摘要:

DALTON PERSPECTIVE J. Chem. Soc., Dalton Trans., 1998, 4071–4080 4071 Solving crystal structures of molecular solids without single crystals: a simulated annealing approach Yuri G. Andreev and Peter G. Bruce School of Chemistry, University of St. Andrews, St. Andrews, Fife, UK KY16 9ST Received 13th July 1998, Accepted 6th October 1998 The ab initio determination of relatively complex crystal structures of flexible molecules without the need for single crystals is discussed. A method is described based on simulated annealing in which the powder diVraction patterns of randomly generated trial structures are calculated and compared with the observed powder diVraction pattern in order to identify the model which provides the best fit and therefore the true structure.By employing simulated annealing both downhill (improved fit) and uphill (reduced fit) moves are possible ensuring escape from local minima in order to find the global minimum in the goodness-of-fit, i.e.the true structure. Key to the successful solution of flexible molecules is the introduction of a geometrical description which specifies atomic positions within the unit cell in terms of bond lengths and angles. In this way only those random structures which are chemically plausible are generated, greatly reducing the number of trial structures and rendering tractable the otherwise impossible task of ab initio determination. It is shown that structures with 37 variable parameters can be solved from only a few milligrams of powder.The limits of structural complexity for this method should be similar to those for refinement using powder data, i.e. around 200 variables. The variables may be those of position, or orientation of the molecule(s) in the unit cell as well as bond lengths, bond angles or torsion angles. 1 Introduction Single crystal X-ray diVraction remains the technique of choice for obtaining accurate structural data. In some quarters the absence of a single crystal is taken to imply that the compound is amorphous and not therefore amenable to crystallographic study, this is generally incorrect. The powdered form of the compound is usually composed of crystals with dimensions in the range 0.1–1 mm.Although too small for single crystal studies, crystal structures may be obtained by analysing the powder X-ray diVraction data. Developments within the last two years have led to robust methods by which crystal structures of relatively high complexity may be solved ab initio from powder diVraction data collected on a few mg of sample.Critically these structures may be of any type and include molecules that are completely flexible. The structural data that may be obtained is of high quality, significantly exceeding that of most other structural techniques available to chemists. We are seeing the dawn of a new and exciting age of crystallography without single crystals. In this paper we describe the simulated annealing (SA) method of ab initio structure solution from powder diVraction data illustrated by examples involving the determination of molecular crystals with up to 37 variable structural parameters. 2 The challenge of ab initio structure solution from powder diVraction data The same basic information is contained in powder and single crystal diVraction data sets. In both cases the intensities arise from reflection of X-rays by the diVerent sets of lattice planes, however in the latter case the intensities are distributed in three dimensions around the crystal whereas in the former they are Peter Bruce was awarded a PhD in chemistry following work with Professor A.R. West on lithium ion conducting solid electrolytes at the University of Aberdeen. After a period of postdoctoral study with Professor J. B. Goodenough then Head of the Inorganic Chemistry Laboratory at the University of Oxford and a period as Lecturer at Heriot-Watt University in Edinburgh he moved to the University of St.Andrews in 1991 where he now holds a Chair in Chemistry and is currently Head of Department. His research interests span the structure and properties of inorganic solids and polymer salt complexes. He has been particularly involved in the structural chemistry, largely using crystallographic techniques, of lithium ion conducting solids, intercalation compounds and polymer electrolytes. He has held research Fellowships from the Royal Society and the Royal Society of Edinburgh.In 1995 he was elected to a Fellowship of the Royal Society of Edinburgh. Yuri Andreev received his MSc in Applied Nuclear Physics in 1981 and PhD in Solid State Physics in 1986 from Moscow Engineering Physics Institute. He has worked as a guest scientist at the University of Uppsala with Professors Torsten Lundström and Josh Thomas. Since 1995, he has been a research fellow at the University of St. Andrews working for Professor Peter Bruce.His research interests extend throughout all aspects of powder diffraction data treatment. Peter Bruce Yuri Andreev4072 J. Chem. Soc., Dalton Trans., 1998, 4071–4080 compressed onto one dimension. A powder diVraction pattern contains data from all the randomly orientated crystals exposed to the X-ray beam and the intensities consist of a set of peaks separated only by the value of the d-spacings between the planes from which they arise. Since the structural information is contained in the intensities of the reflections, the key problem facing structure determination from powder diVraction data is the presence of peak overlap.Although some peaks are isolated many are overlapped either partially or completely (the latter when the d-spacings of diVerent sets of planes coincide). This results in a loss of access to the individual intensities which are critical to the structure solution. Modern high-resolution instruments have improved greatly the separation of peaks however significant numbers of reflections cannot be resolved by any improvement in instrumentation. A number of structures have been determined from powder diVraction data over the last twenty years.These have employed mainly proven single crystal approaches, particularly Direct Methods1 or Patterson synthesis.2 Inevitably they rely on extracting the intensities of individual reflections or groups of reflections using equipartitioning or permutation estimates for grouped reflections, then treating these as a single crystal data set.The combination of a limited number of intensities and the absence of short d-spacing reflections due to the truncated nature of powder diVraction data compared with single crystals, can conspire to frustrate these single crystal methodologies. Nevertheless they have played a significant role in ab initio structure solution from powders and will continue to do so through for example the medium of programs such as EXPO.1 The SA approach described in this paper, diVers fundamentally from the established methods in that no attempt is made to extract individual intensities and treat them in a single crystal sense.Instead eVort is concentrated on generating chemically plausible but random structural models which are tested against the whole powder profile. A brief survey of currently available methods for structure determination from powder diVraction data is given in ref. 3. Rietveld recognised the inevitable limitation of any approach which relies on obtaining individual intensities and in response introduced whole-pattern fitting.4 His method involves calculating the entire powder profile (not just the individual intensities) based on a model structure then varying the structural parameters (e.g. atomic positions) of the model by least-squares until a profile is generated which best fits the observed profile.The best fit is determined by minimising the figure-of-merit function, c2, which acts as a measure of the goodness-of-fit. This approach has been used successfully to refine the details of partially known structures in many hundreds of cases. The only barrier to ab initio structure determination, as opposed to refinement, with the Rietveld approach is the use of the leastsquares method which necessitates a starting structural model that is close to the correct structure because the least-squares routine can only adjust structural parameters in the direction of reducing c2 (downhill move), i.e.increasing the goodness-of-fit between the calculated and observed profiles. In other words the Rietveld method can locate only the local minimum in the goodness-of-fit hence the necessity to start from a structural model for which the local minimum coincides with the global minimum. By definition in ab initio structure determination the starting structural model will bear little relationship to the correct structure. The probability is negligible that starting from such a random structural model, the true structure (corresponding to the global minimum in the goodness-of-fit) can be obtained.Among the minimisation methods capable of finding the global minimum of the goodness-of-fit in the presence of multiple local minima are the methods of SA 5 and genetic algorithm.6 Minimisation by the latter using full-profile fitting of a powder pattern was successfully applied to the solution of the crystal structure of ortho-thymotic acid (2-hydroxy-3- isopropyl-6-methylbenzoic acid), the molecule of which contains only two internal degrees of freedom.7 It is also shown mathematically that the genetic algorithm method, although competitive with SA approaches, can be expected to be orders of magnitude less eYcient.8 In the SA method a Monte Carlo procedure is used to generate random models for the structure.This is achieved by making stepwise increments, random in size and direction, of the structural parameters (e.g.atomic coordinates). The models may yield a better fit (downhill, i.e. lower c2) or worse fit (uphill, higher c2) between the calculated and observed profile. Critically the latter permits escape from local minima. As the minimisation progresses, tolerance for the uphill steps gradually decreases until steps in both directions are exhausted. At this point the set of adjustable atomic coordinates corresponds to the lowest possible value of the figure-of-merit function, i.e.the global minimum. The potential of SA in the realm of structure determination was first demonstrated by solving the previously known crystal structure of benzene using a modified Rietveld method.9 In addition to a SA minimisation the authors included a rigid-body representation of the constituent benzene rings. This allowed significant reduction of the number of variable structural parameters.The crystallographic coordinates of all the constituent atoms used in calculating the powder pattern were, in the case of benzene, computed using only positional and orientation parameters of the rigid body as a whole. The rigid-body approach has been further exploited in the structure solution of molecular structures that are marginally more complex than that of benzene.3,10 However the authors did not utilise minimisation of the full-pattern goodness-of-fit function but instead generated trial rigid-body structures in a Monte Carlo fashion analysing subsequently all the moves to select several low minima.The implementation of such an approach involves multiple subsequent refinements in order to identify the structure that corresponds to the lowest value of the figure-of-merit function. While successful for relatively rigid structures, it is likely that this approach will represent a barrier to solving structures with larger numbers of intermolecular degrees of freedom.If structure solution from powder diVraction data is to become of general utility it is vital to develop a robust approach capable of tackling flexible as well as rigid structures. Determination of flexible structures by SA is much more challenging than rigid bodies, since the number of possible structural permutations appears at first sight to be enormous. Indeed it has been estimated that the computation would take up to 109 years even for a relatively simple structure.11 We have chosen to adopt the SA approach used successfully for benzene but have extended significantly its application to embrace the range of crystals containing highly flexible moieties.Flexibility can be in the bond lengths, bond angles or torsion angles and we shall show that in some cases it is essential to vary all of these in order to achieve a successful structure solution. The modified SA method is also capable of tackling the solution of polymeric structures in which a single molecule straddles more than one asymmetric unit.The problem here is that random models of the asymmetric unit are only valid if they generate chain continuity. Thus the inter-atomic connectivity at the junctions of neighbouring asymmetric units is determined merely by the relevant symmetry operator of the space group and cannot be maintained by random variation of the relevant bond length, bond angle and torsion angle.Key to our approach is the development of a stereochemical description that permits the atomic positions of the structural model to be defined in terms of bond lengths, bond angles and torsion angles, rather than individual atomic coordinates.12 This in turn permits us to restrict attention to chemically plausible structural models thus vastly reducing the number of trial structures and rendering tractable the otherwise impossible taskJ. Chem. Soc., Dalton Trans., 1998, 4071–4080 4073 of solving the crystal structures of flexible molecules.The approach to a geometrical description of flexible molecules via stereochemical parameters is suitable for describing a molecular fragment of any kind. 3 Simulated annealing The principles behind the SA approach and its distinctive features in comparison with other methods of continuous minimisation are best understood by analogy with the process of forming a solid by cooling from a melt. Let us assume that the solid phase can be either amorphous or crystalline.At temperatures above the melting point, atoms have a high mobility and are in chaotic motion, the total energy of the ensemble is also high. The minimum energy of this system corresponds to the crystalline solid. The amorphous phase is at an intermediate position on the energy scale. There are two extreme routes by which the melt may be solidified: slow cooling or quenching. During the latter process, a random atomic configuration is immediately frozen, forming a glass the total energy of which is somewhat higher than in the crystalline state.If the rate of temperature decrease is low enough such cooling corresponds to an annealing process in which the chaotic motion of free atoms in the melt is gradually reduced allowing the ensemble to explore fully the energy space and hence to adopt the most energetically favourable (crystalline) configuration. Applying this thermodynamic reasoning to crystal structure solution from powder data requires substitution of the atoms of the melt by variable structural parameters of the ensemble (e.g.the atomic coordinates or bond lengths) and energy by the value of a goodness-of-fit function (c2). The most frequently used method of minimisation is ‘conventional’ gradient least-squares.13 This is the analogue of quenching. It does not allow the necessary chaotic changes but as described above allows only downhill movement to the local minimum.SA does permit the essential uphill steps necessary to escape local minima in the search for the global minimum corresponding to the true structure. The number of uphill steps representing choatic behaviour of the figure-of-merit function should however be slowly decreased by introducing a varying attenuation factor so that minimisation towards the global minimum may occur. This procedure is analogous to that of slow cooling (annealing) with the attenuation factor acting as the temperature.A convenient way of introducing random steps in the structural model which includes the possibility of uphill moves is known as the standard importance sampling algorithm.14 Application of this procedure to the minimisation of a figureof- merit function, c2(P), whose value is determined by a set of the crystallographic parameters (e.g. bond angles and torsion angles) P, may be outlined as follows. A new set of parameter values Pi (i.e.a new crystal structure model) is accepted if either c2(Pi) < c2(Pi21) or if exp{2[c2(Pi) 2 c2(Pi 2 1)]/ Dc2 cur} > R, where Pi 2 1 is a previously accepted set of parameters, Dc2 cur is a current marginal value of the c2 variation serving as a temperature analogue, R is a random number in the range from 0 to 1. In the case of continuous minimisation each j-th component pj i of the Pi set is calculated via the pj i21 value of the Pi 2 1 vector in a Monte Carlo fashion, where Dpj is a pj i = pj i21 1 rj?Dpj (1) predefined maximum stepwidth and rj is a random number in the range from 21 to 1.Once Pi is accepted then Pi 2 1 = Pi and the process reiterates. An account of various types of the “temperature-reduction” procedure is given in ref. 15. A recently reported SA protocol 16 for structure solution incorporates features of genetic algorithm and, as claimed, facilitates the search for the global minimum. Here we mention a temperature-reduction scheme which was used successfully in the examples of the structure solutions presented in Section 5.At a given value of Dc2 cur the sampling algorithm reiterates for as long as the total number of rejected and accepted Pi sets of parameters (referred to as moves from here on) exceeds the pre-set value of Ntot or until the number of accepted moves becomes greater than f1?Ntot, with the f1 value also chosen in advance. As soon as this happens the value of Dc2 cur is reset to (1 2 f2)?Dc2 cur with a predetermined value of f2 and the whole procedure continues.Minimisation terminates when there are no downhill moves at a current value of Dc2 cur. 4 Constraints and restraints Reduction in the number of structures that must be explored is essential to the success of the minimisation process. This is particularly important in crystal structure solution when the original structural model is likely to be a poor approximation to the true crystal structure and where minimisation of the figure-of-merit function must be performed in an extremely time consuming Monte Carlo fashion.This can be accomplished by imposing rigorous or hard constraints and soft or slack constraints (restraints) on the number of parameters or on their values. By these means only chemically plausible molecules (e.g. no unrealistic bond lengths) need to be explored. 4.1 Non-structural constraints The most obvious constraint to be used in the course of structure solution by a full-pattern fitting approach is fixing the values of those profile-defining parameters that are not directly related to the arrangement of atoms within the unit cell.The values of cell constants, peak-shape and half-width parameters, background, peak asymmetry etc. are readily determined with reasonable accuracy using full-pattern decomposition methods without reference to a structural model.17,18 The only parameter that has no eVect on the structure but cannot be fixed in advance is the scale factor for the calculated pattern.However, its value is easily computed for each new trial structural model using the linear least-squares method. The eVect of introducing these constraints is twofold. First, there is the advantage of a reduction in the number of variables. Second, the disadvantage is that it is unreasonable to expect the fixed values to provide the best fit to the experimental powder pattern when the profile is calculated using the structural parameters instead of the integrated intensities of individual Bragg peaks used during the full-pattern decomposition.In this case the restrained structure solution terminates close to the ‘true’ global minimum rather than in the minimum itself. To reach the minimum the subsequent trivial task of refinement of all parameters using the Rietveld method is required. 4.2 Structural restraints Structural restraints are already an integral part of modern software for structure refinement by the Rietveld method and are a means of taking into account our knowledge about the possible atomic arrangement.19 In most cases the restraints are introduced in calculation of the figure-of-merit function such that models which violate the restraining limits produce a high c2.These restraints do not eliminate unreasonable structural models from the refinement process. Such an approach is rendered unattractive for structure solution because unrealistic models are still generated and this is computationally demanding in SA.An entirely diVerent approach to the imposition of restraints on structures not only preserves all the attributes of the established procedure but, in addition, allows randomised minimisation by SA strictly within the pre-determined limits of the structural parameters. In other words, increments leading to parameter values outside the limits are automatically rejected by SA but they can still be accepted by a refinement procedure4074 J.Chem. Soc., Dalton Trans., 1998, 4071–4080 if they produce a lower overall (pattern fit plus geometryrelated terms) c2 value. In the case of molecular crystals the connectivity of atoms within the molecule is generally known. By describing the positions of the atoms in terms of their bond lengths, bond angles, and torsion angles rather than independent atomic coordinates only chemically plausible structural models need be explored by restraining the values of these stereochemical parameters to be within certain limits.This greatly reduces the number of trial structures; however they may be reduced further by checking for unfavourably close proximity of non-bonding atoms and for the spatial continuity of an infinite, within the crystallite size, moiety (e.g. polymer chain) at the junction of neighbouring asymmetric units. Although it is the stereochemical parameters that are altered to generate each new chemically plausible model, the crystallographic coordinates for each model are still required in order to calculate the powder profile using the conventional mathematical formalism.However these may be readily obtained in terms of the stereochemical parameters (bond lengths lj, j 1 1, bond angles fj, j 1 1, j 1 2, and torsion angles tj, j 1 1, j 1 2, j 1 3) by expressing initially the atomic coordinates for each molecule in a local Cartesian frame following the approach proposed in ref. 20. This type of description requires representation of each molecule as a sequence of chains. In cases when bond angles are more readily constrained than torsion angles it is sometimes more convenient to calculate the coordinates of chosen atoms by rotating bonds around each other.21 Transformation of the atomic coordinates from the local Cartesian to the crystallographic frame introduces a set of additional parameters which determine the position, through crystallographic coordinates of the reference atom positioned at the origin of the local frame, and orientation, through Eulerian angles Q, F, Y, of the molecular fragment, as a whole, in the unit cell.A detailed mathematical account of the stereochemical description of molecules and of the frame transformations is given elsewhere.12 Once this procedure is introduced into a Rietveld-type algorithm in which the original least-squares procedure is substituted by the method of SA, the restraints are imposed in a straightforward manner by allowing the parameters to accept only reasonable values within predetermined limits instead of punishing the value of c2(P) when the limits are violated. The range of values for bond lengths and bond angles between atoms in most types of compounds is readily available, torsion angles vary between 2p and p, the coordinates of the reference atom are kept within the boundaries of the asymmetric unit.The limits on the Eulerian angles are 0 £ Q, Y < 2p and 0 £ F < p.Such a description allows the introduction of further constraints which reduce the total number of parameters to be varied. For example chemical knowledge can indicate that all like bond lengths or bond angles (e.g. all lC–C and all fC–C–C in a benzene ring) can be treated as variable but equal to each other, or that a certain part of the molecule is flat implying that the corresponding torsion angles can be kept at fixed values 0 or p, or the whole molecular fragment is rigid in which case only the values of the Eulerian angles and of the reference-atom coordinates are to be varied.Although introduction of the above constraints is computationally beneficial, it must be used with caution because in certain cases, illustrated below, even a slight reduction of the molecular flexibility can mislead the structure solution. 5 Examples The procedure for structure solution from powder data is best understood by considering some examples.The following structure solutions were performed using X-ray powder diVraction patterns collected in steps of 0.028 in transmission mode on a STOE STADI/P diVractometer with Cu-Ka1 radiation. The computer code implementing the full-profile-fitting procedure with minimisation by the SA method was written in C. The code for pattern calculation was adapted from the CPSR software package.22 All the structures presented here are described in the monoclinic space group, P21/c.Details concerning indexing and space group determination, essential stages of all modern structure solution procedures, are given in refs. 12 and 23. Profile parameters and lattice constants were fixed at the values obtained from profile-fitting using the CPSR program suite. The background was subtracted manually. The set of variable parameters P used in the SA runs included the overall isotropic temperature factor, B. Hydrogen atoms were ignored during the structure solution and were added only at the refinement stage.Unless otherwise stated, the minimisation of the reduced c2(P) by SA was performed using a value of 5 for the initial ‘temperature’ parameter (Dc2 cur), Ntot = 5000, and f1 = f2 = 0.1. The program was mounted on a dual Pentium 100 MHz PC running under Windows NT. Final structure refinements were performed using the Rietveld procedure included in the GSAS program package.24 5.1 3-Hydroxy-2-naphthoic acid The structure of 3-hydroxy-2-naphthoic acid C10H6CO(OH)2 was solved originally using single crystal data.25 In this section we discuss its solution using powder data as a test of the constrained SA approach. The asymmetric unit consists of a single molecule which is Fig. 1 (a) The 3-hydroxy-2-naphthoic acid molecule (hydrogen atoms are not shown) in a local Cartesian frame. (b) Experimental (1) and calculated (———) powder diVraction patterns, the latter from the randomly chosen initial structural model shown in Fig. 2(a). The background has been subtracted.J. Chem. Soc., Dalton Trans., 1998, 4071–4080 4075 shown in Fig. 1(a). Aromaticity imposes planarity on the fused rings hence C1 to C11 and O3 are in a plane which is not necessarily the case for O1 and O2. These two features were taken into account when establishing the stereochemical description of the molecule in a local Cartesian frame with the origin located on the C1 atom [see Fig. 1(a)].On the reasonable assumption that the hexagonal angles are ideal, the coordinates of the C2–C11 atoms are readily expressed via the C–C bond lengths (lCi,Cj) using trivial geometrical relationships. The coordinates of O3 are calculated via lC9,O3 and fC10,C9,O3 as variable parameters with the values of tC2,C1,C10,C9 and tC1,C10,C9,O3 being fixed at 0 and p, respectively, to maintain the oxygen in the plane. The coordinates of O1 and O2 are expressed in a similar fashion with the only diVerence being that tC2,C1,C10,C11 alone is fixed at p while the corresponding torsion angles, tC1,C10,C11,O1 and tC1,C10,C11,O2, are treated as variable parameters together with lC11,O1, fC10,C11,O1 (in the case of O1) and lC11,O2, fC10,C11,O2 (O2).The following additional assumptions were made to reduce the number of variables during the minimisation procedure lC9,O3 = lC11,O1 = lC11,O2 º lC–O, fC10,C9,O3 = fC10,C11,O1 = fC10,C11,O2 º fC–C–O, and all lCi,Cj º lC–C.Together with Q, F, Y, xC1, yC1, zC1, and B the total number of parameters which must be varied to generate random but chemically plausible structural models was 12. The initial position and orientation of the molecule in the asymmetric unit was chosen at random giving the structure shown in Fig. 2(a) which oVered a poor match to the experimental powder pattern [Fig. 1(b)]. After running the SA program during which the value of the figure-of-merit function calculated for each trial structure varied as shown in Fig. 3(a), the molecule ‘froze’ in the asymmetric unit at Dc2 cur = 0.035 after over 140,000 trial structures and produced the structural model shown in Fig. 2(b) which yielded a significantly better fit to the experimental pattern [Fig. 3(b)]. Subsequent Rietveld refinement further improved the quality of the fit, c2 = 1.70, [Fig. 3(c)] without changing significantly the appearance of the structure [Fig. 2(c)] which is in close agreement with that determined from the single-crystal data [Fig. 2(d)]. Fig. 2 Structural models of 3-hydroxy-2-naphthoic acid: (a) randomly chosen model as a starting point for the structure solution by the SA model; (b) after the SA run; (c) after structure refinement of (b) using the Rietveld method (hydrogen atoms are not shown); (d) as determined by the single-crystal-diVraction method (hydrogen atoms are not shown). 5.2 Poly(ethylene oxide)–salt complexes Complexes of this type are composed of salts e.g.LiCF3SO3 dissolved in solid high-molecular-weight polymers, e.g. poly- (ethylene oxide) (PEO). The polymer is a continuous linear chain with the repeat unit (CH2–CH2–O). Previous studies indicate that in complexes with ethylene oxide–salt ratios 3 : 1 and 4 : 1 the polymer chain adopts a helical conformation 26–28 while in the case of complexes with a 1 : 1 ratio the chain forms a stretched zigzag conformation.29,30 The cations are coordinated by the oxygens of the chain, due to their strongly donating lone pairs, and the oxygens of the anion due to their partial negative charges. 5.2.1 (PEO)3–LiN(SO2CF3)2. Similarity between the lattice parameters of (PEO)3–LiN(SO2CF3)2 (a = 12.034 Å, b = 8.660 Å, c = 19.139 Å, b = 128.58) and (PEO)3–LiCF3SO3 (a = 10.064 Å, b = 8.613 Å, c = 16.77 Å, b = 121.08) 27 suggested that the orientation and conformation of the PEO chain in (PEO)3– LiN(SO2CF3)2 may be similar to that found in (PEO)3– LiCF3SO3, where the helical axis is parallel to the b axis and coincides with the 21 screw axis.Despite this, all attempts to refine the structure of (PEO)3–LiN(SO2CF3)2 on the basis of the known structure of (PEO)3–LiSO3CF3, adjusted to the new dimensions of the unit cell, failed, as did attempts to solve the structure by approaches based on direct methods and diVerence-Fourier synthesis. Density measurements suggested the presence of one formula unit in the asymmetric unit of the cell. Initially a SA Fig. 3 (a) Variation of the figure-of-merit function during the structure solution of 3-hydroxy-2-naphthoic acid. (b) Observed (1) and calculated (———) powder diVraction patterns, based on the structural model shown in Fig. 2(b). The background has been subtracted. (c) Observed, calculated and diVerence powder diVraction patterns of 3-hydroxy-2-naphthoic acid after refinement by the Rietveld method (background included). Corresponds to the structural model shown in Fig. 2(c).4076 J.Chem. Soc., Dalton Trans., 1998, 4071–4080 run was performed using fixed crystallographic coordinates for the atoms of the three EO units comprising the PEO chain taken from the (PEO)3–LiCF3SO3 crystal structure but modi- fied to account for the diVerent unit cell dimensions. The Li1 cation was placed inside the helix and its coordinates were allowed to vary. The coordinates of the atoms comprising the imide anion N(SO2CF3)2 2 [Fig. 4(a)] were expressed in terms of stereochemical descriptors.The covalently bonded sequence F1–C1–S1–N–S2–C2–F4 was treated as a chain with the coordinates of the constituent atoms determined using the values of the consecutive bond lengths, bond angles, and torsion angles. Atoms F2, F3, O1, O2, O3, O4, F5, F6 were positioned by rotating C1–F1, S1–N, S2–C2, and C2–F4 bonds. The total number of structural parameters needed to define the structure was reduced from 55 to 24 by invoking the approximation that in the imide anion all bond lengths of a given bond type (e.g.all C–F or S–O bonds) are equal, all bond angles of a given type (e.g. all S–C–F or C–S–O angles) are equal, and all like torsion angles (N–S1–C1–F1 and N–S2–C2–F4) are equal. The set of parameters P used to calculate c2(P) included xLi, yLi, zLi, xC1, yC1, zC1, Qimide, Fimide, Yimide, lC–F, lS–C, lS–O, lS–N, fC–S–O, fS–C–F, fO–S–O, fF–C–F, fC–S–N, fN–S–O, fS–N–S, tN–S–C–F, tC1,S1,N,S2, tS1,N,S2,C2, and B. The SA run analysed ª200,000 trial chemically plausible structural models. Approximately 15,000 of these Fig. 4 (a) Imide anion in a local Cartesian frame. (b) Observed (1), calculated (———) and diVerence X-ray powder diVraction patterns for (PEO)3–LiN(SO2CF3)2 after refinement and following the SA run with fixed coordinates for the atoms belonging to the PEO chain. The insert shows an expansion of the region from 24 to 408 in 2q. were accepted while the rest were rejected either on the basis of the test for closest approach or by the Metropolis algorithm.The structural model, frozen at Dc2 cur = 0.02, after subsequent refinement gave the fit to the experimental pattern shown in Fig. 4(b). Although the quality of fit is reasonably good, a noticeable mismatch is clearly seen in the inset of Fig. 4(b) indicating that the model is still inadequate. To tackle the problem of the unsatisfactory fit the SA procedure was revisited, this time with extra flexibility added to the structural model.The polymer chain was allowed to vary its position and conformation in addition to the set of parameters involved in the first run [Fig. 5(a)]. Such a description added 12 parameters to the P set: xC7, yC7, zC7, QPEO, FPEO, YPEO, tO5,C3,C4,O6, tC3,C4,O6,C5, tC4,O6,C5,C6, tO6,C5,C6,O7, tC5,C6,O7,C7, tC6,O7,C7,C8. The total number of rejected and accepted trial con- figuration at each ‘temperature’ was chosen to be Ntot = 7000 while the initial value of Dc 2 cur was set to 0.5.Over 100,000 random structural models were generated with only 861 being accepted. Approximately 90% of the rejected trial models were discarded on the grounds of breaking the continuity of the PEO chain at the junctions of neighbouring asymmetric units. The best structural model was used in a new refinement which gave an excellent fit to the observed pattern [Fig. 5(b)]. Apart from a diVerent chain conformation [Fig. 6(a)], the second SA run has revealed a diVerent conformation for the SO2CF3 Fig. 5 (a) A fragment of the PEO chain in a local Cartesian frame. (b) Observed (1), calculated (———) and diVerence X-ray powder diVractions patterns for (PEO)3–LiN(SO2CF3)2 after refinement and following the SA run in which the position and conformation of the PEO chain were varied. The insert shows an expansion of the region from 24 to 408 in 2q.J. Chem. Soc., Dalton Trans., 1998, 4071–4080 4077 fragments of the imide group involving rotation about the S2– N bond [Fig. 6(b)], which did not appear during the first run with the chain fixed and could not be established in the course of the first refinement by the Rietveld method. The final structure of (PEO)3–LiN(SO2CF3) is shown in Fig. 7. Similar to other 3 : 1 complexes, one cation is located in each turn of the PEO helix and is coordinated by oxygen atoms. In this case three of the five oxygens coordinating Li1 are from the chain while the other two belong to two imide anions.Each imide bridges neighbouring Li1 ions along the chain by donating one oxygen to each Li1. Note that both oxygens come from the same half of the imide, the other SO2CF3 moiety is not involved in coordination. Further discussion of the structure and details of the final refinement are given in ref. 31. 5.2.2 PEO–NaCF3SO3. Based on the observed density and unit cell volume, the asymmetric unit of PEO–NaCF3SO3 comprises a single EO unit, sodium cation, and triflate, CF3SO3 2, anion.The stereochemical description of the anion coincides with that of the NSO2CF3 moiety [see Fig. 4(a)] of the imide anion but with the nitrogen atom substituted by an oxygen. The initial trial structure for the SA run [Fig. 8(a)] was chosen at random and, as can be seen from Fig. 8(b), did not provide a match between the calculated and observed diVraction patterns. During the minimisation 27 parameters were varied simultaneously with all bond lengths and bond angles associated with particular bond types in the triflate set to be equal.The full list of variable parameters included, for the triflate: xS, yS, zS, Qtrif, Ftrif, Ytrif, lC–F, lS–C, lS–O, fC–S–O, fS–C–F, Fig. 6 (a) The PEO chain from the structural models corresponding to the fits shown in Fig. 4(b) (left) and Fig. 5(b) (right). (b) A single imide anion from the structural models corresponding to the fits shown in Fig. 4(b) (left) and Fig. 5(b) (right). Fig. 7 Left, a portion of the PEO3–LiN(SO2CF3)2 structure showing a single polymer chain with associated ions. Right, view of the structure down the fibre axis. Light blue spheres, lithium; dark blue, nitrogen; yellow, sulfur; green, carbon; red, oxygen; purple, fluorine. fO–S–O, fF–C–F, tO1,S,C1,F1; for the polymer chain: xC2, yC2, zC2, QPEO, FPEO, YPEO, lC2,C3, lO4,C2, fO4,C2,C3; and for the sodium ion: xNa, yNa, zNa; B. This constrained SA run produced a structural model [Fig. 9(a)] with a continuous PEO chain along the shortest cell axis giving a reasonable profile fit after subsequent refinement by the Rietveld method [Fig. 9(b)]. However all attempts to improve the fit further by refinement failed leaving the best c2 equal to 6 and a noticeable misfit in the 2q range from 33 to 558 [see insert in Fig. 9(b)]. The refined model placed fluorines rather than the more negatively charged oxygens of the triflate anion adjacent to the Na1 cation and did not ensure coordination of the sodiums by the chain oxygens [see Fig. 9(a)]. In addition, the separation of adjacent Na1 ions was only 3.26 Å which is unlikely given the cation radius of 1.02 Å. Negative values of the thermal parameter B for some of the atoms provided further evidence indicating the inappropriateness of the structural model. Successful structure determination was achieved only after removing the constraint that all like bond lengths and bond angles in the triflate were equal. A new SA minimisation was performed allowing all such lengths and the angles to vary independently.During this run 37 structural parameters were varied in a random fashion to generate the trial structures but with the imposition of chain continuity. The structural model obtained after further refinement [Fig. 10(a)] revealed sixfold coordination of the Na1 ion by equidistant oxygens from the triflate and the chain and provided an excellent match between Fig. 8 (a) Randomly chosen initial trial structure of PEO–NaCF3SO3 used in the SA minimisation.Black spheres, sodium; the rest of the atom colours are the same as in Fig. 7. (b) Observed (1) and calculated (———) powder diVraction patterns of PEO–NaCF3SO3, the latter based on the above structural model.4078 J. Chem. Soc., Dalton Trans., 1998, 4071–4080 the observed and calculated patterns with c2 = 1.1 [Fig. 10(b)] and all B values positive. The deleterious eVect of averaging the bond lengths and angles of the triflate on the structure solution is dramatic and could not have been anticipated in advance since the same constraint did not negate location of the internal configuration and position of the imide ion (see Section 5.2.1) with almost twice as many like bond lengths and bond angles set to be equal.Nevertheless the distribution of the scattering power among the constituent atomic species in the case of PEO–NaCF3SO3 was such that a random search using the constrained model was biased from the start and could not yield the correct solution.As an illustration, Fig. 11 shows the significant change in the appearance of the calculated diVraction pattern after averaging the bond lengths and bond angles of the triflate in the final structural model. Further computational details and discussion of the structure may be found in ref. 15. Fig. 9 (a) Refined structural model of PEO–NaCF3SO3 after the SA run with all like bond lengths and bond angles in the triflate ion treated as equal.Solid lines connect the Na1 cation to its nearest neighbours. (b) Observed (1), calculated (———) and diVerence powder diVraction profiles for the above structural model of PEO–NaCF3SO3. 6 Conclusion We have demonstrated that by combining the SA method with a geometrical description, which defines the atomic positions in terms of bond lengths and angles, it is possible to determine, ab initio, relatively complex crystal structures of flexible molecules without the need for single crystals.In crystallography, the solution of structures containing diVerent atoms of comparable scattering power (e.g. C, N, O) is often regarded as presenting the greatest challenge. This task is readily tackled for molecular structures by the SA approach. Recently it has been suggested that instead of calculating a complete powder diVraction pattern for each random structure and fitting it directly to the observed pattern, this process could be divided into two.16,32 Stage 1 involves fitting the observed pattern with a series of individual peaks thus yielding a set of integrated intensities.In stage 2 the randomly generated struc- Fig. 10 (a) Refined structural model of PEO–NaCF3SO3 after the SA run in which all like bond lengths and bond angles in the triflate ion were varied independently. Solid lines connect the Na1 cation to its nearest neighbours. (b) Observed (1), calculated (———) and diVerence powder diVraction profiles for the above structural model of PEO–NaCF3SO3.J.Chem. Soc., Dalton Trans., 1998, 4071–4080 4079 tures are tested against the set of intensities rather than the entire profile. Molecular structures, with up to 10 degrees of freedom (torsion angles), were solved using SA minimisation of the figure-of-merit function based on integrated intensities.32 The advantage of this approach is that it oVers a significant reduction of over 100 fold in the computational time.Typically the full profile SA method requires between 24 and 48 hours on a dual 100 MHz Pentium PC. As is frequently the case in modern crystallography, the computational eYciency is much greater than that of the other essential steps in the process of structure determination. Often the time taken to prepare the compound, collect high quality data, index the powder pattern and write the paper, considerably exceeds the computation time! In the case of relatively complex structures, peak overlap is likely to be severe and there has been considerable debate whether, in such circumstances, the fashion in which groupoverlapped intensities are dealt with in the two-stage method is valid or leads to a loss of information compared with fullprofile fitting.It is of course in the structure solution of such complex compounds that the increased computational eYciency would be particularly advantageous. Perhaps the most important question which arises in constrained SA is the degree of flexibility which must be introduced into a molecule in order to determine its crystal structure using either full profile fitting or the two stage process.If the structure is permitted to be fully flexible (i.e. all bond lengths, bond angles and torsion angles in addition to positional and orientational parameters of each fragment in the asymmetric unit are independent variables) then it is possible to reach the global minimum corresponding to the best possible fit to the data.When the asymmetric unit of the structure under determination consists of a single isolated molecule for which interatomic connectivity is well-established and the bond lengths and angles are particularly well defined, such as in the case of 3-hydroxy- 2-naphthoic acid (Section 5.1) or the examples given in refs. 3, 10, 11, 16 and 32, then only a few variables are required in the SA minimisation in order to solve the structure.The required parameters are those defining the position and orientation of the molecule as well as the torsion angles; the bond lengths and angles may be fixed at typical values. However as clearly demonstrated in this paper, such an approach may be insuY- cient in many cases when the asymmetric unit consists of more than a single molecule. If the interactions between separate molecules in the asymmetric unit involve van der Waals force or ionic bonding, then they are largely non-directional and many relative positions of the fragments are possible for any given set of internal conformations for each fragment.As a result there may be many more local minima in the goodness-of-fit function which are suYciently deep to be confused with the global minimum. In such circumstances a final discrimination between dif- Fig. 11 Calculated powder diVraction pattern of PEO–NaCF3SO3 based on atomic coordinates obtained in the final refinement (———).Calculated powder diVraction pattern of the modified PEO–NaCF3SO3 structure by averaging all like bond lengths and bond angles in the triflate ion (1). ferent fits of the calculated and observed data is essential before refinement can be expected to yield the correct structural model. This requires bond lengths and angles to be varied independently in addition to torsion angles. Variation of torsion angles alone will not suYce. The presence of ionic bonding between separate moieties in the structure may be particularly troublesome since such bonding is stronger than van der Waals’ force and can perturb the internal dimensions of covalently bonded moieties compared with the case of two or more neutral molecules. The examples of structure determination we have presented here constitute a particularly severe test of the methodology in the context of these diYculties since the asymmetric unit comprises several independent moieties which interact both via van der Waals forces and ionic bonding.In the case of (PEO)3–LiN(SO2CF3)2 it was suYcient to set all similar bond lengths, bond angles and torsion angles in the imide anion as single variables while changing only the conformation, i.e. the torsion angles, of the polymer chain with fixed bond lengths and angles. A structural model which was obtained using similar constraints during the solution of PEO–NaCF3SO3 was very diVerent from the true model despite giving a reasonably good fit to the experimental data.The correct model was found only when all stereochemical restraints were removed and all the parameters were varied in a random fashion. Of particular interest to those wishing to exploit the constrained SA approach, is the level of structural complexity that may be tackled. The successful structure solution of compounds requiring the independent variation of 37 parameters and up to 25 symmetry-unrelated atoms has been demonstrated in this paper.All possible structural variables (position, orientation, bond lengths, bond angles and torsion angles) are included. The SA method does not discriminate between diVerent types of structural variables. We could have selected a system in which all 37 parameters were torsion angles (had attention been confined to a compound with a single molecule in the asymmetric unit). In fact the upper limit for structure solution by whole pattern fitting using the SA minimisation combined with stereochemical description should be comparable to the limit of structural complexity found for Rietveld refinement which to date is about 60 non-symmetry related atoms. A better measure of complexity is the number of structural variables and the limit in this context should be around 200.Of course these numbers can only be taken as a rough guide since, as these limits are approached, the level of structural complexity that may be tackled will depend on the individual system. SA has been applied previously in the case of energy minimisation.It is important to appreciate that no energy parameterisation of the system is involved here, i.e. there is no requirement for interatomic potentials. Minimisation is carried out with the use of the observed diVraction data alone. 6 Acknowledgements The authors are grateful to Dr P. Lightfoot for his collaboration during the structure solution of (PEO)3–LiN(SO2CF3)2 and for useful discussions. Thanks are due to Mr G.S. MacGlashan and Mr L. J. M. Sawers. The financial support of this work by the EPSRC and The Leverhulme Trust is gratefully acknowledged. 7 References 1 A, Altomare, M. C. Burla, M. Camalli, G. Cascarano, C. Giacovazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori and R. Rizzi, Mater. Sci. Forum, 1998, 278, 284. 2 M. J. Buerger, Vector space, Wiley, New York, 1959, pp. 167–168. 3 K. D. M. Harris and M. Tremayne, Chem. Mater., 1996, 8, 2554. 4 H. M. Rietveld, J.Appl. Crystallogr., 1969, 2, 65. 5 S. Kirkpatrick, C. D. Gelatt and M. P. Vecchi, Science, 1983, 220, 671.4080 J. Chem. Soc., Dalton Trans., 1998, 4071–4080 6 J. H. Holland, Adaptation in Natural and Artificial Systems, The University of Michigan Press, Ann Arbor, 1975. 7 B. M. Kariuki, H. Serrano-Gonzalez, R. L. Johnston and K. D. M. Harris, Chem. Phys. Lett., 1997, 280, 189. 8 L. Ingber and B. Rosen, Math. Comput. Modell., 1992, 16, 87. 9 J. M. Newsam, M. W. Deem and C. M. Freeman, Accuracy in Powder Diffraction II, NIST Special Publication No. 846, 1992, pp. 80–91. 10 K. D. M. Harris, M. Tremayne, P. Lightfoot and P. G. Bruce, J. Am. Chem. Soc., 1994, 116, 3543. 11 K. Shankland, W. I. F. David, T. Csoka and L. McBride, Int. J. Pharm., 1998, 165, 117. 12 Y. G. Andreev, P. Lightfoot and P. G. Bruce, J. Appl. Crystallogr., 1997, 30, 294. 13 International Tables for Crystallography, Kluwer Academic Publishers, Dordrecht, Boston, London, 1995, vol. C, pp. 594–608. 14 N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller and E. Teller, J. Chem. Phys., 1953, 21, 1087. 15 W. H. Press, S. A. Teukolsky, W. T. Vetterling and B. P. Flannery, Numerical Recipes in C: the Art of Scientific Computing, Cambridge University Press, 1992, pp. 451–455. 16 W. I. F. David, K. Shankland and N. Shankland, Chem. Commun., 1998, 931. 17 G. S. Pawley, J. Appl. Crystallogr., 1981, 14, 357. 18 A. Le Bail, H. Duroy and J. L. Fourquet, Mater. Res. Bull., 1988, 23, 447. 19 C. Baerlocher, in The Rietveld Method, IUCr Monographs on Crystallography 5, ed. R. A. Young, Oxford University Press, 1993, pp. 186–196. 20 S. Arnott and A. J. Wonacott, Polymer, 1966, 7, 157. 21 International Tables for X-ray Crystallography, Kynoch Press, Birmingham (Present distributor Kluwer Academic Publishers, Dordrecht), 1959, vol. 2, pp. 62–63. 22 Y. G. Andreev, T. Lundström and N. I. Sorokin, Nucl. Instrum. Methods Phys. Res., Sect. A, 1995, 354, 134. 23 Y. G. Andreev, G. S. MacGlashan and P. G. Bruce, Phys. Rev. B, 1997, 55, 12011. 24 A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory Report Number LA-UR-86-748, 1987. 25 M. P. Gupta and B. P. Dutta, Cryst. Struct. Commun., 1975, 4, 37. 26 Y. Chatani and S. Okamura, Polymer, 1987, 28, 1815. 27 P. Lightfoot, M. A. Mehta and P. G. Bruce, Science, 1993, 262, 883. 28 P. Lightfoot, J. L. Nowinski and P. G. Bruce, J. Am. Chem. Soc., 1994, 116, 7469. 29 M. Yokoyama, H. Ishihara, R. Iwamoto and H. Tadokoro, Macromolecules, 1990, 2, 184. 30 Y. Chatani, Y, Fujii, T. Takayanagi and A. Honma, Polymer, 1990, 31, 2238. 31 Y. G. Andreev, P. Lightfoot and P. G. Bruce, Chem. Commun., 1996, 2169. 32 K. Shankland, W. I. F. David and T. Csoka, Z. Kristallogr., 1997, 220, 550. Paper 8/05437A

