Linear Free Energy Relationships in Dinuclear Compounds. 2. Inductive Redox Tuning via Remote Substituents in Quadruply Bonded Dimolybdenum Compounds

Article abstract of DOI:10.1021/ic960555o

Syntheses and characterizations are reported for dimolybdenum(II) compounds supported by the diarylformamidinate (ArNC(H)NAr^-) ligand, where Ar is XC₆H₄^-, with X as p-OMe (1), H (2), m-OMe (3), p-Cl (4), m-Cl (5), m-CF₃ (6), p-COMe (7), p-CF₃ (8), or Ar is 3,4-Cl₂C₆H₃^- (9) or 3,5-Cl₂C₆H₃^- (10). The (quasi)reversible oxidation potentials measured for the Mo₂⁵⁺/Mo₂⁴⁺ couple were found to correlate with the Hammett constant (ax) of the aryl substituents according to the following equation: ΔE_1/2 = E_1/2(X) - E_1/2(H) = 87(8σ_x) mV. Molecular structure determinations of compounds 1,2,5, and 10 revealed an invariant core geometry around the Mo₂ center, with statistically identical Mo-Mo quadruple bond lengths of 2.0964(5), 2.0949[8], 2.0958(6), and 2.0965(5) A?, respectively. Magnetic anisotropies for compounds 1-10 estimated on the basis of ¹H NMR data were similar and unrelated to σ_x. Similarity in UV-vis spectra was also found within the series, which, in conjunction with the features of both molecular structures and ¹H NMR spectra, was interpreted as the existence of a constant upper valence structure across the series. Results of Fenske-Hall calculations performed for several model compounds paralleled the experimental observations.

Full text of DOI:10.1021/ic960555o

6422
Inorg. Chem. 1996, 35, 6422-6428
Linear Free Energy Relationships in Dinuclear Compounds. 2.† Inductive Redox Tuning
via Remote Substituents in Quadruply Bonded Dimolybdenum Compounds
Chun Lin,‡ John D. Protasiewicz,§ Eugene T. Smith,‡ and Tong Ren*,‡
Departments of Chemistry, Florida Institute of Technology, Melbourne, Florida 32901,
and Case Western Reserve University, Cleveland, Ohio 44106
ReceiVed May 16, 1996X
Syntheses and characterizations are reported for dimolybdenum(II) compounds supported by the diarylformamidinate
(ArNC(H)NAr-) ligand, where Ar is XC6H4-, with X as p-OMe (1), H (2), m-OMe (3), p-Cl (4), m-Cl (5),
m-CF3 (6), p-COMe (7), p-CF3 (8), or Ar is 3,4-Cl2C6H3- (9) or 3,5-Cl2C6H3- (10). The (quasi)reversible oxidation
4+
potentials measured for the Mo25+/Mo2 couple were found to correlate with the Hammett constant (σX) of the
aryl substituents according to the following equation: ∆E1/2 ) E1/2(X) - E1/2(H) ) 87(8σX) mV. Molecular
structure determinations of compounds 1, 2, 5, and 10 revealed an invariant core geometry around the Mo2 center,
with statistically identical Mo-Mo quadruple bond lengths of 2.0964(5), 2.0949[8], 2.0958(6), and 2.0965(5) Å,
1
respectively. Magnetic anisotropies for compounds 1-10 estimated on the basis of H NMR data were similar
and unrelated to σX. Similarity in UV-vis spectra was also found within the series, which, in conjunction with
1
the features of both molecular structures and H NMR spectra, was interpreted as the existence of a constant
upper valence structure across the series. Results of Fenske-Hall calculations performed for several model
compounds paralleled the experimental observations.
Introduction
structure of the dinuclear core plays a key role in these
applications and the ability of fine-tuning the electronic structure
of the dinuclear core may lead to even greater success for the
development of new compounds involved in the above-
mentioned applications.
Since the seminal discovery of the Re-Re quadruple bond
in Re2Cl82-,1 the chemistry of dinuclear compounds containing
metal-metal bond has evolved into an important field of
inorganic chemistry.2 Both the nature of metal-metal interac-
tions and their dependence on the coordinating ligands have
been elaborated through an extensive study of their syntheses
and structural characterization and their spectroscopic, electronic,
and magnetic properties over the past three decades. Mean-
while, efforts focusing on technologically important applications
of dinuclear compounds have led to many promising research
areas, such as inorganic liquid crystals,3 antitumor agents,4
materials based on the extended structures5 and homogeneous6,7
and photolytic catalysis.8 There is no doubt that the electronic
Fine-tuning of the electronic structure is a familiar theme in
the coordination chemistry of mononuclear compounds, espe-
cially in the design and optimization of homogeneous catalysts9
and novel materials.10 Varying the substituents away from the
first coordination sphere (remote substituents) is particularly
enticing since the energy levels of frontier orbitals may be
significantly altered with a minimum change in both the
coordination geometry and the distribution of metal-based
valence orbitals. Furthermore, the degree of tuning can be
quantified by correlating the redox potentials of the compounds
with the Hammett constant (σ) of the substituents,11 which
should enable further fine-tuning by exploiting the well-
established linear free energy relationship (LFER) in organic
chemistry.12 Substituent tuning in metalloporphyrins has been
extensively studied13 and proven to be critical in developing
degradation-resistant metalloporphyrins as the cyt P450 mimics
in hydrocarbon oxygenation.14 Controlling the electronic rich-
ness of square planar Pd organometallic moieties through remote
substituents has attracted intense interest recently,15,16 and the
ramification of such control on kinetics is exemplified through
† For part 1, see ref 23.
‡ Florida Institute of Technology.
§ Case Western Reserve University.
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) (a) Cotton, F. A. Inorg. Chem. 1965, 4, 334. (b) Cotton, F. A.; Harris,
C. B. Inorg. Chem. 1965, 4, 330.
(2) Cotton, F. A.; Walton, R. A. Multiple Bonds between Metal Atoms;
Oxford University Press: Oxford, 1993.
(3) (a) Giroud-Godquin, A.-M.; Maitlis, P. M. Angew. Chem., Int. Ed.
