Electronic and magnetic metal-metal interactions in dinuclear oxomolybdenum(V) complexes across bis-phenolate bridging ligands with different spacers between the phenolate termini: Ligand-centred vs. metal-centred redox activity

View Article Online / Journal Homepage / Table of Contents for this issue

Electronic and magnetic metal–metal interactions in dinuclear

oxomolybdenum(V) complexes across bis-phenolate bridging

ligands with diﬀerent spacers between the phenolate termini:

ligand-centred vs. metal-centred redox activity

Simon R. Bayly,^aElizabeth R. Humphrey,^aHelena de Chair,^aCecilia G. Paredes,^aZoe R. Bell,^a

John C. Jeﬀery,^aJon A. McCleverty,*^aMichael D. Ward,*^aFederico Totti,^bDante Gatteschi,^b

Stephane Courric,^cBarry R. Steele^cand Constantinos G. Screttas^c

^aSchool of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK BS8 1TS.

E-mail: mike.ward@bristol.ac.uk

^bDepartment of Chemistry, University of Florence, Via Maragliano 75/77, 50144 Florence,

Italy

^cInstitute of Organic and Pharmaceutical Chemistry, National Hellenic Research Foundation,

48 Vas. Constantinou Ave., 116 35 Athens, Greece

Received 18th January 2001, Accepted 22nd March 2001

First published as an Advance Article on the web 17th April 2001

A series of dinuclear complexes has been prepared in which two {Mo^V(Tp^Me,Me)(O)Cl} fragments (abbreviated

as Mo; Tp^Me,Me= tris(3,5-dimethylpyrazol-1-yl)hydroborate) are attached to either end of a bis-p-phenolate

bridging ligand [(4,4Ј-OC₆H₄)–X–(4,4Ј-C₆H₄O)]^2Ϫ. The complexes are Mo (C᎐C) (X = CH᎐CH), Mo (C᎐C)

᎐

2

(X = CH᎐CH–CH᎐CH), Mo (C᎐C) (X = CH᎐CH–CH᎐CH–CH᎐CH), Mo (th) (X = 2,5-thiophenediyl),

᎐

2

3

2

᎐

Mo (th) (X = 2,5:2Ј,5Ј-bithiophenediyl), Mo (th) (X = 2,5:2Ј,5Ј:2Љ,5Љ-terthiophenediyl), Mo (C᎐C)

(X = C᎐C), Mo (N᎐N) (X = N᎐N), Mo (CO) [X = C(O)] and Mo (C ꢀC ) [X = CH᎐CH(1,4-C H )CH᎐CH].

᎐

2

3

2

᎐

2

6

4

Electrochemical, UV/VIS/NIR spectroelectrochemical and magnetic measurements have been carried out in

order to see how eﬀectively the diﬀerent spacer groups X mediate electronic and magnetic interactions between

the two redox-active, paramagnetic, Mo centres. The electronic interactions were determined from the redox

separation between the two successive one-electron oxidations which are formally Mo(VI)–Mo(V) couples; it

was found that thienyl units in the bridging ligand are much more eﬀective at maintaining electronic communication

over long distances than p-phenylene or ethenyl spacers of comparable lengths. The azo (N᎐N) linkage aﬀorded

᎐

a much weaker electronic interaction than the ethenyl or ethynyl spacers. UV/VIS/NIR spectroelectrochemical

studies showed that whereas the ﬁrst oxidation is metal-centred to give Mo(VI)–Mo(V) species with characteristic

intense phenolate→Mo(VI) LMCT transitions in the near-IR region, the spectra of the doubly oxidised complexes

are characteristic of quinones: thus, the sequence of species formed on oxidation is [Mo(V)(µ-diolate)Mo(V)]⁰→

[Mo(V)(µ-diolate)Mo(VI)]^ϩ→ [Mo(V)(µ-quinone)Mo(V)]^2ϩ, with an internal charge redistribution associated

with the second oxidation. Semi-empirical ZINDO calculations provide some support for this. Magnetic

᎐

susceptibility measurements on Mo (C᎐C), Mo (th), Mo (N᎐N) and Mo (C᎐C) show that all are weakly

᎐

2

antiferromagnetically coupled, as expected on the basis of a spin-polarisation picture, with the order of strength

of the magnetic interaction being the reverse of the order for electronic coupling, such that Mo₂(th) aﬀords

the strongest electronic interaction but the weakest magnetic interaction.

recently been interested to see how these properties of the

bridging ligand also control the magnetic metal–metal

interaction.

Introduction

We have been interested recently in examining how magnetic

and electronic interactions between metal centres in poly-

nuclear complexes are controlled by the nature of the bridging

ligand which links them.^1–9Electronic interactions between

redox-active metal fragments, as manifested by a separation

between metal-centred redox couples and the consequent

formation of stable mixed-valence states, have been of interest

since the discovery of the Creutz–Taube ion; subsequently,

long-range electron-transfer in mixed-valence complexes has

been extensively studied because of its relevance to the prepar-

ation of ‘molecular wires’ which might permit charge transport

in nanoscopic circuits.^10–14The eﬀects of length, substitution

pattern (i.e. para vs. meta), and conformation of the bridging

ligand on the extent of the electronic metal–metal interactions

have all been studied in some detail. We^2,4–7and others¹⁵have

It has generally been the case that magnetic and electronic

interactions are treated quite separately, despite the fact that

both clearly depend on the nature of the bridging ligand

through which the interaction is transmitted (as long as the

metal centres are far enough apart that direct metal–metal

orbital overlap can be excluded). This reﬂects the fact that

the complexes most commonly used to probe electronic

interactions are based on kinetically inert, diamagnetic metal

fragments such as Ru(II),^10–14whereas studies on magnetic

exchange interactions commonly involve labile metals such as

ﬁrst-row transition-metal and lanthanide() ions in coordin-

ation environments where they do not show reversible redox

interconversions.^16–23In our recent work we have performed

combined studies of electronic and magnetic interactions in

DOI: 10.1039/b100681i

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

This journal is © The Royal Society of Chemistry 2001

1401

View Article Online

dinuclear complexes Mo₂(OO),^2,3,5,6,8using the {Mo^V(Tp^Me,Me)-

(O)Cl} fragment (ﬁrst described by Enemark and co-workers,²⁴

and abbreviated Mo; Tp^Me,Me= tris(3,5-dimethylpyrazol-1-yl)-

hydroborate) attached to either end of a bis-phenolate bridging

ligand (denoted OO). This metal fragment is paramagnetic (d¹

conﬁguration), which has allowed study of magnetic exchange

as a function of the bridging ligand, and we have shown that

the sign of the magnetic exchange interaction can be pre-

dicted according to the number of atoms in the bridging

pathway, according to a simple spin-polarisation picture.^1,2,4–7

The Mo fragment is also redox-active, with reversible Mo(V)–

Mo(IV) and Mo(VI)–Mo(V) couples, such that electronic

interactions can also be studied by measurements of redox

separations in cyclic voltammograms and by spectroscopic

studies on the Mo(V)–Mo(IV) and Mo(VI)–Mo(V) mixed-

valence states.^2,3,8Although magnetic exchange and electronic

delocalisation are quite diﬀerent phenomena, it has become

apparent that there are strong correlations between them

because of their shared dependence on the structure of the

bridging ligand.²

In addition to their electrochemical and magnetic properties,

the dinuclear complexes [{Mo^V(Tp^Me,Me)(O)Cl}₂(µ-OO)] have

also proved to be of interest for their optical spectroscopic

properties in the oxidised Mo(VI)–Mo(V) and Mo(VI)–Mo(VI)

forms.^3,25,26On oxidation of a molybdenum() centre to the

molybdenum() state, the phenolate→Mo(V) LMCT transi-

tion in the visible region is replaced by a much more intense

transition which, in many of the dinuclear complexes, lies in the

near-infrared region (900–1400 nm) and has been assigned

as phenolate→Mo(VI) LMCT.³Thus the complexes act as

electrochromic dyes which can be switched, by a simple

reversible redox process, between transparent and essentially

opaque states in a region of the spectrum which is of interest

for telecommunications and optical data processing. We have

accordingly been exploiting this property of some of these

complexes in optical materials-related applications.²⁶

In this paper we describe the syntheses, electrochemical,

spectroscopic and spectroelectrochemical, and magnetic

properties of further complexes of this type (see Scheme 1 for

the structures of the bridging ligands and their abbreviations):

speciﬁcally, a series of complexes based on para-substituted

bis-phenolate ligands with a variety of diﬀerent spacer groups

separating the two phenolate termini. The ligands H₂th, H₂(th)₂

and H₂(th)₃contain one, two or three 2,5-thienyl spacers

respectively. We were interested in these because oligothiophene

chains are known to be particularly eﬀective at mediating long-

distance electronic interactions because of their favourable

steric and electronic properties.^27–30They are more likely to be

coplanar than oligophenylene chains because the steric inter-

action between the H³/H⁴protons of adjacent thienyl rings is

less than that between the H²/H⁶protons of adjacent C₆H₄

units; this leads to more eﬀective delocalisation, which is fur-

ther enhanced by the participation of π-symmetry S orbitals

Scheme 1 Structural formulae of the bridging ligands used in this

paper and their protected precursors.

eﬀective as potential electrochromic dyes for use in the NIR

region; and (iii) to address fundamental questions of metal-

centred vs. ligand-centred redox activity in complexes where the

bridging ligand is non-innocent.

Results and discussion

Syntheses of ligands

The synthesis of H₂th has been reported by Takahashi et al.,

and involved addition of the aromatic groups to the thienyl core

in two separate metallation/cross-coupling stages. Pd-catalysed

coupling of 2-thienylzinc chloride with 4-iodoanisole aﬀorded

2-(4-methoxyphenyl)thiophene; remetallation of the thienyl

unit with zinc chloride, and a second cross-coupling with

4-iodoanisole, aﬀorded 2,5-bis(4-methoxyphenyl)thiophene

which was demethylated using BBr₃to give H₂th.³¹We have

developed a more convenient synthesis based on Pd-catalysed

coupling of both phenyl substituents simultaneously to 2,5-

bis-(tri-n-butylstannyl)thiophene, to give the protected 2,5-

bis(4-methoxyphenyl)thiophene which was demethylated using

molten pyridinium chloride. We have previously described

the preparation of H₂(th)₂from 5,5Љ-bis(tri-n-butylstannyl)-

2,2Ј-bithiophene using this method,²⁷and it is also eﬀective

for preparation of H₂(th)₃starting from 5,5Љ-bis(tri-n-butyl-

stannyl)-2,2Ј:5Ј,2Љ-terthiophene.

in the conjugated system. The ligands H (C᎐C), H (C᎐C) and

᎐

2

H (C᎐C) were likewise prepared to examine how electronic

᎐

2

3

interactions in this series of complexes are attenuated by double

bond spacers; H (C ꢀC ) is even longer than H (C᎐C) , having

᎐

2

3

(eﬀectively) four double bonds between the phenolate termini,

counting one phenylene spacer as comparable to two double

᎐

bonds in the bridging pathway. Ligands H (C᎐C) (alkynyl

᎐

2

linkage) and H₂(NN) (azo linkage) are two-atom conjugated

spacers to compare with H (C᎐C). In contrast to these, 4,4Ј-

᎐

2

dihydroxybenzophenone (H₂CO) contains only one additional

atom in the direct conjugated pathway between the two phenol

termini.

The aims of this work have been: (i) to compare and contrast

the abilities of these diﬀerent spacer units to mediate electronic

and magnetic interactions; (ii) to evaluate by spectroelectro-

chemical methods the spectroscopic properties of the com-

plexes in all accessible oxidation states, to see which are most

The ethenyl-bridged ligand H (C᎐C) was simply prepared by

᎐

2

a Heck coupling of 4-vinylanisole with 4-bromoanisole to give

1,2-bis(4-methoxyphenyl)stilbene, followed by demethylation

1402

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

View Article Online

Table 1 Analytical and mass spectroscopic data for the new complexes

Analysis (%)^b

Complex^a

Yield (%)

C

H

N

M^ϩby FAB MS m/z

Mo₂(th)

Mo₂(th)₂

Mo₂(th)₃

60

59

35

15

56

39

37

30

65

8

48.1 (47.8)

49.9 (50.7)

48.4 (48.6)

49.1 (49.3)

47.9 (47.8)

47.6 (47.6)

47.3 (47.3)

46.6 (47.2)

4.9 (4.7)

5.1 (5.3)

4.9 (4.8)

4.9 (5.3)

4.8 (4.9)

5.2 (4.9)

5.3 (5.4)

5.2 (5.2)

5.0 (4.6)

14.9 (14.6)

12.6 (12.7)

12.6 (12.2)^c

14.5 (14.7)^d

13.9 (14.4)^e

13.0 (13.6) ^f

13.5 (13.5)^g

16.6 (17.1)^d

13.8 (14.2) ^f

1155

1237

1319

1100

1127

1152

1098

1103

1102

1202

Mo₂(C᎐C)

᎐

Mo₂(C᎐C)₂

᎐

Mo₂(C᎐C)₃

᎐

Mo₂(C᎐C)

᎐

Mo₂(N᎐N)

᎐

Mo₂(CO)

44.5 (44.5)

h

Mo₂(C₂PhC₂)

^aAll complexes are dark green, apart from Mo₂(N᎐N) which is brown and Mo₂(CO) which is maroon. ^bExpected values in parentheses. ^c2 MeOH in

᎐

elemental analysis. ^d0.5 hexane in elemental analysis. ^e0.5 CH₂Cl₂in elemental analysis. ^f1 CH₂Cl₂in elemental analysis. ^g3 MeOH in elemental

analysis. ^hWe could not obtain reliable and repeatable analytical data for this complex.

with BBr₃; this is a considerable improvement on the published

procedure which is nearly 40 years old.³²The syntheses of

Syntheses of complexes; crystal structures of Mo₂(CO) and

Mo₂(NN)

H (C᎐C) and H (C᎐C) (as their protected precursors) are

᎐

2

3

All of the dinuclear complexes were prepared by the same

general route that we have described before, viz. reaction of

[Mo^V(Tp^Me,Me)(O)Cl₂] (>2 equivalents) with the deprotonated

(using Et₃N) bridging ligand in toluene at reﬂux. Puriﬁcation

by column chromatography with silica/CH₂Cl₂aﬀorded pure

products in all cases; salient characterisation data (FAB mass

spectra and elemental analyses) are collected in Table 1. Solu-

tion EPR spectra of all of the complexes conﬁrmed their

based on published methods,^33,34although the use of the

n-propyloxy methyl ether protecting group in the synthesis of

H (C᎐C) is new. The protected (MeO protecting groups)

᎐

2

3

precursor of H (C᎐C) was then demethylated by reaction with

᎐

2

MeMgBr, following the method of Takahashi et al.³⁵Attempts

to use more traditional deprotecting reagents such as BBr₃or

pyridine hydrochloride were in our hands much less successful,

giving mixtures of products which included (according to mass

spectroscopy of the crude products) a dimer of the deprotected

ligand, possibly arising from a Lewis-acid induced Diels–Alder

reaction.³⁶Given the evident fragility of the polyene chain of

formulation as dinuclear molybdenum() species, with g_av

=

1.94 and a hyperﬁne pattern consistent with coupling of both

unpaired electrons to both molybdenum centres (i.e. over-

lapping singlet, sextet and undecet components with

a

H (C᎐C) under the rather harsh conditions used for demethyl-

᎐

2

separation of ca. 25 Gauss between adjacent hyperﬁne signals).

