摘要:

DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1998, 3893–3899 3893 Francis Aston and the mass spectrograph Gordon Squires Cavendish Laboratory, University of Cambridge, Madingley Road, Cambridge, UK CB3 0HE Received 18th June 1998, Accepted 28th July 1998 The chemical determination of atomic weights gives the average weight for an aggregate of a large number of atoms. Although this is useful in many applications, the determination of the masses of individual atoms gives further important information, in particular the stability of the atoms or more precisely of their nuclei.The first accurate determination of the masses of individual atoms was made by Aston in 1919. His measurements demonstrated the existence of isotopes in non-radioactive elements and paved the way for our present picture of the nuclear atom. Early life Francis William Aston was born on 1 September 1877 at Harborne, Birmingham. He was the third child of a family of seven. His father and paternal grandfather were metal merchants and farmers, and Francis was brought up on a small farm.From an early age he showed a keen interest in mechanical toys and scientific apparatus. He had a ‘laboratory’ over a stable and amused his sisters with home-made fireworks and large tissue-paper hot-air balloons. These were dispatched with stamped addressed postcards, which were sometimes returned from great distances.1 Aston entered Malvern College in September 1891 and two years later went to Mason College (which subsequently became the University of Birmingham), where he studied chemistry and physics.The professor in physics was John Poynting (of Poynting’s vector). While at Birmingham Aston acquired skill with tools and glass-blowing which proved invaluable in his later work. Faced with the need to earn a living after graduating he took a course in fermentation chemistry and in 1900 started work in a brewery in Wolverhampton. In his spare time he experimented at home, designing and building new forms of Sprengel and Tœpler vacuum pumps. This experience was again to stand him in good stead later on.In 1903 Aston returned to Birmingham University and physics. He worked on the properties of electrical discharges in Gordon Squires is a retired lecturer in the Department of Physics at the University of Cambridge, and his current interest is the history of the Cavendish Laboratory. His field of research is the scattering of thermal neutrons.He is the author of textbooks on practical physics, quantum mechanics, and the theory of thermal neutron scattering. Gordon Squires gases and measured the variation of the length of the Crookes dark space with current and pressure. In 1908 his father died, and he used a legacy to travel round the world. On his return he was appointed a lecturer in Birmingham University, but after one term he received an invitation from Joseph (J. J.) Thomson to come to the Cavendish Laboratory at Cambridge as his assistant.Poynting, a close friend of Thomson, had recognised Aston’s great gifts as an experimenter and recommended him for the post. Aston accepted the invitation and thus began a career that had momentous consequences for chemistry and nuclear physics. Historical background to Aston’s work To appreciate the significance of the work Thomson was doing and Aston’s subsequent role we need to go back in the history of chemistry.In 1803 John Dalton put forward an atomic theory, which laid the foundations of modern chemistry. One of the postulates was that atoms of the same element are similar to one another and equal in weight. About ten years later William Prout suggested that the atoms of the elements were made up of aggregates of hydrogen atoms. If this were true the weights of atoms would be expressed as whole numbers, i.e. integers, and, on the basis of Dalton’s postulate that all the atoms of an element had the same weight, atomic weights would also be whole numbers.However, experiment showed that although the atomic weights of many of the elements were whole numbers, far more than could be attributed to chance, there were a few, for example, magnesium, atomic weight 24.3, and chlorine, atomic weight 35.5, which were not. Therefore, Dalton and Prout could not both be correct, and around 1900 it was Dalton’s rather than Prout’s hypothesis that was accepted. In 1896 Henri Becquerel discovered radioactivity, and from then until the outbreak of the first World War many radioactive substances were found.An interesting feature was that two and sometimes three of the substances with quite diVerent modes of decay appeared to be chemically similar. For example, in 1906 Bertram Boltwood found that once salts of thorium and ionium were mixed they could not be separated by any chemical means.2 Another example was radium B and lead; not only were their chemical properties the same, but Ernest Rutherford and Edward Andrade found that they had identical X-ray spectra.3 In 1913 Frederick Soddy 4 proposed the word isotopes to describe these chemically similar materials, because they occupy the same place in the Periodic Table of the elements.He observed ‘They are chemically identical, and, save only as regards the relatively few physical properties which depend upon atomic mass directly, physically identical too’. Thomson’s work on positive rays In 1886 Eugen Goldstein was investigating the properties of the electric discharge obtained when a large voltage is applied across a pair of electrodes in a vessel containing a gas at low pressure.He found that if a channel or canal was cut through the cathode a beam of light appeared on the side remote from the anode. He called the beam Kanalstrahlen, canal rays.5 In3894 J. Chem. Soc., Dalton Trans., 1998, 3893–3899 1898 Willy Wien 6 managed to deflect the beam with a strong magnetic field, in a direction which showed it was due to a stream of positively charged particles.They are in fact the positive ions resulting from the atoms in the gas that have lost one or more electrons. In 1907 Thomson7 started to investigate the positive rays. He measured the mass of the particles by deflecting the rays with electric and magnetic fields. His apparatus was the forerunner of Aston’s mass spectrograph and it is instructive to consider it first.The essentials of the apparatus are shown in Fig. 1. The discharge occurs in the spherical tube T. The anode A is located in a side arm, and the positive rays pass through the cathode C, which is a fine tube. The rays then pass between the poles N and S of an electromagnet, the pole pieces P1, P2 of which are insulated from the magnet by thin sheets of mica. By this means a potential diVerence may be applied across the pole pieces, giving an electric field E in the same direction as the magnetic field B, this direction being at right angles to the path of the particles.The particles finally strike the screen H, where they produce a fluorescent spot. In the absence of the two fields the particles travel in a straight line, and the spot is in the centre of the screen in line with the fine tube in the cathode. Take a set of right-handed axes x, y, z, with the initial direction of the particles as the z axis, and the common direction of E and B as the x axis.We consider the electric and magnetic deflections separately. The deflection produced by the electric field is shown by the diagram in Fig. 2. The particles coming from the left with velocity v enter the region between the pole pieces P1 and P2, across which a potential diVerence V is applied. If the pole pieces are a distance d apart this gives an electric field E = V/d, which causes an acceleration eE/m, where e is the charge and m the mass of the particles.If lE is the length of the plates, the particles spend an approximate time lE/u between the plates, and when they emerge from the plates they have acquired a component of velocity in the x direction given by ux = eElE/mu. (1) Since ux ! u, the angle through which the particles are deflected by the field is approximately q = ux u = eElE mu2 . (2) A magnetic field B whose direction is at right angles to the path of the particles deflects them into a circular path of radius R as shown in Fig. 3. The force due to B is Beu, and its direction is at right angles to the directions of both B and v. The acceleration in the circular path is u2/R. Thus mu2/R = Beu, i.e. mu = BeR. (3) Fig. 1 Diagram of Thomson’s positive ray apparatus. Fig. 2 The deflection of positively charged particles by an electric field. The particles are deflected in the y direction through an angle f = lB R = eBlB mu , (4) where lB is the length of the path in the magnetic field.Now let E and B act together. The screen H, which contains the x and y axes, is shown in Fig. 4, with the position O of the spot for the undeflected beam as the origin. The field E deflects the particles in the x direction by an amount proportional to the angle q, while B deflects them in the y direction by an amount proportional to the angle f, both the angles being small. The coordinates of the spot are therefore x = c1 eE mu2 , y = c2 eB mu , (5) where c1 and c2 are constants depending on the geometry of the apparatus.Eliminating the velocity u between these two expressions gives y2 x = c3 e m B2 E , (6) where the constant c3 depends on the geometry of the apparatus. Thus, for a beam of ions with the same value of e/m and varying velocities, the pattern on the screen is a parabola, Fig. 4. Particles with diVerent velocities arrive at diVerent points on the parabola. If ions with diVerent masses are present there will be several parabolas corresponding to the diVerent e/m values.The value of e is the electronic charge, 1.60 × 10219 C, or a simple multiple of it. For singly charged ions the y value of the parabola at a constant value of x is proportional to 1/÷m. So the ratio of two masses is given by the square of the inverse ratio of the two y values at the same value of x. This is independent of the form of the apparatus and of the values of E and B. If an atom of mass m loses two electrons in the discharge tube, the doubly charged ion appears on a parabola corresponding to a singly charged ion of mass m/2.A photographic record is made of the Fig. 3 The deflection of positively charged particles by a magnetic field. The direction of the field is down, perpendicular to the plane of the diagram. Fig. 4 Parabolas obtained with Thomson’s positive ray apparatus. The ratio of the masses of the particles for the two parabolas is given by m1/m2 = (Y2Y29/Y1Y19)2. The parabolas with negative y are obtained by reversing the magnetic field.J.Chem. Soc., Dalton Trans., 1998, 3893–3899 3895 traces. If the mass is known for one of the parabolas, measurements of the y values at constant x give the values of all the other masses. The axis Ox is not marked on the photograph. The magnetic field is therefore reversed for the second half of the exposure, which puts the pattern in the 2y region and allows the y values to be measured. When Aston arrived at the Cavendish Laboratory in 1909 Thomson’s positive ray apparatus was already working.With his assistance the apparatus was greatly improved, and by 1912 parabolas corresponding to mass diVerences of 10% could be resolved. In November of that year some gas containing neon was analysed. The photograph, Fig. 5, showed a strong parabola corresponding to a mass of 20 (on the scale oxygen = 16) and a much weaker one at a mass of 22.8 Various possibilities for the 22 parabola were considered. One was that it was due to doubly charged CO2.However, when the gas was passed through liquid air, the parabola at 44, due to singly charged CO2, disappeared, while the one at 22 was not aVected. Another speculation was that the 22 trace was due to a compound NeH2. From density measurements the atomic weight of neon was known to be 20.2. So the novel, and at the time revolutionary, suggestion was made that neon could exist in two forms, which were isotopes, just like the isotopes suggested by Soddy in radioactive elements.If the isotope of mass 20 was 9 times more abundant than the one of mass 22, that would give the measured atomic weight of 20.2. In other words, neon did not consist of identical atoms of mass 20.2, but of two diVerent atoms of mass 20 and 22, in line with Prout’s hypothesis. Aston set to work to see if he could separate the two constituents of neon. He first tried fractional distillation, but without success. He then tried diVusion through fine pores, using clay tobacco pipes, and after much labour obtained a small eVect.9 Then, in his own words ‘the whole of the lightest fraction was lost by a most unfortunate accident’.(It is said that he dropped the flask containing the specimen!) However, undeterred, he carried on with the heaviest fraction and ultimately obtained two samples with densities 20.15 and 20.28 on Fig. 5 Positive ray parabolas of neon obtained by Thomson in 1912. the scale O2 = 32.These results were just on the borderline of the experimental uncertainty. The work was interrupted by the first World War. Aston was sent to the Royal Aircraft Factory, later the Royal Aircraft Establishment, at Farnborough. Frederick Lindemann, later Lord Cherwell, and George Thomson (J. J. Thomson’s son) were also there. In after years Thomson recollected that Lindemann was sceptical of Aston’s isotope hypothesis, preferring the idea of CO2 or NeH2 for the 22 parabola.10 He said that Lindemann was a much better theoretician than Aston and always won the argument, but Aston ‘had faith and next morning was still of the same opinion’.In 1914 Aston crashed in an experimental aeroplane, but escaped unhurt. He worked at Farnborough as a chemist, studying among other things the properties of the doped canvas with which aeroplanes were then covered. Aston’s first mass spectrograph After the war Aston returned to the Cavendish Laboratory. While at Farnborough he had meditated on an improved form of the apparatus to measure the masses of the positive ions, and in 1919 he built his first mass spectrograph.11 Like Thomson’s parabola apparatus it employed electric and magnetic fields to deflect the particles, but the two fields were in diVerent regions along the path of the particles.Unlike Thomson’s apparatus in which particles with the same e/m value, but diVerent velocities, were distributed along the parabola, in Aston’s spectrograph these particles were focused to the same point on the screen. This was a big advantage.The focused beam was much more intense, thus permitting finer slits to be used, which improved the resolution and accuracy of the instrument. The principle of the instrument is illustrated in Fig. 6. The path of the positive particles emerging from the discharge tube is defined by a pair of narrow slits S1 and S2. The particles then pass between a pair of plates P1 and P2 across which a potential diVerence is applied.The particles are deflected downwards by the electric field towards the negative plate P2. They are deflected continuously in the region between the plates, but as a first approximation we may assume that the paths come from a point Z in the middle of the plates on the line defined by S1 and S2. A group of the rays is allowed to pass through a narrow diaphragm D, which selects those deflected through angles between q and q 1 dq. They then pass between the poles of an electromagnet which has its north pole above the plane of the diagram.This deflects the particles in the opposite direction to that of the electric field. The same notation is used as in the discussion of Thomson’s apparatus. Eqns. (2) and (4) still apply. For particles of velocity v, charge e and mass m, the electric field E gives a deflection q, and the magnetic field B gives a deflection f. The position of the diaphragm D fixes the angle q, and hence the velocity of the particles passing through.The spread dq in q gives rise to a spread du in u, which in turn gives a spread df in the deflection produced by the magnetic field. The relations between dq, du and df are obtained from eqns. (2) and (4). For a constant value Fig. 6 The paths of the particles in Aston’s mass spectrograph. The particles have the same mass value but varying velocities. The path of the fastest particles is shown in blue and that of the slowest in red.For clarity the electric and magnetic deflections are shown as abrupt changes in direction, rather than the actual continuous changes shown in Figs. 2 and 3.3896 J. Chem. Soc., Dalton Trans., 1998, 3893–3899 of e/m, q is proportional to 1/u2, and f is proportional to 1/u. Therefore dq q = 22 du u , df f = 2 du u , (7) whence df/dq = f/2q. (8) The minus signs in eqns. (7) indicate that the faster particles, indicated by the blue path in Fig. 6, are deflected less in both the electric and the magnetic fields than the slower particles indicated in red.Since the electric and magnetic deflections are in opposite directions the rays passing through D are brought together at a point F. It is readily shown that the angle between the line ZF and the initial direction ZC of the particles is equal to q. Fig. 7 shows the mean paths of the particles. The angle between FZ and OZ is denoted by r, where O is the centre of the magnetic field. The angle GOF = f, and ZFO = f 2 r.Now r and f 2 r are small; in Aston’s apparatus they were of the order of 1/10 rad. The line LM in Fig. 6 is therefore almost perpendicular to the lines ZL, ZM, FL and FM, and, to good approximation, its length is given by LM = adq = b(df 2 dq), (9) where a and b are the lengths OZ and OF. Similarly, in Fig. 7, if ON is the perpendicular from O to the line ZF, its length is ON = ar = b(f 2 r). (10) Therefore, from eqns. (10) and (9), a b = f 2 r r = df 2 dq dq , (11) whence f r = df dq = f 2q .(12) The last step follows from eqn. (8). Eqn. (12) shows that r = 2q, i.e. the angle FZC = q. The position of the point F on the line ZB depends on the value of f, and hence from eqn. (4) on the value of e/m. So particles with diVerent e/m values come Fig. 7 Mean paths of the particles in Aston’s mass spectrograph: Z is the centre of the electric field, and O the centre of the magnetic field; OZ = a, OF = b. to a focus at diVerent points along the line ZB.A photographic plate is placed along this line to record the traces. The position of the focus point on the line ZB for a given e/m value may be calculated from the values of E, B, and the geometry of the apparatus. However, the quantities required are the ratios of masses, and these are obtained most accurately by empirical methods. Aston first calibrated the instrument using a set of lines given by atoms and compounds with masses spread over a suitable range, and whose relative masses were known to the accuracy required.An example of such a set was: 6, C21; 8, O21; 12, C; 16, O; 28, CO; 32, O2; 44, CO2. (The integer before each atom or compound is the eVective mass number, i.e. the actual mass number divided by the number of charges on the ion.) This provided a set of points on a calibration curve. He filled in the gaps between the calibration points by taking the spectrum with the same set of ions, which were made to give lines at a diVerent place by changing the value of the magnetic field.Aston gave a preliminary account of the spectrograph in August 1919.11 The instrument was an immediate success. The two isotopes of neon, mass 20 and 22, were easily resolved.12,13 Similarly, chlorine was found to be a mixture of isotopes of mass 35 and 37.14 By the time his first book Isotopes appeared in 1922 he had studied 27 elements.15 Among them were the following (masses, where oxygen is 16, in parentheses): lithium (7, 6), boron (11, 10), magnesium (24, 25, 26), argon (40, 36), krypton (84, 86, 82, 83, 80, 78) and xenon (129, 132, 131, 134, 136, 128, 130).The isotopes are given in the order of the intensities of the lines. Reproductions of the spectra for neon and chlorine are given in Fig. 8. Although the discovery of many isotopes in light nonradioactive elements was of great importance, even more significant was Aston’s result that the masses of all the particles are whole numbers.(The only exception was that of hydrogen whose mass was 1.008, see below.) This whole number rule as it was called gave a simple model for the atomic nucleus. The only particles known at the time were the proton and the electron, with relative masses of 1837. It was therefore proposed that the nucleus of an isotope of mass M and charge Z, both being integers, consisted of M protons and M 2 Z electrons. Thus, for example, the nucleus of 7Li consisted of 7 protons and 4 electrons, while that of 6Li consisted of 6 protons and 3 electrons.Although this model gave the correct mass and charge of the nucleus, and satisfied the whole number rule, it had two grave defects. First, from the uncertainty principle, if an electron were confined to a region as small as an atomic nucleus, its momentum and hence energy would be much larger than the binding energy of the nucleus. Secondly, the spins of some of the nuclei were anomalous on this model.For example, the nucleus of 14N would consist of 14 protons and 7 electrons, giving a total of 21 particles. Since the spin of both the proton and the electron is ��� , the spin of the nucleus with an odd number of particles would be half-integral; in fact the spin of 14N is 1. The discovery of the neutron by James Chadwick16 in 1932 removed these diYculties. The present model is that a nucleus of charge Z and mass number M contains Z protons and M 2 Z neutrons.Isotopes are thus nuclei with the same number of protons and a diVerent number of neutrons. They have the same chemical properties, but diVerent nuclear properties. Fig. 8 Mass spectra obtained by Aston, 1919–1920, of (a) neon showing the 20 and 22 isotopes, and (b) chlorine showing the 35 and 37 isotopes.14 A number of other ions are present in both spectra. For example, the line at 28, prominent in both spectra, is due to CO. The lines at 36 and 38, present in the chlorine spectrum, are due to H35Cl and H37Cl.J.Chem. Soc., Dalton Trans., 1998, 3893–3899 3897 There are no electrons in the nucleus, and the nucleus 14N contains 14, an even number, of particles of spin ��� . Aston’s first mass spectrograph could separate particles with a mass diVerence of 1 in 130, which may be compared with a value of about 1 in 10 for Thomson’s parabola apparatus. The values of the masses were obtained with an accuracy of about 1 part in 1000. The second and third mass spectrographs Aston and other scientists soon grasped the reason for the departure of the mass value of hydrogen from an integral value, namely that it is the only atom with a non-composite nucleus.The masses of all the other atoms are reduced owing to the binding energy of their constituents, which results in hydrogen having a slightly higher mass relative to its mass number. The next step in mass spectrometry was therefore to improve the accuracy of the instrument to measure divergences from the whole number rule for all the atoms, which would give basic information on the binding forces within nuclei.For nuclei with mass numbers greater than about 20, the binding energy per nucleon is roughly constant, with a value between 8 and 9 MeV, which is about 1% of the energy equivalent of the mass of a nucleon. So to determine the binding energy to 1% the mass of the nucleus must be measured to an accuracy of about 1 part in 104.Aston started designing an improved version of his spectrograph in 1921, though he continued to use the original instrument until it was dismantled in 1925. In the second mass spectrograph finer slits were used and they were placed farther apart, thus more accurately defining the paths of the particles.17 The electric deflecting plates were curved, so that the particles remained midway between them as they were deflected. The electric deflection q was doubled to 1/6 rad. The potential for the deflection came from a set of 500 accumulators, each one built by Aston himself.They were charged twice a year and gave a voltage constant to better than 1 part in 105 during a single experiment. To achieve a constant magnetic field with minimum heating, the core of the magnet was wound with over 6000 turns of w over 100 kg. A current of 1 A through the coils produced a magnetic field of 1.6 T, which was suYcient to deflect the heaviest and most energetic particles through an angle f of 2/3 rad.(A detailed calculation shows that, when f = 4q, the position of the line varies linearly with mass, which is convenient for interpolation.) The pole pieces of the magnet were greatly reduced by changing their cross-section from the circular shape of the first instrument to a sickle shape, thereby producing a magnetic field only in the region close to the particle beam. Aston demonstrated the improved resolving power of the second instrument by separating six isotopes of mercury with mass numbers ranging from 198 to 204.18 With the original instrument the lines appeared as an unresolved blur.Aston’s third mass spectrograph in 1937 incorporated further improvements.19 The widths of the collimating slits could be adjusted externally, obviating the laborious opening of the apparatus which was necessary for such adjustments in the first two instruments. The stability of the magnetic field was improved by monitoring its strength with a fluxmeter; in the previous instruments only the exciting current had been kept constant.Any variation in the magnetic field was compensated for by manual adjustment of a spiral mercury resistor. Another advance was in the greatly improved sensitivity of the photographic plates used to record the lines, which resulted after extensive trials carried out with the collaboration of the Ilford photographic company. The biggest advance came in the use of the doublet method for comparing two masses.This consists of measuring the small diVerence in the masses of two ions with the same mass number M. The mass of an atom X with mass number M is denoted by m(MX). Since the value of m is close to the integer M, we may express it as m(MX) = M(1 1 d), (13) where d, known as the packing fraction, is small compared to 1. As an example we show how Aston measured the mass of the hydrogen atom in terms of the mass of the carbon atom.He measured the diVerence in mass D1 between the deuterium atom and the hydrogen molecule (doublet with M = 2), and the difference D2 between the masses of the triatomic deuterium molecule and the doubly charged carbon atom (doublet with M = 6). Then D1 = 2m(1H) 2 m(2H) = 2(1 1 dH) 2 2(1 1 dD) = 2dH 2 2dD, (14) D2 = 3m(2H) 2 m(12C21) = 6(1 1 dD) 2 6(1 1 dC) = 6dD 2 6dC, (15) dH 2 dC = (3D1 1 D2)/6. (16) Aston’s values for D1 and D2 were (15.2 ± 0.4) × 1024 and (423.6 ± 1.8) × 1024 respectively, giving dH 2 dC = (78.2 ± 0.4) × 1024. At the time of Aston’s measurements the atomic mass unit, denoted by u, was defined by taking the mass of the atom 16O to be exactly 16, but in 1962 the definition was changed 20 so that the mass of the atom 12C is taken as exactly 12, i.e.dC = 0. Thus on the present scale Aston’s value for the packing fraction of hydrogen was dH = (78.2 ± 0.4) × 1024, giving m(1H) = 1.00782 ± 0.00004 u. The example shows the intrinsic advantage of the method of measuring the diVerence in mass of doublets with the same mass number.The mass diVerences D1 and D2 are measured to accuracies of the order of 1%, but the mass of the hydrogen atom obtained is accurate to 4 parts in 105. It may be noted that Aston’s value is in complete agreement with the present value,21 m(1H) = 1.00782504 ± 0.00000001 u. The particles whose masses are measured in the spectrograph are ions that have lost one or more electrons, but the mass values quoted relate to the neutral atoms, i.e.the mass of one or more electrons is added to the measured masses. The mass of the electron, 5.486 × 1024 u, is small but not negligible at the accuracy of Aston’s measurements. On the other hand, the binding energies of the atoms in a molecule, being of the order of electronvolts, correspond to mass values of the order of 1029 u. So the mass of a diatomic molecule such as hydrogen may be taken to be twice the mass of the hydrogen atom. Aston improved the resolving power of his spectrographs from 130 for the first instrument to 600 for the second and 2000 for the third.He claimed an accuracy of 1 in 104 for the second instrument and approaching 1 in 105 for the third. If we compare his mass values (changing them to the 12C scale) with the present-day values, which are accurate to about 1 in 106 or better, the diVerences for the values obtained from the 1927 instrument are on average about 1.5 in 104, and for the 1937 values about 2.5 in 105.So his claimed accuracy is well substantiated. The third mass spectrograph, without the magnetic field components, is in the Museum of the Cavendish Laboratory; it is shown in Fig. 9. Aston’s first mass spectrograph is in the Science Museum in South Kensington. Other workers and modern developments Although Aston is recognised as the pioneer in mass spectroscopy there were other major workers in the field from 1918 onwards.Arthur Dempster, at the University of Chicago, con-3898 J. Chem. Soc., Dalton Trans., 1998, 3893–3899 structed a mass spectrograph in 191822 and in the next few years found isotopes in magnesium, lithium, potassium, calcium and zinc. His instrument involved bending monoenergetic ions into a semicircular path by a uniform magnetic field, which gives direction focusing, i.e. ions diverging in direction at the entrance to the magnetic field are brought to a focus after completing a semicircle.Kenneth Bainbridge, at the Franklin Institute, Swarthmore, improved on Dempster’s instrument by using a velocity filter before the magnetic analyser, thereby removing the need for a monoenergetic source.23 With his apparatus he made the first measurement of the mass of the deuterium atom24 and also provided one of the first experimental demonstrations of Einstein’s mass-energy relation.25 The detailed motions of ions in electric and magnetic fields were calculated by Richard Herzog and Josef Mattauch in Vienna 26 and others in the early thirties.The results led to the design of high-resolution double-focusing instruments in which ions with both a spread in velocities and a spread in initial directions were brought to a focus. Instruments making use of double focusing were built by Alfred Nier at the University Fig. 9 Aston’s third mass spectrograph, now in the the Museum of the Cavendish Laboratory. The discharge tube is on the right.The white discs are for adjusting the widths of the slits S1 and S2. The magnet, not shown, acts over the region of the tube in the left-hand wooden support. The photographic plate is placed in the focus plane just before the left end of the tube to record the spectra. The overall length of the instrument, excluding the discharge tube, is 105 cm. Fig. 10 Aston working with his apparatus for the separation of the isotopes of neon by fractional distillation, 1914.of Minnesota 27 and several other workers.28 Further improvements came from replacing photographic plates by electrical detectors and from advances in vacuum technology.29 The value of 2000 for the resolving power of Aston’s third mass spectrograph has been extended to values exceeding 105, and accuracies of the order of 1 part in 108 or 109 have been obtained.30 Mass spectroscopy is now applied in several branches of chemistry, biology, geology, and physics. In many of these applications the classical method of electrostatic and magnetic deflection described in this paper has been replaced by timing methods.The highest precisions are obtained by cyclotron resonance in which the frequencies of ions rotating in a uniform magnetic field are measured and analysed by Fourier transform techniques. Sophisticated ionisation methods have been developed for the analysis of complex biomolecules and macromolecules.31 Aston’s later life Aston’s first mass spectrograph brought him immediate acclaim.He was appointed to a Fellowship in Trinity College, Cambridge in 1920 and was made a Fellow of the Royal Society in 1921. He was awarded the Nobel Prize in Chemistry in 1922 for, in the words of the citation, ‘his discovery, by means of his mass spectrograph, of isotopes in a large number of nonradioactive elements, and for his enunciation of the whole Fig. 11 The research students and staV of the Cavendish Laboratory in 1922, the year that Aston was awarded the Nobel Prize.Aston is fourth from the left in the front row, Thomson is fifth, and Rutherford, the Cavendish Professor, is sixth. Edward Appleton is second from the right in the front row, Patrick Blackett is second from the right in the second row, and Peter Kapitza is at the right-hand end of the back row. Fig. 12 Aston working with his third mass spectrograph, ca. 1937.J. Chem. Soc., Dalton Trans., 1998, 3893–3899 3899 number rule’. In proposing the toast of the laureates at a dinner in December of that year, Svante Arrhenius, the Director of the Nobel Institute, commented that never before had the Nobel Prize been handed over to a group of such distinguished laureates, which, besides Aston, included Niels Bohr, Albert Einstein, and Frederick Soddy.The last two were the 1921 prize winners in Physics and Chemistry respectively, but the awards were made in 1922. Aston never married and for the last 35 years of his life lived in Trinity College. Outside his work his main interests were sport, travel and music.He was a keen cross-country skier, and played tennis up to tournament class. He played golf in a famous foursome with Ernest Rutherford, Ralph Fowler, and GeoVrey (G. I.) Taylor. He was an enthusiastic cyclist, once cycling 200 miles in 22 h. He was also an excellent photographer and combined this hobby with his love of travel to help at several solar eclipse expeditions. He was an omnivorous reader, Sherlock Holmes being his favourite.An acquaintance once described him as the highest lowbrow that he had ever met. Aston died on 20 November 1945. In an obituary in Nature 32 G. P. Thomson wrote ‘Aston was a man in whom a great zest for life was combined with a simplicity of character almost approaching naivety. Though a good occasional lecturer, he had no gift for teaching, and a few early attempts were not persisted in. His attitude to physics was essentially that of the experimenter and visualizer.He preferred the model to the equation, the concrete to the abstract. He was a Conservative in politics as in life, and though he would admit that a change might be good, he preferred it to happen as gradually as possible’. In summary, Aston was not only a one-experiment man, he was eVectively a one-instrument man. However, what he measured was of the highest importance and he did it extremely well. To quote G. P. Thomson once more,33 ‘Aston was a superb experimenter.His first mass spectrograph was a triumph; few but he could have got it to work at all’. Acknowledgements Fig. 8 is reproduced from the Philosophical Magazine, 1920, by permission of Taylor & Francis. Figs. 5 and 9 to 12 are from the Photographic Archives of the Cavendish Laboratory. References 1 G. Hevesy, Obituary Notices Fellows R. Soc., 1948, 5, 635. 2 B. B. Boltwood, Am. J. Sci., 1907, 24, 307. The nomenclature is that used at the time.The thorium referred to here is 232Th; ionium is 230Th; radium B, so called because it is one of the decay products following from radium, is 214Pb. 3 E. Rutherford and E. N. Da C. Andrade, Philos. Mag., 1914, 27, 854. 4 F. Soddy, Nature (London), 1913, 92, 399. The word isotope was suggested to Soddy by Dr. Margaret Todd, a medical doctor with a practical knowledge of Greek (A. Fleck, Biogr. Mem. Fellows R. Soc., 1957, 3, 203). 5 E. Goldstein, Berlin Ber., 1886, 39, 691. 6 W. Wien, Verh. Phys. Gesell. Berlin, 1898, 17, 10. 7 J. J. Thomson, Philos. Mag., 1907, 13, 561. 8 J. J. Thomson, Royal Institution, Weekly Meeting, January 17, 1913. 9 The same method of gaseous diVusion was employed at Oak Ridge during the second World War to separate the uranium isotopes 235 and 238 in the gas UF6. 10 G. P. Thomson, J. J. Thomson and the Cavendish Laboratory in his Day, Nelson, London, 1964, p. 137. 11 F. W. Aston, Philos. Mag., 1919, 38, 707. 12 F. W. Aston, Nature (London), 1919, 104, 334. 13 F. W. Aston, Philos. Mag., 1920, 39, 449. 14 F. W. Aston, Philos. Mag., 1920, 39, 611. 15 F. W. Aston, Isotopes, Edward Arnold, London, 1922. Aston wrote a second book in 1933, entitled Mass-Spectra and Isotopes, that dealt more with the experimental results and less with the general theory. This was followed by a second edition in 1942. 16 J. Chadwick, Proc. R. Soc. London, Ser. A, 1932, 136, 692. 17 F. W. Aston, Proc. R. Soc. London, Ser. A, 1927, 115, 487. 18 F. W. Aston, Nature (London), 1925, 116, 208. 19 F. W. Aston, Proc. R. Soc. London, Ser. A, 1937, 163, 391. 20 B. W. Petley, The Fundamental Physical Constants and the Frontier of Measurement, Adam Hilger, Bristol and Boston, 1985, p. 93. 21 E. R. Cohen and B. N. Taylor, The 1986 Adjustment of the Fundamental Physical Constants, Codata Bulletin, 63, Pergamon, Oxford, 1986. 22 A. J. Dempster, Phys. Rev., 1918, 11, 316. 23 K. T. Bainbridge, Phys. Rev., 1932, 40, 130. 24 K. T. Bainbridge, Phys. Rev., 1932, 42, 1. 25 K. T. Bainbridge, Phys. Rev., 1933, 44, 123. 26 R. Herzog and J. H. E. Mattauch, Ann. Phys., 1934, 19, 345. 27 A. O. Nier, Rev. Sci. Instrum., 1960, 31, 1127. 28 H. E. Duckworth, R. C. Barber and V. S. Venkatasubramanian, Mass Spectroscopy, 2nd edn., Cambridge University Press, 1986, ch. 5. This book gives an excellent account of the subject. 29 The term mass spectrograph is reserved for an instrument that records the spectra on a photographic plate, an instrument that employs electrical detection being termed a mass spectrometer. The last reported use of a mass spectrograph was in 1972. 30 Ref. 28, p. 159. 31 P. G. Gates, World Wide Web, http://www-methods.ch.cam.ac.uk 32 G. P. Thomson, Nature (London), 1946, 157, 290. 33 Ref. 10, p. 139. Paper 8/04629H

ISSN:1477-9226    

DOI:10.1039/a804629h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 3901–3903 3901 Preparation and structures of the mixed-metal clusters Pt2M2Se2Cl2(PPh3)4 (M 5 Cu, Ag) and Pd2Au2Se2(SeH)2(PPh3)4. An entry to ternary clusters Pierre D. Harvey,*† Andreas Eichhöfer and Dieter Fenske * Institut für Anorganische Chemie, Universität Karlsruhe, Engesserstrasse, Gebäude – Nr. 30.45, 76128 Karlsruhe, Germany Received 23rd September 1998, Accepted 19th October 1998 Both cis- and trans-Pt(PPh3)2Cl2 react with Se(SiMe3)2 and CuCl or AgO2CR (R 5 Me, Ph) in THF to form the mixedmetal clusters Pt2M2Se2Cl2(PPh3)4 (M 5 Cu or Ag, respectively), and likewise Au(PPh3)Cl reacts with Pd(acac)2, to form Pd2Au2Se2(SeH)2(PPh3)4. The preparation of copper chalcogenide clusters is of great interest for the synthesis of nanoparticles.1 The preparation of the largest crystallographically characterized copper cluster, Cu146Se73(PPh3)30, was recently reported, along with a number of other clusters with diVerent nuclearities.1 The synthesis of such materials follows this general reaction scheme: 2CuCl 1 Se(SiMe3)2 1 xPR3 æÆ “Cu2Se(PR3)x” 1 2ClSiMe3 (1) where PR3 is either an alkyl, aryl or mixed alkyl/aryl phosphine.This chemistry has also been extended to silver, and similar results have been obtained.1 Binary systems of the type Pt/Se, Pt/Te and Au/Se have been explored by various groups, and the preparations of binuclear and trinuclear species such as (PPh3)2- Pt(m-Se)2Pt(PPh3)2,2 L2Pt(m-Te)2PtL2 (L = PPh3, PEt3; L2 = dppm),3 Pt3Se2(dppe)3,3 [Se(AuPPh3)3]PF6 and Au2Se(PPh3)2,4 have also been reported.In this work we wish to take advantage of the capacity of the Se atom to coordinate an extra M group in order to promote M–M9 bonding, and mixed-metal cluster formation. The synthesis and crystal structures of three novel ternary clusters, Pt2M2Se2Cl2(PPh3)4 (M = Cu 1, Ag 2), and Pd2Au2Se2(SeH)2(PPh3)4 3 which are the first examples of cluster compounds containing Pt/Cu/Se, Pt/Ag/Se,‡ and Pd/Au/ Se are now presented.These compounds represent potential precursors as an entry to larger mixed-metal clusters via the presence of displaceable Cl atoms, and phosphine groups. In relation with this study, we find that examples of Pt–Cu bondcontaining clusters are rather rare, where only six have been characterized from X-ray crystallography.5 For the Pt–Ag analogues, many more examples are known.6 The clusters 1 and 2 can easily be prepared from the following general reactions:§ 2Pt(PPh3)2Cl2 1 2CuCl 1 2Se(SiMe3)2 THF 1 1 4ClSiMe3 (2) 2Pt(PPh3)2Cl2 1 2Ag(O2CR) 1 2Se(SiMe3)2 THF 2 1 4RCO2SiMe3 (3) where R = Me, or Ph, and the Pt(PPh3)2 Cl2 species can be either the cis- or trans-isomers.When the chalcogenide reagent Se(SiMe3)2 is slowly added dropwise into solutions containing Pt(PPh3)2Cl2 and CuCl or AgO2CR (R = Me, Ph) in stoichiometric amount (1:1:1) in the presence of 2 equivalents of PPh3, and at 240 8C, the solutions pass from a deep yellow to a dark brown coloration.Upon slowly warming the solutions to room temperature over several hours, and letting them sit for several days, large orange crystals were readily obtained, and were identified from X-ray crystallography as Pt2M2Se2Cl2(PPh3)4 (M = Cu, Ag). The X-ray structure analysis¶ reveal the isostructural behavior of these two isocentric clusters, where two M–Cl units sit above and under the planar P2Pt(m-Se)2PtP2 fragment, and the Pt2M2Se2 core forms a strongly distorted octahedral structure (C2h symmetry, see Fig. 1). The coordination of the M–Cl units occurs via only one formal Se–M single bond with distances [d(Se–Cu) 2.274(1), d(Se–Ag) 2.548(2) Å] that are normal in comparison with the related binary “M2Se(PR3)x” clusters (M = Cu, Ag).1 Thus the Se atoms adopt a m3-binding mode with Pt–Se bond distances of 2.474(1) and 2.485(1) Å for M = Cu, and 2.509(2) and 2.478(1) Å for M = Ag.The Pt–Se–Pt and Pt–Se–M angles are 97.96(3), and 79.70(4) and 74.44(4)8 for M = Cu, and 98.58(5), and 74.32(5) and 73.56(6)8, for M = Ag. The smaller Pt–Se–M angles are associated with the presence of Pt ? ? ? M interactions. Indeed no formal Pt–M bonding occurs where the Pt–M distances range from ª2.92 to 3.05 Å. This result contrasts greatly with all Pt–Cu bondcontaining clusters for which the Pt–Cu bonds are reported to be ª2.53 < d(Pt–Cu) < ª2.74 Å.5 Despite the long Pt–Cu distance, significant Pt ? ? ? Cu interactions are readily anticipated,7 as these values are located well inside the sum of the van der Waals radii.On the other hand for the Pt2Ag2 analogue, such distances are not uncommon in the literature.6 Such interactions are also felt in the Se–M–Cl angles which deviate slightly from linearity generally encountered in pure sp hybridation. The “(PPh3)2Pt(m-Se)2Pt(PPh3)2” frame is not greatly aVected Fig. 1 Molecular structure for clusters 1 and 2.The H-atoms are omitted for clarity. Selected bond distances (Å) and angles (8) are as follows; 1: Pt–Se 2.474(1), 2.485(1), Pt–P 2.279(2), 2.291(2), Pt–Cu 2.916(1), 3.047(1), Cu–Cl 2.129(3), Cu–Se 2.274(1), Se ? ? ? Se 3.254(2), Pt ? ? ? Pt 3.753(2); Se–Pt–Se 82.04(3), P–Pt–P 100.30(7), Pt–Se–Pt 97.96(3), Se–Cu–Cl 176.90(11)8. 2: Pt–Se 2.509(2), 2.478(1), Pt–P 2.293(4), 2.318(3), Pt–Ag 3.028(2), 3.037(1), Ag–Cl 2.361(5), Ag–Se 2.548(2), Se ? ? ? Se 3.253(3), Pt ? ? ? Pt 3.781(3); Se–Pt–Se 81.42(5), P–Pt–P 96.63(12), Pt–Se–Pt 98.58(5), Se–Ag–Cl 171.59(15)8.3902 J.Chem. Soc., Dalton Trans., 1998, 3901–3903 upon complexation with the M–Cl groups. However, by comparison with the literature data reported for the “free” (PPh3)2Pt(m-Se)2Pt(PPh3)2 dimer,2 some bond distances have increased. Indeed the average Pt–Se and Pt–P bond lengths are 2.449 and 2.277Å for Pt2Se2(Ph3)4,2 2.480 and 2.285 Å for 1, and 2.494 and 2.306 Å for 2, respectively.This eVect is clearly steric on one side, but also some electronic factors such as electronic density change at the Pt atoms promoting Pt ? ? ? M interactions, could also contribute to the bond length variations. Cluster 3 can be prepared in a similar fashion in the dark in reasonable yield according to: 2Au(PPh3)Cl 1 2Pd(acac)2 1 4Se(SiMe3)2 1 2PPh3 1 2H2O THF 3 1 “XSiMe3” (X = Cl, acac, OH) (4) 3 crystallizes as red-orange crystals. This time the excess of PPh3 is not used as a stabilizing/solubilizing agent, but as a reactant.X-Ray crystallographic analysis indicates that 3 is also a centrosymmetric cluster (point group Ci), again showing a strongly distorted Pd2Au2Se2 octahedral (Fig. 2). As for 1 and 2, the d10 electronic configuration metal is bonded to the d8–d8 dimer (PPh3)(SeH)Pd(m-Se)2Pd(PPh3)(SeH) (C2h point group) via a formal Se–Au single bond [2.412(2) Å] leading to weak PdII ? ? ? AuI contacts [3.067(2) and 3.300(2) Å].7 Both Pd and Au carry a single PPh3 ligand which diVers from 1 and 2 where both PPh3’s are bonded to the Pt only.One other diVerence is the presence of SeH groups [1H NMR d(ppm) 0.123] instead of Cl. The fact that clusters 1 and 2 have extra Cl atoms and 3 exhibits SeH centers opens the door to further extension of this chemistry towards larger clusters or the incorporation or another type of metal. Further research in this area is in progress. Acknowledgements We are grateful to the Deutsche Forschungsgemeinschaft (SFB195), to the Fonds der Chemischen Industrie and the EU through the HCM program for financial support.Pierre D. Harvey also thanks the University of Karlsruhe for financial support (guest Professor). Notes and references † Work performed while on sabbatical leave from the Université de Sherbrooke, Canada. Present address: Département de Chimie, Université de Sherbrooke, Sherbrooke, P.Q., Canada, J1K 2R1. E-mail: pharvey@courrier.usherb.ca ‡ According to the Cambridge Data Bank a compound formulated as Fig. 2 Molecular structure for 3. The H-atoms are omitted for clarity. Selected bond distances (Å) and angles (8) are as follows: Au–Se 2.412(2), Au–P 2.261(6), Au–Pd 3.067(2), 3.300(2), Pd–Se(H) 2.446(3) Pd–Se 2.486(3); Se–Au–P 176.4(1), Se–Pd–P 175.5(2), Se–Pd–Se 84.8(1), P–Pd–Se(H) 93.2(2), Pd–Se–Pd 95.2(1)8. [(PPh3)2PtAg2SeCo{(PPh2CH2CH2)3CMe}]BF4 has been described in G. Baldi, M. di Vaira, L.Niccolai, M. Peruzzini and P. Stoppioni, Eur. Cryst. Meeting, 1985, 9, 164, but no formal report of this cluster exits. § Preparation of 1 : 0.35g (0.50 mmol) of either cis- or trans-Pt(PPh3)Cl2 was mixed with 0.050 g (0.50 mmol) of dry CuCl and 0.26 g (1.0 mmol) of PPh3 in 20 ml of dry THF under N2(g) at room temperature. Then the unstirred solution was cooled to ª240 8C using an acetone bath and N2(l), prior to slow addition of ª0.10 ml (1.1 mmol) of Se(SiMe3)2. The solution turned yellow, pale orange, and deep brown during these additions. The solution was then allowed to warm up over several hours until room temperature was reached.After several days in the dark, large orange crystals readily appeared and were collected for X-ray analysis. Yield ª50%. 31P NMR (C6D6) d 28.23 [1J(PaPt) = 3057, 1J(PbPt) = 2964; 3J(PaPt) = 1564, 3J(PbPt) = 1442; 2J(PP) = 88Hz]. Preparation of 2: this cluster was prepared in the same way as described for 1 except that both AgO2CR starting materials (R = Me, Ph) were used (0.085 g, 0.50 mmol, R = Me; 0.11 g, 0.50 mmol, R = Ph), instead of CuCl and the solution was kept in the dark at all times.Orange crystals were obtained in all cases over a period of several weeks. For the crystal reported in this work, an addition of a wet acetone (unpurified)– benzene mixture to the THF solutions was made. Crystallisation appeared more rapidly (ª1 day). Yield ~80%. 31P NMR(C6D6) d 27.94 [1J(PaPt) = 2984, 1J(PbPt) = 2882; 3J(PaPt) = 1615, 3J(PbPt) = 1426; 2J(PP) = 92 Hz].Preparation of 3: 0.30 g (1 mmol) of Pd(acac)2, 0.49 g (1 mmol) of Au(PPh3)Cl and 0.52 g (2 mmol) of PPh3 (excess) were dissolved in 25 ml of THF at room temperature. Then the solution was cooled to ª240 8C and kept in the dark prior to adding 0.50 ml (ª3 mmol) of Se(SiMe3)2. The solution was allowed to warm to room temperature over several hours. After several days in the dark, water was introduced very slowly into the solution over a period of several days and orange-red crystals appeared over this addition period.The crystals are light stable. Yield ª50%. 1H NMR(C6D6) d 0.123 and 0.300 [SeH, 1J(SeH) = 3.3 Hz] for compounds a and b (chemical exchange in solution), 6.8–7.5 (C6H5P, br bands). 31P NMR (C6D6) d 6.6 (free PPh3 in chemical exchange, very br), 16–30 (PdP and AuP, complex, both isomers). ¶ Crystal data. For 1?THF: C80H60Cl2Cu2O2P4Pt2Se2, orange plate, M = 961.62, monoclinic, space group P21/n, a = 14.828(3), b = 14.221(3), c = 18.032(4) Å, b = 102.62(3)8, V = 3710.5(13) Å3, at 200(2) K, Z = 2, Dc = 1.721 g cm23, m = 5.509 mm21, 2qmax = 52.028, 7117 independent reflections measured (Rint = 0.1104) on a STOE IPDS diVractometer.All Pt, Cu, Se, Cl and C atoms were refined anisotropically, to yield R = 0.0632, wR2 = 0.1696, S = 1.013 for 6365 data [Fo > 4sFo]. For 2?C6H6?2H2O: C78H60Ag2Cl2O2P4Pt2Se2, small orange hexagonal plate, M = 1987.88, triclinic, space group P1� , a = 11.386(2), b = 13.499(3), c = 14.169(3) Å, a = 64.73(3), b = 80.95(3), g = 70.48(3)8, V = 1854.7(6) Å3, at 200(2) K, Z = 1, Dc = 1.780 g cm23, m = 5.464 mm21, 2qmax = 52.088, 5205 independent reflections measured (Rint = 0.0529) on a STOE IPDS diVractometer.All C, Ag, Cl, O, P, Pt and Se atoms were refined anisotropically to yield R = 0.0651, wR2 = 0.1850, S = 1.117 for 4434 [Fo > 4sFo]. For 3?2THF: C80H76Au2O2P4Pd2Se4, orange-red plate fragment, M = 2099.73, triclinic, space group P1� , a = 11.004(2), b = 12.939(3), c = 14.452(3) Å, a = 70.27(3), b = 76.32(3), g = 82.27(3)8, V = 1878.6(7) Å3, at 200(2) K, Z = 1, Dc = 1.856 g cm23, m = 6.432 mm21, 2qmax = 45.008, 3693 independent reflections measured (Rint = 0.0732) on a STOE IPDS diVractometer.All C, Au, O, P, Pd, and Se atoms were refined anisotropically, to yield R = 0.0747, wR2 = 0.1826, S = 1.011 for 2657 data [Fo > 4sFo]. CCDC reference number 186/1205.See http:// www.rsc.org/suppdata/dt/1998/3901/ for crystallographic files in .cif format. 1 J. F. Corrigan and D. Fenske, Chem. Commun., 1997, 1837; S. Dehnen and D. Fenske, Chem. Eur. J., 1996, 2, 1407; S. Dehnen, A. Schäfer, R. Ahlrichs and D. Fenske, Chem. Eur. J., 1996 2, 429; S. Dehnen and D. Fenske, Angew. Chem., Int. Ed. Engl., 1994, 33, 2287; H. Krautscheid, D. Fenske, G. Baum and M. Sewmelmann, Angew. Chem., Int. Ed. Engl., 1993, 32, 1303; F. Corrigan and D.Fenske, Chem. Commun., 1996, 943; D. Fenske and J.-C. Steck, Angew. Chem., Int. Ed. Engl., 1993, 32, 238; D. Fenske, T. Langetepe and N. Zhu, Angew. Chem., 1998, 110, 2787, Angew. Chem., Int. Ed. Engl., 1998, 37, 2639. 2 A. Bencini, M. DiVaira, R. Morassi and P. Stoppioni, Polyhedron, 1996, 15, 2079. 3 A. L. Ma, J. B. Thoden and L. F. Dahl, J. Chem. Soc., Chem. Commun., 1992, 1516; R. D. Adams, T. A. Wolfe, B. W. Eichhorn and R. C. Haushalter, Polyhedron, 1989, 8, 701; D.Fenske, H. Fleischer, H. Krautscheid, J. Magull, C. Oliver and S. Weisgerber, Z. Naturforsch., Teil B, 1991, 46, 1384; K. Matsumoto, M. Ikuzawa, M. Kamikubo and S. Ooi, Inorg. Chim. Acta., 1994, 217, 129; H. Wolkers, K. Dehnicke, D. Fenske, A. Khassanov and S. S. Hafner, Acta Crystallogr., Sect. B, 1991, 47, 1627.J. Chem. Soc., Dalton Trans., 1998, 3901–3903 3903 4 C. Lensch, P. G. Jones and G. M. Sheldrick, Z. Naturforsch., Teil B, 1982, 37, 944; P. G. Jones and C. Thöne, Chem. Ber., 1991, 124, 2725. 5 T. G. M. M. Kappen, P. P. J. Schlebos, J. J. Bour, W. P. Bosman, J. M. M. Smits, P. T. Bewiskens and J. J. Steggerda, J. Am. Chem. Soc., 1993, 117, 8327; P. Braunstein, S. Freyburger and Odile Bars, J. Organomet. Chem., 1988, 352, C29; M. F. Hallam, D. M. P. Mingos, T. Adatia and M. McPartlin, J. Chem. Soc., Dalton Trans., 1988, 335; T. G. M. M. Kappen, P. P. J. Schlebos, J. J. Bour, W. P. Bosman, J. M. M. Smits, P. T. Beurskens and J. J. Steggerda, Inorg. Chem., 1995, 34, 2133; M. F. J. Schoondergang, J. J. Bour, P. P. J. Schlebos, A. W. P. Vermeer, W. P. Bosman, J. M. M. Smits, P. T. Beurskens and J. J. Steggerda, Inorg. Chem., 1991, 30, 4704. 6 See for examples, C. Archambault, R. Bender, P. Braunstein, A. De Chian and J. Fisher, Chem. Commun., 1996, 2729; A. Albinati, F. Demartin, L. M. Venanzi and M. K. Wolfer, Angew. Chem., Int. Ed. Engl., 1988, 27, 563; A. Albinati, S. Chaloupka, F. Demartin, T. F. Koetzle, H. Rüegger, L. M. Venanzi and M. K. Wolfer, J. Am. Chem. Soc., 1993, 115, 169; R. Uson, J. Fornies, M. Tomas and J. M. Casas, J. Am. Chem. Soc., 1985, 107, 2556; R. Uson, J. Fornies, M. Tomas, F. A. Cotton and L. R. Falvello, J. Am. Chem. Soc., 1984, 106, 2482. 7 P. Pyykkö, Chem. Rev., 1997, 97, 597. Communication 8/07