Engl. 1991, 30, 375. (b) Baxter, D. V.; Cayton, R. H.; Chisholm, M.
H.; Huffman, J. C.; Putilina, E. F.; Tagg, S. L.; Wesemann, J. L.;
Zwanziger, J. W.; Darrington, F. D. J. Am. Chem. Soc. 1994, 116,
4551. (c) Bonnet, L.; Cukiernik, F. D.; Maldivi, P.; Giroud-Godquin,
A.-M.; Marchon, J.-C.; Ibn-Elhaj, M.; Guillon, D.; Skoulios, A. Chem.
Mater. 1994, 6, 31.
(9) (a) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis; Wiley-
Interscience: New York, 1992. (b) Collman, J. P.; Hegedus, L. S.;
Norton, J. R.; Finke, R. G. Principles and Applications of Organ-
otransition Metal Chemistry; University Science Books: Mill Valley,
CA, 1987. (c) Hegedus, L. S. Transition Metals in the Synthesis of
Complex Organic Molecules; University Science Books: Mill Valley,
CA, 1994.
(10) Inorganic Materials; Bruce, D. W.; O’Hare, D., Eds.; Wiley: Chich-
ester, 1992.
(11) Zuman, P. The Elucidation of Organic Electrode Processes; Academic
Press: New York, 1969.
(12) (a) Hammett, L. P. Physical Organic Chemistry; McGraw-Hill: New
York, 1970. (b) Shorter, J. Correlation Analysis in Organic Chemistry;
Clarendon Press: Oxford, 1973.
(13) Kadish, K. M. Prog. Inorg. Chem. 1986, 34, 435 and earlier references
therein.
(4) Hall, L. M.; Speer, R. J.; Ridgway, H. J. J. Clin. Hematol. Oncol.
1980, 10, 25.
(5) (a) Cotton, F. A.; Kim, Y. J. Am. Chem. Soc. 1993, 115, 8511. (b)
Cotton, F. A.; Kim, Y.; Ren, T. Inorg. Chem. 1992, 31, 2723. (c)
Cayton, R. H.; Chisholm, M. H.; Huffman, J. C.; Lobkovsky, E. B. J.
Am. Chem. Soc. 1991, 113, 8709. (d) Mashima, K.; Nakano, H.;
Nakamura, A. J. Am. Chem. Soc. 1993, 115, 11632. (e) Handa, M.;
Mikuriya, M.; Kotera, T.; Yamada, K.; Nakao, T.; Matsumoto, H.;
Kasuga, K. Bull. Chem. Soc. Jpn. 1995, 68, 2567.
(6) (a) Doyle, M. P. Aldrichimica Acta 1996, 29, 3. (b) Doyle, M. P. In
Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York,
1993; and earlier references therein.
(7) Padwa, A.; Austin, D. J. Angew. Chem., Int. Ed. Engl. 1994, 33, 1797
and earlier references therein.
(8) Nocera, D. G. Acc. Chem. Res. 1995, 28, 209.
S0020-1669(96)00555-1 CCC: $12.00 © 1996 American Chemical Society

Energy Relationships in Dinuclear Compounds
Inorganic Chemistry, Vol. 35, No. 22, 1996 6423
the study of â-alkyl migratory insertion reactions.15 Studies of
metallodiimine compounds, such as the families of Ru(bpy)3
and Pt(diimine)(dithiolate), also demonstrate that a substituent
effect can be critical in optimizing the photophysical properties
of the well-known MLCT excited state.17
Hewlett Packard 8452A diode array UV-vis spectrophotometer. Cyclic
voltammograms were recorded on a BAS CV-50W voltammetric
analyzer with Pt working and auxiliary electrodes and a Ag/AgCl
reference electrode in 0.1 M (n-Bu)4NBF4 CH2Cl2 solution (N2
degassed). In all the measurements, the potential of ferrocenium/
ferrocene couple occurred at 625 mV under the specified experimental
conditions.
Preparation of Ligands. All the ligands were obtained according
to an established procedure.20 In a typical reaction, 12.32 g (0.10 mol)
of p-anisidine and 16.6 mL (0.05 mol) of triethyl orthoformate were
combined in a round-bottom flask equipped with a distillation tube
and heated at 140 °C in air until the distillation of ethanol ceased. The
remaining solid was recrystallized from hot toluene, washed with
hexane, and dried under a dynamic vacuum to yield 11.29 g of white
needles ((p-OMeC6H4)NC(H)N(H)(p-OMeC6H4), 92%). Spectro-
scopic data (NMR and IR) for all the ligands are listed as part of the
supporting information (Table S1).
Preparation of Mo2(ArNC(H)NAr)4. Ar is XC6H4-, with X as
p-OMe (1), H (2), m-OMe (3), p-Cl (4), m-Cl (5), m-CF3 (6), p-COMe
(7), p-CF3 (8), Ar is 3,4-Cl2C6H3- (9) or 3,5-Cl2C6H3- (10). Except 7,
all the compounds were obtained as yellow polycrystalline materials
through refluxing Mo(CO)6 with 3 equiv of the corresponding ligands
in degassed (two freeze-pump-thaw cycles) o-dichlorobenzene.
Yields (based on Mo(CO)6) varied from 70% to 95%. Refluxing the
formamidine bearing the p-acetyl group with Mo(CO)6 led to intractable
brown products. Compound 7 was obtained in 65% yield by treating
Mo2(OAc)4 with the lithiated formamidine (4 equiv) in degassed THF.
Compounds 1 and 4 were submitted for elemental analysis (performed
by Atlantic Microlab, Norcross, GA) with the following results: 1,
calcd (found), C, 59.41 (59.48), N, 9.24 (9.22), H, 4.98 (4.98); 4, C,
50.03 (50.30), N, 8.98 (8.84), H, 2.91 (3.04). Single crystals of X-ray
quality were grown by slow diffusion of hexane into a CH2Cl2 solution
for compounds 1, 2, 5, and 10. The analog with X as p-Me (11), a
known compound, was prepared and purified according to the
literature.24 UV-vis spectral data are given in Table 3 for all the
compounds but 9, for which the data are inaccessible due to very low
solubility of the compound in common organic solvents. 1H and 13C
NMR and IR spectroscopic data are supplied as part of the supporting
information (Table S2).