Such spectra are entirely characteristic of dinuclear oxo-molyb-

denum() complexes in which the two unpaired electrons are

exchange-coupled, as discussed in detail earlier.^1,8

Two of the complexes, Mo₂(CO) and Mo₂(NN), aﬀorded

X-ray quality crystals and the structures are shown in Fig. 1

(see also Tables 2 and 3). In both cases the coordination

geometry around the molybdenum() centres is unremarkable;

both structures show the common^3,5,6,8disorder between the

ating the methyl ether group, the acetal-based n-propyloxy

methyl ether protecting group³⁷was used for the synthesis of

H (C᎐C) , as it could be removed under mild conditions (HCl

᎐

2

3

in MeOH–thf).

The ethynyl-bridged ligand H (C᎐C) has also been reported

᎐

2

before,^38,39although the syntheses are again capable of

improvement in the light of more recent techniques. We

prepared 1,2-bis(4-methoxyphenyl)ethyne by a modiﬁcation of

the Sonogashira–Hagihara coupling of acetylene gas with aryl

halides,⁴⁰in which 2-methyl-3-butyn-2-ol, a protected form of

acetylene, is used instead of acetylene gas.⁴¹This necessitates

addition of a base to the reaction mixture in order to remove

the protecting group in situ, and we used a heterogeneous base

(KOH) in conjunction with a phase-transfer agent in order to

maintain a constant but low concentration of base in the

organic phase throughout the reaction.⁴²We found that initial

stirring of the catalyst precursors in a small amount of piper-

idine,⁴³before addition of the remaining reagents, also

oxo and chloride ligands which renders Mo᎐O and Mo–Cl

᎐

distances inaccurate. The important point in each structure is

the inﬂuence of the bridging ligand. In Mo₂(CO) the bridging

benzophenone unit is buckled with an angle of 52Њ between the

mean planes of the two aromatic rings. This type of conform-

ation is typical of benzophenone units in the solid state.^44–48In

contrast, in the structure of Mo₂(NN) the bridging ligand is

planar, as required by the crystallographic inversion centre at

the centre of the N᎐N linkage. The intermolecular Mo ؒ ؒ ؒ Mo

᎐

separations are 11.87 and 14.61 Å respectively.

improved the yield of 1,2-bis(4-methoxyphenyl)ethyne which

᎐

was demethylated with BBr to give H (C᎐C).

᎐

3

2

Electrochemical properties

H₂(NN) was prepared from 4-aminophenol by (i) protection

i

of the phenol with an Pr₃Si group, (ii) oxidative dimerisation

(i) General considerations. The redox properties of the com-

plexes were examined by cyclic and square-wave voltammetry,

and the results are collected in Table 4. The general pattern that

we have observed in our earlier studies has been preserved, viz.

that the complexes generally show two clearly separated

one-electron processes [previously assigned as Mo(VI)–Mo(V)

couples] at positive potentials, and a single larger wave [corre-

sponding to two coincident Mo(V)–Mo(IV) couples] at an

approximately invariant negative potential (Fig. 2).^2,3,8The fact

that the oxidations are separated but the reductions are coin-

cident, i.e. the oxidised mixed-valence state has a much higher

comproportionation constant than the reduced mixed-valence

state, has been ascribed to the fact that bridging bis-phenolate

ligands of this type are themselves oxidisable to give semi-

quinones and then quinones, whereas they are not reducible.

of the aniline group and (iii) deprotection of the phenol with

ﬂuoride. H₂(C₂ꢀC₂) was prepared via a Heck coupling of

4-vinylanisole (two equivalents) with 1,4-dibromobenzene,

followed by demethylation. The ligand 4,4Ј-dihydroxybenzo-

phenone (H₂CO) is commercially available.

A note on the characterisation is appropriate at this point.

The protected ligand precursors are generally soluble, stable

species which have fully been characterised by the usual

methods. The deprotected bis-phenol ligands, especially the

longer ones, tend to be poorly soluble and not amenable to

e.g. chromatographic puriﬁcation. For this reason they were

usually used crude as isolated from the deprotection reaction,

as we found it much easier to purify the stable, soluble metal

complexes by chromatography and recrystallisation.

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

1403

View Article Online

Fig. 1 Crystal structures of the dinuclear complexes (a) Mo₂(NN) and (b) Mo₂(CO).

Table 3 Selected bond distances (Å) and angles (Њ) for Mo₂(NN)^a

Table 2 Selected bond distances (Å) and angles (Њ) for Mo₂(CO)ؒ

a

0.5C₅H₁₂ؒ1.5CH₂Cl₂

Mo(1)–O(1Ј)

Mo(1)–O(1)

Mo(1)–O(48)

Mo(1)–N(22)

1.782(13)

1.792(19)

1.945(4)

2.172(5)

Mo(1)–Cl(1Ј)

Mo(1)–N(32)

Mo(1)–Cl(1)

Mo(1)–N(12)

2.216(6)

2.221(5)

2.226(6)

2.230(5)

Mo(1)–O(1Ј)

Mo(1)–O(1)

Mo(1)–O(71)

Mo(1)–N(22)

Mo(1)–N(12)

Mo(1)–N(32)

Mo(1)–Cl(1)

Mo(1)–Cl(1Ј)

1.72(2)

1.76(2)

Mo(2)–O(2Ј)

Mo(2)–O(2)

Mo(2)–O(94)

Mo(2)–Cl(2Ј)

Mo(2)–N(52)

Mo(2)–N(62)

Mo(2)–N(42)

Mo(2)–Cl(2)

1.727(14)

1.742(13)

1.924(5)

2.197(7)

2.169(7)

2.224(6)

2.235(7)

2.287(7)

1.952(5)

2.180(7)

2.213(6)

2.223(6)

2.318(8)

2.325(9)

O(1Ј)–Mo(1)–O(48)

O(1)–Mo(1)–O(48)

O(1Ј)–Mo(1)–N(22)

O(1)–Mo(1)–N(22)

99.9(5)

98.3(7)

88.2(5)

89.5(7)

Cl(1Ј)–Mo(1)–N(32)

O(1)–Mo(1)–Cl(1)

O(48)–Mo(1)–Cl(1)

N(22)–Mo(1)–Cl(1)

94.0(2)

100.2(6)

94.5(2)

95.0(2)

O(48)–Mo(1)–N(22) 166.50(15)

N(32)–Mo(1)–Cl(1) 173.9(2)

O(1Ј)–Mo(1)–Cl(1Ј)

O(48)–Mo(1)–Cl(1Ј)

N(22)–Mo(1)–Cl(1Ј)

99.6(5)

93.9(2)

95.4(2)

O(1Ј)–Mo(1)–N(12)

O(1)–Mo(1)–N(12)

O(48)–Mo(1)–N(12)

N(22)–Mo(1)–N(12)

Cl(1Ј)–Mo(1)–N(12) 173.5(2)

N(32)–Mo(1)–N(12)

Cl(1)–Mo(1)–N(12)

86.9(5)

164.3(6)

85.03(17)

84.59(18)

O(1Ј)–Mo(1)–O(71)

O(1)–Mo(1)–O(71)

O(1Ј)–Mo(1)–N(22)

O(1)–Mo(1)–N(22)

O(71)–Mo(1)–N(22)

O(1Ј)–Mo(1)–N(12)

O(1)–Mo(1)–N(12)

O(71)–Mo(1)–N(12)

N(22)–Mo(1)–N(12)

O(1Ј)–Mo(1)–N(32)

O(1)–Mo(1)–N(32)

O(71)–Mo(1)–N(32)

N(22)–Mo(1)–N(32)

N(12)–Mo(1)–N(32)

O(1)–Mo(1)–Cl(1)

O(71)–Mo(1)–Cl(1)

N(22)–Mo(1)–Cl(1)

N(12)–Mo(1)–Cl(1)

N(32)–Mo(1)–Cl(1)

O(1Ј)–Mo(1)–Cl(1Ј)

O(71)–Mo(1)–Cl(1Ј)

N(22)–Mo(1)–Cl(1Ј)

N(12)–Mo(1)–Cl(1Ј)

N(32)–Mo(1)–Cl(1Ј)

100.0(9)

98.5(7)

88.7(9)

89.4(7)

167.2(2)

168.6(8)

90.9(7)

85.8(2)

84.0(2)

90.8(9)

168.8(8)

86.9(2)

83.7(2)

79.7(2)

102.1(7)

95.0(3)

93.1(3)

166.8(3)

87.2(3)

98.7(9)

94.0(4)

93.9(4)

90.6(4)

170.1(4)

O(2Ј)–Mo(2)–O(94)

O(2)–Mo(2)–O(94)

O(2Ј)–Mo(2)–Cl(2Ј)

O(94)–Mo(2)–Cl(2Ј)

O(2Ј)–Mo(2)–N(52)

O(2)–Mo(2)–N(52)

O(94)–Mo(2)–N(52)

Cl(2Ј)–Mo(2)–N(52)

O(2Ј)–Mo(2)–N(62)

O(2)–Mo(2)–N(62)

O(94)–Mo(2)–N(62)

Cl(2Ј)–Mo(2)–N(62)

N(52)–Mo(2)–N(62)

O(2Ј)–Mo(2)–N(42)

O(2)–Mo(2)–N(42)

O(94)–Mo(2)–N(42)

Cl(2Ј)–Mo(2)–N(42)

N(52)–Mo(2)–N(42)

N(62)–Mo(2)–N(42)

O(2)–Mo(2)–Cl(2)

O(94)–Mo(2)–Cl(2)

N(52)–Mo(2)–Cl(2)

N(62)–Mo(2)–Cl(2)

N(42)–Mo(2)–Cl(2)

99.3(5)

98.3(8)

107.4(6)

95.5(3)

88.0(5)

91.0(8)

165.6(3)

94.1(3)

163.3(6)

86.3(7)

85.7(2)

87.9(3)

83.9(3)

83.9(6)

165.6(7)

86.3(3)

168.0(3)

82.1(3)

80.4(2)

94.2(7)

94.5(2)

95.8(2)

179.4(2)

99.0(2)

O(1Ј)–Mo(1)–N(32) 164.0(4)

O(1)–Mo(1)–N(32)

O(48)–Mo(1)–N(32)

N(22)–Mo(1)–N(32)

^aThe oxo and chloride ligands at the Mo are disordered. The site occu-

pancies are as follows: O(1) and Cl(1), 0.46; O(1Ј) and Cl(1Ј), 0.54.

Owing to this disorder, distances and angles involving these atoms will

be relatively inaccurate.

85.3(6)

87.45(16)

82.19(17)

79.55(18)

94.8(2)

Table 4 Electrochemical and magnetic data for the complexes

Redox potentials, V vs. Fc–Fc^ϩ

Mo(V)–

∆E_1/2/

mV

Complex

Mo(IV)^a

A^b

B^b

J/cm^Ϫ1

Ϫ3.6

Mo₂(th)

Ϫ1.15

Ϫ1.11

Ϫ1.12

Ϫ1.18

Ϫ1.16

Ϫ1.20

Ϫ1.15

Ϫ1.13

Ϫ1.07

Ϫ1.15

ϩ0.33^c

ϩ0.44^c

ϩ0.34^c

ϩ0.35^c

ϩ0.28^c

ϩ0.26^c

ϩ0.53^c

ϩ0.60^c

ϩ0.78^c

ϩ0.43^c

ϩ0.70^c

ϩ0.69^d

ϩ0.51^d

ϩ0.69^c

ϩ0.48^d

ϩ0.38^d

ϩ0.88^d

ϩ0.82^c

ϩ0.91^c

ϩ0.56^c

370

250

170

340

200

120

350

220

130

Mo₂(th)₂_e

0

f

Mo₂(th)₃

Mo₂(C᎐C)

Ϫ11.2

᎐

f

^aThe oxo and chloride ligands at each Mo are independently dis-

ordered. The site occupancies are as follows: O(1) and Cl(1), 0.52; O(1Ј)

and Cl(1Ј), 0.48; O(2) and Cl(2), 0.45; O(2Ј) and Cl(2Ј), 0.55. Owing

to this disorder, distances and angles involving these atoms will be

relatively inaccurate.

