摘要:

J. CHEM. SOC. DALTON TRANS. 1985 1077Electron Spin Resonance Studies of Iron( 111) Complexes of Ethylenediamine-tetra-acetate and N-(Z-Hydroxyethyl)ethylenediamine-NN'N'-triacetatein Co-ordinating SolventsCatharina T. Migita," Kotaro Ogura, and Takashi YoshinoDepartment of Applied Chemistry, Faculty of Engineering, Yamaguchi University, 2557 Tokiwadai, Ube 755,JapanRemarkable solvent dependence of the apparent peak-to-peak linewidth (AHpp) has beenobserved for the rhombic type of e.s.r. spectra of the high-spin Fe"'(edta) (edta = ethylene-diaminetetra-acetate) and Fell'( hedta) [ hedta = N- (2- hydroxyethy1)ethylenediamine-N"N'-triacetate] complexes in a series of co-ordinating solvents. Moreover, the order of the solventswhich gave the characteristic AHp for each complex was different for the two complex systems:i.e.H,O 2 dimethyl sulphoxide (dmso) > NN-dimethylformamide (dmf) > CH,OH for Fe"'(edta)and H,O > CH,OH > dmf > dmso for Fe"'(hedta). Solvent dependence of the band maxima(Amax) corresponding to the d-d transition has been also observed for the same series of iron(iii)complex-solvent systems although the hmax values obtained were less sensitive to the solventthan the AHpp values. In this case, the solvent dependence of A,,,, for Fe"'(edta) and Fe"'(hedta)was similar: A,,, was smaller for H,O and CH,OH (<600 nm) than for dmf and dmso (>700 nm).The co-ordination behaviour of the solvent molecules to these iron(it1) complexes is discussed inrelation to both the AH and Amax, values.It is revealed that in the case of rhombic high-spiniron(iii) complexes witrsuch edta and hedta, AHpp values obtained by e.s.r. spectroscopyreflect small changes in the co-ordination circumstances; Amax, values obtained from electronicspectra did not.Iron(1II) in aqueous solution has been known to form variableco-ordination with multidentate ligands such as ethylene-diaminetetra-acetate (edta). ' This multidentate co-ordination isaffected not only by the pH of the solution but also by the typeof solvent, as observed through Mossbauer measurements' andabsorption ~pectroscopy.~ However, the structural informationabout the chelated iron(II1) complexes given by the Mossbauermethod was not direct nor precise. Absorption spectroscopy hasyielded more reliable information, which proposed that theband maximum (IL,~,J corresponding to the d-d transition6A,,+4T2s can be an indicator of the co-ordination numberof Fe"'(edta)(solvent) type complexe~.~ However, the h,,,,values neither depended on the nature of the solvent norcorrelated with the Mossbauer parameters.Therefore, to makeclear the co-ordination behaviour of Fe'"(edta) complexes inco-ordinating solvents, another method is still required.In this study, e.s.r. spectroscopy was applied because this isone of the most powerful methods to analyse the ligand field inparamagnetic transition metal complexes. Since it is known thatFe"'(edta) complexes in frozen aqueous solution are in therhombic high-spin ~ t a t e , ~ .~ . ~ it was hoped that the e.s.r. spectramay be sensitive to changes in the co-ordination environment.The e.s.r. measurements were performed for Fe"'(edta) andFe"'(hedta) [hedta = N-(2-hydroxyethyl)ethylenediamine-NN'N'-triacetate] complexes in the frozen co-ordinatingsolvents water, methanol, NN-dimethylformamide (dmf), anddimethyl sulphoxide (dmso).ExperimentalMaterials.-Iron(m) ammonium sulphate, Na[Fe"'(edta)],Na,(H,edta), and Na,(hedta) were Katayama Chemical reagentgrade. Dimethyl sulphoxide, NN-dimethylformamide, andmethanol solvents were all specially prepared reagents ofNakarai Chemical Limited. Water used as solvent was distilledion-exchanged.Sample Preparation.-For the preparation of Fe"'(edta)complexes, Na[Fe"'(edta)] (1 mmol dmP3) was dissolved into theappropriate solvent.This method ensured that the formationof iron(II1) hydroxide was negligible. Fe"'(hedta) complexsolutions were prepared by dissolving iron(rr1) ammoniumsulphate (1 mmol dm-3) into the Na,(hedta) solution (3 mmoldm-3). Aqueous solutions were adjusted to pH 7 by mixingcitric acid (0.1 mol dm-3) and 0.2 mol dm-3 of disodiumhydrogenphosphate solutions. For the e.s.r. experiments, 0.3cm3 of the complex solution was transferred into a quartz tube(4-mm outside diameter), evacuated, and shielded. In thespectroscopic experiments, complex solutions containing 10mmol dm-3 of Fe3 + were measured using a 10-mm quartz cell.Spectroscopic Method.