assume that the organic phase is dry, which may not be the case as some hydrogen-bonded water molecules may follow the ion into the organic phase (ììwater dragging eÜectœœ).35,36 For the tetrahedral ions, the results obtained with diÜerent conditions (Table 6) con–rm that is larger than *Gwat `~ implying that is less easily transferred than *Gchlor `~, Asz 4 ` from water to chloroform. With the 15 cutoÜ, the Bz4~ Aé calculated is quite large (]6 kcal mol~1). The same **Gt conclusion is obtained with the ESP charges as well as with the charges of set-1 or set-3.It is con–rmed using two other models of chloroform.37 It seems quite paradoxical, if one considers that the anion interacts better than the cation with water, and seems therefore more ììhydrophilicœœ. For the transfer of to acetonitrile, the trend is Asz4 `/Bz4~ opposite to the one in chloroform: is *Gwat `~[*Gaceto `~ negative (from [4.1 to [5.6 kcal mol~1, depending on the simulation conditions), which means that the transfer of to acetonitrile is preferred.Asz4 ` Concerning the spherical ions S` and S~, which best –t the ììTATB hypothesisœœ we calculate again a marked diÜerence in Scheme 1 388 New J. Chem., 1999, 23, 381»391transfer properties : S~ is more easily transferred to acetonitrile ranges from 15 to 17 kcal mol~1), (*Gwat `~[*Gaceto `~ as well as to chloroform kcal (*Gwat `~[*Gchlor `~\20.6 mol~1) than S`.We notice that these numbers are quite large, due mostly to the the better hydration of S`, compared to S~. Discussion and conclusion MD and FEP simulations on and ions and ana- Asz4 ` Bz4~ logues show that sign reversal of the ionic charge leads to marked diÜerences in solvation properties in pure water, acetonitrile, chloroform solutions and at the water»chloroform interface. It appears clearly that, although both tetrahedral ions are similar in size and shape, interacts less than Asz4 ` with all solvents.The eÜect of ion size is nearly neg- Bz4~ ligeable. In the case of the model large spherical S`/S~ ions of identical size, which best –t the criteria for the TATB hypothesis, marked diÜerences are also found in solvation properties and the discrimination is largest in water, where S` is better hydrated than S~.EÜects of charge reversal are mostly ascribable to short-range speci–c interactions, determined by the sign of the ionic charge, the shape of the ììlarge ionœœ and the nature of the solvent. The eÜect of charge reversal cannot be therefore simply assessed by solvent continuum models.Our analysis points out the role of solvent granularity of speci–c interactions in the –rst solvation shells. In the case of recent spectros- Bz4~, copy studies of the HDO water molecules surrounding Bz4~ and the analogue of in aqueous solution have Pz4 ` Asz4 ` been reported.38 They revealed distinct diÜerences in their hydration and concluded that the anion interacts more with water than the cation and that ììthe eÜect of is deter- Bz4~ mined by the anion»water interactions, while the eÜect of is determined by water»water interactions around the Pz4 ` cationœœ.38 This is consistent with our results.Based on the environment analysis of in solid state structures, Bz4~ Marcus quoted that ììthe energetic eÜect of hydrogen bonding between the uncharged water solvent molecules and this anion should not be signi–cantœœ.10 This is not supported by these results.We performed a systematic search of water around the and ions in the Cambridge crystallographic data Bz4~ Asz4 ` base. It is remarkable that the hydration patterns we calculate around are identical to those found in several X-ray Bz4~ structures39,40 in which the HOH molecules display bridging p… … …HwOwH… … …p interactions as depicted in Fig. 3.Such patterns are consistent with the IR spectrococopic results of ref. 38. They contrast with the lack of speci–c interactions between and water in the solid state. Asz4 ` There are many factors that contribute to the solvation thermodynamics of ions (see discussions in refs. 10, 41 and 42 for experimental aspects and in refs. 43»47 for computational aspects). The eÜect of the sign of the charge is not clear. Luzhkov and Warshel compared the hydration energies of the and ions, using two microscopic models, where Pz4 ` Bz4~ the solvent is represented either by polarizable Langevin dipoles or by explicit solvent molecules interacting with the solutes.48 They concluded that is better hydrated than Bz4~ due to the diÜerences in charge distribution and to Pz4 `, ìì steric factors œœ.This is fully consistent with our results. In the case of small spherical ions (e.g., Cl~/ììCl`œœ), RISM-HNC calculations suggest that cations are less hydrated than the anions, due to diÜerences in their ììeÜective size œœ. Other simulations of these Cl~/ììCl`œœ species at a water»dichloroethane interface suggest that Cl~ can approach closer to the interface than ììCl`œœ and is therefore somewhat less surface active.49 The surface activity is generally related to the amphiphilic character of the solute and it is not clear whether Cl~ would be more ììhydrophilicœœ than ììCl`œœ.Our calculations suggest that is more surface active than which is in Bz4~ Asz4 `, agreement with related experiments,50 while S` would more surface active than S~.According to the TATB hypothesis, the and Asz4 ` Bz4~ ions, or (better) the (hypothetical) S` and S~ ions, should display the same energies of transfer from water to any other solvent. This is not found in our simulations. With none of the models are the free energies of solvation identical. There is no simple relationship between solvent polarity and diÜerences in solvation energies.Based on electrochemical measurements across the water»1,2-dichloroethane interface, it was found that is more easily transferred than (the corre- Asz4 ` Bz4~ sponding standard Gibbs energies are [9.6 and [8.6 kcal mol~1, respectively).51 However, these data may not strictly relate to our calculations, as the organic phase is diÜerent, and likely not dry.The eÜect of counter ion, not addressed in our FEP studies, also requires further investigations, as the nature of the and of ion pairs may depend on the X~Asz4 ` Y`Bz4~ nature of the X~ and Y` counter ions as well as on the solvent, as suggested by our computations on the Asz4 `Bz4~ pair. It is worth pointing out that the very low solubility of in water (about 10~8.5 mol l~1) has been determined Asz4Bz4 experimentally by c irradiation,4 which gives no indication of the status of the ion pair (intimate/dissociated). Our simulations deal with a concentration of about 5]10~2 mol l~1, that is, to supersaturated conditions.There might thus be some concentration eÜects on the solvation properties of the cation vs.the anion. On the other hand, our results point out the importance of solvent interactions in the –rst shells, which should be less dependent on the concentration, if the ions are dissociated. Concerning the computations, a quantitative assessment of solvation features and thermodynamic properties certainly requires that the energy representation of the system (ions and solvents) and an ììadequateœœ treatment of electrostatic and internal interactions52 (especially with an explicit representation of non-additivity and polarization eÜects53,54) be tested.In the case of the spherical S` and