Mo₂(C᎐C)₂

᎐

f

Mo₂(C᎐C)₃

᎐

Mo₂(C᎐C)

᎐

Ϫ7

Ϫ12.8

Mo₂(N᎐N)

᎐

Mo₂(CO)

Ϫ1.1

f

Mo₂(C₂ꢀC₂)

^aTwo coincident, unresolved one-electron processes; peak–peak separ-

ation ∆E_ptypically 100–150 mV. ^bSee main text for deﬁnition of the

couples A and B. ^cReversible one-electron process: peak–peak separ-

ation in the range 70–100 mV, equal cathodic and anodic peak currents,

and shown to be chemically reversible by spectroelectrochemistry.

^dIrreversible process: return wave of lower intensity than outward

wave, and results in partial or total decomposition in the spectro-

electrochemistry experiment. ^eA third oxidation is present for this

compound, at ϩ0.98 V vs. Fc–Fc^ϩ, which appears to be reversible by

cyclic voltammetry (see main text). ^fNot measured.

Thus, delocalisation of the oxidised mixed-valence state occurs

via hole transfer through the bridging ligand HOMO because

the states Mo^VI–(L^2Ϫ)–Mo^Vand Mo^V–(L^Ϫ)–Mo^Vare not very

diﬀerent in energy.^3,8In contrast, reduction of the dianionic

bridging ligand is diﬃcult, so delocalisation of the extra

electron in the reduced mixed-valence state via the bridging-

ligand LUMO [Mo^IV–(L^2Ϫ)–Mo^V→ Mo^V–(L^3Ϫ)–Mo^V] is not

feasible.

1404

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

View Article Online

for this compound is consistent with the fact that oligothienyl

fragments become easier to oxidise as they become longer,^58–61

so the appearance of an additional ligand-centred process is to

be expected as more thienyl units are placed in the bridging

ligand.

Each thienyl unit corresponds to two double bonds in the

bridging pathway, yet each causes less of an attenuation in the

electrochemical interaction than does a single ethenyl spacer

(attenuation coeﬃcient 0.59). These values may be compared

with an average attenuation coeﬃcient of 0.43 for the phenylene

spacers in the series [Mo–O(C₆H₄)_nO–Mo] (n = 0, 1, 2). The

attenuation coeﬃcients reﬂect the ability of the diﬀerent spacer

groups to stabilise the mixed-valence complexes (whether metal-

or ligand-based) by delocalisation of the odd electron. It is to

be expected that thienyl units are more eﬀective at this than

phenylene units, for the reasons mentioned earlier: both spacers

are about the same length, an additional two double bonds in

the bridging pathway, but oligothienyl units are more coplanar

than oligophenylene units.^27–30It is surprising however that the

thienyl spacer is also more eﬀective than a single ethenyl unit

at stabilising the mono-oxidised radicals by delocalisation,

especially as polyene spacers are also expected to be approx-

imately coplanar. The thienyl units are therefore acting as more

than just planar butadienyl units. This must reﬂect the elec-

tronic contribution from the S atoms, whose π-symmetry

orbitals will contribute to the delocalised orbital system of the

bridging ligand. The result is that even across three thienyl

spacers (crudely equivalent to six double bonds) the two redox

couples are still signiﬁcantly separated.

Previous studies on polyene-linked series of dinuclear com-

plexes studied by Reimers and Hush,¹²and Launay, Spangler

and co-workers¹³have shown an approximately exponential

decrease in the electronic coupling parameter V_abas the polyene

chain lengthens; values of V_abwere determined from the prop-

erties of the inter-valence charge-transfer (IVCT) band for each

compound. Our results, for both the ethenyl-spacer and thienyl-

spacer series, are consistent with this general expectation. We

have not been able to extract values of V_abfrom the spectra of

the Mo(VI)–Mo(V) complexes, because the expected IVCT

transitions are obscured by much stronger LMCT transitions

involving the molybdenum() centre (vide infra). However as

long as ∆E is dominated by the through-bond electronic

coupling then it should show the same exponential decay with

distance as does V_ab,^13,14as we observe.⁶²

᎐

Fig. 2 Cyclic voltammograms of (a) Mo₂(C᎐C) and (b) [Mo₂(C₂ꢀC₂)]

᎐

(CH₂Cl₂, platinum bead working electrode, scan rate 0.2 V s^Ϫ1).

The non-innocence of the bridging ligands is obviously

related to the properties of the numerous ruthenium() com-

plexes of o-dioxolenes studied by both us^49–54and by Lever and

co-workers^55–57in which there is strong mixing between metal-

and ligand-centred frontier orbitals such that assignment of

redox processes as metal- or ligand-centred can be diﬃcult. The

two well separated oxidations, formally Mo(V) → Mo(VI)

processes, can therefore have some ligand-centred character.

Description of these processes as ‘Mo(VI)–Mo(V) couples’ may

therefore not be entirely appropriate, and to avoid confusion

in the following discussion the successive redox processes are

simply denoted A and B. In contrast the Mo(V) → Mo(IV)

reductions are metal-localised.

(ii) The series Mo (C᎐C) and Mo (th) (n

؍

1, 2, 3). Starting

᎐

2

n

2

n

with the series Mo (C᎐C) (n = 1, 2, 3) we see that the separation

᎐

2

n

(∆E) between redox processes A and B starts at a high value

(340 mV) for n = 1, and steadily diminishes to 200 mV for n = 2

and 120 mV for n = 3. The decay in ∆E is clearly not linear: the

fractional decrease in ∆E is about constant for each increase in

n, and we can accordingly derive an ‘attenuation coeﬃcient’ of

Apart from the separations between the two oxidation

processes, their absolute values are also of interest. For the

about 0.59 for the C᎐C unit, which is the amount by which ∆E

᎐

is multiplied as each additional C᎐C unit is inserted into the

Mo (C᎐C) series the average of the two oxidation potentials

᎐

2 n

᎐

bridging ligand. For the complex with no double bonds in the

pathway, i.e. the biphenyl-bridged complex [Mo–O(C₆H₄)₂O–

shifts to less positive potential as the chain lengthens: 0.52 V vs.

Fc–Fc^ϩfor n = 1; 0.38 V for n = 2; and 0.32 V for n = 3. If these

two couples were totally metal-centred, we would say that this

greater ease of oxidation reﬂected increased electron density at

the metal centre. However this is not consistent with the fact

that the redox potentials of the (metal-centred) Mo(V)–Mo(IV)

couples are almost invariant across the series. This reduction

of the redox potentials of couples A and B as the bridging

ligand lengthens is however consistent with steadily increasing

ligand-centred character for couples A and B. Similar

behaviour, although not so clear-cut, was observed with the

polyphenylene-bridged series of complexes [Mo–O(C₆H₄)_n-

O–Mo] (n = 1–4).⁸The oligothienyl-bridged series Mo₂(th)_ndoes

not show a clear trend at all, with the average potential for

the two couples A and B being ϩ0.52 (n = 1), ϩ0.57 (2) and

ϩ0.43 V (3).

Mo], ∆E is 480 mV. Simply extrapolating from the Mo (C᎐C)

᎐

2

n

(n = 1–3) series, whose bridging ligands are approximately

planar, would suggest a value of ca. 580 mV for this redox

separation; however this does not take into account the twisted

conformation of the biphenyl spacer which will decrease the

interaction. The steady attenuation in ∆E as the number of

ethenyl units increases is equally consistent with A and B being

metal-centred or ligand-centred processes. Although both

waves look symmetric on the timescale of cyclic voltammetry,

spectroelectrochemical measurements (see later) showed that

the second oxidation process (B) became more irreversible as

more ethenyl units are added to the bridge.

For the series Mo₂(th)_n(n = 1, 2, 3) we again see substantial

separations between redox couples A and B, with ∆E = 370, 250

and 170 mV respectively, giving a consistent attenuation coef-

ﬁcient of 0.68. For n = 1 both oxidations appear to be fully

chemically reversible; for n = 2 and 3, the second oxidation is

not fully reversible with the return wave of lower intensity than

the outward wave (reasons for this are discussed later). For

Mo₂(th)₃only, an additional third redox process appears at

ϩ0.98 V vs. Fc–Fc^ϩ. The appearance of an extra redox process

᎐

(iii) The dinuclear complexes Mo (C᎐C), Mo (N᎐N),

᎐

2

᎐

Mo (CO) and Mo (C ꢀC ). For Mo (C᎐C), the couples A and

᎐

2

B are separated by 350 mV, the same as for Mo (C᎐C). In this

᎐

2

case the alkynyl and alkenyl linkages are about equally as eﬀect-

ive as each other at transmitting the electronic interaction.

We ^63,64and others^65,66have previously observed that alkynyl

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

1405

View Article Online

spacers in other dinuclear complexes gave slightly smaller ∆E

values than alkenyl spacers, although the diﬀerences were not

blue-shifted (λ_max= 1017, 1015, 1033 nm respectively) and

tends to be sharper and more intense than in the mixed-valence

state (Fig. 3). In all three complexes the two one-electron oxid-

ations are fully chemically reversible and the dication is

indeﬁnitely stable in solution under the conditions of the

experiment.^3,8

large. The azo-linkage of Mo (N᎐N) however is a much less

᎐

2

eﬃcient conductor than an ethenyl linkage, with a ∆E value

of only 220 mV. The highly twisted conformation of the

bridging ligand in Mo₂(CO) accounts for the low ∆E value of

130 mV.

As mentioned above, the question arises as to the nature of

the two oxidations as metal-based or ligand-based. The limiting

canonical forms of the doubly oxidised species are depicted in

Scheme 2. These can either be Mo(VI)–Mo(VI) complexes with

a bis-phenolate bridge following metal-centred oxidation, or

dinuclear molybdenum(V) complexes spanned by a bridging

quinone following ligand-centred oxidation. The diﬀerent elec-

tronic transitions expected for these extreme forms are included

in Scheme 2. In particular, if the oxidations are metal-centred

giving a Mo(VI)–Mo(V) and then a Mo(VI)–Mo(VI) complex,

then intense, low-energy phenolate→Mo(VI) LMCT transi-

tions are expected; if they are ligand-centred, we expect to see

semiquinone-based and then quinone-based π → π* transi-

tions of the bridging ligand (Scheme 2).^67–73There is no doubt

from a wealth of magnetic and EPR spectroscopic data that

in the starting (neutral) state these complexes are Mo(V)–

(µ-diphenolate)–Mo(V), with the highest-energy occupied

orbitals being the singly occupied d_xyorbital on each metal, and

the bridging ligand HOMO being below these.^1–3On this basis

we expect the ﬁrst oxidation to be metal-centred; however it

would only require a small change in the relative energies of

In contrast the ∆E value of 130 mV across the much longer

bridging ligand of Mo₂(C₂ꢀC₂) is quite impressive given the

distance involved, which is equivalent to four double bonds [cf.

∆E = 120 mV for Mo (C᎐C) ]. It is much larger than would be

᎐

2

3

expected if we just start from [Mo–O(C₆H₄)₂O–Mo] (∆E = 480

mV) and apply the attenuation coeﬃcients for each additional

spacer in turn: with two double bonds and a phenylene unit this

would suggest a ∆E value of about 480 × 0.59 × 0.59 × 0.43 =

72 mV. However, as mentioned above, pure oligophenylene

chains are expected to be highly twisted and the attenuation

eﬀect of an additional phenylene spacer in these cases reﬂects

not only the increased metal–metal separation, but also an addi-

tional twist to disrupt delocalisation in the bridging ligand. In

Mo₂(C₂ꢀC₂) the phenylene spacer is necessarily coplanar with

the rest of the bridging ligand, which results in it having a

smaller attenuation eﬀect than it does in the oligophenylene

chains.

UV/VIS/NIR spectroelectrochemical behaviour (see Table 5)

(i) Properties of the phenylene-bridged complexes [Mo–

O(C₆H₄)_nO–Mo]. Before discussing the spectroelectrochemical

characteristics of the new complexes, it will be useful to sum-

marise brieﬂy the properties of the oligophenylene-bridged

dinuclear complexes [Mo–O(C₆H₄)_nO–Mo] (n = 2, 3, 4) which

we described in a previous paper.³In the Mo(V)–Mo(V)

starting state the spectra of these show the typical features to be

expected for the mononuclear model complex [Mo^V(Tp^Me,Me)-

(O)Cl(OPh)], with the lowest-energy transition being a phen-

olate→Mo(V) LMCT in the region ca. 500–700 nm. On

one-electron oxidation to the (formally) Mo(VI)–Mo(V) state,

an intense near-IR transition appears (λ_max= 1096, 1131, 1047

nm for n = 2, 3, 4 respectively); on further oxidation to the

(formally) Mo(VI)–Mo(VI) complex this transition is slightly

Fig. 3 Electronic spectra of [Mo–O(C₆H₄)₃O–Mo]^nϩ(n = 0, 1, 2) (from

ref. 3).

Scheme 2 Structures of the doubly oxidised forms of the dinuclear complexes arising from (i) purely metal-centred oxidations, and (ii) purely

ligand-centred oxidations. The electronic transitions (hollow arrows) are: (a) phenolate→Mo(V) LMCT; (b) phenolate→Mo(VI) LMCT;

(c) quinone-based π → π*; and (d) Mo(V)→quinone(π*) MLCT.

1406

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

View Article Online

ligand and metal frontier orbitals following the ﬁrst oxidation

to alter this picture for subsequent oxidations.

