1978 577A Microcalorimetric Study of the Macrocyclic Effect. Enthalpies ofFormation of Copper@) and Zinc(ii) Complexes with Some Tetra-azaMacrocyclic Ligands in Aqueous SolutionBy Andrea Anichini, Luigi Fabbrizzi, and Piero Paoletti,” lstituto di Chimica Generale ed Inorganica, UniversitBdi Firenze, and Laboratorio C.N.R., Via J. Nardi 39, 50132 Firenze, ItalyRobert M. Clay, Department of Chemistry, University of Stirling, Stirling FK9 4LAEnthalpies of formation of copper( 11) and zinc(ii) complexes with three tetra-aza macrocyclic ligands of differingring size have been determined directly by rnicrocalorimetry. The values for the copper( 11) complexes, which reach amaximum with the ligand 1.4.8.1 1 -tetra-azacyclotetradecane, have been correlated with the calculated ideal size ofthe aperture in the macrocyclic ligand required to maintain the ligand in a minimum-strain conformation.Themacrocyclic effect has been interpreted as being due to a favourable entropy term and to a normally favo_urableenthalpy term, the magnitude of which is critically dependent on matching the size of the metal ion to that of theaperture in the macrocyclic iigand.THE current interest in metal complexes containingsimple cyclic nitrogen donor ligands can be partlyattributed to the fact that these complexes can beregarded as simple models for naturally occurring metal-macrocycle centres found in proteins.Cyclisation of a linear polyamine ligand producesstriking changes in the properties of the metal com-plexes, when compared to the equivalent complexes ofthe linear ligands.Typical properties of macrocyclicD. K. Csbbiness and D. W. Margerum, J. Amer. Chem. SOC.,1970, 92, 2151.D. H. Busch, K. Farmery, V. Goedken, V. Katovic, A. C.Melnyk, C. R. Sperati, and N. Tokel, Adv. Chem. Ser., 1971, 100,44.complexes include: (a) a marked kinetic inertness bothtowards the formation of the complex from an aqueousmetal ion and ligand and towards decomposition; 1(6) strong metal-nitrogen interactions which are re-flected in the spectrochemical parameters (Dq) ; (c) thepossibility of stabilisation of high oxidation states forthe metal ion; and (d) high ’thermodynamic stability,reflected in the stability constants which can be severalorders of magnitude larger than for the correspondingcomplexes of linear ligand~.~D.C: Olson and J. Vasilevskis, Inorg. Chem., 1969, 8, 1611;1971, 10, 463, 1228. .I D. K. Cabbiness and D. W. Margerum, J. Amer. Chem. SOG.,1969,91, 6540578 J.C.S. DaltonThe purpose of the present work is to investigatefurther the origin of the high thermodynamic stability,which has hitherto remained controversial, to seewhether the role of particular metal ions in some proteinscan be explained in thermodynamic terms.In order to emphasize the greater stability of macro-cyclic complexes over similar complexes of non-cyclicligands, the term ' macrocyclic effect ' was put forwardin 1969 to account for the much higher stability constantfound for the copper(r1) complex of uneso-5,5,7,12,12,14-hexamethyl-l,4,8,1l-tetra-azacyclotetradecane, com-pared to that with the similar non-cyclic ligand 3,7-diazanonane-l,g-diamine in which the chelate rings havethe same sequence.Since then, the enhanced stabilityof other macrocyclic complexes, of different ligands anddifferent metal ions, has been well established.The macrocyclic effect is a Gibbs energy term, referring[ML]"+ + L' --j. [ML']"+ + Id (1)(non-cyclic) (macrocyclic) (macrocyclic) (non-cyclic)to the metathetical reaction ( l ) , and the enhancedstability of the macrocyclic complex may have itsorigins in both enthalpy and entropy terms, The firstattempts to separate the effect into these componentsled to opposing conclusions. Using the