S~ model species, there is no problem of charge distribution in the solute, whose periphery is ìì inert œœ. However, quite large diÜerences in solvation properties are found, especially in water, leading to marked diÜerences in free energies of transfer to chloroform or to acetonitrile.In the case of and ions, the electrostatic Bz4~ Asz4 ` representation is not unique. Even in well-documented examples of small molecules involving C, H, O, N atoms only there is no a priori ììbest choice of chargesœœ. The latter are tested on experimental quantities, including free energies of solvation,55 which are not available from experiment for the individual and ions.The charges chosen in this study were Bz4~ Asz4 ` derived consistently from ESP calculations, but other sets may be derived and repeatedly tested in various solutions. We notice that three other sets of charges (and particularly set-2, which models the periphery of ions as perfectly ìì inert œœ) lead qualitatively to the same conclusions concerning the importance of sign reversal on diÜerences in free energies of solvation of large symmetrical ions.Other critical parameters for this study are those of the solvents, which were derived from pure liquid phase properties, and may not be accurate enough to describe their mutual competition with the solutes properly.It remains that the two models of tetrahedral or two spherical ions interact very diÜerently with a given model of solvent. Three diÜerent models of chloroform give the same trends. We are currently investigating other models of solvents and modi- –ed treatments of the long-range electrostatic interactions. Further improvements involve a mixed MM/QM description of the potential energy,56,57 which again raise the question of a balanced description of the two ions.In principle, annihilation of the pair in the diÜerent solvents should Bz4~Asz4 ` allow one to compare the diÜerences in total solvation energies from one solvent to the other. This introduces new computational problems, due to the large size of the solute, as well New J.Chem., 1999, 23, 381»391 389as physical problems, related to the eÜect of concentration on the status of the ion pair as a function of the solvent. Our computational results are disturbing in many respects, if one refers to classical representations of large hydrophobic ions in solution. Contrary to assumptions made in the TATB hypothesis, these ions are calculated to display marked diÜerences of solvation properties, which depend on the sign of the ionic charge, the nature, hydrogen bonding capabilities and polarity of the solvent.In addition, the precise shape of the ion matters, as tetrahedral ions may display the opposite behaviour compared to spherical ones. As the granularity of the solvent plays an important role (even around the large spherical ions), the latter cannot be modelled solely by a continuum.Our study should stimulate further theoretical treatments along the line described above, as well as experiments on the eÜect of the charge of ions and properties in solution. Fundamentally, they have bearing on our understanding of the hydrophilic/hydrophobic character of neutral and large ionic solutes58,59 and on their behaviour at aqueous interfaces.60h69 Such computations will play an increasing role in the critical analysis of the thermodynamics of solvation and transfer of ions.Acknowledgements authors are grateful to IDRIS and Universiteç Louis The Pasteur for computer resources and to PRACTIS for support. RS thanks the French Ministery of Research for a grant. Notes and references 1 R.A. Robinson and R. H. Stokes, Electrolyte Solutions, Butterworths, London, 1955, ch. 3. 2 A. Ben-Naim and Y. Marcus, J. Chem. Phys., 1984, 81, 2016. 3 N. Bjerrum and E. Larsson, Z. Phys. Chem., 1927, 127, 358. 4 J. I. Kim, J. Phys. Chem., 1978, 82, 191. 5 Y. Marcus, Ion Solvation, Wiley, Chichester, 1985. 6 Y. Marcus, Pure Appl. Chem., 1990, 62, 900. 7 E. Grunwald, G.Baughman and G. Kohnstan, J. Am. Chem. Soc., 1960, 82, 5801. 8 O. Popovych, Anal. Chem., 1966, 38, 558. 9 R. Alexander and A. J. Parker, J. Am. Chem. Soc., 1967, 89, 5549. 10 Y. Marcus, J. Chem. Soc., Faraday T rans. 1, 1987, 83, 2985. 11 A. D. Buckingham, Discuss. Faraday Soc., 1957, 24, 151. 12 R. A. Pierotti, Chem. Rev., 1976, 76, 717. 13 M. Preç vost, I. T. Oliveira, J. P.Kocher and S. J. Wodak, J. Phys. Chem., 1996, 100, 2738. 14 D. A. Pearlman, D. A. Case, J. C. Caldwell, G. L. Seibel, U. C. Singh, P. Weiner and P. A. Kollman, AMBER4, University of California, San Francisco, 1991. 15 W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz, D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell and P. A. Kollman, J. Am. Chem. Soc., 1995, 117, 5179. 16 W. L. Jorgensen, J. Chandrasekhar and J. D. Madura, J. Chem. Phys., 1983, 79, 926. 17 W. L. Jorgensen, J. M. Briggs and M. L. Contreras, J. Phys. Chem., 1990, 94, 1683. 18 P. A. Kollman, Acc. Chem. Res., 1996, 29, 461. 19 The importance of electrostatics is further demonstrated by additional simulations of both ions repeated in a ììpolarœœ chloroform model, where the OPLS charges have been scaled by 1.3.This increases the interaction energies of the solvent with and Bz4~ relative to the unscaled OPLS charges, and also the diÜer- Asz4 `, ence in ììsolvation energiesœœ of the two ions kcal (*Esx\32^5 mol~1). 20 S. Boudon, G. WipÜ and B. Maigret, J. Phys. Chem., 1990, 94, 6056. 21 L. Troxler, J. M. Harrow–eld and G. WipÜ, J. Phys. Chem., 1998, 102, 6821. 22 I. Dance and M. Scudder, Chem. Eur. J., 1996, 2, 481. 23 I. Dance and M. Scudder, J. Chem. Soc., Dalton T rans., 1996, 3755. 24 G. A. JeÜrey, Acc. Chem. Res., 1969, 2, 344. 25 G. A. JeÜrey and W. Saenger, Hydrogen Bonding in Biological Structures, Springer»Verlag, Berlin, 1991. 26 G. A. JeÜrey, in Inclusion Compounds, ed. J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London, 1984, pp. 135»190. 27 J. Lipkowski, in Comprehensive Supramolecular Chemistry, ed. D.D. MacNicol, F. Toda and R. Bishop, Pergamon, New York, 1996, pp. 691»714. 28 J. Lipkowski, in Crystallography of Supramolecular Compounds, ed. G. Tsoucaris, Kluwer Academic Publishers, Dordrecht, 1995, pp. 265»283. 29 A. W. Adamson, Physical Chemistry of Surfaces, 5th edn., Wiley, New York, 1990. 30 T. P. Straatsma and J. A. McCammon, Annu. Rev. Phys. Chem., 1992, 43, 407. 31 W. L. Jorgensen, Acc. Chem. Res., 1989, 22, 184. 32 B. Guillot, Y. Guissani and S. Bratos, J. Chem. Phys., 1991, 95, 3643. 33 The values are close to those expected from the Born *Gwat `~ model ([30 kcal mol~1), whereas the values might seem *Gwat0~ surprisingly low. We checked on the reverse mutation S~ to S0 that the sampling was sufficient kcal mol~1, (*Gwat~0\]13.7 nearly identical to [*Gwat0~). 34 Due to the –nite size of the simulated system, the *Gs are evaluated within the cutoÜ distance and the contributions beyond this distance may not be the same for steps 1 to 3. The –rst series of tests concern the increase from a 11 cutoÜ to a 11/15 twin Aé Aé cutoÜ. In the case of water solutions of tetrahedral and spherical species, this somewhat reduces *G`~ (by less than 1.5 kcal mol~1), and its *G0~ and *G0` components.The question of sampling was checked on the –rst half of the to muta- Bz4~ Bz4 0 tion in water to which displays the largest *Gs, (Bz4~ Bz4~0.5), compared to the other systems (Table 6). The results obtained with 21, 51 and 101 windows, using a 11 cutoÜ, were very close (34.8, Aé 34.5 and 34.0^0.2 kcal mol~1, respectively), indicating that there is no major sampling problem when using 21 windows only per mutation. 