ISSN:1477-9226    

DOI:10.1039/a807414c

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3905–3909 3905 Structure and magnetic behaviour of the one-dimensional [{Ni2(Medien)2(Ï1,1-N3)2(Ï1,3-N3)}n][ClO4]n compound with unusually strong ferro-antiferromagnetic alternating interactions Albert Escuer,*a Ramon Vicente,a M. Salah El Fallah,a Sujit B. Kumar,a Franz A. Mautner b and Dante Gatteschi c a Departament de Química Inorgànica, Universitat de Barcelona, Diagonal, 647, 08028-Barcelona, Spain. E-mail: http:www.ub.es/inorgani/molmag.htm b Institut für Physikalische und Theorestische Chemie, Technische Universität Graz, A-8010 Graz, Austria c Dipartimento di Chimica, Universitá degli Studi di Firenze, Via Maragliano 75-77, 50144 Firenze, Italy Received 7th July 1998, Accepted 29th September 1998 A new monodimensional m-azido-nickel(II) complex of formula [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n][ClO4]n 1 [Medien = bis(2-aminoethyl)methylamine] has been synthesized and crystallographically characterised.The nickel atom is placed in an octahedral environment bridged by two end-on azido ligands and one end-to-end azido bridge to the neighbouring nickel atoms, giving an alternated 1-D system. The magnetic behaviour corresponds to an alternating ferro-antiferromagnetic chain. The superexchange parameters have been calculated as JAF = 234.7 cm21 and JFM = 138.2 cm21, a = 1.10, by means of an improved analytical expression for the magnetic susceptibility of the isotropic ferro-antiferromagnetic S = 1 Heisenberg chain in terms of the alternation parameter a = JFM/|JAF|.Magnetic properties have been correlated to the structural data. Introduction One-dimensional magnetic materials have attracted increasing interest, and have been thoroughly studied from both experimental and theoretical points of view mostly due to their position between high nuclearity clusters and extended 3-D lattices, which opens up new possibilities for understanding phenomena that cannot be explained in a higher dimension.1–4 In recent years the azide bridge has been shown to be able to generate systems from discrete molecules 5 with diVerent nuclearity to 3-D compounds,6 and specifically for nickel(II) a large number of one-dimensional systems have been reported.7 The most common topologies for the 1-D nickel–azide system consist of homogeneous chains with a single end-to-end bridge, with four co-ordination sites of the nickel environment occupied by four polyamine ligands or the double end-to-end bridge with the two remaining co-ordination sites occupied by two diamine blocking ligands.7 The end-to-end kind of compound has been extensively studied from synthetic and magnetic points of view and suitable models were proposed to explain and predict the magnitude of the antiferromagnetic coupling as a function of the bond parameters in the bridge region.5c,8 Less common are the examples of end-on ferromagnetic chains and the alternating systems.9 This latter kind of unusual system may be classified in several types: alternating end-to-end systems, alternating end-to-end/end-on systems and complex alternance of bridges, for which the magnetic analysis becomes more complicated or impossible in some cases.Continuing with our work in this field we report the syntheses and magnetic behaviour of the compound [{Ni2(Medien)2(m1,1-N3)2- (m1,3-N3)}n][ClO4]n [Medien = bis(2-aminoethyl)methylamine] which shows the single end-to-end/double end-on alternance, giving a ferro-antiferromagnetic alternating system.To fit our experimental data an improved expression built on the basis of the increasing ring system calculations has been developed and checked. Experimental Synthesis The complex [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n][ClO4]n was synthesized by the addition of 15 ml of an aqueous solution of sodium azide (0.23 g, 4 mmol) to an aqueous solution obtained by mixing 0.73 g (2 mmol) of nickel perchlorate and 0.24 g (2 mmol) of the Medien ligand.Slow evaporation of the blue resulting solution provided blue crystals suitable for X-ray determination. Yield 60%. [Calc. (Found) for C10H30ClN15- Ni2O4: C, 20.8 (20.8); H, 5.2 (5.3); N, 36.4 (36.4)%]. The IR spectrum shows the typical absorptions corresponding to the Medien ligand, the perchlorate counter anion and the characteristic n(N3) at 2090, 2028 and 2055 (sh) cm21. Spectral and magnetic measurements Infrared spectra (4000–400 cm21) were recorded from KBr pellets on a Nicolet 520 FTIR spectrophotometer. Magnetic measurements were carried out with a DSM8 pendulum susceptometer, working in the temperature range 300–4 K.The applied external magnetic field was 1.5 T. Diamagnetic corrections were estimated from Pascal tables. Crystallographic data collection and refinement The single crystal data were collected on a modified STOE fourcircle diVractometer (crystal size 0.55 × 0.35 × 0.30 mm). The crystallographic data, conditions retained for the intensity data collection, and some features of the structure refinement are listed in Table 1.The accurate unit-cell parameters were determined from automatic centring of 25 reflections (12 < q < 188) and refined by least-squares methods. Intensities were collected with graphite-monochromated Mo-Ka radiation, using the w-scan technique. 5462 Reflections were collected in the range 2.90 < 2q < 27.408 (4917 independent reflections, Rint 0.0242).3906 J.Chem. Soc., Dalton Trans., 1998, 3905–3909 No intensity decay of 3 control reflections (21 3 21; 21 4 23; 0 22 24), collected every hour, was observed. Corrections were applied for Lorentz-polarisation eVects but not for absorption (range of transmission: 0.650–1.000). The structure was solved by direct methods using the SHELXS 86 10 computer program, and refined by full-matrix least-squares methods on F2 using SHELXL 9311 incorporated in the SHELXTL/PC V 5.03 program package.12 Fourier-diVerence maps indicated two split orientations of one partially disordered Medien ligand and the perchlorate counter anion.Refinement of the corresponding population parameters gave values of 0.60(1) and 0.40(1) for the Medien ligand and 0.67(1) and 0.33(1) for the perchlorate. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were located on calculated positions and assigned six common isotropic displacement factors (one for each type of parent C and N atom in each Medien ligand).The final R factor was 0.0398 (wR2 = 0.0999). The number of parameters refined was 378. Goodness of fit 1.048. Maximum shift/e.s.d. = 0.05. Maximum and minimum peaks in the final diVerence synthesis were 0.614 and 20.546 e Å23, respectively. Significant bond parameters are given in Table 2. CCDC reference number 186/1180. Results and discussion Crystal structure The labelled diagram for complex 1 together with a chain perspective are shown in Fig. 1. The structure consists of Ni(Medien) units bridged by azido ligands, giving a monodimensional compound and perchlorate counter anions. Two azide co-ordination modes are present in an alternating form along the chain: each nickel atom has a double end-on bridge with one of the neighbouring nickel atoms whereas the bridge with a second nickel atom consists of a single end-to-end azido linkage. Each Ni(1)–Ni(1A) and Ni(2)–Ni(2B) subunit has an inversion centre, and the structure may thus be envisaged as two non-equivalent Ni(m1,1-N3)2Ni planar units bridged by one m1,3-N3 ligand.The environments of Ni(1) and Ni(2) are quite similar, with the tridentate Medien ligand in fac co-ordination. Bond distances to the N atoms of the end-on bridge [Ni(1)– N(11) 2.110(2), Ni(1)–N(11A) 2.116(3), Ni(2)–N(31) 2.113(3) and Ni(2)–N(31B) 2.112(3) Å] are very similar, and slightly shorter than the bond distances to the N atoms of the end-toend azido bridge [Ni(1)–N(21) 2.159(3) and Ni(2)–N(23) 2.142(3) Å].The Ni(1)–N(11)–Ni(1A) 100.8(1) and Ni(2)– N(31)–Ni(2B) 98.9(1)8 bond angles are smaller than the common bond angles found in previous dimeric Ni(m1,1-N3)2Ni entities, which show values between 101 and 1048.14–17 The endto- end bridge shows asymmetric co-ordination as is normal for this ligand, Ni(1)–N(21)–N(22) 138.1(2) and N(22)–N(23)– Ni(2) 125.7(2)8. The t torsion angle, defined as the dihedral between Ni(1)–N(21)–N(22)–N(23) and N(21)–N(22)–N(23)– Fig. 1 An ORTEP13 drawing of [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n]- [ClO4]n with the atom-labelling scheme. Thermal ellipsoids are at the 50% probability level. Ni(2) mean planes is 162.8(4)8. The Ni ? ? ? Ni distance in the end-on units is 3.257(1) Å for Ni(1)–Ni(1A) and 3.203(1) Å for Ni(2)–Ni(2B), whereas the Ni(1) ? ? ? Ni(2) distance is greater, 6.060(2) Å, due to the end-to-end co-ordination of the bridge.The structure of the [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n][ClO4]n compound shows the same double/single bridge alternance as the previously reported 9b [{Ni2(dpt)2(m1,3-N3)2(m1,3-N3)}n]- [ClO4]n, dpt = bis(3-aminopropyl)amine, Fig. 2. The diVerence between the two systems lies in the kind of bridge, alternating double end-on/single end-to-end for the present compound and only end-to-end bridges for the previous complex. From the synthetic and magnetic points of view, these compounds are a nice example of the extreme versatility of the azido ligand in providing uncommon magnetic systems on the basis of simple synthetic procedures.The model and calculation Ferro-antiferromagnetic systems have attracted much attention in the past few years, mainly focused on their complicated magnetic behaviour derived from the simultaneous exchange coupling interactions. One of the consequences has been the development of suitable models to explain their magnetic properties.For the S = 5/2 system, which permits the treatment as a classical vector, appropriate equations have been proposed,18 whereas for lower S values several authors have solved numerically the magnetic exchange expressions for the alternating anti- Fig. 2 Perspective view of [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n][ClO4]n and comparison with the related [{Ni2(dpt)2(m1,3-N3)2(m1,3-N3)}n][ClO4]n compound. Table 1 Crystal data and structure refinement for [{Ni2(Medien)2- (m1,1-N3)2(m1,3-N3)}n][ClO4]n Formula Formula weight Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 ZT /8C l(Mo-Ka)/Å Dc, Dm/g cm23 m(Mo-Ka)/mm21 R wR2 C10H30ClN15Ni2O4 577.36 Triclinic P1� (no. 2) 7.421(1) 12.339(3) 14.171(5) 113.35(2) 100.80(3) 96.45(2) 1144.9(6) 2 25(2) 0.71069 1.675, 1.66(3) 1.813 0.0398 0.0999J. Chem. Soc., Dalton Trans., 1998, 3905–3909 3907 Fig. 3 Temperature dependence of cr versus Tr for N = 2, 3, 4, 5 rings (solid lines) and N = infinite (dashed line) for the alternating parameter a = 0. 4, 1.5, 3.0 and 6.0. ferromagnetic chain and alternating ferro-antiferromagnetic systems of spins S = 1/2.19 Recently, an expression to fit ferroantiferromagnetic S = 1 chains was proposed by Borras- Table 2 Selected bond lengths (Å) and angles (8) for [{Ni2(Medien)2- (m1,1-N3)2(m1,3-N3)}n][ClO4]n Ni(1)–N(1) Ni(1)–N(2) Ni(1)–N(3) Ni(1)–N(11) Ni(1)–N(11A) Ni(1)–N(21) N(11)–N(12) N(12)–N(13) N(21)–N(22) N(1)–Ni(1)–N(2) N(1)–Ni(1)–N(3) N(1)–Ni(1)–N(11) N(1)–Ni(1)–N(11A) N(1)–Ni(1)–N(21) N(2)–Ni(1)–N(3) N(2)–Ni(1)–N(11) N(2)–Ni(1)–N(11A) N(2)–Ni(1)–N(21) N(3)–Ni(1)–N(11) N(3)–Ni(1)–N(11A) N(3)–Ni(1)–N(21) N(11)–Ni(1)–N(11A) N(11)–Ni(1)–N(21) N(11A)–Ni(1)–N(21) N(12)–N(11)–Ni(1) N(12)–N(11)–Ni(1A) Ni(1)–N(11)–Ni(1A) N(22)–N(21)–Ni(1) N(11)–N(12)–N(13) N(21)–N(22)–N(23) 2.065(3) 2.145(3) 2.076(3) 2.110(2) 2.116(3) 2.159(3) 1.179(3) 1.152(4) 1.163(4) 82.7(1) 98.0(1) 91.8(1) 86.8(1) 174.0(1) 83.7(1) 174.1(1) 102.5(1) 96.0(1) 95.0(1) 172.6(1) 87.7(1) 79.2(1) 89.7(1) 87.8(1) 124.8(2) 134.3(2) 100.8(1) 138.1(2) 179.7(3) 178.1(3) Ni(2)–N(4) Ni(2)–N(5) Ni(2)–N(6) Ni(2)–N(23) Ni(2)–N(31) Ni(2)–N(31B) N(22)–N(23) N(31)–N(32) N(32)–N(33) N(4)–Ni(2)–N(5) N(4)–Ni(2)–N(6) N(4)–Ni(2)–N(31) N(4)–Ni(2)–N(31B) N(4)–Ni(2)–N(23) N(5)–Ni(2)–N(6) N(5)–Ni(2)–N(31) N(5)–Ni(2)–N(31B) N(5)–Ni(2)–N(23) N(6)–Ni(2)–N(31) N(6)–Ni(2)–N(31B) N(6)–Ni(2)–N(23) N(31)–Ni(2)–N(31B) N(23)–Ni(2)–N(31) N(23)–Ni(2)–N(31B) N(32)–N(31)–Ni(2) N(32)–N(31)–Ni(2B) Ni(2)–N(31)–Ni(2B) N(22)–N(23)–Ni(2) N(31)–N(32)–N(33) 2.081(3) 2.118(3) 2.081(3) 2.142(3) 2.113(3) 2.112(3) 1.165(4) 1.194(3) 1.149(4) 82.9(1) 96.1(1) 86.5(1) 91.6(1) 172.8(1) 83.4(1) 100.4(1) 174.1(1) 94.3(1) 175.6(1) 95.0(1) 90.1(1) 81.4(1) 87.5(1) 91.4(1) 125.8(2) 121.0(2) 98.9(1) 125.7(2) 177.8(3) Almenar et al.,20 assuming that the ring of N = 5 pairs of S = 1 spins describes the chain behaviour satisfactorily.In this section, therefore, we felt it appropriate to extend those calculations and attempt to develop such an expression for the N = infinite extrapolation, in order to improve the expression for large a = JFM/|JAF| values.The Hamiltonian for the Heisenberg alternating ferroantiferromagnetic chain can be written as in eqn. (1) where N is H = 2 N21 S i = 1 JAFS2iS2i 1 1 1 JFMS2iS2i 2 1 (1) the number of spin pairs, JAF and JFM are the nearest neighbour antiferro- and ferro-magnetic exchange interactions. By using the usual computational technique, based on the calculation of the properties of finite rings of increasing size, we have determined the magnetic susceptibility of alternating ferro-antiferromagnetic chains for various a values.The series of calculations were made using the computer program CLUMAG, which uses the irreducible tensor operator (ITO) formalism.21 Owing to the diYculties associated with the large dimensions of the calculations and the required computing times, our calculations were only achieved up to 10 spins (N = 5), and by the same reason the influence in the low temperature region of the zero field splitting parameter D has not been considered.Fig. 3 reports the reduced magnetic susceptibility curves of the chains N = 2, 3, 4 and 5 for a = 0.4, 1.5, 3 and 6. Closer examination shows that when a is low (0.4) the two curves N = 4 and 5 are indistinguishable throughout the temperature range, whereas significant diVerences become observable when a increases especially at low temperature (below the reduced temperature, Tr = 1.3 for a = 1.5, 1.6 for 3 and 1.8 for 6), and the maximum deviation becames 2.9% for a = 6.Nevertheless, it seems clear that the unique sequence (N = infinite) will appear in the intermediate region bracketed by the curves for odd N and those for even N, exactly between N = 5 and 4, which led us to assume that the curves of half of the sum between N = 5 and 4,3908 J.Chem. Soc., Dalton Trans., 1998, 3905–3909 Table 3 CoeYcients for the polynominals for 0 £ a £ 2 and 2 £ a < 6 A B C D E 0£a£2 a0 0.610392 a1 22.56528 a2 3.3881 a321.68187 a4 0.310599 b0 0.141083 b1 0.503112 b2 20.687456 b3 0.331817 b4 20.0547117 c0 1.17598 c1 23.14941 c2 3.18665 c3 21.49803 c4 0.269656 d0 1.26115 d1 21.584444 d2 3.30872 d3 1.86917 d4 0.369253 e0 0.282429 e1 21.15824 e2 0.896513 e3 20.383948 e4 0.0743106 2 £ a £ 6 a0 38.728 a1 251.9772 a2 25.4004 a3 25.38647 a4 0.454228 b0 16.4703 b1221.8691 b2 10.4464 b3 22.1991 b4 0.185766 c0 32.1303 c1 242.9283 c2 20.6105 c 24.3098 c4 0.357483 d0 18.3973 d1 223.3778 d2 12.0541 d3 22.66954 d4 0.236077 e0 57.5268 e1 276.6243 e2 36.553 e3 27.6135 e4 0.630971 N• = [(N = 4 1 N = 5)/2], should describe the behaviour of the infinite chain with negligible uncertainty (dashed curves on Fig. 3). Based on this proposition, and applying the same strategy reported in the literature, is possible to generate an expression of cr, which depends on Tr and a, by fitting all the infinite theoretical susceptibility curves, eqn.(2) where cr = 3cm|JAF|/ cr = Tr 2 1 ATr 1 B Tr 3 1 CTr 2 1 DTr 1 E (2) 2Ng2mB 2 and the reduced temperature Tr is given by kT/|JAF|; A–E are the fitting coeYcients, which depend on a, and such dependence can be described by the use of a polynomial expression (3) to the fourth degree in a. (In eqn. (3) x0 etc. corre- X = x0 1 x1a 1 x2a2 1 x3a3 1 x4a4 (3) spond to a0–e0 from Table 3.) Two sets of coeYcients have been proposed according to the value of a, Table 3.The reduced susceptibility expression can be converted into magnetic susceptibility in the habitual form to give eqn. (4) where cm = 2Ng2mB 2 3kT 1 1 Ax 1 Bx2 1 1 Cx 1 Dx2 1 Ex3 (4) x = |JAF|/kT. The expression with the two sets of A–E coef- ficients is valid for kT/|JAF| > 0.22, as shown in Fig. 4, which gives the calculated magnetic susceptibility of the infinite chain with good approximation (R = 1.5 × 1025), even for high values of a.Fig. 4 Temperature dependence of cm (theoretical) versus T for N = infinite. The solid lines are the best fits calculated for diVerent values of the alternating parameter a = 0.2, 0.6, 1.2, 2.0, 3, 4, 5 and 6 (JAF = 250 cm21 and g = 2.35). Magnetic properties and magneto-structural correlations The cmT product and the molar magnetic susceptibilities vs. T in the 300–4 K range for [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n]- [ClO4]n 1 are plotted in Fig. 5, which shows an overall antiferromagnetic behaviour: cmT decreases on cooling from 1.34 cm3 K mol21 at 292 K and tends to zero at low temperature, showing a maximum of susceptibility at 50 K (1.30 × 1022 cm3 mol21). The fit of the experimental data with the above equation gave the best parameters JAF = 234.7, JFM = 138.2 cm21, a = 1.10, g = 2.34 and R = 4.2 × 1025. The strong ferromagnetic contribution of JFM to the global magnetic behaviour is reflected in the very slow decrease of cmT in the high temperature range 300– 150 K.Despite the mathematic quality of the fit, the sign and magnitude of the calculated coupling constants should be chemically reasonable in comparison with those for ferro- or anti-ferromagnetic models or well established experimental data. The best known interaction is the antiferromagnetic superexchange through the single end-to-end azido bridge, which has been modelled as a function of the two Ni–N–N bond angles and the t torsion angle defined as the angle between the mean planes Ni–N–N–N and N–N–N–Ni9.8 Analysis of published data indicates that the best fit JAF value of 234.7 cm21, the Ni–N–N bond angles 138.1/125.78 and the t torsion angle of 162.88 are consistent with the J value expected for a compound with a medium t angle.A useful comparison is oVered by the [{Ni(cyclam)(m1,3-N3)}n][ClO4]n homogeneous chain, which shows practically the same structural and magnetic parameters: Ni–N–N bond angles 140.7/128.28, t = 166.98 and J = 239.2 cm21.22 Less systematized data are available for the double end-on kind of bridge, but the best fit JFM lies in the typical range of J values for this kind of compound, for which J close to 140 Fig. 5 Molar magnetic susceptibility and the cmT product vs. temperature for [{Ni2(Medien)2(m1,1-N3)2(m1,3-N3)}n][ClO4]n. Solid lines show the best fit obtained by applying eqn. (3) (see text).J. Chem.Soc., Dalton Trans., 1998, 3905–3909 3909 cm21 is a normal value. The slightly lower Ni–N–Ni9 bond angle is not expected to be relevant in the light of the data summarized in Table 4, which show a similar J value for all the reported compounds with Ni–N–Ni9 bond angles between 100.8 and 104.98. This result is also consistent with the calculations performed by Ruiz et al.23 by means of density functional methods, which predict a maximum ferromagnetic coupling around the bond angle Ni–N–Ni9 = 1048 and a narrow range of this bond parameter (±88 approximately), in which minor J diVerences should be expected. Concluding remarks A new structurally and magnetically alternating 1-D nickel(II)– azido system has been fully characterized and its superexchange coupling constants have been determined as JAF = 234.7 and JFM = 138.2 cm21, a = 1.1.These results have been calculated by extrapolation of the expected properties for rings of increasing size up to N = 5 pairs of S =1.The best fit parameters are consistent with the values that may be expected on the basis of the structural bond parameters in the bridging region, which confirm the validity of the sign and magnitude of the J values found. One of the most interesting aspects of the present work is the application for the first time of the theoretical expressions to a system in which the two superexchange interactions are strong, the results lying in the optimum range of a, which permits an accurate determination of the J values.Acknowledgements A. E. and R. V. thank the Comisión Interministerial de Ciencia y Tecnologia (Grant PB096/0163) for support of this research. F. A. M. thanks Professor C. Kratky (University of Graz) for use of experimental equipment and the Osterreichische Nationalbank (Jubilaeumsfondsproject 6630) for financial support. References 1 W. E. Hatfield, in Extended Linear Chain Compounds, ed. J. S. Miller, Plenum, New York, 1983, vol. 3. 2 J. C. Bonner, in Magneto-Structural Correlations in Exchange Coupled Systems, eds. R. D. Willet, D. Gatteschi and O. Kahn, NATO ASI Series, Reidel, Dordrecht, 1984. Table 4 Bond angles and J superexchange parameters for the ferromagnetic [NiNi] pairs with a double end-on azido bridge and aliphatic blocking ligands {en = ethane-1,2-diamine, Medt = bis(3-aminopropyl)- methylamine, 232-tet = N,N9-bis(2-aminoethyl)propane-1,3-diamine, Me3[12]N3 = 2,4,4-trimethyl-1,5,9-triazacyclododec-1-ene} Compound [Ni2(en)4(m-N3)2][ClO4]2 [Ni2(Medpt)2(N3)2(m-N3)2](ClO4)2 [Ni2(232-tet)2(m-N3)2][ClO4]2 [Ni2(232-tet)2(m-N3)2][PF6]2 [Ni2(Me3[12]N3)(m-N3)2] Ni2(Medien)2(m-N3)2 fragment Ni–N–Ni/8 104.3 104.0 104.9 104.6 103.8 100.8 98.9 J/cm21 43.4 46.7 33.8 34.3 43.9 38.2 Ref. 14 15 16 17 16 This work 3 Physics in One Dimension, eds. J. Bernasconi and T. Schneider, Springer, Berlin, 1981. 4 Organic and Inorganic Low-Dimensional Crystalline Materials, eds.P. Delhaes and M. Drillon, Plenum, New York, 1987, vol. B, p. 168. 5 (a) M. I. Arriortua, R. Cortés, L. Lezama, T. Rojo and X. Solans, Inorg. Chim. Acta, 1990, 174, 263; (b) R. Cortés, J. I. Ruiz de Larramendi, L. Lezama, T. Rojo, K. Urtiaga and M. I. Arriortua, J. Chem. Soc., Dalton Trans., 1992, 2723; (c) J. Ribas, M. Monfort, C. Diaz, C. Bastos and X. Solans, Inorg. Chem., 1993, 32, 3557; (d) M. Monfort, J. Ribas and X. Solans, Inorg. Chem., 1994, 33, 4271; (e) P. Chauduri, T.Weyhermüller, E. Bill and K. Wieghardt, Inorg. Chim. Acta, 1996, 252, 195; ( f ) A. Escuer, I. Castro, F. A. Mautner, M. S. El Fallah and R. Vicente, Inorg. Chem., 1997, 36, 4633; ( g) J. Ribas, M. Monfort, R. Costa and X. Solans, Inorg. Chem., 1993, 32, 695. 6 M. A. S. Goher and F. A. Mautner, Croat. Chem. Acta, 1990, 63, 559; A. Escuer, R. Vicente, M. A. S. Goher and F. A. Mautner, Inorg. Chem., 1996, 35, 6386; G. De Munno, M. Julve, G. Viau, F. Lloret, J. Faus and D.Viterbo, Angew. Chem., Int. Ed. Engl., 1996, 35, 1807; R. Cortés, L. Lezama, J. L. Pizarro, M. I. Arriortua and T. Rojo, Angew. Chem., Int. Ed. Engl., 1996, 35, 1810; F. A. Mautner, R. Cortés, L. Lezama and T. Rojo, Angew. Chem., Int. Ed. Engl., 1996, 35, 78. 7 J. Ribas, M. Monfort, B. K. Ghosh, R. Cortés, X. Solans and M. Font-Bardia, Inorg. Chem., 1996, 35, 864 and refs. therein. 8 R. Vicente and A. Escuer, Polyhedron, 1995, 14, 2133; A. Escuer, R. Vicente, M. A. S. Goher and F.A. Mautner, Inorg. Chem., 1998, 37, 782. 9 (a) A. Escuer, R. Vicente, J. Ribas, M. S. El Fallah, X. Solans and M. Font-Bardia, Inorg. Chem., 1994, 33, 1842; (b) R. Vicente, A. Escuer, J. Ribas and X. Solans, Inorg. Chem., 1992, 31, 1726; J. Ribas, M. Monfort, B. K. Gosh and X. Solans, Inorg. Chem., 1994, 33, 2087. 10 G. M. Sheldrick, SHELXS 86, Program for the Solution of Crystal Structure, University of Göttingen, 1986. 11 G. M. Sheldrick, SHELXL 93, Program for the Refinement of Crystal Structure, University of Göttingen, 1993. 12 SHELXTL 5.03 (PC-Version), Program Library for the Solution and Molecular Graphics, Siemens Analytical Instruments Division, Madison, WI, 1995. 13 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 14 J. Ribas, M. Monfort, C. Diaz, C. Bastos and X. Solans, Inorg. Chem., 1994, 33, 484. 15 A. Escuer, R. Vicente, J. Ribas and X. Solans, Inorg. Chem., 1995, 34, 1793. 16 R. Vicente, A. Escuer, J. Ribas, M. S. El Fallah, X. Solans and M. Font-Bardia, Inorg. Chem., 1993, 32, 1920. 17 A. Escuer, R. Vicente, M. S. El Fallah, X. Solans and M. Font- Bardia, Inorg. Chim. Acta, 1996, 247, 85. 18 R. Cortés, M. Drillon, X. Solans, L. Lezama and T. Rojo, Inorg. Chem., 1997, 36, 677. 19 J. W. Hall, W. E. Marsh, R. R. Welles and W. E. Hatfield, Inorg Chem., 1981, 20, 1033; J. J. Borras-Almenar, E. Coronado, J. Curely, R. Georges and J. C. Gianduzzo, Inorg. Chem., 1994, 33, 5171; 1995, 34, 2699. 20 J. J. Borras-Almenar, J. M. Clemente-Juan, E. Coronado and F. Lloret, Chem. Phys. Lett., 1997, 275, 79. 21 D. Gatteschi and L. Pardi, Gazz. Chim. Ital., 1993, 123, 231. 22 A. Escuer, R. Vicente, J. Ribas, M. S. El Fallah and X. Solans, Inorg. Chem., 1993, 32, 1033. 23 E. Ruiz, J. Cano, S. Alvarez and P. Alemany, J. Am. Chem. Soc., in press. Paper 8/05247F