We concluded that the former description (metal-centred

oxidations) was more appropriate in the series [Mo–O(C₆H₄)_nO–

Mo], for several reasons. Firstly, ZINDO calculations sup-

ported the description of the near-IR transitions as phen-

olate→Mo(VI) LMCT processes in the doubly oxidised

complexes. Secondly, the strong near-IR absorptions are

characteristic of the low-energy phenolate→Mo(VI) LMCT

process of mononuclear [Mo(Tp^Me,Me)(O)Cl(OPh)]^ϩ,³but are

quite diﬀerent from the spectra of the relevant free poly-

phenylene quinones whose π → π* transitions are much

higher in energy.^67–73Thirdly, the absorption maximum for

the free quinone series becomes progressively red-shifted as the

quinone lengthens,^69–73which does not happen in the series

[Mo–O(C₆H₄)_nO–Mo]^2ϩ(n = 2–4) where the near-IR absorption

maximum hardly changes as the bridging ligand lengthens,

consistent with phenolate→Mo(VI) LMCT character for the

transitions. Fourthly, the observation that λ_maxis similar in

energy for both the mono- and di-oxidised forms (e.g. Fig. 3)

is inconsistent with the behaviour of the free semiquinone and

quinone species whose spectra are quite diﬀerent;⁶⁷it is how-

ever consistent with the presence of one and then two localised

phenolate→Mo(VI) LMCT processes. These factors together

strongly suggest two metal-centred oxidations for the series

[Mo–O(C₆H₄)_nO–Mo].

(ii) Properties of the thienyl-bridged complexes Mo₂(th)_n. The

electronic spectra of these complexes in the Mo(V)–Mo(V)

state are generally similar to those of the other molybdenum(v)

complexes discussed earlier, with the lowest-energy transition

(630–660 nm) being a phenolate→Mo(V) LMCT and the

higher-energy transitions having their usual assignments.^3,8

One-electron oxidation of Mo₂(th) to the monocation

Electronic spectra of (a) [Mo₂(th)]^nϩ(n = 0, 1, 2); (b)

[Mo₂(th)]^ϩgenerates an intense new near-IR transition (λ_max

=

Fig.

4

[Mo₂(th)₂]^nϩ(n = 0, 1); (c) [Mo₂(th)₃]^nϩ(n = 0, 1) [* = detector change].

1199 nm; ε = 56,000 M^Ϫ1cm^Ϫ1) which has a high-energy

shoulder. The intensity and position of this are characteristic of

a phenolate→Mo(VI) LMCT, as observed in the oligophen-

ylene-bridged series,³and this is to be expected given that the

HOMOs are the metal-based d_xyorbitals, each containing one

electron (cf. EPR results, above; and magnetic susceptibility

results, below). On further oxidation to [Mo₂(th)]^2ϩthis transi-

tion disappears and is replaced by one at much higher energy

(684 nm) [Fig. 4(a)]. This behaviour is not consistent with

a second metal-centred oxidation to give a Mo(VI)–Mo(VI)

species, but is more consistent with oxidation of the bridging

ligand to give a quinone (Scheme 2): the π–π* absorption

maximum for the free quinone derived from H₂th is at 531 nm.³¹

We suggest therefore that [Mo₂(th)]^2ϩis more accurately

described as [Mo(V)]₂(µ-quinone) than as [Mo(VI)]₂(µ-diolate).

This requires that the second oxidation results in an internal

redistribution of electrons such that the sequence of species

formed after successive oxidations is Mo(V)(µ-diolate)Mo(V)

→ Mo(V)(µ-diolate)Mo(VI) → Mo(V)(µ-quinone)Mo(V),

in marked contrast to the behaviour of the phenylene-spaced

analogues (cf. Fig. 3); we return to this point later. The lower-

energy transitions such as the shoulder at ca. 1000 nm could

then have Mo(V)→quinone(π*) LMCT character.

consistent with formation of a bridging quinone, for two

reasons. Firstly, the electrostatic interaction between a neutral

quinone O-donor and Mo(V) must be much weaker than

between a negatively charged phenolate and Mo(VI). Others

have observed how 2-electron oxidation of a coordinated

aromatic diolate to a neutral quinone in a dinuclear complex

results in its dissociation from the metal ions for this reason.⁷⁴

Secondly, extended quinones are known to be unstable because

of the presence of a low-lying triplet excited state, in which

unpaired electron density at C²and C⁶of the terminal rings

(next to the oxygen atom) leads to oligomerisation. For this

reason extended quinones such as p-terphenoquinone and

p-quaterphenoquinone are prepared with bulky substituents to

block these C²and C⁶positions which renders them much more

stable.^71,73,75,76

One-electron oxidation of [Mo₂(th)₃] to [Mo₂(th)₃]^ϩagain

results in a broad region of absorbance in the near-IR region

with two maxima resolved at 1304 and 1096 nm [Fig. 4(c)].

These are of rather low intensity compared to other such tran-

sitions in mono-oxidised complexes, but this is consistent

with the substantial drop in intensity for the NIR transition

between [Mo₂(th)]^ϩand [Mo₂(th)₂]^ϩand is possibly related to an

increasing contribution from thienyl radical character, given

the known ease of oxidation of oligothiophenes as they

lengthen.^58–61As we would expect by extrapolation from the

properties of [Mo₂(th)_n] (n = 1, 2) the second oxidation is wholly

irreversible on the spectroelectrochemistry timescale and clearly

results in decomposition of the dication.

Oxidation of [Mo₂(th)₂] to [Mo₂(th)₂]^ϩlikewise results in an

intense near-IR transition (λ_max= 1342 nm; ε = 23,500 M^Ϫ1

cm^Ϫ1) having a shoulder at the high-energy side [Fig. 4(b)],

which we assign as a phenolate→Mo(VI) LMCT process

following metal-centred oxidation. The particularly low energy

of this redox-switchable transition makes it of interest for use

in modulating near-IR lasers and we have recently described

how this [Mo₂(th)₂]^0/ϩcouple can be exploited for near-IR

optical switching.²⁶Further oxidation to [Mo₂(th)₂]^2ϩis however

irreversible on the slow timescale of spectroelectrochemistry,

with spectra showing clear evidence of decomposition of the

dication. This irreversibility of the second oxidation is actually

(iii) Properties of the ethenyl-bridged complexes Mo (C᎐C) .

᎐

2

n

In all cases the spectra of the starting Mo(V)–Mo(V) states in

this series are assignable as before.^3,8For Mo (C᎐C), oxidation

᎐

2

to [Mo (C᎐C)]^ϩresults in the familiar phenolate→Mo(VI)

᎐

2

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

1407

View Article Online

Fig. 5 Electronic spectra of (a) [Mo₂(C᎐C)]^nϩ(n = 0, 1, 2); (b)

᎐

[Mo₂(C᎐C)₂]^nϩ(n = 0, 1, 2).

᎐

LMCT near-IR transition at 1151 nm (ε = 40,000 M^Ϫ1cm^Ϫ1).

Further oxidation to [Mo (C᎐C)]^2ϩresults in evolution of the

᎐

2

spectra in a way similar to that seen on formation of

[Mo₂(th)]^2ϩ, viz. replacement of the NIR transition by a pair

of transitions at much higher energy (500–600 nm), which is

strongly characteristic of a quinone (for comparison, 3,3Ј,5,5Ј-

tetra-tert-butyl-4,4Ј-stilbenequinone has λ_max= 460 nm)⁶⁷but

which is too high in energy to be a phenolate→Mo(VI)

LMCT process [Fig. 5(a)]. As with [Mo₂(th)] we are seeing

Fig.

6

Electronic spectra of (a) [Mo₂(CO)]^nϩ(n = 0, 1, 2); (b)

[Mo₂(C₂ꢀC₂)]^nϩ(n = 0, 1, 2); (c) [Mo₂(N᎐N)]^nϩ(n = 0, 1, 2) [* = detector

᎐

change].

spectrum characteristic of a bridging quinone which requires

both oxidations to be ligand-centred. Secondly, as the ligand

increases in length the doubly oxidised complex decreases

in stability, which is also indicative of double ligand-centred

oxidation as mentioned earlier.

spectroscopic behaviour indicative of

a shift to ligand-

centred oxidations following an internal charge redistribution

associated with the second oxidation. Compared to [Mo–

O(C₆H₄)₂O–Mo] (where both oxidations are metal-centred),³

addition of the double bond between the two phenyl rings

appears to have resulted in a greater degree of ligand-

centred character for the oxidations, although this is not

obviously reﬂected in the electrochemical properties of the

respective free quinones.⁶⁷In addition we note that the presence

of two closely spaced absorption maxima in this case is in

agreement with the electronic spectra of a variety of extended

quinones.⁶⁷

᎐

(iv) Properties of Mo (C᎐C), Mo (N᎐N), Mo (CO) and

᎐

2

Mo₂(C₂ꢀC₂). Mo₂(CO) undergoes two successive metal-centred

oxidations; in this case there is no ambiguity in the assignment

as the bridging ligand cannot be oxidised to a quinone. The

relatively high energy of the phenolate→Mo(VI) LMCT

processes (695 and 698 nm for the mono- and di-cation

respectively) [Fig. 6(a)] reﬂects the fact that the two termini are

eﬀectively decoupled by the twist in the ligand, such that each

end behaves like an isolated molybdenum(VI) group, cf. the

Mo (C᎐C) behaves similarly, with the monocation giving a

᎐

2

strong NIR phenolate→Mo(VI) LMCT transition (λ_max= 1210

nm; ε = 41,000 M^Ϫ1cm^Ϫ1). The spectrum of the dication

however shows a quinone-like absorption maximum at much

higher energy (595 nm) [Fig. 5(b)], as we saw for [Mo (C᎐C)]^2ϩ

LMCT transition of [Mo^VI(Tp^Me,Me)(O)Cl(OPh)]^ϩat 681 nm.³

In Mo (C᎐C) the ﬁrst oxidation is reversible and results in the

typical NIR transition at 1052 nm, but the second oxidation is

irreversible and results in decomposition.

᎐

2

᎐

2

and [Mo₂(th)]^2ϩ. In keeping with our earlier observation that an

increasing degree of quinone character in the doubly oxidised

complex causes irreversibility of the second oxidation, we

The NIR transition arising from the ﬁrst oxidation of

Mo₂(C₂ꢀC₂), at 1554 nm, is the longest-wavelength such

transition we have seen so far and is in the region of the

electromagnetic spectrum which is of particular interest for

electro-optic switching [Fig. 6(b)].²⁶On further oxidation to

[Mo₂(C₂ꢀC₂)]^2ϩthere is a pronounced blue-shift of the new

absorption maxima (1144 and 978 nm), which could be taken

to indicate quinone formation, except that the shift is not as

dramatic as would be expected: a ZINDO calculation suggests

that the π–π* transition of the corresponding free quinone

would be at 539 nm. It may be that the second oxidation in this

case has more metal-centred character, which is consistent with

the observation that both redox processes are fully chemically

reversible.

found that [Mo (C᎐C) ]^2ϩshowed signs of decomposition:

᎐

2

reduction back to Mo (C᎐C) regenerated the starting spectrum

᎐

2

but with a loss of about 20% of its intensity after a few hours at

Ϫ30 ЊC. For Mo (C᎐C) the problem was worse and only the

᎐

2

3

monocation could be spectroscopically characterised (λ_max

=

1272 nm; ε = 21,000 M^Ϫ1cm^Ϫ1for the NIR phenolate→Mo(VI)

LMCT transition).

This series of complexes therefore behaves in a similar

manner to [Mo₂(th)_n] (n = 1, 2), in two respects. Firstly, oxidation

A gives a NIR transition which is ascribable to a phenolate→

Mo(VI) LMCT process following metal-centred oxidations, but

oxidation B (when the product does not decompose) gives a

1408

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

View Article Online

Table 5 UV/VIS/NIR Spectroelectrochemical data (CH₂Cl₂, Ϫ30 ЊC)

In contrast, the azo-bridged complex Mo (N᎐N) shows most

᎐

2

clearly of all those spectroscopic changes which we have

ascribed to the shift from metal-centred to ligand-centred

oxidation following the second redox process [Fig. 6(c)]. There

is a dramatic diﬀerence between the principal absorption

maxima for [Mo (N᎐N)]^ϩ[λ_max= 1268 nm, characteristic of a

Complex

n

λ_max/nm (10^Ϫ3ε/M^Ϫ1cm^Ϫ1

)

[Mo₂(th)]^nϩ

0

1

2

0

1

0

1

656 (7.9),^a414 (sh), 346 (31), 275 (26)

1199 (56),^b673 (36.2), 275 (26.7)

1000 (sh), 845 (37), 684 (87),^c456 (12), 276 (26)

646 (5.2),^a381 (30), 270 (sh)

᎐

2

phenolate→Mo(VI) LMCT process] and [Mo (N᎐N)]^2ϩ

᎐

2

[Mo₂(th)₂]^nϩ

[Mo₂(th)₃]^nϩ

(λ_max= 409 nm, indicative of quinone formation following the

second oxidation). Although the corresponding free quinone

is not known, a ZINDO calculation predicts its π → π*

absorption maximum to be at 410 nm, in excellent agree-

ment with our observations, and this behaviour is com-

pletely inconsistent with the occurrence of one and then two

phenolate→Mo(VI) LMCT transitions following successive

metal-centred oxidations (cf. the behaviour of the series

[Mo–O(C₆H₄)_nO–Mo]).³

1342 (24),^b752 (18), 359 (21), 269 (sh)

636 (3.2),^a411 (13), 337 (sh), 281 (22)

1304 (4.7), 1096 (7.1),^b612 (6.5), 337 (sh), 281

(22)