35 R.-S. Tsai, W. Fan, N. El Tayar, P. A. Carrupt, B. Testa and L. B. Kier, J. Am. Chem. Soc., 1993, 115, 9632. 36 T. Osakai, A. Ogata and K. Ebina, J. Phys. Chem. B, 1997, 101, 8341. 37 As critically determines the diÜerence in free energies of *Gchlor `~ transfer, we made two other estimates of this quantity.The –rst one used a more polar model of chloroform, where the OPLS charges were scaled by 1.3. For the direct mutation of to Asz4 ` (51 windows and 15 cutoÜ), becomes more Bz4~ Aé *Gchlor `~ negative than with the OPLS charges ([15.7 versus [12.3 kcal mol~1). The second test was performed using an all-atom model of chloroform (from ref. 53) and a 15 cutoÜ. All *Gs were similar Aé to those obtained with the OPLS model (Table 6). In particular, was [13.2 instead of [14.6 kcal mol~1 (i.e., still more *Gchlor `~ negative than Thus, these tests strengthen the above *Gwat `~). conclusion that is more easily transferred to chloroform Bz4~ than Asz4 `. 38 J. Stangret and E. Kamieç nska-Piotrowics, J. Chem. Soc., Faraday T rans., 1997, 93, 3463. 39 P. K. Bakshi, A. Linden, B. R. Vincent, S. P. Roe and D. Adhikesavalu, Can. J. Chem., 1973, 72, 1273. 40 P. K. Bakshi, Can. J. Chem., 1994, 72, 1273. 41 Y. Marcus, Pure Appl. Chem., 1987, 9, 1093. 42 M. H. Abraham and J. Liszi, J. Inorg. Nucl. Chem., 1981, 43, 143. 43 B. Roux, H.-A. Yu and M. Karplus, J. Phys. Chem., 1990, 94, 4683. 44 M. Schaefer and M. Karplus, J. Phys. Chem., 1996, 100, 1578. 45 A. Warshel and S. T. Russell, Quart. Rev. Biophys., 1984, 17, 283. 46 A. A. Rashin, J. Phys. Chem., 1989, 93, 4664. 47 J. and T. Hansson, J. Phys. Chem., 1996, 100, 9512. Aé qvist 48 V. Luzhkov and A. Warshel, J. Comput. Chem., 1992, 13, 199. 49 K. J. Schweighofer and I. Benjamin, J. Phys. Chem., 1995, 99, 9974. 50 S. E. Jackson and M. C. R. Symons, Chem. Phys. L ett., 1976, 37, 551. 51 Y. Shao, A. A. Stewart and H. H. Girault, J. Chem. Soc., Faraday T rans., 1991, 87, 2593. 52 X. Daura, P. H. Hué nenberger, A. E. Mark, E. Querol, F. X. Avileç s and W. F. van Gunsteren, J. Am. Chem. Soc., 1996, 118, 6285. 53 T.-M. Chang, K. X. Dang and K. A. Peterson, J. Phys. Chem. B, 1997, 101, 3413. 54 M. H. New and B. J. Berne, J. Am. Chem. Soc., 1995, 117, 7172. 55 G. Kaminski and W. L. Jorgensen, J. Phys. Chem., 1996, 100, 18 010. 56 J. Gao, Acc. Chem. Res., 1996, 29, 298. 57 M. A. Thompson, E. D. Glendening and D. Feller, J. Am. Chem. Soc., 1994, 116, 10 465. 58 B. G. Rao and U. C. Singh, J. Am. Chem. Soc., 1989, 111, 3125. 390 New J. Chem., 1999, 23, 381»39159 W. Blokzijl and J. B. F. N. Engberts, Angew. Chem., Int. Ed. Engl., 1993, 32, 1545. 60 R. Bennes and B. E. Conway, Can. J. Chem., 1981, 59, 1978. 61 W. A. Ducker and L. M. Grant, J. Phys. Chem., 1996, 100, 11 507. 62 M. Hato, J. Phys. Chem., 1996, 100, 18 530. 63 C. Y. Lee, J. A. McCammon and P. J. Rossky, J. Chem. Phys., 1984, 80, 4448. 64 J. S. Nowick, J. S. Chen and G. Noronha, J. Am. Chem. Soc., 1993, 115, 7636. 65 A. R. van Buuren, S.-J. Marrink and J. C. Berendsen, Colloids Surf., 1995, 102, 143. 66 F. Berny, N. Muzet, R. Schurhammer, L. Troxler and G. WipÜ, in Current Challenges in Supramolecular Assemblies, ed. G. Tsoucaris, Kluwer Academic Publishers, Dordrecht, 1998, pp. 221»248. 67 F. Berny, N. Muzet, L. Troxler and G. WipÜ, in Supramolecular Science : W here It Is and W here It Is Going, ed. R. Ungaro and E. Dalcanale, Kluwer Academic Publishers, Dordrecht, 1999, pp. 95» 125. 68 M. Lauterbach, E. Engler, N. Muzet, L. Troxler and G. WipÜ, J. Phys. Chem. B, 1998, 102, 225. 69 N. Muzet, E. Engler and G. WipÜ, J. Phys. Chem. B, 1998, 102, 10 772. Paper 9/00442D New J. Chem., 1999, 23, 381»391 391

ISSN:1144-0546    

DOI:10.1039/a900442d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Synthesis, thermal analysis and crystal structure of lead(II) diaqua 3,6-dicarboxylatopyridazine. Evaluation of performance as a synthetic precursor§ Sophie Sobanska,a Jean-Pierre Wignacourt,*a Pierre Con—ant,a Michel Drache,a Michel Lagreneç ea and Elizabeth M. Holtb a L aboratoire de Cristallochimie et Physicochimie du Solide, CNRS URA 0452, ENSCL and UST L , BP 108, 59652 V illeneuve dœAscq CEDEX, France.E-mail : Jean-Pierre.W ignacourt=ensc.lille.fr b Department of Chemistry, Oklahoma State University, Stillwater, OK 74078, USA. E-mail : betsy=biochem.okstate.edu Received (in Montpellier, France) 25th January 1999, Accepted 27th January 1999 (DCP\3,6-dicarboxylatopyridazine) has been synthesized and characterized by thermal Pb(H2O)2DCP gravimetric analysis and single crystal X-ray dÜraction: tetragonal, space group a\b\15.278(2), I41/a, c\15.218(3) U\3552.3(1) Z\16.The structure consists of columns of lead atoms linked by tetradentate Aé , Aé 3, DCP molecules in a spiral array. The irregular 8-fold coordination sphere of PbII consists of two bidentate DCP molecules, the oxygen atoms of two neighboring ligands and two coordinated water molecules. The thermal decomposition of this complex to yield PbO at low temperatures was observed by powder diÜraction and by thermal gravimetric analysis.Comparison of and as precursors in the synthesis Pb(H2O)2(DCP) Pb(H2O)2(oxalate) of the mixed metal oxide, showed the use of to result in mixed metal oxide product Bi12PbO19 , Pb(H2O)2(DCP) with a smaller grain size. Synthe` se, etude thermique et structure cristalline du plomb(II) diaqua 3,6-dicarboxylatopyridazine.Utilisation potentielle en tant que preç curseur de synthe` se. Le composeç plomb(II) diaqua 3,6-dicarboxylatopyridazine a eç teç preç pareç puis eç tudieç par analyse thermogravime` trique. La structure cristalline a` eç teç [Pb(H2O)2DCP] deç termineç e par diÜraction des rayons X sur monocristal; le syste` me cristallin est quadratique, groupe dœespace avec a\b\15.278(2), c\15.218(3) U\3552.3(1) Z\16.Le mode` le structural est ba� ti a` partir de I41/a, Aé , Aé 3, colonnes dœatomes de Pb, relieç es par deux ligands teç tradentate de DCP et deç veloppeç es en spirales. La coordinance huit du plomb(II) est obtenue a` partir de deux ligands DCP bidentates, de deux atomes dœoxyge` ne provenant de deux ligands voisins et de deux moleç cules dœeau.Lœeç volution thermique de ce complexe a eç teç suive par thermodiÜraction X et analyse thermogravime` trique. Elle conduit a` lœobtention de PbO de` s 275 °C. Un test comparatif dœutilisation des complexes oxalates ou DCP pour la synthe` se de a prouveç lœinteç re� t de ce Bi12PbO19 dernier, qui conduit a` lœobtention de lœoxide mixte avec une faible granulomeç trie.Interest in the use of organometallic complexes as potential precursors of the metal oxides necessary for further synthesis of conducting or superconducting materials by low temperature or ììchimie douceœœ procedures has increased in recent years. The most widely used complexing agents are tartaric acid,1,2 EDTA (ethylenediaminetetraacetic acid),3,4 oxalic or acetic acid,5h9 and also precursors such as b-diketonates.10,11 Subsequent pyrolysis yields –nely divided oxide powders at relatively low temperatures. However, the grain size of these precursors has proved to be linked to the conductivity of potentially superconducting products.Large grain sizes of the metal oxides lead to inhomogeneities within the –nal product and to diminished conductivity.Thus, there is interest in exploration of other ligand systems whose thermal decomposition may lead to more –nely divided metal oxides. DCP (3,6- dicarboxylatopyridazine) is a tetradentate ligand that can be used as a precipitating reagent for a wide variety of elements. 