ISSN:1477-9226    

DOI:10.1039/a805437a

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 4081–4083 4081 Solid state coordination chemistry of the copper cyanide– organoamine system: hydrothermal synthesis and structural characterization of [{Cu2(bpy)2(CN)}2Cu5(CN)7]?0.17H2O Douglas J. Chesnut, Anakarin Kusnetzow and Jon Zubieta * Department of Chemistry, Syracuse University, Syracuse, NY 13244, USA Received 18th August 1998, Accepted 27th October 1998 The hydrothermal reaction of CuCN, KCN, 2,29-bipyridyl (bpy) and water yields [{Cu2(bpy)2(CN)}2Cu5(CN)7]? 0.17H2O 1, a two-dimensional copper–cyano network intercalated by [(bpy)2(Cu)2(CN)]1 cations and water of crystallization.The diversity of properties associated with solid state inorganic materials, which gives rise to applications ranging from heavy construction to integrated optical systems,1 provides a context for the significant contemporary interest in design and synthesis of new structural types. Such materials may exhibit significant architectural complexity and potentially functionality, as a consequence of the incorporation of the inorganic constituent into a hierarchical structure,2,3 evolving from the synergistic interaction between the organic and inorganic substructures.Manipulation of the organic–inorganic interface in the hydrothermal domain provides a powerful synthetic approach for the manipulation of solid phase microstructures, a strategy which has proved fruitful in the preparation of four families of oxide based phases for which organic materials provide structurally important components: zeolites,4 mesoporous oxides of the MCM-41 type,5 microporous octahedral–tetrahedral framework transition metal phosphates 6 and organically templated molybdenum oxides.7 Biomineralized materials represent the fifth class of organically templated oxide phases.8 We have recently extended this concept of templating anionic networks with organic constituents to the copper halide system, represented by [{Cu(en)2}2Cu7Cl11],9 and the copper–cyano system, exemplified by [{Cu2(bpy)2(CN)}Cu5(CN)6] 2, which was synthesized during preliminary investigations of the hydrothermal reactions of the copper–cyano]organoamine system.10 Minor variations in synthetic conditions, such as the addition of KCN, resulted in the synthesis of a new phase, [{Cu2(bpy)2- (CN)}2Cu5(CN)7]?0.17H2O 1, as well as other copper–cyano– organoamine phases to be reported elsewhere.Compound 1† was prepared as clear dark red crystals in ca. 40% yield from the reaction of CuCN, KCN, bpy and water in the mole ratio 1 : 0.23 : 0.22 : 63 at 180 8C and autogenous pressure for 72 h. Orange and yellow solid phases were also observed as reaction products. As shown in Fig. 1a, the structure of 1 ‡ exhibits anionic two-dimensional nets of alternating fused rows of pseudohexagonal {Cu6(CN)6} rings and six-sided {Cu8(CN)8} rings. The 18-membered {Cu6(CN)6} rings, which extend ca. 10.1 Å from vertex to opposing vertex, contain exclusively distorted trigonal planar copper atoms at the vertices linked by disordered cyano groups. The 24-membered {Cu8(CN)8} rings contain disordered cyano groups, linking distorted trigonal planar copper atoms, as well as two linear two-coordinate copper atoms disposed on either elongated edge of the heterocycle. The {Cu8(CN)8} rings have dimensions of ca. 8.8 × 15 Å. The individual rings of both types are puckered which results in the slight folding of an entire layer, illustrated in Fig. 1b. The distortion of the anionic network rings of compound 1 is not as pronounced as was previously observed for compound 2, as shown in Fig. 2. The layers of compound 1 form two distinct interlamellar regions with spacings of 6.5 Å (region A) and 7.0 Å (region B), respectively. The interlamellar regions of compound 1 contain not only the intercalated cations in two diVerent conformations per layer (Fig. 3) but region A contains the water of crystallization as well, giving rise to an ABBABB repeat pattern shown in Fig. 4a.Two sets of cations are disposed such that those of diVerent regions but similar conformation form columnar arrays (Type I) within each region. Each bpy unit of a cation overlaps a bpy group of an adjacent cation in the same columnar array, as illustrated in Fig. 4b. The columns propagate parallel to the crystallographic c-axis. The cations of the second conformation form arrays (Type II) exhibiting overlap of single pyridyl rings of bpy groups of adjacent cations arranged in a zig-zag array propagating parallel to the cationic arrays of Type I within the interlamellar regions (Fig. 4b). At no point do the cations of either type interpenetrate the rings of the anionic layers. The cations composing the Type I arrays are similar in conformation to the cations previously observed for compound 2. Cations of this conformation in compound 1 are observed to Fig. 1 (a) A view down the crystallographic a-axis of 1 perpendicular to a single anionic layer showing the 18- and 24-member rings; (b) an oblique view to the anionic layer showing the distortion of the rings and layer. Black spheres represent cyano groups and gray spheres represent Cu atoms.4082 J. Chem. Soc., Dalton Trans., 1998, 4081–4083 have bpy units centered above and below the planes of the {Cu8(CN)8} rings of the anionic networks. The bpy components of the cations are separated from the anionic layers by distances greater than sum of the van der Waals radii.In contrast, the planes of the rings of the cations composing the zigzag arrays are no longer parallel to the anionic network so as to avoid repulsive interactions of the anionic and cationic substructures by maintaining van der Waals contact distances of ca. 3.5 Å.11 The Type II cations are subsequently significantly distorted, Fig. 4, from the conformation of the Type I cations.The structure of compound 1 demonstrates the power of hydrothermal synthesis in the preparation of organic–inorganic composite materials. Not only are diVerential solubility prob- Fig. 2 (a) A view perpendicular to the anionic layer of 2; (b) a view oblique to the anionic layer showing the extreme distortion from planarity of the rings and the resulting “ruZing” of the layer. Black spheres represent cyano groups and gray spheres represent Cu atoms. Fig. 3 A view of a cation of Type II approximately parallel to the {Cu–CN–Cu} axis.The deviation of the bpy units from the parallel conformation described for compound 2 by ca. 458 is apparent (Cu atoms and C atoms of the bpy and the CN groups are represented by black and dark gray spheres respectively; N atoms of the bpy units are represented by light gray spheres). lems avoided, reducing the tendency to phase segregate, but structurally more complex metastable phases are favored.12 Compounds 1 and 2 also illustrate the rich diversity of structures that can be achieved with only minor variations of reaction conditions.In general, the structural complexity reflects the versatility of the organonitrogen component which may function as counter ion, ligand to the copper cyanide backbone or ligand in a coordination complex ion.10 The role of the organic ligand as a component of a coordination complex cation, a recurring theme in metal oxide–organoamine chemistry, has now been extended to the metal halide 13 and metal pseudohalide phases,10,14 where it is manifested in the large ring sizes of the copper cyanide networks which are required to accommodate the large binuclear cations in compound 2.In addition, compound 1 demonstrates that such cations are non-rigid and capable of displaying a variety of conformations. These cations not only serve a charge compensating role but also participate in a synergistic fashion with the [Cu5(CN)7]22 framework in generating an inorganic microstructure based on both the {Cu6(CN)6} rings and the larger {Cu8(CN)8} rings.This observation suggests that cation dimensions in such structures are not determined by ring size, topology or interlamellar spacing of the anionic substructure, but rather function as agents for the kinetically controlled formation of a variety of structures. Fig. 4 (a) A view perpendicular to the crystallographic ab-plane showing the layers with (region A) and without water of crystallization (region B); (b) a view perpendicular to the crystallographic bc-plane showing the disposition of adjacent cations, their disposition relative to the rings of the anionic layer and the cationic arrays of Type I and Type II which are separated by red bars (Cu atoms, cyano groups, C atoms of the bpy, N atoms of the bpy and waters of crystallization are represented by purple, yellow, black, light blue and red spheres respectively).J.Chem. Soc., Dalton Trans., 1998, 4081–4083 4083 Acknowledgements This work was supported by NSF Grant CHE 9617232.Notes and references † IR spectra (cm21): 3436, (br, O–H stretch), 3122, 3091, 3055, 2778, 2688, 2612, 2545, 2473, 2428, 2119, (s, C–N stretch); 1725; 1596; 1501; 1282; 1143. TGA studies under an oxygen atmosphere show loss of the water of crystallization at ca. 165 8C and the successive loss of the organic components between 170 8C and 390 8C. Satisfactory C, H, N analyses were obtained.‡ Crystal data for 1: C73.5H48.5Cu13.5N25.5O0.25, monoclinic, P2/c, a = 21.6927(1), b = 19.7886(2), c = 17.5075(2) Å, b = 98.12(1)8, V = 7440.07(1) Å3, Z = 4, M = 2150.67, Dc = 1.920 Mg m23; R1 = 0.0458; wR2 = 0.1273 (13015 reflections), T = 150 K, m = 3.834 mm21. CCDC reference number 186/1221. 1 A. K. Cheetham, Science, 1994, 264, 794. 2 S. Mann, J. Chem. Soc., Dalton Trans., 1997, 3953. 3 S. Mann and G. A. Ozin, Nature (London), 1996, 382, 313. 4 J. V. Smith, Chem.Rev., 1988, 88, 149; M. L. Occelli and H. C. Robson, Zeolite Synthesis, American Chemical Society, Washington, DC, 1989. 5 C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature (London), 1992, 359, 710. 6 R. C. Haushalter and L. A. Mundi, Chem. Mater., 1992, 4, 31; M. I. Khan, L. M. Meyer, R. C. Haushalter, C. L. Schweitzer, J. Zubieta and J. L. Dye, Chem. Mater., 1996, 8, 43. 7 D. Hagrman, C. Zubieta, D. J. Rose, J. Zubieta and R. C. Haushalter, Angew. Chem., Int. Ed. Engl., 1997, 36, 873. 8 S. Mann, Nature (London), 1993, 365, 499. 9 J. R. D. DeBord, Y. Lu, C. J. Warren, R. C. Haushalter and J. Zubieta, Chem. Commun., 1997, 1365. 10 D. J. Chesnut and J. Zubieta, Chem. Commun., 1998, 1707. 11 U. Müller, Inorganic Structural Chemistry, John Wiley & Sons, New York, 1993, p. 101. 12 J. Gopalakrishnan, Chem. Mater., 1995, 7, 1265. 13 Y. Zhang, J. R. D. DeBord, C. J. O’Connor, R. C. Haushalter, A. Clearfield and J. Zubieta, Angew. Chem., Int. Ed. Engl., 1996, 35, 989. 14 A. J. Blake, N. R. Champness, M. Crew, L. R. Hanton, S. Parsons and M. Schöder, J. Chem. Soc., Dalton Trans., 1998, 1533. Communication 8/06500D

ISSN:1477-9226    

DOI:10.1039/a806500d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 4085–4086 4085 Organically templated inorganic/organic hybrid materials: hydrothermal synthesis and structural characterization of [C4H12N2][In2(C2O4)(HPO4)3]?H2O Yuh-Feng Huanga and Kwang-Hwa Lii *ab a Institute of Chemistry, Academia Sinica, Taipei, Taiwan 115, R.O.C. b Department of Chemistry, National Central University, Chungli, Taiwan 320, R.O.C. E-mail: liikh@cc.ncu.edu.tw Received 21st September 1998, Accepted 2nd November 1998 The synthesis and characterization of a novel inorganic/ organic hybrid material is described; the 3-D framework consists of InO6 octahedra which are linked by inorganic phosphate and organic oxalate anions to generate intersecting tunnels in which piperazinium cations and water molecules reside.Recently considerable attention has been focused on the organically templated metal phosphates because of their rich structural chemistry and potential applications in heterogeneous catalysis, adsorption, and ion exchange. While phosphate structures constructed from aluminium and gallium represent the most well developed classes of materials, hydrothermal chemistry of phosphates containing transition metals 1–6 or the heavier Group 13 element, In,7–9 has also yielded a large variety of new structures.Another approach to microporous materials has exploited appropriate metal centers that are linked through suitable multidentate organic ligands.10–14 The structures of these coordination polymers have the potential for more precise rational design, through control of the shape, size and functionalization of the pores.As part of our general investigations of the synthesis of oxide phases, we have sought to incorporate both inorganic and organic anions into the framework to provide a new route to open-framework materials. This expectation has been realized in the synthesis and characterization of the organically templated three-dimensional solid [C4H12N2][In2- (C2O4)(HPO4)3]?H2O 1.The material is original in the sense that it is the first time that inorganic phosphate and organic oxalate anions, associated with Group 13 elements, belong to the skeleton of an open framework structure templated with an amine. A few compounds were recently published which had the characteristics of compound 1 containing two anions with metals, namely, the tin compound Sn(O3PCH3)(C2O4) 15 and the rare earth compounds [Ln(H2O)2]2(C2O4)2(CO3)?2.5H2O and [Ln(H2O)2]2(C2O4)(CO3)2.16 However, organic templates are not included in these structures.Hydrothermal treatment of In(NO3)3?5H2O (1 mmol), H3PO4 (5 mmol), oxalic acid (4 mmol), piperazine (6 mmol), and water (10 ml) for 3 days at 165 8C yielded a colorless crystalline product. A colorless bladed crystal was used for singlecrystal X-ray diVraction.† The product is pure 1 as judged by comparison of the X-ray powder pattern of the bulk product to that simulated from the coordinates derived from the singlecrystal study.The yield of 1 was almost quantitative based on indium. Chemical analysis confirmed the stoichiometry (Found: C, 10.18; H, 2.26; N, 3.90. Calc. C, 10.12; H, 2.41; N, 3.94). Thermal gravimetric analysis in flowing oxygen showed a mass loss in two steps in the region 30–700 8C and a sequential decomposition above 800 8C. The first step (ª240 8C) shows a mass loss of 2.75%, which corresponds to the loss of guest water (calc. 2.53%). The second step (ª410 8C, mass loss 17.65%) is correlated to the decomposition of oxalate anion, dehydration of HPO4 groups and deprotonation of piperazinium dication. The decomposition at higher temperatures corresponds to the release of the organic component. Powder X-ray diVraction of the decomposition product at 600 8C shows it to be amorphous. The final decomposition product at 950 8C is essentially pure InPO4.17 A detailed thermal decomposition mechanism has not been derived.The three-dimensional anionic framework of 1 consists of InO6 octahedra connected via coordinating HPO4 22 and C2O4 22 anions to form intersecting tunnels parallel to the [100] and [001] directions. Diprotonated piperazinium cations are located at the intersections of these tunnels. The tunnel parallel to [100] has a window formed by the edges of six InO6 octahedra, four phosphate tetrahedra and two oxalate anions (Fig. 1). The other type of tunnel has an eight-membered window formed by four octahedra and four tetrahedra.An alternative way of describing the structure is that it contains puckered layers of indium phosphate parallel to the (001) plane, which are further connected into a three-dimensional framework via bridging oxalate ligands. The indium phosphate layer is shown in Fig. 2. Each indium atom is coordinated by a bidentate C2O4 22 anion and four HPO4 22 anions. The m coordination of the oxalate leads to highly distorted coordination polyhedra for both In(1) and In(2), as indicated by the wide range of In–O bonds [2.05– 2.21 Å for In(1), 2.09–2.24 Å for In(2)] and the small O–In–O bond angles [76.08 for In(1), 75.98 for In(2)] subtended by the oxalate ligand.The oxalate anion acts as a bis-bidentate ligand to In(1) and In(2). Both HP(1)O4 and HP(2)O4 groups coordinate to three In atoms with the hydroxo group being unshared. HP(3)O4 bonds to two In atoms with a pendant P–O unit being strongly H-bonded to the lattice water molecule.The loss of lattice water molecules in TG analysis occurs well above the boiling point of water, which can be attributed to the extensive Fig. 1 Structure of 1 viewed along the [100] direction. Polyhedra with darker and lighter shading are InO6 octahedra and phosphate tetrahedra, respectively. Solid circles, C atoms; stippled circles, N atoms; open circles, water oxygens.4086 J. Chem. Soc., Dalton Trans., 1998, 4085–4086 hydrogen bonding holding the water in the crystal lattice [Ow? ? ? O(10), 2.45; Ow ? ? ? O(8), 2.62; Ow ? ? ? O(9) 2.63 Å].The piperazinium cation is H-bonded to the phosphate oxygens, as inferred from the N ? ? ? O distances [N(1) ? ? ? O(6), 2.83; N(1) ? ? ? O(10), 3.043; N(2) ? ? ? O(3), 2.83; N(2) ? ? ? O(12), 2.91 Å]. The most important feature of the structure of 1 is that it is the first member of a new class of organically templated inorganic/organic hybrid materials.Both phosphate and oxalate anions occur within the same framework. In the structure, large cavities are occupied by C4H12N2 counter cations and waters of crystallization. This new compound provides a new route for open-framework materials. A series of related materials involving transition metals have been prepared. Also, it may be possible to replace oxalate with extended analogues such as squarate (3,4-dihydroxy-3-cyclobutene-1,2-dionate) and 1,3,5-benzenetricarboxylate.Further work on this theme is in progress. Acknowledgements We thank the Institute of Chemistry, Academia Sinica, the National Science Council, and Chinese Petroleum Corp. of Taiwan for support, and Professor S.-L. Wang and Ms. F.-L. Liao at National Tsing Hua University for X-ray intensity data collection. Notes and references † Crystal data for [C4H12N2][In2(C2O4)(HPO4)3]?H2O: monoclinic, space group P21, a = 6.5052(2), b = 17.5005(2), c = 8.1811(2) Å, Fig. 2 Section of an indium–phosphate layer in 1 viewed along the [001] direction.Oxalate groups are not shown. b = 107.656(1) Å, U = 887.50(5) Å3, Z = 2, Mr = 711.76, Dc = 2.663 g cm23, m(Mo-Ka) = 29.7 cm21, l = 0.71073 Å, graphite monochromator, crystal dimensions 0.12 × 0.05 × 0.01 mm. Of the 3644 unique reflections collected (2qmax = 55.78), 3151 reflections were considered observed [Fo > 4s(Fo)] after empirical absorption correction. Bondvalence calculations 18 indicated that both In atoms are trivalent, O(4), O(6), O(10) and O(11) had valence sums of 1.20, 1.21, 1.27 and 1.11, respectively, and all other oxygen atoms had values between 1.76 and 1.94.A lattice water site, Ow, was located in the structural tunnel. Atoms O(4), O(6) and O(11) are hydroxo oxygens. The valence sum of O(10) can be satisfied by forming a strong hydrogen bond with Ow [O(10) ? ? ? Ow 2.45 Å]. Refinement (271 parameters) was performed by full-matrix least-squares analysis, with anisotropic thermal parameters for all atoms.(Dr)max,min = 0.90, 20.83 e Å23. The reliability factor converged to R1 = 0.0435, wR2 = 0.1012 and S = 0.999. CCDC reference number 186/1228. See http://www.rsc.org/suppdata/dt/1998/4085/ for crystallographic files in .cif format. 1 R. C. Haushalter and L. A. Mundi, Chem. Mater., 1992, 4, 31. 2 M. I. Khan, L. M. Meyer, R. C. Haushalter, A. L. Schweitzer, J. Zubieta and J. L. Dye, Chem. Mater., 1996, 8, 43. 3 M. Cavellec, D.Riou, J. M. Greneche and G. Ferey, Inorg. Chem., 1997, 36, 2187. 4 J. R. D. DeBord, W. M. ReiV, C. J. Warren, R. C. Haushalter and J. Zubieta, Chem. Mater., 1997, 9, 1994. 5 K.-H. Lii and Y.-F. Huang, Chem. Commun., 1997, 1311. 6 K.-H. Lii, Y.-F. Huang, V. Zima, C.-Y. Huang, H.-M. Lin, Y.-C. Jiang, F.-L. Liao and S.-L. Wang, Chem. Mater., 1998, 10, 2599. 7 S. S. Dhingra and R. C. Haushalter, J. Chem. Soc., Chem. Commun., 1993, 1665. 8 A. M. Chippindale and S. J. Brech, Chem. Commun., 1996, 2781. 9 I. D. Williams, J. Yu, H. Du, J. Chen and W. Pang, Chem. Mater., 1998, 10, 773. 10 B. F. Abrahams, B. F. Hoskins, D. M. Michail and R. Robson, Nature (London), 1994, 369, 727. 11 L. R. MacGillivray, S. Subramanian and M. J. Zaworotko, J. Chem. Soc., Chem. Commun., 1994, 1325. 12 G. B. Gardner, D. Venkataraman, J. S. Moore and S. Lee, Nature (London), 1995, 374, 792. 13 O. M. Yaghi, G. Li and H. Li, Nature (London), 1995, 378, 703. 14 S. O. H. Gutschke, M. Molinier, A. K. Powell, R. E. P. Winpenny and P. T. Wood, Chem. Commun., 1996, 823. 15 B. Adair, S. Natarajan and A. K. Cheetham, J. Mater. Chem., 1998, 8, 1477. 16 S. Romero, A. Mosset and J. C. Trombe, Eur. J. Solid State Inorg. Chem., 1997, 34, 209. 17 InPO4, file number 8-52. Joint Committee on Powder DiVraction Standards, International Centre of DiVraction Data, Swarthmore, PA. 18 I. D. Brown and D. Altermatt, Acta Crystallogr., Sect. B, 1985, 41, 244. Communication 8/07345G