[Mo₂(C᎐C)]^nϩ

0

1

666 (7.5),^a420 (18), 333 (26), 306 (27)

1151 (40),^b1006 (sh), 692 (20), 500 (sh), 440

(10), 274 (16)

᎐

2

0

1

1100 (4.6), 739 (sh), 594 (23),^c531 (24), 350 (sh)

692 (7.8),^a357 (28), 274 (19)

[Mo₂(C᎐C)₂]^nϩ

᎐

1210 (41),^b1038 (28), 787 (22), 545 (12), 267

(18)

[Mo₂(C᎐C)₃]^nϩ

0

1

690 (sh),^a600 (10), 388 (37), 264 (23)

1272 (21),^b1055 (18), 789 (18), 665 (32), 271

(23)

᎐

(v) Discussion of spectroelectrochemical results: metal-centred

vs. ligand-centred redox activity. Starting with the monocations,

as we mentioned earlier the obvious assignment for the NIR

transition is a phenolate→Mo(VI) LMCT following metal-

centred oxidation. This accords with the behaviour of the

mononuclear molybdenum() complex [Mo(Tp^Me,Me)(O)-

Cl(OPh)]^ϩ, and also the dinuclear complex series [Mo–

O(C₆H₄)_nO–Mo]^ϩ.³ZINDO calculations^77,78on representative

mono-oxidised complexes assuming a spin doublet state do a

reasonable job of predicting an intense, low-energy LMCT

transition in the NIR region between a bridging-ligand-based

HOMO and a metal-based LUMO: for [Mo–O(C₆H₄)₂O–Mo]^ϩ

we have λ_max(calculated) at 965 nm compared to an observed

value of 1017 nm; for [Mo (C᎐C)]^ϩwe have λ_max(calculated) at

0

1

0

1

602 (5.1),^a376 (15), 285 (47)

nϩ

᎐

[Mo₂(C᎐C)]

᎐

1052 (30),^b592 (29), 471 (13), 362 (14), 282 (23)

644 (sh),^a505 (sh), 403 (27), 255 (22)

1268 (35),^b1150 (sh), 672 (21), 557 (sh), 308

(14), 266 (sh)

[Mo₂(N᎐N)]^nϩ

᎐

2

0

1

2

0

1

2

517 (sh), 409 (38),^c330 (sh)

[Mo₂(CO)]^nϩ

523 (4.6),^a362 (sh), 287 (24.1)

695 (12),^b513 (sh), 360 (sh), 274 (16)

698 (20),^b487 (sh), 367 (14), 270 (17)

660 (6.3),^a550 (sh), 375 (44), 253 (sh)

1554 (23),^b707 (18), 372 (16), 308 (16)

1144 (20), 978 (37), 792 (33), 430 (sh), 266 (15)

[Mo₂(C₂ꢀC₂)]^nϩ

^aPhenolate→Mo(V) LMCT. ^bPhenolate→Mo(VI) LMCT (see main

text). ^cIntense new transition in visible region for doubly oxidised com-

plexes; tentatively ascribed to π → π* transition of bridging quinone

(see main text).

᎐

2

1114 nm compared to an observed value of 1151 nm. The

agreement is less good for [Mo (N᎐N)]^ϩfor which λ_max(calcu-

᎐

2

lated) for the LMCT transition is at 883 nm, compared to

an observed value of 1268 nm, and is even worse for

[Mo₂(C₂ꢀC₂)]^ϩ[λ_max(calculated) = 843 nm; observed, 1554 nm].

We note that the parameters used by ZINDO for second-row

metals such as Mo are not as reliable as those for lighter

elements, and that ZINDO is less accurate at predicting

charge-transfer transitions than it is at predicting d–d and

ligand-centred transitions.⁷⁷Given these limitations, and the

fact that the HOMOs of the neutral complexes are deﬁnitely

the singly occupied metal-based d_xyorbitals, assignment

of the ﬁrst oxidation as formally metal-centred is straight-

forward.

For those dicationic complexes which were stable enough to

be studied by spectroelectrochemistry, the intense transitions

which we see in the UV and visible region are far more similar

to the π–π* transition of the corresponding free quinones than

they are to the phenolate→Mo(VI) LMCT processes which

occur in the NIR region. As mentioned earlier, this requires an

internal charge redistribution associated with the second oxid-

ation, such that one-electron oxidation of Mo(V)(µ-diolate)-

Mo(VI) aﬀords Mo(V)(µ-quinone)Mo(V) (to the extent that

such ‘localised’ descriptions are valid). There are several

precedents for this in the literature, notably amongst ﬁrst-row

following a ligand-centred redox process, e.g. reducing a ligand

will raise the energy of its frontier orbitals. If the metal- and

ligand-based frontier orbitals are close in energy to start with,

which is a deﬁning characteristic of complexes containing non-

innocent ligands, then this can result in an inversion of the

order of the metal- and ligand-based electronic levels and an

internal charge rearrangement. It is also worth mentioning

other work from Pierpont and co-workers who have character-

ised numerous complexes showing ‘valence tautomerism’ in

which e.g. Co^II(semiquinone) and Co^III(quinone) forms of the

same complex (i.e. ligand-oxidised or metal-oxidised) are so

close in energy that they can be interconverted by changes in

temperature or pressure.^84–87A more detailed analysis of the

evolution of the electronic spectra of these complexes through-

out the redox series will clearly require far more sophisticated

methods such as ab initio or DFT calculations; these will be the

subject of future work.

Magnetic exchange interactions, and comparison with electro-

chemical interactions

transition metal complexes with chelating o-dioxolenes of the

The variable-temperature magnetic susceptibility behaviour

of a selection of the complexes was measured with a SQUID

magnetometer and the results are listed in Table 4. It will be

seen that all complexes for which a value of J (the spin

exchange coupling constant) could be determined are anti-

ferromagnetically coupled. Comparison of the values for the

diﬀerent complexes shows some interesting eﬀects.

Ϫ

type [catecholate]^2Ϫ(cat) and [semiquinone] (sq) from the

᝽

work of Pierpont and co-workers. Thus, reversible one-electron

reduction of [Ni^II(sq)₂] aﬀords [Ni^III(cat)₂]^Ϫ, in which both

ligands become reduced but the metal is oxidised,⁷⁹and iden-

tical behaviour is shown by the [Mn^II(sq)₂]–[Mn^III(cat)₂]^Ϫ

couple.⁸⁰A more dramatic charge redistribution occurs in

[V^III(sq)₃], which on one-electron reduction aﬀords [V^V(cat)₃]^Ϫ;

a one-electron reduction of one sq ligand is accompanied by

two-electron oxidation of the metal and one-electron reduction

of each of the two remaining sq ligands.⁸¹Further examples of

such redox rearrangements come from the work of Busch⁸²and

Wieghardt⁸³and co-workers. Behaviour of this type may be

ascribed to a change in the donor character of the ligand

Starting from [Mo–O(C₆H₄)₂O–Mo] for which J = –13.2

cm^{Ϫ1 6}

, the interposed ethenyl spacer of Mo (C᎐C) has only

᎐

2

a slight attenuating eﬀect (J = –11.2 cm^Ϫ1). As with the electro-

chemical interaction, the increase in metal–metal separation

will partly be oﬀset by the more planar conformation of the

bridge which is important for facilitating the exchange inter-

6

᎐

action. The ethynyl spacer of Mo (C᎐C) is signiﬁcantly worse

᎐

2

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

1409

View Article Online

Conclusions

Study of the redox, magnetic and spectroelectrochemical

properties of dinuclear oxomolybdenum() complexes with

bis-phenolate bridging ligands having a variety of spacers

interposed between the phenyl rings has provided many insights

into how the bridging ligands control the metal–metal elec-

tronic and magnetic interactions in the complexes. In particular

we can draw the following conclusions.

(i) The eﬃciency of diﬀerent types of spacer at providing

delocalisation across the complex is thienyl > ethenyl > phenyl,

with thienyl units being particularly eﬀective at maintaining

large redox separations ∆E over long distances. An ethenyl

spacer is about as eﬀective as an ethynyl spacer at mediating

Fig. 7 Alternation of induced spins (small arrows) across the bridging

ligand of Mo₂(C᎐C) resulting in antiferromagnetic exchange between

᎐

the unpaired electrons on the metal centres (large arrows). A similar

᎐

picture can be drawn for Mo₂(NN) and Mo₂(C᎐C).

᎐

at propagating the spin exchange interaction (J = Ϫ7 cm^Ϫ1), but

interestingly the azo spacer of Mo (N᎐N) provides a stronger

᎐

2

electronic interactions; the azo-spacer (N᎐N) is however signiﬁ-

exchange interaction (J = Ϫ12.8 cm^Ϫ1) than the ethenyl spacer

᎐

cantly worse.

of Mo (C᎐C), despite giving a weaker electrochemical inter-

᎐

2

(ii) The oxidations of the complexes may be metal-centred

or ligand-centred depending on how easily the bridging bis-

phenolate bridging ligand can be oxidised to a quinone: the two

types of dication have quite distinct electronic spectra. With

oligophenylene bridges, spectroelectrochemical data (published

earlier) indicate that both oxidations are metal-centred.³With

a variety of other spacers such as ethenyl (1, 2 or 3), thienyl

(1 or 2), the ﬁrst oxidation again is largely metal-centred, but

the second oxidation results in an electronic spectrum which

is consistent with formation of [Mo(V)]₂(µ-quinone) species

following internal charge redistribution associated with the

second oxidation. In general, the doubly oxidised species whose

electronic spectra are characteristic of a bridging quinone

are unstable, which is consistent with dissociation of the

poorly coordinating neutral quinone units and the known

instability of extended quinones. In contrast metal-centred

oxidations to give molybdenum() phenolate species are fully

reversible.

(iii) There is not a good correlation between the magnitudes

of the electronic interactions and the magnetic interactions

across diﬀerent spacer units. In particular, thienyl units, which

are the most eﬀective at electronic delocalisation, are signiﬁ-

cantly worse than the other spacer units at mediating magnetic

exchange.

(iv) The intense NIR transitions which appear in the oxidised

forms of these complexes make them eﬀective as electrochromic

near-IR dyes in the region of the spectrum of interest for ﬁbre-

optic data transmission using silica ﬁbres.

action. The thienyl spacer of Mo₂(th) results in a weak exchange

interaction (J = Ϫ3.6 cm^Ϫ1) despite being particularly eﬀective

at mediating electronic interactions. The twisted conform-

ation of the bridging ligand in Mo₂(CO) results in poor spin

exchange (J = Ϫ1.1 cm^Ϫ1) in agreement with its poor ability to

mediate electrochemical interactions.

There is clearly not a good correlation between the ability of

the diﬀerent bridging ligands to mediate electronic interactions

in the mixed-valence Mo(VI)–Mo(V) state (by delocalisation

through a conjugated orbital system) and magnetic interaction

in the isovalent Mo(V)–Mo(V) state (for which a spin-polaris-

ation mechanism is valid in many cases).^2,4–7The two types of

interaction are presumably related to diﬀerent properties of the

bridging ligand. For those complexes of the type [Mo–O(C₆H₄)–

X–(C₆H₄)O–Mo] for which we have both electrochemical and

magnetic data, the eﬀectiveness of spacer X at transmitting

᎐

electronic interactions is th > (C᎐C ≈ C᎐C) > N᎐N > CO; for

᎐

magnetic interactions, the order is N᎐N > C᎐C > C᎐C > th >

᎐

CO. Apart from the poorest bridging ligand (dihydroxybenzo-

phenone) which is fundamentally diﬀerent from the others,

both in its twisted conformation and the odd number of atoms

in the bridge, the order of eﬀectiveness for the four spacers th,

᎐

C᎐C, C᎐C and N᎐N is reversed.

᎐

We note that for Mo (C᎐C), Mo (N᎐N) and Mo (C᎐C) the

᎐

2

presence of antiferromagnetic coupling is in agreement with

our qualitative expectations on the basis of spin polarisation

(Fig. 7).^2,5–7The presence of an even number of atoms in the

bridging pathway, whichever route is taken through the

bridging ligand, means that the alternating pattern of induced

spins will ensure that the unpaired electrons on the two Mo

fragments will be spin-opposed in the ground state. For Mo₂(th)

the situation is diﬀerent because the thienyl ring contains an

odd number of atoms: one pathway through the thienyl ring has

four C atoms which would aﬀord antiferromagnetic coupling

on the basis of spin polarisation, whereas the other pathway

has three atoms (C–S–C) which would give ferromagnetic

coupling. It is tempting to speculate that the conﬂict between

these two opposing eﬀects means that they will tend to cancel

one another out, an eﬀect which has been suggested before

in the context of electronic interactions,⁸⁸which may partly

explain why the th spacer is so poor at mediating magnetic

exchange when it is so good at mediating electronic delocalisa-

tion. A more detailed study of magnetic exchange interactions

across non-alternant bridging ligands of this type would be of

interest.

We note also that for Mo₂(CO) the spin-polarisation picture

fails: with a single additional atom in the pathway between the

two phenyl rings, spin polarisation would predict ferromagnetic

exchange. However we have found from earlier DFT calcu-

lations that other factors, such as the orientation of the Mo

fragments with respect to the bridging ligand and to each other,

also can make a signiﬁcant contribution to the sign and mag-

nitude of J.⁴These factors can over-ride spin-polarisation

eﬀects and this appears to be occurring with Mo₂(CO).

Experimental

General details

Instrumentation used for routine spectroscopic and electro-

chemical studies has been described recently.⁸⁹Spectroelectro-

chemical studies were performed in CH₂Cl₂solution at Ϫ30 ЊC

using a home-built OTTLE (optically transparent thin-layer

electrode) cell mounted in the sample compartment of a Perkin-

Elmer Lambda 19 spectrophotometer as described previously;⁸⁹

the chemical reversibility of each process was checked by revers-

ing the applied potential after electrolysis and ensuring that

the spectrum of the starting material could be regenerated.