ISSN:1477-9226    

DOI:10.1039/a805247f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3911–3917 3911 Ionic structure in caustic aluminate solutions and the precipitation of gibbsite Helen R. Watling,a Sean D. Fleming,b Wilhelm van Bronswijk b and Andrew L. Rohl b a A. J. Parker Cooperative Research Centre for Hydrometallurgy, CSIRO Division of Minerals, PO Box 90, Bentley, Western Australia 6982 b A. J. Parker Cooperative Research Centre for Hydrometallurgy, School of Applied Chemistry, Curtin University of Technology, PO Box U 1987, Perth, Western Australia 6845 Received 23rd September 1998, Accepted 9th October 1998 The structure of caustic aluminate solutions in relation to the precipitation of gibbsite was investigated using vibrational spectroscopy and molecular dynamics simulations.Results from the molecular dynamics simulations indicate that aluminate ions form clusters as a function of time and that these clusters are stabilised by sodium ions. While the method used has the limitation that bond formation is forbidden, the predicted clustering would certainly facilitate polyaluminate anion formation. It is proposed that observed additional bands in vibrational spectra of concentrated aluminate solutions, as compared with those of dilute solutions, result from vibrations of these clusters (and any polyaluminate ions which arise from them).The absence of spectral features characteristic of a distinct interfacial aluminate layer at the growing crystal surface is explained by clustering throughout the bulk solution, and the participation of such clusters (and polyanions) in the growth process.Historically, comparisons have been made between vibrational spectra of solutions and those of aluminate compounds as a means of describing aluminate ion structures. Most authors concur with the interpretation of Moolenaar et al.1 that monomeric Al(OH)4 2 ions exist in equilibrium with [(OH)3- AlOAl(OH)3]22, the dimeric hydroxo ion.2–4 This conclusion is consistent with 27Al-solution NMR results which show that most of the aluminium in caustic aluminate solution has fourfold coordination.5,6 However, to limit the aluminate species distribution to the above two anions may be too narrow a view.Gibbsite, which is known to precipitate spontaneously from pure, homogeneous caustic solutions supersaturated with aluminium, 7 has a complex structure in which aluminium atoms have six-fold coordination and are linked to one another by double hydroxo bridges in a hexagonal layered structure.8 Clearly the spontaneous precipitation of gibbsite from solution must involve further changes in anionic structure (four-fold to six-fold aluminium coordination and oxo to hydroxo bridging) together with the formation of oligomeric intermediate species.There is little doubt that geometries calculated via quantum mechanics, with the subsequent prediction of vibrational frequencies, has been useful in describing the S4 symmetry of the Al(OH)4 2 monomeric ion.9,10 However, attempts to model other aluminate ions in solution are less convincing, perhaps because ab initio calculations for gas phase molecules are not entirely suitable for the assignment and interpretation of spectra of related solution species, or because researchers are predisposed towards Moolenaar’s1 “dimer” hypothesis. For example, quantum chemical calculations 11,12 have been used to explain the ultraviolet spectra of sodium aluminate solutions (or vice versa) in terms of three anions, Al(OH)4 2, Al2O(OH)6 22 and Al(OH)6 32; however, the UV spectrum of an aluminate solution arises from the soluble impurities rather than the aluminate ions in solution,13 which casts doubt on the authors’ interpretation. The simulation of liquids, one of the earliest applications of computer chemistry,14 has been developed to the point where it is now possible to simulate complex behaviour in complex solutions by using periodic cells containing of the order of several hundred atoms.15 Many recent studies on the nature of aqueous solutions have been accomplished using molecular dynamics.For example, Levitt et al.16 attempted to construct a model (mainly suitable for use with macromolecules) which would reproduce selected physical properties of water. Laaksonen et al.,17 who investigated the methane–water system, demonstrated the use of molecular dynamics to generated spatial distribution functions which describe the local solution structure.Mancera et al.18 focused on the eVect of temperature on the aggregation of methane in aqueous solution. For ionic systems, ion selectivity was examined by plotting the free energy profile of cations to a crown ether in water,19 while the hybrid QM/ MM technique was used to show that a solely classical approach was inadequate to describe the solvation of calcium ions in water.20 In this study, vibrational spectroscopy is combined with molecular dynamics simulations to investigate possible ion associations in aluminate solutions (synthetic Bayer liquors), at concentrations of industrial relevance.[The Bayer process is the most widely used method for the production of gibbsite, Al(OH)3, which is then calcined to alumina and electrolytically reduced to aluminium.] The results are discussed in relation to the mechanism of gibbsite crystal growth. Experimental Spectroscopy Caustic solutions were prepared using analytical grade sodium, potassium and caesium hydroxides. Aluminate solutions were prepared by dissolving aluminium wire (BDH product 10006) in the hot caustic solutions; the resulting liquors were filtered through 0.45 mm pore size membranes and diluted to volume with hot distilled water. Final solution concentrations were determined using inflection point titrimetry.21 Infrared spectra were collected at 4 cm21 resolution using a Bruker IFS-66 spectrometer, with DTGS detector, equipped with an Axiom Dipper-210 ATR probe with zinc selenide element.The useful working range was estimated to be 4000–600 cm21. Raman spectra of crystallizing solutions were collected at 4 cm21 resolution using a Bruker Near-Infrared RFS100 FT spectrometer, with NdYAG-1064 nm laser excitation (900 mW) and Ge-diode detector.3912 J. Chem. Soc., Dalton Trans., 1998, 3911–3917 Raman polarisation measurements were made on an ISA Labram 1B dispersive spectrometer using HeNe excitation (14 mW) and a 600 lines mm21, giving a resolution of ª2.7 cm21.Bands due to solvent and caustic solution were removed from aluminate solution spectra by subtracting the spectra of equivalent caustic solutions. Subtraction was deemed satisfactory when a straight baseline was obtained in the region of interest (450–800 cm21). Both subtracted and unsubtracted spectra were used to calculate polarisation ratios, and gave similar results.Molecular modeling The molecular dynamics simulations reported in this paper were conducted on Silicon Graphics workstations, using the Discover package from Molecular Simulations Incorporated, San Diego. The ESFF potential set 22 was chosen to represent all force field interactions, since the system contains two metallic atom types in addition to oxygen and hydrogen. The Verlet velocity method was employed to integrate the equations of motion. All simulations were conducted with the use of periodic boundary conditions, with the Ewald technique 23 employed to evaluate the van der Waals, and the coulomb interactions.Three systems were constructed, each possessing a composition similar to that of a synthetic Bayer liquor (7 [Al- (OH)4]2 , 5 OH2 , 12 Na1 and 192 H2O molecules). For each simulation, the molecules of the system were placed in a cubic lattice, the dimensions of which were selected to avoid overlap. A randomising algorithm was used to choose the location and orientation of each component whilst preserving the overall required ratio of species.Several hundred 1 fs steps at high pressure in an NPT ensemble (with velocity scaling temperature control) were then undertaken to obtain simulated liquors with the expected density. This gave cubic simulation boxes of approximately equal dimensions (L ª 18.5 Å), corresponding to a solution with [Al(OH)3] ª 1.8 M, [NaOH] ª 3.2 M and a density of approximately 1.2 g cm23. These concentrations are similar to that of a synthetic Bayer liquor at 80 8C.24 The final configurations of these simulations were then used as the starting point for a further 500,000 cycles (step size 1 fs) in an NVT ensemble.The first 200 ps were taken as an equilibration period, with data collection in the final 300 ps. The more robust Andersen temperature control method 15 was used for the equilibration while the more realistic Berendsen method14 was used for the data collection.Results Spectroscopy of aluminate solutions Raman spectra were obtained for caustic and aluminate solutions (Na, K, Cs) with a range of compositions (Fig. 1, Table 1). These exhibit the expected one- or three-band spectra for aluminate ion Al–O vibrations in the low-frequency region. A feature of the Raman spectra collected during an initial study 25 was the increased asymmetry and width of the sidebands (ª695 and ª535 cm21) which flank the most intense aluminate band attributed to Al(OH)4 2 (ª620 cm21).1 This, together with the observed shifts in band frequency at maximum intensity with changed solution composition, tended to support the hypothesis that each side band resulted from two or more vibrations of similar frequency.In the present study, further evidence for this is obtained from spectra of solutions containing one molar aluminium and very high sodium hydroxide concentrations (Fig. 1, g and h). At low caustic concentrations (4 molar) and one molar aluminium, only one aluminate band is observed (ª620 cm21), as expected; at this aluminium concentration the side bands are not apparent.However, a low intensity band centred at 555 cm21 becomes apparent in solution spectra as the caustic concentration is increased to 19 molar. This is not the same band as that which appears in spectra of concentrated aluminate solutions (ª535 cm21), and which exhibits a significant shift to lower frequencies as the aluminium concentration is increased (at constant caustic concentration) (Fig. 1f ). The appearance of the low intensity 555 cm21 band is consistent with the formation of small concentrations of Al(OH)6 32 in these extremely caustic solutions, but such a hypothesis needs to be confirmed by independent methods. In an attempt to resolve the side bands (ª695 and ª535 cm21) into their components, depolarization ratios for each of them were determined, the rationale being that not all component vibrations need have the same degree of symmetry (Table 1).Fig. 1 Raman spectra of aluminate solutions as recorded. a,b) Polarised and depolarised spectra, 5M sodium aluminate with 3M excess NaOH; c,d,e) Cs, K and Na aluminate solutions all 3M in aluminate with 2M excess caustic; f) 5M sodium aluminate with 3M excess NaOH; g) 1M sodium aluminate with 18M excess NaOH; h) 1M sodium aluminate with 4M excess NaOH. Table 1 Calculated depolarisation ratios, r, for aluminate ion vibrational bands Aluminate solution 5M NaAl(OH)4 3M NaOH excess 4M NaAl(OH)4 6M NaOH excess 4M KAl(OH)4 6M KOH excess 3M NaAl(OH)4 2M NaOH excess 3M KAl(OH)4 2M KOH excess 3M CsAl(OH)4 2M CsOH excess 1M NaAl(OH)4 4M NaOH excess 1M NaAl(OH)4 18M NaOH excess r ª695 cm21 ª0.25 ª0.20 ª0.15 ª0.20 ª0.3 ª0.3 — — r ª620 cm21 0.02 0.025 0.02 <0.01 <0.01 <0.01 <0.01 0.03 r ª535 cm21 0.02 0.025 0.02 <0.01 <0.1 <0.1 — <0.1 a a 555 cm21 band.J. Chem.Soc., Dalton Trans., 1998, 3911–3917 3913 At the same time the eVects of possible cation–aluminate ion pairing on spectra were investigated using a series of aluminate solutions prepared with sodium, potassium or caesium hydroxides. In respect of depolarization ratios, illustrated for a sodium aluminate solution (Fig. 1, a and b), it is significant that all three bands in the Raman spectrum of an aluminate solution are the result of symmetrical bond vibrations, irrespective of solution concentration or cation.Only one of them, the 620 cm21 band, can be attributed to the aluminate monomer Al(OH)4 2 which has S4 symmetry.10 The other bands must arise from one or more additional aluminate species (new species) in equilibrium with the monomer in concentrated aluminate solutions. Unfortunately, although the degree of fit obtained by modeling of the spectra is such as to suggest that each of the side bands has more than one component, all components were shown to have a strong degree of symmetry and diVerences in polarised spectra did not enable the deconvolution of the bands.Raman spectra of alkali metal hydroxides also exhibit a very broad low-intensity, low-frequency band at ª300 cm21 which has been attributed to MOH?H2O, in which an almost symmetric O–M–O linkage exists.26 The intensity of this band is reduced when aluminium is introduced into solution, a feature which has been interpreted as indicating that cation aYnity is greater for the aluminate ion than for the hydroxide ion.27 Aluminate and hydroxide solution spectra collected in the present study confirm the above observations and are consistent with the interpretation (Fig. 1, g and f). The eVects of the diVerent cations on aluminate vibrational bands are small (Fig. 1, c, d and e). Apart from a systematic decrease in the 620 cm21 band intensity accompanied by a small increase in width at half height in the order Na > K > Cs (Fig. 2), spectra are remarkably similar across the full frequency range. This relationship is reflected in the combined area under the curve (three bands) for these spectra. Comparisons between spectra of sodium, potassium and caesium aluminate solutions indicate that, if these spectral diVerences are due to cation–- aluminate ion pairing, then pairing is strongest in the sodium system. Bands associated with aluminate ion vibrations were modeled and integrated areas for the three bands were determined for concentrated solutions of sodium and potassium aluminates (Fig. 3). Consistency is achieved in that, across the aluminium concentration range investigated, the observed decrease in the 620 cm21 band area for sodium relative to potassium (at equal aluminium concentration) can be correlated with an increased area in the 535 cm21 band. Thus it appears that one or more new species are promoted at the expense of the Fig. 2 Integrated areas of aluminate Raman-active vibrational bands (combined 535, 620 and 695 cm21 and 620 cm21) for sodium, potassium and caesium aluminate solutions with total cation concentration 5M.aluminate ion in sodium solutions compared with potassium solutions. Results from the previous study25 had led to the belief that both the 535 and 695 cm21 bands arose from vibrations of the same new species, so that changes in the 695 cm21 band area should correlate with those of the 535 cm21 band. This does not seem to be the case, but the analysis should be interpreted with caution because this is the least well defined of the three bands under investigation, which introduces a significant error into the model.The gibbsite–aluminate interface Concentrated alkaline aluminate solutions supersaturated with aluminium are inherently unstable. From the time of preparation changes take place in these solutions which will result in the precipitation of bayerite (a gibbsite polymorph formed at ambient temperature) or gibbsite (>60 8C).In an attempt to monitor these changes, Raman spectra of a freshly prepared, filtered solution were collected as a function of time until precipitation occurred (Fig. 4, a and b). Only bands attributable to the aluminate solution or to bayerite were observed when precipitation occurred and no changes in solution spectra were observed up to that point. However, in the same time frame, the intensity of the Rayleigh line increased significantly (Fig. 5). Indeed, this was the only evidence of any changes taking place in the solution and is attributed to increased Rayleigh scattering by small particles which are being formed. As Rayleigh scatter is proportional to a particle’s polarisability it increases as a function of the square of the volume, i.e. the sixth power of its diameter.28 However, as the number of particles decreases, with the formation of fewer, larger aggregates, the intensity is expected to increase as a function of the square of the diameter of the particle for simple aggregation (i.e.monomer, dimer, tetramer, etc.). Rayleigh scatter is the major scattering mechanism when the particle diameter (D) is less than approximately one twentieth of the wavelength (l) being scattered. As the diameter of the particles increases the Rayleigh–Gans region, 0.1 < D/l < 1, is reached and the 1808 back scatter intensity reduces relative to the forward scatter, until at D @ l the back scatter intensity is typically less than one tenth of the forward scatter.28 The observed initial rapid increase in Rayleigh scatter (Fig. 5) is thus as expected for particle growth and the subsequent decrease in rate indicates that the Rayleigh–Gans region has been reached, or that the rate of particle growth has slowed.Attenuated total reflectance (ATR) infrared spectroscopy Fig. 3 Integrated areas of aluminate Raman-active vibrational bands in spectra of concentrated solutions (total cation concentration 10M).Cations = Na and | = K; frequencies m = 620 cm21, d = 695 cm21 and s = 535 cm21.3914 J. Chem. Soc., Dalton Trans., 1998, 3911–3917 was used to study the gibbsite aluminate solution interface in aqueous solution. Spectra were obtained of supersaturated aluminate solutions (60 8C) using one of two configurations (Fig. 6). In the first, the ATR element was immersed in a slurry of gibbsite particles under conditions known to promote gibbsite crystal growth; in this way the interface was sampled through the solution at the particle surface (Fig. 6a).In the second, a layer of gibbsite scale was deposited on the ATR element, which was then immersed in a solution, again under conditions which would promote further gibbsite deposition onto the scale layer (Fig. 6b). Here, the interface was sampled through the gibbsite scale. With either configuration, the spectrum collected is representative of solution and gibbsite 1–2 mm from the ATR element and must include both the interfacial layer at the growing crystal surface and bulk solution.Spectra collected from both experiments (Fig. 4, c, d and e) are interesting in that they exhibit only those bands associated with aluminate solutions and those associated with gibbsite. In addition, solution infrared spectra collected using a trans- Fig. 4 Raman and infrared spectra obtained from interface studies. a,b) Initial Raman spectrum of a supersaturated sodium aluminate solution and that obtained as precipitation commenced; c) infrared spectrum of freshly prepared sodium aluminate solution; d,e) spectra collected of a gibbsite slurry and of the solution through a layer of gibbsite scale.Fig. 5 Rayleigh scattering as a function of time (state of aggregation). mission cell are identical with those collected using the ATR probe. No spectral features have been identified that can be attributed to an interfacial aluminate layer which is distinguishable from the bulk solution.Molecular dynamics simulations During the 300 ps data collection period, the coordinates of all atoms in the simulation cells were stored at intervals of 100 fs. Thus, each simulation yielded 3000 cell snapshots, or frames, which were analysed for sodium and aluminate structuring. The first measurements made were the distances from the centre of each aluminate ion to all the sodium cations. Analysis of this data yielded the sodium aluminate radial distribution function (RDF) shown in Fig. 7. In order to attach some significance to the result, the RDFs for the first few hundred steps of the equilibration period were constructed for each simulation; these plots contained no significant peaks. Hence the peaks in Fig. 7, which occur at an aluminate to cation separation of around 3.5 Å, are the result of ion pairing. While the RDFs for simulation 1 and simulation 3 exhibit very similar behaviour, the second simulation experiences a greater degree of ion pairing.In addition, past the first large peak in the RDFs, the rest of the distributions appear to possess additional smaller peaks. While these are not as distinct, the deviation from a smooth curve suggests that some longer range solution structuring is also occurring. Further investigation was facilitated by the construction of radial distribution functions for aluminate Al–Al separations (Fig. 8). As in the case of the previous plot, the initial RDF for the equilibration period of each simulation exhibited no prior structuring. In Fig. 8, all three distributions possess a peak at around 5 Å, although it is evident that simulation 2 experiences the greatest degree of aluminate ‘clustering’. It is these clustered aluminates, each with their own shell of paired cations, that are responsible for the smaller peaks in Fig. 7. In Fig. 9, we have constructed a graph of the number of cations paired with each aluminate, versus the proportion of occurrence.This data was obtained by recording the number of cations paired to each aluminate, and summing over every frame in the data collection period. Thus, we normalise the Fig. 6 ATR-FTIR study of the gibbsite–aluminate ion interface. a) Sampling the gibbsite surface through the aluminate solution; b) sampling the aluminate solution through a gibbsite scale layer. Fig. 7 Aluminate to sodium radial distribution functions for the data collection period.J.Chem. Soc., Dalton Trans., 1998, 3911–3917 3915 graph with the product of the number of aluminate ions (7) and the number of frames (3000). In addition, with the use of Fig. 7, we have defined a paired cation as being closer than 4.1 Å to the Al atom of an aluminate ion. The data in Fig. 9 indicate that the aluminate ions in simulations 1 and 3 are paired to an average of between two and three cations. For simulation 2, there is a large proportion of aluminates with three or more associated cations.As each cell has a composition of 7 aluminates and 12 sodiums, the cations are clearly being shared by multiple aluminate ions which have come into proximity. This is particularly evident in simulation 2. Further information emerges from the aluminate cluster size distribution shown in Fig. 10. A cluster is defined as a group of linked aluminate monomers with Al–Al separations of not more than 5.5 Å. The data was collected by cumulatively recording the size of the cluster each aluminate belongs to in Fig. 8 Aluminate separation radial distribution functions for the data collection period. Fig. 9 The extent of simulation time (data collection period) for which the aluminates experience diVerent numbers of paired cations. An ion pair is judged as having a separation of no greater than 4.1 Å. Fig. 10 The extent of simulation time (data collection period) which each aluminate spends in diVerent sized clusters.A cluster is a group of linked aluminate ions, with a separation less than 5.5 Å. each simulation frame. Thus, the distribution was normalised with the same value employed to normalise the data in Fig. 9. This figure suggests that simulations 1 and 3 experience relatively small amounts of aluminate clustering, with a large fraction of aluminates remaining ‘isolated’ during the simulation. However, the second simulation has formed very large clusters that are stable enough to account for almost half of the simulation time.As the second simulation also possesses the highest number of shared cations, it is clear that these sodium–- aluminate bridges are playing an important role in the formation of clusters. The data used to generate Fig. 9 was examined in order to establish how the number of paired cations changes as a function of time. This analysis indicated that, in general, the cations associated with an aluminate are not rigidly bound.Thus, throughout the entire simulation, there is a random fluctuation in the number of cations paired with a particular aluminate. The mean value of these oscillations in associated cations was reported earlier as being between two and three. This mobility of the bridging cations is an important feature, and may have implications in the formation of polyaluminate species. Collectively, the data in Fig. 7 to 10 may be thought to indicate that the sodium and aluminate ions evolve, from initial isolation, into subsequently larger and larger groups. These groups being characterised by the ion pairing and clustering discussed previously.This implies that, if given suYcient time, the simulations would exhibit the form seen in Fig. 11, where all aluminates are bound into a ‘super’ cluster. However, this is not the case. An examination of the changes in cluster sizes during the data collection period, has indicated that simulations 1 and 3 experience varying oscillations in both the sizes and proportions of all clusters.This is also evident in the proportions of isolated aluminates, which fluctuates from between 40% to 80% for each frame, and only overall yielding an average of around 60% (cf. Fig. 10). Thus, at some simulation point, the numbers and sizes of all clusters present are randomly distributed, and not proportional to the age of the simulation. An analysis of the individual motions of cations for each simulation was conducted.This revealed that, in simulation 2, a stable cluster of size 7 was formed from a size 6 cluster at approximately two thirds the way through the data collection period. Whilst the formation of the larger cluster was preceded by an intake of cations, the cluster subsequently underwent a Fig. 11 A snapshot of simulation 2 at 200 ps through the data collection period. This illustrates the maximal clustering of all seven aluminates.3916 J. Chem. Soc., Dalton Trans., 1998, 3911–3917 period of stabilisation, coinciding with a steady increase in the number of free cations in solution.This size 7 cluster in simulation 2 dominated the cell for around 80 ps. However, by the end of the data collection period it had broken up into smaller units. This confirmed the earlier observation that there is no correlation between the age of the simulation and the cluster size distribution. In addition, as cluster destruction and new groupings of monomers occur continuously throughout the simulation, the aluminate–aluminate binding energy must be less than the thermal energy.Discussion The reaction for the dissolution of bauxite or the precipitation of gibbsite from hot alkaline aluminate solutions in the Bayer Process is often represented as Al(OH)4 2 NaOH Al(OH)3 1 OH2 but this equation fails to indicate the complexity of even the simplest case where gibbsite nuclei (Al(OH)3)n form spontaneously in a supersaturated aluminate solution.Homogeneous gibbsite nucleation must involve processes such as (i) ordering of aluminate ions in solution, (ii) formation of hydroxo-bridges between Al atoms, (iii) a change from four- to six-fold Al coordination geometry, (iv) release of sodium and hydroxyl ions from the nucleating crystal to the bulk solution, a consequence of which is (v) the separation of solid phase from solution (dehydration). Even in the Bayer process, where solutions are seeded and growth consists mainly of the incremental addition of gibbsite to seed crystal surfaces (during agglomeration, growth and secondary nucleation), aluminate ions must undergo similar changes with the consequent release of hydroxyl and sodium ions to solution.Thus, any mechanism that is proposed for the precipitation of gibbsite must account for these processes, while being consistent with experimental observations. The data collected in the present study contribute mainly to an understanding of solution ordering.Observed changes in spectra with increasing concentration, such as frequency shifts at maximum intensity and loss of proportionality with band intensities,25 and the appearance of new vibrational bands or the increased asymmetry in the shape of existing ones (all of which are shown to be symmetric vibrations), cannot be explained on the basis of the monomer or its sodium ion pair. Thus it is concluded that, while the monomeric Al(OH)4 2 ion or its sodium ion pair predominates in dilute aluminate solutions, at least one new species, presumably a polyaluminate, forms in concentrated solutions.While the structures of such polyaluminate species cannot be described solely on the basis of the spectroscopic data, it is expected that the aluminium coordination in them remains four-fold and aluminium atoms are linked by oxo-bridges. The maximum concentration of Al(OH)4 2 is estimated to occur at ª4–5 molar aluminium in solution, above which concentration the new species become relatively more abundant at the expense of the monomer.25 Results obtained from the molecular dynamics simulation tend to support the spectroscopic evidence.Whereas it has been assumed that monomeric Al(OH)4 2 ions would be distributed randomly (homogeneously) in solution, the simulations predict that they aggregate into clusters which are stabilised by sodium ions. Analysis has also indicated that the sodium ions are quite mobile with excess cations being ejected from the cluster as it stabilises.The main limitation of the simulation method used is that bonds cannot be formed or broken. So that, although aluminate monomers are predicted to form clusters, facilitating the formation of polyaluminate species, no information about either the bonding between aluminate ions or the aluminium geometry is obtained. However, as noted above, one of the intermediate steps in gibbsite precipitation involves the release of sodium ions to the bulk solution.This may already be partly accomplished through the predicted clustering mechanism. The interfacial layer at the surface of a seed crystal is indistinguishable, spectroscopically, in structure from that of the bulk solution. There are no significant changes in vibrational spectra of aluminate solutions, in the Al–O band frequency region, during the induction period leading to the spontaneous precipitation of gibbsite. An underlying assumption concerning the interface experiments was that such a layer with a “diVerent” structure should exist, and thus the inability to detect it was due to the insensitivity of the spectroscopic technique.However, if aluminate clusters are formed in the bulk solution as predicted, and if these participate in the growth of gibbsite on seed crystal surfaces, the absence of characteristic changes in spectra of the interfacial layer are explained. The additional sidebands in spectra of concentrated aluminate solutions are representative of both polyaluminate clusters in solution and at the gibbsite–solution interface where they participate in the growth process.The one significant change in spectra collected during the induction period up to the point where particles were visible as a faint suspension was the increased intensity of Rayleigh scattering. Recent additional data from multiangle laser light scattering (MALLS)29 also showed a progressive increase in light scattering prior to the appearance of suspended particles, consistent with a nucleation mechanism. In that case, particles were estimated to be ª180 nm diameter, but according to Rossiter et al.7 this might be an overestimation.The estimated diameter is approximately one sixth of the wavelength of the laser used (1064 nm) and confirms that the increased scattering reported in Fig. 5 is due to the Rayleigh and Rayleigh–Gans phenomena and that the Tyndall scattering region (D > l) has not been reached.The important point is that such particles could not form if the aluminate ion clustering, as predicted by the simulation, did not take place. Conclusion The data obtained from a spectroscopic study of caustic aluminate solutions supersaturated with aluminium during the period leading up to the spontaneous nucleation of gibbsite in them, and of the gibbsite–solution interface in seeded aluminate solutions, have been interpreted with the aid of molecular dynamics simulations of a solution of equivalent composition. Predictions from the molecular dynamics simulations include the clustering of aluminate ions in solution, and the stabilisation of these clusters by sodium ions.A limitation of the method is that bond formation is not permitted, but the predicted clustering would certainly facilitate polyaluminate anion formation. It is thought that the additional spectral bands observed in Raman and infrared spectra of concentrated aluminate solutions are due to vibrations of these clusters (and any polyaluminate anions which form from them).The absence of spectral features characteristic of a distinct interfacial aluminate layer at the growing crystal surface is explained by clustering throughout the bulk solution, and the participation of such clusters (and polyanions) in the growth process. Acknowledgements We would like to thank Professor C. R. A. Catlow for invaluable discussions on the modelling techniques used in this work.The financial support of the Australian Government through its Cooperative Research Centres Program is also gratefully acknowledged.J. Chem. Soc., Dalton Trans., 1998, 3911–3917 3917 References 1 R. J. Moolenaar, J. C. Evans and L. D. McKeever, J. Phys. Chem., 1970, 74, 3629. 2 N.-Y. Chen, M.-X. Liu, Y.-L. Cao, B. Tang and M. Hong, Sci. in China, Ser. B, 1993, 36, 32. 3 L. A. Myund, V. M. Sizyakov, M.K. Khripun and A. A. Makarov, Russian J. Gen. Chem., 1995, 65, 826; Zh. Obshch. Khim., 1995, 65, 911. 4 L. A. Myund, V. M. Sizyakov, K. A. Burkov, V. O. Zakharzhevskaya and O. A. Borzenko, Russian J. Gen. Chem., 1995, 68, 1721; Zh. Obshch. Khim., 1995, 68, 1964. 5 J. W. Akitt and W. Gessner, J. Chem. Soc., Dalton Trans., 1984, 147. 6 J. W. Akitt, W. Gessner and M. Weinberger, Magn. Reson. Chem., 1988, 26, 1047. 7 D. S. Rossiter, P. D. Fawell, D. Ilievski and G. M. Parkinson, J. Cryst. Growth, 1998, 191, 521. 8 H. Saalfeld and M. Wedde, Z. Kristallogr., 1974, 139, 129. 9 E. R. Lippincott, J. A. Psellos and M. C. Tobin, J. Chem. Phys., 1952, 20, 536. 10 A. C. Hess, P. F. McMillan and M. O’Keefe, J. Phys. Chem., 1988, 92, 1785. 11 H. Liu, S. Huang and N. Chen, Trans. Nonferrous. Met. Soc. China, 1992, 2(3), 43. 12 N. Chen and H. Liu, J. Mol. Struct. (THEOCHEM.), 1994, 305, 283. 13 P. Sipos, P. M. May, G. T. Hefter and I. Kron, J. Chem. Soc., Chem. Commun., 1993, 2355. 14 D. Frenkel and B. Smit, Understanding Molecular Simulation, Academic Press, San Diego, CA, 1996. 15 M. P. Allen and T. D. Tildesley, Computer Simulation of Liquids, Oxford University Press, Oxford, 1997. 16 M. Levitt, M. Hirshberg, R. Sharon, K. E. Laidig and V. Daggett, J. Phys. Chem. B, 1997, 101, 5051. 17 A. Laaksonen, P. G. Kusalik and I. M. Svishchev, J. Phys. Chem. A, 1997, 101, 5910. 18 R. L. Mancera, A. D. Buckingham and N. T. Skipper, J. Chem. Soc., Faraday Trans., 1997, 93, 2263. 19 L. X. Dang, J. Am. Chem. Soc., 1995, 117, 6954. 20 A. Tongraar, K. R. Liedl and B. M. Rode, J. Phys. Chem. A, 1997, 101, 6299. 21 G. Woods, H. Watling and P. Smith, Analysis of Bayer type liquors by inflection point titrations, CSIRO Division of Mineral Products Confidential Report No IR/P-001, 1994. 22 S. Barlow, A. L. Rohl, S. G. Shi, C. M. Freeman and D. O’Hare, J. Am. Chem. Soc., 1996, 118, 75. 23 D. M. Heyes, J. Chem. Phys., 1981, 74, 1924. 24 A. S. Russell, J. D. Edwards and C. S. Taylor, J. Met., 1955, 7, 1123. 25 H. Watling, Appl. Spectrosc., 1998, 52, 250. 26 M. Moskovits and K. H. Michaelian, J. Am. Chem. Soc., 1980, 102, 2209. 27 S. K. Sharma and S. C. Kashyap, J. Inorg. Nucl. Chem., 1972, 34, 3623. 28 T. Allen, Particle size measurement, Chapman and Hall, London, 4th edn., 1990, p. 487. 29 P. D. Fawell and H. R. Watling, Appl. Spectrosc., 1998, 52, 1115. Paper 8/07420H

ISSN:1477-9226    

DOI:10.1039/a807420h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3919–3925 3919 Structural, spectroscopic and electrochemical studies of binuclear nickel(II) complexes of bis(pentadentate) ligands derived from bis(1,4,7-triazacyclononane) macrocycles Suzanne J. Brudenell,a Leone Spiccia,*a Alan M. Bond,a Peter C. Mahona and David C. R. Hockless b a Department of Chemistry, Monash University, Clayton, Victoria 3168, Australia. E-mail: leone.spiccia@sci.monash.edu.au b Research School of Chemistry, Australian National University, Canberra, ACT 2601, Australia Received 3rd August 1998, Accepted 9th October 1998 Binuclear nickel(II) complexes of bis(pentadentate) ligands generated by functionalisation of the four secondary nitrogens in bis(tacn) macrocycles with 2-pyridylmethyl arms have been prepared, isolated and characterised.The pentadentate compartments are linked by (CH2)2 {tmpdtne, [Ni2(tmpdtne)(OH2)2][ClO4]4?5H2O 1}, (CH2)3 {tmpdtnp, [Ni2(tmpdtnp)(OH2)2][ClO4]4 2}, (CH2)4 {tmpdtnb, [Ni2(tmpdtnb)(OH2)2][ClO4]4 3}, CH2C6H4CH2-m {tmpdtnm-X, [Ni2(tmptdtm-X)(OH2)2][ClO4]4?3H2O 4} and CH2CHOHCH2 {tmpdtnp-OH, [Ni2(tmpdtnp-OH)(OH2)2][ClO4]4 5} bridging units.Single crystal X-ray diVraction studies of 3 have confirmed that each nickel(II) centre is in a distorted octahedral geometry defined by five N-donors from a pentadentate compartment of the ligand and an oxygen donor from a water ligand. Tetragonal elongation of the geometry is evidenced by the axial Ni–N distances, which are ca. 0.05 Å longer than the three equatorial Ni–N distances. The two pentadentate compartments are oriented in an anti configuration with the two aqua ligands pointing away from each other. Cyclic and square-wave voltammetric studies on 1–5 indicate that the complexes undergo oxidation to the nickel(III) state in two overlapping 1e2 processes (i.e. NiIINIII æÆ NiIINiIII æÆ NiIIINiIII). Two partially resolved 1e2 oxidation waves were observed for 1 and 5 while for 2–4 two 1e2 processes were indicated by fact that the signals were broader than expected for a single 2e2 process.Introduction Nickel is recognised as an essential trace element for bacteria, plants and animals.1–4 The active sites of enzymes such as urease, carbon monoxide dehydrogenase, [NiFe]-hydrogenase and methyl-S-coenzyme-M methylreductase are known to contain nickel centres, which are intimately involved in the catalytic cycles.5 Until the recent reports of the structures of urease 6 and [NiFe]-hydrogenase,7 X-ray structural data on nickel-enzymes was lacking.The quest for information about the structures and modes of action of nickel active sites continues to generate interest in the development of synthetic models for the various nickel biosites and in the co-ordination chemistry of nickel in general.5,8–13 Apart from their potential to form the basis of models for the polynuclear active sites of Ni containing enzymes, complexes of macrocyclic ligands and their derivatives are ideal for studying magnetic exchange interactions 9,10 and the redox properties of nickel(II) centres in close proximity.11 Ciampolini et al.11 recognised the potential for homobimetallic complexes of bis(macrocyclic) ligands to be ideal models for the study of mutual electrostatic eVects in twocentre redox systems. Accordingly, a number of binuclear nickel(II) complexes based on bis(cyclam) ligands were synthesized in which the Ni ? ? ? Ni separation could be tuned through the use of aliphatic and aromatic bridges of varying length.Electrochemical studies of the stepwise two-electron oxidation of these dinickel(II) complexes revealed that, as the metal–metal separation increased, a reduction in electrostatic repulsion between proximal nickel(II) centres facilitated the oxidation of the complexes. Investigations of the corresponding bis(tacn) macrocycles (tacn = 1,4,7-triazacyclononane) have revealed that the ethane-linked bis(tacn) ligand, dtne, forms exclusively the mononuclear nickel(II) sandwich complex and there was no evidence for formation of the binuclear compound, even when an excess of nickel(II) salt was used.12 In contrast, the o-xylene bridged bis(macrocycle) is able to behave as either a hexadentate ligand, co-ordinating to one NiII, or as a bis(tridentate) ligand, co-ordinating to two NiII while the m- and p-xylene bridged ligands yield only the binuclear complex.13 Although these types of bis(tacn) ligands have for example, been applied successfully in the synthesis of bridged binuclear copper complexes and in reactivity studies related to oxygen transport and activation by binuclear copper biosites 14 few attempts have been made to prepare bridged binuclear nickel complexes. One exception is the use of a novel bis(Me2tacn)calix[4]arene ligand which, by virtue of the calixarene linking unit, formed a unique ferromagnetic binuclear nickel(II) complex bridged by three azide ligands co-ordinating in the end-on mode.9 Apart from a binuclear complex incorporating a bis(tacn) ligand with alcohol pendant arms,15 no other nickel(II) complexes of bis(tacn) ligands with potentially co-ordinating pendant arms have been described.We have recently reported two series of bis(pentadentate) ligands, obtained by attaching either 2-pyridylmethyl 16,17 or acetate 18 pendant groups to each secondary nitrogen in bis(tacn) macrocycles, and the corresponding binuclear copper(II) complexes. The complexes of ligands bearing 2-pyridylmethyl arms were found to exhibit interesting variations in spectroscopic and electrochemical properties with Cu ? ? ? Cu separation tunable by the type of spacer used.We report here the synthesis and characterisation of the first series of binuclear nickel(II) complexes of bis(pentadentate) ligands incorporating 2-pyridylmethyl pendant arms (Scheme 1).Experimental Materials and reagents Reagent or AR grade materials were used throughout the study. The ligands 1,2-bis[4,7-bis(2-pyridylmethyl)-1,4,7-triazacyclonon- 1-yl]ethane (tmpdtne),16 1,3-bis[4,7-bis(2-pyridyl-3920 J. Chem. Soc., Dalton Trans., 1998, 3919–3925 methyl)-1,4,7-triazacyclonon-1-yl]propane (tmpdtnp),16 1,4- bis[4,7-bis(2-pyridylmethyl)-1,4,7-(triazacyclonon-1-yl]butane (tmpdtnb),16 1,3-[4,7-bis(2-pyridylmethyl)-1,4,7-triazacyclonon- 1-ylmethyl]benzene (tmpdtnm-X) 17 and 1,3-bis[4,7-bis(2- pyridylmethyl)-1,4,7-triazacyclonon-1-yl]propan-2-ol (tmpdtnp- OH)17 were all prepared by published methods.Physical measurements Infrared spectra were measured on a Perkin-Elmer 1600 FTIR spectrometer as KBr pellets or Nujol mulls, and electronic spectra on a Cary 5G spectrometer. Electron microprobe analyses were made with a JEOL JSM-1 scanning electron microscope through an NEC X-ray detector and pulse processing system connected to a Packard multichannel analyser.Microanalyses were performed by Chemical and Micro-Analytical Services (CMAS) of Melbourne, Australia. Room temperature magnetic moments were determined by the Faraday method. Diamagnetic corrections were made using Pascal’s constants. Cyclic (scan rate 100 mV s21) and square-wave (period 30 ms) voltammograms were recorded in dry, nitrogen degassed acetonitrile solutions (ª0.5 mmol dm23) with tetrabutylammonium perchlorate (0.1 mol dm23) as the supporting electrolyte on a Cypress CS 1090 system connected to a platinum macrodisc working electrode (r = 0.8 mm), platinum auxiliary electrode and Ag–Ag1 (10 mmol dm23 AgNO3) reference.The voltammetrically reversible oxidation of ferrocene (Fc) was used as an internal standard and potentials are reported relative to the Fc–Fc1 couple. CAUTION: Although no problems were encountered in this work, transition metal perchlorates are potentially explosive. They should be prepared in small quantities and handled with care.Scheme 1 N N N N N N N N Ni N N N N N Ni N OH H2O OH2 N N N N (1) 2 NiCl2.6H2O (2) NaClO4, CH3CN–water 4+ N N R R = CH2CH2 R = CH2CH2CH2 R = CH2CH2CH2CH2 R = R R = 1 2 3 4 5 Synthesis of complexes [Ni2(tmpdtne)(OH2)2][ClO4]4?5H2O 1. To a dark red-brown solution of tmpdtne (1.70 g, 2.62 mmol) in ethanol (20 cm3) was added a slight excess of NiCl2?6H2O (1.31 g, 5.51 mmol). The resulting brown solution was diluted with water (1 dm3) and loaded on to a Sephadex SP-C25 cation-exchange column.After washing the column thoroughly with distilled water two bands were visible. The first, faint pink, corresponded to [Ni(tacntmp)]Cl2 impurity and was eluted with a 0.1 mol dm23 NaClO4 aqueous solution. The major pink band, corresponding to the desired binuclear nickel(II) complex, was eluted with a 1 mol dm23 NaClO4 solution consisting of a 70 : 30 mixture of water and CH3CN. Acetonitrile was added to prevent the complex from crystallising on the resin.Following elution, the solution was concentrated yielding a pink solid. In some syntheses sodium perchlorate was added to enhance precipitation. This was recrystallised from a CH3CN–water mixture to give a pink-purple crystalline product (yield 1.72 g, 55%) {Found: C, 35.3; H, 5.2; N, 10.9. Calc. for [Ni2C38H52- N10)(OH2)2][ClO4]4?5H2O: C, 35.4; H, 5.1; N, 10.9%}. Electron microprobe: Cl : Ni ratio 2 : 1. Selected IR bands (KBr, cm21): 3418 (br), 2930m, 1608s, 1467m, 1448w, 1088vs, 768m and 629s.Magnetic moment: meff (292 K) = 2.81 mB per NiII. [Ni2(tmpdtnp)(OH2)2][ClO4]4 2. The procedure used to prepare complex 1 was followed except for tmpdtnp (1.10 g, 1.67 mmol) and NiCl2?6H2O (0.83 g, 3.50 mmol). Recrystallisation from a CH3CN–water mixture gave a purple crystalline solid (yield 1.05 g, 52%) {Found: C, 36.6; H, 5.0; N, 10.8. Calc. for [Ni2(C39H54N10)(OH2)2][ClO4]4: C, 36.6; H, 4.5; N, 11.0%}. Electron microprobe: Cl :Ni ratio 2 : 1.Selected IR bands (KBr, cm21): 3421 (br), 2928w, 1609m, 1467s, 1447m, 1088vs, 767m and 627s. Magnetic moment: meff (295 K) = 2.91 mB per NiII. [Ni2(tmpdtnb)(OH2)2][ClO4]4 3. The procedure used to prepare complex 1 was followed except for tmpdtnb (1.00 g, 1.49 mmol) and NiCl2?6H2O (0.74 g, 3.13 mmol). Purple crystals suitable for X-ray crystallography were grown from a CH3CN– water mixture (yield 1.01 g, 57%). {Found: C, 40.0; H, 4.8; N, 11.5.Calc. for [Ni2(C40H56N10)(OH2)2][ClO4]4: C, 39.1; H, 4.9; N, 11.4%}. Electron microprobe: Cl : Ni ratio 2 : 1. Selected IR bands (KBr, cm21): 3419 (br), 2932m, 1609s, 1468m, 1445m, 1090vs, 767m and 628s. Magnetic moment: meff (295 K) = 2.98 mB per NiII. [Ni2(tmpdtnm-X)(OH2)2][ClO4]4?3H2O 4. The procedure used to prepare complex 1 was followed except for tmpdtnm-X (2.06 g, 2.85 mmol) and NiCl2?6H2O (1.42 g, 5.98 mmol). A pink crystalline solid was obtained after recrystallisation from a CH3CN–water mixture (yield 1.21 g, 32%) {Found: C, 39.7; H, 4.8; N, 10.6.Calc. for [Ni2(C44H56N10)(OH2)2][ClO4]4?3H2O: C, 39.7; H, 5.0; N, 10.5%}. Electron microprobe: Cl : Ni ratio 2 : 1. Selected IR bands (KBr, cm21): 3432 (br), 2930w, 1608s, 1470m, 1446m, 1093vs, 764m and 626s. Magnetic moment: meff (295 K) = 2.98 mB per NiII. [Ni2(tmpdtnp-OH)(OH2)2][ClO4]4 5. The procedure used to prepare complex 1 was followed except for tmpdtnp-OH (1.10 g, 1.62 mmol) and NiCl2?6H2O (0.81 g, 3.41 mmol).The product appeared as a pink crystalline solid after recrystallisation from a CH3CN–water mixture (yield 0.66 g, 33%). {Found: C, 37.7; H, 5.0; N, 11.7. Calc. for [Ni2(C39H53N10- OH)(OH2)2][ClO4]4: C, 38.0; H, 4.7; N, 11.4%}. Electron microprobe: Cl : Ni ratio 2 : 1. ESI Mass spectrum: [Ni2L(ClO4)3]1, m/z 1094.0; [Ni2L(ClO4)2]21 497.2; [Ni2L(ClO4)]31, 298.6. Selected IR bands (KBr, cm21): 3508 (br), 3100m, 2939m, 1610s, 1470s, 1448s, 1095vs, 766s and 625s.Magnetic moment: meff (292 K) = 2.99 mB per NiII.J. Chem. Soc., Dalton Trans., 1998, 3919–3925 3921 Crystallography Intensity data for a purple plate crystal of complex 3 of dimensions 0.28 × 0.24 × 0.06 mm were measured on a Rigaku AFC6S diVractometer fitted with graphite-monochromated Mo-Ka radiation. Cell constants and the orientation matrix for data collection were obtained from a least-squares refinement using the setting angles of 18 carefully centered reflections in the range 11.85 < 2q < 18.678.The w–2q scan technique to a maximum 2q value of 50.18 was used to collect 4798 (4501 unique) reflections from which 3084 with I � 3.0s(I) were used in the refinement. The intensities of three representative reflections were measured after every 150 and no decay correction was applied. The data were corrected for Lorentz-polarisation eVects 19 and an empirical absorption correction was applied.20 The structure was solved by direct methods 21 and expanded using Fourier techniques in the DIRDIF 94 program.22 All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were included in calculated positions and not refined. The final cycle of full-matrix least-squares refinement was based on F. Neutral atom scattering factors were taken from ref. 23 and anomalous dispersion eVects 24 were included in Fcalc. Calculations were performed using the TEXSAN crystallographic software package.25 Crystallographic data are given in Table 1.CCDC reference number 186/1196. Results and discussion Synthesis The binuclear nickel(II) complexes of tmpdtne, tmpdtnp, tmpdtnb, tmpdtnm-X and tmpdtnp-OH were all prepared by treating ethanolic solutions of the ‘free’ ligands with an excess of NiCl2?6H2O (Scheme 1). Cation exchange chromatography of the reaction mixtures yielded a major pink band, on elution with water–acetonitrile mixture containing 1 mol dm23 NaClO4, from which the diaqua forms of the binuclear nickel(II) complexes could be crystallised.Satisfactory elemental analyses were obtained for all five complexes and electron microprobe analysis indicated a uniform Ni: Cl ratio of 1 : 2. The IR spectra showed bands indicative of the presence of the ligands and counter ions. The ESI mass spectrum of 5 confirmed the presence of species with composition [Ni2L- (ClO4)3]1, [Ni2L(ClO4)2]21 and [Ni2L(ClO4)]31. This indicated that deprotonation of the propan-2-ol ligand backbone did not occur and hence that the formation of an endogenously bridged binuclear complex was unlikely.Table 1 Crystallographic data for [Ni2(tmpdtnb)(OH2)2][ClO4]4 3 Formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 ZT /K Dc/g cm23 m(Mo-Ka)/cm21 2qmax/8, hkl data collected No. data measured No. unique data No. observed data [I � 3s(I)] RR 9 C40H60Cl4N10Ni2O18 1228.18 Triclinic P1� (no. 2) 9.562(4) 10.414(4) 14.171(4) 82.97(4) 79.65(3) 66.34(4) 1270(1) 1 293(1) 1.606 10.35 50.1, 1h, ±k, ±l 4798 4501 3084 0.048 0.035 Crystal structure of [Ni2(tmpdtnb)(OH2)2][ClO4]4 3 The molecular structure of the cation in complex 3 is shown in Fig. 1. Each nickel(II) centre is in a distorted octahedral geometry defined by five N atoms from the ligand and an oxygen donor from an H2O ligand. The N(1) and N(5) nitrogens form the axial bonds with an average bond length of 2.12 Å, while the N(2), N(3), N(4) and O(1) donor atoms comprise the equatorial bonds with a shorter average length of 2.08 Å.As is observed in the analogous copper(II) complexes, 16,17 the M–N distance to the bridgehead nitrogen, viz. the Ni–N(1) bond of 2.134(4) Å, is longer than the equatorial Ni–N (tacn) distances of 2.066(4) [Ni–N(2)] and 2.085(4) Å [Ni–N(3)] (Table 2). This diVerence, however, is smaller than in the copper(II) complexes.16,17 Although the Ni–N (py) axial bond length of 2.107(4) Å [Ni– N(5)] is marginally longer than the Ni–N (py) equatorial bond distance of 2.082(4) Å [Ni–N(4)], these distances are typical for nickel(II) complexes of pyridyl arm bearing tacn derivatives.10,27–29 The remaining Ni–O(1) equatorial bond of 2.095(3) Å is similar to the Ni–O distance of 2.10 Å in [Ni(py2- tasn)(H2O)][ClO4]2 [py2-tasn = 4,7-bis(2-pyridylmethyl)-1-thia- 4,7-diazacyclononane].27 The two pentadentate compartments of the ligand in 3 are oriented away from each other in an anti configuration, as was found for the manganese(II) analogue 30 and the nickel(II) complex of a related ligand with alcohol pendant arms.15 However, 3 shows considerably less distortion from regular octahedral ry than the manganese(II) complex, viz.the N(1)–Ni–N(5), O(1)–Ni–N(2) and N(3)–Ni–N(4) angles of 164.0(2), 174.8(2) and 165.6(2)8 in 3 are closer to the Fig. 1 An ORTEP26 plot of [Ni2(tmpdtnb)(OH2)2][ClO4]4 3 (thermal ellipsoids are drawn at 30%). Table 2 Selected bond distances (Å) and angles (8) for [Ni2(tmpdtnb)- (OH2)2][ClO4]4 3 Ni–O(1) Ni–N(1) Ni–N(2) Ni–N(3) Ni–N(4) Ni–N(5) N(1)–C(1) N(1)–C(6) N(2)–C(2) O(1)–Ni–N(1) O(1)–Ni–N(3) O(1)–Ni–N(5) N(1)–Ni–N(3) N(1)–Ni–N(5) N(2)–Ni–N(4) N(3)–Ni–N(4) N(4)–Ni–N(5) Ni–N(1)–C(6) Ni–N(2)–C(2) Ni–N(3)–C(13) C(2)–N(2)–C(7) Ni–N(3)–C(4) Ni–N(3)–C(13) C(4)–N(3)–C(13) Ni–N(4)–C(8) C(8)–N(4)–C(12) Ni–N(5)–C(18) 2.095(3) 2.134(4) 2.066(4) 2.085(4) 2.082(4) 2.107(4) 1.477(6) 1.502(6) 1.506(6) 94.8(1) 100.3(2) 86.1(1) 83.8(2) 164.0(2) 81.1(2) 165.6(2) 97.6(2) 107.5(3) 108.8(3) 106.0(3) 106.3(3) 108.7(3) 106.0(3) 112.1(4) 112.5(3) 118.4(5) 129.5(4) N(2)–C(3) N(3)–C(13) N(3)–C(4) N(3)–C(5) N(2)–C(7) N(4)–C(8) N(4)–C(12) N(5)–C(14) N(5)–C(18) O(1)–Ni–N(2) O(1)–Ni–N(4) N(1)–Ni–N(2) N(1)–Ni–N(4) N(2)–Ni–N(3) N(2)–Ni–N(5) N(3)–Ni–N(5) Ni–N(1)–C(1) C(1)–N(1)–C(6) Ci–N(2)–C(3) C(2)–N(2)–C(3) C(3)–N(2)–C(7) Ni–N(3)–C(5) C(4)–N(3)–C(5) C(5)–N(3)–C(13) Ni–N(4)–C(12) Ni–N(5)–C(14) C(14)–N(5)–C(18) 1.473(6) 1.479(6) 1.500(6) 1.479(6) 1.477(6) 1.344(6) 1.330(6) 1.345(6) 1.336(6) 174.8(2) 93.8(2) 85.1(2) 98.3(2) 84.9(2) 95.5(2) 80.3(2) 102.2(3) 111.5(4) 105.5(3) 111.5(4) 112.6(4) 104.6(3) 112.3(4) 112.7(4) 129.1(4) 112.7(3) 117.8(5)3922 J.Chem. Soc., Dalton Trans., 1998, 3919–3925 expected value of 1808 than the corresponding angles in the manganese(II) complex, which are in the 140–1608 range.30 Furthermore, the tacn N–Ni–N chelate angles in 3 (average 84.68) are much closer to 908 than the corresponding N–Mn–N angles (average 76.08).30 The Ni ? ? ? Ni9 separation of 9.563(4) Å in complex 3 is 1.284 Å longer than that observed in the corresponding manganese( II) analogue.30 This is predominantly a consequence of the orientation of the two halves of the complex around the bridgehead nitrogens.In 3 the Ni–N(1)–C(19)–C(20) and Ni9–N(19)–C(199)–C(209) torsion angles, which are a measure of the orientation of pentadentate compartments relative to the axis of the bridging group, are 210.9(8) and 1169.1(8)8, respectively while in the manganese(II) complex they are 159.6(8) and 259.5(8)8.Although both complexes adopt anti configurations, the arrangement of the bridging unit is much more linear in 3. This can be seen in the extended cell diagram for 3 and may explain the longer M ? ? ? M separation when compared with the manganese(II) complex. It is notable that the Cu ? ? ? Cu distance in the copper(II) complex of the same bis(pentadentate) ligand is also much shorter (by 0.88 Å 17) than that in 3.Molecular modelling calculations on this copper(II) complex and those of the other bis(pentadentate) ligands shown in Scheme 1 have predicted that these complexes can adopt a number of configurations, which diVer only in the relative orientation of the pentadentate compartments, with similar strain energies in solution.17 Under these circumstances, it is not surprising to find that variations in crystal packing eVects that result from changes in the metal centre can be responsible for significant variations in solid state structures.Electronic spectra and magnetic behaviour The UV-Visible spectral data for complexes 1–5 recorded in acetonitrile are summarised in Table 3. Two of the three spin allowed transitions expected for NiII in a near octahedral ligand field 31,32 were observed [3A2g æÆ 3T2g (800–900), 3A2g æÆ 3T1g(F) (500–550 nm)] but the third transition [3A2g æÆ 3T1g(P)] normally found in the 300–400 nm region is masked by p æÆ p* and/or CT transitions involving the pyridine rings.The band in the 800–900 nm region, corresponding to the 2A2g æÆ 3T2g transition, is asymmetric as highlighted by a shoulder at ca. 880 nm for each complex. Such behaviour is also evident in the related octahedral nickel(II) complexes mentioned in Table 3 and may be attributed to low symmetry splitting or spin–orbit splitting of the 3T2g state, or the close approach of the 1Eg state such that the spin forbidden 3A2g æÆ 1Eg transition gains intensity through spin–orbit coupling with the 3T2g state.32,33 There is little diVerence in the lmax values for complexes 1–4 Table 3 UV-Visible spectral data of 1–5a and some related nickel(II) complexes Complex 12345[ Ni(dmptacn)- (OH2)][ClO4]2 a [Ni(tmptacn)]- [ClO4]2 a [Ni(dtne)][ClO4]2 b [Ni(tacn)2]Cl2 b [Ni(tatacn)][ClO4]2 b lmax/nm (emax/dm3 mol21 cm21) 519 (42), 812 (64), 878 (sh) (42) 524 (36), 811 (63), 878 (sh) (40) 519 (39), 812 (64), 878 (sh) (42) 519 (35), 806 (60), 881 (sh) (44) 524 (39), 820 (58), 894 (sh) (44) 516 (18), 800 (23), 880 (sh) (22) 515 (27), 810 (36), 886 (sh) (24) 363 (16), 516 (18), 848 (31), 917 (31) 310 (12), 500 (9), 800 (9), 870 (sh) 355 (18), 557 (13), 805 (17), 924 (34) Ref.This work This work This work This work This work 10 36 12 37 28, 29 dmptacn = 1,4-bis(2-pyridylmethyl)-1,4,7-triazacyclononane; tmptacn = 1,4,7-tris(2-pyridylmethyl)-1,4,7-triazacyclononane; dtne = 1,2-bis(1,4,7-triazacyclonon-1-yl)ethane; tatacn = 1,4,7-triazacyclononane- N,N9,N0-triacetate.a In CH3CN. b In water. which range from 519 to 524, 806 to 812 and a shoulder at 878 to 881 nm. However, the values for 5, in which the two halves of the ligand are joined by a propan-2-ol bridge, are the highest in the series. This indicates that the tmpdtnp-OH ligand exerts a weaker ligand field than the other ligands, possibly due to the electron withdrawing power of the OH group.The fact that the absorption maxima for all the complexes 1–5 are at lower energy than those for [Ni(dmptacn)(OH2)][ClO4]2 indicates that the bis(pentadentate) ligands exert a weaker ligand field than does dmptacn. This lower ligand field can be attributed to the presence of an extra tertiary nitrogen per pentadentate compartment in the binucleating ligands. The eVective magnetic moments, meff (per NiII), of complexes 1–5 at room temperature range from 2.81 to 2.99 mB.On average these values are slightly higher than the spin-only value due to second order spin–orbit coupling eVects but they are typical for nickel(II) amine complexes in an octahedral environment. Electrochemistry The redox behaviour of the binuclear nickel(II) complexes, 1–5, has been studied by cyclic and square-wave voltammetry (SWV) at a platinum macroelectrode. The CV and SWV traces of 1, shown in Fig. 2, are composed of two partially resolved signals, but clearly indicate that oxidation of the two nickel(II) centres to NiIII occurs at slightly diVerent potentials, eqn.(1). NiIINiII 1 e2 NiIIINiII 1 e2 NiIIINiIII (1) The potentials for the two oxidation processes for the binuclear nickel complexes are so positive that the second process overlaps with the onset of the background process. Consequently, cyclic voltammetric experiments involve switching the potential in a region where highly reactive species are produced by oxidation of the solvent (electrolyte).It is believed that it is the reaction of these solvent derived species with the dinickel(III) complex which lead to an apparently significant level of Fig. 2 Cyclic (scan rate = 100 mV s21) (a) and square-wave (period = 30 ms) (b) voltammograms of complex 1 in acetonitrile (0.1 mol dm23 Bu4NClO4) using a platinum macrodisc electrode at 20 8C. The potential axis used is relative to the Ag–Ag1 reference.J. Chem.Soc., Dalton Trans., 1998, 3919–3925 3923 chemical irreversibility being associated with the cyclic voltammetric response for the binuclear nickel processes. That is, it is this reaction and not any inherent instability of NiIIINiIII on the CV timescale which leads to the observation of unequal peak heights of the oxidation and reduction components of the nickel oxidation processes. In addition to precluding the assessment of the chemical reversibility of the two oxidation processes, the proximity to the background process prevents well resolved limiting current values being obtained under steady state conditions via the use of microdisc or rotated disc electrodes, as was possible with the binuclear coppersystems.However, fortunately, the square wave technique only requires the measurement of data obtained from scanning the potential in the positive direction and adequate resolution can be obtained from the background response with this technique.Consequently, the SWV method can be used to obtain quantitatively useful data. The E2� 1 values, 11.11 and 11.17 V (Table 4), estimated from both techniques, confirm that a potential separation of approximately 60 mV exists for the two oxidation processes of compound 1. This finding indicates that the metal centres are close enough electronically to influence one another so that, once the first nickel(II) centre has been oxidised to NiIII, the second becomes more diYcult to oxidise due to the additional positive charge that has been introduced.34 This behaviour is also evident in the oxidation of the analogous manganese(II) complex, where two partially resolved oxidation processes are observed in forming the MnIIIMnIII state,30 but not in reduction of the corresponding copper(II) complex, where a decrease in electrostatic repulsion between the two copper centres may occur on reduction to the CuICuI state.17 Notably, Ciampolini et al.11 have reported that oxidation of the nickel(II) complex of the ethane linked bis(cyclam) ligand also occurs in two even more clearly distinguishable processes whose reversible potentials are separated by 100 mV.The CV and SWV responses of the nickel(II) complex of tmpdtnp-OH, 5, displayed in Fig. 3, are similar to those of 1 in that they are broadened and clearly composed of two partially resolved oxidation processes. Nevertheless, despite problems associated with incomplete resolution and proximity to the solvent limit, analysis of data obtained from both the CV and Table 4 Cyclic and square-wave voltammetric data for oxidation of nickel(II) complexes in acetonitrile (0.1 mol dm23 Bu4NClO4)a Com- Cyclic b Square wave c plex Fc 1 2345 E2� 1 /V 0.00 11.11 11.17 11.07 11.06 11.11 11.06 11.17 DEp/mV 58 96 96 84 80 86 55 41 Ep ox/V 20.03 11.16 11.22 11.11 11.10 11.15 11.08 11.19 Ep red/V 10.03 11.06 11.12 11.03 11.02 11.07 11.03 11.15 Ep/V 0.00 11.11 11.17 11.07 11.06 11.11 11.07 11.17 W2� 1 d/mV 100 120 e 120 e 110 110 100 150 e 150 e Ip/mA 14.9 7.9 7.7 14.2 15.2 15.2 8.9 7.0 a Potentials are quoted with respect to the Fc–Fc1 couple used as an internal reference.The concentrations of the complexes used were in the range 0.5–0.55 mmol dm23 but the current data are normalised to 0.5 mmol dm23 to facilitate comparison of values. b Cyclic voltammetric data were obtained at a scan rate of 100 mV s21. The close proximity of the background process and the interference this causes with the voltammetry of the binuclear nickel complexes (see text) means that peak currents for the reduction component on the reverse scan of the CV experiment are perturbed.Consequently, ratios of oxidation and reduction peak heights for nickel processes, usually calculated to assess their chemical reversibility, cannot be obtained for this system. c Square-wave voltammetric data were measured with a period of 30 ms. d W2� 1 is the estimated width of the peak shaped process response at half the wave height.e Inadequate resolution available to provide values for individual processes. Numbers given refer to the width of the total response. SWV techniques shows that in this case (Table 4) the two 1e2 oxidation processes represented by eqn. (1) are separated by approximately 110 mV with E2� 1 values of 11.06 and 11.17 V. Thus, the first nickel(II) centre in 5 is oxidised to NiIII at 11.06 V, which is almost the same average potential recorded for the two very closely spaced 1e2 oxidation processes of the complex of tmpdtnp, 2 (Table 4), while the second centre in 5 is harder than the first as evidenced by the value of 11.17 V.This behaviour parallels that of the analogous copper(II) complex which also showed two partially resolved electrochemical responses. In both complexes the electronic properties of the OH group of the propan-2-ol bridge influence the redox properties of the metal centres.For the nickel(II) complex, the central alcohol group of tmpdtnp-OH may be co-ordinating to the oxidised nickel(III) centre. This brings the two metal centres closer together and, consequently, the remaining nickel(II) centre becomes more diYcult to oxidise. The CV and SWV responses for complexes 2–4 (Fig. 4 displays data for 4) are not even partially resolved like those of 1. However, the value of DEp = 80–86 mV is considerably larger than that of approximately 30 mV expected for a simultaneous 2e2 oxidation step (i.e. 57 mV/n at 25 8C), indicating that two closely separated 1e2 oxidation processes having E2� 1 values separated by 30 ± 10 mV are occurring as shown in eqn. (1). The “average” E2� 1 values for 1–3 of 11.14, 11.07 and 11.06 V, respectively, become less positive with increasing length of the alkyl bridge used to link the two pentadentate ligand compartments (see Table 4). This trend has also been observed for the nickel(II) complexes of the corresponding alkyl bridged bis(cyclam) ligands studied by Ciampolini et al.11 and was rationalised in terms of decreases in electrostatic repulsion between nickel centres with increasing M ? ? ? M separation.This facilitates oxidation to the NiIINiIII and NiIIINiIII states. At the extreme, the mononuclear complex, [Ni(dmptacn)- (OH2)]21, experiences a marked decrease in these electrostatic influences and is therefore more easily oxidised (E2� 1 = 10.98 V 35) than 1–3.Notably, the minimised solution structures of the Fig. 3 Cyclic (a) and square-wave (b) voltammograms of complex 5. Details as in Fig. 2.3924 J. Chem. Soc., Dalton Trans., 1998, 3919–3925 copper(II) complexes of the ethane-, propane- and butanebridged bis(pentadentate) ligands, viz. the analogues of 1–3, confirmed an increase in M ? ? ? M separation with alkyl chain length.17 As was found here, the reduced CuICuI forms of these complexes are more easily oxidised as the M ? ? ? M separation increases.The nickel(II) complex of tmpdtnm-X, 4, exhibits an E2� 1 value (11.11 V) which is more positive than anticipated on the basis of the expected M ? ? ? M separation, i.e. after resolved oxidation processes. This anomalous behaviour was also observed for the corresponding bis(cyclam) complexes where it was found that complexes of the m-xylene bridged ligand have a more positive E2� 1 value than those of the butane bridged ligand which has a shorter M ? ? ? M separation.11 In both systems, a combination of steric and inductive eVects attributable to the xylene group could be reducing the donor strength of the bridgehead nitrogens compared to those with simple aliphatic substituents.11 The decreased donor strength makes the oxidation of 4 more diYcult and oVsets the benefits of a decrease in the charge repulsion within the complex aVorded by the larger Ni ? ? ? Ni separation.The net result is that the E2� 1 value for 4 is more positive than that of 3.The square-wave data, after normalisation to a concentration of 5 × 1024 mol dm23 (Table 4), show the 1e2 ferrocene oxidation step produces a peak current/unit concentration value that is considerably larger than that of the partially resolved processes ame as that of unresolved cases (2–4). However, analysis of the current magnitude of the square-wave currents using the theory reported in ref. 17 and assuming the diVusion coeYcients are significantly less than that for ferrocene, as was demonstrated for the binuclear copper analogues, confirmed that the peak currents are of the correct magnitude for two unresolved or partially resolved 1e2 reversible charge transfer processes.Unfortunately, the inability to obtain quantitatively reliable CV or steady state data, due to the proximity of the background process, meant that more detailed calculations as presented for the copper system were not possible.Fig. 4 Cyclic (a) and square-wave (b) voltammograms of complex 4. Details as in Fig. 2. Conclusion Studies of the co-ordination chemistry of a series of bis(pentadentate) ligands bearing 2-pyridylmethyl pendant groups have been extended to the nickel(II) complexes. Electrochemical analyses of these binuclear complexes established that the potentials associated with the oxidation processes decrease with the length of the alkyl group linking the two ligand compartments.This indicates an increase in Ni ? ? ? Ni separation leads to a decrease in electrostatic repulsion in the oxidised state. For two complexes, 1 and 5, splitting of the oxidation waves was clearly evident, confirming that two closely spaced 1e2 processes rather than a single 2e2 process were taking place (i.e. NiIINiII æÆ NiIINiIII æÆ NiIIINiIII). Acknowledgements This work was supported by Australian Research Council grants to L. S. and A. M. B. S. J. B. was the recipient of a Monash Graduate Scholarship.References 1 R. K. Thauer, G. Diekert and P. Schonheit, Trends Biochem. Sci., 1980, 5, 304. 2 J. O. Nriagu, Nickel in the Environment, Wiley, New York, 1980. 3 W. Mertz, Science, 1981, 213, 1332. 4 F. W. Sunderman (Editor), Nickel in the Human Environment, Oxford University Press, New York, 1985, p. 530. 5 M. A. Halcrow and G. Christou, Chem. Rev., 1994, 94, 2421; A. F. Kolodziej, Prog. Inorg. Chem., 1994, 41, 493. 6 E. Jabri, M. B.Carr, R. P. Hausinger and P. A. Karplus, Science, 1995, 268, 998; D. E. Wilcox, Chem. Rev., 1996, 96, 2435. 7 A. Volbeda, M.-H. Charon, C. Piras, E. C. Hatchikian, M. Frey and J. C. Fontecilla-Camps, Nature, (London), 1995, 373, 580. 8 See, for example, L. Sacconi, F. Mani and A. Bencini, in Comprehensive Coordination Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon Press, Oxford, 1987, vol. 5, p. 1; V. McKee, Adv. Inorg. Chem., 1993, 40, 323; D. E.Fenton and H. Okawa, Chem. Ber., 1997, 130, 433; D. Volkmer, B. Hommerich, K. Griesar, W. Haase and B. Krebs, Inorg. Chem., 1996, 35, 3792 and refs. therein. 9 P. D. Beer, M. G. B. Drew, P. B. Leeson, K. Lyssenko and M. I. Ogden, J. Chem. Soc., Chem. Commun., 1995, 929. 10 G. A. McLachlan, G. D. Fallon, R. L. Martin, B. Moubaraki, K. S. Murray and L. Spiccia, Inorg. Chem., 1994, 33, 4663. 11 M. Ciampolini, L. Fabbrizzi, A. Perotti, A. Poggi, B. Seghi and F. Zanobini, Inorg.Chem., 1987, 26, 3527. 12 K. Wieghardt, I. Tolksdorf and W. Herrmann, Inorg. Chem., 1985, 24, 1230. 13 B. Graham, G. D. Fallon, M. T. Hearn, D. C. R. Hockless, G. Lazarev and L. Spiccia, Inorg. Chem., 1997, 36, 6366. 14 W. B. Tolman, Acc. Chem. Res., 1997, 30, 227; S. Mahapatra, S. Kaderli, A. Llobet, Y.-M. Neuhold, T. Palanche, J. A. Halfen, V. G. Young, Jr., T. A. Kaden, L. Que, Jr., A. D. Zuberbuhler and W. B. Tolman, Inorg. Chem., 1997, 36, 6343. 15 A. J. Blake, T. M. Donlevy, P.A. England, I. A. Fallis, S. Parsons, S. A. Ross and M. Schroder, J. Chem. Soc., Chem. Commun., 1994, 1981. 16 S. J. Brudenell, L. Spiccia and E. R. T. Tiekink, Inorg. Chem., 1996, 35, 1974. 17 S. J. Brudenell, L. Spiccia, A. M. Bond, P. Comba and D. C. R. Hockless, Inorg. Chem., 1998, 37, 3705. 18 F. H. Fry, B. Graham, L. Spiccia, D. C. R. Hockless and E. R. T. Tiekink, J. Chem. Soc., Dalton Trans., 1997, 827. 19 XDISK, Data Reduction Program, Version 4.20.2PC, Siemens Analytical X-Ray Instruments, Inc., Madison, WI, 1989. 20 N. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983, 39, 158. 21 SIR 92, A. Altomare, M. Cascarano, C. Giacovazzo and A. Guagliardi, J. Appl. Crystallogr., 1993, 26, 343. 22 DIRDIF 94, P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman, R. de Gelder, R. Israel, R. O. Gould and J. M. M. Smits, The DIRDIF-94 Program System, Technical Report of the Crystallography Laboratory, University of Nijmegen, 1994. 23 D. T. Cromer and J. T. Waber, International Tables for X-Ray Crystallography, Kynoch Press, Birmingham, 1974, Table 2.3.A. 24 J. A. Ibers and W. C. Hamilton, Acta Crystallogr., 1964, 17, 781. 25 TEXSAN, Crystal Structure, Analysis Package, Molecular Structure Corporation, Houston, TX, 1992.J. Chem. Soc., Dalton Trans., 1998, 3919–3925 3925 26 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 27 K. Wasielewski and R. Z. Mattes, Z. Anorg. Allg. Chem., 1993, 619, 158. 28 K. Wieghardt, U. Bossek, P. Chaudhuri, W. Herrmann and H. J. Kueppers, Inorg. Chem., 1982, 21, 4308. 29 M. J. Van der Merwe, J. C. A. Boeyens and D. R. Hancock, Inorg. Chem., 1985, 24, 1208. 30 S. J. Brudenell, L. Spiccia, A. M. Bond, G. D. Fallon, D. C. R. Hockless and E. R. T. Tiekink, unpublished work. 31 D. Sutton, Electronic Spectra of Transition Metal Complexes, McGraw-Hill, London, 1968. 32 S. M. Hart, J. C. Boeyens and R. D. Hancock, Inorg. Chem., 1983, 22, 982. 33 R. Stranger, S. C. Wallis, L. R. Gahan, C. H. L. Kennard and K. A. Byriel, J. Chem. Soc., Dalton Trans., 1992, 2971. 34 A. Urfer and T. A. Kaden, Helv. Chim. Acta, 1994, 77, 23. 35 G. A. McLachlan, Ph.D. Thesis, Monash University, Melbourne, 1994. 36 K. Wieghardt, E. SchoVman, B. Nuber and J. Weiss, Inorg. Chem., 1986, 25, 877. 37 R. Yang and L. J. Zompa, Inorg. Chem., 1976, 15, 1499. Paper 8/06084C

