摘要:

1975 2149 lmine Hydrolysis Reactions in Copper(i1) Complexes of NN'-Ethylenebis-(thiophen-2-carbaldimine) and -(pyridine-2-carbaldimine) By Andrew C. Braithwaite Clifton E. F. Rickard and T. Neil Waters,' Department of Chemistry University The hydrolysis of copper(l1) complexes containing the ligands ~~'-ethylenebis-(thiophen-2-carbaldimine) and - (pyridine-2-carbaidimine) has been studied. Two distinct reactions involving each imine function in turn, have been observed for the former and with evidence provided by the other ligand have allowed a reassessment of the reasons for the occurrence of hydrolysis in Cu" complexes. An electronic basis with initial polarisation of the imine groups is suggested. of Auckland New Zealand RECENTLY there have been several reports describing solvolysis reactions of copper( 11) complexes containing co-ordinated imine groups which usually produce complexes with amine donors.1-* The reactions have been con-sidered to be a consequence of electronic forces although a different argument based on steric interactions has also been advanced particularly for complexes containing the ligand NW-ethylenebis(pyridine-2-carbaldimine) ? When complexed with CuII this ligand undergoes partial solvolysis involving addition across one or both of its imine functions 2 ~ 3 since it is said the steric strain in the square-planar ligand system is relieved by formation of the more flexible COR-NH linkage.Research4 with similar ligands has been quoted as verifying the hypo-thesis that strain is the major cause of subsequent solvolysis.However the realisation that such reactions are not confined to one type of ligand and that they appear to be specific to copper in many instances suggests that a steric origin may not be of paramount importance. We have now studied the hydrolysis of NN'-ethylenebis-(thiophen-2-carbaldimine) a ligand known to hydrolyse rapidly 5-8 on complex formation with Cuxl and present evidence suggests that such solvolysis reactions are mainly of electronic origin with steric effects being of a minor nature. It has been possible to retard solvolysis sufficiently to observe two distinct cleavage reactions involving each irnine linkage in turn. As subsequent crystal-structure determinations 9~10 have verified that the thiophen sulphur atoms are not co-ordinated to the metal centre it is apparent that steric arguments cannot be applied in this instance and yet similar hydrolysis reactions have been found to occur.EXPERIMENTAL The ligand NN'-ethylenebis(thiophen-2-carbaldimine) (L1) was prepared by condensing thiophen-2-carbaldehyde (2.24 g) with excess of ethylenediamine (en) (0.62 g), recrystallisation being effected from ethanol or diethyl ether to produce straw-coloured needles of m.p. 88-89 "C (lit.,G 92 "C). Complexes of L1 were initially prepared by mixing L1 (2.48 g) with copper(I1) chloride dihydrate (1.71 g} or copper(I1) perchlorate hexahydrate (2.48 g) in 1 1 ratios 1 L. F. Lindoy Quart. Rev. 1971 25 379. C. M. Harris and E. D. McKenzie J . Chem. SOC. ( A ) 1969, * D. W. Busch and J.C . Bailar J . Amer. Chem. SOL 1956 78, M. Cressey E. D. McKenzie and S. Yates J . Chem. SOC. ,A), G. L. Eichhorn and I. M. Trachtenberg J . Amer. Chem. 746. 1137. 1971 2677. SOC. 1964 '76 5183. in methanol. The two salts produced light and dark green needles respectively and these were washed with dry diethyl ether before being placed in a desiccator over silica gel. Recrystallisation was not possible because of the /7 N N L' L2 L3 tendency of the complexes to hydrolyse. Since different stoicheiometries have been reported ' 9 * for these complexes the preparations were repeated a t 1 2 copper to ligand ratios With perchlorate as anion the same dark green precipitate was obtained ; with chloride however no pre-cipitation occurred and subsequent hydrolysis reactions produced a mixture of L1 thiophen-2-carbaldehyde and en salts of CurT.Thus only one product is obtained from each salt regardless of the component ratios. When an excess of solvent was used perchlorate precipit-ation did not immediately occur and the solution turned purple overnight. Careful solvent control allowed crystal-lisation of a purple complex at this stage and i.r. spectra, to be discussed later showed that thiophen moieties were present. Analyses (Table 1) confirmed that a complex of formula [CU(L~)~][C~OJ~ rather than [C~(en)~][C10,] had formed (La = N-2-aminoethylthiophen-2-carbaldimine). The complex [Cu ( L1) 2] [ClO,] also decomposed in the crystalline state exposure to air for 36 h or more producing a purple-grey powder with an elemental analysis consistent with the formula [CU(L~)~][C~O,]~.That the original dark green complex was [CU(L~)~][C~O,] was confirmed in an independent copper analysis on a freshly prepared sample by precipitating copper(1) thiocyanate l1 (Found Cu 8.30. Calc. for [Cu(L1),][CIO,] Cu 8.35. Calc. for [CU(L~)~]-[Clod Cu 11.15%). G. L. Eichhorn and J. C . Bailar jun. J . Amer. Chem. Soc., 1953 75 2905. R. K. Y. Ho and S. E. Livingstone Austral. J . Chem. 1965, 18 669. M. P. Coakley L. H. Young and R. A. Gallagher J . Inorg. Nuclear Chem. 1969 31 1449. @ A. T. Morcom and C. E. F. Rickard unpublished work. lo G. D. Beckingsale and T. N. M. Waters unpublished work 2150 J.C.S. Dalton TABLE 1 Analytical data (%) C H Calc. for [Cu(L1),][CIO& 37.96 3.20 Calc.for [Cu(L2),][C1O4] 29.45 3.55 Found (for decomposed L1 29.8 3.85 Found (for purple crystals) 30.05 4.30 Calc. for [Cu(L3)][C104] 33.6 2.80 Calc. for [Cu,(L3),][C104] 40.7 3.40 Found (green form) 37.2 3.90 Found (blue form) 36.5 4.00 complex) Calc. for [Cu2(LS)C1&1 33.15 2.80 Calc. for [Cu2(L*)ClJCl2* 33.3 3.35 AMeOH F(und (green form) 32.75 3.45 Calc. for [Cu(L3)]Clz.0H 43.0 4.10 Calc. for [Cu(L3)]Cl2*20H 41.15 4.45 Found (blue form) 43.05 4.25 Calc. for [Cu(L3)]C12 35.1 3.75 The chloride complex was confirmed N 7.40 9.80 9.40 9.70 11.2 13.55 12.65 12.45 11.05 10.36 10.65 15.05 14.35 13.7 14.65 c1 8.36 12.4 12.6 12.45 12.7 11.45 11.95 12.15 27.95 26.3 25.85 19.0 18.2 17.36 17.9 by analysis as [Cu(L')]Cl rather than [Cu(L1)JC12.It is stable for several months as a solid but alcoholic solutions slowly turn blue and eventually deposit blue crystals of a mixture of [Cu-(L2) JCl and [Cu(en)]Cl,. It was impossible to crystallise a pure sample of the L2 complex analogous to that obtained with perchlorate. Formation of [Cu(L2)]C12 has neverthe-less been confirmed by recent structural data,@ a dimeric [Cu,(L2) ,Cl,]Cl species with bridging chloride ions being found. The ligand NN'-ethylenebis(pyridine-2-carbaldimine) ( L3) was prepared by condensing pyridine-2-carbaldehyde (2.14 g) with en (0.62 g) in ethan01.~ Rapid cooling in a refrigerator resulted in straw-coloured needles of m.p. 65 "C (lit.,3 68 "C).Copper(I1) complexes using both perchlorate (1.31 g) and chloride (0.85 g) salts were prepared by addition of a methanolic solution to L3 (1.19 g) in 1 1 ratios With perchlorate as anion dark blue crystals were at first obtained but later preparations under supposedly identical conditions produced a black semicrystalline pre-cipitate and a light green powder. 1.r. spectra suggested that these three products were different forms of the same complex and analytical data were consistent with this, although some impurity was apparently present. Harris and McKenzie reported [CU,(L~)~][C~O~]~ as a contaminant and the higher C H and N percentages and lower C1 value found here are consistent with the formation of a complex with a ligand to copper ratio greater than unity and are inconsistent with a simple hydrolytic impurity.In the absence of suitable crystals for X-ray structural analysis, solutions of the CuII complexes eventually decomposing to en salts no further deductions were possible. With chloride as anion a green powder initially precipitated but blue crystals also began to form quite rapidly. To isolate the powder the reaction solution was filtered immediately, elemental analysis then verifying that a complex [Cu2(L3)-Cl,]C12 rather than [Cu(Ls)]Cl had formed in agreement with predictions of Harris and McKenzie but not of Busch and Bailar.3 Analysis of the blue crystals revealed a complex [Cu(L3*OH,)]C1, and indicated that partial hydro-lysis across one C=N linkage had taken place. Thus reactions with the ligand L3 are in general agree-ment with the conclusion of Harris and McKenzie rather than Busch and Bailar.3 However we find no verification for the suggested nature of the contaminants present nor did we have success with the suggested improvement in technique to remove such species.Kinetic data for the hydrolysis of the ligands and their copper@) perchlorate complexes were obtained using 50% methanol-water or spectroscopic (but undried) methanol as solvents. Changes were followed on a Unicam SP 800A visible-u.v. spectrometer. The solution concentrations were ca. 5 x 1 0 - 3 ~ for complexes and ca. l o - 5 ~ for the ligands spectra being obtained a t suitable intervals until the hydrolysis was complete. * RESULTS AND DISCUSSION The synthetic work has revealed that two separate hydrolyses occur with complexes derived from L1 and that these involve each of the C=N linkages in turn.Except for one report of a NiII complex no complexes containing the ligand L2 have been observed in this reaction. Identification of such products has however, an important bearing on later discussions of steric strain. 