Spectroscopic comparisons of MoW(porphyrin)2 heterodimers with homologous Mo2 and W2 quadruple bonds: A dynamic NMR and resonance raman study

Article abstract of DOI:10.1021/ja972130m

The rotational barrier for MoW(meso-monotolyl octaethylporphyrin)₂ ([(TOEP)MoW(TOEP)]) has been determined (ΔG((+))(rot) = 10.6 ± 0.1 kcal/mol) by variable-temperature NMR and complete band shape analysis and is compared with values previously obtained for the analogous homodimers. The overall quadruple bond strengths of these isostructural dimolybdenum, ditungsten, and molybdenum - tungsten porphyrin dimers have also been compared by calculation of the force constants corresponding to each metal - metal bond stretching frequency as observed by resonance Raman spectroscopy. The Raman results are as follows: [Mo(OEP)]₂, ν(MoMo) = 310 cm^-1, k = 2.72 mdyn/A?; [(OEP)MoW(OEP)], ν(MoW) = 279 cm^-1, k = 2.89 mdyn/A?; [Mo(TOEP)]₂, ν(MoMo) = 310 cm^-1, k = 2.72 ν(MoMo) [W(TOEP)]₂ ν(WW) = 275 cm^-1, k = 4.08 mdyn/A?; and [(TOEP)-MoW(TOEP)], ν(MoW) = 278 cm^-1, k = 2.89 mdyn/A?: [MoMoW(TOEP)]. Both the ¹H NMR and Raman specta are consistent with a [(Por)MoW(Por)] structure wherein the Mo(Por) congener experiences a more drastic 'bending-back' distortion of the porphyrin macrocycle.