temperaturedependence of equilibrium data for the system NiII-L3,Hinz and Margerum concluded that the macrocycliceffect was due entirely to an enthalpy term, but Kodamaand Kimura reached the opposite conclusion for CUII-L1, that an entropy term was solely responsible.Then n n nL' L2 L3 LL *L' L6 L7 L8method employed by these workers for determining AHe,from the temperature dependence of equilibrium data, is,at best, only approximate, since a very small error in theequilibrium constants can result in a large error in AHe,particularly if a limited number of constants are usedover a narrow temperature range. The direct calor-imetric determination of the enthalpies of formation ofmacrocyclic complexes had previously appeared impos-sible for several reasons. Both formation and destruc-tion (with H+ or [CNI-) of macrocyclic complexes areextremely slow and could not be followed in conventionalF.P. Hinz and D. W. Margerum, Inovg. Chem., 1974, 13,M. Kodaina and E. Kimura, J.C.S. Dalton, 1976, 116.P. Monk and I. Wadso, Acta Chew. Scand., 1968, 22, 1842.I . Wadso, Acta Chem. Scand., 1968, 22, 927.2941.calorimeters ; the low solubility of many macrocyclicligands in water would have necessitated the use ofextremely dilute solutions, making the measurement ofthe thermal output for a given reaction impossible byisoperibolic calorimetry; the strong proton affinityof most tetra-aza macrocyclic ligands (pK, and pK, FZ11-12) meant that, in aqueous solutions of the ligand,significant amounts of protonated species would havebeen present.In this work, we have overcome these experimentaldifficulties by devising an experimental procedure whichensures the rapid incorporation of the metal ion into themacrocyclic ligand.All the reactions were carried outat pH 14 in 1 mol dmP3 Na[OH], the high pH also elimin-ating the problem of the protonated ligand species. Thesmall heat output was measured with extremely sensitivemicrocalorimeters. Two different experimental proce-dures, one based on flow microcalorimetry7 and theother on batch microcalorimetry,8 were used for the firstsystem studied and gave the same enthalpy valueswithin experimental error. We report directly measuredenthalpies for complexes of both CuT1 and ZnII with theseries of macrocyclic ligands L1, L3, and L4 of differentring size, enabling the enthalpic contribution to themacrocyclic effect to be assessed. The preliminaryresults of this work have been reported previou~ly.~~ loEXPERIMENTALReagents.-Compound L3 was obtained commercially(Strem Chemicals) and purified by double recrystallisationfrom acetonitrile, and .Id4 from the same source was purifiedas the tetrahydrochloride by bubbling HC1 through anethanolic solution of the ligand and recrystallising fromaqueous ethanol.Pure L1*4HCl was a gift from ProfessorKimura (Hiroshima University, School of Medicine) andwas used without further purification. The ligands L5and L6 (Eastman-Kodak) were purified as the tetrahydro-chlorides in the same manner as for L4. Elemental analyses(yo) for these ligands are given below.Found Calc.r-JCC H TL5.4HC1 24.8 7.3 19.8 24.7 7.6 19.2L6.4HC1 28.0 7.5 18.1 27.7 7.3 18.4L1*4HCl 29.0 8.1 17.7 30.2 7.6 17.6L3 60.1 12.7 28.1 60.0 12 1 28.0L4*4HCl 36.6 8.8 15.0 36.7 8.4 15.6------>C H NStandard aqueous solutions of CuT* and ZnII were analysedby conventional gravimetric methods and were used toprepare solutions of the appropriate concentrations in 1 moldm-3 Na[OH].Solutions of the metal complexes wereprepared from the solid complex perchlorates [ML] [C104],(M = Cull or ZnII, L = L3 or L*) obtained by mixingequimolar amounts of M[C10,],*6H20 and ligand in butanol,and recrystallising the product from ethanol. For L1,because of the limited quantity of ligand available, solutionsof the complexes were prepared by mixing equimolaramounts of metal perchlorate and ligand hydrochloride andsubsequently deprotonating with standard Na[OH] .TheA. Anichini, L. Fabbrizzi, P. Paoletti, and R. M. Clay, J.C.S.Chem. Comm., 1977, 244.lo A. Anichini, L. Fabbrizzi, P. Paoletti, and R. M. Clay,Inorg. Chim. Acta Letters, 1977, 22, L261978 5791 mol dm-3 Na[OH] solution was prepared by adding Na[OH](Erba RP) to carbon dioxide-free twice-distilled water.Afifiaratus.