-First-derivative e.s.r. spectra at X-band frequencies (9 150 k 10 MHz) were obtained using aJEOL JES-ME-IX spectrometer system.E.s.r. measurementswere performed at 77 K using a liquid nitrogen dewar. Fieldcalibrations were made with a benzene solution of diphenyl-picrylhydrazyl (dpph; g = 2.003 54).The absorption spectra were measured with a Hitachi model100-50 spectrophotometer at ambient temperature.ResultsE.S.R. Spectra ofFe"'(edta) and Fe"'(hedta) Complexes in Co-ordinating Solven ts.-E.s. r. spectra of Fe"'( ed t a) and Fe"'( hed t a)complexes in various solvents are shown in Figure 1. Theapparent peak-to-peak linewidth (AHpp) of these pseudo-singletspectra varied from 2 to 16 mT depending on the solvent. Inaddition, two complexes in this study gave almost the same e.s.r.spectra in an aqueous solution of pH 7 but gave completelydifferent ones in other solvents (i.e.CH30H, dmf, dmso). As aresult, the order of solvents with decreasing AHpp wasH,O 2 dmso > dmf > CH30H for Fe"'(edta) and H 2 0 >CH30H > dmf 2 dmso for Fe"'(hedta)1078 J. CHEM. SOC. DALTON TRANS. 1985Fe(edta1AH,,I IFe(hed ta)VC00u)Q:nnLL 1 I I I I J100 200 LooHImT1 1 I I 1 1 1 10 100 200 300HImTFigure 1. E.s.r. spectra obtained from Na[Fe"'(edta)] (1 mmol dm-3) inH 2 0 (pH 7) (a), dmso (b), dmf (c), methanol ( d ) and from Fe"'(hedta) inH 2 0 (pH 7) (e), methanol (f), dmf (g), dmso (h). Recorded at 77 KAn additional weak signal was observed to low field of themain signal in every e.s.r. spectrum of these complexes, which isconsidered to originate from the transition between anotherKramers doublet.'The e.s.r.parameters obtained from Figure 1 are summar-ized in the Table.Absorption Spectra-Absorption spectra of Na[Fe"'(edta)]were measured in the same series of solvents used for the e.s.r.experiments (Figure 2). Although the band structure is not asclear beyond 500 nm, similarity of spectral features was foundbetween (a) and (b), both having A,,,. around 800 nm, and alsobetween (c) and (d), and having h,,,. around 600 nm.Absorption spectroscopy was also applied for the Fe"'(hedta)complex but the band structure of the spectra was too broadand ambiguous to estimate the band maximum.The absorption data for the Fe"'(edta) complex are also listedin the Table together with the results for [Fe"'(hedta)(H,O),]reported by Garbett et d3E.S.R.Results.-E.s.r. features of the iron(m) edta and hedtacomplexes in co-ordinating solvents revealed that the ligandfield around Fe3+ was weak and strongly rhombic. Thisrhombic high-spin iron(n1) is described by the spin Hamiltonianof equation (1); precise reports of theoretical analysis of the e.s.r.spectra were presented by Aasa and co-worker~.~*~ Here, S, S,,S,, and S, are the spin quantum number and its (x, y , z )components, D and E are zero-field splitting parameters, pB isthe Bohr magneton, and B is the magnetic flux density at theresonance field.Using the results of their third-order perturbation calculationfor a transition between the middle Kramers doublet, expres-450 550 650 750 850h /nmFigure 2.Absorption spectra of Na[Fe"'(edta)] in dmf (a), dmso (b),H 2 0 (c), and methanol ( d )sions for the principal g values were obtained as in equations(2 )-( 5 1 -g , = 30/7 - (12 240/2 401)q2 - (2 880/2 401)r2 (2)gz,y = 30/7 f- (120/49)q -(1 320/2 401)q2 - (1 620/2 401)r2 (3)where= (1 - 3E/D)/(1 + E/D) (4)andr = -2(pgB/D)/(1 + E/D) ( 5 )Equations (2>-(5) imply that the principal g values vary withq and r, i.e. with the value of E/D. Actually, when the E/D valuechanges from 0.32 to 0.21, contribution of the second terms ofequations (2) and (3) changes from 0.0046 to 0.55 and from 0.07to 0.81. The rest of the terms contribute one order of magnitudeless than these.Wickman et af.' and Aasa5 have presented gversus E/D and bv versus g/g' plots, respectively, which areuseful to predict the ligand-field dependence of the g anisotropy.In order to interpret the large apparent linewidth variation(from 2 to 16 mT) observed in this study, we also estimated the gversus E/D plot which can simulate the experimental results.The best-fit parameters of g (4.5, 4.2, 4.1), AHpp = 13.9 mT,E/D = 0.3, and D = 0.74 cm-' were obtained for theexperimental values of g (4.55,4.36,4.11) and AHpp = 13.8 mT,which were given by Fe"'(edta) complexes made from aerated[Fe"(Hedta)]- in aqueous solution of pH 7.