ISSN:1477-9226    

DOI:10.1039/a807345g

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 4087–4088 4087 A rare uranium(III) complex of a tripodal aromatic amine and its lanthanum analogue Raphaël Wietzke,a Marinella Mazzanti,*a Jean-Marc Latour b and Jacques Pécaut b a Laboratoire de Reconnaissance Ionique, Département de Recherche Fondamentale sur la Matière Condensée, CEA-Grenoble, 17 rue des Martyrs, 38054 Grenoble, Cedex 09, France b Laboratoire de Chimie de Coordination Service de Chimie Inorganique et Biologique, Département de Recherche Fondamentale sur la Matière Condensée, CEA-Grenoble, 17 rue des Martyrs, 38054 Grenoble, Cedex 09, France. E-mail: mazzanti@drfmc.ceng.cea.fr Received 12th October 1998, Accepted 6th November 1998 The crystal structure of the first U(III) complex of the heptadentate tripodal aromatic amine tris[(2,29-bipyridin- 6-yl)methyl]amine and of its La(III) analogue have been determined.While there has been a renewal of interest in the organometallic chemistry of uranium(III),1 the development of the coordination chemistry of U(III) has lagged behind 2 and the number of structurally characterized U(III) complexes containing N-donor ligands is very limited.3 Aromatic amines are ligands of great interest in the coordination chemistry of f block elements for their potential application in actinide(III)/lanthanide(III) separation, a diYcult problem in nuclear waste disposal.4 Indeed, they have been reported to complex actinides(III) more strongly than lanthanides(III), owing to a greater covalent contribution to the metal–amine bonding.5 While an increasing number of lanthanide complexes of aromatic amines have been reported in the last few years, very little is known about the coordination chemistry of actinides(III) with these ligands.2 The preparation of U(III) complexes of 1,10-phenanthroline and 2,29-bipyridine was reported several years ago by Hart and Tajik, who were unable to characterize the complexes due to their extreme reactivity with traces of oxygen or moisture.6 In order to gain more insight into the chemistry of f elements with nitrogen donor ligands, we have started to study the coordination chemistry of Ln(III) and An(III) with the tripodal aromatic amines.7 The synthesis and some physical properties of Ru(II), Fe(II), Cr(II), Cr(III) 8 and Eu(III) 9 complexes of tris[(2,29-bipyridin-6- yl)methyl]amine (tbpa) have been described, but to our knowledge there are no crystallographic reports on tbpa complexes.In this communication we report the synthesis and the molecular structure of the first U(III) complex of an aromatic amine, [U(tbpa)I2]I?py 1 and those of the related lanthanum(III) complex [La(tbpa)(H2O)(h2-ClO4)][ClO4]2?2CHCl3?MeOH 2. Complex 1 was prepared by reacting UI3(thf)4 10 with tbpa in pyridine under argon;† the replacement of the iodide ion and of the four thf ligands of the starting complex UI3(thf)4 by the heptadentate tbpa results in an increased reactivity towards oxygen and water. The crystal structure of 1 ‡ consists of cation–anion pairs and a molecule of pyridine connected by hydrogen bonding to three diVerent cations and one anion. The N N N N N N N tbpa crystal structure of the cation is shown in Fig. 1. The uranium atom is 9-coordinate with tbpa and two iodine anions with approximately a capped square antiprism coordination geometry. The tertiary amine nitrogen N(7) lies on the expanded face formed by the three bipyridyl nitrogen atoms N(1), N(3), N(5) of tbpa and I(1).The tbpa ligand is bound asymmetrically with U–N distances ranging from 2.697(4) Å for N(7) to 2.595(4) Å for the bipyridyl nitrogens. These values are in the range of U(III)–N distances previously reported.3,11 Complex 2 was obtained by reaction of La(ClO4)3?6H2O with tbpa in acetonitrile.§ The crystal structure of the cation in 2 is shown in Fig. 2. The lanthanum ion is 10-coordinate with tbpa, one water molecule and a bidentate perchlorate.The coordination geometry is not regular. Again, tbpa is asymmetrically bound with distortion of the bipyridine moiety. The average La–N distance (2.728 Å) is longer than the value found for the [La(bipy)2(NO3)3] complex [2.66(1) Å],12 and of those found for the 10-coordinate complexes [La(phen)2(NO3)3] and [La- (terpy)2(NO3)2]1 [2.67(3) and 2.68(3) Å].13 NMR studies in acetonitrile solution show that both complexes are stable toward the dissociation of the organic ligand.The NMR spectrum of the La complex shows only one resonance for the methylene protons and seven overlapping resonances for the bipyridine protons. These features are consistent with a three-fold symmetry of the solution species in which all chelating arms of the tbpa ligand are equivalent or in fast Fig. 1 Crystal structure of the cation of [U(tbpa)I2]I?py 1 with thermal ellipsoids at 30% probability. Selected bond lengths (Å): U–N(1) 2.595(4), U–N(2) 2.635(4), U–N(3) 2.643(3), U–N(4) 2.657(4), U–N(5) 2.614(4), U–N(6) 2.635(4), U–N(7) 2.697(4), U–I(1) 3.2568(8), U–I(2) 3.3286(7).4088 J.Chem. Soc., Dalton Trans., 1998, 4087–4088 exchange on the NMR time scale. In addition, the chemical shift equivalence of the methylene protons requires a conformational mobility of the ligand arms in solution which was not observed for the ruthenium complex of tbpa.8 The proton spectrum of the uranium complex at room temperature in acetonitrile shows 8 broad peaks between d 20 and 210.At lower temperature (230 8C) the spectrum shows a large number (at least 24) of narrower signals. These features can be explained by the presence at room temperature of an exchange between diVerent ligand conformations. The number of well characterized coordination compounds of uranium(III) remains very limited because of the extreme ease with which they oxidize to U(IV). Complex 1 is a rare example of a structurally characterized coordination derivative of U(III) and the first, to our knowledge, containing only an aromatic amine as ancillary ligand.The structural data show that the diVerence between the metal to nitrogen distance and the ionic radius of the metal is essentially the same for La (1.458 Å) and U (1.476 Å) [a value of the ionic radius of 1.163 Å for 9- coordinated U(III) was calculated as described by Raymond and Eigenbrot 14]. In addition the C–N and the C–C distances for the bipy rings appear to be very similar in the La and U complexes.Surprisingly, these results appear to indicate the occurrence of a similar ionic type of bonding between tbpa and U or La [though we note that the crystal structure of the iodide complex of La(III) with tbpa would allow a better comparison]. The coordination properties of iodide complexes of lanthanides( III) and uranium(III) complexes with tripodal aromatic amines having diVerent soft character is currently under study together with their extraction properties.Notes and references † Complex 1. A solution of tbpa (0.033 mmol) in pyridine (1 mL) was added to a blue solution of UI3(thf)4 (0.033 mmol) in pyridine (1 mL) to give a brown solution. After standing for two days at room temperature a brown crystalline precipitate formed (0.023 mmol, 71%). Brown crystals of 1 suitable for X-ray analysis were obtained by slow diVusion of hexane into a pyridine–acetonitrile solution {Calc.for [U(tbpa)I2]I, C33H27I3N7U: C, 34.75; H, 2.39; N, 8.60. Found: C, 34.20; H, 2.60; N, 8.29%}. All manipulations were carried out under an inert argon atmosphere using Schlenk techniques and a Braun glovebox equipped with a puri- Fig. 2 Crystal structure of the cation of [La(tbpa)(H2O)(h2-ClO4)]- [ClO4]2?2CHCl3?MeOH 2 with thermal ellipsoids at 30% probability. Selected bond lengths (Å): La–N(1) 2.704(3), La–N(2) 2.693(3), La– N(3) 2.729(3), La–N(4) 2.762(3), La–N(5) 2.688(3), La–N(6) 2.797(3), La–N(7) 2.723(3), La–O(3) 2.763(3), La–O(1) 2.492(3), La–O(2) 2.671(3).fier unit. The water and oxygen level were always kept at less than 1 ppm. The solvents were purchased in their anhydrous form and distilled from K or CaH2 under argon. Solid or solution samples of 1 were stored in the glovebox in glass vessels sealed with silicon greased stoppers. ‡ Crystal data for complex 1. [U(tbpa)I2]I?Py, C38H32I3N8U: M = 1219.45, monoclinic, P21 /c, a = 9.736(2), b = 17,951(4), c = 22.436(5) Å, b = 95.52(3)8, V = 3903.0(13) Å3, Z = 4, Dc = 2.075 g cm23, m = 6.570 mm21. 6901 Independent reflections (2qmax = 52) were collected at 143 K. Refinement using the SHELXTL 5.05 package on all data converged at R1 [F > 4s(F)] = 0.0543, wR2 = 0.1183. Crystals of 1 very quickly oxidised in air and lost solvent rapidly. To prevent oxidation and solvent loss the crystals were mounted in a capillary tube with some of the crystallisation solvent in the glove box.Crystal data for complex 2. [La(tbpa)(H2O)(h2-ClO4)][ClO4]2? 2CHCl3?MeOH, C36H35N7O14Cl9La: M = 1247.67, triclinic, P1� , a = 11.90850(10), b = 13.3371(2), c = 18.2306(3) Å, a = 69.6470(10), b = 87.5360(10), g = 64.0780(10)8, V = 2421.07(6) Å3, Z = 2, Dc = 1.711 g cm23, m = 1.448 mm21. 6844 Independent reflections (2qmax = 52) were collected at 143 K. Refinement using the SHELXTL 5.05 package on all data converged at R1 [F > 4s(F)] = 0.0353, wR2 = 0.0901.The molecules of MeOH and CHCl3 and one of the perchlorate counter ions are disordered with a site occupancy factor of 0.5. CCDC number 186/1234. See http://www.rsc.org/suppdata/dt/1998/4087/ for crystallographic files in .cif format. § Complex 2. The ligand tbpa (0.23 mmol) was added to a solution of La(ClO4)3?7H2O (0.23 mmol) in acetonitrile (5 mL). The resulting pale yellow solution was stirred for 1 h and then concentrated (1 mL). After standing overnight, a white precipitate formed which was collected by filtration, washed with diethyl ether and dried (0.12 mmol, 52 %). 1H NMR (CD3CN, 200 MHz): d 8.29 (d, J = 4.4, 3H), 8.17 (t, J = 6.6 Hz, 3H), 8.08 (m, 9H), 7.50 (m, 6H), 4.57 (s, 6H, CH2). FAB1: m/z 858 [La(tbpa)(ClO4)2]1 {Calc. for [La(tbpa)(H2O)(h2-ClO4)][ClO4]2? 0.5Et2O, C35H32N7Cl3O12.5La: C, 42.21; H, 3.24; N, 9.85. Found: C, 42.55; H, 3.24; N, 10.07%}. Colourless crystals of 2 suitable for X-ray analysis were obtained by leaving a concentrated solution of tbpa and La(ClO4)3?7H2O (0.06 mmol) in MeOH (0.7 mL)–CHCl3 (0.5 mL) standing overnight at room temperature. 1 M. Ephritikine, New J. Chem., 1992, 16, 451. 2 I. Santos, A. Pires de Matos and A. G. Maddock, Adv. Inorg. Chem., 1989, 34, 65. 3 A. J. Amoroso, J. C. JeVery, P. L. Jones, J. C. McCleverty, L. Rees, A. L. Rheingold, Y. Sun, J. Takats, S. Trofimenko, M. D. Ward and G. P. A. Yap, J. Chem. Soc., Chem. Commun., 1995, 1881; A. Carvalho, A. Domingos, P.Gaspar, N. Marques, A. Pires de Matos and I. Santos, Polyhedron, 1992, 11, 1481; P. Roussel and P. Scott, J. Am. Chem. Soc., 1998, 120, 1070. 4 K. N. Nash, Solvent Extr. Ion Exch., 1993, 11, 729. 5 C. Musikas, Actinide-Lantanide Group Separations using Sulfur and Nitrogen Donor Extractants, in Actinide/Lanthanide Separations Proceedings of an International Symposium, Honolulu, Hawaii, 16–22 Dec. 1984, World Scientific, Singapore, 1985, p. 19. 6 F. A. Hart and M. Tajik, Inorg. Chim. Acta, 1983, 71, 169. 7 R. Wietzke, M. Mazzanti, J.-M. Latour, J. Pecaut, P.-Y. Cordier and C. Madic, Inorg. Chem., in the press. 8 R. Ziessel and J.-M. Lehn, Helv. Chim. Acta, 1990, 73, 1149. 9 V. Balzani, E. Berghmans, J.-M. Lehn, N. Sabbatini, R. Terörde and R. Ziessel, Helv. Chim. Acta, 1990, 73, 2083. 10 L. R. Avens, S. G. Bott, D. L. Clark, A. P. Sattelberger, J. G. Watkin and B. D. Zwick, Inorg. Chem., 1994, 33, 2248. 11 A. R. Schake, L. R. Avens, C. J. Burns, D. L. Clark, A. P. Sattelberger and W. H. Smith, Organometallics, 1993, 12, 1497; A. Zalkin and J. G. Brennan, Acta Crystallogr., Sect. C, 1987, 43, 1919; H. J. Wasserman, D. C. Moody, R. T. Paine, R. R. Ryan and K. V. Salazar, J. Chem. Soc., Chem. Commun., 1984, 533. 12 A. R. Al-Karaghouli and J. S. Wood, Inorg. Chem., 1972, 11, 2293. 13 M. Fréchette, I. R. Butler, R. Hynes and C. Detellier, Inorg. Chem., 1992, 31, 1650; M. Fréchette and C. Bensimon, Inorg. Chem., 1995, 34, 3520. 14 K. N. Raymond and C. W. Eigenbrot, Jr., Acc. Chem. Res., 1980, 13, 276. Communication 8/0792

ISSN:1477-9226    

DOI:10.1039/a807921h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 4089–4090 4089 Bis(isonicotinato)iron(II): a rare, neutral three-dimensional iron coordination polymer Ren-Gen Xiong,a Scott R. Wilson b and Wenbin Lin *a a Department of Chemistry, Brandeis University, Waltham, MA 02454, USA. E-mail: wlin@brandeis.edu b School of Chemical Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA Received 18th August 1998, Accepted 9th November 1998 The reaction of iron(III) perchlorate with 4-pyridinecarbaldehyde under hydro(solvo)thermal conditions affords bis(isonicotinato)iron(II) 1, in which both pyridine and carboxylate groups coordinate to iron centers resulting in an unprecedented neutral, three-dimensional polymeric network with a carboxylate-bridged Fe–Fe infinite chain.Coordination polymers with one-, two-, and three-dimensional infinite frameworks have been actively sought in recent years owing to their potential applications as magnetic materials,1 zeolite analogues,2 and catalysts.3 Research eVort in this area has so far been mostly focused on coordination polymers with either neutral donor ligands (e.g., 4,49-bipyridine) 4 or strictly anionic groups (e.g., carboxylates).5 Very little is known about coordination polymers with multifunctional ligands in which both neutral and anionic donor groups are present and can coordinate to metal centers potentially resulting in neutral polymeric structures.We report here the synthesis and characterization of bis(isonicotinato)iron(II), a rare, neutral three-dimensional iron coordination polymer.In this threedimensional network, the isonicotinate group acts as a bifunctional ligand by coordinating to Fe(II) centers through both carboxylate and pyridine functional groups. Iron(II) coordination polymers are also of great interest because they can exhibit interesting spin crossover phenomena if the coordinating ligands have suitable ligand field strength.6 Bis(isonicotinato)iron(II) 1 was synthesized by a hydro- (solvo)thermal reaction between Fe(ClO4)3?6H2O and 4-pyridinecarbaldehyde.† In this reaction, Fe(III) centers were reduced to Fe(II), while 4-pyridinecarbaldehyde was oxidized to isonicotinic acid.The presence of carboxylate groups in 1 was suggested by the strong peaks at 1611, 1552, and 1405 cm21 (C]] O stretching) in the infrared spectrum. Additionally, the IR spectrum indicated that no perchlorate group is present in 1.The three-dimensional polymeric structure of 1 was revealed by an X-ray single crystal diVraction study.‡ The coordination geometry around the Fe(II) center in 1 is a slightly distorted octahedron (Fig. 1). The Fe(II) center is bonded to four oxygen atoms from four diVerent isonicotinate ligands, and to two pyridine nitrogen atoms from two isonicotinate moieties; these oxygen and nitrogen atoms are related by an inversion center lying at the Fe ion.7 The four oxygen atoms bind to the Fe(II) center equatorially, while the two pyridine nitrogen atoms bind to the Fe(II) center in axial positions.All the Fe–O [2.074(1) and 2.142(1) Å] and Fe–N [2.198(2) Å] bond distances are normal. The bond angles around the Fe center range from 86.9(1) to 93.1(1)8, and deviate no more than 3.18 from those expected for an ideal octahedron. The iron center is bridged by four carboxylate groups to two adjacent iron centers to result in an infinite Fe–Fe chain along the crystallographic a axis (Fig. 1). Two adjacent iron centers in this infinite chain form an eight-membered ring with two bridging carboxylate groups. Each carboxylate group of the isonicotinate ligand bridges two iron centers in a syn-anti fashion.8 Consequently, if the central carbon of the carboxylate group is omitted, the two iron and four oxygen atoms in the ring are not coplanar; they instead adopt a chair-like conformation (Fig. 1). The Fe–Osyn bond length of 2.074(1) Å is slightly shorter than the Fe–Oanti bond length of 2.142(1) Å, probably as a result of the higher basicity of the lone pair in the syn carboxylate oxygen.8 This syn-anti carboxylate bridging structure is very rare, and the Fe–Fe distance of 4.954 Å in 1 is significantly longer than those of known diiron complexes with syn-anti carboxylate bridging groups; the previously reported longest Fe–Fe distance in a syn-anti carboxylate bridged diiron complex is 4.612(3) Å.9 Although there are numerous transition metal carboxylate complexes in the literature, to our knowledge, such a doubly carboxylate-bridged, infinite chain with a syn-anti conformation has not been previously observed.The most striking feature of 1 is, however, the connection of the doubly carboxylate-bridged, infinite Fe–Fe chain to the four adjacent Fe–Fe chains by isonicotinate groups to form a threedimensional network (Fig. 2). The Fe–Fe chains are connected to each other along the (0 1 1) and (0 21 1) planes.In the (0 1 1) planes, the Fe–Fe chain is connected to two neighboring Fe–Fe chains with the pyridine ends of the isonicotinate groups. In the (0 21 1) planes, the Fe–Fe chain is connected to two other neighboring Fe–Fe chains with the carboxylate groups of the isonicotinate ligands. Therefore, 1 adopts a very regular threedimensional network structure via coordination of iron centers to both pyridine and carboxylate functionalities of isonicotinate groups.A perspective view of 1 down the a axis is shown in Fig. 3. The centroid-to-centroid distance between adjacent pyridyl rings is rather long at 4.95 Å; there are thus no p–p interactions between the pyridyl rings. Compound 1 does not have any cavity accessible to solvent molecules. This represents the first example of a neutral transition metal isonicotinate with both pyridine and carboxylate functionalities coordinating to the metal centers.10,11 Fig. 1 An ORTEP12 drawing showing the coordination geometry of the iron(II) center in 1 (30% probability). The asymmetric unit is labeled.4090 J. Chem. Soc., Dalton Trans., 1998, 4089–4090 In summary, we have synthesized a rare, neutral threedimensional iron coordination polymer using a hydro(solvo)- thermal method. Preliminary magnetic measurements indicate that 1 is low-spin up to 375 K. The synthesis of other coordination polymers with isonicotinate building blocks is currently under investigation.Fig. 2 Diagram illustrating the connectivity between diVerent Fe chains doubly bridged by the carboxylate end of the isonicotinate group. For clarity, the pyridine ring of the isonicotinate group is represented with a straight line. The open circles with increasing sizes are C, O, and Fe, respectively. Fig. 3 A perspective view of 1 down the a axis. Open circles: carbon; hatched circles: nitrogen; dotted circles: oxygen and iron. Acknowledgements We acknowledge NSF (CHE-9727900) and ACS-PRF for financial support.We also thank Professor Bruce M. Foxman for helpful suggestions, and Ms Teresa Prussak-Wieckowska and the Materials Chemistry Laboratory at University of Illinois at Urbana-Champaign for X-ray data collections. Notes and references † Preparation of Fe(C7H4NO2)2: a heavy walled Pyrex tube containing a mixture of Fe(ClO4)3?6H2O (0.18 g, 0.5 mmol) and 4-pyridinecarbaldehyde (0.14 mL, 1.5 mmol) in ethanol (0.3 mL) and H2O (0.07 mL) was frozen and sealed under vacuum and placed inside an oven at 110 8C.Orange-red rodlike crystals were obtained after 24 hours of heating. Yield: 0.08 g (49.4%). IR (KBr, cm21): 1611 (vs), 1552 (s), 1405 (vs), 1217 (w), 1064 (w), 1018 (w), 875 (vw), 845 (w), 781 (m), 709 (w), 678 (m), 558 (vw). ‡ Crystal data for 1: C14H8N2O4Fe, monoclinic, space group P21/n (no. 14), a = 4.9544(1), b = 13.2443(2), c = 10.4983(1) Å, b = 101.586(1)8, V = 674.84(2) Å3, M = 300.05, F(000) = 304, Z = 2, Dc = 1.477 g cm23, T = 198(2) K, m(Mo-Ka) = 11.3 cm21, l(Mo-Ka) = 0.71073 Å.An orange crystal of approximate dimensions 0.60 × 0.08 × 0.04 mm was mounted on a SMART CCD diVractometer equipped with graphitemonochromated Mo-Ka radiation. The crystal was twinned, and two distinct orientations were identified and integrated. The structure solution was carried out with non-overlapped reflections, while the refinement was by full-matrix least squares on F2 with both non-overlapped and completely overlapped reflections.Final R1, wR2, and S are 0.057, 0.163, and 1.167 for 5981 reflections with I > 2s(I) out of 7105 unique reflections. CCDC reference number 186/1239. See http://www.rsc.org/ suppdata/dt/1998/4089/ for crystallographic files in .cif format. 1 J. L. Manson, C. Campana and J. S. Miller, Chem. Commun., 1998, 251. 2 C. Janiak, Angew. Chem., Int. Ed. Engl., 1997, 36, 1431; G. B. Gardner, D. Venkataraman, J. S. Moore and S.Lee, Nature (London), 1995, 374, 792. 3 M. Fujita, Y. J. Kwon, S. Washizu and K. Ogura, J. Am. Chem. Soc., 1994, 116, 1151. 4 F. Robinson and M. J. Zaworotko, J. Chem. Soc., Chem. Commun., 1995, 2413; S. W. Keller, Angew. Chem., Int. Ed. Engl., 1997, 36, 247; R. W. Gable, B. F. Hoskins and R. Robson, J. Chem. Soc., Chem. Commun., 1990, 1677; D. Hagrman, C. Zubieta, D. J. Rose, J. Zubieta and R. C. Haushalter, Angew. Chem., Int. Ed. Engl., 1997, 36, 873; J. Lu, G. Crisci, T.Niu and A. J. Jacobson, Inorg. Chem., 1997, 36, 5140; M. Kondo, T. Yoshitomi, K. Seki, H. Matsuzaka and S. Kitagawa, Angew. Chem., Int. Ed. Engl., 1997, 36, 1725. 5 O. M. Yaghi, G. Li and H. Li, Nature (London), 1995, 378, 703. 6 O. Kahn and C. J. Martinez, Science, 1998, 279, 44. 7 The four carbon atoms on the isonicotinate ring are disordered, and have been successfully modeled with two mutually perpendicular sets (C1, C2, C4, C5, and C7, C8, C10, C11) with partial occupancies [0.508(3) and 0.492(3)]. Only one of the two orientations (C1, C2, C4, C5) is shown in Figs. 1 and 3. 8 R. L. Rardin, W. B. Tolman and S. J. Lippard, New. J. Chem., 1991, 15, 417. 9 C. Hemmert, M. Verelst and J.-P. Tuchagues, Chem. Commun., 1996, 617. 10 Two reports on the use of the isonicotinate group for the construction of cationic coordination polymers have recently appeared: L. R. MacGillivray, R. H. Groeneman and J. L. Atwood, J. Am. Chem. Soc., 1998, 120, 2676; A. D. Burrows, M. F. Mahon and M. T. Palmer, J. Chem. Soc., Dalton Trans., 1998, 1941. 11 Main group metal coordination polymers with isonicotinate bridging ligands have been reported: S. W. Ng and V. G. K. Das, J. Crystallogr. Spectrosc. Res., 1992, 371; M. B. Cingi, A. G. Manfredotti, C. Guastini and M. Nardelli, Gazz. Chim. Ital., 1972, 1034. 12 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. Communication 8/06499G

