首页   按字顺浏览 期刊浏览 卷期浏览 Lanthanide complexes with O2PH2?, O2P(CH2Cl)2?and O2P(CH2OH)2?ligands: a solid-state tu...
Lanthanide complexes with O2PH2?, O2P(CH2Cl)2?and O2P(CH2OH)2?ligands: a solid-state tubular micelle

 

作者: Roman A. Kresinski,  

 

期刊: CrystEngComm  (RSC Available online 2000)
卷期: Volume 2, issue 33  

页码: 177-182

 

ISSN:1466-8033

 

年代: 2000

 

DOI:10.1039/b008824m

 

出版商: RSC

 

数据来源: RSC

 

摘要:

Lanthanide complexes with O2PH22, O2P(CH2Cl)22 and O2P(CH2OH)22 ligands: a solid-state tubular micelle Roman A. Kresinski,a Andrew W. G. Plattb and Joanne A. Seddonb aSchool of Applied Chemistry, Kingston University, Penrhyn Road, Kingston upon Thames, KT1 2EE, UK. E-mail: r.kresinski@king.ac.uk bSchool of Sciences, Staffordshire University, College Road, Stoke-on-Trent, ST4 2DE, UK Received 1st November 2000, Accepted 20th November 2000 Published on the Web 11th December 2000 Attempts were made to synthesise lanthanide complexes of phosphorus-based ligands so as to incorporate channels which might act as hosts for small molecules; the sharing of ligands in La(O2P(CH2Cl)2)3- (OP(OH)(CH2Cl)2) is between adjacent metals, resulting in a linear stacking arrangement rather than a lattice.Pr(O2P(CH2OH)2)3 crystallises into a highly H-bonded lattice, in which the methylene groups are spontaneously directed so as to line channels oriented along one of the crystal axes in a fashion reminiscent of solution micelles. These channels are, however, apparently too small to accommodate aliphatic hydrocarbons. The structure of Ho(O2PH2)3, which is the same as that of its prototypal Dy analogue, is also pictured and described. Introduction The whole of applied chemistry is concerned with systems which kinetically promote a particular product over another. Most utilisable organic materials, many differing as subtly as being isomeric one to another, lie on a `scale of oxidation' between raw petroleum at one extreme and carbon dioxide at the other.The production of the very many useful intermediate organic compounds depends on kinetic control of the chemical manufacturing processes. Certain means of exercising kinetic control have become familiar. The secondary : primary ratio obtainable from hydroformylation processes of primary alkenes1 is an example of a process determined in part by the electronic preferences of the transition state,2 but also by the bulk exerted locally at the reaction centre.3 More bulky co-ligands at the catalytic site favour the less sterically demanding product,3 albeit at a possible expense in overall rate.1 The logical extension of this principle is to increase steric inØuence by enclosing the reactants at the reaction site, for example by conducting a reaction between small molecules within a channel contained inside another material.The best-known examples of this type of channelled material are the zeolite minerals and their synthetic analogues, whose cavities may accommodate a wide variety of substrates and direct their reactions highly speciÆcally. The synthetic zeolite ZSM-5 is used4,5 industrially to isomerise mixtures of xylenes almost exclusively to the p-isomer, which is more easily accommodated within the zeolite channel.6 Thus, the size of the channel can be seen to be pivotal to the process in question and, indeed, the size of zeolite channels has been closely correlated to the product distribution in zeolite-catalysed hydrocarbon cracking reactions.7 The production of materials in which the size or shape of the channel can be controlled is clearly of academic interest as well as potentially highly industrially relevant.We recently reported8 the syntheses of a class of heteroleptic polymeric phosphite±hypophosphite complexes [Ln(O2PH2)(O3PH)- (H2O)]H2O, which possess channels occupied by water molecules in inÆnite hydrogen-bonded