* For a constant* This system gave the best resolved rhombic e.s.r. spectra among thestudied systems. Accordingly, estimated g values are the most reliable sothat the simulation calculation for this system is worthwhileJ.CHEM. SOC. DALTON TRANS. 1985 1079Table. Absorption and e.s.r. spectroscopic data and solvent parameters E and ETFe"'(edta) Felt'( hed ta) - rSolvent 104AHpp/T Lax./nm 1O4AHPp/T h,,,,"/nm E (25 oC)b €,'/kcal mol-'H2O 157 575 155 540 78.4 63.1Methanol 20 560 47 600 32.6 55.5dmf 60 810 19 725 37 45.0dmso 155 810 18.5 725 46.6 43.8a Ref. 3. Dielectric constant, ref. 9. Ref. 8 (cal = 4.184 J).5.04.5br4.03 . 5I O ~ A H ~ ~ I T187 13936 2 227 il60i 96 9 41 1 1 1 1 l l 1000aoD0A & A '0AA000 @O0.25 0.30 0.35EIDFigure 3. Plot of high-spin iron(ii1) g values [(A) gx, (0) g,, (0) g,] forthe middle Kramers doublet as a function of the rhombic character ofthe field (E/D) and the peak-to-peak linewidth, AHpp, estimated fromthe difference of the resonant field corresponding to the largest and thesmallest principal g values.D = 0.74 cm- ', B = 0.16 Tvalue of E/D, e.g. 0.3, an increase in D from 0.5 to 1.2 cm-' madeeach of the principal g values increase by ca. 0.03. In contrast,slight change in E/D resulted in a large anisotropy change of theg values (Figure 3).The apparent peak-to-peak linewidth observed for Fe"'(edta)and Fe"'(hedta) complexes in this study might be affected by thenon-homogeneous broadening owing to the existence of species,differing for example, in the degree of polymerization, or in thedisplacement of the co-ordinating sites of edta by the solvent.However, these are unlikely because no additional signals wereobserved characteristic of the polymerized iron(rr1) complexesand more strongly co-ordinating solvents such as dmso and dmfyielded smaller AHpp values for Fe"'(hedta) complexes.Moreover, solvent dependence of AH for Fe"'(edta) wascompletely different from that for Fe"(hedta).Therefore, itseems reasonable to consider that the large solvent dependenceof AHpp is mainly due to variation of E/D, especially thevariation of E.*Since indistinguishable e.s.r. spectra for Fe"'(edta) andFe"'(hedta) in a neutral aqueous solution implied that the edtamolecule in this solution was five-co-ordinate, i.e. Hedta3 -, thesame as hedta, the E/D variation in these complexes in differentsolvents probably arose from the type and the number of the co-ordinating solvent molecules.As for the number of the co-ordinating solvent molecules,Garbett et aL3 have reported that both Fe"'(edta) andFe"'(hedta) were seven-co-ordinate in H 2 0 and methanol andsix-co-ordinate in dmf and dmso on the grounds that theabsorption peaks, h,,,., corresponding to the 6 A 1,-+4T2gtransition were characteristic of the co-ordination number: i.e.six-co-ordinate complexes have a h,,,, > 700 nm and the seven-co-ordinate ones (600 nm. As shown in the Table, however,h,,,, values were not dependent on the type of solvent andrepresented no difference between edta and hedta complexes.Variety of the solvent dependence of the g anisotropyobserved as an apparent spectral width indicated that the e.s.r.spectrum is very sensitive to the change in the ligand-fieldsymmetry.According to the estimation shown in Figure 3,complexes with smaller AHpp values have larger rhombicity ofthe ligand field. Since we observed completely different resultsfor Fe"'(Hedta) from those for the Fe"'(hedta) system, it cannotbe concluded simply that more strongly co-ordinating solventscause