ISSN:1477-9226    

DOI:10.1039/a806499g

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON COMMUNICATION J. Chem. Soc., Dalton Trans., 1998, 4091–4093 4091 Chalcogen abstraction from dithiadiazolyl and diselenadiazolyl platinum complexes: crystal structure of a novel metallaheterocycle Neil Feeder, Robert J. Less, Jeremy M. Rawson* and J. Nicholas B. Smith Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1EW Received 9th November 1998 The open-shell complexes, (PhCNEEN)Pt(dppe) (E 5 S or Se) decompose via a chalcogen abstraction process coupled with N-protonation to yield the novel 10� metalla-heterocyclic cations {[PhCN(H)N(H)E]Pt(dppe)}1: the selenium complex is characterised by X-ray crystallography as its chloride salt.The dithiadiazolyl radical RCNSSN, 1, has an extremely varied coordination chemistry;1 it undergoes oxidative addition reactions with a number of low-valent metal species to give mono-, di- or tri-metallic complexes in which the dithiadiazolyl radical ring-opens, with cleavage of the S–S bond.The diversity of the structural types is matched by the variable number of electrons which can be used for ligand–metal bonding (between 2 and 6 e2 depending on the coordination mode).1 The series of monometallic complexes of general formula (RCNSSN)M(P2) [where M = Pt, Pd and P = PPh3 or ��� dppe] in which the unpaired electron associated with 1 is retained in the complex exhibit an unusual reactivity. For example, (PhCNSSN)Pt- (PPh3)2 disproportionates 2 to form the trimetallic complex Pt3(PhCNSSN)2(PPh3)4 and oxidation of (PhCNSSN)Pd- (dppe) with [NO][BF4] proceeds via ring protonation 3 with the formation of [Pd2(PhCNSSNH)(dppe)2][BF4]2.The coordination chemistry of the selenium analogue, 2, has only recently begun to be investigated, but preliminary results 4,5 indicate that the coordination chemistries of 1 and 2 are similar. Now we report a new decomposition pathway for the monometallic platinum dppe complexes, 3 (E = S or Se), which involves the unexpected extrusion of chalcogen and subsequent ring contraction to form a novel five-membered metallaheterocycle, 41.The blue complex, (PhCNSSN)Pt(dppe), 3 (E = S),† was prepared by reaction of Pt(dppe)2 with (PhCNSSN)2 in an analogous fashion to (PhCNSSN)Pd(dppe).3 The solution EPR spectrum of pure 3 (E = S) [EPR (CH2Cl2, 298 K): g = 2.046, aN = 0.55, aP = 0.35 and aPt = 5.48 mT) exhibited well-resolved coupling to N, P and Pt, consistent with extensive p-delocalisation of the unpaired electron.The intensity of the EPR signal of 3 (E = S) slowly decreased over time and a fiveline spectrum [EPR (CH2Cl2, 298 K): g = 2.010, aN = 0.50 mT] typical 6 of free 1 became observed. Subsequent 31P NMR spectra (CDCl3) of the reaction mixture indicated three major P-containing products which were separated by TLC (silica support, 50 : 50 acetone–acetonitrile eluent). An orange band was isolated (Rf ª 0.65) which was assigned to the trimetallic complex, Pt3(PhCNSSN)2(dppe)2 on the basis of the singlet 31P NMR resonance with Pt satellites (dP 149.8, 1JPt = 2299 Hz), UV–VIS spectrum and microanalytical data.This observation, coupled with the identification of 1 (by EPR) are consistent with disproportionation of 3 (E = S) [eqn. (1)] in an entirely analogous fashion to the PPh3 derivative.2 3 (PhCNSSN)Pt(dppe) æÆ Pt3(PhCNSSN)2(dppe)2 1 PhCNSSN 1 dppe (1) A pale pink band (Rf ª 0.5) was characterised by NMR (MeCN, dP 144.8) and a subsequent X-ray diVraction study 7 as 1,2-bis(diphenylphosphino)ethane disulfide, Ph2P(S)CH2- CH2P(S)Ph2 (dppeS2).The dppe, arising as a by-product during the primary decomposition pathway [eqn. (1)] abstracts sulfur (presumably from 3, see below) to form dppeS2. The third, yellow, band (Rf ª 0.2) exhibited two 31P resonances with 2JP couplings and 1JPt satellites (dP 140.9, 2JP = 9, 1JPt = 3051 Hz; dP 144.7, 2JP = 9, 1JPt = 2631 Hz), corresponding to two chemically distinct P environments, i.e.PPtP9. The FIB mass spectrum exhibited a molecular ion peak at m/z = 744.1 indicative of S-extrusion from 3 (E = S) and protonation of both N atoms, i.e. the cation [4]1 (E = S). By comparison with other N,S-bound sulfur–nitrogen ligands 8 such as SNSN22 and SNSNH2, the two 31P NMR resonances (at dP 140.9 and dP 144.7) can be assigned as P trans to N and S respectively. Abstraction of sulfur from organic sulfides and chelating sulfide ligands by nucleophiles such as CN2 or PPh3 is not uncommon,9 e.g.(NH4)2Pt(S5)3 is attacked by PPh3 to give (Ph3P)2PtS4. In this case, the S-abstraction from 3 is carried out by dppe formed by disproportionation of 3 itself [eqn. (1)]. Since no free dppe was observed in the product mixture, the overall reaction can be considered to generate 2 moles of 41 from 5 moles of 3. On this basis, recovered yields of [4]Cl were ca. 50%. The source of protonation is not entirely clear, although the presence of the Cl2 anion in the product indicates that the chlorinated solvents (CH2Cl2 and CDCl3) are intimately involved in the reaction.Similar N-protonation reactions have previously been observed.3,10 Although the formation of [4]1 (E = S) is accelerated in the presence of a source of H1 ions (e.g. carrying out the reaction in wet solvents, with the mixture exposed to the air, or by the addition of silica to the reaction solution) [eqn. (2)], ring-protonation is still observed under 2 (PhCNSSN)Pt(dppe) 1 dppe 2H/2H1 2 {[PhCN(H)SN(H)]Pt(dppe)}1 1 dppeS2 (2)4092 J.Chem. Soc., Dalton Trans., 1998, 4091–4093 rigorously dry conditions. Attempts to grow crystals of 41 (E = S) salts from CH2Cl2 by slow evaporation or layering with hexane or Et2O proved unsuccessful. In a similar manner, addition of excess Pt(dppe)2 to (Ph- CNSeSeN)2 yielded a green complex which exhibited a broad singlet EPR spectrum with 195Pt satellites [EPR (C6H5Me): g = 2.058, aPt = 5.4 mT], which was assigned to 3 (E = Se).Replacement of S by Se leads to line broadening and poorly resolved EPR spectra,4 and in this instance hyperfine coupling to N could not be resolved. The green colouration was rapidly dissipated to produce a yellow solution containing (PhCNSe- SeN)2 [EPR (C6H5Me): g = 2.03] and a small quantity of yellow precipitate. The reaction was repeated on a preparative scale, and a 31P NMR of the yellow precipitate indicated a mixture of two Pt-containing products, with chemical shifts and coupling constants analogous to the sulfur system.The major product, [4]Cl (E = Se) obtained in 20% yield, exhibited two P environments (dP = 141.9, 2JP = 9, 1JPt = 3019 Hz; dP 144.7, 2JP = 9, 1JPt = 2691 Hz), analogous to [4]Cl (E = S) whilst the minor product exhibited a single 31P NMR resonance (dP 147.6, 1JPt = 2363 Hz), consistent with Pt3(PhCNSeSeN)2(dppe)2. The low solubility of both products precluded the observation of Se satellites.The mass spectrum (FAB) exhibited a molecular ion peak at m/z 791.9, consistent with [4]1 (E = Se), analogous to [4]1 (E = S) and also a nNH = 3350 cm21 absorption in the IR. Crystals of [4]Cl (E = Se) suitable for X-ray diVraction‡ were grown by slow diVusion methods (dichloromethane–hexane). The structure of [4]Cl (E = Se) is shown in Fig. 1. The central Pt atom takes up an approximately square-planar geometry with a P2NSe donor set.The chelate nature of both rings leads to some deviation from ideality with both PPtP and NPtSe angles a little less than 908 [84.98(16) and 83.8(4)8 respectively]. The Pt–P bonds are unexceptional [averaging 2.255(5) Å] and the Pt–Se and Pt–N bond lengths are 2.4085(19) and 2.043(15) Å respectively. Whilst derivatives of 3 (E = S) typically exhibit a puckering of the metalla-heterocyclic framework to accomodate the ring strain induced at Pt, the structure of [4]Cl (E = Se) has an almost planar CN2SePt ring (mean deviation < 0.03 Å).The mean C–N bond length [1.31(2) Å] is the same as those observed 12 in both 21 and 2 (R = Ph) which average at 1.36(3) and 1.32(2) Å respectively. In comparison, the Se–N bond length at 1.890(15) Å is longer than that observed 12 in 21 and 2 (R = Ph) [averaging 1.76(2) Å and 1.78(1) respectively]. This can be rationalised in terms of the addition of an extra electron into an N–Se antibonding orbital (described below).The H atom attached to N(2) is hydrogen–bonded to the chloride anion (N ? ? ? Cl 3.07 Å) and the structure can be considered as (PhCNSeNH)Pt(dppe)?HCl (the HCl presumably arises from the CH2Cl2 used during recrystallisation). The second H atom is sterically more protected and does not appear to exhibit any H-bonding, although there is a long intermolecular contact to a Cl atom of a CH2Cl2 solvate molecule (N ? ? ? Cl at 3.78 Å). Fig. 1 Crystal structure of [4]Cl (E = Se) with heteroatom labelling and with 50% probability ellipsoids.Extended Hückel calculations 13 on the parent {[HCN(H)- SeN(H)]Pt(PH3)2}1 cation indicate that the frontier molecular orbitals are both of p-character (Fig. 2) with the LUMO based predominantly on the NCN fragment and the HOMO on the N2SePt unit. These p molecular orbitals are closely related to those observed for the parent heterocycle, 21; the two structures being related by replacement of a Se atom in 2 by a Pt(PH3)2 unit and addition of two H atoms on the N atoms.In 21 and 2, there are 6p and 7p electrons respectively. In 41 the RC(NH)- (NH)Se1 fragment provides 6p electrons and the Pt centre contributes two orbitals (dxz and dyz) and a further 4e2 for p-bonding, producing a formally 10p aromatic system; the dyz orbital contributing to the HOMO of 41. The HOMOs of both 2 and 41 are non-bonding with respect to C–N and antibonding with respect to N–Se. The diVerence in the Se–N bond lengths between 21, 2 and 41, described above, can then be rationalised in terms of the sequential addition of electrons into an N–Se antibonding orbital.Acknowledgements We would like to thank the EPSRC and Ciba-Geigy for studentships (R. J. L. and J. N. B. S. respectively), and the Royal Society for an equipment grant. Notes and references † 3 (E = S): yield = 92%, mp 218 8C (decomp.), UV–VIS lmax = 680 nm (Found: C, 51.3; H, 3.7; N, 3.9. Calc.: C, 51.2; H, 3.8; N, 3.6%).‡ Crystal data: [4]Cl?CH2Cl2: C34H33Cl3N2P2PtSe, Mr = 911.96, monoclinic, P21/c, a = 15.000(5), b = 16.426(5), c = 14.722(5) Å, b = 103.11(2)8, V = 3533(2) Å3, Z = 4, rcalc = 1.715 g cm23, F(000) = 1776, graphite-monochromated Mo-Ka radiation, l = 0.71069 Å, m = 5.346 mm21, T = 180(2) K. Of 11324 reflections collected on a Rigaku R-Axis IIc image plate diVractometer, 6221 were unique data (2q <50.588, Rint = 0.153). The structure was solved by direct methods and refined by full-matrix methods on F2 values for all reflections 11 with anisotropic displacement parameters for all non-hydrogen atoms, except the CH2Cl2 solvent molecule which was refined isotropically. H atoms, including N–H, were added at calculated postions and refined using a riding model.The refinement of 329 parameters on F2 using all 6221 unique reflections converged at wR2 = 0.144, R1 = 0.071 [for Fo > 2s(Fo)] and goodness of fit S = 0.85. Largest residual electron densities were within 11.03/–1.37 e Å23.CCDC reference number 186/1240. See http://www.rsc.org/suppdata/dt/1998/4091 for crystallographic files in .cif format. 1 A. J. Banister, I. May, J. M. Rawson and J. N. B. Smith, J. Organomet. Chem., 1998, 550, 241. 2 A. J. Banister, I. B. Gorrell, J. A. K. Howard, S. E. Lawrence, C. W. Lehman, I. May, J. M. Rawson, B. K. Tanner, C. I. Gregory, A. J. Blake and S. P. Fricker, J. Chem. Soc., Dalton Trans., 1997, 377. Fig. 2 Frontier molecular orbitals of (a) [HCNSeSeN] and (b) {[HCN(H)SeN(H)]Pt(PH3)2}1.J.Chem. Soc., Dalton Trans., 1998, 4091–4093 4093 3 A. J. Banister, J. A. K. Howard, I. May and J. M. Rawson, Chem. Commun., 1997, 1763. 4 J. M. Rawson, A. J. Banister and I. May, Magn. Reson. Chem., 1994, 32, 487. 5 J. E. Davies, R. J. Less, I. May and J. M. Rawson, New. J. Chem., 1998, 763. 6 J. M. Rawson, A. J. Banister and I. Lavender, Adv. Heterocycl. Chem., 1995, 62, 137. 7 J. A. K. Howard, A. S. Batsanov and J. N. B. Smith, unpublished work. 8 I. P. Parkin and J. D. Woollins, J. Chem. Soc., Dalton Trans., 1990, 925; C. A. Mahoney, I. P. Parkin, D. J. Williams and J. D. Woollins, J. Chem. Soc., Dalton Trans., 1989, 1179; (c) R. Jones, P. F. Kelly, D. J. Williams and J. D. Woollins, J. Chem. Soc., Dalton Trans., 1988, 803. 9 See for example: B. Kreutzer, P. Kreutzer and W. Beck, Z. Naturforsch., Teil B, 1972, 27, 461; D. Dudis and J. P. Fackler Jr., Inorg. Chem., 1982, 21, 3577; J. P. Fackler, Jr., J. A. Fetchin and D. C. Fries, J. Am. Chem. Soc., 1972, 94, 7332. 10 R. T. Boeré, K. H. Moock, V. Klassen, J. Weaver, D. Lentz and H. Michael-Schulz, Can. J. Chem., 1995, 73, 1444. 11 G. M. Sheldrick, SHELXTL Manual, Siemens Analytical X-Ray Instruments Inc., Madison, WI, 1990; G. M. Sheldrick, SHELXL 93, Program for crystal structure determination, University of Göttingen, 1993. 12 P. D. B. Belluz, A. W. Cordes, E. M. Kristof, P. V. Kristof, S. W. Liblong and R. T. Oakley, J. Am. Chem. Soc., 1989, 111, 9276. 13 C. Mealli and D. M. Proserpio, J. Chem. Educ., 1990, 67, 399 (PC version 4.0, 1994, using in-laid parameters). Communication 8/08758J

ISSN:1477-9226    

DOI:10.1039/a808758j

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 4095–4100 4095 Cycloauration of 2-substituted pyridine derivatives. Synthesis, structure and reactivity of six-membered cycloaurated complexes of 2-anilino-, 2-phenoxy- and 2-(phenylsulfanyl)-pyridine Yoshio Fuchita,*a Hidenori Ieda,a Arata Kayama,a Junko Kinoshita-Nagaoka,b Hiroyuki Kawano,b Shingo Kamedac and Masahiro Mikuriya c a Division of Molecular Chemistry, Graduate School of Science, Kyushu University at Ropponmatsu, Chuo-ku, Fukuoka 810-8560, Japan.E-mail: fuchita@rc.kyushu-u.ac.jp b Department of Applied Chemistry, Faculty of Engineering, Nagasaki University, Bunkyo-machi, Nagasaki 852-8521, Japan c Department of Chemistry, School of Science, Kwansei Gakuin University, Nishinomiya 662-8501, Japan Received 11th September 1998, Accepted 19th October 1998 At room temperature and in ethanol 2-anilinopyridine reacted with H[AuCl4]?4H2O as well as Na[AuCl4]?2H2O to give directly the six-membered cycloaurated complex [AuCl2(pap-C1, N)] 1a [pap = 2-(2-pyridylamino)phenyl], whereas 2-phenoxypyridine (Hpop) and 2-(phenylsulfanyl)pyridine (Hptp) produced only the salts [H2pop][AuCl4] 2b and [H2ptp][AuCl4] 2c, respectively.The adducts [AuCl3(Hpop)] 3b and [AuCl3(Hptp)] 3c have separately been prepared by the reactions of Na[AuCl4]?2H2O with Hpop and Hptp, respectively, in an acetonitrile–water mixed solvent. The salts 2b and 2c can be converted into the corresponding adducts 3b and 3c when they are stirred in acetonitrile–water at room temperature.The cycloaurated complexes [AuCl2(pop-C1, N)] [pop = 2-(2-pyridyloxy)- phenyl] 1b and [AuCl2(ptp-C1, N)] [ptp = 2-(2-pyridylsulfanyl)phenyl] 1c have been obtained by heating the salts or the adducts in acetonitrile–water. Moreover, complexes 1b and 1c have been synthesized directly by the reaction of H[AuCl4]?4H2O with Hpop and Hptp in refluxing acetonitrile–water, ethanol–water and propan-2-ol–water.The reaction of 1a with an equimolar amount of PPh3 in the presence of NaBF4 gave the cationic complex [AuCl- (pap-C1, N)(PPh3)]BF4 5a, while two equivalents of PPh3 or PEt3 aVorded [AuCl(pap-C1)(PPh3)2]Cl 6a or [AuCl- (pap-C1)(PEt3)2]Cl 7a where the pap ligand co-ordinates through only the carbon atom. On the other hand, the C–N chelates in 1b and 1c are easily cleaved with one equimolar amount of PPh3 to give [AuCl2(pop-C1)(PPh3)] 8b and [AuCl2(ptp-C1)(PPh3)] 8c, respectively. The boat-form structures of the three six-membered auracycles have been confirmed by X-ray diVraction studies of 1b, 1c and 5a.The crystal structure of 7a has also been determined. Cyclometallation is an elegant method used to activate C–H bonds in heterosubstituted molecules.1 However, cycloauration is generally hard to achieve and until now examples have been in principle limited to 2-substituted pyridine derivatives, i.e. 2-phenylpyridine,2 2,9-diphenyl-1,10-phenanthroline,3 4-(4- methoxyphenyl)-6-phenyl-2,29-bipyridine,4 2-benzylpyridine,5,6 6-benzyl-2,29-bipyridine derivatives7 and 6-tert-butyl-2,29- bipyridine.7 With other ligands such as azobenzene,8 N,N-dimethylbenzylamine,9,10 4,4-dimethyl-2-phenyl-1,3-oxazoline, 10 1-(dimethyl- or methyl-aminomethyl)naphthalene,10 1,3-bis(dimethylaminomethyl)benzene 10 and 4-butyl-N-(3,4,5- trimethoxybenzylidene)aniline,11 stable cycloaurated complexes have been synthesized by transmetallation from the corresponding organomercury(II) compounds.We have been challenging the development of new cycloaurations in recent years and have succeeded in the cycloauration of 2-benzoylpyridine.6 As an extension of this work, we wish to report here the cycloauration of 2-anilino-, 2-phenoxy- and 2-(phenylsulfanyl)-pyridine by tetrachloroaurate ion and the X-ray crystallographic analysis of the resulting six-membered cycloaurated complexes. While the present work concerning the cycloauration of 2-anilinopyridine was nearly established, Nonoyama et al.12 reported isolation of the cycloaurated complexes [AuCl2(C–N)] derived from 2-anilino-, 2-(4-toluidino)- and 2-(N-methylanilino)-pyridine. Results and discussion The method of preparation of the six-membered cycloaurated complexes derived from 2-anilino- (Hpap), 2-phenoxy- (Hpop) and 2-(phenylsulfanyl)-pyridine (Hptp) is shown in Scheme 1.Assignment of the 1H NMR spectra was performed with the aid of 1H–1H correlation spectroscopy (COSY) and the data are summarized in Table 1.Reactions of the 2-substituted pyridines, Hpap, Hpop and Hptp, with [AuCl4]2, resulting in the formation of salts, adducts and six-membered cycloaurated complexes 2-Anilinopyridine reacted at room temperature in ethanol with an equimolar amount of H[AuCl4]?4H2O or Na[AuCl4]?2H2O to give directly the six-membered cycloaurated complex [AuCl2(pap-C1, N)] 1a [ pap = 2-(2-pyridylamino)phenyl] in 27 or 62% yield, respectively.It should be noted that the cycloauration proceeds under very mild conditions at room temperature, while Nonoyama et al.12 reported previously that 1a was obtained only when an aqueous mixture containing equimolar amounts of Hpap and Na[AuCl4]?2H2O was heated under reflux. It was found that the yields of 1a depend upon the molar ratio between Hpap and [AuCl4]2, and the highest yields, 88 and 93%, were obtained when the ratio was 3 : 1 for H[AuCl4]?4H2O and 2 : 1 for Na[AuCl4]?2H2O, respectively.These facts indicated that an excess of Hpap accelerates the cycloauration probably by trapping hydrogen chloride generated during the course of the reaction. On the contrary, in ethanol 2-phenoxypyridine and 2- (phenylsulfanyl)pyridine did not cyclometallate by H[AuCl4]? 4H2O at room temperature and even at refluxing temperature,4096 J. Chem. Soc., Dalton Trans., 1998, 4095–4100 only producing the tetrachloroaurate salts, [H2pop][AuCl4] 2b and [H2ptp][AuCl4] 2c, respectively.Interestingly, in an acetonitrile–water (1 :5) mixed solvent the same reaction carried out at room temperature aVorded the adducts [AuCl3(Hpop)] 3b and [AuCl3(Hptp)] 3c. It was also found that the salts could be converted into the corresponding adducts almost quantitatively at room temperature in acetonitrile–water (1 :5). Although Nonoyama et al.12 demonstrated that both 2-(p-tolyloxy)- and 2-(p-tolylsulfanyl)-pyridine do not cyclometallate, novel six-membered cycloaurated complexes [AuCl2(pop-C1, N)] 1b [ pop = 2-(2-pyridyloxy)phenyl] and [AuCl2(ptp-C1, N)] 1c [ ptp = 2-(2-pyridylsulfanyl)phenyl] could be obtained in about 70 and 20% yields respectively by heating the salts or the adducts at 105 8C in acetonitrile–water (1 : 5).It seems reasonable that the cycloauration from the salts proceeds via the formation of the adducts. Moreover, it was also found that cycloaurations of Hpop and Hptp by H[AuCl4]?4H2O occur in refluxing alcohol (ethanol or propan-2-ol) in the presence of water.In addition to this fact, the result that the salts and the adducts did not produce the cycloaurated complexes by refluxing in water-free acetonitrile clearly showed the necessity of water for the cycloauration of Hpop and Hptp. However, the role of water is not clear at the moment. It is also noted that a yellow complex 4c was always obtained in the course of the isolating procedures for 1c.The yield of 4c was about two thirds by weight of that of 1c. However, in spite of the simple 1H NMR and far-IR spectra (see Experimental section), the structure could not be assigned. The far-IR spectra of the cycloaurated complexes 1a, 1b and 1c showed two bands characteristic of v(Au–Cl) frequencies trans to pyridyl nitrogen atom [356 (1a),12 361 (1b) and 360 (1c)] and phenylene carbon atom [284 (1a),12 303 (1b) and 295 (1c)].11 Each 1H NMR spectrum of 1a, 1b and 1c exhibited only eight well separated aromatic protons due to the cycloaurated moiety, and a significant feature of all is the lower field shifts of d(H69) (numbering scheme in Scheme 1) compared with those of the ‘free’ ligands [Hpap (d 8.14), Hpop (8.15) and Hptp (8.40)], the salts (2b and 2c) and the adducts (3b and 3c).Such lower field shifts of d(H69) have been reported for other cycloaurated Scheme 1 (i) H[AuCl4]?4H2O or Na[AuCl4]?2H2O; (ii) H[AuCl4]?4H2O in ethanol; (iii) Na[AuCl4]?2H2O in CH3CN–water (1 : 5); (iv) CH3CN– water (1 : 5); (v) PPh3, NaBF4; (vi) 2PPh3 or 2PEt3; (vii) PPh3.