chains. Whilst these channels are considerably narrower than those in, for example, zeolites, and thus might lend themselves only to highly speciÆc and slow reactions between small inorganics or the termini of aliphatic organics, they are of interest because they are both DOI: 10.1039/b008824m This journal is # The Royal Society of Chemistry Paper chiral and reversibly evacuable.8 Their synthesis is, however, troublesome in that it relies upon in situ aqueous oxidation of hypophosphite, (O2PH22), with continual monitoring of the products by infrared spectroscopy.In a search for similar materials with simpler syntheses, we have returned to homoleptic lanthanide hypophosphite complexes8 and explored means of increasing the cavity size of daughter complexes by derivatisation of the hypophosphite ligands. We describe our Ærst observations herein. Results and discussion Ln(O2PH2)3 One of the materials described in our recent reports9 was the new Pr(O2PH2)3, possessing a polymeric structure with Pr atoms bridged by hypophosphite ligands in a 3-dimensional lattice.On moving across this series of Ln(O2PH2)3 complexes, the lanthanide contraction appears to prompt a sudden transition from 8-coordination for larger Ln, directly to a new series which is comparatively open in structure, being only 6-coordinate but retaining the Ln(O2PH2)3 formulation. No intermediate 7-coordinate species are observed. The report10 of Er(O2PH2)3 prior to ours of Dy(O2PH2)3 had described a structure inconsistent with our own studies of the latter, but we can now support our own observations by reporting the structure of Ho(O2PH2)3, determined by single-crystal X-ray diffraction, as being isomorphous with that of Dy(O2PH2)3. The report of the latter8 contains a more thorough discussion of the structural type and its properties.The Ho complex is seen to best advantage in Fig. 1,11 the view in Fig. 2 emphasising the linked columnar structure and `channels' aligned along the c-axis. Naturally, the spaces in the above Ln(O2PH2)3 (Ln~Dy, Ho) structures are too small to accommodate even the water molecule, as can [Pr(O2PH2)(O3PH)(H2O)]?H2O. Indeed, the Ln(O2PH2)3 (Ln~Dy±Lu) complexes are crystallised from aqueous solution, and no real afÆnity for water is exhibited by the isolated materials, which are consistently anhydrous.8 Ln(O2P(CH2Cl)2)3?nH2O In all Ln(O2PH2)3 complexes heretofore reported, the P±H bonds are never observed to undergo hydrogen bonding or1 CrystEngComm, 2000, 33, 1±6Fig. 1 A view of Ho(O2PH2)3, showing the coordination around the Ho atoms and depicting each unique ligand's bridging mode. The Ho, P and O atoms are depicted as spheres of decreasing arbitrary radius. H atoms are omitted for clarity.Symmetry operators used throughout: i~x, 12y, z; ii~2x, 2y, 2z; iii~2x, 2y, 12z; iv~2x, y,2z;v~2x,12y,2z;vi~2x,y,12z;vii~x,2y,z;viii~2x,12y,12z; 12z; ix~x, y21, z; x~2x, y21, 12z; xi~2x, y21, 2z; xii~x, yz1, z; xiii~12x, 2y, 12z; xiv~x, â2y, z2â; xv~x, â2y, zzâ. Click image or here to access a 3D representation. ligation, despite facile proximity to available groups. The P± H bond of the hypophosphites might thus be regarded as a coordinationally inactive substituent which, if suitably derivatised, might expand the structure generally so as to allow occupation of any cavities so formed by small molecules.We sought to achieve this by the derivatisation of dissolved or suspended Ln(O via addition of formaldehyde across the P±H bonds,12 the resulting Ln(O2P(CH2OH)2)3 being then chlorinated by SOCl2 to afford Ln(O2P(CH2Cl)2)3. The initial stage was achieved for Ln~Pr by reØux of Pr(O2PH2)3 in aqueous formaldehyde solution, but the resulting crystalline Pr(O2P- (CH2OH)2)3 did not react with SOCl2, presumably as a result of strong H-bonding in the solid. The most reliable route to