1.r. spectra obtained for complexes of L1 L2 and en are consistent with their formulations (see Table 2). TABLE 2 1.r. data (cm-l) Compound v(0H) v(NH) (vC=N) 6(NH) v(C10,) 1 630 1616 1100 3260 1626 1680 1140,* 3 225 1 100, 1 080, 1 020 * 1 636, 1616 3250 1620 1670 3 220 [CU (L1)1C12 [CU (L2)1 C l , L3 1 646, 1685 1663 1100 1 600 [Cu2(L3)Cl,]C1,~~MeOH 3 450 1 640, 1 595 [Cu(L3)]Cl,*OH2 3 380 3 250 1596 1690 * Finely resolved doublet.[CU(L~)I [CQI, As already observed the presence of thiophen rings is indicated by strong ' aromatic' bands in the region 700-800 cm-l. The appearance of a broad 6(NH) absorption at 1580 cm-l and a v(NH,) doublet in the region 3 200-3 300 cm-l confirms the initial hydrolysis reaction and final cleavage to en salts is verified by loss of both v(C=N) and ' aromatic ' bands. An interesting feature of the spectrum of [Cu(L2)J[C10,] is the splitting of the perchlorate band at 1 100 cm-l (vJ in place of the single absorption observed with [Cu(L1),][C1O,),. This suggested that the perchlorate groups are bound to the copper centre and a strong bonding interaction has since been confirmed by X-ray crystallography.1° Visible and U.V.spectroscopy have been used to study the hydrolysis reactions in water and methanol solutions. Because of the similarity of the spectra of L1 and thiophen-2-carbaldehyde it was necessary to study the position of the d-d envelope as the main guide to hydro-lysis reactions for L1 and L2 complexes (see Table 3). * 1~ = 1 mol dm-3. 11 A. I. Vogel ' Textbook of Quantitative Inorganic Analysis,' 2nd edn. Longmans Green and Co. London 1959 p. 43 1975 2151 All but one of the complexes produced similar spectra in methanol and water [Cu(L1)J[C10a2 showing evidence of immediate hydrolysis in water. Kinetic data obtained in undried spectroscopic methanol solution and 50% Visible and U.V. spectra (Table 3) did not produce the same clear shifts in absorption maxima found for L1 complexes but rate data were unequivocal in showing that the fast hydrolysis observed for L1 complexes did TABLE 3 U.v.-visible spectra (cm-1) a Complex [WL3)ICl2 ccu (L~)IC~, Ligand Thiophen-2-carbaldeh yde L1 L3 16 100 (40) 15 800 (50) 15 200 (50) 15 400 (50) 15 500 (60) 16 500 (150) 14 400 (100) 16 000 (100) 34 700 (20 500) 18 800 (70) 33 900 (30 000) 18 300 (80) 34 400 (14 900) 18 700 (90) 34 000 (19 300) 34 100 34 000 (12 400) 34 000 (5 100) 33 500 (10 200) 34 800 (15 600) 32 900 (3 000) 35 300 (10 900) 34 800 (12 000) 34 400 (5 600) 35 000 (23 700) 35 000 (5 300) 26 700 C (140) 27 000 c (100) 33 400 (10 000) 38 800 (33 500) 38 600 (29 800) 38 200 (37 000) 27 400 38 000 (24 800) 39 200 (5 600) 36 100 (16 900) 37 600 (15 600) 39 700 (14 800) 43 900 (16 400) 38 800 (10 000) 42 900 (20 000) 38 500 (11 300) 38 800 (26 200) 37 000 (12 600) 42 900 (19 900) 37 900 (55 000) Solvent MeOH G$H 3:H H2O H2O MeOH H2O MeOH MeOH MeOH a Absorption coefficients ( ~ / 1 mol-1 cm-1) are given in parentheses.b Due to partial hydrolysis to [CU(L~)J[C~O~]~. Shoulder on a more intense absorption. methanol-water (Table 4) verified that a fast hydrolysis reaction did indeed occur the d-d absorption band for [Cu(L1)d [ClO,] quickly shifting to higher energies typical of [Cu(L2),] [ClO,] and [Cu(en)d [CIO,],. The rate constants showed good agreement with those obtained for [ Cu ( L1) 2] C1$ TABLE 4 Rate data in a methanol-water (1 1) a t 293 K Compound k1s-l (Wm) Wavelength of measurement [CU (L1) 21 K10412 1.3 x 10-3 570 2.2 x 10-3' 670 2.1 x 10-35 560 1.1 x 10-3 380 ca.10-35 380 4.2 x 10-5 450 1.1 x 10-3 285 L1 L3 2.1 x 10-4 235 [cU(L2) 23 [c10412 [WL3)lCClOJ2 a Obtained in undried methanol solution. 8 Obtained a t pH 8.6. Stable in MeOH over 24 h. A spectral study of complexes involving L3 was also undertaken. 