(ii) 3b, 3c 5a (iii) 5' 6' (iv) (v) Ph3P Au Cl Cl (i) (iv) 6 4' 3 3' 4 5 8b, 8c 2b, 2c (vii) N E E=NH a; O b; S c N E Au Cl Cl E C5H4N BF4 HN E N Au Cl Cl Cl PhE Cl Au PPh3 N HN (vi) 1a, 1b, 1c 6a L=PPh3 7a L=PEt3 AuCl4 L Au Cl L NH C5H4N (iv) Cl complexes containing pyridine ligands2,5–7 and usually observed when a chlorine is in the proximity of the pyridine ring.13 Reactivity of the cycloaurated complexes 1a, 1b and 1c towards triphenylphosphine Complex 1a reacted with an equimolar amount of PPh3 in the presence of NaBF4 to give a cationic complex 5a (LM 169 S cm2 mol21 in acetone).The IR spectrum exhibited a strong band due to BF4 2 at 1060 cm21 and only one band at 313 cm21 assignable to the n(Au–Cl) frequency trans to the phenylene group.11 Moreover, in the 1H NMR spectrum the H69 proton in the pyridine moiety resonated at d 8.79 which is essentially the same chemical shift as that of 1a (d 8.82), indicating that the sixmembered pap–Au ring remains unchanged.On the basis of these data and elemental analysis 5a was assigned as a cationic four-co-ordinate complex [AuCl(pap-C1, N)(PPh3)]BF4. The similar complex [AuCl(pap-C1, N){P(C6H4CH3-4)3}]Cl has been prepared by Nonoyama et al.12 On the other hand, when a two-fold excess of PPh3 was treated with 1a another cationic complex [AuCl(pap-C1)(PPh3)2]Cl 6a (LM 123 S cm2 mol21 in MeOH) was obtained.A v(Au–Cl) band at 295 cm21 in the far- IR spectrum was characteristic of a frequency trans to carbon, supporting the presence of a Au–C bond. The H69 proton resonance observed significantly upfield (d 8.20) compared to d 8.82 for 1a and d 8.79 for 5a confirmed that the second incoming PPh3 cleaved the C–N chelate by dissociating the pyridine– nitrogen co-ordination. Such an upfield shift of d(H69) caused by C–N bond cleavage was also observed for cycloaurated complexes of 2-benzoylpyridine.6 A triethylphosphine analogue [AuCl(pap-C1)(PEt3)2]Cl 7a was also prepared for the X-ray diVraction study (see below).In contrast to the C–N chelate in complex 1a, those in the cycloaurated complexes 1b and 1c were easily cleaved by only one equimolar amount of PPh3 giving neutral complexes [AuCl2(pop-C1)(PPh3)] 8b (LM 1.4 S cm2 mol21 in acetone) and [AuCl2(ptp-C1)(PPh3)] 8c (LM 7.3 S cm2 mol21 in acetone), respectively.Two v(Au–Cl) frequencies of 8b [301 and 325 cm21] and 8c [301 and 316 cm21] lacked the characteristic bands due to chlorine trans to pyridine nitrogen.11 The H69 protons of 8b and 8c appeared significantly upfield d 7.99 and 8.30, respectively, compared with those for 1b (d 9.06) and for 1c (d 9.17). Such diVerent reactivity towards PPh3 of the three cycloaurated complexes 1a, 1b and 1c is probably associated with the stability of the Au–N bonds judging from the basicity of the nitrogen donors in the pyridyl moiety [pKa values:14 2-MeNHC5H4N (7.30), 2-MeOC5H4N (3.55) and 2-MeSC5H4N (4.36)].Crystal structures of complexes 1b, 1c, 5a and 7a The structures of complexes 1b, 1c, 5a and 7a were established by X-ray diVraction and ORTEP15 views of the molecules are shown in Figs. 1–4. Selected bond distances and angles are summarized in Tables 2–5. In complexes 1b, 1c and 5a the gold atoms have essentially square-planar AuCNCl2 and AuCNClP co-ordination with the mean deviation from the best planes of 0.017, 0.017, 0.019 Å, respectively, whereas in 7a the gold atom displays a square-planar AuCClP2 co-ordination with a very slight pyramidal distortion with deviations from the best plane of 20.089, 10.033 and 10.038 Å at Au, Cl(1) and C(1), respectively. The Au–C, Au–N, Au–Cl and Au–P bond distances are very similar to those reported for other gold(III) complexes.3,5,7,10,11,16 In complexes of 1b, 1c and 5a the pop–Au, ptp–Au and pap– Au six-membered auracycles have boat conformations, with atoms N, C(1), C(6) and C(8) essentially coplanar [mean deviations from their best planes of 0.012, 0.020 0.009 Å, respectively].These best planes form dihedral angles with planes C(6)–O–C(7) and N–Au–C(1) of 47.2 and 34.88 for 1b, with planes C(6)–S–C(7) and N–Au–C(1) of 42.5 and 43.48 for 1cJ. Chem. Soc., Dalton Trans., 1998, 4095–4100 4097 Table 1 Proton NMR spectral data of the gold(III) complexes a 2-Substituted pyridine ligand b Complex H69 Other protons Phosphine 1a [AuCl2(pap-C1, N)] 1b [AuCl2(pop-C1, N)] 1c [AuCl2(ptp-C1, N)] 2b [H2pop][AuCl4] 2c [H2ptp][AuCl4] 3b [AuCl3(Hpop)] 3c [AuCl3(Hptp)] 5a [AuCl(pap-C1, N)(PPh3)]- BF4 6a [AuCl(pap-C1)(PPh3)2]Cl 7a [AuCl(pap-C1)(PEt3)2]Cl 8b [AuCl2(pop-C1)(PPh3)] 8c [AuCl2(ptp-C1)(PPh3)] 8.82 (1 H, d) c 9.06 (1 H, d) c 9.17 (1 H, d) f 8.15 (1 H, d) i 8.41 (1 H, d) i 8.15 (1 H, dd) h,i 8.42 (1 H, dd) h,i 8.79 (1 H, d) c 8.20 (1 H, br) 8.08 (1 H, dd) h,i 7.99 (1 H, d) i 8.30 (1 H, dd) h,j 7.02 (1 H, t, H5) d 7.40 (1 H, d, H39) d 10.69 (1 H, s, NH) 7.23 (1 H, t, H5) d 7.69 (1 H, t, H59) e 7.25 (2 H, m, H4, H5) 7.78 (1 H, dt, H59) g,h 7.02 (1 H, d, H39) e 7.4 (2 H, m, Ph) 6.95 (1 H, d, H39) e 7.6 (2 H, m, Ph) d 7.02 (1 H, d, H39) d 7.4 (2 H, m, Ph) 6.97 (1 H, d, H39) d 7.6 (2 H, m, Ph) 6.22 (1 H, t, H4) e 7.12 (1 H, d, H6) e 7.99 (1 H, t, H49) e 6.12 (1 H, t, H4) e 6.95 (1 H, t, H59) d 7.76 (1 H, t, H49) d 6.85 (1 H, t, H59) d 7.23 (1 H, t, H4) e 7.65 (1 H, d, H49) d 6.57 (1 H, dd, H3) d,h 6.90 (1 H, t, H4) d 7.81 (1 H, t, H49) d 6.9 (2 H, m, H5, H39) 7.17 (1 H, d, H6) e 7.1 (2 H, m, H3, H59) 7.56 (1 H, d, H6) d 7.35 (2 H, m, H3, H4) 7.86 (1 H, d, H39) e 7.45 (1 H, d, H3) 8.25 (2 H, m, H39, H49) g 7.15 (3 H, m, Ph) 7.85 (1 H, t, H49) e 7.16 (1 H, dt, H59) h,i 7.66 (1 H, dt, H49) e,h 7.1 (3 H, m, Ph) 7.86 (1 H, dt, H49) d,h 7.19 (1 H, dt, H59) d,h 7.66 (1 H, dt, H49) d,h 6.71 (1 H, d, H3) e 7.19 (1 H, t, H59) e 10.55 (1 H, s, NH) 6.68 (1 H, d, H3) e 7.12 (1 H, d, H6) e 10.13 (1 H, s, NH) 6.92 (1 H, t, H5) e 7.36 (1 H, d, H6) e 9.13 (1 H, s, NH) 6.63 (1 H, t, H5) d 7.06 (1 H, t, H59) d 6.99 (1 H, dt, H4) e,h 7.45–7.65 (1 H, H49) j 7.26 (1 H, t, H4) d 7.97 (1 H, t, H49) d 7.59 (1 H, d, H6) e 8.43 (1 H, t, H49) e 7.58 (1 H, d, H6) c 7.21 (1 H, dt, H59) e,h 7.5 (3 H, m, Ph) 7.21 (1 H, dt, H59) d,h 7.5 (3 H, m, Ph) 6.99 (1 H, t, H5) e 7.46 (1 H, d, H39) e 6.86 (1 H, t, H5) e 7.31 (1 H, d, H39) d 7.06 (1 H, d, H39) d 7.48 (1 H, d, H3) e 6.82 (1 H, d, H39) d 7.17 (1 H, d, H6) d 7.1 (2 H, m, H3, H59) — — — — — — — 7.5–7.85 (15 H, m) 7.45–7.55 (30 H, m) 7.4–7.7 (15 H, m) 7.45–7.65 (15 H, m) j a Measured in DMSO-d6 at 270 MHz and at 23 8C; d in ppm with respect to SiMe4; s = singlet, d = doublet, t = triplet, br = broad, m = multiplet. b For numbering see Scheme 1.c 3J(HH) = 6.4 Hz. d 3J(HH) = 7.5 Hz. e 3J(HH) = 7.8 Hz.f 3J(HH) = 5.9 Hz. g 3J(HH) = 6.8 Hz. h 4J(HH) = 1.0 Hz. i 3J(HH) = 4.9 Hz. j Overlapping signals. and with planes C(6)–N(2)–C(7) and N(1)–Au–C(1) of 35.9 and 43.28 for 5a. The dihedral angles between benzene and pyridine rings are 130.0 (1b), 118.8 (1c) and 133.18 (5a). The bite angles of the cycloaurated ligands are 86.6 (1b), 88.3 (1c) and 85.28 (5a), whose values are wider than those in five- Fig. 1 An ORTEP view of complex [AuCl2(pop-C1, N)] 1b. Hydrogen atoms are omitted for clarity.Fig. 2 An ORTEP view of complex [AuCl2(ptp-C1, N)] 1c. Hydrogen atoms are omitted for clarity. Fig. 3 An ORTEP view of complex [AuCl(pap-C1, N)(PPh3)]BF4 5a. Hydrogen atoms and tetrafluoroborate anion are omitted for clarity. Fig. 4 An ORTEP view of complex [AuCl(pap-C1)(PEt3)2]Cl 7a. Hydrogen atoms and chloride anion are omitted for clarity.4098 J. Chem. Soc., Dalton Trans., 1998, 4095–4100 membered auracycles derived from N,N-dimethylbenzylamine [82.2(4)8],17 4,4-dimethyl-2-phenyl-1,3-oxazoline [81.7(3)8],10 4- butyl-N-(3,4,5-trimethoxybenzylidene)aniline [81.41(14)8] 11 and 4,49-dimethylazobenzene [80.1(2)8] 18 and are comparable to the values in the six-membered auracycles [AuCl(C6H4CH2C5H4NC1, N)(PPh3)]BF4 [85.8(4)8],6 [AuCl2(C6H4CMe2C5H4N-C1, N)] [85.7(1)8],7 [AuCl2(C6H4COC5H4N-C1, N)] [89.5(3)8] 6 and [AuCl2(pap-C1, N)] [87.3(9)8].12 The pap–Au ring structure in 5a was quite similar to that of [AuCl2(pap-C1, N)].12 In complexes 1b and 1c the Au–Cl(2) bond [2.369(5) (1b) and 2.384(4) Å (1c)] is longer than Au–Cl(1) [2.275(4) (1b) and 2.277(4) Å (1c)] owing to the greater trans influence of the aryl carbon atom than the nitrogen atom.Concerning the structure of complex 7a, as expected from spectroscopic data it was confirmed that two PEt3 ligands are located trans to each other and the 2-(2-pyridylamino)phenyl ligand is co-ordinated to Au only through the C(1) atom, forming a neutral complex.The phenyl ring in the 2-(2- pyridylamino)phenyl moiety is located nearly perpendicular to the gold(III) square plane (dihedral angle between two planes is 78.88). There are no gold–nitrogen bonding interactions [N(1), 3.218(8); N(2), 5.109(9) Å], excluding a five-co-ordinate gold(III) configuration. Table 2 Selected bond distances (Å) and angles (8) with estimated standard deviations (e.s.d.s) in parentheses for complex 1b Au–C(1) Au–Cl(1) C(1)–C(6) C(7)–O C(1)–Au–N C(1)–Au–Cl(2) N–Au–Cl(1) 2.03(2) 2.275(4) 1.37(2) 1.41(2) 86.6(6) 177.7(4) 175.2(3) Au–N Au–Cl(2) C(6)–O C(7)–N C(1)–Au–Cl(1) N–Au–Cl(2) Cl(1)–Au–Cl(2) 2.02(1) 2.369(5) 1.41(2) 1.35(2) 89.5(4) 91.7(4) 92.1(2) Table 3 Selected bond distances (Å) and angles (8) with e.s.d.s in parentheses for complex 1c Au–C(1) Au–Cl(1) C(1)–C(6) C(7)–S C(1)–Au–N C(1)–Au–Cl(2) N–Au–Cl(1) 2.04(2) 2.277(4) 1.41(3) 1.76(2) 88.3(6) 178.4(5) 176.8(5) Au–N Au–Cl(2) C(6)–S C(7)–N C(1)–Au–Cl(1) N–Au–Cl(2) Cl(1)–Au–Cl(2) 2.07(1) 2.384(4) 1.77(2) 1.36(2) 89.6(5) 90.2(4) 91.9(1) Table 4 Selected bond distances (Å) and angles (8) with e.s.d.s in parentheses for complex 5a Au–C(1) Au–Cl C(1)–C(6) C(7)–N(2) C(1)–Au–N(1) C(1)–Au–Cl N–Au–P 2.06(1) 2.347(4) 1.37(2) 1.40(2) 85.2(5) 174.9(4) 176.5(3) Au–N(1) Au–P C(6)–N(2) C(7)–N(1) C(1)–Au–P N–Au–Cl Cl(1)–Au–P 2.09(1) 2.319(3) 1.42(2) 1.33(2) 94.5(4) 90.3(3) 89.8(1) Table 5 Selected bond distances (Å) and angles (8) with e.s.d.s in parentheses for complex 7a Au–C(1) Au–P(2) C(1)–C(6) C(7)–N(2) C(1)–Au–P(1) C(1)–Au–Cl P(1)–Au–P(2) 2.039(8) 2.361(3) 1.40(1) 1.37(1) 89.9(2) 173.1(3) 175.2(1) Au–P(1) Au–Cl C(6)–N(2) C(7)–N(1) C(1)–Au–P(2) P(1)–Au–Cl Cl–Au–P(2) 2.365(3) 2.371(3) 1.40(1) 1.32(1) 88.9(3) 87.45(9) 93.2(1) Experimental General The IR spectra were measured on a JASCO FT/IR-420 spectrophotometer, 1H NMR spectra on a JEOL JNM-GX-270 spectrometer using tetramethylsilane as an internal standard.Melting points were determined on a Yanaco MP-500D micro melting-point apparatus and are uncorrected.Conductivity measurements were carried out at 25 8C on a Toa Electronics CM-20E conductometer. 2-Phenoxypyridine 19 and 2-(phenylsulfanyl) pyridine 20 were prepared according to the literature. Other reagents were obtained commercially and used without purification. Syntheses [AuCl2(pap-C1, N)] 1a. An ethanol (5 cm3) solution of 2- anilinopyridine (0.128 g, 0.752 mmol) was added to a solution of H[AuCl4]?4H2O (0.104 g, 0.251 mmol) in the same solvent (5 cm3) and the resulting solution stirred at room temperature.After 15 h, the yellow precipitates obtained were filtered oV and washed with diethyl ether to give complex 1a (0.096 g, 88%), mp 253 8C (decomp.) (Found: C, 30.1; H, 2.1; N, 6.35. C11H9Au- Cl2N2 requires C, 30.25; H, 2.1; N, 6.4%); n& max/cm21 (KBr) 356, 284 (Au–Cl). Complex 1a was also prepared in a similar way using Na- [AuCl4]?2H2O (0.100 g, 0.252 mmol) and 2-anilinopyridine (0.091 g, 0.535 mmol) (yield of 1a: 0.107 g, 93%).[AuCl2(pop-C1, N)] 1b. Method (a). An ethanol solution (5 cm3) of 2-phenoxypyridine (0.045 g, 0.264 mmol) was added to a solution of H[AuCl4]?4H2O (0.102 g, 0.248 mmol) in water (25 cm3), whereupon yellow precipitates appeared. When the resulting suspension was heated at 105 8C for 14 h the precipitates turned to white. They were collected and recrystallized from dichloromethane and hexane to give complex 1b (0.077 g, 71%) as white microcrystals, mp 240 8C (decomp.) (Found: C, 30.15; H, 1.85; N, 3.15.C11H8AuCl2NO requires C, 30.15; H, 1.85; N, 3.2%); n& max/cm21 (KBr) 361, 303 (Au–Cl). Complex 1b was also prepared similarly in 81% yield in propan-2-ol–water (1 : 5). Method (b). An acetonitrile–water suspension (1 : 5, 24 cm3) containing the salt [H2pop][AuCl4] 2b (0.199 g, 0.389 mmol) was heated at 105 8C for 40 h. The resulting precipitates were collected and extracted with hot acetone (20 cm3).The extract was concentrated to give white microcrystals of complex 1b (yield 0.112 g, 66%). Method (c). A yellow suspension of [AuCl3(Hpop)] 3b (0.077 g, 0.163 mmol) in acetonitrile–water (1 : 5, 24 cm3) was heated at 105 8C for 18 h. The resulting white precipitates were collected and recrystallized from dichloromethane and hexane to give complex 1b (0.055 g, 78%). [AuCl2(ptp-C1, N)] 1c. Method (a). An ethanol solution (5 cm3) of 2-(phenylsulfanyl)pyridine (0.048 g, 0.256 mmol) was added to a solution of H[AuCl4]?4H2O (0.102 g, 0.248 mmol) in water (25 cm3), and the resulting mixture heated at 105 8C for 14 h.The resulting yellow suspension was filtered while hot. From the filtrate yellow microcrystals of complex 4c (0.010 g) were precipitated on standing at room temperature. The filter cake was extracted with dichloromethane and the extract concentrated and diluted with hexane to give 1c (0.014 g, 12%). Complex 4c (Found: C, 25.2; H, 1.55; N, 2.65%); n& max/cm21 (KBr) 357 (Au–Cl); dH(DMSO-d6) 7.95 (2 H), 8.16 [1 H, t, 3JHH = 6.8], 8.53 (2 H, m), 8.91 (2 H, m), 10.10 [1 H, d, 3JHH = 6.4 Hz].Complex 1c: mp 239 8C (decomp.) (Found: C, 29.1; H, 1.8; N, 3.1. C11H8AuCl2NS requires C, 29.0; H, 1.85; N, 3.05%); n& max/ cm21 (KBr) 360, 295 (Au–Cl). Complex 1c was also prepared similarly in 13% yield in propan-2-ol–water (1 : 5). Method (b). An acetonitrile–water suspension (1 : 5, 30 cm3)J.Chem. Soc., Dalton Trans., 1998, 4095–4100 4099 containing the salt [H2ptp][AuCl4] 2c (0.200 g, 0.379 mmol) was heated at 105 8C for 20 h. The resulting mixture was filtered while hot. From the filtrate yellow microcrystals of complex 4c (0.023 g) were precipitated, while from the filter cake after washing with hot water using a Soxhlet extraction apparatus yellow microcrystals of 1c were obtained (0.035 g, 20%). Method (c). An acetonitrile–water suspension (1 : 5, 30 cm3) containing the adduct [AuCl3(Hptp)] 3c (0.201 g, 0.409 mmol) was heated at 105 8C for 20 h and then the mixture was filtered while hot.From the filtrate yellow microcrystals of complex 4c (0.022 g) were obtained. The filter cake was extracted with dichloromethane and the extract concentrated and diluted with hexane to give 1c (0.034 g, 18%). [H2pop][AuCl4] 2b. An ethanol (5 cm3) solution of 2- phenoxypyridine (0.171 g, 0.999 mmol) was added to a solution of H[AuCl4]?4H2O (0.203 g, 0.493 mmol) in the same solvent (5 cm3) and the resulting solution stirred at room temperature for 1 d.The resulting mixture was evaporated to dryness and the residue extracted with dichloromethane. The extract was concentrated and diluted with hexane to give yellow microcrystals of complex 2b (0.233 g, 92%), mp 108 8C (decomp.) (Found: C, 26.05; H, 1.9; N, 2.7. C11H10AuCl4NO requires C, 25.85; H, 1.95; N, 2.75%); n& max/cm21 (KBr) 360 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 157 S cm2 mol21.[H2ptp][AuCl4] 2c. Complex 2c was obtained as orange microcrystals in a similar way to that described above by the reaction between 2-(phenylsulfanyl)pyridine (0.185 g, 0.990 mmol) and H[AuCl4]?4H2O (0.196 g, 0.475 mmol) in ethanol (10 cm3), yield 0.182 g (73%), mp 149 8C (decomp.) (Found: C, 25.3; H, 1.95; N, 2.7. C11H10AuCl4NS requires C, 25.05; H, 1.9; N, 2.65%); n& max/cm21 (KBr) 360 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 166 S cm2 mol21.[AuCl3(Hpop)] 3b. Method (a). An acetonitrile (5 cm3) solution of 2-phenoxypyridine (0.023 g, 0.134 mmol) was added to a solution of Na[AuCl4]?2H2O (0.051 g, 0.127 mmol) in water (25 cm3), whereupon bright yellow microcrystals were precipitated. After the resulting suspension was stirred for 15 h at room temperature, the crystals were filtered oV to give complex 3b (0.060 g, 74%), mp 167 8C (decomp.) (Found: C, 27.95; H, 1.95; N, 2.95. C11H9AuCl3NO requires C, 27.85; H, 1.9; N, 2.95%); n& max/cm21 (KBr) 365 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 1.3 S cm2 mol21.Method (b). Water (25 cm3) was added to an acetonitrile solution of complex 2b (0.125 g, 0.245 mmol) and the resulting suspension stirred for 16 h at 25 8C. Yellow precipitates were collected and washed with water to give 3b (0.106 g, 91%). [AuCl3(Hptp)] 3c. Method (a). Complex 3c was obtained as bright orange microcrystals in a similar way to that described above by the reaction between 2-(phenylsulfanyl)pyridine (0.075 g, 0.402 mmol) and Na[AuCl4]?2H2O (0.151 g, 0.380 mmol) in acetonitrile–water (1 : 5, 30 cm3, yield 0.145 g (78%), mp 168 8C (decomp.) (Found: C, 27.0; H, 1.9; N, 2.85.C11H9AuCl3NS requires C, 26.95; H, 1.85; N, 2.85%); n& max/cm21 (KBr) 362 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 0.8 S cm2 mol21. Method (b). Water (25 cm3) was added to an acetonitrile solution of complex 2c (0.103 g, 0.196 mmol) and the resulting suspension stirred for 13 h at 25 8C.Yellow precipitates were collected and washed with water to give 3c (0.106 g, 92%). [AuCl(pap-C1, N)(PPh3)]BF4 5a. Triphenylphosphine (0.063 g, 0.241 mmol) and then sodium tetrafluoroborate (0.028 g, 0.255 mmol) were added to an acetone solution (10 cm3) of complex 1a (0.100 g, 0.230 mmol). The resulting solution was stirred for 15 h and then the volatile materials were evaporated in vacuo. The residue was extracted with dichloromethane and the extract concentrated and diluted with diethyl ether to yield yellow microcrystals of 5a (0.139 g, 80%), mp 153 8C (decomp.) (Found: C, 46.7; H, 3.35; N, 3.75. C29H24AuBClF4N2P requires C, 46.4; H, 3.2; N, 3.65%); n& max/cm21 (KBr) 1060 (BF4 2), 313 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 169 S cm2 mol21. [AuCl(pap-C1)(PPh3)2]Cl 6a.Triphenylphosphine (0.288 g, 1.10 mmol) was added to an acetone solution (10 cm3) of complex 1a (0.201 g, 0.460 mmol). After the mixture was stirred at room temperature for 1 d the volatile materials were evaporated in vacuo.The residue was extracted with dichloromethane and the extract concentrated and diluted with diethyl ether to aVord yellowish white microcrystals of 6a?H2O (0.287 g, 90%), mp 131 8C (Found: C, 57.5; H, 4.25; N, 2.85. C47H41AuCl2N2- OP2 requires C, 57.6; H, 4.2; N, 2.85%); n& max/cm21 (KBr) 295 (Au–Cl); LM(1.0 × 1023 mol dm23, MeOH) 123 S cm2 mol21. [AuCl(pap-C1)(PEt3)2]Cl 7a. Complex 7a was obtained as beige microcrystals in a similar way to that described above by the reaction between 1a (0.105 g, 0.239 mmol) and PEt3 (0.118 g, 1.00 mmol) in acetone (10 cm3), yield 0.128 g (80%), mp 139 8C (Found: C, 41.15; H, 5.9; N, 4.15.C23H39AuCl2N2P2 requires C, 41.0; H, 5.85; N, 4.1%); n& max/cm21 (KBr) 300 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 24 S cm2 mol21. [AuCl2(pop-C1)(PPh3)] 8b. Triphenylphosphine (0.031 g, 0.118 mmol) was added to a dichloromethane solution (10 cm3) of complex 1b (0.050 g, 0.114 mmol).The resulting solution was stirred at room temperature for 18 h and then filtered. The filtrate was concentrated and diluted with diethyl ether to give white microcrystals of 8b (0.068g, 84%), mp 146 8C (Found: C, 49.75; H, 3.45; N, 2.15. C29H23AuCl2NOP requires C, 49.75; H, 3.3; N, 2.0%); n& max/cm21 (KBr) 325, 301 (Au–Cl); LM(1.0 × 1023 mol dm23, MeOH) 1.4 S cm2 mol21. [AuCl2(ptp-C1)(PPh3)] 8c. Complex 8c was obtained as white microcrystals in a similar way to that described above by the reaction between 1c (0.062 g, 0.137 mmol) and PPh3 (0.038 g, 0.144 mmol), yield 0.084 g (86%), mp 170 8C (Found: C, 48.8; H, 3.2; N, 1.9.C29H23AuCl2NPS requires C, 48.6; H, 3.25; N, 1.95%); n& max/cm21 (KBr) 316, 301 (Au–Cl); LM(1.0 × 1023 mol dm23, acetone) 7.3 S cm2 mol21. X-Ray crystallography Suitable crystals of [AuCl2(pop-C1, N)] 1b, [AuCl2(ptp-C1, N)] 1c, [AuCl(pap-C1, N)(PPh3)]BF4 5a and [AuCl(pap-C1)- (PEt3)2]Cl 7a were grown from dichloromethane and hexane except for 1b (dichloromethane and diethyl ether).Details of the crystal data, data collection and refinement are summarized in Table 6. Measurements were made on Rigaku AFC7S (for 1b, 5a and 7a) and Enraf-Nonius CAD4 (for 1c) diVractometers with graphite-monochromated Mo-Ka radiation (l = 0.71069 Å) at 23 8C except for 1c (20 8C). Cell constants were obtained from a least-squares refinement of the setting angles of 25 reflections in the range 29.2 < 2q < 30.18 for 1b, 20.0 < 2q < 30.08 for 1c, 27.0 < 2q < 29.238 for 5a and 33.7 < 2q < 34.88 for 7a.Intensity data were collected by the w–2q scan technique and corrected for Lorentz-polarization eVects and absorption. All the calculations for 1b, 5a and 7a were performed using the TEXSAN software package,21 whereas those for 1c were carried out on a VAX station 4000 90A computer using a MO1EN program package.22 The structures of 1b and 1c were solved by direct methods and expanded using Fourier techniques.The non-hydrogen atoms were refined anisotropically, and the hydrogen atoms included but not refined. The structure of 5a was solved by heavy-atom Patterson methods and expanded using Fourier techniques. All non-hydrogen atoms except for the tetrafluoroborate anion were refined anisotropically. The position of NH was freely4100 J. Chem. Soc., Dalton Trans., 1998, 4095–4100 Table 6 Crystallographic data for complexes 1b, 1c, 5a and 7a Formula M Crystal system Space group a/Å b/Å c/Å b/8 U/Å3 Z Dc/g cm23 Crystal dimensions/mm m(Mo-Ka)/cm21 No.measured reflections No. unique observed reflections [I > 3s(I )] RR9 1b C11H8AuCl2NO 438.06 Orthorhombic P212121 8.22(2) 18.15(2) 7.76(1) 1157(2) 4 2.513 0.20 × 0.40 × 0.40 131.84 1577 1410 0.039 0.048 1c C11H8AuCl2NS 454.13 Monoclinic P21/n 8.488(3) 14.363(3) 10.618(4) 103.74(2) 1257.3(7) 4 2.40 0.41 × 0.32 × 0.22 122.4 2301 1617 0.036 0.045 5a C29H24AuBClF4N2P 750.72 Orthorhombic Pbca 20.899(3) 18.940(2) 14.319(2) 5667(2) 8 1.759 0.35 × 0.20 × 0.30 54.08 7169 3249 0.046 0.062 7a C23H39AuCl2N2P2 673.39 Orthorhombic P212121 15.181(3) 15.873(1) 11.535(2) 2779.6(7) 4 1.609 0.20 × 0.15 × 0.45 56.31 3596 2708 0.028 0.029 refined but its isotropic thermal parameter was fixed.The other hydrogen atoms were included but not refined. As for the refinement of the tetrafluoroborate anion, the atom B(1) was refined isotropically; F(1), F(2), F(3) and F(4) were treated as an idealized rigid group with a common isotropic atomic displacement parameter because the refinement of individual parameters of those atoms failed.The structure of 7a was solved by direct methods and expanded using Fourier techniques. All the non-hydrogen atoms were refined anisotropically. The position of NH was freely refined but its isotropic parameters were fixed. The other hydrogen atoms were included but not refined.CCDC reference number 186/1206. References 1 J. Dehand and M. PfeVer, Coord. Chem. Rev., 1976, 18, 327; M. I. Bruce, Angew. Chem., Int. Ed. Engl., 1977, 16, 73; I. Omae, Chem. Rev., 1979, 79, 287; Coord. Chem. Rev., 1980, 32, 235. 2 E. C. Constable and T. A. Leese, J. Organomet. Chem., 1989, 363, 419. 3 C. W. Chan, W. T. Wong and C. M. Che, Inorg. Chem., 1994, 33, 1266. 4 H. O Liu, T. C. Cheung, S. M. Peng and C. M. Che, J. Chem. Soc., Chem. Commun., 1995, 1787. 5 M. A. Cinellu, A. Zucca, S. Stoccoro, G. Minghetti, M. Manassero and M. Sansoni, J. Chem. Soc., Dalton Trans., 1995, 2865. 6 Y. Fuchita, H. Ieda, Y. Tsunemune, J. Kinoshita-Kawashima and H. Kawano, J. Chem. Soc., Dalton Trans., 1998, 791. 7 M. A. Cinellu, A. Zucca, S. Stoccoro, G. Minghetti, M. Manassero and M. Sansoni, J. Chem. Soc., Dalton Trans., 1996, 4217. 8 J. Vicente, M. T. Chicote and M. D. Bermudez, Inorg. Chim. Acta, 1982, 63, 35. 9 J. Vicente, M. T. Chicote and M. D. Bermudez, J. Organomet. Chem., 1984, 268, 191. 10 P. A. Bonnardel, R. V. Parish and R. G. Pritchard, J. Chem. Soc., Dalton Trans., 1996, 3185. 11 J. Vicente, M. D. Bermudez, F. J. Carrion and P. G. Jones, Chem. Ber., 1996, 129, 1301. 12 M. Nonoyama, K. Nakajima and K. Nonoyama, Polyhedron, 1997, 16, 4039. 13 P. K. Byers and A. J. Canty, Organometallics, 1990, 9, 210. 14 D. Rasala, Bull. Soc. Chim. Fr., 1992, 129, 79. 15 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976; H. O Liu, T. C. Cheung, S. M. Peng and C. M. Che, J. Chem. Soc., Chem. Commun., 1995, 1787. 16 J. Vicente, M. D. Bermudez, M. P. Carrillo and P. G. Jones, J. Organomet. Chem., 1993, 456, 305. 17 J. Vicente, M. T. Chicote, M. D. Bermudez, P. G. Jones and G. M. Sheldrick, J. Chem. Res., 1985, (S) 72; (M) 954. 18 J. Vicente, M. D. Bermudez, M. P. Carrillo and P. G. Jones, J. Chem. Soc., Dalton Trans., 1992, 1975. 19 K. Fujiwara, K. Kondo, I. Yokomichi, F. Kimura, T. Haga and R. Nishiyama, Agr. Biol. Chem., 1970, 34, 68. 20 S. Kondo, M. Nakanishi and K. Tsuda, J. Heterocycl. Chem., 1984, 21, 1243. 21 TEXSAN, Crystal Structure Analysis Package, Molecular Structure Corporation, Houston, TX, 1985 and 1992. 22 C. K. Fair, MOLEN Structure Determination System, Delft Instruments, Delft, 1990. Paper 8/07108J