Ln(O2P(CH2Cl)2)3 was found to be the direct reaction between hydrated Ln(NO3)3 or LnCl3 and the preformed free acid OP(OH)(CH2Cl)2. This route afforded complexes of formulation Ln(O Ln~Gd±Yb, n~1). Their infrared spectra show clear differences between the mono- and dihydrate classes: the OH stretches are of lower intensity and the PO region is more complex for the monohydrates Gd±Yb and a characteristic pair of peaks at 459 and 435 cm21, observed for the dihydrates, reduces to a single peak at 435 cm21 for the monohydrates. Raman spectra show only subtle differences between the two series with PO bands present at medium intensity with the most prominent emission at 662 cm21 assigned to the C±Cl stretch.2PH2)3 to Ln(O2P(CH2Cl)2)3, 2P(CH2Cl)2)3?nH2O (Ln~La±Eu, n~2; A crystal suitable for single-crystal X-ray diffraction study was removed from the bulk La complex. This afforded a poor dataset with a large merging discrepancy (Table 1), and the resulting structure is also disordered with a high discrepancy index, but is nevertheless of interest on account of being 7- coordinate.The actual formulation of the crystal is not that of the bulk material, but is clearly La(O2P(CH2Cl)2)3- (OP(OH)(CH2Cl)2), i.e. there are four anionic ligands to every La3z, of which one ligand is protonated. A packing diagram of the major conformeric form of the structural unit is shown in Fig. 3, pictured along the b-axis. Three of the ligands bridge between only two adjacent La ions so as to produce linear chains, rather than cross-linking La atoms to provide a lattice-like or mesh-like arrangement such as might afford channels. The fourth ligand, containing P1, is 2 CrystEngComm, 2000, 33, 1±6 Fig. 2 The structure of Ho(O2PH2)3 viewed along the c-axis, showing the unit cell (Z~4).H atoms are omitted for clarity. Fig. 3 The packing of La(O2P(CH2Cl)2)3(OP(OH)(CH2Cl)2) units, viewed along the b-axis, showing the unit cell (Z~4). H atoms are omitted for clarity. monodentate. Fig. 4 shows the mode of coordination of the ligands and how the protonated phosphoryl undergoes Hbonding with a neighbouring O atom. This phosphoryl thus seems to play no part in ligation to theLa, leaving the La with its notable coordination number. The structure is disordered, all chloromethyl groups, with the exception of that incorporating C31, being rotationally disordered between two sites. Difference Fourier syntheses suggest alternative placements for most atoms, such as would imply that the disorder might extend across all of each ligand.However, test reÆnements based upon models possessing wider disorder suffer from extreme instability, even when damped, and afford only modest improvements in discrepancy indices. It is likely that the weakness of the dataset effectively gives only a poor data : parameter ratio under these circumstances. Since the X-ray diffraction studies revealed a structure atypical of the bulk material, we returned to the Ln(O2P(CH2Cl)2)3?nH2O series and carried out conductivity measurements and electrospray mass spectral (EMS) studies to establish whether the information these yielded about the solution structures reØected in any way upon the solid-state structure. The complexes dissolve in water but show a pronounced decrease in solubility with increasing atomic number of the lanthanide, such that prolonged stirring is required to produce 1023 molar solutions of the Dy and Yb complexes.The aqueous solutions are conducting, with molar conductance values of 0.023±0.027 V21 m22 mol21, which correspond13 to a 2 : 1 electrolyte.14 These observations implyTable 1 Crystal data and experimental details for all structures. Click here for full crystallographic data (CCDC nos. 152864-152866). La(O2P(CH2Cl)2)3 Pr(O2P (OP(OH)(CH2Cl)2) Ho(O2PH2)3 (CH2OH)2)3 2064.10 Monoclinic 3150.42 Monoclinic P21/c (14) P21/c (14) M 1439.55 System Monoclinic Space group (no.) C2/m (12) Za/A 14.3544(22) b/A 5.7161(9) c/A 12.1035(13) b/� 122.4(5) V/A 3 838.51 T/K 298 m/mm21 9.98 r/g cm23 2.851 Data measured 3861 Unique data 737 int(%) R 6.44 s(%) R 4.04 R [Iw2s(I)](%) 2.62 410.9354(8) 13.9472(11) 11.3916(4) 110.587(8) 1626.47 120 3.34 2.107 6263 2420 5.03 5.28 3.36 8.47 4 418.621(4) 5.249(1) 25.890(5) 90.04(3) 2530.53 150 2.82 2.067 22054 5629 10.35 30.75 13.36 40.33 wR2 (all data)(%) 8.58 that the solid state structure is not retained in solution and is consistent with dissolution predominantly according to eqns.(1) and (2). (1) Ln(O2P(CH2Cl)2)3?â ä Ln(O2P(CH2Cl)2)2 zz â ä{ Ln(O O2(P(CH2Cl)2) 2zz { (2) 2P(CH2Cl)2)3?â ä LnO2P(CH2Cl)2 2P(CH2Cl)2) Ln(O 2 Oâ ä We hoped better to identify individual ions in aqueous solution by electrospray mass spectroscopy (EMS) of representative complexes in solution.Complexes were chosen to provide characteristic isotope distribution patterns to aid assignments, particularly for the identiÆcation of multiply charged ions, where a compression of the isotope proÆle is observed. Spectra run at low cone voltages normally give direct information on the ions present in solution, whilst at higher Fig. 4 A view of the major conformer of La(O2P(CH2Cl)2)3 (OP(OH)(CH2Cl)2), showing the coordination around the La atoms and depicting each unique ligand's bridging and H-bonding modes. Most H atoms are omitted for clarity. Click image or here to access a 3D representation showing the disorder of the chloromethyl groups.voltages fragmentation of the ions gives some indication of the gas phase stability.15,16 The results for the anion spectra are shown in Table 2. The ligand anions O2P(CH2Cl)22 give rise to the base peak in all spectra and signals due to lanthanide containing ions are present only at low intensities, of v5% of the base peak. The cation spectra are shown in Table 3. Here [LnO2P(CH2Cl)2(MeOH)n]2z are the predominant phosphinate ligand containing ions in solution: interestingly the singly charged ions [Ln(O2P(CH2Cl)2)2(MeOH)n]z are generally represented by low intensity signals or are completely absent and the base peak is due to [Ln(OMe)(MeOH)n]2z in the positive ion spectra of Ln(O2P(CH2Cl)2)3. We have previously observed that under electrospray conditions poorly coordinating ions are readily replaced in reactions such as shown in eqn.(3).17 (3) Ln(O2P(CH2Cl)2)3zMeOH?â ä Ln(OMe) 2z zOP(OH)(CH2Cl)2z2O2P(CH2Cl){2 The combined mass spectral and conductimetric data afford no evidence for the adoption of lattice-like bridges by Ln(O2P(CH2Cl)2)3?nH2O. No fragments such as {Ln3L}, which would be typical of a highly-bridged structure, are observable, but only those as might originate from linear chain assemblies. 2P(CH2OH)2)3 In view of the above studies being inconclusive regarding the solid-state structures of Ln(O2P(CH2Cl)2)3?nH2O, we took the opportunity of examining also the crystalline Ln(O2P- (CH2OH)2)3 obtained during the initial attempt at the synthesis of O2P(CH2Cl)22 complexes. Reaction of the preformed ligand free acid O2P(CH2OH)22 with Ln2O3 afforded a more reliable route to the Ln(O2P(CH2OH)2)3 complexes than the treatment of Ln(O2PH2)3 with formaldehyde.The complexes Ln(O2P- (CH2OH)2)3 have an anhydrous composition across the series, their infrared spectra being very similar with strong OH stretches between 3400±3150 cm21 and a series of intense PO stretches between 1200±1000 cm21 numbering six peaks for the lighter lanthanides' complexes, becoming less well resolved for the complexes of the heavier metals giving the appearance of three broad peaks at 1138, 1077 and 1025 cm21 for Yb(O2P(CH2OH)2)3. The reasons for the crystallinity of Ln(O2P(CH2OH)2)3 are clear from the structure of Pr(O2P- (CH2OH)2)3, which is highly bridged and exhibits an