12 Indeed, we have recently reported the preparation § Supplementary material available : infrared spectra of DCP and Available from BLDSC (No.SUP 57493, 4 pp.) See Pb(H2O)2DCP. Instructions for Authors, 1999, Issue 1 (http ://www.rsc.org/njc). and structures of and Cu2(H2O)2DCP Fe2(H2O)2DCP dihydrate13 and dihydrate.14 The structures Cu(H2O)2DCP of Pb(oxalate) dihydrate15 and Pb(oxalate) trihydrate16 have also been reported.We report here the preparation and structure of a new complex, the evaluation of its Pb(H2O)2(DCP), degradation process by thermal gravimetric analysis and a comparison, with that of of its per- Pb(H2O)2(oxalate), formance in the preparation of a mixed metal oxide. Experimental Synthesis A solution of 0.66 g (2 mmol) of in 10 ml of 2 N Pb(NO3)2 aqueous nitric acid and a solution of 0.4 g (2.2 mmol) of 3,6- dicarboxypyridazine monohydrate17 (3,6- in 10 H2DCP…H2O) ml of 2 N aqueous nitric acid were mixed together under stirring, thus producing a white precipitate.The mixture was maintained at room temperature for 1 h. The precipitate was –ltered oÜ, washed successively with hot water and diethyl ether. As a result of thermogravimetric analysis, the collected compound can be formulated as Pb(H2O)2DCP, New J.Chem., 1999, 23, 393»396 393Table 1 Crystal data for Pb(H2O)2DCP Formula C6H6N2O6Pb M 409.32 Space group I41/a a\b/Aé 15.278(2) c/Aé 15.218(3) a, b, c/° 90.0 U/Aé 3 3552(1) Z 16 Dmeas/g cm~3 3.03(1) Dcalc/g cm~3 3.062 F(000) 2976 Crystal dim./mm 0.044]0.044]0.33 l(MoK a )/mm~1 19.011 k(MoK a )/Aé 0.71069 R(int) 0.0561 Param.re–ned 137 R, wR 0.0357, 0.0549 Ext. coef. 0.00048(3) Ext. exp. (x) Fc*\kFc[1]0.001xFc2k3/sin(2h)]~1@4 Re—ns. meas. 4892 Re—ns. [I[2r(I)] 1316 Weighting scheme w\1/[p2(Fo2)](0.0163P)2]16.9366P] where P\(Fo2]2Fc2)/3 h, k, l [14, to 17; [14 to 18; [18 to 14 h range/° 2.67»25.09 yield : 96%. The compound is insoluble in most usual organic solvents. Thermal analysis Thermal gravimetric analysis was carried out on a 951 Dupont thermobalance coupled to a 1090B thermal analyzer and using either 30 or 60 °C/h as the heating rate.X-Ray data were measured using monochromated CuKa(k\1.541 78 Aé ) radiation and a Guinier»De WolÜ camera (room temperature) or a Guinier»Lenneç moving –lm camera (high temperature, 40 h exposure, –lm speed 2 mm h~1). Single crystal X-ray diÜraction A needle-shaped single crystal of was obtained Pb(H2O)2DCP by slow evaporation of a solution of the corresponding powder in nitric acid.Data collection was accomplished using a Nonius CAD4 equipped with monochromated radi- MoKa ation at room temperature. The resulting data were corrected for Lorentz, polarization, background and absorption eÜects. The choice of a space group presented problems. The unit cell dimensions allowed a choice between tetragonal [I41/a, Z\16, U\3552(1) asymmetric Aé 3, unit\Pb(H2O)2DCP], monoclinic Mnon-standard space groups Z\8, I21/c\I2/a, Table 2 Selected bond distances and angles (°) for (Aé ) Pb(H2O)2DCP Pb1wO3 2.532(6) Pb1wO1” 2.668(6) Pb1wO1@ 2.558(6) Pb1wN2@ 2.675(7) Pb1wN3 2.651(7) Pb1wO31 2.694(7) Pb1wO3A 2.664(6) Pb1wO32 2.781(7) O3wPb1wO1@ 112.22(19) O1@wPb1wN2@ 60.5(2) O3wPb1wN3 61.6(2) N3wPb1wN2@ 85.7(2) O1@wPb1wN3 71.1(2) O3AwPb1wN2@ 138.48(19) O3wPb1wO3A 121.78(15) O1”wPb1wN2@ 71.48(19) O1@wPb1wO3A 79.30(17) O3wPb1wO31 76.7(2) N3wPb1wO3A 70.7(2) O1@wPb1wO31 153.2(2) O3wPb1wO1” 79.68(18) N3wPb1wO31 93.6(2) O1@wPb1wO1” 119.90(15) O3AwPb1wO31 74.8(2) N3wPb1wO1” 140.1(2) O1”wPb1wO31 86.0(2) O3AwPb1wO1” 145.51(18) N2@wPb1wO31 142.3(2) O3wPb1wN2@ 69.9(2) Symmetry operations @:y, [x]1/2, z]1/4.A:[y]1/2, x, z[1/4. ”: [x]1/2, [y]1/2, z]1/2. U\3552(1) asymmetric or tri- Aé 3, unit\[Pb(H2O)2DCP]2N, clinic Z\4, U\1776.ymmetric unit\ MP1 6 , Aé 3, space groups. While the structure was [Pb(H2O)2DCP]2N originally solved in the triclinic and later in the mono- P1 6 , clinic and tetragonal groups, the pseudo symmetry between the two Pb atoms in the groups of lower symmetry is incorporated into the space group symmetry of Furthermore, I41/a.the l absence, l\4n, which was apparent in the data, is required only in that space group. Thus, the crystallographic results in space group are reported. Solution and re–ne- I41/a ment (SHELXS,18 SHELXL19) converged to a –nal R factor of 3.57% (Table 1).Selected bond angles and distances, based on the –nal positional parameters, are given in Table 2. CCDC reference number 440/097. See http ://www.rsc.org/ suppdata/nj/1999/393/ for crystallographic –les in .cif format. Results and discussion Thermal analysis The compound formulation results from Pb(H2O)2(DCP) thermal gravimetric analysis (anal.Pb: 0.957 ; 1.95 ; H2O DCP: 0.974). The TGA scan (Fig. 1) clearly indicates successive weight losses leading to a –nal residue identi–ed as pure PbO. If the heating rate is rapid (60 °C/h), one sees a single step at 50 °C corresponding to the loss of two moles of water (1.95 exp. ; 8.5% of the initial sample weight) per mole of PbDCP. With a slower heating rate (30 °C/h), dehydration of the initial organometallic complex proceeds in two steps ; the –rst at 40 °C is equivalent to the loss of 1.5 per H2O molecular formula and a second one at 60 °C, corresponds to the loss of 0.5 mole of water.Between 230 and 390 °C, the ligand decomposes and a loss of about 1 mole of DCP per Pb atom (exp. 0.974 mole) is seen. Analysis of the powder diÜraction spectrum measured at increasing temperatures allows a more detailed analysis. Between room temperature and 40 °C, the pattern is typical of the starting sample, which can be formulated as Between 40 and 60 °C, a diÜerent pattern is Pb(H2O)2(DCP).observed. According to the TGA, this pattern must correspond to Above 60 °C further modi–cation Pb(DCP) … 0.5H2O. in the powder spectrum corresponds to the completely dehydrated material, Pb(DCP).At 390 °C, decomposition of the organic ligand yields a mixture of PbO (massicot,20 yellow, orthorhombic) and PbO (litharge,21 red, tetragonal). This mixture of lead oxides undergoes oxidation in air at 445 °C to yield the typical diÜraction pattern of Pb3O422 [PbIVPb2IIO4], which disappears at 530 °C, leaving the diÜraction pattern of pure PbO (massicot form).However, observation of the evolution of the powder pattern under more controlled conditions reveals the existence of an additional phase between the mixture of massicot and Fig. 1 TGA scan of Pb(H2O)2(DCP). 