-For the calorimetric measurements, LKBBATCH (10 700-2) and FLOW (10 700-1) microcalori-meters were used. The thermal output of both instrumentswas amplified by a Keithley 150B microvolt amplifier andFor theBATCH measurements the heat output was determined bycomparison of the area under the heat curve with a similararea produced by electrical calibration.The areas weredetermined as the mean from six measurements made with anchange, AHle, for reaction (1) between the aqueous metalion and the aqueous ligand the following thermochemicalcycle was employed :,,,2 +(aq) $. L(aq) [ML]2'(aq)recorded as a trace on a potentiometric recorder.1 2 1 ) L IM2+(,q) .+ L ( 1 mol dm-3 Na[OH]) - 13) [ML12+(lmol dm-3 Na[OH] 1CYCLE 1accurate planimeter. The reaction cells were glass, and inorder to prevent the reactants from ' creeping ' over thepartition wall and causing prereaction the inside of thereaction cell was coated with a thin layer of paraffin wax.Solutions were introduced into the cells by weight by meansof hypodermic syringes and the total volume of the reactantswas between 3.0 and 4.0 cm3.The reference cell was emptythroughout all the measurements.For the FLOW instrument, the amplified voltage, whichis proportional to the heat flow, was recorded numericallyon a printer a t predetermined time intervals and the valuea t each steady state was calculated as the mean of manyprinted values. The reactants were passed through theflow cell by means of two high-precision automaticburettes (METTLER DV-10) equipped with 50 cm3 cylin-ders. The flow rate of each burette was regulated by aresistance box and could be accurately varied from 3 x10P-8 x dm3 s-l. Once the reactant solutions hadbeen introduced into the burettes and the calorimetricsystem was stable, a series of experiments was performedby varying the flow rates of each reactant but keeping thesum constant.Both calorimeters were checked for absolute accuracy bydetermining the enthalpy of formation of water l1 (FLOW,AH0 = -13.4; BATCH, AHe = -13.2 kcal mol-l).*Visible spectra of the complexes were recorded on a BecltnianDK2 spectrophotometer equipped with 1-cm quartz cells.Calorimetric Procedure .-The tetra-aza macrocyclic ligandsare strong bases, and an aqueous solution of the ligandcontains mainly the diprotonated species [H,L] ,+.Thereaction between a metal ion and a macrocyclic ligand inaqueous solution is very slow, presumably because of theshort-range electrostatic repulsions.12 However, by de-protonating the ligand in a strongly basic solution, theformation reaction becomes sufficiently fast to enable theenthalpy to be determined calorimetrically. Thus all thecalorimetric measurements were made with solutions of theligand in 1 mol dm-3 Na[OH] a t pH 14.This has the addedadvantage that no corrections need be made for protonatedligand species. (pK, and'pK, values for the ligands are asfollows: L1, 10.7, 9.7; L3, 11.59, 10.62; l3 L4, 11.08,10.38.13) The key calorimetric measurement was thereforethe direct formation of the macrocyclic complex in aqueoussolution a t pH 14, and two separate procedures wereemployed.BATCH technique. In order to obtain the enthalpy* Throughout this paper: 1 cal = 