X-ray Crystallography for Compounds 1, 2, 5, and 10. All the
diffraction data sets were collected with a Siemens P4 instrument (Mo
KR radiation (λ ) 0.710 73 Å). Typically, a brick-shaped yellow crystal
was glued onto the tip of a glass fiber. The crystal was judged to be
acceptable on the basis of omega scans and rotation photography. A
random search located reflections to generate the reduced primitive
cell, cell lengths being corroborated by axial photography. Additional
reflections with 2θ values near 25° were appended to the reflection
array and yielded the refined cell constants. Data were collected as
presented in Table 1 and were corrected for absorption (empirical ψ
scan).
2+
Given the abundance of substituent tuning in mononuclear
coordination chemistry, it is surprising that there are only few
existing examples for the metal-metal bonded dinuclear
compounds.18 Yet the importance of such tuning in dinuclear
compounds has been demonstrated with the Rh2(O2CR)4-
catalyzed intramolecular carbene insertion of 2,3,4-trimethyl-
3-pentyl diazoacetoacetate,7,18c where both the regio- and
chemoselectivities were significantly altered through a variation
in the electron-withdrawing ability of R. To our knowledge,
there has been no systematic report of substituent tuning for
the quadruply bonded species to date. The feasibility, however,
is evident from the earlier work for Mo2Cl4(PR3)4,19 where both
the δ-δ* transition energy and the E1/2(Mo25+/Mo24+) linearly
correlate with the π acidity of PR3. As part of our effort to
explore both the general reactivity and the catalytic properties
of metal-metal bonded dinuclear compounds, we have initiated
extensive studies of the remote substituent effect in dinuclear
paddle wheel compounds. The bidentate ligand diarylforma-
midinate is an excellent candidate for such study, since the
phenyl-substituted derivatives can be readily prepared,20 and its
capability to support dinuclear compounds is well docu-
mented.2,21,22 Preliminary results obtained for a series of
quadruply bonded dimolybdenum compounds, tetrakis(µ-N,N′-
diarylformamidinato)dimolybdenum(II), were communicated
earlier,23 and a full account of the molecular and electronic
structures, spectroscopies, redox properties, and their dependence
on the remote substituents is presented here.
Experimental Section
General Considerations. All the syntheses were performed under
a dry argon atmosphere using standard Schlenk-line techniques unless
otherwise specified. Triethyl orthoformate, p-anisidine, m-anisidine,
p-toluidine, aniline, p-chloroaniline, m-chloroaniline, 3-(trifluorom-
ethyl)aniline, 4′-aminoacetophenone, and 3,5-dichloroaniline were
purchased from Aldrich, 4-(trifluoromethyl)aniline, and 3,4-dichloroa-
niline from ACROS, and molybdenum hexacarbonyl from Strem. 1H
and 13C NMR spectra were recorded on a Bruker AMX-360 NMR
spectrometer, with chemical shifts (δ) referenced to the residual CHCl3
and the solvent CDCl3, respectively. IR spectra were recorded on a
Nicolet system 550 (Magna series) FTIR spectrometer using KBr disks.
UV-vis spectral data were obtained in degassed CH2Cl2, using a
(14) (a) Ostovic, D.; Bruice, T. C. Acc. Chem. Res. 1992, 25, 314. (b)
Grinstadff, M. W.; Hill, M. G.; Birnbaum, E. R.; Schaefer, W. P.;
Labinger, J. A.; Gray, H. B. Inorg. Chem. 1995, 34, 4896.
(15) Rix, F. C.; Brookhart, M.; White, P. S. J. Am. Chem. Soc. 1996, 118,
2436.
(16) (a) Baranano, D.; Hartwig, J. F. J. Am. Chem. Soc. 1995, 117, 2937.
(b) Kapteijn, G. M.; Spee, M. P. R.; Grove, D. M.; Kooijman, H.;
Spek, A. L.; van Koten, G. Organometallics 1996, 15, 1405.
(17) (a) Nazeeruddin, M. K.; Zakeeruddin, S. M.; Kalyanasundaram, K. J.
Phys. Chem. 1993, 97, 9607. (b) Cummings, S. D.; Eisenberg, R. J.
Am. Chem. Soc. 1996, 118, 1949.
(18) (a) Das, K.; Kadish, K. M.; Bear, J. L. Inorg. Chem. 1978, 17, 930.
(b) Bottomley, L. A.; Hallberg, T. A. Inorg. Chem. 1984, 23, 1584.
(c) Pirrung, M. C.; Morehead, A. T. J. Am. Chem. Soc. 1994, 116,
8991.
(19) (a) Cotton, F. A.; Daniels, L. M.; Powell, G. L.; Kahaian, A. J.; Smith,
T. J.; Vogel, E. F. Inorg. Chim. Acta 1988, 144, 109. (b) Hanselman,
D. S.; Smith, T. J. Polyhedron 1988, 7, 2679. Both the inductive effect
and the steric bulkness (Tolman’s cone angle) of R contribute to the
π acidity of PR3.
Computations were performed using SHELXTL PLUS, PC Version
5.1â (Siemens Analytical), and data were refined against F2 (wR2 )
2
[∑[w(Fo2 - Fc2)2]/∑[w(Fo )2]1/2). All of the non-hydrogen atoms were
located with direct methods and refined anisotropically. Hydrogen
atoms were located and refined with Beq ) 1.2 × the value of attached
carbon atoms, except for the methoxyl protons of 1, which were
disordered and idealized atoms that were generated and refined with
Beq ) 0.08. For compounds 1, 5, and 10, there is only one-half of the
molecule in the asymmetric unit, which is related to the other half via
the crystallographic inversion center situated at the centroid of the Mo-
Mo′ vector. There are two half-molecules in the asymmetric units of
2, and each is similarly related to the other half through inversion. The
final least-squares refinement converged at the R-factors reported in
Table 1, along with other procedure parameters. Atomic coordinates
and isotropic displacements are provided as supporting information
(Tables S3-S22). Table 2 lists selected bond lengths and angles for
compounds 1, 2, 5, and 10. Since the geometrical parameters for two
independent molecules of 2 are essentially the same, only those for
one of them are included in Table 2.