extraordinary degree of H-bonding; the disposition of ligands is Table 2 Negative ion mass spectral dataa for LnL3 (L~O2P(CH2X)2) Yb Er Eu Nd La 4]2 673.9 821.5 665.9 815.5 652.8 799.0 643.9 791.5 639(5) 786.6 [LnL X~OH X~Cl 5]22 399.2 395.5 388.9 383.2 382.1 [LnL X~OH 694.0 687.6 673.0 663.6 658.6 [LnL3zCl]2 X~Cl 547.9 517.9 487.8 457.8 549.9 509.9 479.9 449.9 527.0 496.8 465.8 435.8 513.0 483.1 453.0 423.0 For X~OH [M±H]2 [M±H±(CH2O)]2 [M±H±2(CH2O)]2 [M±H±3(CH2O)]2 125.0(100) 161.0(100) [L]2 X~OH X~Cl aData are quoted for the most intense peak expected in the isotope distribution proÆle for a given ion.Numbers in parentheses are the relative intensities of the signal, with valuesv5% not recorded. 3 CrystEngComm, 2000, 33, 1±6Fig. 5 A view of Pr(O2P(CH2OH)2)3, showing the coordination around Pr and depicting each unique ligand's bridging and H-bonding modes. Most H atoms are omitted for clarity. Click image or here to access a 3D representation of an incomplete fragment showing only 7 of the coordinating fragments. shown in Fig. 5 and entails the two ligands incorporating P1 and P2 forming bridges between metals aligned in a mot repeating along the a-axis. The third ligand bridges between metals having the same x-coordinate to afford a mesh-like structure. The full coordination complement of eight is then made up by the bridging of O13 and O23 back to the Ln atom to form 5-membered rings.Even with two of the six phosphoryl O atoms thus occupied in dative bonding, all six hydroxy groups nevertheless Ænd full H-bonding employment in interactions with four phosphoryl O atoms and two hydroxy O atoms from neighbouring residues. Remarkable is the observation that the structure exhibits tubular channels along the c-axis of the crystal, shown in Fig. 6. Despite the structure's Table 3 Positive ion spectra mass spectral dataa for LnL3 (L~O2P(CH2X)2)b [LnL(MeOH)5]2z X~OH X~Cl [LnL(MeOH)6]2z X~OH X~Cl 2]z [LnL X~OH X~Cl 2(MeOH)]z [LnL X~OH X~Cl 2(MeOH)2]z [LnL X~OH X~Cl 2(MeOH)3]z [LnL X~OH X~Cl aData are quoted for the most intense peak expected in the isotope distribution proÆle for a given ion.Numbers in parentheses are the relative intensities of the signal, with valuesv5% not recorded. bThe base peak for all Ln(O2P(CH2Cl)2)3 was due to [Ln(OMe)(MeOH)6]2z. cThe base peak in for this compound (X~OH) at 20 V cone voltage was an unassigned signal at m/z 116.0. dIsotope patterns indicate that clustering is occurring giving higher charged ions in the 20 V spectra. The patterns simplify to correspond to those expected for the monometallic ions at higher cone voltage. 4 CrystEngComm, 2000, 33, 1±6 Nd La 213.6(100) 232.5(50) 211.9(60) 230.0(70) 229.5(90) 248.2(60) 227.9(100) 245.9(80) 410.0d (30) 388.8(5) 462.8d (5) 424.0d (15) 420.9(5) 494.7d 455.9(25) 452.7(20) 526.8 487.9(10) 485.0(5) 558.8 Fig.6 A view of the packing in Pr(O2P(CH2OH)2)3, viewed along the caxis, showing the unit cell (Z~4). Non-hydroxy H atoms are omitted for clarity. architectural debts to electrostatic and H-bonding interactions, the channels are lined mainly with methylene groups: the ligands adopt conformations so as to project their hydrophobic moieties into this tubular channel in a manner reminiscent of solution micelles. When projected along the c-axis, the channel is only apparently about 3.62 A wide. However, since the channel `lining' is not smoothly continuous, originating instead from several different ligands, a molecule such as an aliphatic hydrocarbon might still be accommodated by `weaving' along the channel's general axis.Any molecule putatively within this channel would be afforded 8 close contacts with methylene groups per each unit cell c-axis repeat length of some 11.4 A . Thus, to assess the afÆnity of this channel for hydrophobic molecules, we suspended the solid materials in hexane