Processes described as ‘chemically reversible’ fulﬁlled this

criterion and also aﬀorded isosbestic points in the overlaid

spectra recorded during the redox conversion.

Magnetic susceptibilities were measured in the temperature

range of 2–250 K in an applied ﬁeld of 1 T using a Metronique

Ingéniérie MS03 SQUID magnetometer; diamagnetic correc-

tions were estimated from Pascal’s constants.^90,91The values

of the parameters obtained by the ﬁtting procedure using

the exchange spin Hamiltonian H = ϪJ S₁ؒS₂(with positive

J indicating ferromagnetism and negative J indicating anti-

ferromagnetism) are included in Table 4.

[Mo(Tp^Me,Me)(O)Cl₂] was prepared according to the method

of Enemark and co-workers;²⁴2,5-bis(tri-n-butylstannyl)thio-

1410

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

View Article Online

Table 6 Crystallographic data for the two structures^a

Mo₂(CO)ؒ0.5C₅H₁₂ؒ1.5CH₂Cl₂

Mo₂(NN)

Empirical formula

M

Crystal system, space group

C₄₇H₆₁B₂Cl₅Mo₂N₁₂O₅

1264.8

Monoclinic, C2/c

C₄₂H₅₂B₂Cl₂Mo₂N₁₄O₄

1101.4

Triclinic, P1

¯

a/Å

b/Å

c/Å

α/Њ

33.059(4)

19.737(4)

18.568(2)

11.060(5)

11.916(5)

11.985(6)

119.69(2)

114.15(4)

90.46(4)

β/Њ

98.070(9)

γ/Њ

V/Å³

11996(3)

1206.9

Z

8

1

µ/mm^Ϫ1

0.692

0.687

Reﬂections collected: total/independent/R_int

Data/restraints/parameters

Final R1, wR2^b

31988, 10500, 0.0608

10500/78/747

0.0746, 0.2493

12455, 5481, 0.0899

5481/0/317

0.0592, 0.1603

2

^aData in common: T = 173 K, λ = 0.71073 Å. ^bStructure was reﬁned on F_ousing all data; the value of R1 is given for comparison with older

reﬁnements based on F_owith a typical threshold of F ≥ 4σ(F ).

phene and 5,5Љ-bis(tri-n-butylstannyl)-2,2Ј:5Ј,2Љ-terthiophene

were prepared according to the method of Miller and Yu.⁹²The

ligand H₂(th)₂was available from previous work.²⁷Organic

starting materials were purchased from the usual commercial

sources (Aldrich, Lancaster, Avocado) and used as received.

ZINDO calculations were carried out with INDO/1 param-

eters using the CAChe suite of programs on a Silicon Graphics

Indy computer.⁷⁸Structures were initially energy-minimised

using the molecular mechanics method with MM3 parameters.

The Mo–O–C–C dihedral angles were then set to 45Њ such that

13.9 mmol), dry Et₃N (2 cm³), Pd(OAc)₂(0.048 g, 0.21 mmol)

and PPh₃(0.112 g, 0.428 mmol) in a sealed Schlenk tube under

N₂was heated to 100 ЊC for 3 days.⁹⁵The resulting grey solid

was partitioned between CH₂Cl₂and water; the organic layer

was separated, washed with water, and dried over MgSO₄.

Removal of the solvent in vacuo aﬀorded a crude green solid

which was puriﬁed by column chromatography (alumina,

CH Cl ) to give 1,2-bis(4-methoxyphenyl)ethene [Me (C᎐C)] as

᎐

2

a yellow solid in 58% yield. EIMS: m/z 240 (M^ϩ). ¹H NMR (300

MHz, CDCl₃): δ 7.43 (4 H, d, J = 8.8; phenyl), 6.93 (2 H, s;

vinylic CH), 6.89 (4 H, d, J = 8.8 Hz; phenyl), 3.83 (6 H, s;

OCH₃). Found: C, 79.6; H, 7.1 (required for C₁₆H₁₆O₂: C, 80.0;

H, 6.7%).

the plane of the phenolate ring bisected the Mo᎐O and Mo–Cl

᎐

axes. This angle varies considerably in diﬀerent crystal struc-

tures and we found that the results of the calculation are quite

sensitive to this geometry as it aﬀects the Mo[d(π)]–O[p(π)]

overlap; accordingly we chose a representative geometry

and ﬁxed it for all of the molecules studied. For diamagnetic

molecules the electronic spectra calculations (restricted

Hartree–Fock, RHF) were performed with a conﬁguration

interaction level of 20; for paramagnetic species the electronic

spectra calculations (restricted open-shell Hartree–Fock,

ROHF) used a conﬁguration interaction level of 18 which is the

maximum allowed by the CAChe software.

H (C᎐C) [1,2-Bis(4-hydroxyphenyl)ethene]. Demethylation

᎐

2

of Me (C᎐C) was performed using BBr in CH Cl according

᎐

2

3

2

to a published method.³¹After quenching the reaction mixture

with water the product precipitated; the solid was ﬁltered oﬀ,

washed (water and CH Cl ) and dried leaving H (C᎐C) as a red

᎐

2

solid (45% yield). EIMS: m/z 212 (M^ϩ). H NMR (300 MHz,

CD₃OD): δ 7.33 (4 H, d, J = 8.6; phenyl), 6.88 (2 H, s; vinylic

CH), 6.75 (4 H, d, J = 8.8 Hz; phenyl).

1

H (C᎐C) [1,4-Bis(4-hydroxyphenyl)butadiene]. The methyl-

ated precursor 1,4-bis(4-methoxyphenyl)butadiene [Me₂-

X-Ray crystallography

᎐

2

Crystals of Mo₂(CO) and Mo₂(NN) were grown by diﬀusion of

hexane into concentrated solutions of the complexes in CH₂Cl₂.

In each case a suitable crystal was coated with hydrocarbon oil

and attached to the tip of a glass ﬁbre, which was then trans-

ferred to a Siemens SMART diﬀractometer under a stream of

cold N₂at 173 K. Details of the crystal parameters, data collec-

tion and reﬁnement for each of the structures are collected

in Table 6. After collection of a full sphere of data in each case

an empirical absorption correction (SADABS) was applied,⁹³

and the structures were then solved by conventional direct

methods and reﬁned on all F²data using the SHELX suite of

programs.⁹⁴In all cases, non-hydrogen atoms were reﬁned with

anisotropic thermal parameters; hydrogen atoms were included

in calculated positions and reﬁned with isotropic thermal

parameters which were ca. 1.2 × (aromatic CH) or 1.5 × (Me)

the equivalent isotropic thermal parameters of their parent

carbon atoms.

(C᎐C) ] was prepared from 4-methoxycinnamaldehyde and

᎐

2

diethyl-4-methoxybenzylphosphonate according to a published

procedure.³⁴Demethylation of this to aﬀord H (C᎐C) was

᎐

2

carried out using MeMgI as follows. To a solution of MeMgI

[prepared under N₂from Mg (0.3 g, 12 mmol) and MeI (0.75

cm³, 1.71 g, 12 mmol) in Et₂O (20 cm³)] was added with stirring

a solution of Me (C᎐C) (0.20 g, 0.75 mmol) in Et O (10 cm³).

᎐

2

The solvent was removed in vacuo, and the residue heated to

190 ЊC for 1 h. After cooling to room temperature, saturated

aqueous NH₄Cl (50 cm³) was added and the suspension further

acidiﬁed by 1 M HCl. The solid which precipitated was

extracted into Et₂O (4 × 50 cm³), dried over MgSO₄, and the

solvent evaporated to give H (C᎐C) as a brown powder in 76%

᎐

2

1

yield. EI-MS: m/z 238 (M^ϩ). H NMR (300 MHz, d⁶-DMSO):

δ 9.54 (2 H, broad; OH), 7.30 (4 H, d, J = 8.6; phenyl), 6.82

(2 H, dd, J = 12.1, 2.8; butadiene H¹/H⁴), 6.74 (4 H, d, J = 8.6;

phenyl), 6.54 (2H, dd, J = 12.1, 2.8 Hz; butadiene H²/H³).

CCDC reference numbers 157071 and 157072.

See http://ww w .rsc.org/suppdata/dt/b1/b100681i/ for crystal-

lographic data in CIF or other electronic format.

(PrOCH ) (C᎐C) [1,6-Bis{4-(propoxymethoxy)phenyl}hexa-

᎐

2

3

1,3,5-triene]. This follows a known method.³³Freshly distilled

glyme (1,2-dimethoxyethane) (50 cm³) was slowly added under

an argon atmosphere to a suspension of sodium hydride in

glyme (80%, 3.75 g, 125 mmol). A mixture of freshly distilled

4-(propoxymethoxy)benzaldehyde (11.7 g, 60 mmol) and

Syntheses of ligands

Me (C᎐C) [1,2-Bis(4-methoxyphenyl)ethene]. A mixture of

᎐

2

4-bromoanisole (2.0 g, 10.7 mmol), 4-methoxystyrene (1.87 g,

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

1411

View Article Online

1,4-but-2-enediylbis(diethyl phosphonate)³³(9.90 g, 30 mmol)

in glyme (25 cm³) was added dropwise at room temperature.

The reaction is exothermic during the initial addition. The

reaction mixture was stirred at room temperature for 72 h and

then heated under reﬂux for 2 h. After cooling, ⁱPrOH (10 cm³)

was added to destroy unchanged NaH. The dark brown suspen-

sion was then poured onto ice–water (700 cm³) to give a yellow

ﬂaky solid. After ﬁltration and drying in vacuo, the crude prod-

uct was dissolved in CH₂Cl₂, puriﬁed by ﬂash chromatography

(alumina, CH₂Cl₂) and recrystallised from methylcyclohexane.

Yield: 35%. EIMS: m/z 408 (M^ϩ). ¹H NMR (300 MHz, CDCl₃):

δ 0.90 (6 H, t, J = 7.4; CH₃), 1.60 (4 H, m, CH₃CH₂), 3.62 (4 H,

t, J = 6.6; CH₂CH₂O), 5.22 (4 H, s; OCH₂C₆H₄), 6.50 (4 H, m;

alkenyl), 6.75 (2 H, m; alkenyl), 6.99 (4 H, d, J = 8.7; phenyl),

7.33 (4 H, d, J = 8.7 Hz; phenyl). Found: C, 76.6; H, 7.4

(required for C₂₆H₃₂O₄: C, 76.5; H, 7.8%).

methoxystyrene (2.25 g, 16.8 mmol), Pd(OAc)₂(0.027 g, 0.12

mmol), PPh₃(0.063 g, 0.24 mmol) and dry Et₃N (2.5 cm³) in a

sealed Schlenk tube under N₂was heated to 100 ЊC for 3 days,

aﬀording a green solid.⁶⁸This was suspended in water, which

dissolved the [Et₃NH]Br and inorganic salts, and ﬁltered; the

resulting grey solid was washed copiously with water and then

with a small amount of CH₂Cl₂to give nearly pure Me₂(C₂ꢀC₂)

in 32% yield. EIMS: m/z 342 (M^ϩ). ¹H NMR (300 MHz,

CDCl₃): δ 7.46 (4 H, d, J = 8.4 Hz; phenyl), 7.30 (4 H, m;

phenyl), 7.10 (2 H, m; vinylic CH), 6.90 (6 H, m; 4 phenyl and

2 vinylic CH). Found: C, 84.2; H, 6.8 (required for C₂₄H₂₂O₂: C,

84.2; H, 6.4%).

H₂(C₂ꢀC₂)

{1,4-Bis[2-(4-hydroxyphenyl)ethenyl]benzene}.

Demethylation of Me₂(C₂ꢀC₂) was performed with BBr₃in

CH₂Cl₂according to a published method.³¹After quenching the

reaction by addition of water, a precipitate formed which was

ﬁltered oﬀ, washed with water and CH₂Cl₂, and dried to give

H₂(C₂ꢀC₂) as a grey solid in 50% yield. EIMS: m/z 314 (M^ϩ). It

is too insoluble in common solvents for NMR spectroscopy.

H (C᎐C) [1,6-Bis(4-hydroxyphenyl)hexatriene]. To a stirred

᎐

2

3

solution of (PrOCH ) (C᎐C) (0.753 g, 1.9 mmol) in MeOH–

᎐

2

3

thf (6 cm³; 1 : 2, v/v) was added concentrated HCl (1 cm³).

The mixture was stirred overnight and then water was added,

resulting in formation of a yellow precipitate which was

collected by ﬁltration, washed with further water, and dried.

t

᎐

Me (C᎐C) [1,2-Bis(4-methoxyphenyl)ethyne]. [ Bu N][HSO ]

᎐

2

4

(0.115 g, 0.34 mmol) and KOH (0.673 g, 12 mmol) were ground

up together and added under N₂to a stirred mixture of CuI

(0.190 g, 1.0 mmol), PPh₃(0.393 g, 1.5 mmol), PdCl₂(PPh₃)₂

(0.197 g, 0.28 mmol) and piperidine (1 cm³). To this was added

a solution of 2-methyl-3-butyn-2-ol (0.9 cm³, 0.78 g, 9.3 mmol)

and 4-iodoanisole (3.93 g, 17 mmol) in benzene–thf (9 : 1,

70 cm³), and the resultant mixture heated to 80 ЊC for 24 h.