(20) Bradley, W.; Wright, I. J. Chem. Soc. 1956, 640.
(21) (a) Barker, J.; Kilner, M. Coord. Chem. ReV. 1994, 133, 219. (b)
Umakoshi, K.; Sasaki, Y. AdV. Inorg. Chem. 1993, 40, 187.
(22) Cotton, F. A.; Ren, T. J. Am. Chem. Soc. 1992, 114, 2237.
(23) Lin, C.; Protasiewicz, J. D.; Smith, E. T.; Ren, T. J. Chem. Soc., Chem.
Commun. 1995, 2257.
Computational Procedure. Fenske-Hall molecular orbital calcula-
(24) Cotton, F. A.; Feng, X.; Matusz, M. Inorg. Chem. 1989, 28, 594.

6424 Inorganic Chemistry, Vol. 35, No. 22, 1996
Lin et al.
Table 1. Crystallographic Data for Compounds 1, 2, 5, and 10
1
2
5
10
formula
formula weight
cryst syst
space group
a, Å
b, Å
c, Å
C60H60N8O8Mo2
1213.04
triclinic
P1h
10.314(1)
10.405(1)
13.996(1)
80.65(1)
75.76(1)
81.55(1)
1427.4(3)
1
1.411
0.500
0.71073
2.00-25.00
293(2)
C52H44N8Mo2
972.83
monoclinic
P21/n
18.101(1)
10.404(1)
24.498(2)
C52H36Cl8N8Mo2
1248.37
triclinic
P1h
9.962(3)
11.555(3)
12.463(3)
87.80(3)
75.90(3)
65.85(2)
1266.6(5)
1
1.637
0.963
0.71073
1.94-25.00
293(2)
C52H28Cl16N8Mo2
1523.90
triclinic
P1h
9.885(1)
11.773(1)
14.463(1)
97.58(1)
104.67(1)
113.53(1)
1440.3(2)
1
1.757
1.223
0.71073
1.96-23.99
293(2)
R, deg
â, deg
101.99(1)
γ, deg
volume, Å3
Z
4512.7(7)
4
1.432
F
calc, g cm-3
µ, mm-1
0.601
λ, Å
0.71073
1.96-24.00
293(2)
0.0377
0.0763
1.070
θ range, deg
T, K
R1a
0.0395
0.1127
1.427
0.0310
0.0903
1.281
0.0279
0.0691
1.216
wR2b
goodness of fit on F2c
2
2
2
a R1 ) ∑||Fo| - |Fc||/∑|Fo|. b wR2 ) [∑[w(Fo - Fc2)2]/∑[w(Fo )2]]1/2
Table 2. Selected Geometric Parameters for Molecules 1, 2, 5, and 10
.
c goodness of fit ) [∑[w(Fo - Fc2)2]/(n - p)]1/2
.
1
2
5
10
Mo(1)-Mo(1)′
Mo(1)-N(1)
Mo(1)-N(2)′
Mo(1)-N(3)
Mo(1)-N(4)
N(1)-C(1)
2.0964(5) Mo(1)-Mo(1)′
2.0944(8)
2.161(4)
2.152(4)
2.167(4)
2.138(4)
1.318(6)
1.330(6)
1.329(6)
1.333(6)
Mo(1)-Mo(1)′
Mo(1)-N(1)
Mo(1)-N(2)′
Mo(1)-N(3)
Mo(1)-N(4)′
N(1)-C(1)
2.0958(6) Mo-Mo′
2.0965(5)
2.157(2)
2.163(2)
2.174(2)
2.170(2)
1.325(4)
1.324(4)
1.327(4)
1.325(4)
2.154(3)
2.163(3)
2.150(3)
2.166(3)
1.327(4)
1.328(4)
1.324(4)
1.333(4)
Mo(1)-N(5)
Mo(1)-N(6)
Mo(1)-N(7)
Mo(1)-N(8)
N(5)-C(4)′
N(6)-C(3)′
N(7)-C(3)
N(8)-C(4)
2.165(2)
2.152(2)
2.165(2)
2.163(2)
1.326(3)
1.327(3)
1.330(3)
1.323(3)
Mo-N(11)
Mo-N(21)
Mo-N(31)
Mo-N(41)
N(11)-C(3)′
N(21)-C(4)′
N(31)-C(3)
N(41)-C(4)
N(2)-C(1)
N(2)-C(1)
N(3)-C(2)′
N(4)-C(2)
N(3)-C(2)
N(4)-C(2)
Mo(1)′-Mo(1)-N(1) 91.44(7)
Mo(1)′-Mo(1)-N(2)′ 93.99(7)
Mo(1)′-Mo(1)-N(3) 92.86(7)
Mo(1)′-Mo(1)-N(4) 92.72(7)
Mo(1)′-Mo(1)-N(5) 91.57(10)
Mo(1)′-Mo(1)-N(6) 94.44(10)
Mo(1)′-Mo(1)-N(7) 91.28(10)
Mo(1)′-Mo(1)-N(8) 93.88(10)
91.34(14)
Mo(1)′-Mo(1)-N(1) 92.82(6)
Mo(1)′-Mo(1)-N(2)′ 92.53(6)
Mo(1)′-Mo(1)-N(3) 92.22(6)
Mo(1)′-Mo(1)-N(4)′ 93.27(6)
Mo′-Mo-N(11)
Mo′-Mo-N(21)
Mo′-Mo-N(31)
Mo′-Mo-N(41)
N(11)-Mo-N(21)
N(11)-Mo-N(41)
92.26(6)
92.80(6)
93.15(6)
92.42(6)
89.93(9)
89.59(9)
174.54(8)
89.12(9)
174.77(8)
90.86(9)
117.9(2)
117.4(2)
116.2(2)
117.3(2)
N(3)-Mo(1)-N(1)
N(3)-Mo(1)-N(2)′
N(1)-Mo(1)-N(2)′
N(3)-Mo(1)-N(4)
N(1)-Mo(1)-N(4)
N(2)′-Mo(1)-N(4)
C(1)-N(1)-Mo(1)
C(1)-N(2)-Mo(1)′
C(2)′-N(3)-Mo(1)
C(2)-N(4)-Mo(1)
N(1)-C(1)-N(2)
N(3)′-C(2)-N(4)
89.90(10) N(6)-Mo(1)-N(5)
N(1)-Mo(1)-N(4)′
89.22(9)
90.73(9)
88.88(10) N(5)-Mo(1)-N(7)
174.49(9) N(6)-Mo(1)-N(7)
174.31(9) N(8)-Mo(1)-N(5)
91.09(10) N(8)-Mo(1)-N(6)
89.60(10) N(8)-Mo(1)-N(7)
118.6(2)
115.8(2)
117.4(2)
116.5(2)
120.2(3)
120.5(3)
177.12(14) N(1)-Mo(1)-N(3)
88.04(14)
88.78(14)
N(2)′-Mo(1)-N(1)
N(2)′-Mo(1)-N(3)
174.62(7) N(11)-Mo-N(31)
89.68(9)
89.87(9)
N(21)-Mo-N(31)
N(21)-Mo-N(41)
171.66(14) N(2)′-Mo(1)-N(4)′
91.43(14)
117.7(3)
115.6(3)
117.8(3)
116.4(3)
120.7(5)
120.4(4)
N(4)′-Mo(1)-N(3)
C(1)-N(1)-Mo(1)
C(1)-N(2)-Mo(1)′
C(2)-N(3)-Mo(1)
C(2)-N(4)-Mo(1)′