4.184 J.11 J . D. Hall, R.M. Izatt, and J. J. Christensen, J . Phys. Chem.,1963,67, 2605; C. E. Vanderzeeand J. A. Swanson, ibid., p. 2608.l2 C. Lin, D. B. Rorabacher, G. C. Cayley, andD. W. Margerum,Inorg. Chern., 1975, 14, 919.l3 M. Micheloni and A. Vacca, unpublished work.l4 L. Sacconi, P. Paoletti, and M. Ciampolini, J . Chem. SOC.,1961, 5115.AH3e was obtained by mixing an aqueous solution of MCl,(0.4 cm3, 0.05 mol ~ l m - ~ ) with an alkaline solution of theligand (3.0 cm3, 0.01 mol dmP3) using a ca. 50% excess ofligand. A comparatively small volume of metal solutionwas used in order to minimise the endothermic enthalpy ofdilution of the Na[OH]. The last quantity was determinedseparately (0.4 cm3 of H,O + 3.0 cm3 of mol dm-3 1 Na[OH])and used to correct the heat output, giving AH3e.Thiscorrection is ca. 7% of the heat output from the reaction.AH4* was determined by adding an aqueous solution of themetal complex (0.4 cm3) to 1 niol dm-3 Na[OH] (3.0 cm3)and correcting for dilution as above. AHZe cannot bedetermined due to the impossibility of forming an aqueoussolution containing solely the pure unprotonated species L.It has been taken as zero throughout, and thus all theenthalpies of formation are calculated with reference to thepure ligand i n aqueous solution a t pH 14. The assumptionthat AH,O is zero was tested by determining the enthalpy offormation, AHe, of [CuL5l2+ by the above cycle (-21.4kcal mol-l) and comparing it with the value previouslydetermined by isoyeribolic calorimetry ( - 2 1.6 kcal mol-l) ,14the small difference being within the limits of experimentalerror.AHIe was determined from Cycle 2.FLOW technique.M ?* (aq) t L(aq1M2+(1 mol dm-3 Na[OH]) -t L ( 1 mol d r n - ~ N a [ O H ] ) ~ [ M ~ l 2 ~ i l , o l dm'NabH1)CYCLE 2The key measurement, that of was made in the flowcalorimeter by passing solutions of the metal ion and theligand in 1 mol dmP3 Na[OH] simultaneously through themixing cell. For determinations with CuII, the solutionswere mol dm-3, the limiting factor being the solubilityof copper(I1) hydroxide in Na[OH]. The solubility ofzinc(I1) hydroxide is much higher, and for the zinc(I1)determinations lo-, mol dm-3 metal and ligand solutionswere used.The flow rates of the two burettes were changed in such away that the total flow rate remained constant, while theL : M ratio was varied from 0.6 to 1.7 : 1.For L3 and L4AHGe was independent of the I, : M ratio; however, for L1,both with CulI and ZnII, although AH6e was independentof this ratio below L : M = 1 : 1 it increased rapidly abovethis value, becoming constant again when L : M > 1.5 : 1.This suggests that a further polymeric species is formedwith the smallest macrocycle ligand, possibly [M,L1,I4+.Therefore., for L1, AHee was determined for a mean of steadystates obtained with L : M < 1 : 1 and based on the ligandas the limiting reagent. For the other two ligands, AHs580 J.C.S. Daltonwas determined with L : M > 1 : 1 and based on the metal Cycle 1. The consistency of the two methods can beion as the limiting reagent.seen in the excellent agreement between AHle obtainedAHZe and AH4e have already been discussed in connection via Cycle 1 and Cycle 2 for the ~3 complexes of bothwith the BATCH technique. In order to make the measure- cU11 and Z ~ I I . The accuracy is demonstrated by theand t o remove the uncertainty arising from the nature of the from both cycles (-21.4 and -21.5 kcal mol-l), and the species present in strongly alkaline solutions of CuIl andznn, A H ~ U was determined as follows. For ZnII i t was previously determined literature value (-21.6 kcalmeasured directly by adding a solution of ZnC1, (0.4 cni3, mo1-1)-140.05 mol dm-3) to Na[OH] (3.0 cm3, 1 