ISSN:1477-9226    

DOI:10.1039/a807108j

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 4101–4108 4101 Synthesis and crystal structure of gold(I) complexes with triazole and triphenylphosphine ligands: monomeric complex [Au(1,2,3-L)- (PPh3)] and dimeric complex [Au(1,2,4-L)(PPh3)]2 (HL 5 triazole) through an Au–Au bond in the solid state Kenji Nomiya,* Ryusuke Noguchi, Katsunori Ohsawa and Kazuhiro Tsuda Department of Materials Science, Faculty of Science, Kanagawa University, Tsuchiya, Hiratsuka, Kanagawa 259-1293, Japan.E-mail: nomiya@info.kanagawa-u.ac.jp Received 10th August 1998, Accepted 21st October 1998 Two novel gold(I)–triphenylphosphine complexes with nitrogen-containing heterocycles, [Au(1,2,3-L)(PPh3)] 1 and [Au(1,2,4-L)(PPh3)]2?xH2O (x = 0.5–1.0) 2 (HL = triazole) were synthesized from stoichiometric reactions of a precursor complex [AuCl(PPh3)] with HL in acetone in the presence of NaOH, and isolated as colorless needles and cubic crystals, respectively. The crystal structures of 1 and 2 were determined by single-crystal X-ray diVraction.Complexes 1 and 2 were also fully characterized by complete elemental analyses, TG/DTA and FT-IR in the solid state and by solution NMR (31P, 1H and 13C) spectroscopies and solution molecular-weight measurements. Complex 1 consisted of a monomeric 2-coordinate AuNP core both in the solid state and in solution, while, in contrast, 2 comprised a dimeric (AuNP)2 core through an Au–Au bond in the solid state, but a monomeric AuNP core in solution. Within the two gold(I) complexes composed of very closely related nitrogen-containing heterocycles and a common bulky PPh3 ligand, it was found that aggregation through the Au ? ? ?Au interaction in 2 was overruled in 1.The molecular structures of 1 and 2 were also compared with those of the corresponding silver(I) analogs, [Ag(1,2,3-L)(PPh3)2]n 3 and [Ag(1,2,4-L)(PPh3)2]n 4, the molecular structures of which have been recently determined as helical polymers in the solid state.Over the past 30 years, interest has increased in the coordination chemistry of silver(I) and gold(I) complexes with biological or medicinal activities.1–7 The molecular design and structural determination of such silver(I) and gold(I) complexes with common ligands are an intriguing aspect of bioinorganic chemistry, inorganic syntheses and metal-based drugs. For example, the thiosulfato complexes of silver(I) and gold(I), [Ag- (S2O3)2]32 (STS) with anti-ethylene activity 8 and [Au(S2O3)2]32 (Sanocrisin) with anti-arthritic activity,1,9 are a classical, but prototypic case; the former has 4-coordinate silver(I) with tetrahedral geometry caused by a bridging thiosulfate ligand,10 while the latter has been established as a 2-coordinate linear structure with an AuS2 core.9c Recently, we have realized a combination of several silver(I) and gold(I) complexes with common thiol ligands, such as {Na[Ag(Htma)]?0.5H2O}n (n = 24–34; H3tma = thiomalic acid) 11a and gold thiomalate, {Na2[Au- (tma)]?1.75H2O}n (n = 3–10),3,9a,12a and of {Na[Ag(tsa)]?H2O}n (n = 21–27; H2tsa = thiosalicylic acid) 13 and Na3[Au(tsa)2]? 5H2O.14 These oligomeric silver(I) complexes have displayed eVective antimicrobial activities against selected bacteria, yeast and mold,11a,13 whereas the corresponding gold(I) complexes have shown eVective anti-arthritic activities.15 Structure determination with X-ray analysis of these complexes has been unsuccessful because most of them were hard to crystallize, 3,9a,11b the exception being gold thiomalate, {Na2Cs[Au2- (Htma)(tma)]}n, the crystal structure of which was very recently solved.12k On the other hand, tertiary phosphine ligands have been utilized in order to limit polymerization of the [Au(SR)] unit (HSR = thiol ligand).16 Thus, we have also substantiated another combination of silver(I) and gold(I) complexes consisting of a thiol or N-containing heterocyclic ligand with an auxiliary ligand (PPh3) such as in [Ag(Htsa)(PPh3)3] and [Au(Htsa)(PPh3)],17 and [Ag(im)(PPh3)3] (Him = imidazole) and [Au(im)(PPh3)].18,19 Recent advances in gold chemistry have highlighted the flexible metal–metal bonding modes exhibited by this element.20 Particular attention has been paid to compounds which contain weak intermolecular metal–metal interactions with gold–gold separations typically in the range 2.5–3.5 Å, which are less than 3.60 Å, twice the van der Waals radii for gold.21,22 Schmidbaur et al.have suggested that steric eVects play a decisive role, since the weak forces associated with the Au ? ? ?Au contacts are easily overruled by steric repulsion and other factors such as packing forces. In solution, it is the solvation by solvent molecules which overrules the aggregation through Au ? ? ? Au contacts.23 The aYnity of gold for the nitrogen atom, in comparison with sulfur and phosphorus atoms, is very low indeed, and most compounds with gold–nitrogen bonds are of limited stability.23 Nevertheless, a number of neutral or ionic gold(I) complexes with nitrogen centers in the presence of the auxiliary P-donor ligands have been prepared, e.g., as neutral complexes such as [Au(pyrmd)(PPh3)] (Hpyrmd = 5-fluoro-1-(tetrahydrofuran- 2-yl)-(1H,3H-pyrimidine-2,4-dione),24 [Au(im-2-R)(PPh3)] (R = H, Me, i-Pr, Ph),25 and [Au(pz)(PPh3)] (Hpz = pyrazole),26 and as ionic species such as [Au(Him)(PPh3)]1[Z]2 (HZ = picric acid),27a [Au(NMe3)(PPh3)]ClO4,28 [Au(qncd)(PPh3)]BF4 (qncd = quinuclidine),29 [Au(Q)(PPh3)]ClO4 [Q = 2,6-dimethylpyridine, naphthyridine, 2-(2-pyridyl)benzimidazole].30 However, in gold(I) complexes with N-donor ligands, Au ? ? ? Au interactions are scarce and only a few examples have been recently found such as in [Au(NH2But)(PMe3)]BF4 and [{Au- (PMe3)}2NH(CH2Ph)]BF4,23 and [{Au(PPh3)}4(m-bbzim)]- (ClO4)2 (H2bbzim = 2,29-bibenzimidazole).31 On the other hand, examples of the corresponding neutral silver(I) complexes with an AgN(PPh3) unit have been recently reported, e.g., such as [Ag2(pz)2(PPh3)2] and [Ag2(pz)2- (PPh3)3].32a Very recently, we have isolated two other neutral silver(I) complexes with a 4-coordinate Ag(N)2(PPh3)2 core using 1,2,3-triazole and 1,2,4-triazole ligands, i.e., [Ag(1,2,3-L)-4102 J.Chem. Soc., Dalton Trans., 1998, 4101–4108 (PPh3)2]n 3 and [Ag(1,2,4-L)(PPh3)2]n 4 (HL = triazole), as crystals and characterized them both to be helical polymers in the solid state by single-crystal X-ray diVraction (Fig. 1).33 In a separate account, the development of crystalline AuI compounds with a 2-coordinate AuSP core stems from the discovery of auranofin [(tetraacetylthioglucose)(triethylphosphine) gold(I)],4 which has been used in chemotherapy as an eVective anti-arthritic agent for oral administration, although the mechanism of the action is not established yet. We have also been interested in the anti-arthritic activity of the crystalline nitrogen-centered gold(I) complexes described here.Thus, in the present work, we have aimed at (i) preparing the neutral gold(I) complexes [Au(1,2,3-L)(PPh3)] 1 and [Au(1,2, 4-L)(PPh3)]2?xH2O (x = 0.5–1.0) 2 with two nitrogen-containing heterocyclic ligands, 1,2,3- and 1,2,4-triazole, in the presence of an auxiliary P-donor ligand, (ii) determining by their crystal Fig. 1 (a) Molecular structure of the local coordination around each silver(I) center of [Ag(1,2,3-L)(PPh3)2]n 3 with 50% probability ellipsoids (symmetry operations; i; x, 2y, z 2 0.5, ii; x, 2y, z 1 0.5) and its helical polymer structure.(b) Molecular structure of the local coordination around each silver(I) center of [Ag(1,2,4-L)(PPh3)2]n 4 with 50% probability ellipsoids (symmetry operations iii; 1.5 2 x, y 2 0.5, 0.5 2 z, iv; 1.5 2 x, y 1 0.5, 0.5 2 z) and its helical polymer structure.33 In both polymeric structures, black circles represent silver atoms, and small and large gray circles represent nitrogen and phosphorus atoms, respectively.N N HN N N HN 3 5 5 4 1,2,3-triazole 1,2,4-triazole structure analysis whether an Au ? ? ?Au interaction occurs in them, (iii) comparing their molecular structures with those of the silver(I) analogs 3 and 4 and (iv) elucidating the solution behavior of 1 and 2. Herein we report full details of the synthesis and isolation of 1 as colorless needle crystals and 2 as colorless cubic crystals.The compositional characterization of 1 and 2 in the solid state has been achieved by complete elemental analyses, FT-IR, thermogravimetric and diVerential thermal analyses (TG/DTA) and the structural characterization with single-crystal X-ray crystallography. The Au ? ? ?Au interaction found in 2 in the solid state is a rare example. Also reported are the characterization of 1 and 2 by solution NMR (31P, 1H and 13C) spectroscopies and solution molecular-weight measurements.Results and discussion Compositional characterization Two gold(I) complexes with triazole (HL) and triphenylphosphine ligands, 1 and 2, were synthesized by stoichiometric reactions in acetone of the precursor complex [AuCl(PPh3)] with HL in the presence of NaOH, and isolated in 51.0% (0.29 g) and 53.6% (0.29 g) yields, respectively, as colorless crystals soluble in most organic solvents, by a vapor diVusion method with benzene/hexane as the internal/external solvents.Complete elemental analyses of 1 and 2 for C, H, N, P and Au, for the samples dried overnight at room temperature under 1023–1024 Torr, showed that their compositions had molar ratios of AuI:L:PPh3 = 1:1:1. Their TG/DTA measurements done under atmospheric conditions confirmed the absence of any solvated molecules for 1 because no weight loss was observed below the decomposition temperature 198 8C and the presence of 0.5–1.0 hydrated water for 2 because 1.31% weight loss was observed below the decomposition temperature 195 8C.IR measurements confirm the presence of coordinated PPh3 molecules in 1 as typical vibrational bands at 1479, 1435, 748, 711, 691, 547 and 504 cm21 and at 1487, 1436, 747, 712, 694, 544 and 500 cm21 in 2. In both complexes, the IR measurements also show that L coordinates to the gold(I) atom as a triazolate anion, but not as a neutral triazole, because multiple vibrational bands due to N–H stretchings observed in free HL in the 3100–2600 cm21 region disappear.Thus, molecular formulae of 1 and 2 in the solid state can then be represented as having a general formula of [AuL(PPh3)]n. From single-crystal X-ray analysis, described later, the gold(I) compounds 1 and 2 are a monomer and dimer in the solid state, respectively, which is in contrast with the silver(I) analogs 3 and 4 which are helical polymers in the solid state (Fig. 1).33 On the other hand, molecular weight measurements in acetone solution revealed that both 1 and 2 were present as a monomeric species in solution.Probably these complexes are also monomeric in CH2Cl2 solution, because their solution 31P NMR spectra are virtually identical. Molecular structures Single crystals suitable for single-crystal X-ray analysis were obtained for 1 and 2. The molecular structures of 1 and 2 with the atom numbering scheme are depicted in Figs. 2 and 3, respectively. Selected bond distances and angles with their estimated standard deviations are listed in Table 1.The neutral complex 1 is a monomer consisting of an AuL(PPh3) core coordinated by a triazolate anion, in which the geometry around the gold(I) atom is described as almost linear coordination, the P–Au–N angle being 178.7(5)8. Among the three nitrogen atoms of 1,2,3-triazolate, the N(2) and N(3) atoms do not participate in the coordination. In Table 2, bond distances and angles of the previously reported neutral and ionic gold(I) complexes are listed.The bond distance Au–N [1.98(2) Å] in 1 is slightly shorter, but the distance Au–PJ. Chem. Soc., Dalton Trans., 1998, 4101–4108 4103 [2.229(5) Å] in 1 is similar to those in the reported neutral complexes; [Au(L1)(PPh3)] (HL1 = 1-methylthymine) [Au–N and Au–P, 2.20(1) and 2.240(5) Å, respectively],36n [Au(L2)(PPh3)] (HL2 = 6-methylpyridone) [2.077(9) and 2.236(3) Å],27b [Au(pyrmd)(PPh3)] [ 2.042(24) and 2.235(7) Å],24 [Au(L3)- (PPh3)] (HL3 = 3,7-dihydro-1,3-dimethyl-1H-purine-2,6-dione) [2.047(6) and 2.231(2) Å] 35 and [Au(L4)(PPh3)] (HL4 = 7-azaindole) [2.033(5) and 2.233(1) Å].34 The most interesting feature in 2 is the presence of an intramolecular Au(1)–Au(2) distance of 3.1971(6) Å, which is significantly less than twice the van der Waals radii for gold, 3.60 Å, indicating a weak intramolecular interaction between the two AuL(PPh3) cores.This complex with one water of hydration in the crystals has an isolated dimeric unit without any intermolecular interaction.In this dimeric complex 2, the N(2) and N(4) atoms of 1,2,4-triazolate similarly do not participate in the coordination to the gold(I). The two Au–N distances [Au(1)–N(1a) 2.026(7), Au(2)–N(1b) 2.037(7) Å] and the two Au–P distances [Au(1)–P(1) 2.243(2), Au(2)–P(2) 2.238(2) Å] are comparable with those of 1. The geometry around each gold(I) atom was also described as almost linear coordination, the P(1)–Au(1)–N(1a) and P(2)–Au(2)–N(1b) angles being 177.1(2) and 172.7(3)8, respectively.In the very closely related Fig. 2 Molecular structure of [Au(1,2,3-L)(PPh3)] 1 with 50% probability ellipsoids. Fig. 3 Molecular structure of [Au(1,2,4-L)(PPh3)]2 2 with 50% probability ellipsoids. complex 1, there exists neither an intermolecular nor intramolecular gold(I)–gold(I) interaction. The AuNP core with the gold(I)–gold(I) interaction is not very common as listed in Table 3; the only precedents are the ionic complexes, [Au(NH2But)- (PMe3)]1BF2 4 with the Au–Au distance of 3.047(1) Å and [{Au(PMe3)}2NH(CH2Ph)]1BF4 2 with 3.171(1) and 3.143(1) Å, Table 3.23 Thus, the gold(I)–gold(I) interaction within the neutral N–Au–P complex found in 2 is a very rare case.The pairing of gold(I) atoms either intra- or intermolecularly and the formation of polymers via metal–metal interactions are well-established phenomena in gold(I) coordination chemistry.21a As to the gold(I)–gold(I) interaction, it has been suggested that neither is there a correlation between the nature of the gold(I)–gold(I) interaction, i.e., intramolecular vs.intermolecular, and the Au–Au distance. Recently it has been suggested, however, that the steric eVects play a decisive role in the formation of gold(I)–gold(I) contacts.23 In the present complexes 1 and 2 with a common bulky PPh3 ligand, the diVerence between their ligands is only in the position of one nitrogen atom, i.e., N(3) or N(4), within the triazolate ring.The diVerence in basicity of their nitrogen atoms should not be very large, because both nitrogen atoms can coordinate to silver(I) atoms as recently found in the corresponding silver(I) analogs 3 and 4, where the triazolate ligands bridge two silver(I) atoms to form helical polymer structures. In 2, one triazolate ring (N1b–C5b) nearly overlaps with one (C31–C36) of phenyl rings in the PPh3 group with a separation in the range of 3.4–4.0 Å, suggesting the possibility of stabilization by a stacking interaction between them.However, since this complex is not symmetrical, the stacking of the other triazolate ring (N1a–C5a) is not observed. Thus, the features observed in 1 and 2 are in contrast to those of [Au(NH2But)- (PMe3)]1BF4 2 and [Au(NH2But)(PMePh2)]1BF4 2 with and Table 1 Selected bond lengths (Å) and angles (8) for complexes 1 and 2 1 Au–P1 Au–N1 P1–C11 P1–C21 P1–C31 N1–N2 N2–N3 N1–C5 N3–C4 P1–Au–N1 Au–P1–C11 Au–P1–C21 Au–P1–C31 Au–N1–N2 Au–N1–C5 N1–N2–N3 C11–P1–C21 C11–P1–C31 C21–P1–C31 N2–N1–C5 N2–N3–C4 2.229(5) 1.98(2) 1.80(2) 1.83(2) 1.82(3) 1.31(2) 1.33(2) 1.36(3) 1.35(4) 178.7(5) 111.9(8) 114.2(7) 112.6(9) 121(1) 130(1) 111(1) 108(1) 103(1) 104(1) 107(2) 104(2) 2 Au1–Au2 Au1–P1 Au2–P2 Au1–N1a Au2–N1b P1–C11 P1–C21 P1–C31 P2–C41 P2–C51 P2–C61 N1a–N2a N1a–C5a N4a–C3a N4a–C5a N1b–N2b N1b–C5b N2b–C3b N4b–C3b N4b–C5b P1–Au1–N1a P2–Au2–N1b Au1–Au2–P2 Au2–Au1–P1 Au1–Au2–N1b Au2–Au1–N1a Au1–N1a–N2a Au1–N1a–C5a Au2–N1b–N2b Au2–N1b–C5b Au1–P1–C11 Au1–P1–C21 Au1–P1–C31 Au2–P2–C41 Au2–P2–C51 Au2–P2–C61 3.1971(6) 2.243(2) 2.238(2) 2.026(7) 2.037(7) 1.826(9) 1.807(8) 1.795(9) 1.831(9) 1.803(9) 1.826(8) 1.364(9) 1.32(1) 1.31(1) 1.32(1) 1.367(10) 1.34(1) 1.33(1) 1.32(1) 1.30(1) 177.1(2) 172.7(3) 101.06(6) 98.21(6) 85.0(2) 83.7(2) 121.6(6) 132.4(8) 122.7(7) 130.1(8) 114.8(3) 110.2(3) 113.1(3) 108.7(3) 117.0(3) 113.3(3)4104 J.Chem.Soc., Dalton Trans., 1998, 4101–4108 Table 2 Comparison of Au–N and Au–P distances (Å) and P–Au–N angle (8) for structures containing AuNP core [Au(NMe3)(PPh3)]ClO4 [Au(qncd)(PPh3)]BF4 [Au(dmpy)(PPh3)]ClO4 [Au(napy)(PPh3)]ClO4 [Au(pbzim)(PPh3)]ClO4 [{Au(PPh3)}4(m-bbzim)](ClO4)2 [Au(NH2But)(PMe3)]BF4 [Au(NH2But)(PPh2Me)]BF4 [{Au(PMe3)}2NH(CH2Ph)]BF4 [Au(L1)(PPh3)] [Au(L2)(PPh3)] [Au(pyrmd)(PPh3)] [Au(L4)(PPh3)] [Au(L3)(PPh3)] [{Au(PPh3)}2(m-bbzim)] [Au(1,2,3-L)(PPh3)] 1 [Au(1,2,4-L)(PPh3)]2 2 Au–N 2.108(7) 2.11(1) 2.091(3) 2.093(13) 2.075(4) 2.03(2) 2.05(1) 2.06(1) 2.06(2) 2.13(1) 2.11(1) 2.105(8) 2.071(8) 2.073(8) 2.20(1) 2.077(9) 2.042(24) 2.033(5) 2.047(6) 2.053(9) 1.98(2) 2.026(7) 2.037(7) Au–P 2.231(2) 2.240(4) 2.233(4) 2.230(4) 2.238(1) 2.236(6) 2.233(5) 2.239(5) 2.233(5) 2.236(4) 2.235(3) 2.235(3) 2.246(3) 2.248(3) 2.240(5) 2.236(3) 2.235(7) 2.233(1) 2.231(2) 2.228(3) 2.229(5) 2.243(2) 2.238(2) P–Au–N 179.3(2) 173.0(3) 178.8(3) 174.3(4) 172.4(1) 172.3(5) 173.9(5) 171.2(4) 174.7(5) 175.7(4) 174.8(3) 176.4(2) 178.7(4) 173.4(3) 174.5(6) 176.6(2) 176.1(2) 176.9(3) 178.7(5) 177.1(2) 172.7(3) Ref. 28 29 30 30 30 31 23 23 23 36(n) 27(b) 24 34 35 31 a a dmpy = 2,6-dimethylpyridine; napy = 1,8-naphthyridine; pbzim = 2-(2-pyridyl)benzimidazole. a This work. without the gold(I)–gold(I) interaction, respectively, because it has been considered that steric eVects of the bulky But and PMePh2 groups prevent the intimate approach required for metal–metal contact in the [Au(NH2But)(PMePh2)]1BF4 2 complex.23 Solution NMR (31P, 1H and 13C) The 31P NMR spectra (Fig. 4a, 4c) measured at room temperature in CD2Cl2 of 1 and 2 show only one resonance at d 31.1 and 31.5, respectively, the chemical shifts of which are in the region usually observed for the PPh3 ligands coordinated to gold(I) and can be compared with those of related compounds: [Au(im)(PPh3)] at d 32.46 and [AuCl(PPh3)] at d 33.40.18,37 These resonances are observed at much lower field than those of PPh3 coordinated to a silver(I) atom, e.g., 3 at d 6.48,33 4 at d 4.84,33 [Ag(im)(PPh3)3] at d 3.71,18 [Ag2(pz)2(PPh3)2] at d 9.87 and [Ag2(pz)2(PPh3)3] at d 6.13.32a The low-temperature 31P NMR measurements showed that (i) the single peak in 1 in CD2Cl2 observed at d 31.1 at room temperature was split into two peaks at d 30.6 and 30.0 with an intensity ratio of 2 : 1 at 290 8C (Fig. 4b), and (ii) the original single-line spectrum was recovered at room temperature from the low-temperature splitting 31P NMR spectrum.On the other hand, the single peak in 2 in CD2Cl2 at d 31.5 at room temperature was observed as a single peak at d 30.5 at 290 8C (Fig. 4d). Both complexes should be present as a monomer in solution, because of their solution molecular weight measurements. Thus, the low-temperature 31P NMR measurements indicate that, in solution at room temperature, the coordination of nitrogen atoms on the triazolate ligand to the gold(I) atom exchange dynamically; in 1 all three nitrogen atoms [N(1), N(2) and N(3)] participate in coordination at room temperature, while in 2 only two nitrogen atoms [N(1) and N(2)] do, but N(4) does not. The 1H and 13C NMR spectra of 1 and 2 measured at room temperature in CD2Cl2 exhibit only one resonance for the coordinating triazolate anion, respectively, also as a result of dynamic exchange.However, in response to the lowtemperature 31P NMR spectra, the single 1H NMR peak at d 7.80 at room temperature for the 1,2,3-triazolate ligand of 1 is changed to the splitting peaks at d 7.75 and 7.82 at 290 8C.The solution NMR (31P, 109Ag, 1H and 13C) spectra at room temperature of the corresponding silver(I) complexes 3 and 4 have been also interpreted as averaged signals resulting from the rapid dynamic exchange among several unequivalent 109Ag Fig. 4 31P-{1H} NMR spectra measured in CD2Cl2 of [Au(1,2,3-L)- (PPh3)] 1 (a) at room temperature and (b) at 290 8C, and of [Au- (1,2, 4-L)(PPh3)]2 2 (c) at room temperature and (d) at 290 8C.J.Chem. Soc., Dalton Trans., 1998, 4101–4108 4105 Table 3 Comparison of Au–Au interactions and distances (Å) for structures containing AuYP core (Y = Cl, SR and N) a Entry Complexes Au–Au Ref. Intramolecular Au–Au interactions 1234567 89 10 11 12 13 [Au2(Cl)2(dppm)] [Au2(Cl)2{[(Ph2P)2C]PMe3}] [Au2(Cl)2{1,19-bis(diphenylphosphino)-1,19-bicyclopropyl}] [Au2(Cl)2(Ph2PCHCHPPh2)] [Au2(SCH2CH2PEt2)2] [Au2(3,4-S2C6H3CH3)(PPh3)2] [Au3(3,4-S2C6H3CH3)(PPh3)3] [Au2{S2CN(C2H4OMe)2}2] [Au4(m-S2C6H3CH3)2(PEt3)2] [Au4Ag(CH2SiMe3)4(m-dppm)2]SO3CF3 [{Au(PPh3)}4(m-bbzim)](ClO4)2 [{Au(PMe3)}2NH(CH2Ph)]BF4 [Au(1,2,4-L)(PPh3)]2 2 3.351(2) 3.000(1) 3.085(1) 3.05(1) 3.104 3.096(2) 2.9624(12) 3.1966(14) 2.7902(6) 3.104(6) 3.058(3) 3.116(6) 3.017(3) 3.2170(9) 3.2773(12) 3.157(1) 3.222(1) 3.171(1) 3.1971(6) 36(c) 36(b) 36(d) 36(e) 36(f) 22 22 20 36(o) 36(q) 31 23 b Intermolecular Au–Au interactions 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 [Au(Cl){2,4,6-(But)3C6H2Ph2}]2 [Au2(Cl)2(dppe)]2 [Au2(Cl)2(dppe)]2 [Au2(Cl)2{Ph2P(CH2)2AsPh2}]2 [Au2(Cl)2{Ph2PCH2As(Ph)CH2PPh2}]2 [Au2(1,3-S2C6H4)(PPh3)2] [Au2{S2CN(C2H4OMe)2}2] [AuSPh(o-OMe)(TPA)] [AuSPh(3,5-Cl2)(TPA)] [Au(NH2But)(PMe3)]BF4 [{Au(PMe3)}2NH(CH2Ph)]BF4 [Au(NH]] CMe2)2]CF3SO3 [Au(C]] ] CSiMe3)(CNBut)] [AuCl(Ph2C]] NH)] {[Au(Ph2C]] NH)2][AuCl2]}2 3.440(1) 3.189(1) 3.187(1) 3.221(1) 3.21 3.141(1) 3.0834(8) 3.1572(7) 3.263(2) 3.341(2) 3.0468(10) 3.047(1) 3.143(1) 3.1663(5) 3.1705(5) 3.1244(10) 3.3633(5) 3.1944(5) 36(g) 36(h) 36(i) 36(j) 36(a) 22 20 36(p) 36(p) 23 23 36(r) 36(r) 36(s) 36(s) Polymers 29 30 31 32 33 [Au2(Cl)2(Ph2PCHCHPPh2)]n [Au2(Cl)2{Ph2PCH2C(CH2)CH2PPh2}]n [Au2(Cl)2(dppp)]n [Au2(p-tc)2(dppb)]n [Au2(p-tc)2(dpppn)]n 3.043(1) 3.294 3.316 3.094(1) 3.200(1) 36(k) 36(l) 36(m) 21(a) 21(a) dppp = 1,3-bis(diphenylphosphino)propane; dppb = 1,4-bis(diphenylphosphino)butane; dpppn = 1,5-bis(diphenylphosphino)pentane; p-tc = pthiocresol; TPA = 1,3,5-triaza-7-phosphaadamantanetriylphosphine; SPh(o-OMe) and SPh(3,5-Cl2) are benzenethiolate ligands with substituent groups as indicated. a This table is made by supplementing many additional data to the table previously reported by Narayanaswamy et al.21a b This work.species containing ionic species such as [Ag(PPh3)4]1L2 and [Ag(PPh3)2]1L2.33 Conclusion Using nitrogen-containing heterocycles, 1,2,3-triazole and 1,2,4-triazole (HL), in the presence of PPh3, two novel neutral complexes [Au(1,2,3-L)(PPh3)] 1 and [Au(1,2,4-L)(PPh3)]2? xH2O (x = 0.5–1.0) 2 were isolated as colorless crystals in good yields and their crystal structures determined by single-crystal X-ray diVraction.Two heterocyclic ligands in these complexes act as anionic monodentate ligands. Complex 1 consisted of a monomeric 2-coordinate AuNP core in the solid state, while 2 comprised of a dimeric (AuNP)2 core through an Au–Au bond in the solid state.The Au ? ? ?Au interaction found in 2 in the solid state is rare. These features were in contrast to the fact that the two corresponding silver(I) analogs [Ag(1,2,3-L)(PPh3)2]n 3 and [Ag(1,2,4-L)(PPh3)2]n4 are helical polymers constituted by bridging triazolate ligands in the solid state. On the other hand, 1 and 2 in solution were present as a monomeric species. Their solution (31P, 1H and 13C) NMR spectra measured at room temperature were interpreted based on the presence of an equilibrium due to the rapid dynamic exchange on the NMR timescale, the presence of which was evidenced from lowtemperature 31P NMR measurements.The title complexes are also of interest as a possible new type of metal-based drug; studies of their biological activities are planned. Experimental Materials The following were used as received: 1,2,4-triazole, NaAuCl4? 2H2O, NaOH, triphenylphosphine, dichloromethane, ethanol, methanol, diethyl ether, hexane, light petroleum (bp: 30–60 8C), acetone, benzene (all from Wako); 1,2,3-triazole (Aldrich); CD2Cl2, acetone-d6 (Isotec).[AuCl(PPh3)] was prepared according to the literature.37 Benzene should be used in a fume hood as it is toxic.4106 J. Chem. Soc., Dalton Trans., 1998, 4101–4108 Instrumentation/analytical procedures Elemental analyses after overnight drying under 1023–1024 Torr were carried out by Mikroanalytishes Labor Pascher (Remagen, Germany).Thermogravimetric (TG) and diVerential thermal analysis (DTA) were carried out using a Rigaku TG 8101D and TAS 300 data processing system. TG/DTA measurements were run under air with a temperature ramp of 1 8C min21 between 20 and 500 8C. Infrared spectra were recorded on a Nicolet 510 FT-IR spectrometer in KBr disks at room temperature. Molecular weight measurements in acetone solutions based on the vaporimetric method using a vapor pressure osmometer were done by Mikroanalytishes Labor Pascher (Remagen, Germany) and evaluated for 11.285 mg of the complex 1 dissolved in 0.9367 g of acetone and for 11.309 mg of the complex 2 dissolved in 0.8301 g of acetone. 1H NMR (399.65 MHz), 13C-{1H} NMR (100.40 MHz) and 31P-{1H} NMR (161.70 MHz) spectra in solution were recorded at 22 8C in 5 mm outer diameter tubes on a JEOL JNM-EX 400 FT-NMR spectrometer with a JEOL EX-400 NMR data processing system. 1H and 13C-{1H} NMR spectra of the complexes were measured in CD2Cl2 solution with reference to internal SiMe4.Chemical shifts are reported on the d scale and resonances downfield of SiMe4 (d 0) are recorded as positive. 31P-{1H} NMR (161.70 Hz) spectra were measured in CD2Cl2 or acetone-d6 solution with reference to an external standard of 25% H3PO4 in H2O in a sealed capillary. Chemical shifts are reported as negative for resonances upfield of H3PO4 (d 0). Preparations [Au(1,2,3-L)(PPh3)] 1. 0.495 g (1.00 mmol) of [AuCl(PPh3)] and 0.072 g (1.04 mmol) of 1,2,3-HL were dissolved in 100 mL acetone.To the solution 1.0 mL of 1.0 M NaOH aqueous solution (1.00 mmol) was added. During 1 h stirring, white powder of NaCl was produced and it was filtered oV using a folded filter paper (Whatman No. 2). The obtained colorless clear filtrate was evaporated to dryness at 50 8C with a rotaryevaporator. The residue was dissolved in 30 mL benzene. The clear filtrate obtained through a folded filter paper (Whatman No. 2) was added dropwise to 200 mL hexane. White precipitates formed, which were collected on a membrane filter (JG 0.2 mm), washed twice with 20 mL light petroleum and dried thoroughly by suction. Crystallization was performed by vapor diVusion method. The obtained white powder was redissolved in 15 mL benzene and the solution was filtered through a folded filter paper (Whatman No. 2). The colorless filtrate was placed in an internal small vial and hexane used as an external solvent within a screw-capped vial for the vapor diVusion.After 6 h at room temperature, colorless needle crystals began to form. After a few days, the crystals were collected on a membrane filter (JG 0.2 mm), washed twice with 100 mL light petroleum, and dried in vacuo for 2 h. Yield was 0.29 g (51.0%). Relatively light- and thermally-stable, colorless needle crystals obtained as compound 1 were soluble in methanol, ethanol, acetone, dichloromethane, chloroform, benzene and DMSO, but insoluble in diethyl ether, light petroleum, hexane and water.For the sample dried overnight at room temperature under 1023–1024 Torr: Found: C, 45.33; H, 3.22; N, 7.92; P, 5.80; Au, 37.40, total 99.67%. Calc. for C20H17N3PAu or [Au(1,2,3- L)(PPh3)]: C, 45.56; H, 3.25; N, 7.97; P, 5.87; Au, 37.35%. TG/ DTA data under atmospheric conditions: no weight loss was observed below the decomposition temperature; decomposition began around 198 8C with an endothermic peak at 198 8C and exothermic peaks at 212 and 258 8C. Molecular weight measurement: 512 in acetone; calc. 527.3 for [Au(1,2,3-L)(PPh3)]. Some prominent IR bands in the 1700–400 cm21 region (KBr disk): 1479m, 1435vs, 1400w, 1309w, 1226w, 1184w, 1101s, 1070w, 1043s, 998m, 795m, 748s, 711s, 691vs, 547vs, 504vs cm21. 1H NMR measured in CD2Cl2 with reference to internal SiMe4 at room temperature: d 7.50 (15H, m, aryl protons), 7.80 (2H, s, H4 1 H5 of L). 1H NMR in CD2Cl2 at 290 8C: d 7.52, 7.61 (15H, m, aryl), 7.82, 7.75 (2H, s, H4 1 H5 of L). 13C NMR measured in CD2Cl2 with reference to internal SiMe4 at room temperature: d 131.51 (C4 1 C5 of L), 128.79 (d, JCP 64.3, phenyl), 129.74 (d, JCP 11.0, phenyl), 132.53 (s, phenyl), 134.67 (d, JCP 13.1 Hz, phenyl). 31P NMR measured at room temperature with reference to an external 25% aqueous H3PO4 in a sealed capillary: d 31.1 in CD2Cl2 and d 32.1 in acetone-d6. 31P NMR measured in CD2Cl2 at 290 8C by a substitution method: d 30.6, 30.0. [Au(1,2,4-L)(PPh3)]2?xH2O (x 5 0.5–1.0) 2.Compound 2 was isolated in a similar manner to the work-up described above using 0.040 g (1.00 mmol) of solid NaOH, 0.069 g (1.00 mmol) of 1,2,4-HL in 10 mL methanol and 0.495 g (1.00 mmol) of [AuCl(PPh3)] suspended in 20 mL methanol, instead of 1.0 M aqueous NaOH, 1,2,3-HL and 100 mL acetone as the initial solvent. Yield was 0.29 g (53.6% with respect to one hydrated species). Relatively light- and thermally-stable, colorless cubic crystals were soluble in methanol, ethanol, acetone, dichloromethane, chloroform, benzene and DMSO, but insoluble in diethyl ether, light petroleum, hexane and water.For the sample dried overnight at room temperature under 1023–1024 Torr: Found: C, 45.24; H, 3.05; N, 7.90; P, 5.80; Au, 37.20, total 99.19%. Calc. for C20H17N3PAu or [Au(1,2, 4-L)(PPh3)] as a monomer unit: C, 45.56; H, 3.25; N, 7.97; P, 5.87; Au, 37.35%. TG/DTA data under atmospheric conditions: 1.31% weight loss was observed below the decomposition temperature; calc. 0.85% for x = 0.5 and 1.68% for 1.0 in [Au(1,2,4- L)(PPh3)]?xH2O; decomposition began around 195 8C with an endothermic peak at 195 8C and exothermic peaks at 324 and 470 8C. Molecular weight measurement: 545 in acetone; calc. 527.3 for [Au(1,2,4-L)(PPh3)] as a monomer unit. Some prominent IR bands in the 1700–400 cm21 region (KBr disk): 1487s, 1436s, 1375w, 1284w, 1255m, 1192m, 1154m, 1101s, 1082m, 1027w, 997m, 964w, 867m, 854m, 747s, 712m, 694vs, 544vs, 500s cm21. 1H NMR measured in CD2Cl2 with reference to internal SiMe4 at room temperature: d 7.50 (15H, m, aryl protons), 8.05 (2H, s, H3 1 H5 of L), 2.15 (H2O). 13C NMR measured in CD2Cl2 with reference to internal SiMe4 at room temperature: d 150.79 (C3 1 C5 of L), 128.80 (d, JCP 62.3, phenyl), 129.77 (d, JCP 13.1, phenyl), 132.53 (s, phenyl), 134.66 (d, JCP 15.1 Hz, phenyl). 31P NMR measured at room temperature with reference to an external 25% aqueous H3PO4 in a sealed capillary: d 31.5 in CD2Cl2 and d 32.4 in acetone-d6. 31P NMR measured in CD2Cl2 at 290 8C by a substitution method: d 30.5. X-Ray crystallography Compounds 1 and 2 formed colorless needle crystals and colorless cubic crystals, respectively, by vapor diVusion of the benzene–hexane system. During a few days standing of the solutions at room temperature, crystals of suYcient quality for single-crystal X-ray diVraction studies were grown.Each single-crystal of 1 and 2 was mounted on glass fiber and transferred to a Rigaku AFC5S diVractometer. Cell contents and orientation matrix of 1 and 2 were obtained from the leastsquares refinement of 18 and 25 reflections, respectively. The reflection data were collected using w–2q scan with graphitemonochromated Mo-Ka radiation at room temperature. The intensities of three standard reflections which were measured after every 150 reflections remained constant throughout data collection.The data were corrected for Lorentz and polarization eVects and empirical absorption corrections based on y scans were applied to the data. For the overall averaged transmission curve, the transmission factors of 1 and 2 were in the range of 0.49–1.00 and 0.65–1.00, respectively. The structures were solved by direct methods followed by subsequent diVer-J. Chem. Soc., Dalton Trans., 1998, 4101–4108 4107 ence Fourier calculation and refined by a full-matrix leastsquares procedure using TEXSAN package.38 All nonhydrogen atoms, except carbon atoms in the phenyl group of 1, were refined anisotropically, and all hydrogen atoms and the phenyl carbon atoms of 1 isotropically.A summary of crystal data, data collection, and refinement for 1 and 2 is given in Table 4. CCDC reference number 186/1210. References 1 W. Kaim and B. Schwederski, Bioinorganic Chemistry: Inorganic Elements in the Chemistry of Life, John Wiley, New York, 1994, p. 373. 2 M. J. Abrams and B. A. Murrer, Science, 1993, 261, 725. 3 R. C. Elder and M. K. Eidsness, Chem. Rev., 1987, 87, 1027. 4 E. J. Corey, M. Mehrotra and A. U. Khan, Science, 1987, 236, 68; D. L. B. Bryan, Y. Mikuriya, J. C. Hempel, D. Mellinger, M. Hashim and R. F. Pasternack, Inorg. Chem., 1987, 26, 4180; C. K. Mirabelli, R. K. Johnson, D. T. Hill, L. F. Faucette, G. R. Girard, G. Y. Kuo, C. M. Sung and S. T. Crooke, J. Med. Chem., 1986, 29, 218; M. K. Chan and J. O. Minta, Immunopharmacology, 1985, 10, 61; M.T. Razi, P. J. Sadler, D. T. Hill and B. M. Sutton, J. Chem. Soc., Dalton Trans., 1983, 1331; D. T. Hill and B. M. Sutton, Cryst. Struct. Commun., 1980, 9, 679; B. M. Sutton, E. McGusty, D. T. Walz and M. J. DiMartino, J. Med. Chem., 1972, 15, 1095. 5 D. H. Brown and W. E. Smith, Chem. Soc. Rev., 1980, 3, 217. 6 C. F. Shaw, Inorg. Perspect. Biol. Med., 1978, 2, 287. 7 P. J. Sadler, Struct. Bonding (Berlin), 1976, 29, 171. 8 H. Veen and A.A. M. Kwakkenbos, Sci. Hortic. (Amsterdam), 1982, 1983, 18, 277; H. Veen, Sci. Hortic. (Amsterdam), 1983, 20, 211. 9 (a) M. A. Mazid, M. T. Razi, P. J. Sadler, G. N. Greaves, S. J. Gurman, M. H. J. Koch and J. C. Phillips, J. Chem. Soc., Chem. Commun., 1980, 1261; (b) H. Brown, J. Am. Chem. Soc., 1927, 49, 958; (c) H. Ruben, A. Zalkin, M. O. Faltens and D. H. Templeton, Inorg. Chem., 1974, 13, 1836. 10 R. Stromberg, I.-B. Svensson and A. A. G. Tomlinson, Acta Chem. Scand., 1973, 27, 1192; R. Stomberg, I.-B.Svensson, A. A. G. Tomlinson and I. Persdotter, Acta Chem. Scand., Ser. A, 1982, 36, 579; M. Nakahara, Dictionary of Inorganic Compounds & Complexes, Kodansha Scientific, Japan, 1997, p. 762. 11 (a) K. Nomiya, K. Onoue, Y. Kondoh, N. C. Kasuga, H. Nagano, Table 4 Summary of crystal data Formula M Crystal system Space group a/Å b/Å c/Å b/8 V/Å3 F(000) Z Dc/g cm21 Crystal size/mm No. of reflections used for unit cell dimension (2q range/8) Radiation (l/Å) Scan mode Scan width/8 Scan speed/min21 2q Range/8 m/cm21 Total reflections Unique reflections Observed reflections R,R9 Goodness of fit [Au(1,2,3-L)(PPh3)] C20H17N3PAu 527.31 Orthorhombic P212121 (no. 19) 12.689(4) 13.660(7) 10.93(1) 1894(1) 1008 4 1.849 0.3 × 0.2 × 0.2 18 (20.1–25.9) Mo–Ka (0.71069) w–2q 1.26 1 0.30 tan q 4 6–55 78.85 2492 2492 1561 [I > 2.00s(I)] 0.052, 0.055 1.46 [Au(1,2,4-L)(PPh3)]2?H2O C40H36N6P2Au2O 1072.64 Monoclinic P21/c (no. 14) 13.575(3) 11.623(3) 24.739(2) 101.35(1) 3827(1) 2056 4 1.861 0.3 × 0.3 × 0.3 25 (26.6–29.1) Mo-Ka (0.71069) w–2q 0.73 1 0.30 tan q 8 6–55 78.08 9602 9225 4609 [I > 2.00s(I)] 0.037, 0.028 1.31 R = S[|Fo| 2 |Fc|]/S|Fo|, R9 = [S(w[|Fo| 2 |Fc|]2/[Sw(|Fo|)2]� �� , with w = 4Fo 2/ [s2(Fo 2)].M. Oda and S. Sakuma, Polyhedron, 1995, 14, 1359. The described n = 15–19 should be corrected to n = 24–34 as shown in Polyhedron, 1996, 15, 2303; (b) K. Nomiya, Y. Kondoh, H. Nagano and M. Oda, J.Chem. Soc., Chem. Commun., 1995, 1679; (c) G. R. Lenz and A. E. Martell, Inorg. Chem., 1965, 4, 378; (d ) F. Secheresse, J. Lemerle and J. Lefebvre, Bull. Soc. Chim. Fr., 1974, 2423; (e) K. J. Ellis and A. McAuley, J. Inorg. Nucl. Chem., 1975, 37, 567; ( f ) O. P. Agrawal, K. K. Verma and S. Bhayana, Curr. Sci., 1989, 58, 1201. 12 (a) K. Nomiya, H. Yokoyama, H. Nagano, M. Oda and S. Sakuma, Bull. Chem. Soc. Jpn., 1995, 68, 2875 and refs. therein; (b) M.Delepine, US Pat., 1 994 213, 1935; (c) A.A. Isab and P. J. Sadler, J. Chem. Soc., Dalton Trans., 1976, 1051; (d) C. F. Shaw III, G. Schmitz, H. O. Thompson and P. Witkiewicz, J. Inorg. Biochem., 1979, 10, 317; (e) A. A. Isab and P. J. Sadler, J. Chem. Soc., Dalton Trans., 1982, 135; ( f ) D. H. Brown, M. Paton and W. E. Smith, Inorg. Chim. Acta, 1982, 66, L51; ( g) G. Otiko, M. T. Razi, P. J. Sadler, A. A. Isab and D. L. Rabenstein, J. Inorg. Biochem., 1983, 19, 227; (h) D. T. Hill, B. M. Sutton, A.A. Isab, T. Razi, P. J. Sadler, J. M. Trooster and G. H. M. Calis, Inorg. Chem., 1983, 22, 2936; (i) S. M. Cottrill, H. L. Sharma, D. B. Dyson, R. V. Parish and C. A. McAuliVe, J. Chem. Soc., Perkin Trans. 2, 1989, 53; ( j) M. D. Rhodes, P. J. Sadler, M. D. Scawen and S. Silver, J. Inorg. Biochem., 1992, 46, 129; (k) R. Bau, J. Am. Chem. Soc., 1998, 120, 9380. 13 K. Nomiya, Y. Kondoh, K. Onoue, N. C. Kasuga, H. Nagano, M. Oda, T. Sudoh and S. Sakuma, J. Inorg. Biochem., 1995, 58, 255. The described n = 12–14 should be corrected to n = 21–27. 14 K. Nomiya, H. Yokoyama, H. Nagano, M. Oda and S. Sakuma, J. Inorg. Biochem., 1995, 60, 289. 15 K. Dairiki, Proc. Jpn. Soc. Immunology, 1995, 25, 316 (in Japanese). 16 J. L. Clement and P. S. Jarrett, J. Inorg. Biochem., 1993, 51, 105; P. D. Cookson and E. R. T. Tiekink, J. Coord. Chem., 1992, 26, 313; C. S. W. Harker, E. R. T. Tiekink and M. W. Whitehouse, Inorg. Chim. Acta, 1991, 181, 23; P. D. Cookson and E.R. T. Tiekink, J. Chem. Soc., Dalton Trans., 1993, 259; B. F. Hoskins, L. Zhenrong and E. R. T. Tiekink, Inorg. Chim. Acta, 1989, 158, 7; E. R. T. Tiekink, Z. Kristallogr., 1989, 187, 79; P. D. Cookson and E. R. T. Tiekink, J. Crystallogr. Spectrosc. Res., 1993, 23, 231; F. Bonati, A. Burini, B. R. Pietroni and E. Giorgini, Inorg. Chim. Acta, 1987, 137, 81. 17 K. Nomiya, N. C. Kasuga, I. Takamori and K. Tsuda, Polyhedron, 1998, 17, 3519. 18 K. Nomiya, K. Tsuda, Y. Tanabe and H.Nagano, J. Inorg. Biochem., 1998, 69, 9. 19 K. Nomiya, K. Tsuda, T. Sudoh and M. Oda, J. Inorg. Biochem., 1997, 68, 39. 20 P. Bishop, P. Marsh, A. K. Brisdon, B. J. Brisdon and M. F. Mahon, J. Chem. Soc., Dalton Trans., 1998, 675. 21 (a) R. Narayanaswamy, M. A. Young, E. Parkhurst, M. Ouellette, M. E. Kerr, D. M. Ho, R. C. Elder, A. E. Bruce and M. R. M. Bruce, Inorg. Chem., 1993, 32, 2506; (b) W. B. Jones, J. Yuan, R. Narayanaswamy, M. A. Young, R. C. Elder, A. E. Bruce and M.R. M. Bruce, Inorg. Chem., 1995, 34, 1996. 22 M. C. Gimeno, P. G. Jones, A. Laguna, M. Laguna and R. Terroba, Inorg. Chem., 1994, 33, 3932. 23 K. Angermaier and H. Schmidbaur, J. Chem. Soc., Dalton Trans., 1995, 559. 24 T. Amagi, T. K. Miyamoto, H. Ichida and Y. Sasaki, Bull. Chem. Soc. Jpn., 1989, 62, 1078. 25 M. Felici, B. R. Pietroni and A. Burini, Gazz. Chim. Ital., 1982, 112, 5. 26 G. Minghetti, G. Banditelli and F. Bonati, Inorg. Chem., 1979, 18, 658. 27 (a) F. Bonati, A.Burini, B. R. Pietroni, E. Giorgini and B. Bovio, J. Organomet. Chem., 1988, 344, 119; (b) F. Bonati, A. Burini, B. R. Pietroni and B. Bovio, J. Organomet. Chem., 1985, 296, 301. 28 J. Vicente, M.-T. Chicote, R. Guerrero and P. G. Jones, J. Chem. Soc., Dalton Trans., 1995, 1251. 29 A. Grohmann, J. Riede and H. Schmidbaur, Z. Naturforsch., Teil B, 1992, 47, 1255. 30 M. Munakata, S.-G. Yan, M. Maekawa, M. Akiyama and S. Kitagawa, J. Chem. Soc., Dalton Trans., 1997, 4257. 31 B.-C. Tzeng, D. Li, S.-M. Peng and C.-M. Che, J. Chem. Soc., Dalton Trans., 1993, 2365. 32 (a) G. A. Ardizzoia, G. La Monica, A. Maspero, M. Moret and N. Masciocchi, Inorg. Chem., 1997,321; (b) N. Masciocchi, M. Moret, P. Cairati, A. Sironi, G. A. Ardizzoia and G. La Monica, J. Chem. Soc., Dalton Trans., 1995, 1671; (c) N. Masciocchi, M. Moret, P. Cairati, A. Sironi, G. A. Ardizzoia and G. La Monica, J. Am. Chem. Soc., 1994, 116, 7668. 33 K. Nomiya, K. Tsuda and N. C. Kasuga, J.Chem. Soc., Dalton Trans., 1998, 1653. 34 C.-K. Chan, C.-X. Guo, K.-K. Cheung, D. Li and C.-M. Che, J. Chem. Soc., Dalton Trans., 1994, 3677.4108 J. Chem. Soc., Dalton Trans., 1998, 4101–4108 35 E. Colacio, A. Romerosa, J. Ruiz, P. Roman, J. M. Gutierrez- Zorrilla and M. Martinez-Ripoll, J. Chem. Soc., Dalton Trans., 1989, 2323. 36 (a) A. L. Balch, E. Y. Fung and M. M. Olmstead, J. Am. Chem. Soc., 1990, 112, 5181; (b) H. Schmidbaur, W. Graf and G. Muller, Angew. Chem., Int. Ed. Engl., 1988, 27, 417; (c) H. Schmidbaur, A. Wohlleben, F. Wagner, O. Orama and G. Huttner, Chem. Ber., 1977, 110, 1748; (d) K. Dziwok, J. Lachmann, D. L. Wilkinson, G. Muller and H. Schmidbaur, Chem. Ber., 1990, 123, 423; (e) P. G. Jones, Acta Crystallogr., Sect. B, 1980, 36, 2775; ( f ) W. S. Crane and H. Beall, Inorg. Chim. Acta, 1978, 31, L469; ( g) H. Schmidbaur, G. Weidenhiller, O. Steigelmann and G. Muller, Chem. Ber., 1990, 123, 285; (h) P. A. Bates and J. M. Waters, Inorg. Chim. Acta, 1985, 98, 125; (i) D. S. Eggleston, D. F. Chodosh, G. R. Girard and D. T. Hill, Inorg. Chim. Acta, 1985, 108, 221; ( j) O. M. Ni Dhubhghaill, P. J. Sadler and R. Kuroda, J. Chem. Soc., Dalton Trans., 1990, 2913; (k) D. S. Eggleston, J. V. McArdle and G. E. Zuber, J. Chem. Soc., Dalton Trans., 1987, 677; (l) H. Schmidbaur, C. Paschalidis, O. Steigelmann and G. Muller, Chem. Ber., 1989, 122, 1851; (m) M. K. Cooper, L. E. Mitchell, K. Henrick, M. McPartlin and A. Scott, Inorg. Chim. Acta, 1984, 84, L9; (n) R. Faggiani, H. E. Howard-Lock, C. J. L. Lock and M. A. Turner, Can. J. Chem., 1987, 65, 1568; (o) R. M. Davila, A. Elduque, T. Grant, R. J. Staples and J. P. Fackler, Jr., Inorg. Chem., 1993, 32, 1749; (p) J. M. Forward, D. Bohmann, J. P. Fackler, Jr. and R. J. Staples, Inorg. Chem., 1995, 34, 6330; (q) M. Contel, J. Garrido, M. C. Gimeno and M. Laguna, J. Chem. Soc., Dalton Trans., 1998, 1083; (r) J. Vicente, M.-T. Chicote, M.-D. Abrisqueta, R. Guerrero and P. G. Jones, Angew. Chem., Int. Ed. Engl., 1997, 36, 1203; (s) W. Schneider, A. Bauer and H. Schmidbaur, J. Chem. Soc., Dalton Trans., 1997, 415. 37 M. I. Bruce, B. K. Nicholson and O. B. Shawkataly, Inorg. Synth., 1989, 26, 324; N. C. Baenziger, W. E. Bennett and D. M. SoboroV, Acta Crystallogr., Sect. B, 1976, 32, 962; D. M. L. Goodgame, C. A. O’Mahoney, S. D. Plank and D. J. Williams, Polyhedron, 1993, 12, 2705. 38 TEXSAN, Crystal Structure Analysis Package, Molecular Structure Corporation, Houston, TX 1985 and 1992. Paper 8/06283H