ISSN:1477-9226    

DOI:10.1039/a806084c

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3927–3933 3927 Synthesis and crystal structures of silver(I) and palladium(II) complexes of new bis(2-pyridyloxy)benzenes and methylene extended analogues Chris M. Hartshorn and Peter J. Steel * Chemistry Department, University of Canterbury, Christchurch, New Zealand. E-mail: p.steel@chem.canterbury.ac.nz Received 22nd September 1998, Accepted 24th September 1998 Four new ligands, 1,3-bis(2-pyridyloxy)benzene II, 1,2-bis(2-pyridyloxy)benzene III, 1,4-bis(2-pyridyloxymethyl)- benzene IV and 1,4-bis(2-pyridylmethoxy)benzene V, have been prepared, and their complexes with silver(I) nitrate and palladium(II) chloride synthesized.The silver complex of ligand II is shown, by a crystal structure determination, to be a M2L2 20-membered macrocycle with close p–p stacking of the benzene rings of the ligands, similar to the analogous complex of the para substituted isomer of this ligand. The structure of the mononuclear palladium complex of ligand III shows the ligand co-ordinated to the metal in a nine-membered chelate ring, with evidence for a weak interaction between the metal and the central benzene ring. Structures of the silver complexes of ligands IV and V reveal that these are polymeric species, with the molecular packing controlled by p–p stacking interactions between aromatic rings.The crystal structure of the palladium complex of ligand V shows that it self-assembles into a M2L2 26-membered macrocycle, but without intramolecular p–p stacking.Introduction The term metallosupramolecular chemistry was introduced by Constable 1 to describe the self-assembly of larger molecular aggregates from combinations of transition metal ions and polydentate ligands. Often this involves the use of metal ions that have defined co-ordination numbers and stereochemical preferences, to encode the rational assembly of specific molecular architectures by recognition of the inherent properties of logically designed ligands.In this way, many recent reports 2 have described the formation of numerous metallosupramolecular species with novel topological structures, such as squares, cages, ladders, bricks, helicates and polycatenates, some of which contain interlocking rings and interpenetrating 3-D networks, much of which has provided useful information for the rapidly expanding area of crystal engineering.3 However, this approach imposes natural restrictions on the architectures available, due to the limited range of angular motifs available from stereorigid metals.Thus, we are presently studying the chemistry of metal ions that have less well defined coordination numbers and geometries. For example, the d10 metal silver(I) forms complexes with various co-ordination numbers, and, within a given co-ordination number, has very flexible geometrical requirements. Previously, we reported the synthesis of the new ligand I and its reaction with silver nitrate to produce the novel M2L2 dimetalloparacyclophane 1 containing a p–p stacked subunit (Scheme 1).4 We were curious to know whether this was simply a serendipitous event or a more general reaction that would apply to other structurally related ligands.Specifically, we were interested to determine the extent to which the p–p stacking was responsible for the self-assembly of this dimeric structure and whether the oxygen atoms of the ligand played a role in the dimer formation.Accordingly, we have now synthesized four new ligands and studied their reactions with silver(I) nitrate and, for comparative purposes, palladium(II) chloride. Since our previous report of the reaction of I, a number of structurally related M2L2 dimers of various ligands with silver 5 and other metals 6 have been reported; however none of these displays such intimate p–p stacking of aromatic rings. Results and discussion Two approaches were adopted for structural modification of the ligand I, in order to assess the generality of the dimer formation.First, the para substitution in the benzene ring was replaced by the meta and ortho isomers and, secondly, methylene spacer groups were introduced on either side of the oxygen atoms. Thus, the four new ligands II–V were synthesized, all by nucleophilic substitution reactions. By analogy with the method used to prepare the para isomer I,4 2-bromopyridine was treated with 1,3-dihydroxybenzene (resorcinol) and 1,2-dihydroxybenzene (catechol), in the presence of potassium carbonate, to give the meta and ortho isomers II and III, respectively.The methylene extended homologue IV was prepared by reaction of 2-bromopyridine with 1,4- benzenedimethanol, in an adaptation of a reported method for the synthesis of benzyl 2-pyridyl ether.7 The isomer V was prepared by reaction of 1,4-dihydroxybenzene (hydroquinone) with 2-chloromethylpyridine under phase-transfer catalysed conditions.The 1H and 13C NMR spectra of all four ligands were fully assigned, by a combination of 1- and Scheme 1 O O N N O O N N O O N N Ag Ag OH2 H2O ONO2 O2NO I 1 AgNO33928 J. Chem. Soc., Dalton Trans., 1998, 3927–3933 2-D NMR techniques, and are given in the Experimental section. The ligands II–V reacted smoothly with silver nitrate, under similar conditions to those used to prepare 1, to give complexes 2–5, in excellent yields. The crystal structures of three of these are described below.In order to determine whether stereochemically less flexible metals would also form the M2L2 dimers, palladium(II) was chosen as a representative square-planar co-ordinating metal, which has been much used in metallosupramolecular self-assemblies.8 Reactions of I and II–V with palladium chloride aVorded complexes 6–10 of 1 : 1 stoichiometry, in excellent yields. Unfortunately, only two of these complexes were suYciently soluble for NMR studies and for recrystallisation to furnish X-ray quality crystals.The crystal structures of these are also described below. The silver nitrate complex 2, of 1,3-bis(2-pyridyloxy)benzene II, crystallises in the monoclinic space group C2/c. Like 1, it is a M2L2 dimeric macrocycle, but in this case has C2 crystallographic symmetry, rather than a centre of inversion. Fig. 1 shows a perspective view of the structure, with atom labelling of the asymmetric unit.Selected interatomic distances and angles are also listed. To each silver atom are co-ordinated two pyridine nitrogens and a water oxygen atom. The silver–donor bond distances are similar to those of 1, although the Ag– O(100) distance [2.486(2) Å] is slightly shorter than the analogous bond in 1 [2.528(2) Å]. The geometry at the silver atom is distorted T-shaped with a similar N–Ag–N angle [158.08(6)8] to that in 1. However, whereas in 1 the water oxygen bisects this angle, in the case of 2 it subtends two very diVerent O–Ag–N angles [109.17(6) and 92.75(6)8].As with 1, the co-ordinated water hydrogen atoms are hydrogen-bonded to oxygen atoms of two diVerent nitrate anions, with O ? ? ? O separations of 2.878(3) and 3.034(3) Å. The intramolecular Ag ? ? ? Ag separation [9.405(1) Å] in the 20-membered macrocycle 2 of the meta Fig. 1 Perspective view and atom labelling of the crystal structure of complex 2. Selected interatomic distances (Å) and angles (8): Ag(1)– N(11) 2.196(2), Ag(1)–N(31A) 2.218(2), Ag(1)–O(100) 2.486(2), Ag(1) ? ? ? O(10) 3.063(2), Ag(1) ? ? ? O(30A) 2.899(2); N(11)–Ag(1)– N(31A) 158.06(6), N(11)–Ag(1)–O(100) 109.17(6), N(31A)–Ag(1)– O(100) 92.75(6).O O N N O O N N O O N N O O N N IV II III V isomer is considerably less than that [10.384(1) Å] in the 22- membered macrocyclic complex of the para disubstituted ligand.4 The benzene and two pyridine rings are each planar, but, as was observed with complex 1, the oxygen linking atoms, O(10) and O(30), are both significantly out of the plane of the benzene ring [0.088(4) and 0.163(4) Å, respectively].Again, this distortion is towards the proximate silver atom, which, in combination with the O ? ? ? Ag distances [2.899(2) and 3.063(2) Å], suggests the existence of weak ether–oxygen–silver interactions. The change in topology of the ligand from 1,4- to 1,3- disubstitution also changes the relative inclinations of the various aromatic rings.The mean planes of the pyridine rings are now inclined at angles of 83.6(1) and 103.9(1)8 to that of the linking benzene ring, while the two pyridine ring mean planes at each silver atom are inclined to one another at an angle of 17.5(1)8. Again, the silver atom is significantly out of the extended mean planes of the co-ordinated pyridine rings to the extent of 0.164(2) and 0.121(2) Å. An important feature of this structure is the preservation of the p–p stacking of the benzene rings.In particular, it is interesting that this complex crystallises in a conformation with the two meta disubstituted benzene rings oriented in the same direction (with C2v molecular symmetry), rather than an alternative possible orientation with one of the benzene rings inverted (with C2h symmetry). We believe that the observed orientation occurs in order to maximise the favourable p–p stacking interaction. Nevertheless, unlike the centrosymmetric isomer 1, in the structure of 2 the benzene rings are not parallel, with mean planes inclined at an angle of 9.18.However, despite this slight splaying of the benzene rings away from each other, they are still displaced from one another in the same manner as in 1, with the centroid of one ring lying above an atom of the other ring, as shown in the alternative view of the structure displayed in Fig. 2. We have previously noted that this seems to be the most favourable orientation for parallel stacked aromatic rings, as has been suggested by computational studies.9 The spacing between the two parallel benzene rings in 1 is 3.33(1) Å; in 2 the slight splaying of the rings has the result that the corresponding values range from 3.343(3) Å, for C(2), to 3.777(3) Å, for C(5).Thus, the formation of a second M2L2 dimer has demonstrated that the structure 1 was not a fortuitous singularity and may represent a general self-assembly process, controlled by the p–p stacking interaction.Attention, therefore, was turned to the ortho isomer III. Unfortunately, the silver nitrate complex of this ligand failed to provide crystals suitable for X-ray analysis. Elemental analysis suggested an intriguing M5L4 stoichiometry; however, this complex was not able to be characterised further. Reaction of palladium chloride with III provided a 1 : 1 complex 8 which was more amenable to study, being soluble in dichloromethane, suggestive of a monomeric structure.The 1H NMR spectrum in CD2Cl2 shows only six signals, which indicates the complex has two-fold symmetry. Significant coordination- induced changes in chemical shift are observed in comparing the spectra of III and 8 in CD2Cl2. In particular, the two pairs of benzene hydrogens have the same chemical shift (d 7.27) for free III, while diVerent chemical shifts downfield of Fig. 2 Top view of complex 2 showing the relationship of the p–p stacked benzene rings.J.Chem. Soc., Dalton Trans., 1998, 3927–3933 3929 this are observed for complex 8 (d 7.32 and 7.59). Such changes in chemical shift can result from a number of factors such as a change in the conformation of the ligand upon co-ordination and through-space ring-current anisotropy eVects.10 Assuming cis co-ordination by the two pyridine nitrogens, this complex would possess a nine-membered chelate ring. Furthermore, two distinct structures are possible, depending upon whether the benzene ring is proximate (8a) or distal (8b) to the palladium atom.Inspection of molecular models suggested that interconversion between these two isomers would not be a facile process. Significantly, no NOE enhancement of any pyridine hydrogen signal was observed upon irradiation of either benzene hydrogen signal. Inspection of molecular models showed that this would be so only for isomer 8a, which in turn raises the interesting possibility that there may also be an interaction between the palladium atom and the benzene ring.Thus a crystal structure determination of 8 was carried out. The palladium complex 8 crystallises in the monoclinic space group P21/c. Fig. 3 shows a perspective view and atom labelling of the structure, along with selected bonding geometry. The ligand, III, is indeed cis-chelated to the palladium, producing a nine-membered chelate ring. A search of the Cambridge Structural Database revealed that this is the first structure involving palladium in a nine-membered chelate ring with N,N-donors, although analogous structures with P,P-donors are known.11 The Pd–N and Pd–Cl distances are within the range of values previously reported for related structures.12 In the solid state, the twofold symmetry observed in solution is not crystallographically present, with the mean planes of the two pyridine rings inclined at angles of 90.6(1) and 119.8(1)8 to the benzene ring, and at 79.7(1)8 to each other.As suggested by the NMR experiments, the complex exists as isomer 8a.We believe that this is due to a weak interaction Fig. 3 Perspective view and atom labelling of the crystal structure of complex 8. Selected interatomic distances (Å) and angles (8): Pd(1)– N(11) 2.059(2), Pd(1)–N(21) 2.040(2), Pd(1)–Cl(1) 2.2831(7), Pd(1)– Cl(2) 2.2871(7); N(11)–Pd(1)–N(21) 88.33(8), N(11)–Pd(1)–Cl(1) 90.77(6), N(21)–Pd(1)–Cl(2) 89.07(6), Cl(1)–Pd(1)–Cl(2) 91.67(3), N(11)–Pd(1)–Cl(2) 176.63, N(21)–Pd(1)–Cl(1) 175.78(6). O O N N Pd Cl Cl O N Pd Cl Cl O N 8a 8b between the palladium atom and the benzene ring, as indicated by certain features of the structure.The palladium atom lies at a distance of 2.879(3) Å from the centre of the C(1)–C(2) bond. This interaction is suYcient slightly to pyramidalise the squareplanar palladium, by pulling it out of the co-ordination plane and towards the benzene ring. In turn, the oxygen atoms are displaced, by 0.134(4) and 0.142(4) Å [O(10) and O(20), respectively], on the opposite side of the plane of the benzene ring, indicating pyramidalisation of the attached carbons.This intramolecular interaction between the benzene p orbitals and the palladium dz2 orbital is analogous to those proposed to account for the intermolecular packing of palladium complexes containing aromatic ligands. Thus III represents an interesting new chelating ligand. As described above, the two methylene extended ligands IV and V were prepared to determine whether they would selfassemble into larger dimeric macrocycles, and to examine the possible role of the oxygen atoms in this process.Each of these ligands was treated with silver nitrate to produce, in excellent yields, crystalline complexes (4 and 5), suitable for structure determination. The complex 4 crystallises in the monoclinic space group C2/c, and is a metallopolymer, rather than a dimetallocycle. The asymmetric unit contains half of the ligand IV, positioned about a centre of inversion, that is co-ordinated to a silver atom, which, in turn, lies on a twofold rotation axis.The chelated nitrate ion, which is disordered over two orientations, also lies with two atoms on this axis. Fig. 4 shows the labelled contents of the asymmetric unit, together with adjacent groups and selected interatomic distances and angles. The silver atom is coordinated to two pyridine nitrogens with non-linear geometry [N(11)–Ag(1)–N(11B) 139.6(2)8].The silver–donor bond distances are similar to those in related literature compounds.13 The mean plane of the pyridine ring is inclined to that of the benzene at an angle of 46.5(6)8, while the two pyridine ring mean planes at each silver atom are inclined to one another at an angle of 62.8(6)8. Although this ligand does not produce a dimeric structure, it is interesting that both of the types of interaction discussed above are still found in the extended structure of complex 4.Specifically, the ether oxygen shows a similar weak interaction with the silver atom [Ag(1) ? ? ? O(10) 2.874(4) Å], being pulled towards the silver as seen in the distortion from ideal geometry around C(12) [O(10)–C(12)–N(11) 111.4(7), O(10)–C(12)– C(13) 125.1(8)8]. Also, the packing of the metallopolymeric structure still appears to be controlled in part by p–p interactions, but, in this case, between pyridine and benzene rings of adjacent units, rather than the benzene–benzene stacks of 1 and 2.As shown in Fig. 5, the zigzag shaped chain of the polymer is Fig. 4 Perspective view and atom labelling of the crystal structure of complex 4. Selected interatomic distances (Å) and angles (8): Ag(1)– N(11) 2.216(5), Ag(1)–O(1) 2.545(3), Ag(1)–O(1A) 2.503(4), Ag(1) ? ? ? O(10) 2.874(4); N(11)–Ag(1)–N(11B) 139.6(2), N(11)–Ag(1)– O(1) 97.0(2), N(11)–Ag(1)–O(1A) 120.0(2).3930 J. Chem. Soc., Dalton Trans., 1998, 3927–3933 such that each benzene ring exhibits p–p stacking with two pyridine rings that are inclined at an angle of 17.9(8)8.The structure of the silver complex 5 of the isomeric ligand V is somewhat similar. It crystallises in the monoclinic space group P21/c and is also a one dimensional metallopolymer. Fig. 6 shows the labelled contents of the asymmetric unit, along with selected adjacent groups. The silver atom is co-ordinated to two pyridine nitrogens and in half of the units to an oxygen of the nitrate counter ion, which is again disordered over two equally occupied sites.The silver–donor bond distances are again in agreement with the distances of related literature compounds. 13 The pyridine–silver–pyridine co-ordination geometry is non-linear, with the half occupancy co-ordinated oxygen bisecting the larger N–Ag–N angle. The mean planes of the pyridine rings are inclined at angles of 48.4(3) and 2.4(3)8 to that of the linking benzene ring. Since the ether oxygen atoms are no longer directly attached to the pyridine ring they do not interact with the silver atom.Once again the metallopolymer describes a zigzag shape, which results from p–p stacking interactions. As shown in Fig. 7, there is a weak p–p interaction between alternating pyridine rings in the polymer chain. Stronger interactions exist between diVerent strands of the polymer, with crystallographically equivalent pyridines of adjacent strands, related by a centre of inversion, being coplanar and separated by only 3.33(1) Å.Once again, the rings are displaced such that an atom of one ring lies over the centroid of the other ring, a relationship similar to that existing in complexes 1 and 2. Hence, p–p interactions appear to contribute to the assembly of the complex, and certainly to its crystal packing. The palladium chloride complex 10 of this ligand V was soluble and stable in DMSO, which allowed for characterisation by 1H NMR spectroscopy.The spectrum, with co-ordination induced shifts ranging between 0.10 and 0.57 ppm downfield from those of the non-co-ordinated ligand, showed that, in solution, the two pyridine rings are equivalent. Although the Fig. 5 View down the b axis of the polymeric chain structure of complex 4, with hydrogen and nitrate atoms omitted, showing the interligand p–p stacking of aromatic rings. Fig. 6 Perspective view and atom labelling of the crystal structure of complex 5. Selected interatomic distances (Å) and angles (8): Ag(1)–N(11) 2.203(3), Ag(1)–N(41A) 2.195(3), Ag(1)–O(3A) 2.557(5); N(11)–Ag(1)–N(41A) 150.3(1), N(11)–Ag(1)–O(3A) 104.9(1), N(41A)– Ag(1)–O(3A) 104.2(1).silver complexes of IV and V were polymeric, the solubility of 10 suggested that this was not a polymeric species and, hence, a crystal structure determination seemed worthwhile. Suitable crystals were obtained by vapour diVusion of acetone into a DMSO solution of the complex. The palladium complex 10 crystallises in the monoclinic space group C2/c, with the asymmetric unit containing one PdLCl2 moiety and half an acetone solvate molecule, the latter lying on a twofold rotation axis.The complex is a dimetalloparacyclophane incorporating a 26 membered macrocycle. Fig. 8 shows the structure with the asymmetric unit labelled and lists selected interatomic distances and angles. Unfortunately, the structure exhibits considerable disorder. As shown in Fig. 8, each chlorine atom is disordered over three sites. High thermal displacement parameters for some atoms of the ligand suggested that these too are disordered, but this was not included in the refinement model. Consequently, the bonding Fig. 7 Perspective view of complex 5, with hydrogen and nitrate atoms omitted, showing the intra- and inter-polymer p–p stacking of pyridine rings. Fig. 8 Perspective view and atom labelling of the crystal structure of complex 10. Selected interatomic distances (Å) and angles (8): Pd(1)– N(11) 2.03(1), Pd(1)–N(41A) 2.06(1), Pd(1)–Cl range 2.267(6)– 2.383(6); N(11)–Pd(1)–N(41A) 172.1(8), N(11)–Pd(1)–Cl range 86.9(4)–92.5(4), N(41A)–Pd(1)–Cl range 87.1(6)–92.1(4), Cl(1A)– Pd(1)–Cl(1B) 178.4(3), Cl(2A)–Pd(1)–Cl(2B) 171.8(7), Cl(3A)–Pd(1)– Cl(3B) 178.6(5).J.Chem. Soc., Dalton Trans., 1998, 3927–3933 3931 geometry parameters are less well determined than in the other structures. Each palladium atom is approximately square planar and is trans co-ordinated to two pyridine nitrogens and two disordered chlorine atoms.The Pd–N and Pd–Cl distances are within the range of reported values for related structures.14 Within the ligand, the mean planes of the pyridine rings are inclined at angles of 25(1) and 47(1)8 to that of the benzene ring, while two pyridines co-ordinated to the same palladium atom are inclined at an angle of 15(1)8. The intramolecular Pd ? ? ? Pd separation in the macrocycle is 11.269(2) Å.This is significantly larger than those of the silver structures described above, as a result of the additional methylene groups incorporated in the ligand. The two non-stacked benzene rings are inclined at an angle of 44(1)8 to each other, with their centroids separated by 6.45(3) Å, which indicates that this dimension of the macrocycle is also enlarged from the earlier structures. The overall shape of the macrocycle is curved (Fig. 9). This is as a result of hydrogen bonding interactions between a co-ordinated chlorine and the acetone solvate molecule [closest Cl ? ? ? H distance 2.59(2) Å].This has the eVect of wrapping the macrocycle around the solvate to give the concave curvature. Although there are no intramolecular p–p interactions, the molecular packing involves weak intermolecular p–p stacking interactions between the rings of adjacent molecules. Conclusion The four new ligands, 1,3-bis(2-pyridyloxy)benzene II, 1,2-bis- (2-pyridyloxy)benzene III, 1,4-bis(2-pyridyloxymethyl)benzene IV and 1,4-bis(2-pyridylmethoxy)benzene V, have been shown to form complexes with silver nitrate and palladium chloride, in excellent yields.Crystal structure determinations have shown a variety of modes of co-ordination by these ligands, including the formation of a mononuclear chelated complex, bridged binuclear cyclic dimers and bridged polymeric structures. Further experiments designed to rationalise the behaviour of these ligands will be reported subsequently.For example, the accompanying paper describes the silver complexes of structurally related ligands incorporating sulfur atoms rather than oxygen linking groups. Experimental General The 1H NMR spectra were recorded on a Varian 300 Unity spectrometer with a 3 mm probe operating at 300 MHz, 13C NMR spectra on a Varian 300 Unity or XL-300 spectrometer with a 3 or 5 mm probe, respectively, operating at 75 MHz. Spectra were referenced relative to internal Me4Si.Melting points were determined using an Electrothermal melting point apparatus and are uncorrected. Elemental analyses were performed by the Chemistry Department, University of Otago, Dunedin. Solvents were purified according to literature pro- Fig. 9 Side view of complex 10 showing the macrocycle wrapping around the hydrogen-bonded acetone. All hydrogens except those of the acetone are omitted for clarity. cedures. Unless otherwise stated, reagents were obtained from commercial sources.Ligand preparations 1,3-Bis(2-pyridyloxy)benzene II. A mixture of 1,3- dihydroxybenzene (1.15 g, 10.4 mmol), 2-bromopyridine (3.31 g, 21.0 mmol) and potassium carbonate (2.90 g, 21.0 mmol) was heated, with stirring, at 210–220 8C for 5 h. The resulting tar was extracted several times with diethyl ether; the extracts were combined and washed with aqueous sodium hydroxide (40%), then water. The solvent was then removed to give a brown oil.Crystallisation of this oil from light petroleum–ethyl acetate (2 : 1) aVorded pure compound II (1.06 g, 39%) as colourless crystals, mp 50–51 8C (Found: C, 72.90; H, 4.76; N, 10.56. C8H6NO requires C, 72.72; H, 4.58; N, 10.60%). 1H NMR (CDCl3): d 6.93 (2 H, d, H39), 6.96 (1 H, s, H2), 6.99 (2 H, dd, H4,6), 7.00 (2 H, t, H59), 7.40 (1 H, t, H5), 7.68 (2 H, t, H49), 8.21 (2 H, d, H69). 13C NMR (CDCl3): d 111.48 (C39), 113.88 (C2), 116.81 (C4,6), 118.56 (C59), 129.92 (C4), 139.26 (C49), 147.53 (C69), 154.98 (C1,3), 163.07 (C29). 1,2-Bis(2-pyridyloxy)benzene III. 1,2-Dihydroxybenzene was treated as above and gave crude compound III as a white solid. Recrystallisation from light petroleum (bp 40–60 8C)–ethyl acetate (2 : 1) aVorded pure III (0.87 g, 32%) as colourless crystals, mp 97–98 8C (Found: C, 72.55; H, 4.69; N, 10.70). 1H NMR (CDCl3): d 6.70 (2 H, d, H39), 6.90 (2 H, t, H59), 7.28 (4 H, s, H3–6), 7.56 (2 H, t, H49), 8.10 (2 H, d, H69). 13C NMR (CDCl3): d 110.73 (C39), 118.17 (C59), 123.60 (C3,6), 125.73 (C4,5), 138.96 (C49), 145.66 (C1,2), 147.31 (C69), 163.06 (C29). 1,4-Bis(2-pyridyloxymethyl)benzene IV. A mixture of 1,4- benzenedimethanol (0.57 g, 4.1 mmol), 2-bromopyridine (1.31 g, 8.3 mmol) and potassium hydroxide (1.86 g, 33.2 mmol) was refluxed in toluene (20 ml) for 42 h. The solvent was then removed and the residue divided between chloroform (15 ml) and aqueous (30 ml) layers. The organic layer was then separated, dried (Na2SO4) and concentrated to give crude compound IV.Unchanged 2-bromopyridine was removed under vacuum at room temperature for 30 min. Trituration of the remaining residue with ice-cold methanol gave a white precipitate which was filtered oV to give IV (0.41 g, 34%), mp 67–68 8C (Found: C, 72.03; H, 5.55; N, 9.17. C18H16N2O2?0.5H2O requires C, 71.75; H, 5.69; N, 9.30%). 1H NMR (CDCl3): d 5.38 (4 H, s, CH2), 6.80 (2 H, d, H39), 6.87 (2 H, t, H59), 7.47 (4 H, s, H2,3,5,6), 7.57 (2 H, t, H49), 8.17 (2 H, d, H69). 13C NMR (CDCl3): d 67.11 (CH2), 111.19 (C39), 116.79 (C59), 127.96 (C2,3,5,6), 136.87 (C1,4), 138.48 (C49), 146.70 (C69), 163.45 (C29). 1,4-Bis(2-pyridylmethoxy)benzene V. A mixture of 1,4- dihydroxybenzene (2.05 g, 18.6 mmol), 2-chloromethylpyridine hydrochloride (6.11 g, 37.2 mmol) and 40% aqueous tetrabutylammonium hydroxide (4 drops) was refluxed in benzene (40 ml) and 40% aqueous sodium hydroxide (8 ml) for 24 h. The organic layer was then separated, dried (Na2SO4) and concentrated to give crude compound V.Recrystallisation from light petroleum–ethyl acetate (10 : 1) gave pure V (2.63 g, 48%), mp 106 8C (Found: C, 73.55; H, 5.51; N, 9.46. C9H8NO requires C, 73.96; H, 5.52; N, 9.58%). 1H NMR (CDCl3): d 5.16 (4 H, s, CH2), 6.92 (4 H, s, H2,3,5,6), 7.23 (2 H, t, H59), 7.53 (2 H, d, H39), 7.72 (2 H, t, H49), 8.60 (2 H, d, H69). 13C NMR (CDCl3): d 71.14 (CH2), 115.69 (C2,3,5,6), 121.24 (C39), 122.54 (C59), 136.78 (C49), 149.16 (C69), 152.81 (C1,4), 157.43 (C29).Silver nitrate complexes [Ag2(II)2(H2O)2][NO3]2 2. Reaction of compound II (70 mg, 0.26 mmol), dissolved in methanol (8 ml), with silver nitrate (54 mg, 0.31 mmol), dissolved in water (5 ml), gave, over a period of several days at room temperature, complex 2 as colourless crys-3932 J. Chem. Soc., Dalton Trans., 1998, 3927–3933 Table 1 Crystal data and details of data collections and structure refinements for complexes 2, 4, 5, 8 and 10 Formula Formula weight Crystal system a/Å b/Å c/Å b/8 V/Å3 Space group Z Dc/Mg m23 F(000) T/K Crystal form Crystal size/mm m/mm21 2q range/8 Reflections collected Unique reflections (Rint) Parameters DiVerence peaks/e Å23 Goodness of fit Ra [I > 2s(I)] wRb (all data) 2 C32H28Ag2N6O12 904.34 Monoclinic 20.751(4) 9.311(1) 18.659(3) 113.21(1) 3313.4(9) C2/c 4 1.81 1808 158(2) Colourless block 0.56 × 0.49 × 0.31 1.26 4–55 4620 3797 (0.023) 243 0.427 1.074 0.0235 0.0623 8 C16H12Cl2N2O2Pd 441.58 Monoclinic 8.512(1) 17.845(1) 11.172(1) 106.41(1) 1627.9(3) P21/c 4 1.80 872 168(2) Orange block 0.60 × 0.19 × 0.16 1.48 4–50 3076 2874 (0.014) 208 0.428 0.940 0.0213 0.0523 4 C18H16AgN3O5 462.21 Monoclinic 15.793(2) 12.699(1) 10.959(1) 128.283(8) 1725.3(4) C2/c 4 1.78 928 188(2) Colourless block 0.52 × 0.28 × 0.19 1.21 4–50 1570 1513 (0.034) 126 0.766 1.068 0.0475 0.1182 5 C18H16AgN3O5 462.21 Monoclinic 12.235(1) 9.050(1) 15.820(2) 97.36(1) 1737.3(3) P21/c 4 1.77 928 130(2) Colourless block 0.31 × 0.15 × 0.12 1.20 4–50 3209 3057 (0.024) 280 0.654 0.766 0.0281 0.0441 10 C39H38Cl4N4O5Pd2 997.34 Monoclinic 15.428(4) 7.921(2) 32.63(1) 97.86(3) 3950(2) C2/c 4 1.68 2000 132(2) Yellow plate 0.68 × 0.41 × 0.03 1.23 4–48 3553 3083 (0.069) 254 0.964 0.807 0.0622 0.1874 a R = S(|Fo| 2 |Fc|)/S|Fo|.b wR = [Sw(Fo 2 2 Fc 2)2/Sw(Fo 2)2]� �� . tals suitable for single crystal structure determination (96 mg, 83%), mp 185–186 8C (Found: C, 42.61; H, 3.29; N, 9.16.C32H24Ag2N6O10?2H2O requires C, 42.50; H, 3.12; N, 9.29%). [Ag(NO3)5](III)4(H2O)2 3. Reaction of compound III (70 mg, 0.26 mmol), dissolved in methanol (5 ml), with silver nitrate (54 mg, 0.31 mmol), dissolved in methanol (5 ml), gave a colourless solution. This was concentrated to approximately 5 ml. Subsequent vapour diVusion of diethyl ether into this solution gave microcrystals of complex 3 (101 mg, 80%), mp > 130 8C (Found: C, 39.76; H, 2.50; N, 9.21.C64H48Ag5N13O23?2H2O requires C, 39.57; H, 2.70; N, 9.37%). [Ag(IV)(NO3)]n 4. Reaction of compound IV (50 mg, 0.17 mmol) dissolved in acetone (7 ml) with silver nitrate (29 mg, 0.17 mmol), dissolved in water (3 ml), gave, after slow evaporation at room temperature, colourless crystals of complex 4, suitable for single crystal structure determination (63 mg, 79%), mp 187–188 8C (Found: C, 46.68; H, 3.39; N, 9.15. C18H16Ag- N3O5 requires C, 46.78; H, 3.49; N, 9.09%).[Ag(V)(NO3)]n 5. Reaction of compound V (50 mg, 0.17 mmol), dissolved in methanol (10 ml), with silver nitrate (29 mg, 0.17 mmol), dissolved in water (10 ml), gave colourless needles of complex 5, after 15–30 min, which were filtered oV and washed with methanol (65 mg, 82%). Slow evaporation of an acetonitrile solution of 5 gave crystals suitable for single crystal structure determination, mp > 190 8C (Found: C, 46.72; H, 3.53; N, 9.23. C18H16AgN3O5 requires C, 46.78; H, 3.49; N, 9.09%).Palladium chloride complexes Pd(I)Cl2 6. Reaction of compound I (46 mg, 0.17 mmol), dissolved in hot methanol (5 ml), with palladium chloride (31 mg, 0.17 mmol), dissolved in hot aqueous hydrochloric acid (5 ml, 2 M), gave complex 6 as an orange precipitate (67 mg, 88%), mp > 300 8C (Found: C, 43.35; H, 2.80; Cl, 15.87; N, 6.26. C16H12Cl2N2O2Pd requires C, 43.52; H, 2.74; Cl, 16.06; N, 6.34%). Pd(II)Cl2 7. Reaction of compound II (60 mg, 0.23 mmol), dissolved in hot methanol (5 ml), with palladium chloride (40 mg, 0.23 mmol), dissolved in hot aqueous hydrochloric acid (5 ml, 2 M), gave crude complex 7 as an orange precipitate.This was subsequently recrystallised by vapour diVusion of acetone into a DMSO solution of the crude product to give 7 (82 mg, 81%), mp > 235 8C (decomp.) (Found: C, 41.29; H, 3.41; Cl, 14.72; N, 5.98. C16H12Cl2N2O2Pd?H2O?0.25(CH3)2SO requires C, 41.36; H, 3.26; Cl, 14.80; N, 5.85%). Pd(III)Cl2 8.Reaction of compoI (61 mg, 0.23 mmol), dissolved in hot methanol (5 ml), with palladium chloride (41 mg, 0.23 mmol), dissolved in hot aqueous hydrochloric acid (5 ml, 2 M), gave complex 8 as orange crystals suitable for single crystal structure determination (79 mg, 78%), mp > 260 8C (decomp.) (Found: C, 43.25; H, 2.65; Cl, 16.00; N, 6.38. C16H12Cl2N2O2Pd requires C, 43.52; H, 2.74; Cl, 16.06; N, 6.34%). 1H NMR (CD2Cl2): d 7.12 (2 H, t, H59), 7.16 (2 H, d, H39), 7.32 (2 H, dd, H4,5), 7.59 (2 H, dd, H3,6), 7.84 (2 H, t, H49), 8.69 (2 H, d, H69).Pd(IV)Cl2 9. Reaction of compound IV (60 mg, 0.21 mmol), dissolved in hot methanol (5 ml), with palladium chloride (36 mg, 0.21 mmol), dissolved in hot aqueous hydrochloric acid (5 ml, 2 M), gave complex 9 as an orange precipitate, which was filtered oV and washed with hot ethanol (89 mg, 87%), mp > 220 8C (decomp.) (Found: C, 44.26; H, 3.59; Cl, 14.62; N, 5.60. C18H16Cl2N2O2Pd?H2O requires C, 44.22; H, 3.72; Cl, 14.54; N, 5.74%). Pd(V)Cl2 10.Reaction of compound V (45 mg, 0.15 mmol), dissolved in hot methanol (5 ml), with palladium chloride (27 mg, 0.15 mmol), dissolved in hot aqueous hydrochloric acid (5 ml, 2 M), gave complex 10 as an orange precipitate (59 mg, 82%), mp > 300 8C. Vapour diVusion of acetone into a dimethyl sulfoxide solution of 10 gave crystals suitable for single crystal structure determination (Found: C, 46.68; H, 3.63; Cl, 14.40; N, 5.70.C36H32Cl4N4O4Pd2?CH3COCH3 requires C, 46.97; H, 3.84; Cl, 14.22; N, 5.62%). 1H NMR (DMSO): d 5.47 (8 H, s, CH2), 7.16 (8 H, s, H2,3,5,6), 7.93 (4 H, t, H59), 8.06 (4 H, d, H39), 8.49 (4 H, t, H49), 8.92 (4 H, d, H69). X-Ray crystallography The crystal data and details of the data collections and refinements for the five structures are listed in Table 1. All measure-J. Chem. Soc., Dalton Trans., 1998, 3927–3933 3933 ments were made with a Nicolet P4s diVractometer using graphite monochromatized Mo-Ka (l = 0.71073 Å) radiation.Cell parameters were determined by least-squares refinement on diVractometer angles for at least 20 accurately centred reflections. Throughout data collections (w scan mode, ±h 1k 1l) the intensities of three standard reflections were monitored at regular intervals and in no case showed variations of >7%. Intensities were corrected for Lorentz-polarisation eVects and for minor absorption using a technique based on azimuthal y scans.The structures were solved by direct methods using SHELXS15 and refined on F 2 using all data by full-matrix least-squares procedures with SHELXL 93.16 All non-hydrogen atoms were refined with anisotropic displacement parameters. Hydrogen atoms were included in calculated positions with isotropic displacement parameters 1.3 times the isotropic equivalent of their carrier atoms. The functions minimised were Sw(Fo 2 2 Fc 2), with w = [s2(Fo 2) 1 aP2 1 bP]21, where P = [max(Fo)2 1 2Fc 2]/3.CCDC reference number 186/1178. References 1 E. C. Constable, Chem. Ind., 1994, 56. 2 P. J. Stang, Chem. Eur. J., 1998, 4, 19; D. S. Lawrence, T. Jiang and M. Levett, Chem. Rev., 1995, 95, 2229 and refs. therein; B. F. Hoskins, R. Robson and D. A. Slizys, J. Am. Chem. Soc., 1997, 119, 2952; E. C. Constable, Chem. Commun., 1997, 1073; P. N. W. Baxter, J.-M. Lehn, B. O. Kneisel and D. Fenske, Chem. Commun., 1997, 2231; P. L. Jones, K. J. Byrom, J.C. JeVery, J. A. McCleverty and M. D. Ward, Chem. Commun., 1997, 1361; A. J. Blake, N. R. Champness, A. N. Khlobystov, D. A. Lemenovskii, W.-S. Li and M. Schröder, Chem. Commun., 1997, 1339; M. Fujita and K. Ogura, Bull. Chem. Soc. Jpn., 1996, 69, 1471; C. J. Jones, Chem. Soc. Rev., 1998, 289. 3 D. Braga, F. Grepioni and G. R. Desiraju, Chem. Rev., 1998, 98, 1375; T. L. Hennigar, D. C. MacQuarrie, P. Losier, R. D. Rogers and M. J. Zaworotko, Angew. Chem., Int. Ed. Engl., 1997, 36, 972; K.N. Power, T. L. Hennigar and M. J. Zaworotko, Chem. Commun., 1998, 595. 4 C. M. Hartshorn and P. J. Steel, Inorg. Chem., 1996, 35, 6902. 5 A. Del Zotto, P. Rigo and G. Nardin, Inorg. Chim. Acta, 1996, 247, 183; M. Konrad, F. Meyer, M. Büchner, K. Heinze and L. Zsolnai, Chem. Ber./Recueil, 1997, 130, 95; K. A. Hirsch, S. R. Wilson and J. S. Moore, J. Am. Chem. Soc., 1997, 119, 10401. 6 G. Hogarth and T. Norman, J. Chem. Soc., Dalton Trans., 1996, 1077; M. E. van der Boom, M.Gozin, Y. Ben-David, L. J. W. Shimon, F. Frolow, H.-B. Kraatz and D. Milstein, Inorg. Chem., 1996, 35, 7068; M. J. Hannon, C. L. Painting and W. Errington, Chem. Commun., 1997, 307. 7 J. S. Duggan, E. J. Grabowski and W. K. Russ, Synthesis, 1980, 573. 8 C. M. Hartshorn and P. J. Steel, Chem. Commun., 1997, 541 and refs. therein. 9 W. L. Jorgensen and D. L. Severance, J. Am. Chem. Soc., 1990, 112, 4768; C. A. Hunter, Angew. Chem., Int. Ed. Engl., 1993, 32, 1584; T. Dahl, Acta Chem. Scand., 1994, 48, 95. 10 P. J. Steel and E. C. Constable, J. Chem. Soc., Dalton Trans., 1990, 1389; A. J. Downard, G. E. Honey and P. J. Steel, Inorg. Chem., 1991, 30, 3733. 11 H. M. Buch and C. Kruger, Acta Crystallogr., Sect. C, 1984, 40, 28; K. Tani, M. Yabuta, S. Nakamura and T. Yamagata, J. Chem. Soc., Dalton Trans., 1993, 2781; M. Sawamura, H. Hamashima, M. Sugawara, R. Kuwano and Y. Ito, Organometallics, 1995, 14, 4549; D. Seebach, E. Devaquet, A. Ernst, M. Hayakawa, F. N. M. Kuhnle, W. B. Schweizer and B. Weber, Helv. Chim. Acta, 1995, 78, 1636; J. Cermak, S. D. Perera, B. L. Shaw and M. Thornton-Pett, Inorg. Chim. Acta, 1996, 244, 115. 12 G. Annibale, L. Canovese, L. Cattalini, G. Natile, M. Biagini-Cingi, A.-M. Manotti-Lanfredi and A. Tiripicchio, J. Chem. Soc., Dalton Trans., 1981, 2280; G. R. Newkome, F. R. Fronczek, V. K. Gupta, W. E. Puckett, D. C. Pantaleo and G. E. Kiefer, J. Am. Chem. Soc., 1982, 104, 1782; W. L. Darby, R. J. Butcher and L. M. Vallarino, Inorg. Chim. Acta, 1992, 194, 113; A. J. Canty, N. J. Minchin, B. W. Skelton and A. H. White, Aust. J. Chem., 1992, 45, 423. 13 L. M. Engelhardt, C. Pakawatchai, A. H. White and P. C. Healy, J. Chem. Soc., Dalton Trans., 1985, 117; S. Gotsis and A. H. White, Aust. J. Chem., 1987, 40, 1603; G. Smith, A. N. Reddy, K. A. Byriel and C. H. L. Kennard, Polyhedron, 1994, 13, 2425 and refs. therein. 14 M. W. Mulqi, F. S. Stephens and R. S. Vagg, Inorg. Chim. Acta, 1982, 63, 197; M. C. Navarro-Ranninger, S. Martinez-Carrera and S. Garcia-Blanco, Acta Crystallogr., Sect. C, 1985, 41, 21; M. A. Makhyoun, N. A. Al-Salem and M. S. El-Ezaby, Inorg. Chim. Acta, 1986, 123, 117; B. Viossat, N. Dung and F. Robert, Acta Crystallogr., Sect. C, 1993, 49, 84; J. Vicente, M.-T. Chicote, M.-C. Lagunas, P. G. Jones and E. Bembenek, Organometallics, 1994, 13, 1243. 15 G. M. Sheldrick, Acta Crystallogr., Sect A, 1990, 46, 467. 16 G. M. Sheldrick, SHELXL 93, University of Göttingen, 1993. Paper 8/07397J