mol dm-3) in the Previously determined enthalpy values for theBATCH Calorimeter, and niaking an appropriate correction systems CuIr-L1 and -L4 have been reported by Kodamafor dilution.For CuII, due to the limited solubility, this and Kimura,6716 who made polarographic determinationsmerits with those from the BATCH technique agreement between our AH,e values for CuII-LS, obtainedTABLE 1Experimental results (kcal mol-l) at 25 "CAHler L7 Otherliteraturevalues System AH3e A H 4 e AH,e AH,e Cycle 1 Cycle 2CUII--L5 -21.4 f 0.3 0.0 -& 0.1 -10.9 f 0.1 -10.6 f 0.2 -21.4 f 0.4 -21.5 f 0.3 -21.6 l4CUII-L6 - 10.9 0.1 a - 16.8 f 0.2 -27.7 l5CUII-Ll 0.0 & 0.1 -10.9 & 0.1 a -11.8 & 0.2 - 22.7 0.3 - 18.3CUII--L3 -32.2 f 0.3 0.0 f 0.1 -10.9 0.1 -21.6 0.2 -32.2 f 0.4 -32.5 & 0.3CU'I--L4 0.0 -+ 0.1 -10.9 f 0.1 a -15.6 IJI- 0.2 -26.5 f 0.3 -26.5 lG1.2 f 0.1 -7.7 & 0 .2 b -8.2 & 0.1 -14.5 0.3Z ~ I I - L ~ -15.6 & 0.2 0.9 & 0.1 -7.7 f 0.2 -8.3 f 0.1 -14.7 f 0.3 -14.9 f 0.30.3 f 0.1 -7.7 f 0.2 -9.3 f 0.1 -16.5 f 0.3ZnlI-Ll~ ~ 1 1 4 4a Ca~cukdted by inserting t h e value l5 of AHIQ for Cu1I--LG into Cycle 2. Measured directly on the BATCH instrument.was not possible and the system Cu11-L6 was studied bymeans of Cycle 2, and AHle, which had been determinedpreviously, l5 was inserted thereby indirectly giving AHSe.This procedure is justified by the good agreement betweenvalues for AHIe obtained for the CuI1-L3 system by Cycles1 and 2, since the value obtained zria Cycle 1 does notdepend onFLOW versus BA TCH Techniq.ues.-The FLOW techniquehas the decided advantage over the RATCH in that, oncethe correct experimental conditions have been found, a seriesof measurements can be made in a single session, simplyby varying the flow rates.In comparison, the proceduresinvolved with the BATCH technique are extremely labori-ous, requiring the complete emptying, cleaning, and refillingof the cell, and the consequent delay in waiting for thermalequilibrium t o be attainecl before the next measurementcan be made. On the other hand, it is not possible to mixsolutions of significantly differing densities in the FI,OWcalorimeter, and the BATCH instrument must be used fordetermining AH,u and The two instruments aretherefore essentially complementary.We have usedCycle 1 (based entirely on BATCH measurements) for thesystems M-L3 (M = CuII or ZnT1) to confirm the reliabilityof the experimentally easier Cycle 2 (in which the keymeasurement is made on the FLOW instrument). There-after we used Cycle 2 to determine the enthalpies of forma-tion.RESULTS AND DISCUSSIONThe experimental results are given in Table 1. Datafor AH3e, AH4u, and (for ZnII) from the BATCHcalorimeter are given as the mean from 2-4 separatemeasurements, and for AHGe from the FLOW calorimeteras the means of at least five steady-state values. AHS0(-10.9 kcal mol-l) has been obtained by inserting thepreviously determined l5 value of AHle for Cu'1-L6 intoof the stability constants at different temperaturesbetween 10 and 35 "C.Our results agree extremely wellfor L4 but differ by 4.4 kcal mot1 for L1. For thesystem CuIr--L1 several criticisms of the polarographicresults can be made. First the previous workersreported only the formation of a 1 : 1 complex. Secondly,they did not correct their calculations of the stabilityconstants at different temperatures for the temperaturedependence of pK, and pK,. Thirdly, the reversibilityof the electrode process is not satisfactorily establishedsince the log plot gives a gradient (-36 to -40 mV)considerably different from that required by a reversibletwo-electron process (-29.6 mV). For L4 these critic-isms are not as serious. Only a 1 : 1 complex is formed,the pKl and pK, values were corrected