and Er Eu 218.8(100) 237.0(70) 225.5d (100) 244.2(90) 241.5d (50) 260.5(90) 233.9d (60) 253.0(90) 402.6d (20) 475.0(20) 489.8 448.0(5) 523.7 435d (10) 507.0(5) 479.9(20) 555.8(5) 467.0d (30) 539.0(20) 497.0(20) 571.0(20) 511.9d (15) 587.8(5) Ybc 229.1d (60) 246.9(30) 245.0d (50) 263.1(35) 456.0(5) 488.0d (10) 519.9(10)Table 4 Complex characterisation data Nd(O2P(CH2OH)2)3 Dy(O2P(CH2OH)2)3 Yb(O2P(CH2OH)2)3 La(O2P(CH2Cl)2)3?2H2O Nd(O2P(CH2Cl)2)3?2H2O Eu(O2P(CH2Cl)2)3?2H2O Dy(O2P(CH2Cl)2)3?H2O Yb(O2P(CH2Cl)2)3?H2O heavy petroleum, but no lasting weight gain was observed upon drying of the recovered solids in either case.Interestingly, the largest peak, of some 1.62 e A 23, in the Ænal difference Fourier synthesis is located centrally within the channel, and makes no bonds with any other peak, or any established atom. Whether this truly represents an atom is not clear since, when the `site' is treated as a water O atom, its occupancy reÆnes to only about 13%. In a further attempt to Ænd possible structural correlations with the Ln(O2P(CH2Cl)2)3?nH2O series of complexes, conductimetric and EMS data were collected for Ln(O2P- (CH2OH)2)3; molar conductances of 0.016±0.020 V21 m22 mol21 fall in the range13 intermediate between 2 : 1 and 1 : 1 electrolytes and imply that the second ligand dissociation is not complete for Ln(O2P(CH2OH)2)3.In negative ion EMS the O2P(CH2OH)22 anions again give rise to the base peak in all spectra at low cone voltages, signals due to lanthanide containing ions being present only at v5%. The O2P(CH2OH)22 ion appears to be less stable than O2P(CH2Cl)22 in the gas phase (as do its complexes), presumably due to the presence of a low energy decomposition route unavailable to the latter ion and its complexes. The main fragmentation of the O2P(CH2OH)22 anion is due to successive loss of formaldehyde, and at 250 V H2PO22 becomes the base peak, at m/z~65, with [O2P(H)(CH2OH)]2 evident at m/z~95.At 290 V the base peak at m/z~63 is presumably due to [PO2]2. At higher cone voltages similar fragmentation is observed to a smaller extent for Ln(O2P(CH2OH)2)3, with loss of Hz and formaldehyde apparent for all compounds studied. The results for cation EMS are shown in Table 3. Here [LnO2P(CH2OH)2(MeOH)n]2z and [Ln(O2P(CH2OH)2)2- (MeOH)n]z are both abundant ions, in contrast to the case of their chloromethyl analogues. The conductimetric and EMS studies seem to conÆrm that the coordinating ability of O2P(CH2Cl)22 to lanthanides is lower as compared to O2P(CH2OH)22 for which a stabilising interaction of the 2P(CH2OH)2)2containing ions in Fig.7 Proposed stabilisation of (O the gas phase. Required; Found(%) C 13.88; 13.94 H 3.49; 3.53 C 13.40; 13.21 H 3.37; 3.22 C 13.15; 13.37 H 3.31; 3.17 C 10.91; 11.18 H 2.44; 2.41 C 10.82; 10.59 H 2.42; 2.35 C 10.70; 11.21 H 2.39; 2.33 C 10.82; 10.91 H 2.12; 1.90 C 10.65; 10.75 H 2.08; 1.97 Required; Found(%) Eu(O2P(CH2OH)2)3 Ho(O2P(CH2OH)2)3 C 13.67; 14.03 H 3.44; 3.34 C 13.34; 13.82 H 3.36; 3.49 Pr(O2P(CH2Cl)2)3?2H2O Sm(O2P(CH2Cl)2)3?2H2O C 10.87; 10.83 H 2.43; 2.41 C 10.72; 10.54 H 2.40; 2.34 Er(O2P(CH2Cl)2)3?H2O C 10.74; 11.06 H 2.10; 1.99 hydroxy group with the metal can occur (Fig. 7). A similar interaction is seen in the solid state structure. Experimental General Infrared spectra were recorded as KBr discs on an ATI Mattson Genesis single-beam FT instrument and Raman spectra on crystalline solids using a Foster and Freeman Foram 685 Raman microscope.Electrospray mass spectra were recorded on a VG Quattro II triple-quadrupole mass spectrometer. Samples in solution were loop injected into a stream of water : methanol (1 : 1) passing through a steel capillary held at high voltage (z3.5 kV for positive mode, 23.0 kV for negative mode). Nebulisation of the resulting spray was pneumatically assisted by a concentric Øow of