394 New J. Chem., 1999, 23, 393»396litharge phases and the oxidized product, High tem- Pb3O4 . perature powder diÜraction, that is heating from 200 to 300 °C over a ten-hour period, followed by 30 h at 300 °C, shows the formation of PbO (both massicot and litharge forms) at temperatures as low as 270 °C.Then, after 5 h at 300°C in air, PbO undergoes oxidation to give PbO1.44 .23 Continual observation of the powder diÜraction pattern of a sample heated directly at 300 °C (temperature at which decomposition of the ligand begins) for 10 h leads to the identi –cation of as the solid residue at room temperature. PbO1.44 The compound has also been noted in the PbO1.44h1.57 thermal decomposition scheme of Pb oxalate.24 The cubic unit cell dimensions reported there agree with those observed in this work.The decomposition sequence observed for is in PbO1.44 agreement with literature observations that the red, tetragonal form (litharge) is of greater thermodynamic stability than the yellow, orthorhombic (massicot) form.Literature reports also suggest that the yellow massicot form converts to the red form, which is then oxidized to Pb3O4 . Crystal structure Columns of Pb atoms, [Pb… … …Pb separation, 3.994(4) Aé ] extend parallel to the c axis of the cell. The arrangement of Pb atoms is only approximately linear.Each Pb atom is bound to a nitrogen atom [PbwN 2.651(7)»2.675(7) to the ortho Aé ], carboxylate oxygen atom [PbwO 2.532(6)»2.668(6) of two Aé ] diÜerent ligand molecules, to the carboxylate oxygen atom of two additional organic molecules and to two water molecules [PbwO, 2.694(7) and 2.781(7) (Fig. 2). The Aé ] geometry of the eight-fold coordination is irregular.Each successive pair of lead atoms is bridged by a PbwNwNwPb link and by two bridging oxygen atoms. Carboxylate groups are coplanar with the aromatic ring (ligand dev: 0.045) and successive ligand planes subtend an interplanar angle of 87.7°. A projection view of the structure (Fig. 3) shows the polymeric units packed in a parallel array without signi–cant channels between columns.Columns are held in close proximity by networks of hydrogen bonds involving the water molecules. Examination of the solid state structures of and of the red and yellow forms of PbO and Pb(H2O)2(DCP) of (the single crystal structure of is Pb3O4 PbO1.44 unreported) in an eÜort to discover a structural progression during decomposition was unfruitful.Evaluation of performance as a synthetic precursor Synthesis of Stoichiometric quantities of Bi12PbO19 . Bi2O3 and PbO [PbO/(PbO] were dissolved in a Bi2O3)\0.143] minimum volume of concentrated nitric acid to give the nitrates of the two metals. The pH of the solution was adjust- Fig. 2 View of the Pb coordination (@: y, [x]1/2, z]1/4 ; A: [y]1/2, x, z[1/4 ; ”: [x]1/2, [y]1/2, z]1/2).Fig. 3 Projection view of on the 101 plane. Pb(H2O)2(DCP) ed with an appropriate quantity of distilled water to pH 1.40. At the same time, the stoichiometric quantity of ligand (metal : ligand\1 : 1, ligand\DCP) was dissolved in a minimum quantity of hot nitric acid. The two solutions were mixed and stirred for 12 h. On cooling, product appeared in the form of a white powder, which was dried using a rotary evaporator.The remaining material was then ground to a –ne powder and dried in a dessicator under vacuum. A similar preparation process was carried out using oxalic acid and the performance of the ligands was evaluated. Thermal decomposition of mixed metal ligand complexes. The thermal decomposition of the mixed metal complexes leading to formation of was followed both by Bi12PbO19 TGA and by powder diÜraction under changing thermal conditions.Both methods were in agreement on the progress of the decomposition. The powder pattern of the bismuth»lead oxalate [Fig. 4(a)] shows a more complicated evolution in three steps : between room temperature and 335 °C dehydration and ligand decomposition take place ; between 335 and 488 °C, tetragonal Bi2O3 is formed; at 488 °C (i.e.about 50 °C above the corresponding Fig. 4 Evolution of powder diÜraction pattern for (a) (Bi,Pb)oxalate and (b) (Bi,Pb)DCP with temperature. \ and indicate Au and Pb L grid diÜraction lines, respectively. New J. Chem., 1999, 23, 393»396 395Fig. 5 Electron microscope photographs of prepared Bi12PbO19 from (a) oxalate and (b) DCP precursors. temperature for the DCP compound) the sillenite type oxide, is seen in the presence of Bi12PbO19, Bi2 O3 .The corresponding DCP complex undergoes a simpler decomposition process : ligand decomposition [Fig. 4(b)] occurs up to 443 °C; appears along with Bi12PbO19 Bi2O3 above that temperature. Electron microscopy. The grain size of the products resulting from decomposition of (Bi,Pb)oxalate or (Bi,Pb)DCP was evaluated using a scanning electron microscope.In both cases the grain size was smaller than that observed for products prepared by solid state fusion methods. Moreover, the grain size was observed to be inversely related to the mass of the ligand. The prepared using the oxalate precur- Bi12PbO19 sor [Fig. 5(a)] displayed a larger grain size than that prepared using mixed metal DCP [Fig. 5(b)]. Conclusions A new compound, lead(II) diaqua 3,6-dicarboxylatopyridazine, has been synthesized and structurally characterized by single crystal X-ray diÜraction methods. The structure consists of polymeric chains of Pb atoms bridged in successive pairs by DCP molecules. These adopt a spiral conformation about the polymeric chain.The irregular coordination sphere of Pb is completed by coordination to two water molecules. Under dynamic conditions, thermal decomposition of the compound progresses via a mixture of the massicot and litharge forms of PbO, to and –nally shows complete conversion to the Pb3O4 , yellow or massicot form. prepared by thermal decomposition of precursor Bil2PbOl9 DCP complexes is seen to be prepared at lower temperatures and with a smaller grain size than that prepared by thermal decomposition of oxalate precursors.Thus, DCP is most promising as a precursor for synthesis of lead-based metal compounds. Acknowledgements acknowledges the support of the IMCC EC/US E.M.H. consortium. References 1 J. Garcia-Jaca, J. L. Pizarro, J. I. Larramendi, L. Lozama, M.I. Arriortua and T. Rojo, J. Coord. Chem., 1993, 30, 327. 2 J. Garcia-Jaca, J. L. Pizzaro, A. Gotti and M. I. Arriortua, J. Mater. Chem., 1995, 5, 227. 3 E. Escriva, D. Beltran and J. Beltran, An. Quim., 1980, 77, 330. 4 M. Insauti, J. L. Pizzaro, L. Lezema, R. Cortes, E. H. Bocanegra, M. I. Arriortua and T. Rojo, Chem. Mater., 1994, 6, 707. 5 K. Vidysager, J. Gopalakrishnan and C.N. R. Rao, Inorg. Chem., 1984, 23, 1206. 6 A. Rousset, J. Mater. Sci., 1986, 21, 3111. 7 D. Dollimore, T hermochim. Acta, 1992, 198, 395. 8 N. Deb, N. N. Dass and P. K. Gogoi, T hermochim. Acta, 1992, 198, 395. 9 A. Otsuki and K. Ogi, Jap. Pat., JP98195086 A2 980728, Kokai T okkyo Koho, 1997. 10 M. A. Crouch and T. C. DeVore, Chem. Mater., 1996, 8, 32. 11 Q. Zhou, J. Zhang, L.Zhang and X. Yao, in Proceedings of the 9th IEEE International Symposium on Applied Ferroelectronics, 1994, p. 419. 12 P. Con—ant, M. Drache, J. P. Wignacourt and M. Lagreneç e, J. Alloys Compa., 1992 188, 165. 13 B. Mernari, F. Abraham, M. Lagrenee and C. Bremard, New J. Chem., submitted. 14 S. Sobanska, M. Lagreneç e, J. P. Wignacourt and E. M. Holt, 1998, Acta Crystallogr., Sect. C, in the press. 15 A. V. Virovets, D. Y. Naumov, E. V. Boldyreva and N. V. Podberezska, Acta Crystallogr., Sect. C, 1993, 49, 1882. 16 S. Huang, R. Wang and T. C. W. Mak, J. Crystallogr. Spectrosc. Res., 1990, 20, 1029. 17 S. Sueur, M. Lagreneç e, F. Abraham and C. Bremard, J. Hetrocycl. Chem., 1987, 24, 1285. 18 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467. 19 G. M. Sheldrick, SHEL XL Program for the Re–nement of Crystal Structures, University of Goé ttingen, Germany, 1997. 20 Powder DiÜraction File No 38-1477, International Centre for Diffraction Data, Newton Square, PA, USA. 21 Powder DiÜraction File No 05-0561, International Centre for Diffraction Data, Newton Square, PA, USA. 22 Powder DiÜraction File No 41-1493, International Centre for Diffraction Data, Newton Square, PA, USA. 23 Powder DiÜraction File No 27-1201, International Centre for Diffraction Data, Newton Square, PA, USA. 24 P. Pascal, Nouveau de Chimie Masson, Paris, T raiteç Mineç rale, 1963, vol. VIII, p. 767. Paper 9/00679F 396 New J. Chem., 1999, 23, 393»396