ISSN:1477-9226    

DOI:10.1039/a806283h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 4109–4112 4109 Synthesis and characterization of a novel one-dimensional iron phosphate: [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2(PO4)]?0.5H2O Vítìzslav Zima † and Kwang-Hwa Lii ‡ Institute of Chemistry, Academia Sinica, Taipei, Taiwan, R.O.C. Received 2nd June 1998, Accepted 30th October 1998 An organically templated FeIII phosphate, [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2(PO4)]?0.5H2O, has been synthesized under hydrothermal conditions and characterized by single-crystal X-ray diVraction, Mössbauer spectroscopy and thermogravimetric analysis.The compound crystallizes in the triclinic space group P1� (no. 2) with a = 6.3347(2), b = 13.0075(4), c = 13.7810(5) Å, a = 62.834(1), b = 81.404(1), g = 82.688(1)8, U = 996.67(8) Å3 and Z = 2. The compound is unusual in that it is not only one of the few examples of 1-D organically templated iron phosphate, but also has a new type of chain structure which is built up from tetranuclear iron–oxygen clusters.The tetramer involves an edge-sharing dimer which further links at the shared corners to two more octahedra. These clusters are connected into the 1-D chain by PO4 tetrahedra and have terminal HPO4 and H2PO4 groups, with the piperazinium cations between the chains. Introduction Organically templated transition metal phosphates are of intense current interest because of their novel structures and potential applications as solid catalysts.The first such metallophosphates were prepared with vanadium and molybdenum.1,2 Recently, a number of iron phosphates and fluorophosphates have been reported.3–5 They exhibit a wide structural diversity; many adopt 3-D frameworks and 2-D sheets, whilst few have 1-D chain structures. It makes sense in terms of bond-valence theory that 1-D structures are the least common. For P51 every oxide ligand receives a bond-valence contribution of 1.25 v.u. from the central cation, leaving 0.75 v.u.to be supplied by the rest of the structures. Thus most of the phosphates are highly polymerized, being dominated by frameworks and sheets, irrespective of any other variables that may also influence the stoichiometry and structure.6 To our knowledge [(1R,2R)- C6H10(NH2)2][Fe(OH)(HPO4)2]?H2O, [trans-1,2-C6H10(NH2)2]- [Fe(OH)(HPO4)2]?H2O, and [C3H12N2][FeF(HPO4)2]?xH2O are the only 1-D organically templated iron phosphates (FePOs).7,3 Their structures consist of [FeX(HPO4)2]22 n (X = OH, F) chains of the tanocoite type, with the diprotonated amines inserted in between.In this paper we report the synthesis and structural characterization of a new 1-D FePO templated with piperazinium cations, [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2- (PO4)]?0.5H2O, which contains a novel infinite chain constructed from phosphate groups and tetranuclear iron–oxygen clusters, and show through Mössbauer measurements that it is an iron(III) compound. The tetranuclear cluster has been observed in the iron phosphate mineral leucophosphite 8 and in the synthetic iron phosphate [H3N(CH2)3NH3]2- [Fe4(OH)3(HPO4)2(PO4)3]?xH2O.5 Experimental Synthesis The synthesis was carried out in Teflon-lined acid digestion bombs with an internal volume of 23 cm3 under autogeneous † On leave from Joint Laboratory of Solid State Chemistry, University of Pardubice, Czech Republic.‡ Present address: Department of Chemistry, National Central University, Chungli, Taiwan 320, R.O.C.pressure by heating the starting mixtures at 110 8C for 4 d followed by slow cooling to room temperature at 5 8C h21. Small colorless prismatic crystals of [C4H12N2]1.5[Fe2(OH)- (H2PO4)(HPO4)2(PO4)]?0.5H2O and large colorless crystals of piperazinium hydrogenphosphate monohydrate, (C4H12N2)? HPO4?H2O,9 were obtained as the major products from a reaction mixture of FeCl3?6H2O (1 mmol), H3PO4 (6 mmol), piperazine (6 mmol), and 10 ml of H2O. Subsequently, hydrothermal treatment of FeCl3?6H2O (1 mmol), H3PO4 (6 mmol), piperazine (6 mmol), and 10 ml of a mixture of water and ethylene glycol (volume ratio 1 : 1) under the same reaction conditions gave a pure product of the title compound, as indicated by a comparison of the X-ray powder pattern to that simulated from the atomic coordinates derived from a single-crystal study.The IR spectrum was also measured. However, we are unable to clearly identify the three types of phosphates from the spectrum because the n(P–O) bands overlap, n(OH) bands are weak and broad, and all the phosphates and the water molecule are involved in an extensive hydrogen bonding network.Elemental analysis confirmed the stoichiometry (Found: C, 11.32; H, 3.73; N, 6.22; P, 18.81; Fe, 16.89. Theoretical: C, 11.02; H, 3.70; N, 6.43; P, 18.95; Fe, 17.08%). The yield was 46% based on iron. TGA and Mössbauer spectroscopy Thermogravimetric analysis was performed on a Perkin-Elmer TGA 7 thermal analyzer.The sample was heated to 900 8C at 10 8C min21 in flowing oxygen. The final decomposition products were identified by powder X-ray diVraction. The 57Fe Mössbauer measurements were made on a constantacceleration instrument at 300 K. Isomer shift is reported with respect to an iron foil standard. Single-crystal X-ray diVraction A colorless prismatic crystal of dimensions of 0.1 × 0.05 × 0.05 mm was selected for indexing and intensity data collection on a Siemens Smart-CCD diVractometer equipped with a normal focus, 3 kW sealed tube X-ray source. Intensity data were collected in 2082 frames with increasing w (width of 0.38 per frame).Number of measured reflections and observed unique reflections (Fo > 4s(Fo)): 10301 and 2546. Rint = 0.0786. Empirical absorption corrections were applied using the SADABS program10 for Siemens area detector (Tmin,max = 0.772,4110 J. Chem. Soc., Dalton Trans., 1998, 4109–4112 0.928). On the basis of statistics of intensity distribution and successful solution and refinement of the structure, the space group of the title compound was determined to be P1� (no. 2). The structure was solved by direct methods. The metal and phosphorus atoms were first located, and the oxygen, nitrogen, and carbon atoms were found in diVerence Fourier maps. The hydrogen atoms were not located. The piperazinium cation, which is centered at (0,1/2,0), is disordered over two positions with equal occupancy.The water oxygen Ow initially showed a very large thermal parameter. If the occupancy of Ow is refined, the site occupancy factor obtained is 0.52(1), indicative of a half occupancy of the water of crystallization. The final cycles of least-squares refinement including atomic coordinates and anisotropic thermal parameters for all atoms converged at R1 = 0.0512 and wR2 = 0.1226. Secondary extinction corrections were applied. Structure solution and refinement were performed by using SHELXTL PC, Version 5.11 CCDC reference number 186/1227.Results and discussion TGA and Mössbauer measurements Several mass loss regions are seen in TGA of the title compound (Fig. 1). The first mass loss (ª77 8C) corresponds to loss of lattice water molecule. It is interesting that this mass loss occurs below the boiling point of water. We can attribute this facile loss to the low degree of hydrogen bonding holding the water in the crystal lattice. The observed mass loss between 30 and 150 8C is 1.40% and corresponds to 0.5 mol of H2O per formula unit (calc. 1.38%). The next mass loss (ª325 8C), which is not well resolved from the final steps, is probably due to dehydration of OH2 and hydrogen phosphate groups and deprotonation of piperazinium dications (loss calculated, 11.0%; observed between 150 and 350 8C, 11.9.%). The final steps correspond to the release of organic components, giving Fe(PO3)3 and Fe4(P2O7)3 as the final decomposition products as indicated by powder X-ray diVraction.12 The observed total weight loss of 33.8% is close to that calculated for the loss of 27 H2O and 9 piperazine molecules (32.16%), as indicated by the following equation. 6 [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2(PO4)]?0.5H2O æÆ 4 Fe(PO3)3 1 2 Fe4(P2O7)3 1 27 H2O 1 9 C4H10N2 The room temperature Ml;ssbauer spectrum (Fig. 2) was least-squares fitted by two doublets with a constraint on the area ratio of 1 : 1. The obtained parameters are d (isomer shift) = 0.417 mm s21, DEQ (quadrupole splitting) = 0.279 mm s21 and G (full width at half height) = 0.30 mm s21 for Fe(1) and d = 0.433 mm s21, DEQ = 0.583 mm s21 and G = 0.27 mm s21 for Fe(2).The isomer shifts of both contributions are characteristic Fig. 1 Thermogravimetric analysis of [C4H12N2]1.5[Fe2(OH)(H2PO4)- (HPO4)2(PO4)]?0.5H2O in flowing oxygen at 10 8C min21. of high spin FeIII. The higher value of quadrupole splitting of Fe(2) is because of the greater octahedral distortion due to edge-sharing.Therefore, the composition of the title compound is further defined by TG analysis and Mössbauer spectroscopy. Crystal structure The crystallographic data are listed in Table 1. The bond lengths, and bond-valence sums are given in Table 2. All atoms are in general positions. There are three independent piperazinium cations. The occupancy factors of N(3a), N(3b), C(6) and C(7) are 0.5 because the piperazinium cation which is centered at (0,1/2,0) is disordered over two positions. The site occupancy factor of the water of crystallization is 0.5.Valence sum calculations 13 show the Fe atoms to possess an oxidation state of 31. Both Fe atoms are octahedrally coordinated. Atoms O(1), O(3), O(4), O(5), O(8), O(13), O(16) and O(17) have valence sums of 1.42, 1.15, 1.13, 1.33, 1.08, 1.33, 1.08 and 1.00, respectively, and all other oxygen atoms have values close to 2. The valence sums of O(1), O(5) and O(13) are satisfied by forming hydrogen bonds (see below).Atoms O(3), O(4), O(8), O(16) and O(17) are hydroxo oxygens. Therefore the compound contains H2P(1)O4, HP(2)O4, P(3)O4 and HP(4)O4 groups. Atom O(17) is the hydroxo oxygen bridging three iron atoms. Fig. 2 Mössbauer spectrum of [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2- (PO4)]?0.5H2O at 300 K. Table 1 Crystallographic data for [C4H12N2]1.5[Fe2(OH)(H2PO4)- (HPO4)2(PO4)]?0.5H2O Formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 V/Å3 Z Dc/g cm23 F(000) m(Mo-Ka)/cm21 T/8C l/Å Maximum 2q/8 Reflections collected Unique reflections Observed unique reflections [I > 2s(I)] Number of parameters R1a wR2b Goodness of fit (Dr)max,min/e Å23 C6Fe2H24N3O17.5P4 653.86 Triclinic P1� 6.3347(2) 13.0075(4) 13.7810(5) 62.834(1) 81.404(1) 82.688(1) 996.67(8) 2 2.179 666 18.7 23 0.71073 57.6 10301 4708 2546 317 0.0512 0.1226 1.114 0.90, 20.55 a R1 = S|Fo| 2 |Fc|/S|Fo|.b wR2 = {S[w(Fo 2 2 Fc 2)2]/S[w(Fo 2)2]}� �� , where w = 1/[s2(Fo 2) 1 (0.0592P)2 1 2.36P] with P = (maxFo 2 1 2Fc 2)/3.J.Chem. Soc., Dalton Trans., 1998, 4109–4112 4111 The structure consists of 1-D infinite chains of iron phosphate parallel to the [100] direction, which are H-bonded with the amine groups of piperazinium cations (Fig. 3). The water molecule is also H-bonded as inferred from Ow ? ? ? N(3a) (2.70 Å) and Ow ? ? ? O(16) (2.87 Å). The basic building unit of the chain is a tetranuclear iron–oxygen cluster, which is shown in Fig. 4. The cluster consists of two central FeO6 octahedra that share a common edge, with the two hydroxo oxygens involved in the shared edge serving as corners for two additional FeO6 octahedra. The tetramer possesses 1� symmetry located at the midpoint of the shared edge. Based on the Fe–O distances, the Fe(2)O6 octahedron appears more distorted than Fe(1)O6 because of edge-sharing. These tetranuclear clusters Fig. 3 Structure of [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2(PO4)]? 0.5H2O viewed along the [100] direction. The piperazinium cation, which is centered at (0,1/2,0), is disordered over two positions.Solid circles, C atoms; stippled circles, N atoms; open circles, water oxygens. Table 2 Selected bond lengths (Å) and bond-valence sums (Ss) for [C4H12N2]1.5[Fe2(OH)(H2PO4)(HPO4)2(PO4)]?0.5H2O Fe(1)–O(2) Fe(1)–O(9ii) Fe(1)–O(14) 1.978(5) 1.942(5) 2.012(5) Fe(1)–O(7i) Fe(1)–O(10) Fe(1)–O(17) 2.007(5) 1.997(5) 2.179(5) Ss(Fe(1)–O) = 3.02 Fe(2)–O(6i) Fe(2)–O(12) Fe(2)–O(17) 1.918(5) 2.006(5) 2.144(5) Fe(2)–O(11iii) Fe(2)–O(15iii) Fe(2)–O(17iii) 1.999(5) 1.898(5) 2.172(5) Ss(Fe(2)–O) = 3.05 P(1)–O(1) P(1)–O(3) 1.493(5) 1.569(5) P(1)–O(2) P(1)–O(4) 1.520(5) 1.573(5) Ss(P(1)–O) = 5.00 P(2)–O(5) P(2)–O(7) 1.511(5) 1.528(5) P(2)–O(6) P(2)–O(8) 1.514(5) 1.581(6) Ss(P(2)–O) = 5.00 P(3)–O(9) P(3)–O(11) 1.509(5) 1.550(5) P(3)–O(10) P(3)–O(12) 1.525(5) 1.552(5) Ss(P(3)–O) = 5.00 P(4)–O(13) P(4)–O(15) 1.512(5) 1.524(5) P(4)–O(14) P(4)–O(16) 1.524(5) 1.584(5) Ss(P(4)–O) = 4.97 N(1)–C(1) C(1)–C(2) N(2)–C(4) N(3a)–C(5vi) N(3b)–C(5) C(5)–C(6) 1.51(1) 1.52(1) 1.50(1) 1.56(2) 1.47(2) 1.51(3) N(1)–C(2iv) N(2)–C(3ii) C(3)–C(4v) N(3a)–C(6vii) N(3b)–C(7) C(5)–C(7viii) 1.502(9) 1.51(1) 1.53(1) 1.50(3) 1.46(3) 1.59(2) Symmetry codes: (i) x, y 1 1, z; (ii) x 1 1, y, z; (iii) 2x, 2y 1 2, 2z 1 1; (iv) 2x 1 2, 2y 1 1, 2z 1 1; (v) 2x 1 1, 2y 1 2, 2z; (vi) x, y, z 2 1; (vii) 2x, 2y 1 1, 2z 1 1; (viii) 2x, 2y 1 1 , 2z12.are connected in a 1-D chain by PO4 tetrahedra and have terminal H2PO4 and HPO4 groups (Fig. 5). The phosphate tetrahedra that complete the building unit are of three types. H2P(1)O4 has one oxygen, O(2), bridging to Fe(1) and extends away from the cluster as a pendant group. Of the three remaining P(1)–O bonds, one receives hydrogen bonds from neighboring piperazinium cations [O(1) ? ? ? N(1) 2.657 Å, O(1) ? ? ? N(2) 2.734 Å], whilst the other two with longer bonds [P(1)–O(3) 1.569 Å, P(1)–O(4) 1.573 Å] constitute P–OH groups.HP(2)O4 and HP(4)O4 groups share two corners of two octahedra within a tetramer with one corner as a terminal hydroxo group and the remaining corner serving as H-bond acceptor from piperazinium cations and H2PO4 groups [O(5) ? ? ? O(3) 2.502 Å, O(5) ? ? ? N(3a) 2.688 Å, O(13) ? ? ? O(4) 2.519 Å, O(13) ? ? ? N(3b) 2.780 Å]. The third type of phosphate, P(3)O4, shares three corners with three octahedra within a tetramer, and one corner with one octahedron in a neighboring tetramer. A topologically identical iron–oxygen cluster exists in the 3-D networks of the mineral leucophosphite, K2[Fe4(OH)2- (H2O)2(PO4)4]?2H2O, and the synthetic compound [H3N(CH2)3- NH3]2[Fe4(OH)3(HPO4)2(PO4)3]?xH2O.In leucophosphite the tetramer is coordinated by four PO4 tetrahedra, each sharing either two or three corners with the FeO6 octahedra of a Fig. 4 Tetranuclear iron cluster in [C4H12N2]1.5[Fe2(OH)(H2PO4)- (HPO4)2(PO4)]?0.5H2O.Filled lines and single lines are for Fe–O and P–O bonds, respectively. Thermal ellipsoids are shown at 60% probability. Fig. 5 Section of the infinite chain in [C4H12N2]1.5[Fe2(OH)(H2PO4)- (HPO4)2(PO4)]?0.5H2O showing the connectivity between tetranuclear clusters.4112 J. Chem. Soc., Dalton Trans., 1998, 4109–4112 tetramer. One vertex of each corner-sharing octahedron is occupied by a terminal water molecule. In [H3N(CH2)3NH3]2- [Fe4(OH)3(HPO4)2(PO4)3]?xH2O, the terminal aqua ligand in leucophosphite is replaced by a doubly bridging OH group which connects adjacent tetramers via –Fe–O(H)–Fe– bonds forming orthogonal, but non-intersecting infinite chains.Conclusion The hydrothermal synthesis described in this work produces another example of iron phosphate with an encapsulated piperazinium cation. The molecule piperazine is a versatile agent which directs the formation of 3-D framework, 2-D layered and 1-D chain structures.7 There was only one known chain iron phosphate structure containing octahedrally coordinated iron centers.The 1-D FePOs [(1R,2R)-C6H10(NH2)2]- [Fe(OH)(HPO4)2]?H2O, [trans-1,2-C6H10(NH2)2][Fe(OH)- (HPO4)2]?H2O and [C3H12N2][FeF(HPO4)2]?xH2O all contain [FeX(HPO4)2]22 n (X = OH, F) chains of the tanocoite type. The tind presents a new type of chain architecture which is formed from tetranuclear iron–oxygen clusters. Acknowledgements We thank the Institute of Chemistry, Academia Sinica, National Science Council (NSC-87-2113-M-001-019), and Chinese Petroleum Corp.for support, Ms. F.-L. Liao and Professor S.-L. Wang at the National Tsing Hua University for X-ray intensity data collection, and Professor T.-Y. Dong at the National Sun Yat-Sen University for Mössbauer spectroscopy measurements. References 1 R. C. Haushalter and L. A. Mundi, Chem. Mater., 1992, 4, 31. 2 M. I. Khan, L. M. Meyer, R. C. Haushalter, A. L. Schweitzer, J. Zubieta and J. L. Dye, Chem. Mater., 1996, 8, 43. 3 M. Cavellec, D. Riou, J. M. Greneche and G. Ferey, Inorg. Chem., 1997, 36, 2187. 4 J. R. D. DeBord, W. M. ReiV, C. J. Warren, R. C. Haushalter and J. Zubieta, Chem. Mater., 1997, 9, 1994. 5 K.-H. Lii and Y.-F. Huang, Chem. Commun., 1997, 839. 6 F. Hawthorne, Z. Kristallogr., 1990, 192, 1. 7 K.-H. Lii, Y.-F. Huang, V. Zima, C.-Y. Huang, H.-M. Lin, Y.-C. Jiang, F.-L. Liao and S.-L. Wang, Chem. Mater., 1998, 10, 2599. 8 P. B. Moore, Am. Mineral., 1972, 57, 397. 9 D. Riou, T. Loiseau and G. Ferey, Acta Crystallogr., Sect. C, 1993, 49, 1237. 10 G. M. Sheldrick, SADABS, Empirical Absorption Corrections Program, University of Göttingen, 1997. 11 G. M. Sheldrick, SHELXTL PC, Version 5, Siemens Analytical X-Ray Instruments Inc., Madison, WI, 1994. 12 Fe(PO3)3, file number 38–109; Fe4(P2O7)3, 36–318; Joint Committee on Powder DiVraction Standards, International Center of DiVraction Data, Swarthmore, PA. 13 I. D. Brown and D. Altermatt, Acta Crystallogr., Sect. B, 1985, 41, 244. Paper 8/04143A