Full text of DOI:10.1021/ja972130m

1456
J. Am. Chem. Soc. 1998, 120, 1456-1465
Spectroscopic Comparisons of MoW(porphyrin)2 Heterodimers with
Homologous Mo2 and W2 Quadruple Bonds: A Dynamic NMR and
Resonance Raman Study
James P. Collman,*,† S. T. Harford,† Stefan Franzen,‡,§ T. A. Eberspacher,†
Richard K. Shoemaker, and William H. Woodruff‡
Contribution from the Department of Chemistry, Stanford UniVersity, Stanford, California 94305,
Los Alamos National Laboratory, Los Alamos, New Mexico 87545, and UniVersity of NebraskasLincoln,
Lincoln, Nebraska 68588
ReceiVed June 26, 1997
Abstract: The rotational barrier for MoW(meso-monotolyl octaethylporphyrin)2 ([(TOEP)MoW(TOEP)]) has
been determined (∆G q ) 10.6 ( 0.1 kcal/mol) by variable-temperature NMR and complete band shape
rot
analysis and is compared with values previously obtained for the analogous homodimers. The overall quadruple
bond strengths of these isostructural dimolybdenum, ditungsten, and molybdenum-tungsten porphyrin dimers
have also been compared by calculation of the force constants corresponding to each metal-metal bond stretching
frequency as observed by resonance Raman spectroscopy. The Raman results are as follows: [Mo(OEP)]2,
νMoMo ) 310 cm-1, k ) 2.72 mdyn/Å; [(OEP)MoW(OEP)], νMoW ) 279 cm-1, k ) 2.89 mdyn/Å; [Mo-
(TOEP)]2, νMoMo ) 310 cm-1, k ) 2.72 νMoMo [W(TOEP)]2, νWW ) 275 cm-1, k ) 4.08 mdyn/Å; and [(TOEP)-
1
MoW(TOEP)], νMoW ) 278 cm-1, k ) 2.87 mdyn/Å. Both the H NMR and Raman specta are consistent
with a [(Por)MoW(Por)] structure wherein the Mo(Por) congener experiences a more drastic “bending-back”
distortion of the porphyrin macrocycle.
Introduction
crystallography,3,6 photoelectron spectroscopy,7 and resonance
Raman7 studies of these molecules suggests that the unbridged
heteronuclear dimers exhibit properties which are the aVerage
of the homonuclear dimers. Evaluation of the apparently deviant
properties of bridged heteronuclear dimers became the subject
of a review by Morris,8 in which the “enhanced stability” of
bridged heterodimers was attributed to ligand steric preferences.
The first reported heteronuclear quadruple bonds between
molybdenum and tungsten appeared to be some of the strongest
bonds found in coordination complexes.1 Cotton et al. have
reported that the compound MoW(mhp)4 (mhp ) 2-hydroxy-
6-methylpyridine anion) exhibits a larger force constant for the
metal-metal stretch and a formally shorter bond length than
either of the corresponding homodimers.2 McCarley’s labora-
tory has also found a shorter bond in the MoW(O2CCMe3)4
dimer relative to its homonuclear analogues.3 These observa-
tions led to a proposition that polar, heteronuclear metal-metal
bonds exhibit a “special stability” relative to the homonuclear
congeners.
We have recently developed a single-step preparation of the
heterometallic porphyrin dimer [(OEP)MoW(OEP)] (1) contain-
ing an unsupported quadruple bond between molybdenum and
tungsten. We report here the relative bond strengths for
homologous dimolybdenum, ditungsten, and molybdenum-
tungsten porphyrin dimers. The novel heterodimer [(TOEP)-
MoW(TOEP)] (2) (TOEP ) meso-(4′-tolyl)octaethylporphyrin)
has been synthesized to calculate the barrier to rotation about
the mixed-metal δ-bond. Rotational barriers for [Mo(TOEP)]2
(3) and [W(TOEP)]2 (4) have been previously reported.9
More recently, the independent syntheses of MoWCl4(PR3)4
(PR3 ) PMePh2, PMe2Ph, or PMe3) by Morris4 and McCarley5
provided the first family of unbridged molybdenum-tungsten
quadruple bonds. The preliminary information available from
electronic absorption spectroscopy,4 cyclic voltammetry,4 X-ray
Vibrational spectroscopy has been employed to determine
ν
MM, and the corresponding force constants have been calculated
† Stanford University.
for these same group VIB dimers. With these techniques a
single component of the metal-metal quadruple bond as well
as the overall multiple metal-metal bond may both be directly
studied for evidence of an enhanced stability in the heteronuclear
species.
‡ Los Alamos National Laboratory.
§ Current address: North Carolina State University, Raleigh, NC 27695.
University of NebraskasLincoln.
(1) (a) Bursten, B. E.; Cotton, F. A.; Cowley, A. H.; Hanson, B. E.;
Lattman, M.; Stanley, G. G. J. Am. Chem. Soc. 1979, 101, 6244. (b) Cotton,
F. A.; Hanson, B. E. Inorg Chem. 1978, 17, 3237.
(2) (a) Katovic, V.; Templeton, J. L.; Hoxmeier, R. J.; McCarley, R. E.
J. Am. Chem. Soc. 1975, 97, 5300. (b) Katovic, V.; McCarley, R. E. J. Am.
Chem. Soc. 1978, 100, 5586.
(6) Bancroft, G. M.; Bice, Jo-Ann; Morris, R. H.; Luck, R. L. J. Chem.
Soc., Chem. Commun. 1986, 898.
(3) Luck, R. L.; Morris, R. H. J. Am. Chem. Soc. 1984, 106, 7978.
(4) Carlin, R. T. Ph.D. Thesis, Iowa State University, 1982.
(5) (a) Luck, R. L.; Morris, R. H.; Sawyer, J. F. Inorg. Chem. 1987, 26,
2422.
(7) Carlin, R. T.; McCarley, R. E. Inorg. Chem. 1989, 28, 280.
(8) Morris, R. H. Polyhedron 1987, 6, 793.
(9) Collman, J. P.; Garner, J. M.; Hembre, R. T.; Ha, Y. J. Am. Chem.
Soc. 1992, 114, 1292.
S0002-7863(97)02130-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/06/1998