ISSN:1477-9226    

DOI:10.1039/a807397j

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3935–3940 3935 Metallosupramolecular silver complexes of bis- and tetrakis- (2-pyridylsulfanylmethyl)benzenes Chris M. Hartshorn and Peter J. Steel * Chemistry Department, University of Canterbury, Christchurch, New Zealand. E-mail: p.steel@chem.canterbury.ac.nz Received 22nd September 1998, Accepted 24th September 1998 The three isomeric bis(2-pyridylsulfanylmethyl)benzenes and 1,2,4,5-tetrakis(2-pyridylsulfanylmethyl)benzene have been prepared by reactions of pyridine-2-thiol with the appropriate poly(bromomethyl)benzene, in the presence of triethylamine, in good yields.Each of these new ligands reacts with silver nitrate, in excellent yield, to give co-ordination complexes in which the ligands bridge two or more metal centres. The crystal structures of these complexes have been determined. The para-disubstituted isomer self-assembles into a M2L2 macrocyclic dimer. The other ligands form one-dimensional metallopolymers containing complex networks of interconnected macrocyclic rings.In the crystal packings p–p stacking interactions between aromatic rings are observed. Introduction We are currently involved in the synthesis and study of a large range of ligands represented by the generalised structure I, in which various numbers of heterocyclic donors are attached via spacer groups, X to a central arene core. For example, we have reported the ability of polypyrazolylmethylbenzenes to encapsulate metal ions,1 to form large metallosupramolecular cages,2 to bridge metal–metal bonded species 3 and to undergo double cyclometallation reactions.4 We have also shown that 1,4-bis(2-pyridyloxy)benzene self-assembles in the presence of silver nitrate into a M2L2 dimetalloparacyclophane 1 with intimate p–p stacking of the central benzene rings.5 In the accompanying paper we described experiments designed to assess the factors controlling this self-assembly process.6 One such factor that was considered was the possible role of the aryl ether oxygen atoms. We reasoned that incorporation of sulfur atoms in related ligands would lead to greater chalcogen–metal interactions, as a result of the high thiophilicity of the silver(I) ion.In contrast to the extensive literature on the co-ordination chemistry of heterocyclic thiolates,7 less attention has been paid to pyridyl thioethers,8 even though thioethers are well established ligands in co-ordination 9 and metallosupramolecular chemistry.10 Ligands containing pyridyl substituents directly linked to sulfide spacer groups might be expected to form interesting complexes with silver(I) ions.For example, a thiomethylterpyridine has O O N N O O N N Ag Ag N X n I 1 recently been shown to self-assemble into aggregated metallosupramolecular boxes in the presence of silver(I) ions, with coordination by all N and S donors.11 Furthermore, we have recently shown that reaction of silver nitrate with di-2-pyridyl sulfide, a ligand that normally acts as a N,N-bidentate chelating ligand, results in the formation of a novel binuclear complex in which the ligand co-ordinates in an N,N,S-tridentate bridging mode to form a structure with the same topology as those of the well studied organic photodimers of anthracene.12 The study of such assembly processes can provide useful information for use in the rapidly expanding area of crystal engineering. 13 We now report the preparation of four new bridging ligands that contain 2-pyridylsulfanyl groups attached to xylylene cores, and the synthesis and crystal structures of their respective complexes with silver(I) nitrate. Results and discussion The three isomeric bis(2-pyridylsulfanylmethyl)benzenes (II–IV) were prepared by adaptation of a literature procedure for the preparation of benzyl 2-pyridyl sulfide.14 Reaction of pyridine-2-thiol with the appropriate bis(bromomethyl)benzene, in the presence of triethylamine, gave II–IV in good yields, and the products were fully characterised by 1H and 13C NMR, using a combination of 1-D and 2-D NMR techniques.Each of these ligands reacted smoothly with one equivalent of silver nitrate, in aqueous methanol, to give good yields of the corresponding complexes 2–4. Crystals suitable for structure determination were obtained by slow evaporation of acetonitrile solutions of these complexes. The complex 2 crystallises in the triclinic space group P1� and is a centrosymmetric M2L2 dimetalloparacyclophane.Fig. 1 shows a perspective view of the structure, with atom labelling of the asymmetric unit. Selected interatomic distances and angles are also listed. Each silver atom is co-ordinated to two pyridine nitrogens and an oxygen atom of the nitrate anion, with all three silver–donor bond distances within the range expected for such co-ordination.15 The distances between the silver and the sulfur atoms of the ligand are greater than >3.16 Å, which in structurally related structures has been considered non-interacting.16 The co-ordination geometry of the silver is distorted T-shaped with the co-ordinated oxygen slightly displaced towards one of the pyridines.The mean planes of the pyridine rings are inclined at angles of 91.9(2) and 125.2(2)8 to3936 J. Chem. Soc., Dalton Trans., 1998, 3935–3940 that of the linking benzene ring, while the two pyridine ring mean planes at each silver atom are inclined to one another at an angle of 57.0(2)8.An important feature of this macrocycle is the absence of an intramolecular p–p stacking interaction, as the two benzene rings are significantly displaced from one another, in contrast to the situation in complexes such as 1.5,6 Thus, such p–p stacking is not a necessary feature for the formation of these M2L2 macrocycles. Within the 26-membered macrocycle the Ag ? ? ? Ag separation is 12.196(2) Å, which is considerably greater than the corresponding value [11.269(2) Å] in the structure of a concave dipalladium complex of the same ring size.6 Whilst there is an absence of intramolecular p–p stacking, the packing of the macrocycles is controlled, in part, by such interactions.As shown in Fig. 2, the dimetallomacrocycles selfassemble into a two-dimensional array within which there is p–p stacking between pyridine rings of adjacent macrocycles (average plane-to-plane distance = 3.8 Å).As observed in other cases,5,6 this stacking occurs with the aromatic rings disposed such that an atom of one ring lies approximately over the centroid of another ring. The silver complex 3 of the meta-disubstituted ligand III crystallises in the triclinic space group P1� and is a one dimensional metallopolymer. Fig. 3 shows a perspective view and atom labelling of the asymmetric unit of the structure, along with selected interatomic distances and angles.The silver atom is co-ordinated to two pyridine nitrogens and a sulfur, each Fig. 1 Perspective view and atom labelling of the crystal structure of complex 2. Selected interatomic distances (Å) and angles (8): Ag(1)– N(11) 2.166(2), Ag(1)–N(41A) 2.165(2), Ag(1)–O(1) 2.592(2), Ag(1) ? ? ? S(10) 3.160(1), Ag(1) ? ? ? S(40A) 3.164(1); N(11)–Ag(1)– N(41A) 169.70(8), N(11)–Ag(1)–O(1) 98.36(8), N(41A)–Ag(1)–O(1) 91.60(8). S S S S S S N N N N N N IV III II from diVerent ligands, as well as an oxygen from the nitrate counter ion.The geometry of these four donors around the silver is distorted tetrahedral [donor–Ag–donor angles between 88.8(3) and 119.9(3)8]. The distortion from tetrahedral geometry is possibly due to an additional semi-co-ordinative interaction with a second sulfur atom [viz. S(10)]. The mean planes of the pyridine rings are inclined at angles of 84.2(9) and 63.3(9)8 to that of the linking benzene ring, while the two pyridine ring mean planes are inclined to one another at an angle of 62.1(9)8.Co-ordination by the silver atom to donors from three diVerent ligands results in the assembly of an intriguing extended polymeric structure. Fig. 4 shows a perspective view and Fig. 2 Packing diagram of complex 2 showing p–p stacking interactions ben the pyridine rings of the macrocycles. Fig. 3 Perspective view and atom labelling of the crystal structure of complex 3. Selected interatomic distances (Å) and angles (8): Ag(1)– N(11) 2.285(7), Ag(1)–N(31A) 2.327(6), Ag(1)–S(30A) 2.658(3), Ag(1)– O(2) 2.418(7); N(11)–Ag(1)–N(31A) 119.9(3), N(11)–Ag(1)–S(30A) 118.5(2), N(11)–Ag(1)–O(2) 112.5(3), N(31A)–Ag(1)–S(30A) 106.6(2), N(31A)–Ag(1)–O(2) 88.8(3), S(30A)–Ag(1)–O(2) 105.9(2).Fig. 4 Perspective view (top), with hydrogens omitted, and schematic representation (bottom) of a section of the extended polymeric structure of complex 3. Nitrate anions are not shown.J. Chem. Soc., Dalton Trans., 1998, 3935–3940 3937 schematic representation of a section of the polymeric chain, which consists of an alternating array of two diVerent sizes of macrocycle fused by a common bond.The smaller of the two is a centrosymmetric eight-membered ring containing two silver atoms [Ag ? ? ? Ag 4.340(2) Å], bridged by a pair of pyridyl thioether moieties, with the two pyridine rings co-ordinated to diVerent silver atoms. This eight-membered ring adopts a chair conformation with the two pyridine rings transoid with respect to the Ag2S2 plane, as necessitated by the crystallographic centre of inversion. Such a substructure is also present in the silver nitrate complex of di-2-pyridyl sulfide.12 Fused to this macrocycle is a twenty-membered centrosymmetric dinuclear ring.Again, the ring is formed by coordination of a pyridine nitrogen and a sulfur atom to each of two silver atoms. However, in this case, instead of the co-ordinating sulfur bonding directly to the co-ordinating pyridine, they are separated by the arene spacer of the ligand.As a result, the macrocycle is enlarged to twenty members with an Ag ? ? ? Ag interatomic distance of 7.920(2) Å. While there are no p–p stacking interactions within this polymer chain, inspection of the molecular packing reveals that there does exist a series of such interactions between pyridine rings of adjacent chains. Crystallographically related pyridine rings are parallel stacked with a plane-to-plane separation of only 3.18(1) Å.Once again, the arrangement of this stacking is such that an atom of one ring lies over the centroid of the adjacent ring. Similar intermolecular interactions have recently been shown to exist between molecular squares derived from silver nitrate and pyrimidine.17 The silver complex 4 of the ortho-disubstituted ligand IV crystallises in the monoclinic space group P21/n and is also a one dimensional metallopolymer. Fig. 5 shows a perspective view of the labelled contents of the asymmetric unit and selected interatomic distances and angles.The silver atom is coordinated to two pyridine nitrogens and a sulfur atom, each from diVerent ligands, as well as to an oxygen from the nitrate counter ion. The silver atom further interacts with a silver atom related by a crystallographic centre of inversion [Ag–Ag distance 2.9874(5) Å]. Similar silver–silver interactions have been observed in complexes of other nitrogen-containing heterocyclic ligands.3,18 The geometry of the five atoms around the silver is distorted trigonal bipyramidal, with the central silver, sulfur and two nitrogens approximately coplanar and with the Fig. 5 Perspective view and atom labelling of complex 4. Selected interatomic distances (Å) and angles (8): Ag(1)–Ag(1A) 2.9874(5), Ag(1)– N(11A) 2.288(2), Ag(1)–N(21A) 2.377(2), Ag(1)–S(10) 2.4984(8), Ag(1)–O(2) 2.571(2); Ag(1A)–Ag(1)–N(11A) 90.79(6), Ag(1A)–Ag(1)– N(21A) 117.51(6), Ag(1A)–Ag(1)–S(10) 78.35(2), Ag(1A)–Ag(1)– O(2) 155.46(6), N(11A)–Ag(1)–N(21A) 103.82(8), N(11A)–Ag(1)– S(10) 154.44(6), N(11A)–Ag(1)–O(2) 85.09(8), N(21A)–Ag(1)–S(10) 101.71(6), N(21A)–Ag(1)–O(2) 86.91(8), S(10)–Ag(1)–O(2) 95.21(5).second silver and nitrate oxygen occupying the axial sites. The mean planes of the pyridine rings are inclined at angles of 27.9(2) and 77.9(2)8 to that of the linking benzene ring, while the two pyridine ring mean planes are inclined to one another at an angle of 62.7(2)8. The silver atom lies out of the extended planes of the co-ordinated pyridine rings by 0.214(3) and 0.226(3) Å.Fig. 6 shows a perspective view and a schematic representation of the extended polymeric structure of the complex 4, which, like that of 3, consists of a complex system of interconnected rings. In both structures a binuclear bis(pyridyl sulfide) moiety is present, and the pyridyl ring at the opposite end of these ligands co-ordinates to a silver atom of the next binuclear unit.However, there is a diVerence between the two structures as to which silver of the binuclear pair is bonded to the “bridging” pyridine. In 3 the “bridging” pyridine coordinates to the closer silver of the next binuclear pair, while in 4 it co-ordinates to the more distant silver of the next pair. As a result, the twenty-membered macrocycle linking two binuclear units in 4 incorporates four silver atoms, with an Ag ? ? ? Ag transannular interatomic distance of 8.138(1) Å.Once again, p–p stacking interactions exist between pyridine rings of adjacent chains, although in this case the separation between the rings is somewhat greater (3.55 Å). The three isomeric bis(2-pyridylsulfanylmethyl)benzenes II–IV react with silver nitrate to form a dimeric structure, for the para isomer, and one-dimensional polymeric structures, for the meta and ortho isomers. A ligand that incorporates all three of these substitution patterns as subunits is the tetrapodal ligand 1,2,4,5-tetrakis(2-pyridylsulfanylmethyl)benzene V, which might also oVer the possibility of forming metallopolymers of higher dimensionality.This new ligand was synthesized in excellent yield from pyridine-2-thiol and 1,2,4,5- tetrakis(bromomethyl)benzene (Scheme 1) and treated with two equivalents of silver nitrate to give, in high yield, a complex 5, the crystal structure of which was determined. The complex 5 crystallises in the triclinic space group P1� with the asymmetric unit containing half a ligand, a silver nitrate unit and a water molecule, and is another one-dimensional polymer.Fig. 7 shows a perspective view and atom labelling of the structure. To each silver atom is co-ordinated a chelated pyridyl moiety and a pyridine nitrogen from a diVerent ligand. The nitrate counter ion and the water molecule are disordered, approximately evenly, over two sites. For half of the silver atoms, a water oxygen is also co-ordinated [Ag(1)–O(100) Fig. 6 Perspective view (top), with hydrogens omitted, and schematic representation (bottom) of a section of the extended polymeric struture of complex 4.The nitrate anions are not shown.3938 J. Chem. Soc., Dalton Trans., 1998, 3935–3940 2.360(8) Å], and for the other half two nitrate oxygens are weakly co-ordinated to the silver [Ag(1) ? ? ? O(1A) 2.639(9), Ag(1) ? ? ? O(2A) 2.793(9) Å]. The mean planes of the pyridine rings are inclined at angles of 87.1(8) and 94.7(8)8 to that of the linking benzene ring, and at 7.9(8)8 to each other. Despite the fact that the ligand V is co-ordinated to four diVerent silver atoms, the complex is still a one-dimensional polymer, a section of which is shown in Fig. 8. The polymeric structure is a chain of twenty-membered macrocycles joined by benzene rings which are shared by adjacent macrocycles. The Ag ? ? ? Ag separation [7.447(1) Å] across the macrocycles is shorter than those of the previously described twentymembered binuclear rings, possibly as a consequence of a Fig. 7 Perspective view and atom labelling of complex 5. Selected interatomic distances (Å) and angles (8): Ag(1)–N(11) 2.271(3), Ag(1)– N(31B) 2.227(3), Ag(1)–S(10) 2.956(1), Ag(1)–O(100) 2.360(8), Ag(1) ? ? ? O(1A) 2.639(8), Ag(1) ? ? ? O(2A) 2.793(8); N(11)–Ag(1)– N(31B)45(1), N(11)–Ag(1)–S(10) 57.03(8), N(11)–Ag(1)–O(100) 91.0(2), N(31B)–Ag(1)–S(10) 88.80(8), N(31B)–Ag(1)–O(100) 119.5(2), S(10)–Ag(1)–O(100) 133.4(2).Scheme 1 S S N N S S N N CH2Br BrH2C CH2Br BrH2C N SH V Et3N CH3CN 2 AgNO3 5 p–p interaction between the two co-ordinated pyridines within the ring, the mean planes of which are separated by 3.34(1) Å. Weak p–p interactions exist between pyridine rings of adjacent polymer chains, which also exhibit short S ? ? ?S contacts [3.412(2) Å]. Such sulfur–sulfur interactions have recently been shown to exist in other supramolecular silver networks.19 Conclusion Reaction of silver nitrate with each of the four new ligands described above results in the assembly of co-ordination complexes within which the ligands bridge two or more metal centres.In general, co-ordination of the ligand nitrogen and sulfur donors results in the formation of extended onedimensional polymeric structures containing a complex system of interconnected metallomacrocyclic rings. The nature of co-ordination complexes formed from these and related ligands appears to result from a delicate interplay of thermodynamic and kinetic factors, which are now recognised to control the self-assembly of various metallosupramolecular architectures.20 Experimental General General experimental details are given in the accompanying paper.6 The isomeric bis(bromomethyl)benzenes 21 and 1,2,4,5- tetrakis(bromomethyl)benzene 22 were prepared by literature procedures.Ligand preparations 1,4-Bis(2-pyridylsulfanylmethyl)benzene II. 1,4-Bis(bromomethyl) benzene (0.59 g, 2.24 mmol) was added, with stirring, to an ice-cooled solution of pyridine-2-thiol (0.50 g, 4.48 mmol) and triethylamine (0.57 g, 5.6 mmol) in acetonitrile (10 ml). The mixture was stirred for 24 h at room temperature, filtered and concentrated to give a yellow residue that was dissolved in chloroform (15 ml). This solution was washed with water (2 × 30 ml), dried (Na2SO4), and concentrated to give crude compound II as a white solid.Recrystallisation from light petroleum (bp 40–60 8C) gave pure II (0.40 g, 55%), mp 76 8C (Found: C, 66.26; H, 4.90; N, 8.71. C9H8NS requires C, 66.63; H, 4.97; N, 8.63%). 1H NMR (CDCl3): d 4.41 (4 H, s, CH2), 6.98 (2 H, t, H59), 7.15 (2 H, d, H39), 7.33 (4 H, s, H2,3,5,6), 7.46 (2 H, t, H49), 8.45 (2 H, d, H69). 13C NMR (CDCl3): d 33.92 (CH2), 119.44 (C39), 121.92 (C59), 128.96 (C2,3,5,6), 135.82 (C49), 136.71 (C1,4), 149.23 (C69), 158.59 (C29). 1,3-Bis(2-pyridylsulfanylmethyl)benzene III. Reaction of 1,3- bis(bromomethyl)benzene (0.59 g, 2.24 mmol) with pyridine-2- thiol (0.50 g, 4.48 mmol) and triethylamine (0.57 g, 5.6 mmol), as described above for II, gave compound III as a yellow oil (0.61 g, 84%) (Found: M~1, 324.0755. C18H16N2S2 requires M~1 324.0755). 1H NMR (CDCl3): d 4.40 (4 H, s, CH2), 6.98 (2 H, t, H59), 7.14 (2 H, d, H39), 7.22 (1 H, t, H5), 7.28 (2 H, d, H4,6), 7.45 (3 H, H2,49), 8.44 (2 H, d, H69). 13C NMR (CDCl3): d 34.06 Fig. 8 Perspective view of a section of the extended polymeric structure of complex 5.The nitrate anions are not shown.J. Chem. Soc., Dalton Trans., 1998, 3935–3940 3939 Table 1 Crystal data and details of data collections and structure refinements for complexes 2–5 Formula Formula weight Crystal system a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 Space group Z Dc/Mg m23 F(000) T/K Crystal form Crystal size/mm m/mm21 2q range/8 Reflections collected Unique reflections (Rint) Parameters DiVerence peaks/e Å23 Goodness of fit Ra [I > 2s(I)] wRb (all data) 2 C36H32Ag2N6O6S4 988.66 Triclinic 7.803(2) 10.173(1) 13.206(2) 104.39(1) 103.84(2) 104.66(1) 929.9(3) P1� 1 1.77 496 168(2) Colourless block 0.26 × 0.21 × 0.13 1.33 4–55 4509 4194 (0.018) 244 0.553 0.903 0.0301 0.0598 3 C18H16AgN3O3S2 494.33 Triclinic 8.281(3) 10.011(3) 11.488(4) 106.16(2) 97.04(3) 91.75(3) 905.8(5) P1� 2 1.81 496 158(2) Colourless plate 0.54 × 0.52 × 0.12 1.37 4–50 3393 3152 (0.242) 244 2.901 0.966 0.0795 0.2145 4 C18H16AgN3O3S2 494.33 Monoclinic 8.556(1) 10.617(1) 19.827(2) 97.25(1) 1786.7(3) P21/n 4 1.84 992 169(2) Colourless block 0.64 × 0.46 × 0.27 1.39 4–50 3384 3153 (0.022) 244 1.836 1.027 0.0253 0.0643 5 C30H30Ag2N6O8S4 946.58 Triclinic 9.360(1) 9.712(1) 10.042(1) 92.42(1) 100.10(1) 109.98(1) 839.5(2) P1� 1 1.87 474 169(2) Colourless block 0.42 × 0.31 × 0.21 1.48 4–50 3148 2944 (0.022) 272 0.696 0.876 0.0320 0.0658 a R = S(|Fo| 2 |Fc|)/S|Fo|.b wR = [Sw(Fo 2 2 Fc 2)2]/Sw(Fo 2)2]� �� .(CH2), 119.36 (C39), 121.74 (C59), 127.46 (C4,6), 128.36 (C5), 129.29 (C2), 135.77 (C49), 137.81 (C1,3), 149.04 (C69), 158.36 (C29). 1,2-Bis(2-pyridylsulfanylmethyl)benzene IV. Reaction of 1,2- bis(bromomethyl)benzene (0.59 g, 2.24 mmol) with pyridine-2- thiol (0.50 g, 4.48 mmol) and triethylamine (0.57 g, 5.6 mmol), as described above for II, gave compound IV as a yellow solid that was recrystallised from light petroleum (0.61 g, 84%), mp 82 8C (Found: C, 66.42; H, 5.27; N, 8.61%). 1H NMR (CDCl3): d 4.61 (4 H, s, CH2), 6.96 (2 H, t, H59), 7.15 (2 H, d, H39), 7.19 (2 H, dd, H4,5), 7.43 (2 H, dd, H3,6), 7.45 (2 H, t, H49), 8.42 (2 H, d, H69). 13C NMR (CDCl3): d 31.71 (CH2), 119.41 (C39), 121.93 (C59), 127.53 (C3,6), 130.46 (C4,5), 135.80 (C49), 136.18 (C1,2), 149.24 (C69), 158.70 (C29). 1,2,4,5-Tetrakis(2-pyridylsulfanylmethyl)benzene V. Reaction of 1,2,4,5-tetra(bromomethyl)benzene (0.50 g, 1.11 mmol) with pyridine-2-thiol (0.50 g, 4.44 mmol) and triethylamine (0.57 g, 5.6 mmol), as described above for II, gave a precipitate of compound V directly from the reaction mixture.This was filtered oV, washed with water (3 × 10 ml) and recrystallised from acetonitrile to give pure V (0.52 g, 82%), mp 114 8C (Found: C, 62.84; H, 4.79; N, 9.98. C15H13N2S2 requires C, 63.12; H, 4.59; N, 9.82%). 1H NMR (CDCl3): d 4.52 (8 H, s, CH2), 6.95, (4 H, t, H59), 7.12 (4 H, d, H39), 7.44 (4 H, t, H49), 7.48 (2 H, s, H3,6), 8.38 (4 H, d, H69). 13C NMR (CDCl3): d 31.38 (CH2), 119.40 (C39), 121.95 (C59), 132.75 (C3,6), 135.41 (C1,2,4,5), 135.79 (C49), 149.23 (C69), 158.62 (C29). Silver nitrate complexes [Ag(II)(NO3)]2 2. Reaction of compound II (71 mg, 0.22 mmol), dissolved in hot methanol (12 ml), with silver nitrate (37 mg, 0.22 mmol), dissolved in water (3 ml), gave complex 2 as a white precipitate (73 mg, 68%), mp >162 8C (decomp.) (Found: C, 43.60; H, 3.42; N, 8.37. C18H16AgN3O3S2 requires C, 43.73; H, 3.26; N, 8.50%).[Ag(III)(NO3)]n 3. Reaction of compound III (79 mg, 0.24 mmol), dissolved in methanol (10 ml), with silver nitrate (41 mg, 0.24 mmol), dissolved in water (3 ml), gave complex 3 as a white precipitate (98 mg, 83%), mp 151–152 8C (Found: C, 43.64; H, 3.26; N, 8.57. C18H16AgN3O3S2 requires C, 43.73; H, 3.26; N, 8.50%). [Ag(IV)(NO3)]n 4. Reaction of compound IV (82 mg, 0.25 mmol), dissolved in hot methanol (10 ml), with silver nitrate (43 mg, 0.25 mmol), dissolved in hot methanol (5 ml), gave a colourless solution that was filtered.Diethyl ether was diVused into this solution to give complex 4 as a white solid (86 mg, 71%), mp 152–153 8C (Found: C, 43.51; H, 3.33; N, 8.59. C18H16AgN3O3S2 requires C, 43.73; H, 3.26; N, 8.50%). [Ag(NO3)]2(V)?2H2O 5. Reaction of compound V (70 mg, 0.12 mmol), dissolved in hot acetone (10 ml), with silver nitrate (42 mg, 0.24 mmol), dissolved in water (4 ml), gave complex 5 as a white precipitate (98 mg, 86%), mp 153–154 8C (Found: C, 38.24; H, 3.18; N, 8.94.C30H26Ag2N6O6S4?2H2O requires C, 38.07; H, 3.19; N, 8.88%). X-Ray crystallography The crystal data and details of the data collections and refinements for the four structures are listed in Table 1. All measurements were made with a Nicolet P4s diVractometer using graphite-monochromatized Mo-Ka (l = 0.71073 Å) radiation. Cell parameters were determined by least-squares refinement on diVractometer angles for at least 20 accurately centred reflections.Throughout data collections (w scan mode) the intensities of thrreflections were monitored at regular intervals and in no case showed variations of >5%. Intensities were corrected for Lorentz-polarisation eVects and for minor absorption using a technique based on azimuthal y scans. The structures were solved by direct methods using SHELXS23 and refined on F2 using all data by full-matrix least-squares procedures with SHELXL 93.24 All non-hydrogen atoms were refined with anisotropic displacement parameters. Hydrogen atoms were included in calculated positions with isotropic displacement parameters 1.3 times the isotropic equivalent of their carrier atoms.The functions minimised were Sw(Fo 2 2 Fc 2), with w = [s2(Fo 2) 1 aP2 1 bP]21, where P = [max(Fo)2 1 2Fc 2]/3. CCDC reference number 186/1179.3940 J. Chem. Soc., Dalton Trans., 1998, 3935–3940 References 1 C. M. Hartshorn and P. J. Steel, Angew.Chem., Int. Ed. Engl., 1996, 35, 2655. 2 C. M. Hartshorn and P. J. Steel, Chem. Commun., 1997, 541. 3 C. M. Hartshorn and P. J. Steel, Aust. J. Chem., 1997, 50, 1195. 4 C. M. Hartshorn and P. J. Steel, Organometallics, 1998, 17, 3487. 5 C. M. Hartshorn and P. J. Steel, Inorg. Chem., 1996, 35, 6902. 6 C. M. Hartshorn and P. J. Steel, preceding paper. 7 E. S. Raper, Coord. Chem. Rev., 1997, 165, 475. 8 J. Vicente, M.-T. Chicote and C. Rubio, Chem. Ber., 1996, 129, 327; C.Vinas, P. Angles, G. Sanchez, N. Lucena, F. Teixidor, L. Escriche, J. Casabo, J. F. Piniella, A. Alvarez-Larena, R. Kivekas and R. Sillanpaa, Inorg. Chem., 1998, 37, 701. 9 S. G. Murray and F. R. Hartley, Chem. Rev., 1981, 81, 365; A. Müller and E. Diemann, in Comprehensive Coordination Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon, Oxford, 1987, vol. 2, p. 551; H. A. Jenkins, S. J. Loeb and A. Malats i. Riera, Inorg. Chim. Acta, 1996, 246, 207; J.Casabo, T. Flor, M. N. S. Hill, H. A. Jenkins, J. C. Lockhart, S. J. Loeb, I. Romero and F. Teixidor, Inorg. Chem., 1995, 34, 5410; C. J. Mathews, W. Clegg, S. L. Heath, N. C. Martin, M. N. S. Hill and J. C. Lockhart, Inorg. Chem., 1998, 37, 199. 10 See, for example, J. R. Black, N. R. Chambers, W. Levason and G. Reid, J. Chem. Soc., Chem. Commun., 1995, 1277. 11 M. J. Hannon, C. L. Painting and W. Errington, Chem. Commun., 1997, 1805. 12 R. J. Anderson and P. J. Steel, Acta Crystallogr., Sect. C, 1998, 54, 223. 13 D. Braga, F. Grepioni and G. R. Desiraju, Chem. Rev., 1998, 98, 1375; T. L. Hennigar, D. C. MacQuarrie, P. Losier, R. D. Rogers and M. J. Zaworotko, Angew. Chem., Int. Ed. Engl., 1997, 36, 972; K. N. Power, T. L. Hennigar and M. J. Zaworotko, Chem. Commun., 1998, 595; P. J. Stang, Chem. Eur. J., 1998, 4, 19. 14 N. Furukawa, F. Takahashi, T. Kawai, K. Kishimoto, S. Ogawa and S. Oae, Phosphorus Sulphur, Relat. Elem., 1983, 16, 167. 15 L. M. Engelhardt, C. Pakawatchai, A. H. White and P. C. Healy, J. Chem. Soc., Dalton Trans., 1985, 117; S. Gotsis and A. H. White, Aust. J. Chem., 1987, 40, 1603; G. Smith, A. N. Reddy, K. A. Byriel and C. H. L. Kennard, Polyhedron, 1994, 13, 2425 and refs. therein. 16 B. de Groot, H. A. Jenkins and S. J. Loeb, Inorg. Chem., 1992, 31, 203; R. Alberto, W. Nef, A. Smith, T. A. Kaden, M. Neuburger, M. Zehnder, A. Frey, U. Abram and P. A. Schubiger, Inorg. Chem., 1996, 35, 3420. 17 C. V. K. Sharma, S. T. GriYn and R. D. Rogers, Chem. Commun., 1998, 215. 18 See, for example, N. C. Baenziger and A. W. Struss, Inorg. Chem., 1976, 15, 1807; J. P. Fackler, C. A. Lopez, R. J. Staples, S. Wang, R. E. Winpenny and R. P. Latimer, J. Chem. Soc., Chem. Commun., 1992, 146; E. C. Constable, A. J. Edwards, G. R. Haire, M. J. Hannon and P. R. Raithby, Polyhedron, 1998, 17, 243. 19 J. Dai, T. Kuroda-Sowa, M. Munakata, M. Maekawa, Y. Suenaga and Y. Ohno, J. Chem. Soc., Dalton Trans., 1997, 2363. 20 C. J. Jones, Chem. Soc. Rev., 1998, 289. 21 W. Wenner, J. Org. Chem., 1952, 17, 523. 22 J. T. Stapler and J. Bornstein, J. Heterocycl. Chem., 1973, 10, 983. 23 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467. 24 G. M. Sheldrick, SHELXL 93, University of Göttingen, 1993. Paper 8/07398H