larger rhombicities. In the hedta system, however, thisseems to be true since solvent dependency of h,,,, and AHppshowed an inverse correlation for the studied solvents: thestronger axial ligand will destabilize the 6 A l g . ground statewhich has half-filled d orbitals more than the excited 4T2e statewhich has a vacant e, orbital, so that the larger h,,,.valueresulted. Generally, the stronger ligand at the sixth co-ordination site leads to the smaller rhombicity because it wouldoccupy as close a position as possible to the bisector of the x andy axes and destabilize the dyz and dz, orbitals to about the samedegree. However, when the configuration of the sixth ligand isfixed out of the bisector, the situation could be inverse becausethe dyz orbital would be destabilized more than dzx. The co-ordination of the 'stronger' ligand makes the rhombic splittingenlarge more than the tetragonal splitting; that is, the 'stronger'axial ligand possibly causes the larger rhombicity and will givethe smaller AHpp- This seems to be the case for Fe"'(hedta).Interestingly, in the Fe"'(hedta) system, the solvationparameter,' ET, showed a good correlation with A,.for thestudied solvents (Table). Solvation may work to increase thecubic field strength and destabilize the 4T2g state more by theincreased pairing energy, and consequently results in a lowerh,,,, for a larger ET. The smaller rhombicity in the solvents oflarger ET is accord with this.Unexpected results for the Fe"'(Hedta) system can beattributed to the unco-ordinated portion of edta, i.e.* Since the studied systems did not provide three accurate principal gvalues, we could not perform a perfect simulation calculation. However,using the apparent g values, which were estimated at the saddle,crossover, and bottom points of the spectrum, we obtained reasonablefigures of E/D: Fe"'(edta) in H,O and in dmso, E/D = 0.295; in dmf,0.318; in methanol, 0.328; Fe"'(hedta) in H,O, E/D = 0.295; inmethanol, 0.320; in dmf and in dmso, 0.3281080 J.CHEM. SOC. DALTON TRANS. 1985-CH,COOH (cf: -CH,CH,OH in hedta). It is likely that thefree carboxymethyl group in edta repels polar solvent moleculesmore strongly than the less polar ones and interferes with tightco-ordination of the solvent. Supporting this, the dielectricconstant ( E ) ~ showed good correspondence with the AHppvalues in the Fe"'(Hedta) system (see Table). This is in contrastto the Fe"'(hedta) system whose non-co-ordinating group isless polar and which did not show such a correspondence.On the basis that AHpp of Fe"'(hedta) in a methanol solutionwas 2.5 times larger than that of Fe*"(Hedta) as well as A,.forthe former being larger than for the latter, we suspect that theFe'"(hedta) complex is a mixture of six- and seven-co-ordinatecomplexes in methanol, different from Garbett's proposal of asingle species with the seven-co-ordination.In conclusion, it can be stated that as a result of the solventco-ordination to Fe"'(Hedta) or its derivative complexes in arhombic high-spin state, where the solvents have slightlydifferent co-ordinating ability, an apparent peak-to-peaklinewidth of the e.s.r. spectra reflects the ligand-field symmetryvery sensitively, whereas the &d transition band only reflects adifference in the ligand-field strength due to different co-ordination numbers.References1 H. Schugar, C. Walling, R. B. Jones, and H. B. Gray, J. Am. Chem.2 G. Lang, R. Aasa, K. Garbett, and R. J. P. Williams, J. Chem. Phys.,3 K. Garbett, G. Lang, and R. J. P. Williams, J. Chem. SOC. A, 1971,3433.4 W. E. Blumberg, 'Magnetic Resonance in Biological Systems,'5 R. Aasa, J. Chem. Phys., 1970,52,3919.6 G. Lang, R. Aasa, K. Garbett, and R. J. P. Williams, J. Chem. Phys.,7 H. H. Wickman, M. P. Klein, and D. A. Shirley, J. Chem. Phys., 1965,8 K. Dimroth and C. Reichardt, Ann. Chim. (Paris), 1963,611,l.9 J. G. Kirkwood, J. Chem. Phys., 1934,2,351.Soc., 1967,89,3712.1965,42,2113.Pergamon Press, Oxford, 1964, p. 119 et seq.1971,554539.42, 2113.Received 2nd June 1983; Paper 3189

ISSN:1477-9226    

DOI:10.1039/DT9850001077

出版商:RSC    

年代:1985

数据来源: RSC