nitrogen, and desolvation aided by a Øow of nitrogen bath gas and heated source (70 �C). Declustering and molecular fragmentation were promoted by increasing the cone voltage from 20 to 90 V. Isotope distribution proÆles were calculated using the ShefÆeld University ChemPuter18 and all assignments are within 0.5 Da of the calculated values.Crystallography Ho(O2PH2)3. X-Ray diffraction data were collected using previously described procedures,19 merged, and the structure solved by isomorphous replacement.8 Non-H atoms were reÆned20 anisotropically and H atoms placed at calculated angles at a common, reÆned distance and bearing a common, reÆned isotropic displacement parameter. On convergence of reÆnement, an absorption correction21 was administered and reÆnement recommenced to reconvergence. La(O2P(CH2Cl)2)3(OP(OH)(CH2Cl)2). Data were collected22 and corrected for absorption.23 The structure was solved by Patterson methods24 and the data then merged.H atoms were reÆned20 isotropically in theoretical positions and with a common, reÆned displacement parameter. Disorder of seven of the eight chloromethyl groups was evident, but the disordered groups required group restraints on all C±Cl, P±C and P±Cl distances so as to provide satisfactory angles and distances at C. One aberrant low-angle reØection with Fo vv Fc was suppressed. The comparatively high thermal parameters of all of the carbon atoms, and certain peaks in the Ænal difference Fourier synthesis, indicated thathe overall disorder may extend to the whole of each ligand. However, no beneÆt was5 CrystEngComm, 2000, 33, 1±6gained by modelling the ligands, so the weakness of the dataset (Rs~30.75%) leading to instability when the number of reÆned parameters was increased. Also for this reason, only the La, P, Cl and O atoms were allowed to reÆne anisotropically in the major conformeric form, and only La and P atoms in the minor form.Whilst the overall disposition of bridging ligands into a columnar array is not in doubt, the poor quality of the data and the necessarily simpliÆed nature of the chosen model require that absolute values of bond distances and angles should be treated with due caution. Furthermore, there might possibly arise systematic bias in the quoted error values themselves. Pr(O2P(CH2OH)2)3. Data were collected,7,25 the structure was solved by Patterson methods24 and the data then merged. Anisotropic non-H atoms and isotropic H atoms with a common, reÆned displacement parameter were used.20 The methylene H atoms were placed in theoretical positions based on the P±C±O angle, and the hydroxy H atoms constrained to positions 30% of the distance across the interoxygen vector of the H-bond.An absorption correction21 based upon the converged model was applied, and the model reÆned to reconvergence. Synthesis The acid OP(OH)(CH2OH)2 was prepared by the reaction of 50% H3PO2 with paraformaldehyde, in aqueous HCl. Chlorination with SOCl2 followed by reduced pressure distillation gave the acid chloride OP(Cl)(CH2Cl)2 which was hydrolysed with water to give the free acid OP(OH)(CH2Cl)2 as a lowmelting waxy solid.26 2PH2)3 was prepared as reported previously.8 Other Ho(O complexes were prepared according to similar procedures, by digesting Ln2O3 with OP(OH)(CH2OH)2 in aqueous suspension on a steam bath, or from ethanolic solutions of Ln(NO3)3 or LnCl3 with OP(OH)(CH2Cl)2, a route necessitated by the low water solubilities of Ln(O2P(CH2Cl)2)3 for the later lanthanides leading to difÆculties in isolation of pure materials due to contamination with unreacted metal oxides.Characterising analytical data are collected in Table 4. Preparations of 2P(CH2Cl)2)3?H2O, Yb(O 2P(CH2Cl)2)3?2H2O Nd(O and Er(O2P(CH2Cl)2)3?H2O by the two methods are given as representative of the general procedures. Yb(NO3)3?5H2O (0.59 g, 