N(1)-C(1)-N(2)
N(4)-C(2)-N(3)
174.51(7) N(41)-Mo-N(31)
C(4)′-N(5)-Mo(1)
C(3)′-N(6)-Mo(1)
C(3)-N(7)-Mo(1)
C(4)-N(8)-Mo(1)
N(7)-C(3)-N(6)′
N(5)′-C(4)-N(8)
116.9(2)
117.8(2)
117.3(2)
116.6(2)
120.0(2)
120.5(2)
C(3)′-N(11)-Mo
C(4)′-N(21)-Mo
C(3)-N(31)-Mo
C(4)-N(41)-Mo
N(11)′-C(3)-N(31) 120.4(3)
N(21)′-C(4)-N(41) 119.9(3)
tions25 were carried out with a VAX station 4000 VLC. Basis functions
used were generated by the numerical XR atomic orbital program26 in
conjunction with an XR-to-Slater basis program.27
Results and Discussion
Synthesis. The synthetic routes (A and B) applied in the
present study (sketched below) are the same as that reported
for the p-Me analog,24 and the yields were excellent. The failure
to isolate p-acetyl derivative (7) via route A is due to the thermal
instability of the ligand. All of the compounds except 9 are
fairly soluble in common polar solvents, which facilitated both
the spectroscopic and electrochemical characterizations. Sat-
Calculations were performed for compounds 1, where -OMe was
replaced with -OH due to the limit on the maximum number of basis
functions, 4, 5, and 10. Due to the invariance of Mo2 core geometry
(see Results and Discussion), the following metric parameters were
kept identical in all the model compounds: Mo-Mo, 2.095 Å; Mo-
N, 2.17 Å; Mo-Mo-N, 92.4°; and an eclipsed configuration. The
dihedral angle between the phenyl ring and the adjacent N-Mo-Mo
plane was set to 30° in keeping with the canted orientation of phenyl
groups observed in the X-ray structures, and all the model compounds
possess a D4 point symmetry. Typical distances for C-C, C-H, C-Cl,
and O-H bonds were assumed.28
o-Cl2C6H4
Mo(CO)6 + Hform
8
A
Mo2(form)4 7THF Mo2(OAc)4 + Li+form-
B
(25) Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768.
(26) Herman, F.; Skillman, S. Atomic Structure Calculations; Prentice-
Hall: Englewood Cliffs, NJ, 1963.
isfactory elemental analysis was obtained for two of the
(27) (a) Bursten, B. E.; Fenske, R. F. J. Chem. Phys. 1977, 67, 3138. (b)
Bursten, B. E.; Jensen, J. R.; Fenske, R. F. J. Chem. Phys. 1978, 68,
3320.
(28) International Tables for X-Ray Crystallography; Kynoch: Birming-
ham, 1968; Vol. 3.

Energy Relationships in Dinuclear Compounds
Inorganic Chemistry, Vol. 35, No. 22, 1996 6425
Figure 1. Structural diagrams for the molecular structures determined: (a) 1, (b) 2, (c) 5, (d) 10.
compounds (1 and 4), while the purity of the rest of compounds
is obvious from their H NMR spectra.
1H NMR Spectra. All the compounds in the series display
only one set of aromatic chemical shifts in their 1H NMR
spectra, which implies an effective 8-fold molecular symmetry,
either D4 or C4h. Although compounds 3, 4, and 6-9 have not
been crystallographically characterized, the observed trend
indicates that they are isostructural to 1, 2, 5, and 10.
It has been demonstrated that the chemical shifts of the
protons in the vicinity of the dinuclear core are sensitive to the
magnetic anisotropy (∆ø) produced by the metal-metal multiple
bond and may thus serve as a sensitive probe of the bond
multiplicity.22,29 ∆ø for a variety of dinuclear compounds were
estimated on the basis of the δ of methine proton (-NCHN-)
of the formamidinate group according to the following
equation:30
1
Molecular Structures. In contrast to the p-Me analog which
crystallizes in tetragonal space group (P4/n),24 all the compounds
crystallographically characterized in the present work crystallize
in lower symmetries, which reflects the sensitivity of the packing
pattern toward the identity and position of phenyl substituents
and reflects, perhaps, the choice of crystallization solvents as
well.