ISSN:1477-9226    

DOI:10.1039/a807398h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3941–3946 3941 Reactions of transition-metal nitrido compounds with B(C6F5)3 : crystal structure of [Re{NB(C6F5)3}(PMe2Ph)(S2CNMe2)2] Linda H. Doerrer, Andrew J. Graham and Malcolm L. H. Green* Inorganic Chemistry Laboratory, South Parks Road, Oxford, UK OX1 3QR. E-mail : malcolm.green@chem.ox.ac.uk Received 13th July 1998, Accepted 2nd October 1998 The transition-metal nitrido complexes [Re(N)(PR3)(S2CNR92)2] (PR3 = PMe2Ph, R9 = Me; PR3 = PMePh2, R9 = Et 1), [Re(N)(Cl)(PMePh2)2(S2CNMe2)] 2, [Mo(N)(S2CNR2)3] (R = Me or Et) and [NBun 4][Os(N)(1,2-S2C6H4)2] reacted with the strong Lewis acid B(C6F5)3 to yield the adducts [Re{NB(C6F5)3}(PR3)(S2CNR92)2] (PR3 = PMe2Ph, R9 = Me 3*; PR3 = PMePh2, R9 = Et 4), [Re{NB(C6F5)3}(Cl)(PMePh2)2(S2CNMe2)] 5, [Mo{NB(C6F5)3}(S2CNR2)3] (R = Me 6 or 7) and [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8 (* indicates that the compound has been structurally characterised).Reactions of 3, 6 and 8 with competing strong Lewis bases have revealed diVerences in the stability of the M]] ] N–B interaction depending on the steric crowding around the metal centre.Reaction of 8 with MeO3SCF3 causes the formation of [Os{NB(C6F5)3}{1,2-(S)(SMe)C6H4}(1,2-S2C6H4)] 9. Recently we have been exploring the varied chemistry of the strong Lewis acid B(C6F5)3 which is crystalline and easily synthesized. As well as mediating unusual and unexpected reactions, 1,2 it has been shown to form relatively stable adducts with transition metal oxo complexes.3,4 Transition metal nitrido complexes are another class of nucleophiles that might be expected to react with strong Lewis acids; indeed reactions between rhenium nitrido compounds and boron trihalides have been reported.5,6 However, adducts with triarylboranes have only previously been demonstrated by indirect reaction 7 and electrophilic attack at nitrido moieties bound to other metals has been confined to carbocationic Lewis acids.8,9 We here describe studies into the reactivity of B(C6F5)3 with transition metal nitrido complexes.† Results and discussion The new rhenium(V) nitrido complexes [Re(N)(PMePh2)- (S2CNEt2)2] 1 and [Re(N)(Cl)(PMePh2)2(S2CNMe2)] 2 were prepared by reaction of [Re(N)Cl2(PMePh2)3] 11 with 2 equivalents of NaS2CNEt2?3H2O and 1 equivalent of NaS2CNMe2? H2O respectively in refluxing methanol.The preparations used were analogous to those used by Ritter and Abram12 to prepare [Re(N)(PMe2Ph)(S2CNEt2)2] and [Re(N)(Cl)(PMe2Ph)2(S2- CNMe2)]. Characterisation was undertaken by means of microanalysis and IR and NMR spectroscopies; these data are summarised in Table 1.Assignments were straightforward; the Re]] ] N stretches in the IR spectra were assigned by analogy with those of previously reported complexes. Treatment of each of the transition-metal nitrido complexes [Re(N)(PR3)(S2CNR92)2] (PR3 = PMe2Ph, R9 = Me; PR3 = PMe- Ph2, R9 = Et 1), [Re(N)(Cl)(PMePh2)2(S2CNMe2)] 2, [Mo(N)- (S2CNR2)3] (R = Me or Et) and [NBun 4][Os(N)(1,2-S2C6H4)2] with an excess of B(C6F5)3 in dichloromethane at ambient temperature yields the nitridometal–Lewis acid adducts as purple [Re{NB(C6F5)3}(PR3)(S2CNR92)2] (PR3 = PMe2Ph, R9 = Me 3; PR3 = PMePh2, R9 = Et 4), orange [Re{NB(C6F5)3}(Cl)- (PMePh2)2(S2CNMe2)] 5, cream and red-brown [Mo{NB- (C6F5)3}(S2CNR2)3] (R = Me 6 or Et 7) and olive-green [NBun 4]- † During the revision of this manuscript a report of work demonstrating diVerent reactivity between arylborane species and osmium nitrido compounds was published.10 [Os{NB(C6F5)3}(1,2-S2C6H4)2] 8 respectively (Scheme 1).These compounds are all highly soluble in dichloromethane but virtually insoluble in hydrocarbon solvents and removal of the excess of borane by washing thoroughly with hexanes was generally suYcent to obtain analytically pure product. The compounds 3–8 are reasonably air- and moisture-tolerant in the solid state and can be stored indefinitely under an inert atmosphere without decomposition.Yields were generally quite high (ca. 60–90%). Compounds 3–8 have been characterised by standard techniques, namely 1H, 11B, 13C, 19F and 31P NMR and IR spectroscopies, microanalysis and, in the case of 3, 4 and 6, FAB (Fast Atom Bombardment) mass spectrometry (Table 1). A single crystal determination of compound 3 has been carried out; crystals of 4 were also obtained but they proved to be of insuf- ficient quality to obtain a properly refined structure although the connectivity between atoms could be verified.As expected, an upfield shift of the 11B-{1H} NMR spectroscopic resonance of B(C6F5)3 (d 51) to d 23 to 26 is observed on adduct formation. This is consistent with the presence of a four-co-ordinate boron species and hence with the expected formation of an M]] ] N–B dative bond. Infrared spectroscopy was in the main rather uninformative as regards the strength of the M]] ] N–B interaction since pentafluorophenyl groups display strong absorptions in the range 900–1100 cm21, preventing unambiguous assignment of the M]] ] N stretch.One might expect that the nitridometal complex–Lewis acid interaction would weaken the M]] ] N bond, as is found to be the case with Lewis acid adducts of oxometal complexes.13 However, in practice, an increase in the M]] ] N stretching frequency is usually observed, e.g.the Re]] ] N stretch in [AsPh4][Re(N)Br4] is observed to move from 1099 to 1170 cm21 on addition of BBr3.14 This is often attributed to resonance between the B–N and M]] ] N stretches; however one of the referees has suggested an alternative explanation based on Molecular Orbital theory. The “nitrogen lone pair” MO in the parent nitrido compound has significant M–N s-antibonding character. Upon co-ordination of the Lewis acid this orbital gains some B–N bonding character hence increasing the M–N stretching frequency.In general, IR stretching frequencies are more sensitive to such changes in antibonding/bonding character than are bond distances. Tentative assignments of the M]] ] N stretch in compounds 3 to 8, based on the premise that they increase upon borane co-3942 J. Chem. Soc., Dalton Trans., 1998, 3941–3946 Scheme 1 Reagents and conditions: (i) excess of B(C6F5)3 in CH2Cl2; (ii) excess of MeO3CF3 in CH2Cl2. N Re S PR3 S S S C CNR'2 N Re S PR3 S S S C CNR'2 N Re S Ph2MeP S Ph2MeP N Mo S S S S S S CNR2 R2NC CNR2 N Mo S S S S S S CNR2 R2NC CNR2 N Os S S S S N Os S S S S N Os S S S S Me CNMe2 Cl N Re S Ph2MeP S Ph2MeP CNMe2 Cl B(C6F5)3 B(C6F5)3 B(C6F5)3 B(C6F5)3 [Bun 4N] [Bun 4N] i i i i PR3 = PMe2Ph, R' = Me; PR3 = PMePh2, R' = Et 1 PR3 = PMe2Ph, R' = Me 3; PR3 = PMePh2, R' = Et 4 R'2N R'2 N 2 5 R = Me 6, Et 7 8 B(C6F5)3 9 ii ordination, are given in Table 1.The C6F5 regions of the 13C-{1H} and 19F NMR spectra of compounds 3 to 9 have been assigned by analogy with those of B(C6F5)3 15 and also with those of other published adducts.16 For compounds 3 to 5, the phenyl resonances in the 13C-{1H} NMR spectra were assigned by examination of the magnitude of nJPC; for compound 4, a 13C–1H correlation experiment was performed to assign the downfield region of the 1H NMR spectrum.Other assignments are straightforward and are not discussed further. The molecular structure of compound 3 is shown in Fig. 1; principal bond distances and angles are given in Table 2 . The rhenium centre has an octahedral co-ordination environment with the nitride–borane unit occupying an axial site and the phosphine cis to it. The starting nitrido complex has not been structurally characterised, however a single crystal X-ray study has been performed on the closely related compound [Re(N)- (PMe2Ph)(S2CNEt2)2].12 Comparison of the two structures reveals the expected features, namely an extremely small increase in Re–N bond distance on Lewis acid co-ordination [from 1.666(6) to 1.700(4) Å], an almost linear Re]] ] N–B moiety [170.9(3)8] and a significant diminishing of the trans influence of the nitrido ligand (reduction in the diVerence between the Re–Strans and Re–Scis distances from ca. 0.35 to ca. 0.25 Å). Table 3 shows the structurally characterised borane adducts of octahedral rhenium nitrido complexes found in the literature and demonstrates the generality of all three of these features.Crystallographically characterised starting metal nitrido compounds are included for comparison purposes.J. Chem. Soc., Dalton Trans., 1998, 3941–3946 3943 Table 1 Analytical and spectroscopic data for compounds 1–9 Compounda 1 [Re(N)(PMePh2)(S2CNEt2)2] Orange-brown C, 40.2 (39.6); H, 4.7 (4.8); N, 5.8 (6.0); P, 4.6 (4.4) IR: 1358s, 1302s, 1273s, 1212s, 1146s, 1096s, 1076s, 1054s [n(Re–N)], 914s, 888s NMR Data b 1H: 1.04 (dd, 6 H, 3JHH = 7.1, 7.2, NCH2CH3), 1.36 (dd, 6 H, 3JHH = 7.2, 7.2, NCH2CH3), 2.31 (d, 3 H, 2JPH = 8.8, PCH3), 3.36, 3.58, 3.71 and 3.87 (m, 2 H each, NCH2CH3), 7.3– 7.8 (m, 10 H, PC6H5) 13C-{1H}: 12.45 and 12.84 (s, NCH2CH3), 18.02 (d, 1JCP = 36.7, PCH3), 45.14 and 46.03 (s, NCH2CH3), 128.38 (d, 2JCP = 11.2, PC6H5, Co), 131.56 (s, PC6H5, Cp), 134.04 (d, 1JCP = 48.5, PC6H5, Cipso), 135.17 (d, 3JCP = 8.9, PC6H5, Cm), 202.66 and 223.87 (s, S2CNEt2) 31P-{1H}: 25.05 (s) 2 [Re(N)(Cl)(PMePh2)2(S2CNMe2)] Yellow-brown C, 46.5 (46.1); H, 4.1 (4.3); N, 4.0 (3.7) IR: 1261s, 1094s, 1060s [n(Re–N)], 1020s, 800s 1H: 2.13 (d, 6 H, 2JPH = 8.9, PCH3), 3.24 (s, 6 H, NCH3), 7.1–7.7 (m, 20 H, PC6H5) 31P-{1H}: 210.77 (s) 3 [Re{NB(C6F5)3}(PMe2Ph)(S2CNMe2)2] Lavender C, 35.4 (35.2); H, 2.1 (2.1); B, 0.7 (1.0); N, 3.2 (3.85) Mass: 579, [M 2 B(C6F5)3]1 IR: 2727s, 1304s, 1281s, 1156s, 1092s [n(Re–N)] d 1H: 1.81 and 1.97 (d, 3 H each, 2JPH = 9.6, PCH3), 2.67, 2.98, 3.28 and 3.35 (s, 3 H each, NCH3), 7.3–7.4 (m, 5 H, PC6H5) 11B-{1H}: 23.9 (s) 13C-{1H}: 15.82 (d, 1JCP = 33.6, PCH3), 16.04 (d, 1JCP = 39.3, PCH3), 39.07,c 39.17 and 39.71 (s, NCH3), 117.5 (br s, BC6F5, Cipso), 127.32 (d, 2JCP = 9.7, PC6H5, Co), 130.03 (s, PC6H5, Cp), 131.37 (d, 3JCP = 8.3, PC6H5, Cm), 134.28 (d, 1JCP = 53.1, PC6H5, Cipso), 136.89 (d, 1JCF = 257, BC6F5, Cm), 139.42 (d, 1JCF = 247, BC6F5, Cp), 147.77 (d, 1JCF = 241, BC6F5, Co), 201.75 and 228.20 (s, S2CNMe2) 31P-{1H}: 226.57 (s) 4 [Re{NB(C6F5)3}(PMePh2)(S2CNEt2)2] e Purple C, 41.3 (41.2); H, 3.05 (3.0); B, 0.85 (0.9); N, 3.4 (3.4); P, 2.6 (2.5) Mass: 1209, M1; 1042, [M 2 C6F5]1; 842, [M 2 PMePh2 2 C6F5]1; 697, [M 2 B(C6F5)3]1; 581, [M 2 B(C6F5)3 2 4Et]1; 549, [M 2 B(C6F5)3 2 S2- CNEt2]1; 497, [M 2 B(C6F5)3 2 PMePh2]1; 399, [M 2 B(C6F5)3 2 2S2CNEt2]1 IR: 1303m, 1276m, 1147m, 1089m [n(Re–N)],d 978s 1H: 0.86 [d, 3 H, 3JHH = 6.5 (CH3)2CHOH], 0.94, 1.10, 1.25 and 1.37 (dd, 3 H each, 3JHH = 7.0, 7.0, NCH2CH3), 2.20 (d, 3 H, 2JPH = 9.0, PCH3), 3.22, 3.29, 3.42, 3.52, 3.59, 3.72, 3.78 and 3.81 (m, 1 H each, NCH2CH3), 3.73 [m, 0.5 H, (CH3)2CHOH], 7.22 (dd, 2 H, 3JPH = 8.5, 3JHH = 8.5, PC6H5, Ho), 7.29 (t, 1 H, 3JHH = 8.5, PC6H5, Hp), 7.36 (dd, 2 H, 3JHH = 8.5, 8.5, PC6H5, Hm), 7.42 (m, 2 H, PC6H5, Ho), 7.43 (m, 1 H, PC6H5, Hp), 7.65 (dd, 2 H, 3JHH = 8.5, 8.5, PC6H5, Hm) 11B-{1H}: 23.4 (s) 13C-{1H}: 1.10 [s, (CH3)2CHOH], 12.01, 12.43, 12.58 and 12.67 (s, NCH2CH3), 17.22, (d, 1JCP = 36.8, PCH3), 22.68 [s, (CH3)2CHOH], 44.72, 44.96, 45.00 and 45.98 (s, NCH2CH3), 119.0 (br s, BC6F5, Cipso), 128.02 and 128.38 (d, 2JCP = 11.0, PC6H5, Co), 130.23 and 130.93 (s, PC6H5, Cp), 132.2 and 133.40 (d, 3JCP = 9.2, PC6H5, Cm), 133.7 and 136.88 (d, 1JCP = 49.2, PC6H5, Cipso), 136.81 (d, 1JCF = 271, BC6F5, Cm), 139.52 (d, 1JCF = 245, BC6F5, Cp), 148.02 (d, 1JCF = 241, BC6F5, Co), 199.95 and 229.78 (s, S2CNEt2) 19F: 2168.79 (dd, 6 F, 3JFF = 22.6, 20.7, BC6F5, Fm), 2163.90 (t, 3 F, 3JFF = 20.7, BC6F5, Fp), 2133.80 (d, 6 F, 3JFF = 22.6, BC6F5, Fo) 31P-{1H}: 212.86 (s) 5 [Re{NB(C6F5)3}(Cl)(PMePh2)2(S2CNMe2)] Orange C, 44.5 (44.5); H, 3.3 (2.5); N, 1.9 (2.2) IR: 1099w [n(Re–N)],d 970w, 895m, 722m 1H: 1.79 (d, 6 H, 2JPH = 9.5, PCH3), 3.16 (s, 6 H, NCH3), 7.0–7.8 (m, 20 H, PC6H5) 11B-{1H}: 22.7 (s) 13C-{1H}: 15.31 (d, 1JCP = 36.6, PCH3), 19.17 (d, 1JCP = 38.3, PCH3), 38.58 and 39.65 (s, NCH3), 120.1 (br s, BC6F5, Cipso), 127.84, 128.39, 129.01, 130.17, 130.92, 131.28, 132.18, 132.62 and 133.32 (PC6H5), 134.37 (d, 1JCP = 59.6, PC6H5, Cipso), 135.82 (d, 1JCP = 58.9, PC6H5, Cipso), 136.68 (d, 1JCF = 252, BC6F5, Cm), 139.27 (d, 1JCF = 263, BC6F5, Cp), 148.03 (d, 1JCF = 253, BC6F5, Co), 191.15 (s, S2CNMe2) 31P-{1H}: 216.73 (s) 6 [Mo{NB(C6F5)3}(S2CNMe2)3] Cream C, 33.6 (33.0); H, 1.9 (1.85); B, 1.0 (1.1); N, 4.9 (5.7) Mass: 984, M1; 773, [M 2 C6F5 2 MNe2]1; 697, [M 2 C6F5 2 S2CNMe2]1; 472, [M 2 B(C6F5)3]1; 352, [M 2 B(C6F3)3 2 S2CNMe2]1 IR: 1645w, 1559m, 1514m, 1305m, 1156m, 1082m (br) [n(Re–N)],d 979m (br) 1H: 3.19 (s, 3 H, NCH3), 3.25 (s, 6 H, NCH3), 3.34 (s, 6 H, NCH3), 3.38 (s, 3 H, NCH3) 11B-{1H}: 26.5 (s) 13C-{1H}: 35.63, 37.36, 40.68 and 41.10 (s, NCH3),f 119.0 (br s, BC6F5, Cipso), 136.86 (d, 1JCF = 241, BC6F5, Cm), 139.28 (d, 1JCF = 247, BC6F5, Cp), 147.81 (d, 1JCF = 239, BC6F5, Co), 200.78 and 203.62 (s, S2CNMe2) g 19F: 2168.98 (m, 6 F, BC6F5, Fm), 2163.87 (t, 3 F, 3JFF = 26.3, BC6F5, Fp), 2134.17 (d, 6 F, 3JFF = 18.8, BC6F5, Fo) 7 [Mo{NB(C6F5)3}(S2CNEt2)3] Red-brown C, 38.7 (37.2); H, 3.1 (2.8); B, 1.0 (1.0); N, 4.6 (5.25) IR: 1302s, 1262s, 1209m, 1152m, 1089m [n(Re–N)],d 976s 1H: 1.11–1.46 (m, 18 H, NCH2CH3), 3.61–3.81 (m, 12 H, NCH2CH3) 11B-{1H}: 26.6 (s) 13C-{1H}: 11.77, 12.23, 12.36 and 12.48 (2, NCH2CH3),f 43.31, 44.46, 45.58 and 46.64 (s, NCH2CH3),f 119.8 (br s, BC6F5, Cipso), 136.87 (d, 1JCF = 255, BC6F5, Cm), 139.24 (d, 1JCF = 248, BC6F5, Cp), 147.81 (d, 1JCF = 240, BC6F5, Co), 199.39 and 202.38 (s, S2CNEt2) g 8 [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] Olive-green C, 44.9 (44.6); H, 3.25 (3.6); B, 1.2 (0.9); N, 2.0 (2.3) IR: 1644m, 1515s, 1283m, 1275m, 1097s, 979s, 794m 1H: 0.90 [t, 12 H, 3JHH = 7.1, N(CH2)3CH3], 1.26 [m, 8 H, N(CH2)2CH2CH3], 1.37 (m, 8 H, NCH2CH2CH2CH3), 2.63 [m, 8 H, NCH2(CH2)2CH3], 7.03 and 7.68 (m, AA9BB9 spin system, 8 H, S2C6H4) 11B-{1H}: 23.7 (s) 13C-{1H}: 13.32 [s, N(CH2)3CH3], 19.72 [s, N(CH2)2CH2CH3], 23.68 (s, NCH2CH2- CH2CH3), 58.98 [s, NCH2(CH2)2CH3], 114.0 (br s, BC6F5, Cipso), 124.69 and 127.90 (s, S2C6H4), 136.67 (d, 1JCF = 238, BC6F5, Cm), 139.56 (d, 1JCF = 228, BC6F5, Cp), 147.56 (d, 1JCF = 246, BC6F5, Co), 149.77 (s, S2C6H4, Cipso) 19F: 2168.80 (m, 6 F, BC6F5, Fm), 2162.99 (t, 3 F, 3JFF = 20.7, BC6F5, Fp), 2134.72 (d, 6 F, 3JFF = 24.0, BC6F5, Fo) 9 [Os{NB(C6F5)3}{1,2-(S)(SMe)C6H4}(1,2-S2C6H4)] Dark green oil h 1H: 3.06 (s, SCH3), 7.0–7.8 (br, C6H4) 11B-{1H}: 22.3 (s) 13C-{1H}: 33.5 (br s, SCH3), 118.7 (br s, BC6F5, Cipso), 122.17, 127.01, 128.36, 128.97, 130.47 and 132.27 [br s, S2C6H4 and (S)(SCH3)C6H4], 136.90, (d, 1JCF = 246, BC6F5, Cm), 140.30 (d, 1JCF = 267, BC6F5, Cp), 148.12 (d, 1JCF = 240, BC6F5, Co) a Analytical data given as found (calculated) in %.Mass spectral data (Fast Atom Bombardment) given as m/z (assignment), selected IR data (cm21) as Nujol mulls. b At probe temperature. Data given as: chemical shift (d) (multiplicity, relative intensity, J in Hz, assignment).All obtained in CD2Cl2. c Two coincident resonances. d Tentative assignment, see text. e Crystallised with 0.5 molecule of PriOH. f Resonances in 2:2:1:1 intensity ratio. g Resonances in 2 : 1 intensity ratio. h Oil too sensitive to obtain microanalytical data.3944 J. Chem. Soc., Dalton Trans., 1998, 3941–3946 The NMR spectra of compounds 3–5 reveal that there is a lowering of symmetry upon co-ordination of the Lewis acid. For instance, compound 1 displays 2 methyl and 4 methylene resonances in its 1H NMR spectrum whereas 4 and 8 signals respectively are observed for the B(C6F5)3 adduct 4.This may be attributed to restriction of free rotation about the Re–P bond upon addition of the bulky triarylborane. A further point of interest is provided by compound 5 where there is apparently only one dithiocarbamate methyl environment and one phosphine methyl environment in the 1H NMR spectrum but two distinct resonances for each in the 13C-{1H} NMR spectrum.This has been attributed to the diVerent timescales involved in 1H and 13C-{1H} NMR spectroscopy. The adducts 6 and 7 display 4 distinct dithiocarbamate resonances in their 1H and 13C-{1H} NMR spectra with intensity ratios 1:2:2:1; this is in agreement with published data for the compound [Mo(NCPh3)- (S2CNMe2)3][BF4].8 The NMR spectra of compound 8 are similar to those of the parent nitrido complex and are not discussed further.In order to test the strength of the M]] ] N–B interaction, the reactions of compounds 3, 6 and 8 with a series of competing Lewis bases were attempted. The bases used were NEt3, PMe3 Fig. 1 Molecular structure of [Re{NB(C6F5)3}(PMe2Ph)(S2CNMe2)2] 3 showing the atom numbering scheme. Hydrogen atoms are omitted for clarity. Thermal ellipsoids are drawn at the 50% probability level. Table 2 Selected bond distances (Å) and angles (8) for compound 3 B(1)–N(3) Re–N(3) Re–S(1) Re–S(2) Re–S(3) Re–S(4) Re–P(1) 1.548(7) 1.700(4) 2.3769(12) 2.4350(15) 2.6090(12) 2.4608(13) 2.4154(15) B(1)–N(3)–Re C(20)–B(1)–C(30) C(30)–B(1)–C(40) C(20)–B(1)–C(40) C(20)–B(1)–N(3) C(30)–B(1)–N(3) C(40)–B(1)–N(3) 170.9(3) 103.3(4) 114.5(4) 114.7(4) 109.0(4) 115.2(4) 100.4(4) and THF and in all cases an excess of base was added to a dichloromethane solution of the adduct and after 1 h stirring the reaction residues were analysed by NMR spectroscopy.Tetrahydrofuran was shown to cause no alteration to either the 1H or 11B-{1H} NMR spectra of all 3 compounds, however NEt3 and PMe3 had diVering eVects depending on the metal centre.For the rhenium complex 3 no eVect was observed on Lewis base addition, whereas for the osmium compound 8, addition of L (L = PMe3 or NEt3) caused quantitative formation of the parent nitrido compound and L?B(C6F5)3 within 1 h. The molybdenum compound 6 displayed intermediate stability with approximately 50% displacement of the parent nitrido complex by the competing Lewis base over 1 h.The diVerence in stability of these 3 metal nitrido complex–Lewis acid adducts is probably due to steric factors since the Re]] ] N–B moiety in 3 is protected by the bulky cis tertiary phosphine and the Os]] ] N–B linkage in 8 is exposed by the ‘tied back’ dithiolate ligands. Compound 6 displays intermediate steric hindrance. Following the work of Sellmann et al.9 who demonstrated the presence of two nucleophilic sites on the compound [NBun 4]- [Os(N)(1,2-S2C6H4)2], the B(C6F5)3 adduct of this metal nitrido complex was treated with 2 competing Lewis acids, namely [Ph3C][BF4] and MeO3SCF3.As expected, the bulky trityl cation displaced the borane at the less hindered nitrido moiety to yield the known compound [Os(NCPh3)(1,2-S2C6H4)2].9 However, reaction of [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8 with an excess of methyl triflate in dichloromethane yields, after extraction into toluene, the olive-green oil [Os{NB(C6F5)3}{1,2- (S)(SMe)C6H4}(1,2-S2C6H4)] 9 in which Lewis acids are coordinated to both nucleophilic sites.This adduct has been characterised by NMR spectroscopy only since the oil was too sensitive to obtain meaningful microanalytical data. This indicates that the methyl group is relatively mobile and can move between the 4 sulfur donors. In conclusion, we have demonstrated that the nitrido group in transition metal nitrido complexes is suYciently nucleophilic to form a dative bond with the Lewis acid B(C6F5)3.These adducts are the first reported from direct reactions between a triarylborane and a transition metal nitrido complex. The M]] ] N–B interaction is reasonably strong as demonstrated by the relative stability of the adducts towards atmospheric oxygen and moisture, although stability towards competing strong Lewis bases seems to vary from metal to metal and is probably a function of steric crowding. Experimental All preparations and manipulations of air and/or moisture sensitive materials were carried out under an atmosphere of dinitrogen using standard Schlenk line techniques or in an inert-atmosphere glove-box containing dinitrogen.Dinitrogen was purified before use by passage through a drying column filled with activated molecular sieves (4 Å) and a deoxygenating column filled with either manganese(II) oxide suspended on vermiculite (Schlenk line) or BASF catalyst (glove-box).Solvents were predried over activated 4 Å molecular sieves and then distilled from sodium (toluene), sodium–potassium alloy Table 3 Comparison of cis and trans metal–ligand distances and Re–N–B bond angles for some octahedral rhenium(V) nitrido complexes and their Lewis acid adducts Compound 3 [Re{NB(C6F5)3}(PMePh2)(S2CNMe2)2] [Re(N)(PMe2Ph)(S2CNEt2)2] 12 [Re(NBCl3)(PMe2Ph)(S2CNEt2)2] 6 [Re(NBPh3)(PMe2Ph)(S2CNEt2)2] 7 [Re(N)Cl2(PMe2Ph)3] 17 [Re(NBCl3)Cl2(PMe2Ph)3] 18 Re–Ltrans/Å 2.6090 2.793(2) 2.565(2) 2.579(4) 2.633(2) 2.439(3) Re–Lcis/Å 2.3769(12)–2.4608(13) 2.396(1)–2.449(1) 2.376(2)–2.455(2) 2.362(4)–2.431(4) 2.442(2) b 2.394(3) b B–N–Re/8 170.9(3) a 170.5(3) 170.9(9) a 176.5(6) Re–N/Å 1.700(4) 1.666(6) 1.704(3) 1.653(12) 1.660(8) 1.728(7) a Not applicable. b Re–Clcis distance.J.Chem. Soc., Dalton Trans., 1998, 3941–3946 3945 [pentane and light petroleum (bp 40–60 8C)], potassium (THF) or calcium hydride (dichloromethane) under a slow continuous stream of dinitrogen.The Analar solvents methanol and PriOH were used as supplied without drying and degassed by bubbling dinitrogen through them for 15 min. Deuteriated dichloromethane for NMR spectroscopy was dried over calcium hydride and deoxygenated by three freeze–pump–thaw cycles. Deuteriochloroform was used as supplied. The NMR spectra were recorded on either a Varian Unity- Plus 500 (1H, 11B, 13C, 19F and 31P at 499.87, 160.38, 123.70, 470.28 and 202.35 MHz respectively) or a Bruker AM300 spectrometer (1H, 11B, 13C and 31P at 300.13, 96.25, 75.5 and 121.6 MHz respectively).They were referenced internally using the residual protio-solvent (1H) and solvent (13C) resonances and measured relative to tetramethylsilane (d 0), or referenced externally to BF3?Et2O (11B, d 0), CFCl3 (19F, d 0) or 85% H3PO4 (31P, d 0). Chemical shifts are quoted in d (ppm); a positive sign indicates a downfield shift relative to the standard. Fast atom bombardment mass spectra were obtained by the EPSRC Mass Spectrometry Service at the University College of Swansea under the supervision of Dr J.A. Ballantine; infrared spectra as Nujol mulls between NaCl plates on a Perkin-Elmer 1710 FTIR spectrometer in the range 400 to 4000 cm21. Elemental analyses were obtained by the microanalytical department of the Inorganic Chemistry Laboratory. The compounds [Re(N)Cl2(PMePh2)3],11 [Re(N)(PMe2Ph)- (S2CNMe2)2],12 [Mo(N)(S2CNR2)3] (R = Me or Et),19 [NBun 4][Os(N)(1,2-S2C6H4)2]9 and B(C6F5)3 1,20 were prepared by literature methods.Preparations [Re(N)(PMePh2)(S2CNEt2)2] 1. To a stirred solution of [Re(N)Cl2(PMePh2)3] (0.300 g, 0.34 mmol) in methanol (30 cm3), NaS2CNEt2?3H2O (0.233 g, 1.03 mmol) in methanol (15 cm3) was added. The reaction mixture, which immediately changed from yellow to orange, was heated to reflux for 1 h before being allowed to cool to room temperature. The solvent was removed in vacuo and the resulting, rather oily, orange solid washed with PriOH (20 cm3). Recrystallisation from PriOH and dichloromethane (15 cm3 of a 1 : 1 mixture) at 280 8C aVorded complex 1 analytically pure.Yield: 0.116 g (48%). [Re(N)(Cl)(PMePh2)2(S2CNMe2)] 2. The complex [Re(N)- Cl2(PMePh2)3] (0.436 g, 0.500 mmol) was dissolved in methanol (20 cm3) and a solution of NaS2CNMe2?H2O (72 mg, 0.500 mmol) in methanol (15 cm3) was added causing immediate darkening of the reaction mixture.The mixture was heated to reflux for 1 h after which time it was cooled to ambient temperature and concentrated to half volume. This led to precipitation of the product which was isolated analytically pure by filtration and washing with PriOH (2 × 10 cm3). Yield: 0.201 g (53%). [Re{NB(C6F5)3}(PMe2Ph)(S2CNMe2)2] 3. To a stirred solution of [Re(N)(PMe2Ph)(S2CNMe2)2] (0.289 g, 0.500 mmol) in dichloromethane (20 cm3) was added B(C6F5)3 (0.280 g, 0.547 mmol) in dichloromethane (15 cm3).Upon stirring overnight the solution darkened somewhat and the solvent was removed under vacuum to yield a grey oily solid. The product was aVorded analytically pure by trituration with pentane (2 × 15 cm3) and drying overnight in vacuo. Single crystals suitable for analysis by X-ray diVraction were grown by slow vapour diffusion of pentane into a dichloromethane (20 cm3) solution of complex 3 (ca. 20 mg). Yield: 0.300 g (55%). [Re{NB(C6F5)3}(PMePh2)(S2CNEt2)2] 4. To a stirred solution of [Re(N)(PMePh2)(S2CNEt2)2] 1 (90 mg, 0.129 mmol) in dichloromethane (15 cm3), B(C6F5)3 (72 mg, 0.140 mmol) in dichloromethane (5 cm3) was added.An immediate change from yellow to violet was observed. After 1 h of stirring, light petroleum (30 cm3) was added but no solid precipitated overnight. Hence the volatiles were removed under vacuum and the resulting oily solid triturated with light petroleum (20 cm3) to yield the product as a lavender powder. Crystallisation by slow evaporation of a solution of the product in PriOH and dichloromethane (10 cm3 of a 1 : 1 mixture) led to its isolation as purple single crystals which proved to be of insuYcient quality for analysis by X-ray diVraction.Microanalysis and NMR spectroscopy identified this product as pure complex 4?0.5 PriOH. Yield: 0.104 g (65%). [Re{NB(C6F5)3}(Cl)(PMePh2)2(S2CNMe2)] 5. To a stirred solution of [Re(N)(Cl)(PMePh2)2(S2CNMe2)] 2 (0.180 g, 0.238 mmol) in dichloromethane (15 cm3) a solution of B(C6F5)3 (0.144 g, 0.281 mmol) in dichloromethane (5 cm3) was added dropwise.An immediate change from yellow to orange-red was observed. Stirring was maintained for 3 h after which time the solvent was removed under vacuum and the resulting red oily solid triturated with pentane (20 cm3). This aVorded the product as a dark orange powder which was dried in vacuo overnight and shown to be pure by microanalysis. Yield: 0.199 g (66%). [Mo{NB(C6F5)3}(S2CNMe2)3] 6.To a stirred suspension of [Mo(N)(S2CNMe2)3] (0.200 g, 0.425 mmol) in dichloromethane (30 cm3) a solution of B(C6F5)3 (0.250 g, 0.488 mmol) in dichloromethane (15 cm3) was slowly added. Stirring was maintained for 48 h after which time all solid material had dissolved and the solution had changed from orange to red. Volatiles were removed in vacuo to yield a brown oily solid which was rendered as an orange powder by trituration with pentane (40 cm3). The yellow microcrystalline solid was obtained analytically pure by cooling a solution in dichloromethane (15 cm3) to 280 8C.Yield: 0.251 g (60%). [Mo{NB(C6F5)3}(S2CNEt2)3] 7. To a stirred suspension of [Mo(N)(S2CNEt2)3] (0.250 g, 0.451 mmol) in dichloromethane (30 cm3) a solution of B(C6F5)3 (0.245 g, 0.479 mmol) in dichloromethane (15 cm3) was slowly added. Stirring was maintained for 48 h after which time all solid material had dissolved and the solution had changed from brown to red.Volatiles were removed in vacuo to yield a brown-red oil. This failed to crystallise from a solution in dichloromethane–pentane (30 cm3 of a 1 : 4 mixture) and was rendered as a solid by sonication for 15 min in pentane (30 cm3). The red-brown foamy solid product was isolated by filtration and dried overnight in vacuo. Yield: 0.101 g (21%). [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8. To a stirred solution of [NBun 4][Os(N)(1,2-S2C6H4)2] (0.450 g, 0.619 mmol) in dichloromethane (40 cm3), B(C6F5)3 (0.450 g, 0.879 mmol) in dichloromethane (20 cm3) was added.An immediate change from yellow to deep red was observed and after 1 h of stirring the volatiles were removed under vacuum. The resulting oily green solid was triturated with pentane (20 cm3) to yield the product. It was isolated analytically pure as an olive-green powder by filtration and drying under vacuum. Yield: 0.675 g (88%). [Os{NB(C6F5)3}{1,2-(S)(SMe)C6H4}(1,2-S2C6H4)] 9. The complex [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8 (200 mg, 0.160 mmol) and MeO3SCF3 (50 mg, 0.305 mmol) were combined in a dry-box and dissolved in dichloromethane (15 cm3).The reaction mixture was stirred for 1 h without any obvious colour change and then the volatiles were removed under vacuum. Extraction of the oily residues with toluene yielded a red solution. Removal of the solvent from this aVorded the product as an olive-green oil. Vigorous washing with pentane (30 cm3) failed to yield a solid so the product was characterised3946 J.Chem. Soc., Dalton Trans., 1998, 3941–3946 by NMR spectroscopy. It was too air-sensitive to allow satisfactory microanalytical data to be obtained. Reactions with L 5 THF, NEt3 or PMe3 [Re{NB(C6F5)3}(PMe2Ph)(S2CNMe2)2] 3. To a solution of [Re{NB(C6F5)3}(PMe2Ph)(S2CNMe2)2] 3 (20 mg, 0.018 mmol) in dichloromethane (10 cm3) was added an excess of Lewis base; either THF (3 drops), NEt3 (3 drops) or PMe3 (1 cm3 of a 0.105 M solution in light petroleum, 0.105 mmol).No immediate colour change was observed and the reactions were allowed to proceed for 1 h. Volatiles were then removed in vacuo and the residues dried for 2 h. The products were analysed by 1H and 11B-{1H} NMR spectroscopy in CDCl3 and this showed that in all three experiments no reaction of compound 3 had occurred. [Mo{NB(C6F5)3}(S2CNMe2)3] 6. To a solution of [Mo{NB- (C6F5)3}(S2CNMe2)3] 6 (20 mg, 0.020 mmol) in dichloromethane (10 cm3) an excess of Lewis base was added as above.Little immediate colour change was observed and the reactions were allowed to proceed for 1 h. Further treatment and NMR analysis as above showed that in the case where L = THF no decomposition of compound 6 had occurred whereas when L = NEt3 or PMe3 the residue consisted of ca. 50% 6 and 50% [Mo(N)(S2CNMe2)3] plus B(C6F5)3?L. NMR data (CDCl3, 298 K): B(C6F5)3?NEt3, 1H, d 1.46 (m, 9 H, NCH2CH3) and 3.60 (m, 6 H, NCH2CH3); 11B-{1H}, d 24.2 (s); B(C6F5)3?PMe3, 1H, d 1.79 (d, 2JPH = 13.1 Hz, PCH3); 11B-{1H}, d 24.4 (s) [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8.To a solution of [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8 (20 mg, 0.016 mmol) in dichloromethane (10 cm3) an excess of Lewis base was added as above. Little immediate colour change was observed and the reactions were allowed to proceed for 1 h. Further treatment and NMR analysis as above showed that in the case where L = THF no decomposition of compound 8 had occurred whereas when L = NEt3 or PMe3 the residue consisted entirely of [NBun 4][Os(N)(1,2-S2C6H4)2] 9 and B(C6F5)3?L (NMR spectroscopic data as above).Reaction of [NBun 4][Os{NB(C6F5)3}(1,2-S2C6H4)2] 8 with [Ph3C][BF4] Complex 8 (50 mg, 0.040 mmol) and [Ph3C][BF4] (50 mg, 0.151 mmol) were combined as solids and dissolved in dichloromethane (15 cm3). The mixture was stirred for 30 min after which it appeared to have darkened slightly. The volatiles were removed in vacuo and the whole of the reaction residues analysed by 1H, 11B-{1H} and 13C-{1H} NMR spectroscopy in CDCl3. This confirmed the formation of the known compound [Os(NCPh3)- (1,2-S2C6H4)2],9 B(C6F5)3 and [NBun 4][BF4].Crystallography Single crystals of complex 3 suitable for analysis by X-ray crystallography were grown by slow vapour diVusion of pentane into a solution of 3 in dichloromethane at 298 K. A plate-shaped crystal was selected for diVraction, covered with paratone-N oil under an inert atmosphere and mounted on the end of a glass fibre.Crystal data. C32H23BF15N3PReS4 3, M = 1090.75, triclinic, space group P1� , a = 8.871(5), b = 13.233(1), c = 16.7220(13) Å, a = 80.73(4), b = 77.13(5), g = 86.05(5)8, V = 1887.6 Å3, Z = 2, Dc = 1.92 g cm23, m = 3.62 mm21, purple crystals, crystal dimensions 0.25 × 0.05 × 0.05 mm. Data collection and processing. Data were collected at 125 K on an Enraf-Nonius DIP2000 image plate diVractometer with graphite monochromated Mo-Ka radiation (l = 0.71069 Å). 11663 Reflections were measured (1 < q < 268, ±h, ±k, 1l, 6103 unique giving 5408 with I > 3s(I). The images were processed with the DENZO and SCALEPACK programs.21 Structure solution and refinement. All solution, refinement and graphical calculations were performed using the CRYSTALS22 and CAMERON23 software packages. The structure was solved byirect methods using the SIR 92 program24 and refined by a full-matrix least squares procedure on F.All non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were generated and allowed to ride on their corresponding carbon atoms with fixed thermal parameters. A Chebychev weighting scheme with the parameters 2.84, 0.362 and 2.06 was applied as well as an empirical absorption correction.25 This yielded R = 0.042 and R9 = 0.051 with maximum residual electron density of 1.61 e Å23. CCDC reference number 186/1188. See http://www.rsc.org/suppdata/dt/1998/3941/ for crystallographic files in .cif format.Acknowledgements We thank St John’s College for a Junior Research Fellowship (to L. H. D.) and the EPSRC for support of this work. References 1 A. N. Chernega, A. J. Graham, M. L. H. Green, J. Haggitt, J. Lloyd, C. P. Mehnert, N. Metzler and J. Souter, J. Chem. Soc., Dalton Trans., 1997, 2293. 2 J. R. Galsworthy, M. L. H. Green, N. Maxted and M. Müller, J. Chem. Soc., Dalton Trans., 1998, 387. 3 J.R. Galsworthy, M. L. H. Green, M. Müller and K. Prout, J. Chem. Soc., Dalton Trans., 1997, 1309. 4 J. R. Galsworthy, J. C. Green, M. L. H. Green and M. Müller, J. Chem. Soc., Dalton Trans., 1998, 15. 5 J. Chatt and B. T. Heaton, J. Chem. Soc. A, 1971, 705. 6 S. Ritter and U. Abram, Z. Anorg. Allg. Chem., 1996, 622, 965. 7 S. Ritter and U. Abram, Inorg. Chim. Acta, 1995, 231, 245. 8 M. W. Bishop, J. Chatt, J. R. Dilworth, B. D. Neaves, P. Dahlstrom, J. Hyde and J. Zubieta, J. Organomet. Chem., 1981, 213, 109. 9 D. Sellmann, M. W. Wemple, W. Donaubauer and F. W. Heinemann, Inorg. Chem., 1997, 36, 1397. 10 T. J. Crevier and J. M. Mayer, Angew. Chem., Int. Ed. Engl., 1998, 37, 1891. 11 J. Chatt, J. D. Garforth, N. P. Johnson and G. A. Rowe, J. Chem. Soc., 1964, 1012. 12 S. Ritter and U. Abram, Z. Anorg. Allg. Chem., 1994, 620, 1443. 13 C.-H. Yang, J. A. Ladd and V. L. Goedken, J. Coord. Chem., 1988, 18, 317. 14 W. Kafitz, F. Weller and K. Dehnicke, Z. Anorg Allg. Chem., 1982, 490, 175. 15 A. G. Massey and A. J. Park, J. Organomet. Chem., 1966, 5, 218. 16 J. Karl, G. Erker and R. Fröhlich, J. Am. Chem. Soc., 1997, 119, 11165. 17 E. Forsellani, U. Casellato, R. Graziani and L. Magon, Acta Crystallogr., Sect. B, 1982, 38, 3081. 18 R. Dantona, E. Schweda and J. Strähle, Z. Naturforsch., Teil B, 1984, 39, 733. 19 J. Chatt and J. R. Dilworth, J. Indian Chem. Soc., 1977, 54, 13. 20 A. G. Massey and A. J. Park, J. Organomet. Chem., 1964, 2, 245. 21 Z. Otwinowski and W. Minor, Methods Enzymol., 1996, 276, 307. 22 D. J. Watkin, C. K. Prout, J. R. Carruthers and P. W. Betteridge, CRYSTALS User Guide, Issue 10, Chemical Crystallography Laboratory, University of Oxford, 1996. 23 D. J. Watkin, C. K. Prout and L. J. Pearce, CAMERON, Chemical Crystallography Laboratory, University of Oxford, 1996. 24 A. Altomare, G. Cascarano, G. Giacovazzo, A. Guagliardi, M. C. Burla, G. Polidori and M. Camalli, SIR 92, Program for automatic solution of crystal structures by direct methods, J. Appl. Crystallogr., 1994, 27, 435. 25 N. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983, 39, 158. Paper 8/05425H