1.r. spectra for the three perchlorate complexes (Table 2) showed two v(C=N) bands at 1600 and 1653 cm-l consistent with partial cleavage across only one of the imine linkages. No splitting in the perchlorate bands was observed for these complexes. The green chloride complex showed similar bands at 1 595 and 1 640 cm-l; however only one band at I595 cm-l remained for the blue chloride complex.It was the absence of a second v(C=N) band that prompted Busch and Bailar to propose a complex in which both aliphatic imine groups had been partially hydrolysed, but this is incompatible with the analytical data.2 Thus a shift in the second v(C=N) band to 1 600 cm-l where it would be hidden under the aromatic v(C=N) peak as suggested by Harris and McICenzie,2 is supported. not occur. It is pertinent at this stage however to compare both the rate constants for hydrolysis of the ligands L1 and L3 in company with their complexes before making an assessment in terms of ' electronic ' arguments all rate constants (first order) being shown in Table 4. It has already been noticed that hydrolysis of [Cu-(L1)2]2+ to [Cu(L2)J2+ and subsequently to [Cu(en)J2+ proceeds at a far greater rate than for [Cu(L3)J2+ the first-order rate constants differing by almost two orders of magnitude.Comparison with ligand solvolysis shows that the hydrolysis is accelerated when L1 is complexed particularly when smaller amounts of water are present whereas L3 itself hydrolyses at a faster rate than its copper complex. Such an observation is easily explained by noting that bond rupture absent for the L1 complex must occur in the L3 complex to allow hydrolysis to occur. The retardation rate does not, however imply a steric origin for hydrolysis whereas the acceleration in L1 complexes does support an electronic mechanism. We had also intended to make a study of the reaction at pH values >7 to verify that the same mechanism of hydrolysis applied to both complexes and in particular that OH- attack would occur to cause an increase in the rate constant.Unfortunately L3 complexes formed dark solutions in basic conditions even at pH 8 and no hydrolysis data could be obtained. Initial reaction, however appeared to be more rapid as it did with L1 complexes where an almost two-fold increase in rate constant was observed at pH 8.9. This preliminary kinetic evidence can now be added to the many unusual reactions of imines co-ordinated to CuII and an assessment of such reactions can be made. The electronic argument is based on the charge separ-ation that occurs across the imine linkage as a result o 9152 J.C.S.Dalton electron donation to the central metal ion; the imine carbon atom is then susceptible to nucleophilic attack by, for example OH- or OR- from solvents. Thus many CuII complexes are hydrolysed even though the ligands themselves are relatively stable. (In some instances where the reverse is true special stabilising features of complex formation are present which outweigh hydroly-tic tendencies a feature discussed in detail by Lindoyl in a recent review.) Steric causes have been postulated largely as a result of studies of L3 and derivatives where it is said that partial hydrolysis to the more flexible COR-NH linkage occurs to relieve the strain imposed by quadridentate co-ordination. Consideration of the reactions of the L1 and L2 complexes throws considerable doubt on such a theory however.The establishment of the two distinct hydrolysis reactions suggests that the only difference between L1 and L3 systems is the more rapid hydrolysis of the former which can be explained by noting that a chelate ring does not have to be broken in the bidentate complexes of L1 since the sulphur atoms unlike the ring nitrogens of L3 do not co-ordinate. Steric arguments do not therefore apply to L1 because there is no strain in the ligand system. Introduction of the known reactions of NN’-ethylenebis (pyrrole-2-carbaldimine) (L4) also contradicts the steric hypothesis. Molecular models built for L1 L3 and Lp show that when complexed with CuIr in a square-planar arrange-ment the order of steric strain is L4 > L1 > L3 although L1 does not in practice act as a quadridentate ligand.Little adjustment in the normal bond lengths and angles is needed to bond L4 as a quadridentate ligand even though 1 1 complexes of L4 are not only stable with the pyrrole nitrogen donors apparently bonded to the copper centre but also show no cleavage across their C=N linkages.l*le This resistance to hydrolysis in the presence