ISSN:1144-0546    

DOI:10.1039/a900679f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Nanoporous three-dimensional networks topologically related to Cooperite from the self-assembly of copper(I) centres and the ììsquare-planarœœ building block 1,2,4,5-tetracyanobenzene Lucia Carlucci, Gianfranco Ciani,* Dorothea W.v. Gudenberg and Davide M. Proserpio Dipartimento di Chimica Strutturale e Stereochimica Inorganica and Centro CNR, V ia G. V enezian 21, 20133 Milano, Italy. Received (in Strasbourg, France) 23rd November 1998, Accepted 16th February 1999 The deliberate assembly of a CuI 3D coordination network with the Cooperite (PtS) topology, namely [Cu(l4- 1, has been accomplished by using the pseudo-square-planar building block 1,2,4,5- TCNB)][PF6] tetracyanobenzene (TCNB).It has been obtained by reacting with TCNB (in molar ratio 1 : 1) [Cu(NCMe)4][PF6] in solvent mixtures of and of a poorly basic solvent, like ethanol, isopropanol, benzyl alcohol or acetone.CH2Cl2 The crystals are orthorhombic, space group Cccm, with a\8.157(1), b\18.634(2), c\11.613(1) Z\4. The Aé , framework is based on tetrahedral CuI centres and tetradentate ligands and presents cavities containing l4-TCNB guest solvent molecules. On performing the same reaction in a 3D three-connected polymeric CH2Cl2»THF species, 2, is obtained.Compound 2 gives monoclinic crystals, space group with [Cu(l3-TCNB)(THF)][PF6] P21/c, a\21.617(2), b\11.127(3), c\16.294(2) b\111.61(2)°, Z\8. In 2 the TCNB ligands are tridentate and the Aé , copper centres achieve a tetrahedral coordination on forming a bond with a THF molecule. The network presents an unprecedented topology, (8210)-b, and the crystals undergo, by thermal activation, an unusual crystal-to-crystal transformation, losing the THF molecules to give the same frame as 1.Many eÜorts in the crystal engineering of networked materials based on polymeric coordination compounds1,2 have been devoted to the use of novel suitably tailored polydentate ligands. Polycyano ligands, neutral and anionic, are attracting a growing interest,3 both for their electronic properties and also for their diÜerent bonding modes and variable hapticity.In particular, some tetracyano species like TCNE (tetracyanoethene) and TCNQ (7,7,8,8-tetracyanoquinodimethane) can potentially act as planar l4-bridging groups, as, indeed, found in some previously reported polymeric species.4,5 The self-assembly of ligands like these with tetrahedral metal centres can produce open nanoporous frameworks related to the prototypical structure of Cooperite (PtS, Fig. 1, top), containing a 1 : 1 ratio of tetrahedral and square-planar nodes. Networked materials with this topology (4284),6 have been deliberately assembled by using tetrahedral CuI ions together with square-planar metal centres, like [Pt(CN)4]2~ 7 or metalloporphyrins,8 but also a suitable organic group can play the same role of a square-planar coordination centre, as found in the case of the two-fold interpenetrated PtS-like network of [Ag(TCNQ)],9 schematically shown in Fig. 1 (bottom). Our previous attempts to build a PtS-like net using TCNE together with AgI or CuI ions were unsuccessful, resulting in frames of diÜerent topology because of the transformation of the ligand, under the reaction conditions, into a non-planar carbanionic acetonyl derivative.10 We report here on the reactions of with 1,2,4,5-tetracyanobenzene [Cu(NCMe)4][PF6] (TCNB), leading to a Cooperite-like framework, [Cu(l4- 1, in solution mixtures of with diÜer- TCNB)][PF6] CH2 Cl2 ent poorly basic solvents (like acetone, ethanol, isopropanol, benzyl alcohol), and to a three-connected net of uncommon topology, 2, in [Cu(l3-TCNB)(THF)][PF6] CH2 Cl2»THF Fig. 1 The prototypal network of Cooperite (PtS) viewed down the [1 0 0] direction (top) and a schematic view of the two-fold interpenetrated PtS-like net of [Ag(TCNQ)] (bottom, black balls\baricentres of the TCNQ ligands). New J. Chem., 1999, 23, 397»401 397(THF\tetrahydrofuran). Compounds 1 and 2 can be interconverted and the noteworthy crystal-to-crystal transformation of 2 into 1 under thermal activation is also demonstrated.During the preparation of this manuscript another study on the reactivity of with [Cu(NCMe)4][PF6] TCNB has been reported by Munakata et al.,11 leading to quite diÜerent polymeric products, of composition due to the use of a large stoichiomet- [Cu2(TCNB)3][PF6]2 , ric excess of the TCNB ligand. Results and discussion Reactivity of with TCNB [Cu(NCMe)4 ] [PF6 ] The reactions of with TCNB were per- [Cu(NCMe)4][PF6] formed in diÜerent solvents using a 1 : 1 molar ratio of the reagents.The metal complex was dissolved in and CH2Cl2 treated with solutions of TCNB in ethanol, isopropanol, benzyl alcohol, acetone or THF.A yellow microcrystalline product is rapidly formed under stirring, which gives, after –ltration and drying, rather variable analyses corresponding approximately to the 1 : 1 adduct 1. Crys- [Cu(TCNB)][PF6] tals were grown in few days by slow diÜusion, upon carefully layering the solution of the ligand on that of the metal complex.Similar elongated yellow crystals were obtained from the acetone, ethanol, benzyl alcohol or isopropanol solutions, containing the same 3D polymeric network, as con–rmed by X-ray diÜraction analyses performed on crystals of all the samples. From the system, on the other hand, by CH2Cl2»THF slow diÜusion beautiful —attened hexagonal shaped orangeyellow crystals separated, corresponding to [Cu(l3- 2, containing a diÜerent 3D polymeric TCNB)(THF)][PF6] network, characterized by X-ray analysis.It is interesting to note that the reactions of with a large excess (–ve-fold) of TCNB [Cu(NCMe)4][PF6] give two diÜerent polymeric species on varying the solvent, i.e. in acetone and [Cu2(l2-TCNB)(l4-TCNB)2][PF6]2 Æ 4Me2CO in methylethyl ketone.11 [Cu2(l2-TCNB)(l4-TCNB)2][PF6]2 The former contains 2D undulated layers while the latter consists of a two-fold interpenetrated 3D net topologically related to the polymorph moganite.12 SiO2 The three-dimensional networks of 1 and 2 The crystal structure of 1 consists of a 3D net based on an equal number of planar tetradentate organic centres and distorted tetrahedral CuI ions [Cu»N 1.969(4) N»Cu»N Aé , 95.2(3)»117.2(2)°], with the desired PtS-like topology (see Fig. 2). The TCNB ligands (as well as TCNE and TCNQ) can be better described as ììrectangularœœ rather than square centres (with intraligand N… … …N long and short rectangular edges having a ratio of ca. 1.8 : 1). The channels of the network are occupied by the