ISSN:1477-9226    

DOI:10.1039/a805425h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3947–3951 3947 The reactions of heterocyclic organotellurium and selenium compounds with triiron dodecacarbonyl Zulfiqar Majeed,a William R. McWhinnie,a Keith Paxton b and Thomas A. Hamor b a Chemical Engineering and Applied Chemistry, Aston University, Aston Triangle, Birmingham, UK B4 7ET b School of Chemistry, The University of Birmingham, Edgbaston, Birmingham, UK B15 2TT Received 28th July 1998, Accepted 2nd October 1998 The reaction of 2-telluraphthalide, C8H6OTe, with [Fe3(CO)12] gave as the major characterised species compound 1, [C6H4CH2TeFe(CO)3]2.An iron atom has inserted into the telluracyclic ring, and it is probable that one co-ordinated CO ligand arises from the initially organic carbonyl group. X-Ray single crystal investigations revealed a dimeric structure containing an Fe2Te2 core. The reaction of 2-telluraphthalic anhydride, C8H4O2Te, with [Fe3(CO)12] gave a known, but unexpected, organic product phthalide, C8H6O2.X-Ray crystallography confirmed its isolation; the main feature of interest is the significant double bond character of C(8)–O(1) at 1.352(4) Å. 2-Selenaphthalic anhydride, C8H4O2Se, gave intractable products on reaction with [Fe3(CO)12], but 2-selenaphthalide, C8H6OSe, on reaction with the carbonyl gave a major product 2, [Fe(CO)3{C6H4CH2SeFe(CO)3}] and a minor product 3, [Fe(CO)2{h6-C6H4CH2SeFe2(CO)6}] which is an intermediate in the formation of 2.Compound 2 was shown by X-ray methods to be very similar to 1 except that the 18 electron rule is satisfied by co-ordination of an Fe(CO)3 moiety, rather than by dimerisation. Compound 3, also studied by X-ray crystallography, diVers from 2 mainly in the addition of an h6-bonded Fe(CO)2 moiety, but the selenaferrole ring is more distorted. It is proposed that comparative studies of reactions of selenium and tellurium compounds with [Fe3(CO)12] may assist the development of an understanding of the complex reaction pathways.Introduction In three recent papers it has been shown that heterocyclic compounds containing only tellurium,1,2 or containing both tellurium and nitrogen,3 may act as precursors for some interesting new organometallic derivatives of iron1,3 or rhodium.2 That several, although not all, of the products obtained involve detelluration of the heterocyclic molecule probably reflects the relatively low carbon–chalcogen bond strengths in the compounds, compared with the sulfur analogues which are generally less reactive under similar conditions.1,4,5 Reactions are more likely to proceed to completion with the tellurium containing heterocycles; if, however the corresponding selenium containing materials were to be considered, it is possible that products corresponding to an earlier stage of the reaction sequence may be isolable.Angelici and co-workers 6 have extended their studies of thiophene–transition metal interactions to selenophenes, where, amongst other factors, the ability to measure 77Se NMR spectra was valuable (similar advantages accrued from our studies of tellurium heterocycles where 125Te NMR measurements have been valuable 7).In this brief paper we consider reactions of directly related selenium and tellurium heterocycles with [Fe3(CO)12]. Experimental and results 2-Tellura-2,3-dihydroinden-1-one (2-telluraphthalide),8 2- tellura-2,3-dihydroindene-1,3-dione (2-telluraphthalic anhydride), 9 2-selena-2,3-dihydroinden-1-one (2-selenaphthalide) 10 and 2-selena-2,3-dihydroindene-1,3-dione (2-selenaphthalic anhydride) 9 were prepared using literature methods.Triiron dodecarbonyl [Fe3(CO)12] was obtained from Aldrich and used as received. All manipulations involving reactions of tellurium and selenium compounds were carried out under an atmosphere of pure argon with the use of Schlenk techniques. The TLC was carried out on UV active silica gel plates.Reactions of triiron dodecarbonyl With 2-oxo-2,3-dihydrobenzotellurophene. 2-Telluraphthalide (1 g, 4 mmol) and [Fe3(CO)12] (2.05 g, 4 mmol) were refluxed, with stirring, in the dark, in toluene (25 cm3) for 4.5 h. The reaction mixture was cooled to room temperature and filtered to give a dark red filtrate and a residual black solid which adhered to the sides of the flask. The solvent was removed from the filtrate in vacuo to give a red solid. The red solid was recrystallised from hexane to give bright red needles, 1, which were dried under vacuum.The red crystals decomposed at temperatures greater than 220 8C {0.39 g, 24% based on [Fe3- (CO)12]} (Found: C, 33.7; H, 1.97. Calc. for C10H6FeO3Te: C, 33.5; H, 1.67%). FTIR (KBr, cm21): n(CO) 2049, 1980 and 1953. NMR (CDCl3): 1H, d 7.87, 7.30, 7.06 (C6H4), 4.30, 4.31, 3.99, 3.97 AB pattern (CH2); 13C, d 209.4, 205.3, 202.3 (CO), 153.0, 154.4, 147.4, 129.5, 127.3, 124.1 (C6H4) and 15.6 (CH2); 125Te, d (decoupled) 2729.7, JTeC 17.95 Hz, (coupled) 2727.9, 2728.1, 2728.4, 2728.6 (AB pattern, JTeH 50.27, 21.55 Hz).EI mass spectrum: m/z = 277, [C7H4FeTe]1; 91, [C7H7]1. Crystals suitable for X-ray diVraction measurements were grown by cooling a concentrated hexane solution. With 2-telluraphthalic anhydride. 2-Telluraphthalic anhydride (0.52 g, 2.0 mmol) and [Fe3(CO)12] (1 g, 2.0 mmol) were refluxed, with stirring, in the dark, with toluene (25 cm3) over 4.5 h.On cooling to room temperature the reaction mixture was filtered to give an orange-red filtrate and a residual black solid which adhered to the side of the flask. The solvent was removed from the filtrate in vacuo to give an orange solid. Hot hexane was used to extract the light orange solution, leaving a very3948 J. Chem. Soc., Dalton Trans., 1998, 3947–3951 small amount of a dirty white residue. The latter was discarded, and the hexane removed from the light orange solution giving a light brown (tan) product.The light brown solid was recrystallised from hexane to give light brown (tan) needle like crystals of phthalide which were dried under vacuum, mp 2 70–71 8C (lit., 72–74 8C) {0.11 g, 41% based on [Fe3(CO)12]} (Found: C, 71.4; H, 4.55. Calc. for C8H6O2: C, 71.6; H, 4.47%). FTIR (KBr, cm21): n(CO) 1753. NMR (CDCl3): 13C, d 170.5, 145.8, 133.3, 128.3, 125.0, 121.4 (C6H4) and 68.9 (CH2). EI mass spectrum: m/z = 134, [C8H6O2]1; 106, [C7H6O]1.Crystals suitable for X-ray diVraction measurements were grown by cooling a concentrated acetone solution. 2-Telluraphthalic anhydride (0.52 g, 2.0 mmol) in toluene (25 cm3) was refluxed in the dark for 4.5 h. The work-up of the reaction mixture was carried out as above. Only the starting materials were recovered. With 2-selenaphthalide. 2-Selenaphthalide (0.79 g, 4 mmol) and [Fe3(CO)12] (2.05 g, 4 mmol) were refluxed, with stirring, in the dark, with toluene (25 cm3) over 4.5 h.On cooling to room temperature the reaction mixture was filtered to give a deep red filtrate and a residual black solid which adhered to the sides of the flask. The solvent was removed from the filtrate in vacuo to give a deep red oil. This was chromatographed on a column of silica gel (pore diameter ca. 6 nm) giving a red band, followed by another smaller red band and finally a pale yellow band. Elution with chloroform–hexane (1 : 1) and removal of the solvent gave a deep red solid, 2, from the first eluate, a black solid, 3, from the second eluate, and pale orange crystals from the third eluate (unchanged 2-selenaphthalide).The deep red solid, 2, recrystallised from boiling hexane yielding red crystals; mp 190–192 8C (decomp.) {0.90 g, 50% based on [Fe3(CO)12]} (Found: C, 34.6; H, 1.56. Calc. for C13H6Fe2O6Se: C, 34.7; H, 1.34%). FTIR (KBr, cm21): n(CO) 2062, 2021 and 1962. NMR (CDCl3): 1H, d 8.19, 7.35, 6.93 (C6H4), 3.78 (CH2); 13C, d 209.5 (CO), 159.2, 139.7, 130.5, 127.8, 123.5, 112.4 (C6H4) and 35.5 (CH2).EI mass spectrum: m/z = 450, [C13H6O6Fe2Se]1: 394, [C11H6O4Fe2Se]1. The black solid, 3, recrystallised from boiling hexane yielding black crystals, mp 159–160 8C (decomp.) {0.048 g, 2% based on [Fe3(CO)12]} (Found: C, 33.02; H, 1.80. Calc. for C15H6Fe3O8Se: C, 32.13; H, 1.08%). FTIR (KBr, cm21): n(CO) 2106, 2061, 2024, 1983 and 1956. NMR (CDCl3): 1H, d 7.22 (br, C6H4), 2.15 (br, CH2); 13C, d 213.0, 212.0, 211.2, 209.5 (CO), 132.8, 130.2, 128.2, 125.4, 111.7 (C6H4) and 34.7 (CH2).EI mass spectrum: m/z = 562, [C15H6O8Fe3Se]1; 506, [C13H6O6Fe3Se]1; 450, [C13H6- O6Fe2Se]1: 422, [C12H6O5Fe2Se]1. Crystals suitable for X-ray diVraction measurements were grown by cooling a concentrated chloroform–hexane (1 : 1) solution for compound 2 and a solution of chloroform for 3. With 2-selenaphthalic anhydride. 2-Selenaphthalic anhydride (0.49 g, 2.3 mmol) and [Fe3(CO)12] (1.15 g, 2.3 mmol) were refluxed, with stirring, in the dark, with toluene (25 cm3) over 4.5 h.On cooling to room temperature the reaction mixture was filtered to give a deep red filtrate and a residual black solid which adhered to the sides of the flask. The solvent was removed from the filtrate in vacuo to give a red solid. This was chromatographed on a column of silica gel (pore diameter ca. 6 nm) giving a red band followed by a colourless band. Elution with chloroform–hexane (2 : 1) and removal of the solvent gave a red solid from the first eluate and pale yellow crystals from the second, which were shown to be unchanged 2-selenaphthalic anhydride.The first component has to date defied attempts to characterise it (Found: C, 27.73; H, 2.40%): FTIR (KBr, cm21): n(CO) 2072 and 1958 cm21. X-Ray crystallography The crystal structures of compounds 1–3 and phthalide were determined; crystal parameters and experimental data are listed in Table 1. Cell dimensions and intensity data for all four structures were measured on a Rigaku R-AXIS II area detector diVractometer at 293(2) K using graphite-monochromated Mo- Ka radiation, l = 0.7107 Å.Conventional absorption corrections were not applied since, on average, each unique reflection intensity is the mean of three intensities measured at diVerent orientations of the crystal, thus minimising absorption eVects. The structures were determined11 by direct methods and refined12 by least squares on F2 using anisotropic thermal parameters for non-hydrogen atoms.Hydrogen atoms were placed in calculated positions, riding on their respective bonding atoms. Diagrams were drawn with ORTEP;13 thermal ellipsoids are at the 30% probability level. Selected bond lengths and angles are in Table 2. CCDC reference number 186/1186. See http://www.rsc.org/suppdata/dt/1998/3947/ for crystallographic files in .cif format. Physical measurements Infrared spectra were obtained for KBr discs with a Bio-Rad FTS-40A FTIR spectrometer, proton (300.133), 13C (75.469), 125Te (78.580 MHz) NMR spectra with a Bruker AC 300 spectrometer at the indicated frequencies; references used were TMS (1H, 13C) and Me2Te (125Te).The electron impact (EI) and fast atom bombardment (FAB) mass spectra were obtained via the EPSRC mass spectrometry service, University College, Swansea. Discussion The reaction of 2-telluraphthalide with [Fe3(CO)12] gave compound 1 (Scheme 1). It is likely that an initial stage of the reaction involved insertion of iron into the Te–C(O) bond, and that this was followed by what is eVectively the reverse of a carbonyl insertion reaction.Thus one of the carbonyl groups on iron is probably of “organic” origin. The monomeric unit (with respect to iron and tellurium) does not satisfy the 18 electron rule, hence dimerisation occurs by co-ordination of tellurium (Lewis base) to the iron in the neighbouring isostructural fragment. The room temperature NMR data (1H, 13C) are entirely consistent with the structure determined by X-ray crystallography (see below).The reaction of [Fe3(CO)12] with 2-telluraphthalic anhydride gives an unexpected product (phthalide), confirmed by X-ray crystallography; it is a case of the mode of origin of the material being of more interest than the actual product. 2-Telluraphthalic anhydride was pure and correctly characterised (mp, FTIR, NMR, C,H,N analysis,9a and X-ray crystallographic analysis 9b), thus the origin of the product cannot be attributed to impurities in the starting materials; the yield would also suggest that this explanation is precluded.In further experiments the 2-telluraphthalic anhydride was recovered unchanged from refluxing toluene; nor did [Fe3(CO)12] react with toluene under the reaction conditions used. This renders it improbable that the carbon skeleton originates from toluene. It is diYcult to explain the origin of the product but the postulation of an initial insertion of iron into a Te–C(O) bond is reasonable.It is apparent that removal of Te requires a multimetal centre, hence at least a second iron is expected to be present at this stage of the reaction. Rear attack on one carbonyl by the oxygen of the second carbonyl occurring synchronously with elimination of FeTe is then possible. The resulting carbene may be temporally stabilised by co-ordination to the second iron (a recent example of an iron stabilised carbene has been reported 14); decomposition by adventitious moisture is possible, although a secondary alcohol may be expected if this were so.Alternatively, proton extraction from the solventJ. Chem. Soc., Dalton Trans., 1998, 3947–3951 3949 may occur. Further speculation is not merited at this juncture. Reaction of 2-selenaphthalide with [Fe3(CO)12] yields compounds 2 and 3. Their relative yields suggest that 3 (2%) is an intermediate in the formation of 2 (50%). Initial insertion of iron into a Se–C(O) bond, followed by the loss of carbonyl from the iron as shown in Scheme 1 can be postulated, giving product 3. On prolonged heating of compound 3 the loss of the h6- bound Fe(CO)2 moiety must occur giving 2.It is of interest that the ring iron atom in 2 satisfies the 18 electron rule by retention of a Fe(CO)3 unit, whereas in the related compound 1 loss of this moiety and co-ordination to tellurium of a neighbouring fragment is preferred.The use of the selenium analogue has then given materials which can very reasonably be considered models for intermediate stages of what is doubtless a complex reaction sequence, and to that extent the objectives of the study have been achieved. The room temperature NMR data (1H, 13C) for both 2 and 3 appear too simple to be consistent with the static structures determined by X-ray methods (see below). Thus, for 2, the CH2 protons give a sharp singlet rather than the AB pattern demanded by the static structure; also the 13C data imply the equivalence of the six carbonyl ligands.A rapid interchange of the two iron atoms rendering the methylene protons equivalent on the NMR timescale is probable. Cooling a CDCl3 solution of the complex incrementally to 218 K produces significant broadening of the CH2 resonance, but even at this temperature complete resolution into the AB pattern is not observed. It is clear that 3 must also be fluxional in solution at room temperature; only four resonances for the carbonyl ligands are observed and a singlet is again seen for the CH2 protons, although in this case the resonance is broad at room temperature and sharpens on heating to 323 K.Under these Scheme 1 Reaction of organotellurium and organoselenium compounds with triiron dodecacarbonyl. circumstances, it is surprising that cooling to 233 K, although inducing further broadening of the methylene signal, fails to achieve resolution into the expected AB pattern..The fluxional behaviour of 3 is complex and must certainly involve, at the least, interchange of the two Fe(CO)3 groups. The reaction of the 2-selenaphthalic anhydride with [Fe3- (CO)12] aVorded only unchanged starting material and an intractable red product (see Experimental section). Only the starting materials were recovered when the experiments involving the four organo-selenium and -tellurium compounds were repeated with triruthenium dodecarbonyl [Ru3(CO)12].X-Ray crystallography The dimeric complex 1 has crystallographic 1 (Ci) symmetry. The central Fe2Te2 core is approximately square-shaped, with iron–tellurium distances 2.572(1) and 2.575(1) Å, angle at iron 83.22(4)8, angle at tellurium, 96.78(4)8 (see Fig. 1 and Table 2). The Fe–Te bond lengths fall near the lower end of the range, 2.54–2.67 Å, found3,15 in a number of other complexes containing analogous atomic groupings.The co-ordination geometry at iron is approximately octahedral, maximum angular deviation 15.38, mean deviation 5.58. The tellurium–carbon bond length, Te(1)–C(1) 2.139(9) Å, is slightly shorter than the mean value given16 for Te–C(sp3) bonds, 2.158 Å, but falls within the range 2.119–2.154 Å found17 in a number of more recent structure determinations. The eight-atom grouping C(1)–C(7), Fe(1) is coplanar to within 0.03 Å (root mean square, r.m.s. deviation 0.019 Å). The tellurium atom which completes a benzotelluraferrole ring system is, however, displaced by 0.83(1) Å.The plane of the Fe2Te2 core is oriented at 80.4(1)8 to this plane, forming a step to the centrosymmetrically related benzotelluraferrole ring, so that the perpendicular distance between these parallel planes is approximately 3.34 Å. The non-hydrogen atoms of phthalide are essentially coplanar, r.m.s. atomic deviation 0.014 Å (Fig. 2). The maximum deviation of 0.026(2) Å is that of the carbonyl oxygen atom, O(2), and omitting this atom from the calculations gives a significantly better plane, r.m.s.deviation 0.009 Å, with O(2) displaced by 0.049(5) Å. Noteworthy is the diVerence in bond lengths between those at the saturated carbon atom, C(1), and those at the trigonally hybridised C(8). Thus, while O(1)–C(1) at 1.455(4) Å approximates to a C–O single bond, the O(1)– C(8) length of 1.352(4) Å indicates significant double bond character, and, also, C(1)–C(2) is significantly longer than C(7)–C(8).The O(2)]] C(8) formal double bond, 1.208(4) Å, is similar in length to the corresponding bonds in phthalic anhydride, 1.192(4) Å.18 Fig. 1 View of the crystal structure of complex 1. Starred atoms are related to the corresponding unstarred atoms by an inversion centre.3950 J. Chem. Soc., Dalton Trans., 1998, 3947–3951 Table 1 Crystallographic data for compounds 1–3 and phthalide Formula M Crystal system Space group a/Å b/Å c/Å b/8 U/Å3 Z m(Mo-Ka)/mm21 Reflections collected [I > s(I)] Unique reflections Rint R, wR2 Observed reflections [I > 2s(I)] R, wR2 1 C20H12Fe2O6Te2 715.2 Monoclinic P21/n 10.863(2) 7.245(2) 14.380(3) 105.49(2) 1090.6(4) 2 3.985 5937 1888 0.0583 0.0579, 0.1060 1781 0.0533, 0.1042 phthalide C8H6O2 134.1 Monoclinic P21/c 7.760(2) 10.799(3) 8.234(3) 112.88(2) 635.7(3) 4 0.101 2740 966 0.0598 0.0986, 0.2246 772 0.0739, 0.1944 2 C13H6Fe2O6Se 448.8 Monoclinic P21/n 9.479(2) 15.339(3) 10.637(2) 97.73(1) 1532.5(5) 4 4.290 8792 2680 0.0929 0.0741, 0.1422 2190 0.0564, 0.1307 3 C15H6Fe3O8Se 560.7 Monoclinic P21/n 8.264(1) 15.591(2) 13.628(2) 97.34(1) 1741.5(4) 4 4.597 10382 2899 0.0478 0.040, 0.101 2733 0.037, 0.098 In the benzoselenaferrole moiety of complex 2, the two iron atoms appear to be in chemically identical environments (Fig. 3). However, stereochemically they are not identical; Fe(1) lies close [0.18(1) Å] to the plane of the organic residue and may, therefore, be considered to be the iron constituent of the benzoselenaferrole, whereas Fe(2) is displaced by 1.99 Å from this plane and may be considered to be p-bonded to the selenium atom, Fe(1) and C(7).Bond lengths are consistent with this, Fe(1)–C(7) and Fe(1)–Se(1) each being shorter than the corresponding bonds involving Fe(2) (see Table 2). The selenium atom is displaced by 0.72 Å from the organic plane, on the same Fig. 2 View of the crystal structure of phthalide. Fig. 3 View of the crystal structure of complex 2. side as Fe(2). The crystal structure of the analogous m-[1,2-hselanylcyclohex- 1-ene-1-carbaldehydato(2 2)-m-Se]-bis(tricarbonyliron) has been determined.19 Here the out-of-plane iron atom is bonded to the selenium atom and both unsaturated carbon atoms of the cyclohexene at distances of 2.353(1), 2.214(8) and 2.132(7) Å, respectively, while the in-plane iron is bonded to selenium at 2.327(1) Å and to the carbonyl carbon atom of the selenaferrole ring at 1.990(8) Å.The Fe–Fe distance of 2.631(2) Å is greater than our distance of 2.482(1) Å which is, however, similar to such distances generally found in Fe2- (CO)6 residues.19 For comparison, in the complex [Cp*(OC)2- Re{m-h6-SeC4H4Fe(CO)3}Fe(CO)3] the Fe2Se triangle has Fe–Se and Fe–Fe distances 2.357, 2.367 and 2.558 Å,6a similar to the values cited above and those measured in complex 3 (see below), although the distinction between the s- and p-bonded iron atoms is less obvious.Complex 3 diVers from 2 only by the addition of an h6- bonded Fe(CO)2 moiety. The selenaferrole ring is more distorted than in 2, with the “in-plane” iron atom, Fe(1), displaced by 0.407(5) Å from the plane of atoms C(1)–C(7) (Fig. 4). The out-of-plane Fe(2) and the selenium atom are displaced by 2.05 and 0.70 Å, respectively, on the opposite side of the C(1)–C(7) plane, so that the selenaferrole ring [C(1), C(2), C(7), Fe(1), Se(1)] has a half-chair conformation The Fe–Fe, Fe–Se, Fe–C and Se–C bond distances are similar to those in 2, apart from Fe(2)–C(7), which is longer by 0.26 Å here, and Fe(1)–Fe(2), which is longer by 0.07 Å.The increase in the Fe(2)–C(7) bond Fig. 4 View of the crystal structure of complex 3.J. Chem. Soc., Dalton Trans., 1998, 3947–3951 3951 length may be a consequence of the presence of the h6-bonded Fe(3), which is situated 2.836(1) Å from Fe(2) and 2.226(4) Å from C(7). In a similar situation, an Fe–Fe distance of 2.822 Å has been considered 20 to indicate a metal–metal bond.The Fe(3)–C (phenyl) distances are similar to those measured 20 previously; apart from the Fe(3)–C(7) distance, they are all in the range 2.12–2.14 Å. Table 2 Selected bond lengths (Å) and angles (8) for compounds 1–3 and phthalide Compound 1 Te(1)–Fe(1) Te(1)–Fe(1*) Te(1)–C(1) Fe(1)–C(7) Fe(1)–Te(1)–Fe(1*) Fe(1)–Te(1)–C(1) Fe(1*)–Te(1)–C(1) Te(1)–Fe(1)–C(7) Te(1)–Fe(1)–C(8) Te(1)–Fe(1)–C(9) Te(1)–Fe(1)–C(10) Te(1)–Fe(1)–Te(1*) C(7)–Fe(1)–Te(1*) 2.575(1) 2.572(1) 2.139(9) 2.062(8) 96.78(4) 88.9(2) 106.0(3) 85.6(2) 91.6(3) 89.3(3) 172.0(3) 83.22(4) 80.7(2) Fe(1)–C(8) Fe(1)–C(9) Fe(1)–C(10) C(7)–Fe(1)–C(8) C(7)–Fe(1)–C(9) C(7)–Fe(1)–C(10) C(8)–Fe(1)–Te(1*) C(8)–Fe(1)–C(9) C(8)–Fe(1)–C(10) C(9)–Fe(1)–Te(1*) C(9)–Fe(1)–C(10) C(10)–Fe(1)–Te(1*) 1.767(10) 1.806(10) 1.777(9) 84.5(3) 174.9(4) 92.0(4) 164.7(3) 96.3(4) 95.8(4) 98.1(3) 93.0(4) 88.9(3) Phthalide O(1)–C(1) O(1)–C(8) O(2)–C(8) C(1)–O(1)–C(8) O(1)–C(1)–C(2) O(1)–C(8)–C(7) 1.455(4) 1.352(4) 1.208(4) 110.7(3) 104.2(3) 108.1(3) C(1)–C(2) C(7)–C(8) O(2)–C(8)–O(1) O(2)–C(8)–C(7) 1.491(5) 1.459(5) 121.4(4) 130.5(4) Compound 2 Se(1)–Fe(1) Se(1)–Fe(2) Se(1)–C(1) Fe(1)–Fe(2) Fe(1)–C(7) Fe(1)–C(8) Fe(1)–Se(1)–C(1) Fe(2)–Se(1)–C(1) Fe(1)–Se(1)–Fe(2) Se(1)–Fe(1)–C(7) Fe(2)–Fe(1)–C(7) 2.347(1) 2.378(1) 1.967(6) 2.482(1) 2.046(5) 1.774(6) 96.3(2) 90.6(2) 63.36(3) 82.1(1) 57.0(1) Fe(1)–C(9) Fe(1)–C(10) Fe(2)–C(7) Fe(2)–C(11) Fe(2)–C(12) Fe(2)–C(13) Se(1)–Fe(1)–Fe(2) C(7)–Fe(2)–Se(1) C(7)–Fe(2)–Fe(1) Se(1)–Fe(2)–Fe(1) 1.781(7) 1.783(6) 2.194(5) 1.789(5) 1.757(7) 1.798(6) 58.93(3) 78.4(1) 51.5(1) 57.70(3) Compound 3 Se(1)–Fe(1) Se(1)–Fe(2) Se(1)–C(1) Fe(1)–Fe(2) Fe(1)–C(7) Fe(1)–C(8) Fe(1)–C(9) Fe(1)–C(10) Fe(2)–C(7) Fe(2)–C(11) Fe(2)–C(12) Fe(1)–Se(1)–C(1) Fe(2)–Se(1)–C(1) Fe(1)–Se(1)–Fe(2) Se(1)–Fe(1)–C(7) Fe(2)–Fe(1)–C(7) 2.355(1) 2.383(1) 1.945(4) 2.551(1) 2.001(4) 1.794(5) 1.811(5) 1.769(4) 2.452(4) 1.757(5) 1.764(4) 95.2(1) 99.6(1) 65.15(2) 81.6(1) 63.8(1) Fe(2)–C(13) Fe(3)–C(2) Fe(3)–C(3) Fe(3)–C(4) Fe(2)–C(5) Fe(3)–C(6) Fe(3)–C(7) Fe(3)–C(14) Fe(3)–C(15) Fe(3) ? ? ?Fe(2) Se(1)–Fe(1)–Fe(2) C(7)–Fe(2)–Se(1) C(7)–Fe(2)–Fe(1) Se(1)–Fe(2)–Fe(1) 1.791(5) 2.140(4) 2.122(4) 2.125(4) 2.129(4) 2.124(4) 2.226(4) 1.765(5) 1.762(4) 2.836(1) 57.96(2) 72.5(1) 47.1(1) 56.90(2) In all three iron carbonyl complexes the Fe–C–O moieties are near linear, angular ranges 175.8–178.98 in 1, 177.7–179.48 in 2 and 175.1–179.68 in 3.The Fe–CO bond lengths are normal, 1.77–1.81 Å in 1, 1.76–1.80 Å in 2 and 1.76–1.81 Å in 3. There is some evidence that Fe–CO bonds in Fe(CO)2 moieties, mean lengths 1.763(2),20 1.736(18) 21 and 1.763(2) Å in 3 are shorter than those in Fe(CO)3 moieties, mean lengths 1.795(12),19 1.791(4),20 1.783(12) in 1, 1.780(6) in 2 and 1.781(9) Å in 3. Acknowledgements Z. M. thanks the EPSRC for a studentship.The EPSRC and the University of Birmingham are thanked for funds with which to purchase the X-ray diVractometer. The authors are also grateful to the EPSRC for access to the Mass Spectroscopy Service at University College, Swansea. References 1 K. Singh, W. R. McWhinnie, H. L. Chen, M. Sun and T. A. Hamor, J. Chem. Soc., Dalton Trans, 1996, 1545. 2 K. Badyal, W. R. McWhinnie, H. L. Chen and T. A. Hamor, J. Chem. Soc., Dalton Trans., 1997, 1579. 3 K. Badyal, W. R. McWhinnie, T.A. Hamor and H. Chen, Organometallics, 1997, 16, 3194. 4 T. B. Rauchfuss, Prog. Inorg. Chem., 1991, 39, 259. 5 R. J. Angelici, Coord. Chem. Rev., 1990, 105, 61. 6 (a) M. G. Choi and R. J. Angelici, J Am. Chem. Soc., 1991, 113, 5651; (b) C. J. White, T. L. Wang, R. A. Jacobsen and R. J. Angelici, Organometallics, 1994, 13, 4474; (c) C. J. White and R. J. Angelici, Organometallics, 1994, 13, 5132. 7 K. Badyal, W. R. McWhinnie. J. Homer and M. C. Perry, J. Organomet. Chem., 1998, in the press. 8 L. Engman and M. P. Cava, J. Org. Chem., 1981, 46, 4194. 9 (a) J. Bergman and L. Engman, Org. Prep. Proced. Int., 1978, 10, 289; (b) Z. Majeed, W. R. McWhinnie, K. Paxton and T. A. Hamer, J. Organomet. Chem., in the press. 10 W. H. H. Gunther, J. Org. Chem., 1967, 32, 3929. 11 TEXSAN, Single Crystal Structure Analysis Software, version 1.6, Molecular Structure Corporation, Houston, TX, 1993. 12 G. M. Sheldrick, SHELXL 93, Program for Crystal Structure Refinement, University of Göttingen, 1993. 13 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 14 A. Klose, E. Solari, C. Floriani, N. Re, A. Chiesivila and C. Rizzoli, Chem. Commun., 1997, 2297. 15 P. Mathur, D. Chakrabarty, M. M. Hossain, R. S. Rashid, V. Rugmini and A. L. Rheingold, Inorg. Chem., 1992, 31, 1106; G. Gervasio, J. Organomet. Chem., 1992, 441, 271; L. C. Roof, D. M. Smith, G. W. Drake, W. T. Pennington and J. W. Kolis, Inorg. Chem., 1995, 34, 337; J. R. Eveland and K. H. Whitmire, Angew. Chem., Int. Ed. Engl., 1996, 35, 741. 16 F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen and R. Taylor, J. Chem. Soc., Perkin Trans., 1987, S1. 17 K. Y. Abid, N. I. Al-Salim, M. Greaves, W. R. McWhinnie, A. A. West and T. A. Hamor, J. Chem. Soc., Dalton Trans., 1989, 1697; H. B. Singh, N. Sudha, A. A. West and T. A. Hamor, J. Chem. Soc., Dalton Trans., 1990, 907. 18 R. B. Bates and R. S. Cutler, Acta Crystallogr., Sect. B, 1977, 33, 893. 19 R. C. Petterson, K. H. Pannell and A. J. Mayr, Acta Crystallogr., Sect. B, 1980, 36, 2434. 20 N. S. Nametkin, V. D. Tyurin, A. I. Nekhaev, Yu. P. Sobolev, M. G. Kondrateva, A. S. Batsanov and Yu. T. Struchkov, J. Organomet. Chem., 1983, 243, 323. 21 J. Chen, J. Yin, Z. Fan and W. Xu, J. Chem. Soc., Dalton Trans., 1988, 2803. Paper 8/05881D