of steric strain is clearly inconsistent with a steric argument for hydrolysis. Three further hydrolysis reactions are significant in this context Hydrolysis of NN’-ethylenebis(quin0-line-8-carbaldimine) l7 is said to occur only where the ligand is required to be planar by metal-ion constraints because of consequent steric clashes between hydrogen atoms in the 2-positions of the quinoline rings.These, it is supposed cause the imine functions to twist and subsequently cleave because of the loss of x overlap. However it is difficult to understand why the flexible en l2 R. E. Clarke and J. H. Weber J . Inorg. Nuclear Chem. 1968, 18 A. Chakravorty and T. S. Kannan J. Inorg. Nuclear Chem., l4 R. H. Holm A. Chakravorty and VV. J. Theriot Inorg. l5 J. H. Weber Inovg. Chem. 1967 6 268. 80 1837. 1967 29 1691. Chem. 1966 5 626. rings cannot twist to relieve some of the strain and why the non-planar complexes should not hydrolyse when, presumably strain and lack of x delocalisation is also present. It seems far more likely that when the ligand is planar a greater charge separation occurs across the imine linkages the maximum st interaction in the chelate ring then being possible and that for the planar geo-metry only this is just sufficient to cause hydrolysis to occur.The most convincing argument for an electronic origin to account for cleavage of bound imine groups can be found in reactions l8 of the ligand L5. This contains three aliphatic imine three aromatic imine and a tertiary amine nitrogen as potential donors. With many metal ions it co-ordinates through the six imine donors, but for CuII only a subsequent hydrolysis occurs which involves two of the three equivalent aliphatic donors. In addition the tertiary amine nitrogen atom is brought into the co-ordination sphere occupying an axial position of a trigonal bipyramid as in (I). The suggested reason 2+ I Ill for this observation namely that the driving force for the reaction is the tendency of CuII to adopt a five-co-ordinate stereochemistry is then an electronic one.The hypothesis is supported by the stability of one imine group whereas a steric argument would require that one or all be hydrolysed. A further example of the hydrolysis reaction occurring only in a CuII complex has recently been reported.lg It involves the thiazole ligand La. On complex formation with CuII 1 mol of ligand is reduced to a thiazoline, whereas a further mol is hydrolysed across the C=N linkage and oxidised to a pyridinecarboxylate. The five-co-ordinate complex shown in (11) then results, co-ordination being through three imine linkages (two aromatic and one aliphatic) and an oxygen from the carboxylate ligand and from a co-ordinated water molecule.This peculiar redox reaction occurs only with CuII and again supports the view that hydrolysis reactions are electronic in origin. Thus in all instances where steric strain has been suggested as the fundamental cause of hydrolysis several doubts have arisen and alternative electronic mechanisms can be proposed. These involve polarisation of the C=N linkage. This view also accounts for the stability of 16 K. N. Yeh and R. H. Baker Inorg. Chem. 1967 6 830. l7 J. Dekkers and H. A. Goodwin Austral. J. Chem. 1966 19, 18 E. C. Lingafelter L. C. Andrews R. M. Kirchner N. J. Rose, 19 A. Mangia M. Nardelli C. Pelizzi and G Pelizzi J.C.S. 2241. and L. J. Wilson Co-ordination Chem. Rev. 1972 8 55. Dalton 1972 2483 1975 2153 L4 complexes since L4 on co-ordination loses two is absent a charge-balancing mechanism to satisfy protons and is effectively a 2- ion with the charge Pauling's electroneutrality principle2* must involve a largely remaining on the pyrrole nitrogen atoms. The considerable degree of polarisation of the imine donors two negative charges effectively cancel the charge of the and a greater susceptibility to attack by nucleophiles. + copper centre and polarisation of the irrtine groups does Thus a simple electronic explanation can be advanced not occur to any great extent. By contrast in the to account for hydrolysis reactions in copper(I1)-imine neutral ligands L1 and L3 where an electrostatic effect systems, which is not to say that structural effects are 2o L. Pauling ' The Nature of the Chemical Bond,' 2nd edn., Oxford University Press 1940. [4/1878 Received 16th September 19741 Of no consequence

ISSN:1477-9226    

DOI:10.1039/DT9750002149

出版商:RSC    

年代:1975

数据来源: RSC