Structural diagrams of compounds 1, 2, 5, and 10 are shown
in Figure 1 parts a-d. Similar to the p-Me analog (11), all the
compounds adopt the paddle wheel motif. Mo-Mo distances
are 2.0964(5), 2.0949[8], 2.0958(6), and 2.0965(5) Å for 1, 2,
5, and 10, respectively, which are typical for a Mo-Mo
quadruple bond and longer than, but statistically identical to,
that of 11 (2.085(4) Å).24 Similarity in the Mo-N bond lengths
also exists among 1, 2, 5, 10, and 11: the averaged values are
2.158[3], 2.155[4], 2.161[2], 2.166[2], and 2.171[14] Å, re-
spectively. An essentially eclipsed coordination geometry for
the Mo2 core was also observed for all the compounds, where
the averaged torsional angles N-Mo-Mo′-N′ are 1.1°, 1.4°,
0.4°, and 0.5° for 1, 2, 5, and 10, respectively. Considering
the large variance of σX among the compounds structurally
characterized, it is clear that there is little, if any, substituent
effect on the geometry about the dimolybdenum core.
12πr3∆δ
(1 - 3 cos2 θ)
∆ø ) ø| - ø )
(1)
where ∆δ ) δ(-NCHN- in M2) - δ(-NCHN- in Ni2) and
the r is the estimated distance between the methine proton
(29) (a) Cotton, F. A.; Kitagawa, S. Polyhedron 1988, 7, 1673. (b) Cotton,
F. A.; Kitagawa, S. Inorg. Chem. 1987, 26, 3463.
(30) Harris, R. K. Nuclear Magnetic Resonance Spectroscopy; Pittman:
London, 1983.

6426 Inorganic Chemistry, Vol. 35, No. 22, 1996
Lin et al.
Table 3. Spectroscopic and Electrochemical Data of Mo2(diarylformamidinate)4
compound
(X)
1
11
2
3
4
5
6
7
8
10
(3,5-Cl2)
a
(p-OCH3) (p-CH3 )
(-H)
(mOCH3)
(p-Cl)
(m-Cl)
(m-CF3)
(p-COCH3)
(p-CF3)
σ
E
-0.27
244
-0.17
333
0
418
0.12
458
0.23
601
0.37
683
0.43
762
0.50
778
0.54
795
0.74
1/2, mVb
(∆Ep (mV), ip,c/ip,a
)
(158, 0.98) (191, 0.97) (93, 0.96) (197, 1.01) (250, 1.19) (177, 0.99) (155, 0.75)
(137, 0.73) (155, 0.77)
4750
∆ø (10-36
4950
5060
5000
5040
4900
4890
4820
4730
m3/molecule)
λ
λ
max (δ-δ*), nmc 430 (sh)
438 (sh)
270 (sh)
290
440 (sh)
270 (sh)
288
442 (sh)
280 (sh)
294
438 (sh)
270 (sh)
294
442 (sh)
270 (sh)
290
442 (sh)
436 (sh)
446 (sh)
444 (sh)
max (others), nm, 274 (sh)
272 (52 500) 308 (sh)
288 336
(69 600) (53 500) (106200)
268 (54 400) 272 (sh)
296 292
(70100) (53800)
(ꢀ, M-1 cm-1
)
288
(67 900)
c
(63 200)
(46 700)
(56 000)
(74 300)
310 (sh)
384 (sh)
316 (sh)
375 (sh)
312 (sh)
370 (sh)
318 (sh)
384 (sh)
320 (sh)
384 (sh)
312 (sh)
384 (sh)
308 (sh)
384 (sh)
314 (sh)
386 (sh)
320 (sh)
382 (sh)
a Spectroscopic and electrochemical data for this compound were reported in ref 24, but were redetermined here to ensure a consistent experimental
condition across the series. Only minor deviation was noticed. b Measurement was carried out in CH2Cl2 with Bu4NBF4 as the supporting electrolyte,
Pt working and auxiliary electrodes, Ag/AgCl reference electrode, [Mo2] ≈ 1 mM, and scan rate 100 mV/s. Under these conditions, the E1/2(Fc+/
Fc) was consistently measured at +625 mV. E1/2 ) (Ep,a + Ep,c)/2 and ∆Ep ) Ep,a - Ep,c.
c UV-vis spectra were recorded in degassed CH2Cl2.
and the centroid of the M-M vector.22 As discussed earlier,22
the change of the chemical shift (∆δ) due to the magnetic
anisotropy is best referenced to the dinickel analogs (Ni2(form)4),
where the magnetic anisotropy vanishes due to the absence of
Ni-Ni covalent bonding. Most of the dinickel analogs corre-
sponding to the present dimolybdenum series have been
synthesized and characterized,31 and hence the ∆ø for the
dimolybdenum center can be estimated according to eq 1 and
the values are listed in Table 3. The magnetic anisotropy does
vary across the series, but the variance is relatively small and
seems unrelated to the Hammett constant of substituents. Thus
the accumulation of the bonding electron density along the Mo-
Mo vector is essentially a constant, reinforcing our belief that
the Mo-Mo quadruple bond is not affected by remote substit-
uents.
Redox Properties. Early studies established that the p-Me
analog undergoes a reversible one-electron (1e) oxidation, which
led to the isolation of the oxidized monocation. Structural
Figure 2. Some typical cyclovoltammograms for dimolybdenum
compounds: (a) 1, (b) 7, (c) 10. Peaks corresponding to the ferrocene
standard are marked with “Fc”. Ferrocene standard is not shown in (b)
since its peaks severely overlap with those of the sample.
-
Mo2(L)4 y-e z [Mo2(L)4]+
(2)
+ e-
comparison between the neutral parent molecule and the
daughter cation led to the unambiguous assignment that the
oxidation corresponds to the removal of an electron from the
Mo-Mo δ bonding orbital,24 and hence the half-wave potential
(E1/2) determined from a cyclic voltammogram (CV) should
provide an accurate assessment of the absolute energy of δ
bonding orbital. Similarly, the first eight compounds (1-8) in
the current series were found to undergo 1e (qausi)reversible
oxidation in dichloromethane solution. Typical CV profiles of
1, 8, and 10 are shown in Figure 2, while the measured
parameters (E1/2, ipc/ipa) are listed in Table 3. The measured
E1/2 values range from +244 mV for 1 to +795 mV for 8 and
are shifted anodically as σX increases. Therefore, the energy
of δ bonding orbital decreases significantly upon increasing the
electron-withdrawing ability of the substituent.