ISSN:1477-9226    

DOI:10.1039/a805881d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 3953–3960 3953 Synthesis, structures and magnetochemistry of binuclear cobalt(II), nickel(II) and copper(II) complexes of 2,6-diformyl-4-methylphenol dioxime Daniel Black,a Alexander J. Blake,a Keith P. Dancey,b Andrew Harrison,c Mary McPartlin,b Simon Parsons,c Peter A. Tasker,*c,d Gavin Whittaker c and Martin Schröder *a a School of Chemistry, The University of Nottingham, University Park, Nottingham, UK NG7 2RD b School of Chemistry, University of North London, London, UK N7 8DB c Department of Chemistry, The University of Edinburgh, West Mains Road, Edinburgh, UK EH9 3JJ d Zeneca Specialities Plc, Hexagon House, Blackley, Manchester, UK M9 3DA Received 5th May 1998, Accepted 8th September 1998 Reaction of 2,6-diformyl- and 2,6-diacetyl-4-methylphenol with a large excess of both NH2OH?HCl and CH3CO2K in EtOH aVords high yields of 2,6-diformyl-4-methylphenol dioxime (2-hydroxy-5-methylbenzenedicarbaldehyde dioxime) (H3L1) and 2,6-diacetyl-4-methylphenol dioxime (H3L2), respectively.The crystal structure of (H3L2) shows intramolecular hydrogen bonding with long-range intermolecular p-stacking interactions and an extended intermolecular hydrogen-bonding network. The binuclear complexes of CoII, NiII and CuII—[Co2(H2L1)2(MeOH)2- (H2O)2]Cl2?2MeOH, [Ni2(H2L1)2(H2O)4][ClO4]2?2H2O and [Cu2(H2L1)2(ClO4)2], respectively—derived from the dioxime ligand (H3L1) have been synthesized and characterised and their single-crystal structures determined.The structure of [Co2(H2L1)2(MeOH)2(H2O)2]21 shows each high-spin CoII to be six-co-ordinate and bound to an N2O4- donor array presented by two dioxime ligands and axially co-ordinated H2O and MeOH molecules, the dioxime ligands co-ordinating via the imino N- and phenoxy O-donors. The structure of [Ni2(H2L1)2(H2O)4]21 shows two octahedrally co-ordinated NiII each with an N2O4 donor set similar to that in [Co2(H2L1)2(MeOH)2(H2O)2]21 except that the co-ordination sphere of each NiII is completed by axial ligation to two H2O molecules.The structure of [Cu2(H2L1)2(ClO4)2] confirms N2O4 donation at CuII with two bidentate ClO4 2 anions, Cu ? ? ? O 2.51(2), 2.76(2) Å, interacting with the metal centres on either side of the planar oxime–phenolate array. In all three complexes the two dioxime ligands are monodeprotonated at the phenolic oxygen, and the oximes are linked by hydrogen bonds, which results in a pseudo-macrocyclic framework.Magnetic susceptibility measurements on the complexes over the range 2.5–340 K confirm that the complexes are antiferromagnetically coupled with values for the magnetic exchange constant J of 26.9 ± 0.1, 216.0 ± 0.6, and 2452 ± 4 cm21 for [Co2(H2L1)2(MeOH)2(H2O)2]Cl2, [Ni2(H2L1)2(H2O)4][ClO4]2 and [Cu2(H2L1)2(ClO4)2], respectively. Introduction Since the mid-1960s ortho-hydroxyoximes have been utilised commercially for the solvent extraction of copper from aqueous solutions of acid-soluble copper ores.1 Such reagents bind to copper(II) ions to give a neutral 2 : 1 complex from which the metal is liberated by back extraction into a strongly acidic aqueous phase.2 As part of an investigation into the solvent extractant abilities of analogous binucleating dioxime ligands we have synthesized 2,6-diformyl-4-methylphenol dioxime (2-hydroxy-5-methylbenzenedicarbaldehyde dioxime) (H3L1) and studied its reaction with various salts of CoII, NiII and CuII.Formation of binuclear transition complexes with oxime ligands has been observed previously, notably with AgI,3 CuII 3–5 and NiII.6 There have been relatively few reports dealing with the co-ordination chemistry of 2,6-diformyl-4-methylphenol dioxime and its derivatives.7–10 In 1973 Okawa et al.7 reported the reaction of 2,6-diformyl-4-methylphenol with NH2OH in the presence of Cu(O2CMe)2?H2O and NiCl2?6H2O. In both cases insoluble complexes were obtained which, on the basis of microanalytical, magnetic susceptibility, spectroscopic and mass spectral data, were assigned the formulations [Cu2(HL1)2] and [Ni2(HL1)2]?3H2O.Very recently, Thompson and co-workers 9 have reported magnetochemical and structural data on related nickel(II) oxime complexes, while Busch and coworkers 10 have prepared asymmetric iminooxime compartmental species. In addition, Chaudhuri and co-workers 8 have reported the reaction of H3L1 with Ni(O2CMe)2?4H2O and [FeLCl3] (L = 1,4,7-trimethyl-1,4,7-triazacyclononane) in basic MeOH solution. After treatment with Bu4NBF4 a solid of formulation [Fe2Ni2(m-O2CMe)2L2(L1)2(MeOH)2][BF4]2 was isolated and characterised by X-ray crystallographic and magnetic susceptibility measurements.This tetranuclear complex exhibits a nearly linear FeIIINiIINiIIFeIII core in which each metal centre adopts an octahedral co-ordination geometry. We report herein the synthesis and structures of oxime ligands derived from 2,6-diacetyl- and 2,6-diformyl-4-methylphenol, and the synthesis, structures and magnetochemistry of their complexes of CoII, NiII and CuII.OH N N R R HO OH R = H: H3L1 H3L2 R = Me:3954 J. Chem. Soc., Dalton Trans., 1998, 3953–3960 Results and discussion Reaction of either 2,6-diformyl-4-methylphenol or 2,6-diacetyl- 4-methylphenol with a large excess of both NH2OH?HCl and CH3CO2K in EtOH aVords, after aqueous work-up, high yields of 2,6-diformyl-4-methylphenol dioxime (H3L1) and 2,6- diacetyl-4-methylphenol dioxime (H3L2), respectively.The IR spectrum of H3L1 shows sharp bands at 1623 and 1604 cm21 assigned to the C]] N stretching vibration, while H3L2 shows the corresponding absorption at 1647 cm21. The mass spectra of the compounds show molecular ion peaks at m/z 194 and 223, corresponding to [H3L1]1 and [H3L2 1 H]1 respectively, while 1H NMR spectroscopy of H3L1 in (CD3)2CO shows single resonances at d 2.27, 7.37 and 8.38 assigned to the CH3, aromatic CH and imine protons respectively.Additional single resonances are observed at d 10.48 (phenolic OH) and 10.71 (oxime OH). The 13C DEPT NMR spectrum shows a methyl carbon resonance at d 18.54 as well as resonances at d 117.94 and 127.64 due to quaternary aromatic carbon centres. Resonances at d 128.68 (aromatic CH), 146.71 (imino C) and 152.29 (COH) are also observed. These data indicate either the presence in solution of one symmetrical isomer, or a mixture of rapidly interconverting isomers.Whilst the interconversion of oxime isomers in solution has been reported,11 steric factors would perhaps be expected to favour the formation of a E con- figuration at each C]] N bond. It has not proven possible to grow crystals of H3L1 of suitable quality for crystallographic studies and so the solid-state structure of the ligand has not been established. The 1H NMR spectrum of H3L2 in CD3OD shows three single peaks.Two distinct methyl proton resonances are observed at d 2.26 and 2.29 assigned to aliphatic and ring methyl groups respectively, while the aromatic proton resonance appears at d 7.18. No resonances are seen for the phenolic and oxime protons because of rapid exchange with the solvent. The 13C DEPT NMR spectrum shows methyl carbon resonances at d 11.79 and 19.22, quaternary aromatic carbon resonances at d 122.50 and 127.30, and an aromatic CH resonance at d 129.32.The spectrum is completed by the COH and imino C resonances at d 157.39 and 153.68. These spectra also indicate either the presence of one symmetrical isomer or a mixture of rapidly interconverting isomers in solution. Pale yellow crystals of H3L2 of diVraction quality were obtained by slow evaporation from a (CD3)2CO solution of the compound. A single-crystal structure determination was undertaken to establish which isomer of H3L2 is present in the solid state. The molecular structure confirms [Fig. 1(a), Table 1] that each C]] N double bond has a E configuration with the oxime hydroxyl groups and the aromatic ring on opposite sides of the double bond. An intramolecular hydrogen bond is observed within each dioxime molecule between the phenolic H atom and an N-donor of the oxime group. The crystal structure reveals that the dioxime molecules associate through extensive intermolecular hydrogen bonding [Fig. 1(b)] to produce extended molecular chains. This contrasts with the structure of salicylaldoxime 12 in which only hydrogen-bonded dimers are formed.Association between adjacent chains occurs via longrange p-stacking interactions, the separation between adjacent ring centroids being ca. 3.7 Å. Metal complexation Cobalt. Treatment of H3L1 with 1 equivalent of CoCl2?6H2O in MeOH aVords an orange solution from which an orange solid can be isolated. The FAB mass spectrum of this solid shows a peak at m/z 503, corresponding to [59Co2(H2L1)2 2 H]1 with the correct isotopic distribution. The IR spectrum shows absorption peaks at 1636 and 1617 cm21 assigned to the C]] N stretching vibration.In order to determine unambiguously the structure of the complex a single-crystal structure determination was undertaken. Suitable crystals were grown by slow evaporation from a MeOH solution of the complex. The structure of [Co2(H2L1)2(H2O)2(MeOH)2]Cl2?2MeOH shows (Fig. 2, Table 2) a centrosymmetric binuclear structure with each octahedral cobalt(II) centre bound to an N2O4-donor set from two dioxime ligands, Co–N 2.071(4), 2.073(4), Co–O 2.055(4), 2.075(4) Å, one axial H2O molecule, Co–O 2.102(4) Å, and one axial MeOH molecule, Co–O 2.133(4) Å.A crystallographic centre of inversion lies at the midpoint of the Co–Co vector, Co ? ? ? Co 3.092(2) Å. The two CoII lie in the plane defined by the four N- and two O-donor atoms of the dioxime ligands. The complex has a pseudo-macrocyclic structure which results from hydrogen bonding between the two dioxime ligands, O(2) ? ? ? O(3) 2.740(6) Å.Opposite pairs of oxime hydroxyl groups lie marginally above and below this plane as a result of accommodating this hydrogen bond. Nickel. The reaction of H3L1 with a slight excess of Ni(ClO4)2? 6H2O in thf–MeOH aVords a green solution. Crystals of diffraction quality were grown directly from this solution by slow evaporation at room temperature. The FAB mass spectrum of a sample of the crystalline material shows a peak at m/z 501, corresponding to [58Ni2(H2L1)2 2 H]1 with the correct isotopic Fig. 1 (a) View of the structure of H3L2 with numbering scheme adopted. H(1) ? ? ? N(2) 1.83 Å, O(1)–H(1)–N(2) 1458, O(1) ? ? ? N(2) 2.542(6) Å. (b) View of the packing diagram of H3L2. Table 1 Selected bond lengths (Å) and angles (8) with e.s.d.s for H3L2 O(1)–C(1) O(2)–N(2) O(6)–N(6) C(2)–C(20) N(2)–C(20) C(20)–N(2)–O(2) C(60)–N(6)–O(6) N(2)–C(20)–C(2) N(2)–C(20)–C(21) 1.368(4) 1.392(4) 1.407(4) 1.475(5) 1.288(5) 113.0(3) 111.1(3) 115.4(3) 123.4(4) C(20)–C(21) C(6)–C(60) N(6)–C(60) C(60)–C(61) C(4)–C(40) C(2)–C(20)–C(21) N(6)–C(60)–C(61) N(6)–C(60)–C(6) C(61)–C(60)–C(6) 1.492(5) 1.495(5) 1.287(5) 1.480(5) 1.519(5) 121.3(4) 125.3(4) 113.5(4) 121.1(4)J.Chem. Soc., Dalton Trans., 1998, 3953–3960 3955 distribution. The IR spectrum of the crystalline material shows absorption bands at 1648 and 1621 cm21 assigned to the C]] N stretching vibration.The presence of the ClO4 2 counter anion was confirmed by three very intense absorptions at 1145, 1112 and 1086 cm21. Electronic spectroscopy suggested that the product incorporates octahedral nickel(II) centres. To establish fully the nature of the complex a single-crystal structure determination was undertaken. The structure of [Ni2(H2L1)2(H2O)4][ClO4]2?2H2O shows (Fig. 3, Table 3) a centrosymmetric binuclear structure with both NiII centres having octahedral stereochemistries.A crystallographic centre of inversion lies at the midpoint of the Ni–Ni vector. Each NiII is bound to a N2O4 donor set from two dioxime ligands, Ni–N 2.012(2), 2.024(2), Ni–O 2.0167(13), 2.0206(14) Å, and to two axially co-ordinated water molecules, Fig. 2 View of the structure of [Co2(H2L1)2(H2O)2(MeOH)2]21 with numbering scheme adopted. The chloride counter anions and non-coordinating methanol molecules are omitted for clarity. Primed atoms are related to their unprimed equivalents by the symmetry operation (2x, 2y 1 1, 2z).Fig. 3 View of the structure of [Ni2(H2L1)2(H2O)4]21 with numbering scheme adopted. The perchlorate counter anions and non-coordinating water molecules are omitted for clarity. Primed atoms are related to their unprimed equivalents by the symmetry operation (2x 1 1, 2y, 2z 1 1). Table 2 Selected bond lengths (Å) and angles (8) with e.s.d.s for [Co2(H2L1)2(H2O)2(MeOH)2]Cl2?2MeOH Co(1)–O(1) Co(1)–N(1) Co(1)–N(2) Co(1)–O(19) O(1)–Co(1)–N(1) O(1)–Co(1)–N(2) N(1)–Co(1)–N(2) O(1)–Co(1)–O(19) N(1)–Co(1)–O(19) N(2)–Co(1)–O(19) O(1)–Co(1)–O(5) N(1)–Co(1)–O(5) 2.055(4) 2.071(4) 2.073(4) 2.075(4) 169.3(2) 87.3(2) 103.2(2) 83.1(2) 86.5(2) 170.3(2) 96.3(2) 86.5(2) Co(1)–O(5) Co(1)–O(49) Co(1) ? ? ? Co(19) N(2)–Co(1)–O(5) O(19)–Co(1)–O(5) O(1)–Co(1)–O(49) N(1)–Co(1)–O(49) N(2)–Co(1)–O(49) O(19)–Co(1)–O(49) O(5)–Co(1)–O(49) Co(1)–O(1)–Co(19) 2.102(4) 2.133(4) 3.092(2) 88.8(2) 91.6(2) 89.8(2) 88.2(2) 88.0(2) 92.7(2) 172.9(2) 96.9(2) Primed atoms are related to their unprimed equivalents by the symmetry operation (2x, 2y 1 1, 2z).Ni–O 2.086(2), 2.154(2) Å. These bond lengths are similar to those reported for the Ni2O2N4 core of [Fe2Ni2(m-O2CMe)2L2- (L1)2(MeOH)2][BF4]2.8 As in [Co2(H2L1)2(H2O)2(MeOH)2]Cl2? 2MeOH, the metal ions in [Ni2(H2L1)2(H2O)4]21 lie almost in the plane defined by the four N- and two O-donor atoms of the two dioxime ligands, Ni ? ? ? Ni 3.0495(8) Å.The opposing pairs of oxime hydroxyl groups lie slightly above and below this plane as a result of accommodating a hydrogen bond, O(20) ? ? ? O(60i) 2.634(3) Å, i 1 2 x, 2y, 1 2 z, between them. Again, this association results in the complex having a pseudomacrocyclic structure. Such linkages between oxime hydroxyl groups have been reported previously.5,13 In addition, there is a hydrogen bond between the two water molecules above each face of the complex. Reaction between H3L1 and other nickel(II) salts such as NiCl2?2H2O and Ni(O2CMe)2?4H2O aVords extremely insoluble pale green solids in all cases.The FAB mass spectral data indicate the presence of [Ni2(H2L1)2]21 units within these products. We speculate that the marked insolubility of these solids arises either from the presence of a very extensive hydrogenbonding network within the structure, or from stacking of the binuclear NiII 2 complexes such that each NiII achieves an octahedral co-ordination by axial ligation of hydroxyl oxygen atoms from adjacent oxime groups.These observations concerning the relative insolubility of these compounds are in accord with those reported by Okawa et al.7 Copper. We have found that reaction between H3L1 and a range of copper(II) salts aVords similarly insoluble solids. The FAB mass spectral data indicate the presence of binuclear [Cu2(H2L1)2]21 units within these materials although the precise nature of these products is unknown.One equivalent of H3L1 was dissolved in thf. The solution was frozen and then allowed to thaw by gradual warming to room temperature. A solution of 2 equivalents of Cu(ClO4)2? 6H2O in MeOH was added to the freshly thawed thf solution of H3L1. Stirring of the pale green solution overnight at room temperature led to the precipitation of a khaki-green solid which was collected, washed with diethyl ether and dried under suction. The IR spectrum of the product shows absorptions at 1612 and 1592 cm21 which are assigned to C]] N bond stretching vibrations, and also reveals the presence of ClO4 2 counter ions.However, the FAB mass spectrum of the material did not show any peaks which could be assigned conclusively. Crystals of this product were grown from thf–MeOH solution. A single crystal structure determination confirms (Fig. 4, Table 4) a binuclear species [Cu2(H2L1)2(ClO4)2] with both CuII centres displaying octahedral co-ordination geometries.The two dioxime ligands and the two CuII form an approximately planar array. A crystallographic two-fold axis passes through the methyl carbons and phenolate oxygens which results in there being only two independent Cu–N and two independent Table 3 Selected bond lengths (Å) and angles (8) with e.s.d.s for [Ni2(H2L1)2(H2O)4][ClO4]2?2H2O Ni(1)–N(20) Ni(1)–O(10) Ni(1)–O(109) Ni(1)–N(609) N(20)–Ni(1)–O(10) N(20)–Ni(1)–O(109) O(10)–Ni(1)–O(109) N(20)–Ni(1)–N(609) O(10)–Ni(1)–N(609) O(109)–Ni(1)–N(609) N(20)–Ni(1)–O(2) O(10)–Ni(1)–O(2) 2.012(2) 2.017(1) 2.021(1) 2.024(2) 88.70(6) 170.58(6) 81.88(6) 100.79(7) 170.49(6) 88.62(6) 89.26(6) 87.04(6) Ni(1)–O(2) Ni(1)–O(1) Ni(1) ? ? ? Ni(19) O(109)–Ni(1)–O(2) N(609)–Ni(1)–O(2) N(20)–Ni(1)–O(1) O(10)–Ni(1)–O(1) O(109)–Ni(1)–O(1) N(609)–Ni(1)–O(1) O(2)–Ni(1)–O(1) Ni(1)–O(10)–Ni(19) 2.086(2) 2.154(2) 3.0495(8) 90.14(6) 93.51(6) 92.07(6) 88.68(6) 87.85(6) 90.48(6) 175.49(5) 98.12(6) Primed atoms are related to their unprimed equivalents by the symmetry operation (2x 1 1, 2y, 2z 1 1).3956 J.Chem. Soc., Dalton Trans., 1998, 3953–3960 Cu–O distances (Table 4), Cu–N 1.95(2), 1.99(2) Å; Cu–O (phenoxy) 1.941(10), 1.965(12) Å, and which dictates that the Cu2O2 core is planar. The estimated standard deviations (e.s.d.s) are relatively high due to crystal decomposition during data collection and solvent disorder, but it is clear that these bond lengths are slightly shorter than those observed in the structures of [Ni2(H2L1)2(H2O)4][ClO4]2?2H2O and [Co2(H2L1)2(MeOH)2- (H2O)2]Cl2?2MeOH.The Cu ? ? ? Cu separation is 2.994(4) Å and the two Cu–O–Cu angles are 99.2(8) and 100.9(7)8. Two symmetry-related bidentate ClO4 2 anions bridge in an asymmetric manner between the two CuII, one above and one below the plane of the hydroxydioxime ligands, Cu–O (ClO4 2) 2.51(2), 2.76(2) Å. This co-ordination mode has been found for many related perchlorate complexes.14 The elongated axial interactions demonstrate the Jahn–Teller distortion expected for octahedrally co-ordinated CuII.Opposing pairs of hydroxyl groups lie fractionally above and below the plane of the two hydroxydioxime ligands and are linked via an intramolecular hydrogen bond, O(2A) ? ? ? O(2B) 2.59(3) Å. Two severely disordered thf solvent molecules accompany each complex. Magnetic measurements Nickel. The molar magnetic susceptibility of [Ni2(H2L1)2- (H2O)4][ClO4]2?2H2O was measured over the temperature range 2.5–340 K.The data are displayed in Fig. 5. The distinct cusp at about 40 K indicates significant antiferromagnetic exchange and the secondary rise in susceptibility at lower temperatures implies that there is a small amount of a paramagnetic impurity. One expects that the electronic ground state of NiII in an octahedral, or slightly distorted octahedral, ligand field has Fig. 4 View of the structure of [Cu2(H2L1)2(ClO4)2] with numbering scheme adopted. Non-co-ordinating thf molecules are omitted for clarity.Primed atoms are related to their unprimed equivalents by the symmetry operation (x, ��� 2 y, 1– 4 2 z). Table 4 Selected bond lengths (Å) and angles (8) with e.s.d.s for [Cu2(H2L1)2(ClO4)2]?2thf Cu–O(1A) Cu–N(1A) Cu–O(1B) N(1A)–Cu–N(1B) O(1A)–Cu–O(1B) N(1A)–Cu–O(1B) N(1B)–Cu–O(1B) 1.941(10) 1.95(2) 1.965(12) 100.2(7) 79.9(5) 167.9(6) 90.5(7) Cu–N(1B) Cu ? ? ? Cu9 N(1A)–Cu–O(1A) O(1A)–Cu–N(1B) Cu–O(1A)–Cu9 Cu–O(1B)–Cu9 1.99(2) 2.994(4) 90.5(6) 166.1(6) 100.9(7) 99.2(8) Primed atoms are related to their unprimed equivalents by the symmetry operation (x, ��� 2 y, 1– 4 2 z).electronic spin S = 1, and that the susceptibility c will be described well by the appropriate expression for an exchange coupled dimer with isotropic exchange. The data for [Ni2- (H2L1)2(H2O)4][ClO4]2?2H2O were fitted by least squares using eqn. (1) where c is defined per mol of NiII in the sample, r is the c = (1 2 r) C T F (J,T ) 1 r C9 T 1 cTIP (1) fraction of a paramagnetic nickel(II) impurity with Curie constant C9, and J is the intramolecular exchange constant for the binuclear complex, defined for the isotropic exchange Hamiltonian (2).For S = 1, C = Ng2b 2/k, and F(J,T) = (2e2x 1 H = 22J S1?S2 (2) 10e6x)/(1 1 3e2x 1 5e2x), where x = J/kT; cTIP represents any temperature independent paramagnetism, arising from the second-order Zeeman eVect, and is of the order of 240 × 1026 emu mol21.A least-squares fit of expression (1) to the data over the full temperature range yielded J = 217.3 ± 0.6 cm21, g = 2.06 ± 0.01 and r = 0.03 ± 0.003, while the TIP contribution was found to be (2.9 ± 0.1) × 1025 emu mol21. The susceptibility of a series of binuclear nickel(II) complexes with similar co-ordination and geometry has been treated recently in some detail,9 and it has been shown that further influences on the form of the susceptibility may also need to be taken into account.In particular, the non-cubic components of the ligand field may act on the S = 1 ground state to produce a zero-field splitting which may be of the same order of magnitude as J.15,16 We fitted our data using the explicit expression 9 for a dimer of S = 1 ions with a zero-field splitting D and intramolecular dimer exchange J. The optimised values of D and J were 17.9 ± 0.9 cm21 and 216.0 ± 0.6 cm21 respectively. It is conceivable that there is also an intermolecular exchange interaction J9 which will perturb the optimised values of D and J.However, there was little improvement in fit when J9 was introduced. In the absence of evidence for such an eVect, or any obvious structural feature suggesting that interdimer exchange should be significant, we chose not to explore this possibility further with more sophisticated fitting procedures. The results when J9 is set to zero are reproduced in Fig. 5. Cobalt. Samples of [Co2(H2L1)2(MeOH)2(H2O)2]Cl2?2MeOH could be prepared as crystals suYciently large for single crystal magnetic susceptibility measurements.Measurements were made on a flat piece of this material mounted on a silica strand with nail varnish over the temperature range 2.5–290 K. The susceptibility of the sample was found not to change signifi- Fig. 5 Molar magnetic susceptibility per NiII for [Ni2(H2L1)2- (H2O)4][ClO4]2?2H2O (open circles) with a fit by the expression for a dimer of exchange-coupled spins S = 1 with a zero-field splitting D as described in the text.Closed circles denote the eVective moment per NiII ion.J. Chem. Soc., Dalton Trans., 1998, 3953–3960 3957 cantly with its orientation for three diVerent settings relative to the flat plane of the flake, and a separate measurement on a powder sample held in a gelatine capsule yielded very similar data. The data are displayed in Fig. 6. The cusp centred at approximately 20 K indicates significant antiferromagnetic exchange, and the secondary rise in the susceptibility at lower temperatures implies that a paramagnetic impurity is present. The 3d 7 configuration of CoII in a octahedral ligand field and in a high-spin state has a 4T1 ground term which bestows an orbital contribution on the moment, raising it from the spinonly value of ÷15mB to a value that lies typically in the range 4.5–5.1mB.15 The precise calculation of the moment not only requires reliable values of ligand-field and spin–orbit coupling parameters, but also of the degree of electron delocalisation, and we treat it here merely as an empirical term to be estimated from our data.The molar susceptibility data were fitted by the expression (3) used for the binuclear nickel(II) complex with F(J,T) = (2e2x 1 10e6x 1 28e12x) (1 1 3e2x 1 5e63) F(J,T) modified for two centres each of spin S = ��� . The leastsquares fit yielded the values J = 26.9 ± 0.1 cm21, g = 2.49 ± 0.05 and r = 0.03 ± 0.0006, assuming the paramagnetic impurity to have a similar eVective moment as the cobalt(II) ions in the binuclear complex.This value of g yields a moment of 4.82mB, which is in the range of typical values given above for high-spin CoII in an octahedral field. The fit and the dependence of the eVective moment on temperature are displayed in Fig. 6. Deviations of the susceptibility from the theoretical expression probably reflect the eVect of single-ion anisotropy15,17 or anisotropic exchange,18 both of which are likely to be significant for a 4T1 term subjected to non-cubic structural distortions.Copper. The molar magnetic susceptibility of a powder sample of [Cu2(H2L1)2][ClO4]2?2thf was measured over the temperature range 2.5–330 K. The data are displayed in Fig. 7. The magnitude of the eVective moment is very small 7 except at the highest temperatures where the susceptibility rises on warming. This implies that the exchange coupling in the binuclear complex is strong and that the cusp in susceptibility lies well above the maximum experimental temperature.There was also a very small contribution to the susceptibility at lower temperatures, reflecting the presence of a small amount of paramagnetic impurity. Finally, a weak cusp centred at about 70 K reflects a small concentration of an unidentified impurity. We regard contamination with O2 as unlikely since this feature is not observed in the absence of the sample, it is not observed for Fig. 6 Molar magnetic susceptibility per CoII for [Co2(H2L1)2- (MeOH)2(H2O)2]Cl2?2MeOH (open circles) with a fit by the expression [eqns. (1) and (3)] for a dimer of exchange-coupled spins S = ��� as described in the text. Closed circles denote the eVective moment per CoII. other complexes investigated in the present study, and it appears to be a consistent feature for this particular complex. The 3d9 configuration of CuII in an octahedral ligand has a 2E1 ground term with a magnetic moment given approximately by the spin-only value of ÷3mB; perturbation by spin–orbit coupling with excited ligand-field states and a second-order Zeeman contribution modify this value to lie typically in the range 1.9–2.0mB.15 The molar susceptibility data were fitted by expression (1) used for the NiII 2 complex with F(J,T) modified for a dimer of spins S = ��� , i.e.as in eqn. (4). F(J,T) = 2e2x (1 1 3e2x) (4) Data in the region 40–100 K were not included in the fitting procedure because we did not know what to use to model the feature at 70 K.It should be noted, however, that the removal of this portion of the data did not have a significant influence on the fit at higher temperatures. It was not possible to produce a stable fit to the data if both J and the eVective moment were unconstrained, so we fixed the g value for the moments in the dimer to be 2.00, 2.20 and 2.40, spanning a range of values observed for binuclear copper(II) species.17 In all cases the TIP contribution was found to be 0.00064 ± 0.00002 emu mol21, and the fraction of paramagnetic impurity (assumed to be monomeric CuII with g = 2.00) to be 0.20(1)%.Values of J of 2415 ± 5, 2435 ± 5 and 2452 ± 4 cm21 were found for g = 2.00, 2.20 and 2.40 respectively. The X-band EPR spectrum of this complex is consistent with relatively strong antiferromagnetic coupling with a very broad signal being observed.The value for TIP is anomalously high, values of approximately 0.00006 emu mol21 being more typical. The feature at 70 K will have some contribution to this term in the fit, but is unlikely to account in full for this discrepancy. Magnetostructural trends in metal(II) dioxime complexes Magnetostructural correlations in dimeric transition metal compounds have been extensively studied and rationalised,19 particularly in the case of copper(II) ions linked through hydroxide ions.20 Detailed analysis of the relation between J and the structure and bonding in the compound requires a thorough consideration of the MOs constructed from valence orbitals on the metal ions and mediating ligands and is beyond the scope of this paper, so we will confine our discussion to a comparison with related compounds and general features of the magnetochemistry.The net exchange in all these materials is composed of competing ferro- and antiferro-magnetic Fig. 7 Molar magnetic susceptibility per CuII for [Cu2(H2L1)2- (ClO4)2]?2thf (open circles) with a fit by a modified Bleaney–Bowers expression [eqns. (1) and (4)] for a dimer of spins S = ��� with g fixed at 2.40. The feature centred at approximately 70 K is assumed to be an impurity and so data over the temperature range 40–100 K were not included in the fitting procedure.3958 J. Chem. Soc., Dalton Trans., 1998, 3953–3960 Table 5 Magnetic and structural parameters Complex [Co2(H2L1)2(H2O)2(MeOH)2]Cl2?2MeOH [Ni2(H2L1)2(H2O)4][ClO4]2?2H2O [Cu2(H2L1)2(ClO4)2]?2thf J/cm21 26.9 216.0 2452 S ��� 1� �� 4JS2/cm21 262.1 272 2452 M–O/Å 2.055(4), 2.075(4) 2.017(1), 2.021(1) 1.941(10), 1.965(12) M? ? ?M/Å 3.092(2) 3.0495(8) 2.994(4) M–O–M/8 96.9(2) 98.12(6) 99.2(8), 100.9(7) components.The antiferromagnetic component arises from the formation of MOs primarily involving s bonding between valence orbitals on M and O, and diVerences in the energy of these orbitals favours pairing of electrons to form a spin singlet.In this series of binuclear complexes the principal interaction will involve p orbitals on O and dx2 2 y2 on M, for which, in the first approximation, the MOs will be degenerate when the bridging angle q is 908 and hence there will be no antiferromagnetic term. In principle, however, accidental orthogonality should occur at higher angles due to s–p orbital mixing. This is compounded by a ferromagnetic term arising from quantum mechanical exchange of electrons which is independent of q to a first approximation.Thus, as q is either raised or lowered from 908 a crossover from ferro- to antiferro-magnetic J is expected at an angle that will depend, among other factors, on the strength of the M–O covalent bond and on the particular orbitals that contribute to the MOs. In the case of binuclear copper(II) complexes linked through two phenoxide bridges, J is found empirically to depend on q above 908 as J (cm21)= 215.98q (degrees) 1 1231 21 while the expression for a series of phenoxide-coupled nickel(II) dimers approximates to J (cm21) = 27.4q (degrees) 1 720.22 These expressions predict J = 26 and 2360 cm21 respectively for the binuclear complexes of NiII and CuII described herein, with deviations from these predicted values anticipated on the grounds that J will also depend on other ligands present and on the precise co-ordination geometry of the metal ion.In the case of the copper(II) complex this interpretation should be treated with caution.It has been pointed out 21 that the empirical relation between q and J predicts that the ferro- to antiferromagnetic crossover will occur at q = 798 for diphenoxy-bridged species, quite diVerent from the universal value of 97–97.58 observed in other O-bridged CuII dimers.9,10,21,23 The complexes in the present study involve diphenoxide bridges as part of a larger tetraimino pseudo-macrocycle. This may provide an alternative exchange pathway through the conjugated p framework of the azomethine nitrogen and the benzene ring.10 Such a term should also be antiferro-magnetic. Thus, it is possible that the phenoxide linkage in isolation would cross from ferro- to antiferro-magnetic exchange at an angle similar to that for other Cu–O–Cu dimers.There is far less work published on binuclear cobalt(II) complexes of this or indeed any sort, and no such empirical relations between J and q have been drawn up; small values of J have been observed in related exchange-coupled dimers.24 It should be noted that a true comparison for ions with diVerent numbers of unpaired electrons 4JS2,25 and this term is given in Table 5 together with appropriate bond lengths and bridging angles.The increase in 4JS2 with atomic number reflects the increase in M–O bond strength, supplemented by the increase in M–O–M angle. Further, as the number of singly occupied d orbitals increases from CuII to NiII to CoII, so too does the ferromagnetic contribution to the exchange, consistent with the observed value of J in the current study.Current work is aimed at the study of these and related systems in order to probe magnetochemical–structural correlations in polynuclear macrocyclic complexes. Experimental Infrared spectra were recorded as KBr discs using a Perkin- Elmer 1600 Series FTIR spectrometer over the range 400– 4000 cm21.Microanalyses were performed by the Edinburgh University and Nottingham University Chemistry Department Microanalytical Services. Proton and 13C NMR spectra were recorded on Bruker WP250, WP300 and WH300 instruments, fast atom bombardment (FAB) and electron impact (EI) mass spectra on a Kratos 50TC spectrometer with FAB spectra run using a 3-nitrobenzyl alcohol matrix. The molar magnetic susceptibilities of the complexes were measured using a Quantum Design MPMS2 SQUID magnetometer with an applied magnetic field of 0.1 T.The data were corrected for the diamagnetic contributions of the gelatine sample holder and of the constituent atoms.26 2,6-diformyl-4-methylphenol 27 and 2,6-diacetyl-4-methylphenol 28 were prepared by the literature methods. All solvents were of reagent grade and were used as received. Synthesis of 2,6-diformyl-4-methylphenol dioxime (H3L1) To a solution of 2,6-diformyl-4-methylphenol (0.255 g, 1.55 mmol) in EtOH (100 cm3) were added NH2OH?HCl (1.084 g, 15.60 mmol) and CH3CO2K (1.529 g, 15.58 mmol).The system was refluxed for 3 h during which time the inorganic solids remained in suspension. The mixture was allowed to cool to room temperature prior to removal of the inorganic solids by filtration. Concentration of the filtrate by rotary evaporation followed by addition of deionised water resulted in the precipitation of a white solid which was collected by filtration and dried under suction. Yield = 70%.The product was purified by chromatography on silica (Merck, Kieselgel 60) using CH2Cl2– (CH3)2CO (2: 1 v/v) as eluent (Found: C, 55.67; H, 5.00; N, 14.20. Calc. for C9H10N2O3: C, 55.67; H, 5.15; N, 14.43%). EI mass spectrum: m/z 194 (100%); Calc. for [C9H10N2O3]1 194. 1H NMR [250.13 MHz, (CD3)2CO, 298 K]: d 2.27 (s, CH3, 3 H), 7.37 (s, aromatic CH, 2 H), 8.38 [s, C(NOH)H, 2 H], 10.48 (s, phenolic OH, 1 H) and 10.71 (s, oxime OH, 2 H). 13C DEPT NMR [62.90 MHz, (CD3)2CO, 298 K]: d 18.54 (CH3), 117.94, 127.63 (quaternary aromatic), 128.68 (aromatic CH), 146.71 [C(NOH)H] and 152.29 (COH).IR (KBr disc): 3293 (br) vs, 2919w, 1623m, 1604m, 1464s, 1377m, 1307s, 1265s, 1222w, 1176w, 1061s, 1027s, 974m, 934s, 863m, 793s, 745s, 697s and 569w cm21. Synthesis of 2,6-diacetyl-4-methylphenol dioxime (H3L2) The method used was as above except that it was not necessary to purify the final product by chromatography. Yield = 62% (Found: C, 59.61; H, 6.39; N, 12.60.Calc. for C11H14N2O3: C, 59.46; H, 6.31; N, 12.61%). FAB mass spectrum: m/z 223 (100%); Calc. for [C11H14N2O3 1 H]1 223. 1H NMR (300.13 MHz, CD3OD, 298 K): d 2.26 [s, C(NOH)CH3, 6 H], 2.29 (s, CH3, 3 H) and 7.18 (s, aromatic CH, 2 H). 13C DEPT NMR (75.48 MHz, CD3OD, 298 K): d 11.79, 19.22 (CH3), 122.50, 127.32 (quaternary aromatic), 129.32 (aromatic CH), 153.68, 157.39 [COH and C(NOH)CH3]. IR spectrum (KBr disc): 3376 (br) s, 2922w, 1773w, 1647s, 1612w, 1452vs, 1368s, 1342m, 1300w, 1265vs, 1249vs, 1204w, 1190s, 1103w, 1035s, 957vs, 912m, 899m, 867m, 831m, 790vs, 724m, 684s, 600m, 576m and 518w cm21.Single-crystal structure determination of 2,6-diacetyl-4-methylphenol dioxime (H3L2) Slow evaporation of a (CD3)2CO solution of the compoundJ. Chem. Soc., Dalton Trans., 1998, 3953–3960 3959 Table 6 Summary of crystal data Formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 Z m(Mo-Ka)/mm21 T/K Reflections used Parameters refined R1, wR2 (SHELXL 93)a R, R9 (SHELX 76)b C11H14N2O3 222.24 Triclinic P1� 7.707(2) 8.574(3) 9.597(2) 108.56(2) 110.70(2) 92.07(2) 554.4 2 0.098 298 1463 149 0.0545, 0.162 C18H26Cl2N4Ni2O18?2H2O 810.78 Monoclinic P21/c 8.672(2) 10.565(2) 16.568(2) 89.91(2) 1518 2 1.51 150 3468 229 0.0284, 0.0726 C20H30Co2N4O10?2CH3OH 739.32 Triclinic P1� 9.033(6) 9.224(5) 10.085(6) 71.26(2) 82.69(3) 71.90(4) 756 1 1.34 150 1934 204 0.0407, 0.104 C18H18Cl2Cu2N4O14?2C4H8O 818.45 Tetragonal I4� 2d 21.527(3) 21.527(3) 14.759(8) 6839 8 1.471 295 1031 135 0.0753, 0.0727 a R1 is based on F > 4s(F ), wR2 on all F 2 data.b R and R9 are based on data with F > 5s(F ). aVorded yellow crystals of diVraction quality. The selected crystal was mounted on a Stoe Stadi-4 four-circle diVractometer (Table 6). The structure was solved by direct methods using SHELXS 8629 and refined on F 2 using SHELXL 93.30 All non-H atoms were refined with anisotropic thermal parameters and all H atoms were located from DF synthesis introduced at calculated positions and refined as part of a rigid group or using a riding model.Synthesis of [Ni2(H2L1)2(H2O)4][ClO4]2?2H2O A solution of Ni(ClO4)2?6H2O (0.119 g, 0.35 mmol) dissolved in MeOH (8 cm3) was layered on top of a frozen solution of 2,6- diformyl-4-methylphenol dioxime (0.063 g, 0.32 mmol) in thf (8 cm3), contained in a Schlenk tube. The whole system was frozen by immersion in liquid nitrogen, placed in an empty narrow-necked dewar and allowed slowly to warm to room temperature. Thawing and mixing of the two solutions resulted in the formation of a green solution.Crystalline [Ni2(H2L1)2- (H2O)4][ClO4]2?2H2O was obtained by slow evaporation of the MeOH–thf. Yield = 65%. FAB mass spectrum: m/z 501 (100%); Calc. for [58Ni2(H2L1)2 2 H]1 501 with correct isotopic distribution. IR (KBr disc): 3322 (br) s, 3114 (br) w, 1648m, 1621s, 1560s, 1508m, 1440vs, 1353m, 1316vs, 1270w, 1234s, 1145vs, 1112vs, 1086vs, 984s, 869w, 814s, 772m, 700s, 637s, 529m and 502m cm21.Electronic spectrum (in CH3CN): lmax 560 (emax = 15), 365 (9210), 254 (49630) and 209 nm (38340 dm3 mol21 cm21). Single-crystal structure determination of [Ni2(H2L1)2(H2O)4]- [ClO4]2?2H2O Slow evaporation of the thf–MeOH solution of the complex aVorded large, rust-brown block-like crystals. A lath was cut from one of these large crystals and sealed in a Lindemann glass capillary tube to prevent decomposition.The crystal was cooled to 150 K using an Oxford Cryosystems open-flow cryostat 31 mounted on a Stoe Stadi-4 four-circle diVractometer (Table 6). The structure was solved by direct methods using SHELXS 8629 and refined on F 2 using SHELXL 93.30 All non- H atoms were refined with anisotropic thermal parameters. The H atoms on the H2O of crystallisation and the co-ordinated H2O were located in a diVerence map following several cycles of least squares weighted towards high-angle data and then refined with restraints on O–H distances and H–O–H angles and with Uiso(H) = 1.2Ueq(O).All other H atoms were placed geometrically or located from diVerence maps and refined as rigid groups or using a riding model. Synthesis of [Co2(H2L1)2(MeOH)2(H2O)2]Cl2?2MeOH To a solution of 2,6-diformyl-4-methylphenol dioxime (0.147 g, 0.76 mmol) dissolved in MeOH (30 cm3) was added CoCl2?6H2O (0.180 g, 0.76 mmol). On dissolution of the CoCl2?6H2O an orange solution was produced which was stirred overnight.Removal of the MeOH by rotary evaporation aVorded an emerald green solid. This partially redissolved on the addition of more MeOH to aVord an orange solid suspended in an orange solution. The addition of deionised water caused the orange solid to flocculate. The solid was collected by filtration, washed with the mother-liquor and dried in vacuo. Yield = 60% {Found: C, 40.23; H, 3.73; N, 10.24. Calc. for [Co2(H2L1)2(MeOH)2(H2O)2]Cl2?H2O: C, 40.47; H, 3.77; N, 10.49%}.FAB mass spectrum: m/z 503 (45%); Calc. for [59Co2(H2L1)2 2 H]1 503 with correct isotopic distribution. IR (KBr disc): 334219s, 3002s, 2931s, 2778m, 1636m, 1617s, 1554vs, 1499m, 1443vs, 1350m, 1319vs, 1272w, 1234vs, 1193w, 1073m, 980vs, 872w, 812m, 767m, 697m, 637w, 518m and 501m cm21. Single-crystal structure determination of [Co2(H2L1)2(MeOH)2- (H2O)2]Cl2?2MeOH Slow evaporation of a MeOH solution of the complex aVorded small orange block-like crystals of diVraction quality.The determination was carried out as for the previous compound. Synthesis of [Cu2(H2L1)2(ClO4)2]?2thf A solution of 2,6-diformyl-4-methylphenol dioxime (0.1164 g, 0.60 mmol) in thf (20 cm3) was added to a 100 cm3 beaker and frozen by careful immersion of the beaker in liquid N2. The frozen solution was allowed to thaw by warming to room temperature. To the freshly thawed solution of 2,6-diformyl-4- methylphenol dioxime was added a solution of Cu(ClO4)2? 6H2O (0.2488 g, 0.67 mmol) in MeOH (20 cm3).The solution was stirred overnight at room temperature during which time a khaki-green solid precipitated. This was collected, washed with diethyl ether and dried under suction. Yield = 40% {Found: C, 34.3; H, 3.6; N, 6.6. Calc. for [Cu2(H2L1)2(ClO4)2]?2thf: C, 34.5; H, 4.0; N, 6.5%}. IR (KBr disc): 3260, 1642, 1628, 1612, 1592, 1370, 1350, 1313, 1268, 1243, 1200, 1100, 1003, 972, 956, 922, 882, 820, 762, 708, 682, 564 and 514 cm21.Single-crystal structure determination of [Cu2(H2L1)2(ClO4)2]? 2thf Crystals of diVraction quality were obtained from thf– methanol solution. The selected crystals were coated in araldite resin and mounted on a Phillips PW1100 four-circle diVractometer. Owing to crystal decay, data from two crystals were collected and merged (Table 6). The structure was solved for the copper atoms by heavy-atom methods using SHELX 76:32 the3960 J.Chem. Soc., Dalton Trans., 1998, 3953–3960 remaining atoms were located in successive Fourier-diVerence syntheses and the structure was refined on F using SHELX 76.32 The Cu and Cl atoms were refined with anisotropic thermal parameters. The H atoms on carbons were introduced at calculated positions and allowed to ride at a fixed distance of 1.08 Å, while two alternative sites for each of the hydroxyl H atoms were located in a Fourier-diVerence map. These were positionally refined with site occupation factors of 0.5.CCDC reference number 186/1155. Synthesis of [Cu2(HL1)2] A solution of 2,6-diformyl-4-methylphenol dioxime (0.1 g, 0.52 mmol) in dmf (50 cm3) was added to a solution of [Cu(O2- CMe)2] (0.24 g, 1.2 mmol) in dmf to produce a light coloured precipitate. The reaction solution was filtered and dried to give green microcystals of the product. The solution was stirred overnight at room temperature during which time a khaki-green solid precipitated.Yield = 90% {Found: C, 41.8; H, 3.4; N, 10.9. Calc. for [Cu2(HL1)2]: C, 42.1; H, 3.4; N, 10.9%}. IR (KBr disc): 3450, 3025, 3005, 2930, 1620, 1600, 1588, 1402, 1350, 1302, 1237, 1190, 1098, 1070, 1010, 960, 930, 906, 865, 822, 762, 707, 686, 582, 568, 519, 504, 477 and 430 cm21. Acknowledgements We thank Zeneca Specialties plc for a CASE Award (to D. B.), the EPSRC for support and the EPSRC National Mass Spectometry Service, University of Swansea. We thank the referees for helpful comments and suggestions.References 1 Comprehensive Co-ordination Chemistry, eds. G. Wilkinson, R. D. Gillard and J. A. McCleverty, Pergamon, Oxford, 1987, vol. 6. 2 B. McCudden, P. O’Brien and J. R. Thornback, J. Chem. Soc., Dalton Trans., 1983, 2043; Hydroxyoximes and Copper Hydrometallurgy, CRC Press, Boca Raton, FL, 1993. 3 S. O. Sommerer, B. L. Westcott, A. J. Jircitano and K. A. Abboud, Inorg. Chim. Acta, 1995, 238, 149. 4 C. Onindo, T. Yu. Sliva, T Kowalik-Jankowska, I.O. Fritsky, P. Buglyo, L. D. Pettit, H. Kozlowski and T. Kiss, J. Chem. Soc., Dalton Trans., 1995, 3911. 5 B. Mernari, F. Abraham, M. Lagrenee, M. Dillon and P. Legoll, J. Chem. Soc., Dalton Trans., 1993, 1707; F. Abraham, M. Lagrenee, S. Sueur, B. Mernari and C. Bremard, J. Chem. Soc., Dalton Trans., 1991, 1443. 6 M. M. Aly, A. O. Baghlaf and N. S. Ganji, Polyhedron, 1985, 4, 1301. 7 H. Okawa, T. Tokii, Y. Muto and S. Kida, Bull. Chem. Soc. Jpn., 1973, 46, 2464. 8 C. Krebs, M. Winter, T. Weyhermüller, E. Bill, K. Wieghardt and P. Chaudhuri, J. Chem. Soc., Chem. Commun., 1995, 1913. 9 K. K. Nanda, A. W. Addison, N. Paterson, E. Sinn, L. K. Thompson and U. Sakaguchi, Inorg. Chem., 1998, 37, 1028. 10 E. V. Rybak-Akimova, D. H. Busch, P. K. Kahol, N. Pinto, N. W. Alcock and H. J. Clase, Inorg. Chem., 1997, 36, 510. 11 E. Buehler, J. Org. Chem., 1967, 32, 261; E. Buehler and G. B. Brown, J. Org. Chem., 1967, 32, 265; E. Falco and G. B. Brown, J. Med. Chem., 1968, 11, 142. 12 C. E. Pfluger and R. L. Harlow, Acta Crystallogr., 1964, 17, 1109. 13 A. Chakravorty, Coord. Chem. Rev., 1974, 13, 1 and refs. therein. 14 W. E. Hatfield, Comments Inorg. Chem., 1981, 1, 105. 15 A. P. Ginsberg, R. L. Martin, R. W. Brookes and R. C. Shewood, Inorg. Chem., 1972, 11, 2884. 16 See also, B. N. Figgis, Introduction to Ligand Fields, Wiley, New York, 1966; A. P. Ginsberg and M. E. Lines, Inorg. Chem, 1972, 11, 2289; R. L. Carlin and J. N. McElearney, Inorg. Chem., 1972, 11, 2291; A. B. P. Lever, L. K. Thompson and W. M. ReiV, Inorg. Chem., 1972, 11, 2291. 17 R. L. Carlin, Magnetochemistry, Springer, Berlin, 1986. 18 S. Nakatsuka, K. Osaki and N. Uryu, Inorg. Chem., 1982, 21, 4332. 19 Magnetostructural Correlations in Exchange Coupled Systems, eds. D. Gatteschi, O. Kahn and R. D. Willett, Reidel, Dordrecht, 1984; O. Kahn, Molecular Magnetism, VCH, Weinheim, 1995. 20 V. H. Crawford, H. W. Richardson, J. R. Wasson, D. J. Hodgson and W. E. Hatfield, Inorg. Chem., 1976, 15, 2107. 21 L. K. Thompson, S. K. Mandal, S. S. Tandon, J. N. Bridson and M. K. Park, Inorg. Chem., 1996, 35, 3117. 22 K. K. Nanda, L. K. Thompson, J. N. Bridson and K. Nag, J. Chem. Soc., Chem. Commun., 1994, 1337. 23 A. J. Blake, J. P. Danks, A. Harrison, S. Parsons, P. Schooler, G. Whittaker and M. Schröder, J. Chem. Soc., Dalton Trans., 1998, 2335; A. J. Blake, J. P. Danks, I. A. Fallis, A. Harrison, W.-S. Li, S. Parsons, S. A. Ross, G. Whittaker and M. Schröder, J. Chem. Soc., Dalton Trans., 1998, 3969. 24 S. L. Lambert and D. N. Hendrickson, Inorg. Chem., 1979, 18, 2683. 25 R. K. Nesbet, Ann. Phys., 1958, 4, 87. 26 C. J. O’Connor, Prog. Inorg. Chem., 1982, 29, 203. 27 F. Ullman and K. Brittner, Berichte, 1909, 42, 2539. 28 S. K. Mandal and K. Nag, J. Chem. Soc., Dalton Trans., 1983, 2429. 29 G. M. Sheldrick, SHELXS 86, program for crystal structure solution, Acta Crystallogr., Sect. A, 1990, 46, 467. 30 G. M. Sheldrick, SHELXL 93, program for crystal structure refinement, University of Göttingen, 1993. 31 J. Cosier and A. M. Glazer, J. Appl. Crystallogr., 1986, 19, 105. 32 G. M. Sheldrick, SHELX 76, program for crystal structure refinement, University of Cambridge, 1976. Paper 8/07031H

ISSN:1477-9226    

DOI:10.1039/a807031h

出版商:RSC    

年代:1998

数据来源: RSC