anions and by highly disordered PF6~ solvent molecules, which are difficult to rationalize (see below).In contrast to what was observed for the two-fold interpenetrated [Ag(TCNQ)] (Fig. 1, bottom), compound 1 contains a single net, very probably due both to the smaller dimensions of the TCNB ligand and to the presence of the counter ions. The deliberate contruction of a PtS-like framework, using proper building blocks, has thus been successfully carried out.The same topology has also been obtained by the selfassembly of TCNB and Signi–cantly, these net- AgSbF6 .13 works (both in the CuI and in the AgI species) are not interpenetrated. The structure of compound 2, [Cu(l3-TCNB)(THF)][PF6], reveals a diÜerent 3D frame. The TCNB ligands use only three of the four nitrile groups for bonding to CuI ions, which are connected to three such groups [Cu»N 1.915(7)»1.999(7) Aé ] and to the oxygen atom of a THF molecule [Cu»O 2.117(7) Fig. 2 A SCHAKAL perspective view of the framework of 1 down [1 1 0], oriented to show the PtS topology (cf. Fig. 1 top). Hatched spheres represent the P atoms of the anions. PF6~ and 2.119(7) in a distorted tetrahedral geometry. In this Aé ] case, in contrast to 1, the solvent molecules are able to enter into the metal coordination sphere, replacing a bound CN group.The coordination is illustrated in Fig. 3, which shows the shortest circuit in the net, consisting of alternated four CuI ions and four TCNB ligands. From a topological point of view, the 3D frame is comprised of three-connected metallic and organic centres in equal ratio, which form a system of alternating and intercon- 41 43 nected parallel helices, extending in the b direction. It was classi–ed by Wells6 as the Archimedian 3D net (8210)-b (see Fig. 4) and, to the best of our knowledge, no example of this type has been previously reported.In spite of the diÜerent topology, the structure of 2 displays a great resemblance to that of 1, as can be seen on comparing the projection of the frame of 1 down c with that of 2 down b (Fig. 5). This structural similarity can account for the crystalto- crystal transformation of the two species, that is described below. The nanoporous network of compound 1 When compound 1 is rapidly precipitated under stirring it gives polycrystalline samples whose X-ray powder diÜraction patterns are almost identical whichever solvent system is used; this is also the case for the mixture, if the pre- THF»CH2Cl2 cipitate is immediately removed from the mother-liquor.Single crystals are grown within a few days by slow mixing of the solution of the metal complex in and the solution CH2Cl2 of TCNB in the second solvent. All these single crystals give very similar cell parameters, with the exception of those obtained from (compound 2).Structure THF»CH2Cl2 Fig. 3 The shortest circuit in the network of 2, showing the triconnected TCNB ligands and the distorted tetrahedral coordination of the CuI ions. 398 New J. Chem., 1999, 23, 397»401Fig. 4 The ideal triconnected (8210)-b network (top) and the topologically equivalent schematized net of 2 (bottom), in which the black balls represent the baricenters of the TCNB ligands.analyses of crystals of 1 from diÜerent solvents, both at room temperature and at [30 °C, gave similar results, showing the presence of some variable residual electron density peaks due to highly disordered solvent molecules, that are difficult to rationalize on the basis of the X-ray data only, located within the cavities (of approximate volume 239 illustrated in Aé 3)14 Fig. 6 by big spheres. These solvent regions are disposed down [0 0 1], along rhombic channels topologically equivalent to the square ones down the tetragonal axis of PtS. DiÜerent experiments have been carried out in order to characterize the guest solvent molecules. IR microscopy on the crystals shows bands due to the presence of both and the second CH2Cl2 solvent, and, in the case of the crystals obtained from monitoring of the acetone l(CO) stretching CH2Cl2»acetone, reveals that this band rapidly decreases and disappears on heating the sample to 80 °C. 1H NMR spectra of diÜerent samples from collected after drying the crys- CH2Cl2»ethanol, tals under —ux and dissolution in show signals N2 acetone-d6 , at d 8.93 (2H, TCNB), 5.63 (2H, and 1.5 (t, 3H, CH2Cl2) CH3 ethanol), of variable relative intensities from (TCNB»CH2Cl2 1 : 1 to 1 : 033; TCNB»ethanol from 1 : 0.33 to 1 : 0.20).Ther- Fig. 5 A comparison of the networks in the structures of 1, down c (top) and of 2, down b (bottom). The dashed lines in the case of 2 represent the directions of Cu»N bond formation in the course of the 2]1 transformation (see text).mogravimetric analyses of 1 show a loss of solvent (not constant), not exceeding ca. 4% of the mass, in the range 80» 100 °C. The network is stable up to ca. 280 °C and above this temperature it loses the TCNB ligands. Almost constant elemental analyses for 1, corresponding to the solvent-free network, can be obtained after heating the samples in an oven at 80 °C for some hours.The X-ray powder diÜraction spectra of these samples show patterns quite similar to the original one, con–rming that the polymeric network is not destroyed on removing the solvents. Crystal-to-crystal 2«1 conversion Thermal analyses (DSC and TGA) have shown that 2 loses the coordinated THF molecule at 110»130 °C and starts to decompose above 280 °C. Examination under the microscope of the samples previously heated to ca. 140 °C and cooled at room temperature revealed that the macroscopic crystal shape and colour seem almost unchanged in the process. These crystals, submitted to single crystal X-ray diÜraction, showed marked broadening of the peaks and lack of diÜraction above h ca. 10°, but gave an orthorhombic cell [a\8.30(1), b\18.66(5), c\11.60(4) V \1797(8) similar to that of Aé , Aé 3] 1.The same conversion has been con–rmed on bulk samples New J. Chem., 1999, 23, 397»401 399Fig. 6 A view of the frame of 1 that illustrates the rhombic channels down c containing the regions (big spheres) occupied by the disordered solvent molecules. The anions are also shown with van PF6~ der Waals radii. Fig. 7 X-Ray powder diÜraction spectra showing the transformation of a polycrystalline sample of 2 (see top) left at 110 °C for one night to give 1 (see bottom).Each XRPD trace is in the 5\2h\30° range (Cu-Ka) ; observed (top) and calculated (bottom) patterns. Starred peaks arise from mutual sample contaminations. of 2 by monitoring the heating process by X-ray powder diffraction methods (Fig. 7). A crystal-to-crystal tranformation takes therefore place, leading from 2 to 1, with the cavities and channels along the c axis (see Fig. 5, top and Fig. 6) left empty. The process is made possible by the limited extent of the required rearrangements: the displacement of the THF molecules of 2, which can travel through the crystal channels, is accompanied by the formation of the fourth Cu»N bond (the relevant Cu… … …N contact varying from ca. 3.8 in 2 to ca 2.0 Aé in 1, along the directions illustrated by the dashed lines in Aé Fig. 5, bottom). So Fig. 5 gives the extreme situations of the solid-state displacement reaction path CN… … …Cu… … …THF. Crystal-to-crystal reactions are in general rare,15 particularly in coordination polymer chemistry.16,17 Some evidence of the reverse 1]2 solid state transformation has also been achieved: for instance, samples of 1 left in THF at room temperature for a week gave X-ray powder diÜraction patterns which can be ascribed to a ca. 50% mixture of 1 and 2. Experimental All manipulations were carried out under a nitrogen atmosphere with use of standard Schlenk techniques. Solvents were distilled under nitrogen from (THF) or Na»Ph2CO P2O5 was prepared and puri–ed (CH2Cl2).