for temperaturedependence, and the gradient of the log plot (-32 mV)is much closer to the theoretical value.We thereforebelieve that our calorimetric value for CuII-Ll isthe more reliable and we have recalculated the entropychange on complex formation using the log KML valueobtained by Kodama and Kimura l6 and our own AH*value. VC7hile this log KMI, value must still remaindoubtful, the AS* value of 36 cal K-l mol-l seems muchmore reasonable than the 51 cal K-l mol-l originallyreported, when compared to the other tetra-aza macro-cyclic ligands, and the ASe values now diminish steadilywith increasing size and decreasing rigidity of the macro-cyclic ligands (Figure 1). This suggests that the greatererror is manifest in AH*, and emphasizes the unreli-ability of determining AHe by this method, particularly ifthe stability constants themselves are doubtful.15 L.Fabbrizzi, R. Barbucci, and P. Paoletti, J . C . S . Dalton,1972, 1529.16 M. Kodama and E. Kimura, J . C . S . Dalton, 1976, 23411978 58 1The best available thermodynamic parameters forcomplexes of CuII and ZnII with macrocyclic and asso-ciated non-cyclic ligands are given in Table 2, togetherI I ICyclic L’- L2 L3 LLL6 L7 LB Non-cyctic L~FIGURE 1 Enthalpies (----) and entropies (- - - -) of formationfor cyclic ( 0 ) and non-cyclic (m) polyamine copper complexeswith their origins and the estimated uncertainties in theAH* values.Co$$er(Ir) Cowz$Zexes.-Of the ligands studied, L3reacts most exothermically with CuT1.I t has previouslybeen demonstrated l7 that the ideal M-N distance forTABLE 2Thermodynamic parameters for macrocyclic and non-cyclictetra-aza ligand complexes of CuII and ZnIICUIII h ZnlIAHe/ AS$ (AHe/Ligand kcal mol-l log KMI, cal K1 mol-l kcal mol-l)L1 -22.7 a 24.8 36.2 - 14.5 aL2 -29.2 29.1 33.7L3 - 32.4 - 14.8f 0 . 3 4 0 . 3h 0 . 3 a,d f 0 . 3f 0 . 3 f 0.3L4 - 26.5 a 24.4 +22.7 - 16.5 ’L5 -21.6f 20.2 + 19.5 - 8.9L7 -25.9f 21.8 + 13.1 - 10.6L6 - 27.7 f 23.9 + 16.5 -11.9L8 - 19.5 f 17.3 + 12.8 - 7.4a This work. Ref. 6 ; calculated using AHe from this work.c Ref. 23. Mean of Cycle 1 and Cycle 2 determinations.Ref. 16. f Ref. 19 and reference therein.this ligand to adopt its strain-free conformation with allthe four nitrogens coplanar is 2.07 A.The Cu-N bonddistance found in tetragonal polyamine complexesranges from 2.03 to 2.10 A,lS and the Cu2+ ion is thereforel7 L. Y. Martin, L. J. Dehayes, L. J. Zompa, and D. H. Busch,J . Amer. Chem. Soc., 1974, 96, 4046.l8 Y . Komiyama and E. C. Ljngafelter, Acta C~yst., 1964, 17,1145; A. Pajunen, Suomen Kem., 1969, B42, 5.l9 P. Paoletti, L. Fabbrizzi, and R. .Barbucci, Inovg. Chim.Acta Rev., 1973, 7, 43.tailor-made to form a strain-free complex with L3.This is reflected in the large AHe value. Furthermore,the complex [CuL3I2+ contains a system of fused alternat-ing 5,6,5,6 membered chelate rings, an arrangementwhich has previously been found to favour the greatestexothermicity in complexes of non-cyclic tetra- and tri-amines.lg ’For L4 the Cu-N distance is now smaller than theideal l7 (M-N 2.22 A) required to maintain the minimumstrain energy in planar co-ordination, and less strongmetal-donor interactions will result, which are reflectedin the lower AHe.Similarly, L1 is too small (ideal M-N1.83 A) to accommodate the metal ion in square co-ordination, and in this case, where the mismatch islargest, a distortion from square co-ordination may result.This could take the form of ligand folding to form a cis-octahedral structure, rings folding to form a trigonal-bipyramidal structure, or the removal of the metal ionfrom the nitrogen plane to give a square-pyramidalstructure. Examples of all these distortions are knownwith other metal ions and macrocyclic ligands.20-22 Inany case the strong in-plane Cu-N interaction will bereduced resulting in a lower AHe, as confirmed by ourcalorimetric results.PA similar situation exists for L2,for which AH0 has been determined by Kodama and22 l I - 2 - L 1.80 1.90 2.00I I- 20L- 190 2,49 I m.