After cooling to room temperature, saturated aqueous NH₄Cl

(150 cm³) was added, and the mixture stirred for 1 h and then

extracted with toluene (4 × 100 cm³). The combined organic

extracts were dried over MgSO₄, ﬁltered, and the solvent was

removed in vacuo. The crude product was dissolved in hexane

and once more ﬁltered and evaporated to dryness. Pure pale

yellow 1,2-bis(4-methoxyphenyl)ethyne was produced by

recrystallisation from hot ethanol (61% yield). EI-MS: m/z 238

Recrystallisation from hot dmf–water aﬀorded pure H (C᎐C)

᎐

2

3

as a yellow crystalline powder in 89% yield. EIMS: m/z 264

1

(M^ϩ). H NMR (300 MHz, d⁶-DMSO): δ 9.62 (2 H, broad;

OH), 7.33 (4 H, d, J = 8.6; phenyl), 6.86–6.73 (6 H, m; four

phenyl H and hexatriene H²/H⁵), 6.54 (2 H, d, J = 15.4; hexa-

triene H¹/H⁶), 6.49 (2 H, dd, J = 7.0, 3.0 Hz; hexatriene H³/H⁴).

(ⁱPr Si) (N᎐N) [4,4Ј-Bis(triisopropylsiloxy)azobenzene]. To an

᎐

3

2

ice-cold solution of 4-aminophenol (2.55 g, 23.4 mmol) in dry

thf (150 cm³) under N₂was added NaH (60% dispersion in

mineral oil: 1.2 g, 30 mmol) which resulted in the solution

becoming cloudy. Triisopropylsilyl chloride (4.51 g, 23.4 mmol)

was added and the mixture stirred overnight at room temper-

ature. Unchanged NaH was destroyed by addition of MeOH

(10 cm³); the solvents were removed in vacuo leaving a brown

liquid which was partitioned between CH₂Cl₂and water. The

organic layer was separated, washed with water, and dried over

MgSO₄. The resulting brown oil [crude 4-(triisopropyl-

siloxy)aniline] was dissolved in toluene (200 cm³) and MnO₂(8

g, a large excess) was added.⁹⁶The mixture was stirred at room

temperature overnight and became deep red. After removal of

MnO₂by ﬁltration, evaporation of the solvent in vacuo left a

dark red oil, which was puriﬁed by column chromatography

(alumina, CH₂Cl₂) to give 4,4Ј-bis(triisopropylsiloxy)azobenz-

ene as a dark orange oil which solidiﬁed on cooling (34% yield

1

(M^ϩ). H NMR (300 MHz, CDCl₃): δ 7.45 (4 H, d, J = 9.0;

phenyl), 6.87 (4 H, d, J = 9.0 Hz; phenyl), 3.82 (6H, s, CH₃).

᎐

H (C᎐C) [1,2-Bis(4-hydroxyphenyl)ethyne]. Demethylation

᎐

2

᎐

of Me (C᎐C) was performed with BBr as follows. To a solution

of Me (C᎐C) (0.200 g, 0.84 mmol) in CH Cl (5 cm ) was added

᎐

2

3

᎐

2

BBr₃(1.0 M solution in CH₂Cl₂: 1.6 cm³, 1.6 mmol). The deep

purple solution so obtained was stirred for 24 h and then

poured into ice–water (250 cm³). The suspended solid which

formed was extracted into Et₂O (3 × 150 cm³); and then the

combined organic fractions were re-extracted with aqueous

NaOH (2 M, 2 × 150 cm³). The basic solution was acidiﬁed by

dropwise addition of concentrated HCl to pH 2, and then

extracted a ﬁnal time with Et₂O (3 × 150 cm³). These organic

fractions were then dried over MgSO₄and the solvent was

1

based on aminophenol). EIMS: m/z 526 (M^ϩ). H NMR (300

MHz, CDCl₃): δ 7.80 (4 H, d, J = 9.0; phenyl), 6.97 (4 H, d,

i

J = 9.0 Hz; phenyl), 1.1–1.3 (42 H, m; Pr₃Si). Found: C, 68.8;

H, 9.6; N, 4.9 (required for C₃₀H₅₀N₂O₂Si₂: C, 68.4; H, 9.5; N,

5.3%).

evaporated to produce a red-brown solid, which after recrys-

᎐

tallisation from hot aqueous EtOH gave H (C᎐C) in 53% yield.

᎐

2

H (N᎐N) (4,4Ј-Azophenol). To a solution of (ⁱPr Si) (N᎐N)

EI-MS: m/z 210 (M^ϩ, 100%). H NMR (300 MHz, CD₃OD):

1

᎐

2

3

2

(0.81 g, 1.54 mmol) in dry thf (50 cm³) was added an excess of

tetra(n-butyl)ammonium ﬂuoride under N₂at room temper-

ature. On addition of the ﬂuoride the orange solution became

dark brown, and after 20 minutes a yellow precipitate formed.

The mixture was stirred for 30 minutes, after which time the

yellow solid was ﬁltered oﬀ, washed with thf, and dried. The

crude material was puriﬁed by column chromatography (silica,

10 : 1 ethyl acetate–CH₂Cl₂) to aﬀord pure 4,4Ј-azophenol in

43% yield. EIMS: m/z 214 (M^ϩ). ¹H NMR (300 MHz, CD₃OD):

δ 7.74 (4 H, d, J = 9.2; phenyl), 6.89 (4 H, d, J = 9.1 Hz; phenyl).

Found: C, 67.1; H, 4.7; N, 13.0 (required for C₁₂H₁₀N₂O₂: C,

67.3; H, 4.7; N, 13.1%).

δ 7.29 (4 H, d, J = 8.8; phenyl), 6.75 (4 H, d, J = 8.8 Hz; phenyl).

Me₂th [2,5-Bis(4-methoxyphenyl)thiophene]. To a solution of

2,5-bis(tri-n-butylstannyl)thiophene (4.35 g, 6.6 mmol) in dry

toluene (50 cm³) under N₂was added Pd(PPh₃)₄(0.10 g,

87 µmol) and 4-bromoanisole (2.25 cm³, 3.36 g, 18 mmol); the

mixture was stirred at 100 ЊC for 24 h. The cooled mixture was

poured into sat. aqueous NH₄Cl (100 cm³) and the two phases

were separated. The aqueous layer was further extracted with

CH₂Cl₂(3 × 100 cm³), the combined organic extracts were dried

over MgSO₄, and the solvent was evaporated to leave an oily

yellow solid. The was dissolved in acetone and reprecipitated by

addition of MeOH, leaving 2,5-di(4-methoxyphenyl)thiophene

as a yellow-green powder in 26% yield. EIMS: m/z 296 (M^ϩ). ¹H

NMR (300 MHz, CDCl₃): δ 7.55 (4 H, dd, J = 6.8, 2.2; phenyl),

Me₂(C₂ꢀC₂) {1,4-Bis[2-(4-methoxyphenyl)ethenyl]benzene}.

A mixture of 1,4-dibromobenzene (1.42 g, 6.0 mmol), 4-

1412

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

View Article Online

7.15 (2H, s; thienyl), 6.93 (4 H, dd, J = 6.6, 2.2 Hz; phenyl), 3.84

(6H, s; CH₃). Found: C, 72.6; H, 5.0 (required for C₁₈H₁₆O₂S: C,

73.0; H, 5.4%).

and gave correct molecular ions in their FAB mass spectra;

elemental analyses were generally consistent with incorpor-

ation of some recrystallisation solvent, as detailed in the foot-

notes to Table 1. The sole exception to this was Mo₂(C₂ꢀC₂)

for which we could not obtain reliable analytical data; we have

encountered this problem before with complexes of highly

extended aromatic ligands.

H₂th [2,5-Bis(4-hydroxyphenyl)thiophene]. Demethylation of

Me₂th to give H₂th was accomplished with pyridine hydro-

chloride, as follows. Pyridine (technical grade, 10 cm³) and

conc. HCl (10 cm³) were stirred together and heated to 200 ЊC

in an open reaction ﬂask under a stream of N₂. After ca. 1.5 h

solid pyridinium chloride began to sublime around the neck of

the ﬂask. At this point Me₂th (0.20 g, 0.68 mmol) was added

and the ﬂask sealed. Heating and stirring were continued for

3 h. The reaction was allowed to cool and hot water (10 cm³)

added. The resultant mixture was poured into cold water (50

cm³) and after cooling to room temperature the precipitate was

collected by ﬁltration. Recrystallisation of the crude material

by addition of diethyl ether to an acetone solution of the solid

aﬀorded H₂th as a ﬂuﬀy yellow solid (54% yield). EIMS: m/z

Acknowledgements

We thank the European Community TMR Network pro-

gramme (contract no. EC-CHRX-CT96-0047) and the EPSRC

(UK) for ﬁnancial support. M.D.W. is the Royal Society of

Chemistry Sir Edward Frankland fellow for 2000/2001.

References

1 J. A. McCleverty and M. D. Ward, Acc. Chem. Res., 1998, 36, 3447.

2 S. Bayly, J. A. McCleverty, M. D. Ward, D. Gatteschi and F. Totti,

Inorg. Chem., 2000, 39, 1288.

3 N. C. Harden, E. R. Humphrey, J. C. Jeﬀery, S.-M. Lee,

M. Marcaccio, J. A. McCleverty, L. H. Rees and M. D. Ward,

J. Chem. Soc., Dalton Trans., 1999, 2417.

4 A. Bencini, D. Gatteschi, F. Totti, D. N. Sanz, J. A. McCleverty and

M. D. Ward, J. Phys. Chem. A, 1999, 102, 10545.

5 V. A. Ung, S. M. Couchman, J. C. Jeﬀery, J. A. McCleverty, M. D.

Ward, F. Totti and D. Gatteschi, Inorg. Chem., 1999, 38, 365.

6 V. A. Ung, A. M. W. Cargill Thompson, D. A. Bardwell,

D. Gatteschi, J. C. Jeﬀery, J. A. McCleverty, F. Totti and M. D.

Ward, Inorg. Chem., 1997, 36, 3447.

7 A. M. W. Cargill Thompson, D. Gatteschi, J. A. McCleverty, J. A.

Navas, E. Rentschler and M. D. Ward, Inorg. Chem., 1996, 35, 2701.

8 V. A. Ung, D. A. Bardwell, J. C. Jeﬀery, J. P. Maher, J. A.

McCleverty, M. D. Ward and A. Williamson, Inorg. Chem., 1996,

35, 5290.

9 A. Wlodarczyk, J. P. Maher, J. A. McCleverty and M. D. Ward,

J. Chem. Soc., Dalton Trans., 1997, 3287.

1

268 (M^ϩ). H NMR (300 MHz, d⁶-acetone): δ 7.38 (4 H, dd,

J = 6.6, 2.2; phenyl), 7.09 (2 H, s; thienyl), 6.76 (4 H, dd, J = 6.6,

2.2 Hz; phenyl).

Me₂(th)₃

[5,5Љ-Bis(4-methoxyphenyl)-2,2Ј:5Ј,2Љ-terthio-

phene]. To a solution of 5,5Љ-bis(tri-n-butylstannyl)-2,2Ј:5Ј,2Љ-

terthiophene (1.96 g, 2.4 mmol) in dry toluene (25 cm³) under

N₂was added Pd(PPh₃)₄(65 mg, 56 µmol) and 4-bromoanisole

(0.50 cm³, 0.75 g, 4.0 mmol). The mixture was stirred and

heated to 100 ЊC for 3 h. The cooled mixture was quenched

by addition of water (40 cm³) and the green-yellow precipitate

collected by ﬁltration. The ﬁlter-cake was slurried in thf (100

cm³), ﬁltered, and the residue extracted with further thf (700

cm³). Evaporation of the solvent in vacuo left a brown residue

which was washed with CH₂Cl₂(100 cm³) to leave 5,5Љ-bis(4-

methoxyphenyl)-2,2Ј:5Ј,2Љ-terthiophene in 46% yield. EIMS:

m/z 460 (M^ϩ). Found: C, 67.2; H, 4.4 (required for C₂₆H₂₀O₂S₃:

C, 67.8; H, 4.2%).

10 M. D. Ward, Chem. Soc. Rev., 1995, 24, 121.

11 M. D. Ward, Chem. Ind., 1996, 568.

12 J. R. Reimers and N. S. Hush, Inorg. Chem., 1990, 29, 3686.

13 A.-C. Ribou, J.-P. Launay, M. L. Sachtleben, H. Li and C. W.

Spangler, Inorg. Chem., 1996, 35, 3735.

14 C. E. B. Evans, M. L. Naklicki, A. R. Revzani, C. A. White, V. V.

Kondratiev and R. J. Crutchley, J. Am. Chem. Soc., 1998, 120,

13096.

15 F. Lloret, G. De Munno, M. Julve, J. Cano, R. Ruiz and

A. Caneschi, Angew. Chem., Int. Ed., 1998, 37, 135.

16 O. Kahn, Molecular Magnetism, VCH publishers, Inc., New York,

1993.

17 T. C. Brunold, D. R. Gamelin and E. I. Solomon, J. Am. Chem. Soc.,

2000, 122, 8511.

18 S. Benard, P. Yu, J.-P. Audiere, E. Riviere, R. Clement, J. Guilhem,

L. Tchernatov and K. Nakatani, J. Am. Chem. Soc., 2000, 122, 9444.

19 A. Escuer, J. Cano, M. A. S. Goher, Y. Journaux, F. Lloret, F. A.

Mautner and R. Vicente, Inorg. Chem., 2000, 39, 4688.

20 C. Ruiz-Perez, M. Hernandez-Molina, P. Lorenzo-Luis, F. Lloret,

J. Cano and M. Julve, Inorg. Chem., 2000, 39, 3845.

21 S. Triki, F. Berezovsky, J. S. Pala, E. Coronado, C. J. Gomez-Garcia,

J. M. Clemente, A. Riou and P. Molinie, Inorg. Chem., 2000, 39,

3771.