[Cu(NCMe)4][PF6] according to the literature method.18 All the other reagents were used as purchased from Aldrich. IR spectra in Nujol and IR microscopy data were collected on a Perkin»Elmer FT-IR Paragon 1000 spectrometer. 1H NMR spectra were recorded on a Bruker AC200 instrument. Thermal analyses were performed on DSC 7 and TGA 7 Perkin»Elmer instruments with a heating rate of 10 °C per min.X-Ray powder diÜraction spectra were collected on a Rigaku D/Max horizontal-scan diÜractometer. Elemental analyses were carried out at the Microanalysis Laboratory of this University. Synthesis of the compounds Compound 1. In a typical preparation a solution of TCNB (0.021 g, 0.12 mmol) in ethanol (5»6 mL) was added to a solution of (0.044 g, 0.12 mmol) in (8 [Cu(NCMe)4][PF6] CH2 Cl2 mL).The solution was left to stir for 1 h. A yellow precipitate of 1 was obtained, which was –ltered oÜ, washed with ethanol and then with hexane and dried under vacuum (ca. 0.025 g, yield 55%). Crystals were grown by slow diÜusion of a solution of TCNB in ethanol into a solution of the copper complex in The crystals are air stable for few days.CH2Cl2 . Constant elemental analyses, corresponding to the solvent-free product, were obtained after heating at 80 °C for some hours (Calc. for C 31.06, H 0.52, N 14.49. Found: C10H2CuF6N4P: C 30.94, H 0.75, N 14.10%. IR (Nujol, cm~1) : CH 3117, 3044, CN 2269, 832. PF6~ Compound 2. A solution of TCNB (0.028 g, 0.16 mmol) in THF (5 mL) was added to a solution of [Cu(NCMe)4][PF6] (0.058 g, 0.16 mmol) in (7»8 mL).After stirring for CH2Cl2 one night the orange-yellow precipitate of 2 was –ltered oÜ, washed with ethanol and then with hexane and dried under vacuum (ca. 0.052 g, yield 71%) (Calc. for C14H10CuF6N4OP: C 36.65, H 2.20, N 12.21. Found: C 36.12, H 2.06, N 11.96%. IR (Nujol, cm~1) : CH 3128, 3059; CN 2262, 2252; THF 1043, 833.PF6~ X-Ray crystallography Single crystals of 1 and 2 were mounted under a coating of cyanoacrylic glue, to prevent decomposition, on a Siemens SMART CCD area-detector diÜractometer. Crystal data for 1: M\386.67, orthorhombic, space group C10H2CuF6N4P, Cccm (no. 66), a\8.157(1), b\18.634(2), c\11.613(1), V \1765.1(3) Z\4, Mg m~3, Mo-Ka radi- Aé 3, Dc\1.455 400 New J. Chem., 1999, 23, 397»401ation (j\0.71073 k\1.381 mm~1.Crystal data for 2: Aé ), M\458.77, monoclinic, space group C14H10CuF6N4OP, (no. 14), a\21.617(2), b\11.127(3), c\16.294(2) P21/c Aé , b\111.61(2)°, V \3643.8(11) Z\8, Mg Aé 3, Dc\1.673 m~3, Mo-Ka radiation (j\0.71073 k\1.356 mm~1. The Aé ), data collections were performed at 243 K, within the limits 2\h\28° for 1 and at 293 K within the limits 2\h\26° for 2.The structures were solved by direct methods (SIR97)19 and re–ned by full-matrix least-squares (SHELX-97)20 against In both cases the anions were found disordered and a Fo2 . suitable doubled model [F atoms with 83»17% occupancy (1) and 50% (2)] was re–ned. Final residual peaks do not exceed 0.795 e for 1 and 0.440 e for 2. All the results here Aé ~3 Aé ~3 reported for 1 refer to a sample obtained from Three other single crystal analyses, based on CH2Cl2»ethanol.data collected at room temperature on samples obtained from diÜerent solvent systems, gave similar results with greater residual peaks around 2/m special positions which were not easily rationalized, indicative of disordered solvent molecules. The –nal agreement indexes were 0.0614 for 735 indepen- R1 dent signi–cant absorption corrected data for 1 [Fo[4p(Fo)] and 0.0647 for 3365 independent signi–cant [Fo[4p(Fo)] absorption corrected data for 2.Anisotropic thermal parameters were assigned to all the non-hydrogen atoms, but to the atoms of the minor component of the disordered anion in 1. All the drawings were produced with SCHAKAL 97.21 CCDC reference number 440/103.Notes and references 1 (a) B. F. Hoskins and R. Robson, J. Am. Chem. Soc., 1990, 112, 1546; (b) R. Robson, B. F. Abrahams, S. R. Batten, R. W. Gable, B. F. Hoskins and J. Liu, Supramolecular Architecture, ACS publications, Washington DC, 1992, ch. 19; (c) S. R. Batten and R. Robson, Angew. Chem., Int. Ed., 1998, 37, 1461. 2 For some recent reviews see : M.J. Zaworotko, Chem. Soc. Rev., 1994, 283; C. L. Bowes and G. A. Ozin, Adv. Mater., 1996, 8, 13; O. M. Yaghi, H. Li, C. Davis, D. Richardson and T. L. Groy, Acc. Chem. Res., 1998, 31, 474. 3 See for example, K. R. Dunbar, Angew. Chem., Int. Ed. Engl., 1996, 35, 1619 and refs. therein. 4 F. A. Cotton and Y. Kim, J. Am. Chem. Soc., 1993, 113, 8511. 5 C. Campana, K. R. Dunbar and X.Ouyang, Chem. Commun., 1996, 2427. 6 A. F. Wells, T hree-dimensional Nets and Polyhedra, Wiley, New York, 1977. 7 R. W. Gable, B. F. Hoskins and R. Robson, J. Chem. Soc., Chem. Commun., 1990, 762. 8 B. F. Abrahams, B. F. Hoskins, D. M. Michail and R. Robson, Nature (L ondon), 1994, 369, 727. 9 L. Shields, J. Chem. Soc., Faraday T rans., 1985, 81, 1. A rationalization of the topology of this species is given in ref. 1(c). 10 L. Carlucci, G. Ciani, D. M. Proserpio and A. Sironi, Angew. Chem., Int. Ed. Engl., 1996, 35, 1088. 11 M. Munakata, G. L. Ning, T. Kuroda-Sowa, M. Maekawa, Y. Suenaga and T. Horino, Inorg. Chem., 1998, 37, 5651. 12 We have rationalized the topology of this binodal network, that has a symbol (426282)(4864). This is the same topology of the ììhoneycombœœ frame described in ref. 16, and of a polymorph of called moganite (see M. OœKeeÜe and B. G. Hyde, Crystal SiO2 structures. I. Patterns and symmetry, Mineralogical Society of America, Washington, DC, 1996). 13 L. Carlucci, G. Ciani and D. M. Proserpio, unpublished work. 14 Obtained by using the PLATON program: A. L. Spek, PLATON, A Multipurpose Crystallographic Tool, Utrecht University, 1998. 15 See for example, T. Wagner, U. Eigendorf, G. E. Herberich and U. Englert, Struct. Chem., 1994, 5, 233 and refs. therein. 16 B. F. Abrahams, M. J. Hardie, B. F. Hoskins, R. Robson and G. A. Williams, J. Am. Chem. Soc., 1992, 114, 10641. 17 A photo-induced solid state diacetylene polymerization has been observed in the networked polymer [L\1,4-bis(4- [Cd(CN)2L] pyridyl)butadiyne)], but the photo-product has not been characcterized : B. F. Abrahams, M. J. Hardie, B. F. Hoskins, R. Robson and E. E. Sutherland, J. Chem. Soc., Chem. Commun., 1994, 1049. 18 G. J. Kubas, B. Nozyk and A. L. Crumbliss, Inorg. Synth., 1979, 2, 90. 19 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C. Burla, G. Polidori and M. Camalli, J. Appl. Crystallogr., 1994, 27, 435. 20 G. M. Sheldrick, SHELX-97, Program for structure re–nement, University of Goé ttingen, 1997. 21 E. Keller, SCHAKAL 97, A computer program for the graphical representation of crystallographic models, University of Freiburg, 1997. Paper 8/09144G New J. Chem., 1999, 23, 397»401 401

ISSN:1144-0546    

DOI:10.1039/a809144g

出版商:RSC    

年代:1999

数据来源: RSC