∆E1/2 ) E1/2(X) - E1/2(H) ) F8σX, F ) 87.2 mV (3)
The reactivity constant F is comparable to those determined
for the metal-based redox processes in various substituted
metalloporphyrins13 and is larger than that estimated for the
oxidation of Rh2(O2CR)4 (64 mV).18 However, F(Mo2) is
significantly smaller than the reactivity constant determined for
the Ni2(form)4 analogs (114 mV),31 which contrasts the pattern
observed for metalloporphyrins where the redox reactivity
constant decreases as the number of d orbital electrons
increases.13 This discrepancy probably originates from the
difference in the Ni-based HOMO between the current series
and Ni-porphyrine series. On the basis of eq 3, a (quasi)-
reversible oxidation around 944 mV would be anticipated for
compound 10 (σ ) 2σm(Cl) ) 0.74; assuming that the additivity
of Hammett constants is valid, see ref 12), but only an anodic
peak was observed at 1029 mV.
A linear least-squares fitting of the E1/2 - 8σX plot (coefficient
8 is the number of substituents per compound, see ref 23 for
the plot) revealed the following Hammett equation11 with a
correlation coefficient of 0.993:
Consistent with the measured E1/2 values, the resistance
toward aerobic oxidation gradually increases with increasing σ
value across the series. For instance, a solution of 1 in
anhydrous CH2Cl2 turns to brown within minutes when exposed
to air, while the solution of 10 displays an identical UV-vis
spectrum even after weeks of exposure to air.
(31) Lin, C.; Protasiewicz, J. D.; Ren, T. Inorg. Chem. submitted for
publication. δ(-NCHN-) (ppm, in CDCl3) are 6.15 (p-OCH3), 6.17
(p-CH3), 6.28 (-H), 6.31 (m-OCH3), 6.22 (p-Cl), 6.33 (m-Cl), 6.45
(m-CF3), 6.48 (p-CF3), 6.41 (3,5-Cl2).

Energy Relationships in Dinuclear Compounds
Inorganic Chemistry, Vol. 35, No. 22, 1996 6427
Table 4. Upper Valence Molecular Orbitals for Dimolybdenum Compounds 1, 4, 5, and 10
compound
1
4
5
10
assign.
π*(e)
σ*(a2)
E (eV)
characters (%)
E (eV)
characters (%)
E (eV)
characters (%)
E (eV)
characters (%)
1.12 dxz,yz (76)
0.59 dxz,yz (57)
-0.03 dxz,yz (62)
0.04 dxz,yz (89)
2
2
2
2
-1.36 dz (41), Mo pz (33), -1.87 dz (42), Mo pz (34), -2.13 dz (45), Mo pz (37), -2.19 dz (45), Mo pz (36),
Mo s (10)
Mo s (10)
Mo s (11)
Mo s (10)
δ* (b1, LUMO) -2.20 dxy (76), Npx,y (14)
-2.82 dxy (76), Npx,y (13)
-3.30 dxy (77), Npx,y (12)
-3.38 dxy (78), Npx,y (12)
δ (b2, HOMO)
-7.12 dxy (86),
Cmethinepx,y (14)
-7.66
d
xy (86),
-7.99
d
xy (86),
-8.05
d
xy (87),
Cmethinepx,y (14)
Cmethinepx,y (13)
Cmethinepx,y (13)
(a1)
(e)
π(e)
σ(a1)
-8.54 Npx,y (48)
-9.20 Npx,y (45)
-10.43 Npx,y (59)
-10.63 Npx,y (58)
-8.93 Npx,y (42)
-9.55 Npx,y (38)
-10.78 Npx,y (47)
-10.96 Npx,y (41)
-10.12
dxz,yz (93)
-10.67 dxz,yz (90)
-11.28 dz2 (54), Mo s (26)
-11.05 dxz,yz (92)
-11.64 dz2 (54), Mo s (26)
-11.11 dxz,yz (81)
-11.60 dz2 (36), Mo s (19)
-10.74 dz2 (55), Mo s (26)
Least-squares fittings of ∆E1/2 plots vs either σp- or σp+ (taken
from the tabulation of Hansch et al.32) were also attempted with
less satisfactory results: the correlation coefficients are 0.95
jugation between the N-C-N group and the p-acetyl group
through the phenyl ring.
Electronic Structures. In order to evaluate the substituent
influence, all the atoms/functional groups were retained in the
model calculation with one exception: replacement of the
-OMe group with -OH in 1. Although a comprehensive
analysis is practically impossible since there are 160-200
occupied valence MOs, the feature of most low-energy MOs
being C-C, C-H, C-N, and C-X σ bonding and C-C and
C-N π bonding orbitals enables us to focus on those of
dominant Mo contributions. Relevant information for these
orbitals are summarized in Table 4. In all of the model
calculations, the Mo-Mo (anti)bonding orbitals in order of the
ascending energies are σ(a1), π(e), δ(b2, HOMO), δ*(b1,
LUMO), σ*(a2), and π*(e).34 Thus the existence of a Mo-Mo
quadruple bond (σ2π4δ2) in all the compounds is further
confirmed on the basis of the orbital occupancy.
+
for the ∆E1/2 - σp plots of 1, 4, 7, and 11 and 0.97 for the
∆E1/2 - σp- plots of 1, 4, 7, 8, and 11. This behavior parallels
the early observation in the metallotetraphenylporphorins,
where the influence of the phenyl substitution is best described
- 13
by σ, while that of â-substitution correlates better with σp
.
The rotational freedom of the aryl group about the Mo2
coordination sphere in solution is probably responsible for the
absence of a substantial delocalized contribution from the remote
substituent.