22 Y. Cui, G. Chen, J. Ren, Y. T. Qian and J. S. Huang, Inorg. Chem.,

2000, 39, 4165.

23 Q. D. Liu, S. Gao, J. R. Li, Q. Z. Zhon, K. B. Yu, B. Q. Ma, S. W.

Zhang, X. X. Zhang and T. Z. Jin, Inorg. Chem., 2000, 39, 2488.

24 W. E. Cleland, Jr., K. M. Barhhart, K. Yamanouchi, D. Collison,

F. E. Mabbs, R. B. Ortega and J. H. Enemark, Inorg. Chem., 1987,

26, 1017.

H₂(th)₃[5,5Љ-Bis(4-hydroxyphenyl)-2,2Ј:5Ј,2Љ-terthiophene].

Demethylation of Me₂(th)₃(0.20 g, 0.43 mmol) to give H₂(th)₃

was carried out with molten pyridinium chloride, exactly

according to the method described above for H₂th except that

the reaction temperature was maintained for 5 h. After cooling

to below 100 ЊC, hot water (10 cm³) was added and the mixture

then poured into cold water (50 cm³). The resulting suspension

was saturated with NaCl and extracted with Et₂O–thf (3 : 1,

4 × 200 cm³). The cloudy organic extracts were ﬁltered, dried

over MgSO₄, ﬁltered once more, and evaporated to dryness to

leave H₂(th)₃as a green solid in 57% yield. EI-MS: m/z 432

(M^ϩ).

Syntheses of complexes

All complexes were prepared by the same general method. A

mixture of the bridging ligand (typically 0.3 mmol) and dry

Et₃N (1 cm³) in dry toluene under N₂was heated to reﬂux for

15 minutes. After this time, [Mo(Tp^Me,Me)(O)Cl₂] (0.7 mmol,

slightly over two equivalents) was added and the mixture heated

at reﬂux under N₂until TLC analysis (silica, CH₂Cl₂) showed

that no further change occurred. Typically this took 2 hours,

but with the more insoluble bridging ligands such as H₂(th)₃this

could take up to 5 hours. After removal of the solvent in vacuo,

the crude residue was puriﬁed by column chromatography on

silica using CH₂Cl₂–hexane mixtures, varying from pure

CH₂Cl₂to 1 : 1 CH₂Cl₂–hexane according to how fast the

principal fraction eluted. After collection of the main fraction

and removal of solvents in vacuo, the complexes were puriﬁed

by recrystallisation from CH₂Cl₂, by slow diﬀusion of either

hexane or MeOH into the CH₂Cl₂solution of the material.

Yields and characterisation data are collected in Table 1. All

complexes were chromatographically pure by TLC analysis

25 S.-M. Lee, M. Marcaccio, J. A. McCleverty and M. D. Ward, Chem.

Mater., 1998, 10, 3272.

26 A. McDonagh, S. R. Bayly, D. J. Riley, M. D. Ward, J. A.

McCleverty, M. A. Cowin, C. N. Morgan, R. Varrazza, R. V. Penty

and I. H. White, Chem. Mater., 2000, 12, 2523.

27 J. Hock, A. M. W. Cargill Thompson, J. A. McCleverty and M. D.

Ward, J. Chem. Soc., Dalton Trans., 1996, 4257.

28 A.-C. Ribou, J.-P. Launay, K. Takahashi, T. Nihira, S. Tarutani and

C. W. Spangler, Inorg. Chem., 1994, 33, 1325.

29 P. Bäuerle, T. Fischer, B. Bidlingmeier, A. Stabel and J. P. Rabe,

Angew. Chem., Int. Ed. Engl., 1995, 34, 303.

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

1413

View Article Online

Harding and M. G. Hutchings, J. Chem. Soc., Chem. Commun.,

1992, 1796.

30 J. Guay, A. Diaz, R. Wu and J. M. Tour, J. Am. Chem. Soc., 1993,

115, 1869.

31 K. Takahashi, T. Suzuki, K. Akiyama, Y. Ikegami and Y. Fukazawa,

J. Am. Chem. Soc., 1991, 113, 4576.

32 W. H. Laarhoven, R. J. F. Nivard and E. Havinga, Recl. Trav. Chim.

Pays-Bas, 1961, 80, 775.

33 C. W. Spangler, R. K. McCoy, A. A. Dembek, L. S. Sapochak and

B. D. Gates, J. Chem. Soc., Perkin Trans. 1, 1989, 151.

34 K. Friedrich and W. Hartmann, Chem. Ber., 1961, 94, 840.

35 M. Takahasi, Y. Shioura, T. Murakami and K. Ogasawara,

Tetrahedron Asymmetry, 1997, 8, 1235.

36 M. Santelli and J.-M. Pons, Lewis Acids and Selectivity in Organic

Synthesis, CRC press, Boca Raton, FL, 1996.

37 P. J. Kocienski, Protecting Groups, Georg Thieme, Stuttgart, 1994.

38 J. Trpin and B. G. Zupancic, Monatsh. Chem., 1969, 100, 114.

39 R. J. Kaufman and R. S. Sidhu, J. Org. Chem., 1982, 47, 4941.

40 K. Sonogashira, Y. Tohda and N. Hagihara, Tetrahedron Lett., 1975,

50, 4467.

41 D. E. Ames, D. Bull and C. Takundwa, Synthesis, 1981, 364.

42 C. Pugh and V. Percec, Mol. Cryst. Liq. Cryst., 1990, 178, 193.

43 M. Alami, F. Ferri and G. Linstrumelle, Tetrahedron Lett., 1993, 34,

6403.

44 P. K. A. Shonﬁeld, A. Behrendt, J. C. Jeﬀery, J. P. Maher, J. A.

McCleverty, E. Psillakis, M. D. Ward and C. Western, J. Chem. Soc.,

Dalton Trans., 1999, 4341.

45 H. M. Doesburg and J. H. Noordik, Cryst. Struct. Commun., 1979,

8, 377.

63 A. J. Amoroso, A. M. W. Cargill Thompson, J. P. Maher, J. A.

McCleverty and M. D. Ward, Inorg. Chem., 1995, 34, 4828.

64 A. Das, J. P. Maher, J. A. McCleverty, J. A. Navas Badiola and

M. D. Ward, J. Chem. Soc., Dalton Trans., 1993, 681.

65 J. E. Sutton and H. Taube, Inorg. Chem., 1981, 20, 3125.

66 L. T. Cheng, W. Tam, S. H. Stevenson, G. R. Meredith, G. Rikken

and S. R. Marder, J. Phys. Chem., 1991, 95, 10631.

67 J. Zhou and A. Rieker, J. Chem. Soc., Perkin Trans. 2, 1997, 931.

68 K. Takahashi, A. Gunji, K. Yanagi and M. Miki, J. Org. Chem.,

1996, 61, 4784.

69 W. J. Detroit and H. Hart, J. Am. Chem. Soc., 1952, 74, 5215.

70 O. B. Lantratova, A. I. Prokof’ev, I. V. Khudyakov, V. A. Kuzmin

and I. F. Pokrovskaya, Nouv. J. Chim., 1982, 6, 365.

71 R. West, J. A. Jorgensen, K. L. Stearley and J. C. Calabrese, J. Chem.

Soc., Chem. Commun., 1991, 1234.

72 P. Boldt, D. Bruhnke, F. Gerson, M. Scholz, P. G. Jones and F. Bär,

Helv. Chim. Acta, 1993, 76, 1739.

73 A. Rebmann, J. Zhou, P. Schuler, H. B. Stegmann and A. Rieker,

J. Chem. Res. (M), 1996, 1765.

74 M. Kurihara, I. Saito and Y. Matsuda, Chem. Lett., 1996, 1109.

75 J. Bourdon and M. Calvin, J. Org. Chem., 1957, 22, 101.

76 K. Dimroth, W. Umbach and K. H. Blöcher, Angew. Chem., Int. Ed.

Engl., 1963, 2, 620.

77 See:

http://btcpxx.che.uni-bayreuth.de/COMPUTER/Software/

MSI/insight972/zindo/ZindoTOC.doc.html.

46 G. P. M. van der Velden and J. H. Noordik, J. Cryst. Mol. Struct.,

1980, 10, 83.

47 G. M. Lobanova, Kristallograﬁya, 1968, 13, 984.

48 M. More, G. Odou and J. Lefebvre, Acta Crystallogr., Sect. B, 1987,

43, 398.

49 L. F. Joulié, E. Schatz, M. D. Ward, F. Weber and L. J. Yellowlees,

J. Chem. Soc., Dalton Trans., 1994, 799.

50 M. D. Ward, Inorg. Chem., 1996, 35, 1712.

78 CAChe version 3.2, Oxford Scientiﬁc Ltd, Oxford, 1999.

79 C. W. Lange and C. G. Pierpont, Inorg. Chim. Acta, 1997, 263, 219.

80 S. K. Larsen, C. G. Pierpont, G. De Munno and G. Dolcetti, Inorg.

Chem., 1986, 25, 4828.

81 M. E. Cass, N. R. Gordon and C. G. Pierpont, Inorg. Chem., 1986,

25, 3962.

82 N. Takvoryan, K. Faramery, V. Katovic, F. V. Lovecchio, E. S. Gore,

L. B. Anderson and D. H. Busch, J. Am. Chem. Soc., 1974, 96, 731.

83 B. de Bruin, E. Bill, E. Bothe, T. Weyhermüller and K. Wieghardt,

Inorg. Chem., 2000, 39, 2936.

51 A. M. Barthram, R. L. Cleary, J. C. Jeﬀery, S. M. Couchman and

M. D. Ward, Inorg. Chim. Acta, 1998, 267, 1.

52 A. M. Barthram, R. L. Cleary, R. Kowallick and M. D. Ward,

Chem. Commun., 1998, 2695.

53 A. M. Barthram and M. D. Ward, New J. Chem., 2000, 24, 501.

54 A. M. Barthram, Z. R. Reeves, J. C. Jeﬀery and M. D. Ward,

J. Chem. Soc., Dalton Trans., 2000, 3162.

84 O. S. Jung, D. H. Jo, Y. A. Lee, Y. S. Sohn and C. G. Pierpont, Inorg.

Chem., 1998, 37, 5875.

85 A. S. Attia and C. G. Pierpont, Inorg. Chem., 1998, 37, 3051.

86 O. S. Jung, D. H. Jo, Y. A. Lee, B. J. Conklin and C. G. Pierpont,

Inorg. Chem., 1997, 36, 19.

55 S. I. Gorelsky, E. S. Dodsworth, A. B. P. Lever and A. A. Vlcek,

Coord. Chem. Rev., 1998, 174, 469.

87 O. S. Jung, D. H. Jo, Y. A. Lee, Y. S. Sohn and C. G. Pierpont,

Angew. Chem., Int. Ed. Engl., 1996, 35, 1694.

88 V. Marvaud, J.-P. Launay and C. Joachim, Chem. Phys., 1993, 177,

23.

89 S.-M. Lee, R. Kowallick, M. Marcaccio, J. A. McCleverty and

M. D. Ward, J. Chem. Soc., Dalton Trans., 1998, 3443.

90 C. J. O’Connor, Prog. Inorg. Chem., 1982, 29, 203.

91 R. L. Carlin, Magnetochemistry, Springer Verlag, New York, 1986.

92 L. L. Miller and Y. Yu, J. Org. Chem., 1995, 60, 6813.

93 G. M. Sheldrick, SADABS, A program for absorption correction

with the Siemens SMART area-detector system, University of

Göttingen, 1996.

56 A. B. P. Lever, H. Masui, R. A. Metcalfe, D. J. Stufkens, E. S.

Dodsworth and P. R. Auburn, Coord. Chem. Rev., 1993, 125, 317.

57 R. S. da Silva, S. I. Gorelsky, E. S. Dodsworth, E. Tfouni and

A. B. P. Lever, J. Chem. Soc., Dalton Trans., 2000, 4078.

58 J. Guay, A. Diaz, R. Wu and J. M. Tour, J. Am. Chem. Soc., 1993,

115, 1869.

59 J. Guay, P. Kasai, A. F. Diaz, R. Wu, J. M. Tour and L. H. Dao,

Chem. Mater., 1992, 4, 1097.

60 M. G. Hill, J.-F. Penneau, B. Zinger, K. R. Mann and L. L. Miller,

Chem. Mater., 1992, 4, 1106.

61 K. Takahashi and T. Suzuki, J. Am. Chem. Soc., 1991, 111, 5483.

62 This behaviour contrasts with a linear decrease in ∆E between two

Mo(I)–Mo(0) couples across a bis(4-pyridyl)polyene ligand series,

with up to 5 double bonds in the bridging pathway: J. A. Thomas,

C. J. Jones, J. A. McCleverty, D. Collison, F. E. Mabbs, C. J.

94 G. M. Sheldrick, SHELXS 97 and SHELXL 97, programs for

crystal structure solution and reﬁnement, University of Göttingen,

1997.

95 R. F. Heck, Org. React., 1981, 27, 345.

96 I. Bhatnagar and M. V. George, J. Org. Chem., 1968, 33, 2407.

1414

J. Chem. Soc., Dalton Trans., 2001, 1401–1414

Article Doi

Electronic and magnetic metal-metal interactions in dinuclear oxomolybdenum(V) complexes across bis-phenolate bridging ligands with different spacers between the phenolate termini: Ligand-centred vs. metal-centred redox activity

DOI: 10.1039/b100681i

Source and publish data:

Authors:

Article abstract of DOI:10.1039/b100681i

Full text of DOI:10.1039/b100681i

Products guided by the article

R&D Labs maybe for 352462-92-1

Relevant to this article

Hot Product