1.3 mmol) in ethanol (15 ml) was treated with a solution of OP(OH)(CH2Cl)2 (0.77 g, 4.7 mmol) in ethanol (10 ml).A white precipitate formed on mixing and the suspension was digested on a steam bath for 3 h. The product was Æltered and dried at the pump to give 0.74 g (86%). Nd2O3 (1.10 g 3.27 mmol) was suspended in a solution of OP(OH)(CH2Cl)2 (3.32 g, 20.5 mmol in 20 ml water). The mixture was digested on a steam bath for 16 h after which time a lilac solution with undissolved lilac solid was obtained. Concentration of the Æltrate to 5 ml gave rhipidate crystals (1.24 g). Soxhlet extraction of the residual solid with water for 24 h followed by crystallisation as above gave a further 2.05 g. Overall yield 74%. Er2O3 (0.80 g 2.13 mmol) was suspended in a solution of the acid (1.00 g, 6.14 mmol) and digested for 16 h on a steam bath.6 CrystEngComm, 2000, 33, 1±6 Soxhlet extraction of the reaction mixture with water for 24 h gave 0.28 g (10%) of rhipidate crystals with 1.51 g pink powder undissolved in the Soxhlet thimble. Acknowledgement We thank the EPSRC and the Universities of Wales and Southampton for use of the National X-ray Crystallography and the National Mass Spectrometry Services. We also thank Dr Tom Hamor for the collection of the Ho dataset. One of us (A. W. G. P. ) also thanks the Royal Society of Chemistry for support under the Research Fund. References 1 R. Sanchez-Delgado, J. S. Bradley and G. Wilkinson, J. Chem. Soc., Dalton Trans., 1976, 399. 2 M. Orchin, Adv. Catal., 1966, 16, 1. 3 C. K. Brown and G. Wilkinson, J. Chem. Soc. (A), 1970, 2753.4 C. D. Chang and A. Silvestri, J. Catal., 1977, 47, 249. 5 L. D. Rollmann and E. W. Valyocsik, Inorg. Synth., 1981, 22, 61. 6 L. B. Yong, S. A. Butter and W. E. Keading, J. Catal., 1982, 76, 418. 7 J. A. Darr, S. R. Drake, M. B. Hursthouse and K. M. A. Malik, Inorg. Chem., 1993, 32, 5704. 8 J.A. Seddon, A. R. W. Jackson, R. A. Kresinski and A. W. G. Platt, J. Chem. Soc., Dalton Trans., 1999, 2189. 9 J.A. Seddon, A. R. W. Jackson, R. A. Kresinski and A. W. G. Platt, Polyhedron, 1996, 15, 1899. 10 L. A. Aslanov, V. M. Ionov, M. A. Porai-Koshits, V. G. Lebedev, B. N. Kulikovskii, O. N. Gilyarov and T. L. Novoderezhkina, Izv. Acad. Nauk. SSSR Neorg. Mater., 1975, 11, 96. 11 E. K. Davies, SNOOPI, Chemical Crystallography Laboratory, University of Oxford, UK, 1982. 12 Organic Phosphorus Chemistry, ed. G. M. Kosolapoff and L. Maier, Wiley, New York, 1972. 13 J. D. Lee, Concise Inorganic Chemistry, Chapman and Hall, London, 1991, 4th edition. 14 W. J. Geary, Coord. Chem. Rev., 1971, 7, 81. 15 D. K. Walanda, R. C. Burns, G. A. Lawrence and E. I. Von Nagy- Felsobuki, J. Chem. Soc., Dalton Trans., 1999, 311. 16 C. C. H. Chin, J. S. L. Yeo, Z. H. Loh, J. J. Vittal, W. Henderson and T. S. A. Hor, J. Chem. Soc., Dalton Trans., 1998, 3777. 17 A. W. G. Platt, J. Fawcett, R. S. Hughes and D. R. Russell, Inorg. Chim. Acta, 1999, 295, 146. 18 M. Winter, ShefÆeld ChemPuter, http://www.shef.ac.uk/chemistry/ chemputer/isotopes.html. 19 J. D. McKinney, H. Chen, T. A. Hamor, K. E. Paxton and C. J. Jones, J. Chem. Soc., Dalton Trans., 1998, 2163. 20 G. M. Sheldrick, SHELXL 93, Program for Crystal Structure ReÆnement, University of Go» ttingen, 1993. 21 N. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983, 39, 158. 22 Z. Otwinowski and W. Minor, Methods in Enzymology, ed. C. W. Carter, Jr and R. M. Sweet, Academic Press, New York, 1997. 23 R. H. Blessing, J. Appl. Crystallogr., 1997, 30, 421. 24 G. M. Sheldrick, SHELXS 86, Program for the Solution of Crystal Structures, University of Go» ttingen, 1986. 25 J. W. PØugrath and A. Messerschmidt, MADNES. Program for Area-Detector Data Interpretation, Delft Instruments, Delft, The Netherlands, 1990. 26 B. E. Ivanov and A. R. Pantereeva, J. Gen. Chem. USSR, 1967, 37, 1768.

 



返 回