The E1/2 value for all compounds remained essentially scan
rate-independent from 5 to 400 mV/s, with the largest variance
determined as 18 mV. In addition, ip - (scan rate)1/2 profiles
were all linear except for 6 at high scan rate. On the basis of
these observations, we believe that the values listed in Table 3
obtained under identical conditions and a scan rate of 100 mV/s
represent the practical equilibrium formal reduction potentials
(E°’s). Although it is apparent that the reversibility is less than
ideal for most of the compounds on the basis of the difference
in their anodic and cathodic peak potentials (i.e., ∆Ep > 60
mV), it is not entirely evident that this is due to sluggish electron
transfer rates. It is not uncommon for ∆Ep- scan rate profiles
to be influenced by uncompensated solution resistance in
nonaqueous solvents.33
Fine features of the Mo-Mo (anti)bonding interactions and
Mo-L bondings around the dimolybdenum core with quadruple
bond are well documented,2 and those of the model compound
supported by the [HNC(H)NH]- bridge were addressed in length
on the basis of XR calculations24 and will not be discussed here.
The unique feature of the present results is that the energy levels
of both Mo-Mo bonding and antibonding orbitals do depend
on the substituent, and all of them decrease as the σX increases.
While the variance of energy levels is not exactly linear with
σX, the variance of E(δ*) is essentially the same as that of E(δ),
and hence a nearly constant δ-δ* (HOMO-LUMO) gap (4.92,
4.84, 4.69, and 4.67 eV for 1, 4, 5, and 10, respectively) is
Electronic Spectroscopy. Consistent with the pale yellow
color of the compounds, there is no significant absorption peak
in the visible region except for a shoulder appearing in the range
of 430-446 nm (2.88-2.78 eV), which is, as discussed earlier,24
1
assigned as the dipole-allowed δ-δ* (1A1 f A2) transition.
maintained, which parallels the invariance in the observed λmax
-
(δ-δ*) across the series. It is worth pointing out that there is
no orbital contribution toward Mo-Mo bonding orbitals from
the substituents, confirming the absence of the substituent
contribution through delocalization. Furthermore, the Mulliken
charges on the molybdenum center for 1, 4, 5, and 10 are 0.505,
0.540, 0.556, and 0.571, respectively. This demonstrates that,
despite the absence of direct orbital contribution to the upper
valence orbitals, the electron-withdrawing substitutent does
induce a buildup of positive charge on the Mo2 core. On the
basis of these observations we postulate that polar substitution
away from the first coordination sphere is equivalent to applying
a spherical electrostatic potential, which can significantly shift
the absolute energy levels of the Mo2 core with a minimum
alternation in both the relative distribution and the composition
of the upper valence molecular orbitals.
The narrow distribution of the δ-δ* transition across the series
implies that the increments in both E(δ) and E(δ*) must be
approximately the same when σX is increased. With the
exception of compound 7, the overall features in the UV region
for all the compounds are very similar: there are one peak
(around 290 nm) and three shoulders (around 380, 315, and
271 nm) which can be attributed to the LMCTs from the (-N-
C-N-)-based high-energy π orbitals to both δ* and π* orbitals
at the Mo2 center.24 Since the observed variance for each
transition across the series (except 7) is negligible compared
with the corresponding transition energy, it is clear that the
relative energy distribution of the orbitals involved in these
LMCTs, i.e., δ*(Mo2), π*(Mo2) orbitals, and nitrogen-based π
orbitals on the formamidinate ligands, is unaffected by the polar
substitution. The p-acetyl derivative (compound 7) is unique
and has only one well-resolved absorption and one shoulder in
the UV region. It is possible that there exists some hypercon-
(34) The reversal of the conventional E(π*) < E(σ*) is not unusual for
the Fenske-Hall calculation of metal-metal bonded dinuclear species,
and it is likely caused by the empirical approximations inherent in
the Fenske-Hall method (personal communication, Dr. Xuejun Feng,
Texas A&M University).
(32) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165.
(33) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; Wiley: New
York, 1980.

6428 Inorganic Chemistry, Vol. 35, No. 22, 1996
Lin et al.
Conclusions and Future Work
Acknowledgment. Donors of Petroleum Research Fund are
thanked for partial support of this research. The NMR facility
at Florida Tech is supported through an NSF grant (CHE-
9013145). The authors acknowledge Dr. A. B. Brown for
introducing us to ref 11b and the discussion about Hammett
correlation, Dr. M. B. Hall for a copy of the Fenske-Hall
program, Dr. J. Rokach and his group for various assistance,
and Dr. T. R. Webb for suggesting the study of p-acetyl-
derivatized formamidine.
Our study demonstrates that through the remote substituent
both the energy levels of frontier orbitals and the electron
richness of dimolybdenum core can be significantly attenuated.
The generality of this observation within the family of paddle
wheel compounds is being probed with the study of forma-
midinate compounds of middle and late transition metals.
Similar studies will be carried out with the diruthenium systems
bearing both labile and less bulky bridging ligands which may
catalyze reactions such as carbene-transfer, hydrogenation, and
olefin metathesis.35
Supporting Information Available: Tables of spetroscopic data
for ligands (S1), 1H and 13C NMR and IR data for dimolybdenum
compounds (S2), positional parameters for all atoms, anisotropic tem-
perature factors, bond distance and angles for compounds 1, 2, 5 and
10 (S3-S22) and linear free energy plot E1/2 vs σX (Figure S1) (33
pages). Ordering information is given on any current masthead page.
(35) (a) Lindsay, A. J.; McDermott, G.; Wilkinson, G. Polyhedron 1988,
7, 1239. (b) Noels, A. F.; Demonceau, A.; Carlier, E.; Hubert, A. J.;
Marquez-Silva, R.-L.; Sanchez-Delgado, R. A. J. Chem. Soc., Chem.
Commun. 1988, 783. (c) Noels, A. F.; Demonceau, A.; Saive, E. AdV.
Catal. Des., Proc. Workshop 2nd. 1992, 2, 73.
IC960555O