ISSN:1477-9226    

DOI:10.1039/a804143a

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, 4113–4118 4113 Synthesis and complexation of Gd31, Ca21, Cu21 and Zn21 by 3,6,10-tri(carboxymethyl)-3,6,10-triazadodecanedioic acid Yun-Ming Wang,*a Chien-Hsun Lee,a Gin-Chung Liu b and Reu-Sheng Sheu b a School of Chemistry, Kaohsiung Medical College, No. 100 Shih-Chuan 1st Road, Kaohsiung, Taiwan 807, Republic of China b Department of Radiology, Kaohsiung Medical College, No. 100 Shih-Chuan 1st Road, Kaohsiung, Taiwan 807, Republic of China Received 27th July 1998, Accepted 12th October 1998 3,6,10-Tri(carboxymethyl)-3,6,10-triazadodecanedioic acid (H5L), was synthesized and its protonation constants were determined by potentiometric titration in 0.10 mol dm23 Me4NNO3 and by NMR pH titration at 25.0 ± 0.1 8C.Stability and selectivity constants have been measured to evaluate the possibility of using the corresponding gadolinium(III) complex as a magnetic resonance imaging contrast agent. The formations of gadolinium(III), copper(II), zinc(II) and calcium(II) complexes were investigated quantitatively by potentiometry.The stability constant for the gadolinium(III) complex is larger than those of CaII, ZnII and CuII for this octadentate ligand. The selectivity constants and modified selectivity constants of the ligand for Gd31 over endogenously available metal ions were calculated. The spin–lattice relaxivity R1 for the gadolinium(III) complex was also determined. It was found to decrease with increasing pH below 4 and became invariant with respect to pH over the range 4–10.Oxygen-17 NMR shifts showed that the [DyL]22 complex had one inner-sphere water molecule. The water proton spin–lattice relaxation rate for the [GdL]22 complex was also consistent with one inner-sphere co-ordination position. There is a great interest in the synthesis and characterization of new gadolinium(III) complexes of poly(aminocarboxylate) ligands as contrast agents in magnetic resonance imaging (MRI).1,2 Among paramagnetic contrast agents, stable water soluble gadolinium(III) chelates have the ideal properties of high water relaxivity, chemical stability, and low toxicity in vivo.At present, the octa-chelating ligands carboxymethyliminobis( ethylenenitrilo)tetraacetic acid (H5dtpa), N,N0-di(methylcarbamoylmethyl) carboxymethyliminobis(ethylenenitrilo)diacetic acid (H3dmdtta), 10-(2-hydroxypropyl)-1,4,7,10-tetraazacyclododecane- 1,4,7-triacetic acid (H3hpdotra) and 1,4,7, 10-tetraazacyclododecane-N,N9,N0,N--tetraacetic acid (H4- dota) are eVective MRI contrast agents when complexed with trivalent gadolinium ion.3–6 These gadolinium(III) chelates possess suYcient paramagnetism and high stability.The toxic eVect of uncomplexed Gd31 and free pro-ligand arising from dissociation of the metal complex is one of the major concerns in MRI.7–13 The acute toxicity of gadolinium(III) complexes of the poly(aminocarboxylates) correlates well with the selectivity of the latter for Gd31.The release of Gd31 is related to the stability constants of the gadolinium(III) chelates.14,15 The characterization of the complexes of gadolinium(III) with N,N9-bis(amide) derivatives of H5dtpa had been investigated in a series of our studies.16–18 In the continuing search for chelates for MRI, we have modified the ethylene group of H5dtpa to introduce a methylene group and explored both the stability and relaxivity of its gadolinium(III) complex.Therefore, this report describes the synthesis of one derivative of H5dtpa (Scheme 1), i.e. H5L = 3,6,10-tri(carboxymethyl)-3,6,10-triazadodecanedioic acid. Its protonation constants, thermodynamic and conditional stability constants of complexes with Gd31, Cu21, Zn21 and Ca21 and its selectivity for Gd31 over endogenously available metal ions are discussed. The 17O NMR shifts of the [DyL]22 complex are investigated. Finally, the spin–lattice relaxivity R1 of [GdL]22 is also described. Experimental Materials Gadolinium chloride (>99.9%) was obtained from Aldrich Chemical Co.and oven dried at 110 8C for at least 24 h before use. All other reagents used for the synthesis of the ligand were from commercial sources unless otherwise noted. Proton NMR spectra and elemental analyses were used to confirm the composition of the products. Preparation of 3,6,10-tri(carboxymethyl)-3,6,10-triazadodecanedioic acid (H5L) A suspension of 3.0 g (25.64 mmol) N-(2-aminoethyl)propane- 1,3-diamine, 25.56 g (136.19 mmol) tert-butyl bromoacetate, and 18.41 g (133 mmol) of anhydrous potassium carbonate in 125 cm3 of acetonitrile was stirred for 20 h.Removal of the solvent at reduced pressure on a rotary evaporator gave a residue which was partitioned between 100 cm3 of water and 100 cm3 of chloroform. The aqueous layer was separated and then extracted with two 100 cm3 portions of chloroform. The chloroform portions were combined and dried over MgSO4.Filtration and evaporation of solvent gave an amber oil, which was puri- fied by chromatography on silica gel using 75% ethyl acetate in methanol as the eluent to give a yellow oil. The oil was then Scheme 14114 J. Chem. Soc., Dalton Trans., 1998, 4113–4118 treated with 45 cm3 of concentrated aqueous hydrochloric acid (12 mol dm23) and stirred at room temperature for 6 h. The acid was removed by rotary evaporation and the residue taken up in water (20 cm3).The solution was loaded onto an AG 50W × 8 cation exchange resin column (200–400 mesh, H1 form, 3.5 × 20 cm) and washed with distilled water (1 dm3). The crude product was eluted with 0.5 mol dm23 NH3(aq). The solution was concentrated by rotary evaporation and the white residue applied to an AG1 × 8 anion exchange resin column (200–400 mesh, HCO2H form, 3.5 × 20 cm). The column was washed with distilled water and eluted with 0.5 mol dm23 formic acid solution to give the white hygroscopic free acid.Yield: 6.58 g (58%) (Found: C, 40.61; H, 6.84; N, 9.31. C15H25N3O10?2H2O requires C, 40.63; H, 6.59; N, 9.47%); dH 3.83 (s, 4 H, NCH2- COOH), 3.80 (s, 2 H, NCH2COOH), 3.51 (s, 4 H, NCH2- CH2N), 2.28 (t, 4 H, NCH2CH2CH2N) and 2.16 (m, 2 H, NCH2CH2CH2N). 13C NMR(D2O): d 176.52, 175.01, 173.45, 60.46, 60.23, 58.76, 56.37, 55.03, 54.41, 53.56 and 23.04. General techniques Solutions of H5L (0.1 mmol dm23) for NMR pH titration were made up in D2O, and the pD was adjusted with DCl or CO2- free NaOD.Proton NMR spectra were measured in D2O solution on a Varian Unity Plus 400 spectrometer. The final pD of the ligand solutions was determined with a microelectrode, pD = pH 1 0.40.19 The hydrogen electrode used in the present study allows a reliable and accurate determination of the proton activity over an extended pH range. The 17O NMR spectra were recorded by a Varian Gemini 300 spectrometer at 21 8C.The induced 17O shift (d. i. s.) measurements were determined with respect to water as external standard. The hydration number of [DyL]22 was determined by the method of Alpoim et al.20 An equimolar solution of Dy31 and ligand was prepared, and a stoichiometric amount of standardized NaOH was added so that the complex was fully formed. Six solutions of diVering dysprosium concentrations were prepared by serial dilution of the stock solution. Solution preparations Stock solutions of Ca21, Zn21, Cu21 and Gd31 were prepared between 0.015 and 0.02 mmol dm23 from the nitrate salts with demineralized water (obtained by a Millipore/Milli-Q system) and standardized by titration with Na2H2edta (disodium salt of ethylenedinitrilotetraacetic acid) or atomic absorption spectrophotometry.A stock solution was prepared by dissolving 4.65 g reagent grade Na2H2edta and diluting it to 250 cm3 with demineralized water. This was used as a titrant to standardize the solution of Gd31 and Ca21.A weakly acidic gadolinium chloride titrant solution was prepared at pH 5 by using a 0.5 mol dm23 acetate buVer and one drop of pyridine. Six drops of xylenol orange were added as an indicator, followed by titration with Na2H2edta solution until the solution changed from purple to yellow. This gadolinium(III) solution was used to standardize solutions of the linear poly(aminocarboxylates). Titrant solutions of the latter consisted of approximately 2.0– 0.6 mmol dm23 solute, to which acetate buVer pH 5 and one drop of pyridine were added.Six drops of indicator solution (xylenol orange) were added followed by titration with stock gadolinium(III) solution until a change from yellow to purple was observed.21 Stock gadolinium(III) complex solutions (henceforth identified as GdL and having a concentration range of 1.5–0.5 mmol dm23) were prepared by mixing equimolar amounts of stock solution of gadolinium(III) and ligand.A slight excess (2%) of ligand was used to ensure total complexation of gadolinium(III). Potentiometric measurements Potentiometric titrations were performed with an automatic titrator system to determine the protonation constants of the ligand and the stability constants of the metal complexes. The autotitrating system consists of a 702 SM Titroprocessor, a 728 stirrer, and a PT-100 combination pH electrode (Metrohm). The pH electrode was calibrated using two standard buVer solutions and all calibrations and titrations were carried out under a CO2-free nitrogen atmosphere to avoid any contact with carbon dioxide in a sealed glass vessel (20 cm3) thermostatted at 25.0 ± 0.1 8C, and an ionic strength of 0.10 mol dm23 Me4NNO3.A CO2-free 0.100 mol dm23 NaOH solution was used as the titrant to minimize ionic strength change during the titration. The purity of the ligand was also confirmed by potentiometric titration with standard NaOH. Oxygen and carbon dioxide were excluded from the reaction mixtures by maintaining a positive pressure of purified nitrogen in the titration cell.More than 200 data points were collected for each experiment. Each titration was performed at least three times. Since the Gd31 chelate is completely or almost completely formed at low pH, its stability constant could not be determined from the normal potentiometric titration method. Therefore, it was evaluated by a ligand–ligand competition potentiometric titration.22–24 A 1 : 1 : 1 molar ratio of Gd31, ligand, and a reference ligand with a known metal chelate stability was titrated.A good reference ligand for the Gd31 systems was found to be H4edta15 which forms a complex with Gd31 whose stability constant is accurately known. The electromotive force of the cell is given by E = E98 1 Q log[H1] 1 Ej and both E98 and Q were determined by titrating a solution of known hydrogen-ion concentration at the same ionic strength, using the acid range of the titration.The liquidjunction potential, Ej, was found to be negligible under the experimental conditions used. The potentiometric equilibrium studies were made on solutions of ligand, in the absence of metal ions, and then in the presence of each metal ion with the M :L ratio 1 : 1. The E data were obtained after additions of 0.005 cm3 increments of standard 0.100 mmol dm23 NaOH solution, and after stabilization in this direction equilibrium was then approached from the other direction by adding standard 0.100 mol dm23 acid solution.The equilibria were slow to attain and about 15 min were required for each point of the titration where the ligand– ligand competition took place. However, complexation was usually rapid (1–5 min per point to give a stable pH reading) with CuII, CaII and ZnII. The same values of the stability constants were obtained either by using the direct or the back titration. Computational method The protonation constants of the ligand were calculated using a FORTRAN computer program PKAS25 written for polyprotonic weak acid equilibria.The overall stability constants of the various metal complexes formed in aqueous solution were determined from the titration data with the FORTRAN computer program BEST.25 The average diVerence between observed and calculated 2log [H1] was <0.04 throughout all titrations. A value of 13.78 was employed for the pKw at 25 8C.The species distribution diagrams were calculated with the FORTRAN programs SPE and SPEPLOT.25 Relaxation time measurement A gadolinium(III) chelate solution was prepared by combining equimolar amounts of the stock GdCl3 and the ligand solution. A slight excess (3%) of the ligand was used and the solution was allowed to react for at least 2 h at room temperature to ensure completion of the complexation. Gadolinium(III) chelate solutions at various pH values were prepared by combining the buVer solution with an appropriately diluted complex solution in a 1 : 1 (v/v) ratio.The following buVer systems (all 0.10 mol dm23) were used: chloroacetic acid–NaOH (pH 2 and 3), aceticJ. Chem. Soc., Dalton Trans., 1998, 4113–4118 4115 acid–NaOH (pH 4 and 5), H2PIPES (piperazine-N,N9-bis- (ethane-2-sulfonic acid))–NaOH (pH 6.8), and ammonia–HCl (pH 9 and 10).21 These buVer solutions were used to maintain constant ionic strength (i.e. 0.10 mol dm23). The 0.10 mol dm23 buVers were suYcient to keep the solution pH within the desired range in most cases.The buVered gadolinium(III) chelate solutions were all allowed to equilibrate for at least 2 h. Their pH was determined immediately before relaxation time T1 measurements. Relaxation times T1 of aqueous solutions of the gadolinium( III) complex of H5L were measured to determine the relaxivity R1. All measurements were made using an NMR spectrometer operating at 20 MHz and 37.0 ± 0.1 8C (NMS 120 Minispec, Bruker). Before each measurement the spectrometer was tuned and calibrated.The value of T1 was measured from eight data points generated by an inversion–recovery pulse sequence. The slopes of plots of 1/T1 versus concentration give R1 in dm3 mmol21 s21. Results and discussion Protonation constants The ligand protonation constants are expressed as in eqn. (1). Kn = [HnL]/[Hn 2 1L][H1] (1) Table 1 summarizes the protonation constants of H5L, H3- dmdtta, H4edta and H5dtpa measured in the range pH 2–10.The titration curve of H5L shows one sharp increase between about pH 9.0 and 5.0 (mols of base per mol ligand present = 3). This is due to the large diVerence between the second (log K2) and third protonation constant (log K3), i.e. 8.92 and 5.12. The log K4 (fourth protonation constant) value is 2.80. The first and second protonation constants of H5L are very similar to those of H5dtpa (log K1 = 10.49, log K2 = 8.60 in 0.1 mol dm23 NaClO4).26 The third protonation constant decreases in the order H5L > H5dtpa.The replacement of the one ethylene group in H5dtpa by the one trimethylene group results in an increase in log K2 (i.e. 0.31 unit), log K3 (i.e. 0.83 unit), log K4 (i.e. 0.15 unit) and SpKa values (i.e. 1.43 unit). This can be explained by considering the chain length between the amino groups. In general, the protonation constant increases with the chain length between the amino groups.27 The protonation constants of the ligands given in Table 1 decrease in the order H5L > H5dtpa > H4edta > H3dmdtta.NMR pH titration The macroscopic protonation constants of the ligands in Table 1 determined by the potentiometric titration technique do not give a clue to the specific preference of the protonation sites. However, the microscopic protonation scheme that is obtained by NMR spectroscopy coupled with pH titration will. This is constructed by measuring the chemical shifts of the methylenic protons as a function of pH, and is based on previous observation that the protonation of a basic site of a poly(aminocarboxylate) in acidic solution leads to a deshielding of the adjacent methylene protons.28 The NMR chemical shifts at Table 1 Thermodynamic data for the successive protonation of H5L at 25.0 ± 0.1 8C in aqueous Me4NNO3 (I = 0.10 mol dm23) Species log b H 1234 L 1111 H5L 10.60 (0.02) 19.52 (0.02) 24.64 (0.02) 27.44 (0.03) H5dtpa a 10.49 19.09 23.37 26.01 H3dmdttab 9.37 13.75 17.06 H4edta a 10.17 16.28 18.96 20.01 a Data were obtained from ref. 26. b Ref. 15. diVerent pH values were assigned on the basis of signal multiplicities and the absence of signal crossovers over the whole pH range. These show that the central nitrogen atom is the most basic. Plots of the chemical shift values (d) of the methylenic resonance of H5L as a function of pH are given in Fig. 1. The observed deshielding of the methylene protons of the ligand is correlated with the percentages of protonation of the amino or carboxylate groups.29–31 The protonation fractions of H5L (%), fj, for the nitrogen atoms (f1, f2 and f3) and carboxylate groups (f4 and f5) labelled in Fig. 1 were calculated for integer values of n (1–3, number of mols of acid added per mol of polyaminopolycarboxylate). When 1 equivalent of acid is added to the fully unprotonated form of the ligand, the values for H5L (n = 1, f1 = 27.4, f2 = 50.9, f3 = 19.2) are similar to those for H5dtpa.29 These results indicated that the central nitrogen is more strongly basic than the terminal nitrogen atoms.For n > 1 the protonation fraction values obtained for H5L are as follows: at n = 2 ( f1 = 89.9, f2 = 41.0, f3 = 69.0) and at n = 3 ( f1 = 99.5, f2 = 93.1, f3 = 95.1, f4 = 23.3, f5 = 25.0). There is a preference for the terminal trimethylene and ethylene nitrogens relative to the central nitrogen for n = 2. The protonated forms with the terminal nitrogen atoms preferentially protonated are stabilized by internal hydrogen bonding between terminal carboxylates and nitrogen atoms, leading to high values of the second and third protonation constants. Thermodynamic stability constants The stability of the diVerent gadolinium(III) complexes can be expressed in four ways: (1) the thermodynamic stability constant of the complex, (2) the conditional stability constants at pH 7.4,15 (3) the selectivity constant, Ksel [the diVerence between the thermodynamic stability constant of the gadolinium complex and that of endogenously available metal ions (KZnL, KCaL and KCuL)],11 and (4) the modified selectivity constant, Ksel9 (the stability corrected for competition between the endogenously available metal ion and H1).15 The normal chelate thermodynamic stability constants (KML) are expressed as in eqn.(2) where M represents the KML = [ML]/[M][L] (2) Fig. 1 Proton NMR titration curves for H5L.4116 J.Chem. Soc., Dalton Trans., 1998, 4113–4118 free, unhydrolysed aquametal ion, L the uncomplexed, totally deprotonated form of the ligand and ML is the normal unprotonated and unhydrolysed complex. All potentiometric titration curves have an inflection point at 5 mol base added per mol ligand. The [CaL]32, [CuL]32 and [ZnL]32 curves increase rapidly from pH 4 to 8, 4 to 10 and 7 to 10, respectively. The stability constant of the complex of Gd31 with H5L was derived from the competition reaction of H4edta.In Table 2 the thermodynamic stability constants are presented for the linear poly(aminocarboxylates) H5L, H3dmdtta, and H5dtpa.15,26,32 The weaker stability of H3dmdtta chelates when compared to the H5L and H5dtpa chelates is assigned to the weaker donor ability of the amide group and the lower basicity of the terminal nitrogen atoms. The higher stability of H5L chelates when compared to the H3dmdtta chelates is assigned to the higher basicity of the nitrogen atoms.The thermodynamic stability constants of Ca21 complexes follow the order [CaL]32 (14.45) > [Ca(dtpa)]32 (10.75) > [Ca(dmdtta)]2 (7.17). Since the stability constants of calcium(II) complexes with H5dtpa (log KCaL = 10.75) and H4edta (log KCaL = 10.61) are similar, it appears that the co-ordination behavior of the H5dtpa type ligand is similar to that of H4edta. However, the co-ordination number of calcium(II) in the crystal structure of [Ca(edta)]22 is eight, six donor atoms from H4edta and two co-ordinated waters.33 Six-co-ordination of Ca21 with H3dmdtta, H5dtpa and H4edta has been also proposed.34 In other words, calcium(II) does not take advantage of all donor atoms in the case of H5dtpa and H3dmdtta.Therefore, the stability of [CaL]32 complexes is higher than those of [Ca(dtpa)]32 and [Ca(dmdtta)]2 due to their lower basicity. Conditional stability constants and selectivity constants For biological studies the conditional stability of a metal chelate under physiological conditions (pH 7.4) 11,15 is more important than the thermodynamic stability constant.In Table 2 the conditional stability constants at pH 7.4 are presented for the three poly(aminocarboxylates) H5L, H3dmdtta and H5dtpa. Their order is [GdL]22 ª [Gd(dtpa)]22 > [Gd(dmdtta)]. Fig. 2 shows the pH dependence of the conditional stability for the complexes [GdL]22, [Gd(dtpa)]22 and [Gd(dmdtta)]. The results for [GdL]22 and [Gd(dtpa)]22 are very similar. Stability constants do not provide directly comparable bases for measuring the total ion sequestering abilities of the ligands under physiological conditions (pH 7.4), and therefore they were used to calculate pM values (pM = 2log [Mf]), where [Mf] is the concentration of the free aqua metal ion that would be present at equilibrium pH 7.4.35 The advantage of comparing pM values rather than stability constants of the complexes is Table 2 Stability constants and selectivity constants of the complexes of Gd31, Zn21, Ca21 and Cu21 at 25.0 ± 0.1 8C in aqueous Me4NNO3 (I = 0.10 mol dm23) Parameter log ([GdL]/[Gd][L]) log KGdL9 (pH 7.4) log ([CaL]/[Ca][L]) log bCaHL log KCaL9 (pH 7.4) log ([CuL]/[Cu][L]) log bCuHL log KCuL9 (pH 7.4) log ([ZnL]/[Zn][L]) log bZnHL log KZnL9 (pH 7.4) Selectivity [log K(Gd/Zn)] [log K(Gd/Ca)] [log K(Gd/Cu)] log Ksel9 H5L 22.77 (0.03) 18.04 14.45 (0.05) 20.51 (0.05) 9.72 19.31 (0.01) 24.83 (0.01) 14.58 18.59 (0.03) 26.36 (0.04) 13.86 4.18 8.32 3.46 8.44 H3dmdttaa 16.85 (0.05) 14.84 7.17 (0.04) 11.62 5.11 13.03 (0.03) 16.39 11.06 12.04 (0.03) 16.08 10.02 4.81 9.73 3.78 9.03 H5dtpa b 22.46 18.14 10.75 — 6.43 21.38 — 17.06 18.70 — 14.38 3.76 11.71 1.08 7.04 a Data were obtained from ref. 15. b Refs. 26 and 32. that the pM values reflect the influence of ligand basicity and metal chelate protonation. The larger the pM value the higher is the aYnity of that ligand for the metal ion under the specified conditions.The relative order of the pM values may change if a diVerent set of conditions (concentration and pH) is used to calculate the pM values. The results given in Table 3 indicate that H5L is a much stronger gadolinium(III) chelating agent than H3dmdtta. The pM values of [GdL]22 and [Gd(dtpa)]22 are larger than those of [CaL]32 and [Ca(dtpa)]32, [CuL]32 and [Cu(dtpa)]32, and [ZnL]32 and [Zn(dtpa)]32, respectively. Therefore, the competition among Gd31, Ca21, Cu21 and Zn21 with H5L is seen to favor Gd31 at pH 7.4, indicating that the gadolinium( III) complex should be stable enough to avoid interference by Ca21, Cu21 and Zn21.The pGd value for [GdL]22 is slightly higher than that of [Gd(dtpa)]22. Even though the stability constants for H5L chelates are significantly larger than those of the corresponding H3dmdtta chelates, the pGd value for [GdL]22 is about 4.0 log units larger than that of [Gd(dmdtta)], because H5L has high protonation constants, and therefore the formation of its complex is subject to stronger hydrogen ion competition. Species distribution curves of [GdL]22 shown in Fig. 3, generated from the potentiometric data given in Table 2, indicate that there is still some free Gd31 at pH 1 but by pH 3 the complex is fully formed. However, [GdL]22 is the dominant species at physiological pH 7.4. The logarithmic selectivity constant 11,15 of H5L, H5dtpa and H3dmdtta for Gd31 over Zn21, Ca21 and Cu21 is the diVerence between log KGdL and log KML (M = Zn21, Ca21 or Cu21), i.e.log K(Gd/Zn), log K(Gd/Ca) and log K(Gd/Cu). The selectivity constants are also given in Table 2. From these, H5L shows more favourable selectivity toward Gd31 over Zn21 and Cu21 than does H5dtpa. The consequences of the selectivity for Gd31 over other endogenous metal ions (Cu21, Ca21 and Zn21) and H1 for a ligand can be calculated by using eqn. (3).15 This equation is Ksel9 = KML(aH 21 1 aCaL 21 1 aCuL 21 1 aZnL 21)21 (3) Fig. 2 Variation of the conditional stability constants for [Gd(dtpa)]22 (1), [GdL]22 (2) and [Gd(dmdtta)] (3) with pH. Table 3 The pMa values of the complexes of Gd31, Zn21, Ca21 and Cu21 at pH 7.4 pGd pCu pCa pZn H5L 17.03 13.58 9.73 13.22 H5dtpa 17.14 16.06 5.45 13.39 H3dmdtta 13.88 10.05 4.19 9.06 a pM = 2log [M]free at pH 7.4, [metal ion]total = 1 mmol dm23, and [ligand]total = 1.1 mmol dm23.J. Chem. Soc., Dalton Trans., 1998, 4113–4118 4117 obtained by the incorporation of ligand equilibria with Cu21, Ca21, Zn21 and H1 where a is a side reaction coeYcient defined as in eqns. (4)–(7).aH 21 = 1 1 K1[H1] 1 K1K2[H1]2 1 K1K2K3[H1]3 1 . . . . (4) aCaL 21 = 1 1 KCaL[Ca21] (5) aCuL 21 = 1 1 KCuL[Cu21] (6) aZnL 21 = 1 1 KZnL[Zn21] (7) Table 2 shows the modified selectivity constants of H5L, H5dtpa and H3dmdtta at pH 7.4. The concentrations of Ca21, Cu21 and Zn21 used were 2.5, 1.0 × 1023 and 5.0 × 1022 mmol dm23, respectively.15 The log Ksel9 of H5L (8.44) is higher than that of H5dtpa (7.04), but slightly lower than that of H3dmdtta (9.03). The ligands H5L and H3dmdtta appear to have comparable selectivity and should have comparable toxicity due to metal ion displacement in vivo.Thus, H5L forms a gadolinium( III) complex that is slightly more stable than [Gd- (dtpa)]22 toward transmetallation with endogenous metal ions at pH 7.4. DyIII-induced 17O water NMR shifts The DyIII-induced water 17O NMR shifts versus chelate concentration for a solution of DyCl3 and [DyL]22 in D2O at 21 8C are shown in Fig. 4.A hydration number of dysprosium(III) ion of eight has been proposed.36–38 For the [DyL]22 complex the slope is 248.8 ppm. Under the conditions applied in the present study the slope per DyIII-bound water molecule is 255.70 ppm. It can be concluded that the [DyL]22 complex contains 0.90 inner-sphere water molecule per DyIII. The result is in good agreement with that for H3dmdtta using the same technique.20 The number of LnIII-bound water molecules in this complex provides information on the co-ordination mode of the ligand.One co-ordination site of each LnIII is occupied by one water molecule and eight sites are available for the ligand molecule. By binding of three amine nitrogen atoms and five carboxylates, a similar co-ordination mode as found for the previously studied H5dtpa is attained.39 Relaxometric studies of the gadolinium(III) complex The inner sphere relaxation mechanism could be influenced by the rate of chemical exchange of water from the co-ordination water to the bulk water.The paramagnetic contribution of the solvent longitudinal relaxivity is obtained from eqn. (8),40 where Fig. 3 Species distribution curves for a 7.0 × 1023 mol dm23 [GdL]22 system containing a 1 : 1 molar ratio of GdIII to ligand. T = 25.0 ± 0.1 8C; I = 0.10 mol dm23 (Me4NNO3); % = percentage relative to 7.00 × 1023 mol dm23 total ligand species = 100%.R1 = Nq/[55.6(T1M 1 tM)] (8) N is the molar concentration of the gadolinium(III) complex, q the number of water molecules bound per metal ion, T1M the relaxation time of the bound water protons, and tM the residence lifetime of the bound water. Owing to the opposite temperature dependences of T1M and tM, two cases can be considered: (1) fast chemical exchange (T1M @ tM), R1 increases with decreasing temperature, and (2) slow chemical exchange (T1M ! tM), R1 decreases by decreasing the temperature. Fig. 5 shows the temperature dependence of the relaxivity for the complex [GdL]22 at 20 MHz. A monoexponential decrease of the observed relaxivity upon increasing the temperature in the range 5–60 8C was found. This is characteristic of the fast chemical exchange behavior which occurs when the residence lifetime of the co-ordinated water molecule (tM) is much shorter than the relaxation time of the bound water proton (T1M).The spin–lattice relaxivity R1 of [GdL]22 is given in Table 4. It is similar to those of H5dtpa and its bis(amide) derivatives under the same experimental conditions.6,16–18 The relaxivity of a paramagnetic metal complex consists of two components: the inner-sphere and outer-sphere relaxivities. Since the basic skeleton of the ligands studied and the shapes and sizes of the gadolinium(III) complexes are similar, it is assumed that the Fig. 4 The DyIII-induced water 17O NMR shift versus chelate concentration for solutions of (1) [DyL]22 and (2) DyCl3 in D2O at 25 8C.Fig. 5 Temperature dependence of the longitudinal relaxation rate for a 1 mmol dm23 solution of [GdL]22, measured at 20 MHz, pH 6.8.4118 J. Chem. Soc., Dalton Trans., 1998, 4113–4118 outer-sphere relaxivities are similar. Thus, the observed relaxivity is primarily attributed to the variation in the inner-sphere contribution. The inner-sphere relaxivity is mainly dependent on the hydration number of the gadolinium(III) complex.A larger hydration number leads to a higher relaxivity R1. However, the q value of the complex of gadolinium(III) with H5L is the same as those of [Gd(dtpa)]22 and [Gd(dmdtta)] 20 leading to almost identical R1 values. In other words, the similarity in the relaxivity R1 of [GdL]22, [Gd(dmdtta)] and [Gd(dtpa)]22 confirms that the number of inner sphere water molecules is identical. The relaxivities R1 for the complexes [GdL]22 and [Gd- (dmdtta)] at various pH values are given in Fig. 6. The relaxivity curve exhibits no pH dependence over the range 4–10. Therefore, no ligand dissociation occurred with this pH range and the hydration number remains constant. High relaxivity (R1) and high stability of the paramagnetic metal chelate are the most important prerequisites for a magnetopharmaceutical drug. The fact that the gadolinium( III) complex of H5L is quite stable in aqueous solution, does not dissociate under physiological conditions (pH 7.4), and does not exchange with CaII, CuII and ZnII to an appreciable extent shows that the ionic chelate [GdL]22 may be considered a promising MRI contrast agent.Acknowledgements We are grateful to the National Science Council and Department of Health of the Republic of China for financial support. Fig. 6 pH Dependence of the relaxivity for the complexes [GdL]22 (1) and [Gd(dmdtta)] (2), all in 0.10 mol dm23 buVers at 20 MHz and T = 37.0 ± 0.1 8C.Table 4 Relaxivities R1 of gadolinium(III) complexes at 37.0 ± 0.1 8C and 20 MHz Complex [GdL]22 [Gd(dtpa)]22 [Gd(dmdtta)] pH 7.5 ± 0.1 limiting a 7.6 ± 0.1 limiting a 7.5 ± 0.1 limiting a Relaxivity R1 a/ dm3 mmol21 s21 3.85 ± 0.03 4.50 ± 0.05 3.89 ± 0.03 4.65 ± 0.05 3.85 ± 0.03 4.60 ± 0.05 a Average of relaxivity values over the pH range 3–10. References 1 R. B. LauVer, Chem. Rev., 1987, 87, 901. 2 S. C. Quay, U.S. Pat., 4 687 659, 1987. 3 D. D. Dischino, E.J. Delaney, J. E. Emswiler, G. T. Gaughan, J. S. Prasad, S. K. Srivastava and M. F. Tweedle, Inorg. Chem., 1991, 30, 1265. 4 H. J. Weinmann, R. C. Brasch, W. R. Press and G. Wesbey, Am. J. Roentgenol., 1984, 142, 619. 5 J. C. Bousquet, S. Saini, D. D. Stark, P. F. Hahn, M. Nigam, J. Wittenberg and J. T. Ferrucci, Radiology, 1988, 166, 693. 6 C. A. Chang, Invest. Radiol., 1993, 28, S21. 7 K. Kumar and M. F. Tweedle, Pure Appl. Chem., 1992, 65, 515. 8 K. Kumar, C. A. Chang and M.F. Tweedle, Inorg. Chem., 1993, 32, 587. 9 K. Kumar and M. F. Tweedle, Inorg. Chem., 1993, 32, 4183. 10 K. Kumar, C. A. Chang, L. C. Francesconi, D. D. Dischino, M. F. Malley, J. Z. Gougoutas and M. F. Tweedle, Inorg. Chem., 1994, 33, 3567. 11 K. Kumar, M. F. Tweedle, M. F. Malley and J. Z. Gougoutas, Inorg. Chem., 1995, 34, 6472. 12 C. Paul-Roth and K. N. Raymond, Inorg. Chem., 1995, 34, 1408. 13 H.-J. Weinmann, W.-R. Press and H. Gries, Invest. Radiol., 1990, 25, S49. 14 M. F. Tweedle, J. J. Hagan, K. Kumar, S. Mantha and C. A. Chang, Magn. Reson. Imaging, 1991, 9, 409. 15 W. P. Cacheris, S. C. Quay and S. M. Rocklage, Magn. Reson. Imaging, 1990, 8, 467. 16 Y. M. Wang, T. H. Cheng, G. C. Liu and R. S. Sheu, J. Chem. Soc., Dalton Trans., 1997, 883. 17 Y. M. Wang, T. H. Cheng, R. S. Sheu, I. T. Chen and M. Y. Chiang, J. Chin. Chem. Soc., 1997, 44, 123. 18 Y. M. Wang, S. T. Lin, Y. J. Wang and R. S. Sheu, Polyhedron, 1998, 17, 2021. 19 K. Mikkelsen and S. O. Nielsen, J. Phys. Chem., 1960, 64, 632. 20 M. C. Alpoim, A. M. Urbano, C. F. G. C. Geraldes and J. A. Peters, J. Chem. Soc., Dalton Trans., 1992, 463. 21 C. A. Chang, H. G. Brittain, J. Telser and M. F. Tweedle, Inorg. Chem., 1990, 29, 4468. 22 W. R. Harris and A. E. Martell, Inorg. Chem., 1976, 15, 713. 23 Y. Li, A. E. Martell, R. D. Hancock, J. H. Reibenspies, C. J. Anderson and M. J. Welch, Inorg. Chem., 1996, 35, 404. 24 C. H. Taliaferro, R. J. Motekaitis and A. E. Martell, Inorg. Chem., 1984, 23, 1188. 25 A. E. Martell and R. J. Motekaitis, Determination and Use of Stability Constants, 2nd edn., VCH, New York, 1992. 26 R. M. Smith and A. E. Martell, Critical Stability Constants, Plenum, New York, 1975–1977, vols. 1–4. 27 J. Clark and D. D. Q. Perrin, Chem. Rev., 1964, 18, 295. 28 A. D. Sherry, W. P. Cacheris and K.-T. Kuan, Magn. Reson. Med., 1988, 8, 180. 29 J. L. Sudmeier and C. N. Reilley, Anal. Chem., 1964, 36, 1698. 30 C. F. G. C. Geraldes, A. M. Urbano, M. C. Alpoim, A. D. Sherry, K. T. Kuan, R. Rajagopalan, F. Maton and R. N. Muller, Magn. Reson. Imaging, 1995, 13, 401. 31 B. Achour, J. Costa, R. Delgado, E. Garrigues, C. F. G. C. Geraldes, N. Korber, F. Nepveu and M. I. Prata, Inorg. Chem., 1998, 37, 2729. 32 D. L. Wright, J. H. Holloway and C. N. Reilley, Anal. Chem., 1965, 37, 884. 33 B. L. Barnett and V. A. Uctman, Inorg. Chem., 1979, 18, 2674. 34 E. N. Rizkalla, G. R. Choppin and W. P. Cacheris, Inorg. Chem., 1993, 32, 582. 35 W. R. Harris, K. N. Raymond and F. L. Weitl, J. Am. Chem. Soc., 1981, 103, 2667. 36 T. Kowall, F. Foglia, L. Helm and A. E. Merbach, J. Am. Chem. Soc., 1995, 117, 3790. 37 C. Cossy, L. Helm, D. H. Powell and A. E. Merbach, New J. Chem., 1995, 19, 27. 38 C. Cossy, A. C. Barnes, J. E. Enderby and A. E. Merbach, J. Chem. Phys., 1989, 90, 3254. 39 J. A. Peters, Inorg. Chem., 1988, 27, 4686. 40 T. J. Swift and R. E. Connick, J. Chem. Phys., 1962, 37, 307. Paper 8/05837G

ISSN:1477-9226    

DOI:10.1039/a805837g

出版商:RSC    

年代:1998

数据来源: RSC