--18i3 f-172-10 2-20 2.30C ~ - N 1 aF~GUI~E 2 Enthalpies of formation ( 0 ) and energies of thed-d band (m) plotted against the Cu-N distance calculated forminimum strain energy, in an ideal square co-ordination foraqueous macrocyclic complexes. The hatched area indicatesthe normal range of Cu-N distances in polyamine complexesK i m ~ r a , ~ ~ but the distortion would not be expected to beso severe as for L1 and AH@ will consequently be higher.The experimental AHe values are plotted in Figure 22o Y.Iitaka, M. Shina, and E. Kimura, Inorg. Chenz., 1974, 13,2886.21 N. F. Curtis, D. A. Swann, and T. N. Waters, J.C.S. Dalton,1973, 408.22 M. J. D’Aniello, M. T. Mocella, F. Wagner, E. K. Barefield,and I. C. Paul, J . Amev. Chem. Soc., 1975, 97, 192.23 M. Kodama and E. Kimura, J.C.S. Chem. Comm., 1975, 891582 J.C.S. Daltonagainst the M-N ideal distance 17 for the ligand tominimise its strain energy in planar co-ordination.The maximum in the enthalpies is also paralleled bythe energies of the d-d bands in the visible spectra (Figure2). Previously, a linear relation had been found be-tween AHe and v(d-d) for over 30 copper(I1) complexes ofnon-cyclic amines 24 and, interestingly, of the three macro-cyclic complexes studied only that of L3 fits this relation.It therefore appears that the adherance to this relationmay be a sensitive test for complexes containing a CuN,planar geometry, with Cu-N bond distances in thenormal range (2.03--2.10 A).The other significantexceptions to this relation were tripod-like ligands suchas tris(2-aminoethy1)amine in which the four nitrogendonors cannot co-ordinate in a plane. We can surmiseat this point that significant distortions from squareplanarity will occur in [CuL]2+ (L = L1 or L2), that[CuL3]2+ is almost certainly planar, and [CuL4I2+ isprobably planar, with a lengthened Cu-N bond andconsequently reduced AHe and v(d-d) values. Anydistortion or bond lengthening has a more significanteffect in v(d--d) than on AHe, since for the complexeswith both L1 and L4 the experimental AHe is well inexcess of the value calculated from the linear relation.Zinc(r1) Com$Zexes.-The enthalpies do not show apeak like the copper(r1) system, but decrease slowlyalong the series L4, L3, L1.The Zn2+ ion is larger thanCu2+ and this will favour the larger macrocyclic ligand.On the other hand zinc is known to have a preference fortetrahedral co-ordination of polyamines and thisstereochemistry can be most easily achieved by thelargest macrocycle. However, with no reliable structuralevidence it would be dangerous to draw firm conclusions.The Macrocyclic EfecL-The origin of the term' macrocyclic effect ' is described in the Introduction.Both enthalpy and entropy contributions are possibleto the added stability of the macrocyclic complex and itis the relative importance of these two contributionswhich has for some time been contr~versial.~?