Study of protonation of 1,4,7-tris(2-hydroxyethyl)-1,4,7-triazacyclononane, and its complexes with metal ions, by crystallography, polarography, potentiometry, molecular mechanics and NMR

The unusually high first protonation constant (pK(1)) of 1,4,7-tris(2-hydroxyethyl)-1,4,7-triazacyclononane (L) is analysed. A H-1 NMR study of the protonation of L in D2O confirms the high value of pK(1) reported previously (C.M. Madeyski, J.P. Michael and R.D. Hancock, Inorg. Chem., 23 (1984) 1487). The crystal structure of L . HBr is reported: monoclinic, space group P2(1)/n, a = 8.511(5), b = 13.301(5), c = 13.604(4) Angstrom, beta = 90.73(4)degrees, Z = 4, R = 0.0548. The structure shows that the proton, attached to one nitrogen of the macrocyclic ring, is hydrogen bonded to the other two nitrogens of the ring, plus an oxygen of one pendant 2-hydroxyethyl donor group. Molecular mechanics (MM) calculations suggest that this hydrogen bonding is probably not a major cause of the high first protonation constant of L. Rather, the MM analysis suggests that pK(1) for [9]-aneN(3) (1,4,7-triazacyclononane) itself is low because in the free ligand one of the N-H hydrogens is strongly hydrogen bonded within the macrocyclic cavity. This means that when the ligand is protonated, the proton is added to the outside of [9]-aneN(3), and is not held within the macrocyclic cavity. Derivatives of [9]-aneN(3) which have three N-alkyl groups, such as L, or Me(3)-[9]-aneN(3) (N,N',N ''-trimethyl-1,4,7-triazacyclononane) will all have high first protonation constants because they have no N-H hydrogens to occupy the macrocyclic cavity in the free ligand, so when the proton is added it is added within the macrocyclic cavity. MM calculations on free and monoprotonated forms of [9]-aneN(3) and Me(3)-[9]-aneN(3) are used to support this suggestion. The synthesis and structure of the quaternary salt, N,N,N',N ''-tetrakis(2-hydroxyethyl)-[9]aneN(3) bromide, is reported: monoclinic, space group, P2(1)/c, a = 20.036(2), b = 19.075(5), c = 9.470(3) (Angstrom), beta = 101.52(2)degrees, Z = 8, R = 0.0725. The structure shows that the fact that the 2-hydroxyethyl groups of the salt cannot occupy the macrocyclic cavity means that the ring is forced to adopt an irregular structure unlike that adopted by all known [9]aneN(3) rings. The structure of the Fe(III) complex of L, [Fe-2(L)(2)H-3](ClO4)(3), is reported: orthorhombic, space group Pnnm, a = 13.566(3), b = 14.981(5), c = 9.517(4) Angstrom, Z = 4, R = 0.0855. The structure consists of hydrogen-bonded dimers of Fe(III)LH(1.5) complex units, which between them have lost three protons from the oxygens of the alcoholic oxygens of L. The pairs of partly deprotonated Fe(III)LH(-1.5) units are held together face-to-face, by three short O-H-O hydrogen bonds between the faces formed by the three alcoholic oxygens of each complex cation. The hydrogen-bonding O-O distances are short at 2.42 Angstrom. The coordination geometry around the Fe(III) is distorted to close to trigonal prismatic rather than octahedral. The three oxygens donors are twisted, relative to the three nitrogens, around the Cg axis Of the complex, by 45 degrees, relative to the positions expected for regular octahedral coordination. The role of d(pi)-p(pi) pi-bonding in shortening Fe-O bonds, by up to 0.23 Angstrom in the examples considered, is discussed. It is shown that the trans effect that the shortened Fe-O bonds exert on the Fe-N bonds leads to an inverse relationship between Fe-O and Fe-N bond length in complexes of Fe(III) with N-3-O-3 donor sets. Near normal Fe-O and Fe-N bond lengths are found for ligands where N is sp(2) hybridized and capable of pi-bonding in competition with the oxygen donors. The formation constants of L with Cd(II), Bi(III), and Pb(II) are reported, determined by both polarography and glass electrode potentiometry. For these large metal ions, log K-1 for L is higher than for [9]aneN(3), the analogue without 2-hydroxyethyl donors. It had previously been shown for the small Cu(II) and Zn(II) ions that log K-1 for L is not significantly higher than for [9]aneN(3) (C.M. Madeyski, J.P. Michael and R.D. Hancock, Inorg. Chem., 23 (1984) 1487). This is as expected from the effect that neutral oxygen donors have on increasing log K-1 for larger metal ions relative to smaller ions (R.D. Hancock, in A.P. Williams, C. Floriani and A.E. Merbach (eds.), Perspectives in Coordination Chemistry, VCH, Weinheim, 1992, p. 129).