~ Unfor-tunately, the choice of the appropriate non -cyclic refer-ence ligand is not easy.For copper(rr) complexes theenthalpies of the non-cyclic tetramine ligands also gothrough a maximum while the entropy changes oncomplexation decrease slowly with the size of the ligand(Figure 1). The highest AHe for a non-cyclic complexoccurs with an alternating 5,6,5 chelate-ring sequence,and the lower enthalpies of the other complexes are dueto internal strain energies associated with the introduc-tion of bridgeheads between either two five- or two six-membered chelate rings.In order to correctly assessthe macrocyclic enthalpy, AHe for the macrocycliccomplex should be compared to AHe for a complexcontaining the same degree of internal strain, the remain-ing difference between these two enthalpies beingassociated with ring closure. A proper estimate cantherefore only reastically be made for the complexeswith L1 and L P which can be compared to the complexes2p I,. Fabbrizzi, P. Paoletti, and A. €3. P. Lever, Inorg. Chem.,1976,15, 1502.with Id5 and L6 respectively, maintaining the same ringsequence between cyclic and non-cyclic ligands. Themacrocyclic enthalpies are given in Table 3.TABLE 3Macrocyclic enthalpiesAH*macrocyc~icRing sequence Reaction kcal mo1-1- 1.1[ZnL5I2+ + L1 -5.6[ZnL6I2+ + Id3 - 2.9[CUL5]2+ + L' All five-memberedAlternating five and six [CUL6]2+ + L3 -4.7memberedFor these two systems accurate entropy data areunfortunately not available for any of the cyclic com-pounds.However, an estimate of AS* for the complex[CuL1I2+ has already been made using Kodama andKimura's AGe value and our own value of AHe. ASefor [ C U L ~ ] ~ ~ can be estimated from Figure 1 as 29 calK-l mol-l probably to within *2 cal K-l mol-l, since theentropy changes on complex formation for the cyclicligands parallel their non-cyclic counterparts by decreas-ing slowly down the series. Thus ASe (macrocyclic) canbe estimated for the two metathetical copper systemswith L1 and Ira3 as 17 and 13 cal K-l mol-l respectively.While these values are very tentative, they reinforceour belief that there will always be a positive and there-fore favourable entropy contribution to the macrocycliceffect.This arises from the fact that the macrocyclicligand, before co-ordination, is already rigid and pre-orientated, unlike its non-cyclic counterpart, and willtherefore not lose configurational entropy after co-ordination to nearly the same extent. In fact, all thereported entropy values for macrocyclic complexes are inexcess of the equivalent non-cyclic complexes with theexception of some nickel(r1) complexes 5 including[NiL3I2+. It is extremely desirable to have a calori-metric value of AHe for this complex, since much of theoriginal thermodynamic interpretation of the macro-cyclic effect has been based on the original work of Hinzand Marger~m,~ who found that the macrocyclic effectwas solely due to an enthalpy term and that the macro-cyclic entropy change was negative (- 16 cal K-l mol-l).In the examples studied in this work it is clear that amacrocyclic entlialpy is present, the magnitude of whichdepends on both the matching of the metal-ion size tothe size of the aperture in the macrocyclic ligand in itsstrain-free conformation and on the stereochemicalpreference of the metal ion. We would thereforeexpect the macrocyclic enthalpy to vary considerablyL9for different systems. Because of the importance ofmatching the size of metal ion to the size of the ligandaperture, very large macrocyclic enthalpies can b1978expected when the match is good. However, in casesof extreme mismatch the macrocyclic enthalpy maywell become unfavourable, and, for example, we wouldentropy, however, should always remain positive andfavourable, its magnitude changing only slightly fromligand to ligand.predict, by extrapolating the enthalpy data in Figure 2 We thank Professor Kimura for the gift of the ligand L1, to theis 2'38 that for [cuL9i2+ Probably be less the award of a European Fellowship (to R. M, C.),with L99 for which the M-N distance the Italian C.N.R. for support, and the Royal Society forthan -20 kcal mol-l, or in other words less than thatfor any of the non-cyclic tetramines. The macrocyclic [7/1323 Received, 22nd July, 1977