A Dynamic NMR and Resonance Raman Study
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1457
Preparation of [(OEP)MoW(OEP)] (1). In the glovebox, a 100
mL round-bottom flask equipped with a Teflon vacuum valve and no.
9 O-ring sidearm was charged with 10 mL of Decalin, H2OEP (250
mg, 0.468 mmol), W(CO)6 (800 mg, 2.27 mmol), Mo(CO)6 (100 mg,
0.379 mmol), and a stir bar. The headspace was evacuated at 10-2
Torr for 10-15 min to remove adventitious oxygen and water. The
sealed flask was heated at 180 °C for 6 h, cooled to ambient
temperature, and transferred to the glovebox and held at -20 °C
overnight. (Caution! Although heating a closed system in this manner
may often be hazardous and ill-advised, at this temperature the vapor
pressure of Decalin is still less than 1 atm and the reaction is at a slightly
negative pressure.) The mixture of [(OEP)MoW(OEP)] (1) (82 mg,
13.7%), [Mo(OEP)]2 (5) (197 mg, 34.1%), and [W(OEP)]2 (6) (25 mg,
4.5%) was collected by filtration and washed with cold hexane. The
ditungsten species, 6, was removed as the monocation (EPR gav ) 1.87)
by titration of a benzene solution of the three dimers with a
stoichiometric amount (5.80 mg, 0.0175 mmol) of ferricinium hexafluo-
rophosphate (as determined by integration of Hmeso resonances in the
1H NMR and total mass of the mixture) followed by filtration and
Experimental Section
Materials. H2OEP and H2TOEP were synthesized simultaneously
by an adaptation10 of the published procedure for H2OEP.11 Dieth-
ylpyrrole was donated by Pharmacyclics and distilled immediately prior
to use. Mo(CO)6 and W(CO)6, cobaltocene, and ferricinium hexafluo-
rophosphate were purchased from Strem and used as received.
Benzene-d6 and toluene-d8 were purchased from Cambridge Isotope
Laboratories and vacuum distilled from sodium-benzophenone ketyl
immediately prior to use. Solvents used for the metalation (Decalin)
and manipulation (benzene) of the dimers were distilled from sodium
benzophenone ketyl under argon before introduction into the glovebox.
[Mo(OEP)]2,12 [W(OEP)]2,13 [Mo(TOEP)]2,10 and [W(TOEP)]210 were
prepared according to published procedures.
Physical Measurements. A nitrogen-filled Vacuum Atmospheres
drybox equipped with a Dri-Train inert gas purifier was employed for
manipulations carried out under anaerobic conditions. 1H NMR spectra
were recorded on a Varian XL-400 or General Electric 500 Omega
FT-NMR spectrometer using benzene-d6 or toluene-d8 as a solvent.
Resonances in the 1H NMR were referenced versus the residual proton
signal of the solvent. The mixing time for the 2D NOESY/EXCHSY
spectra was 100 ms. Resonance Raman samples were prepared in the
glovebox and flame sealed under vacuum.
Excitation for the RR experiments was provided by an Ar+ ion laser
(Spectra Physics Model 171). Typical laser powers were 20-30 mW
on resonance. Scattered light was collected by an f/1 5 cm focal length
lens and focused onto the slit of a SPEX 0.6 operating as a spectrograph
to disperse the light onto a Photometrics CCD camera. Typical data
acquisition times were 20 min. Because the precision of the depolar-
ization ratios (F) was (0.1, only general information such as whether
bands were polarized (a1g modes), depolarized (b1g and b2g), or
anomalously polarized (a2g) could be determined.
washing of the collected solid with benzene. [(OEP)MoW(OEP)]+PF6
-
(7) was then isolated by a second titration with ferricinium hexafluo-
rophosphate (16.1 mg, 0.0487 mmol) and collected by filtration and
washed with benzene. The yield was 73 mg (9.8% based on porphyrin).
EPR (frozen CH2Cl2): g ) 1.89, g ) 1.94. Mass spectrum: m/e
(OEP)MoW(OEP)+ (cluster) calcd 1344.6, found 1345 (see the Sup-
porting Information).
7 (12 mg, 8.06 mmol) was suspended in 5 mL of stirring benzene,
and a stock solution of cobaltocene (5 mg, 26.4 mmol in 2 mL of
benzene) was slowly added dropwise. The solution immediately
darkened and was allowed to stir for 2 h. The solution was filtered to
remove insoluble cobaltocinium hexafluorophosphate and the solvent
removed in vacuo. Unreacted cobaltocene was removed by sublimation.
The product was dried at 50 °C and 10-2 Torr for 4 h. The yield was
10 mg, (95%). UV-vis (nm) (log ꢀ, C6H6): Soret 382 (4.86), 435
(4.56), 532 (3.50), 560 (3.66). 1H NMR (ppm, C6D6): δ 9.06 (s, 4H,
Mo Hmeso); 8.68 (s, 4H, W Hmeso); 4.22-4.35 (m, 8H, J ) 7.6 Hz, Mo
-CH2CH3), 3.82-3.95 (m, 8H, J ) 7.6 Hz, Mo -CH2CH3); 4.15-
4.27 (m, 8H, J ) 7.4 Hz, W -CH2CH3), 3.75-3.87 (m, 8H, J ) 7.4
Hz, W -CH2CH3); 1.74 (t, 24H, J ) 7.6 Hz, Mo -CH2CH3); 1.72 (t,
24H, J ) 7.4 Hz, W -CH2CH3). Mass spectrum m/e (cluster) calcd
1344.6, found 1345.
Mass spectrometry was performed at the Mass Spectrometry Facility
at the University of CaliforniasSan Francisco and by Dr. Doris Hung
of the Analytical Services Division at Northwestern University.
Preparation of 5-(4′-Methylphenyl)-2,3,7,8,12,13,17,18-octaethyl-
porphyrin, H2TOEP.11 A 3000 mL, three-neck, round-bottom flask
was wrapped in aluminum foil and equipped with a Dean-Stark trap,
stir bar, and argon inlet. The flask was charged with 3,4-diethylpyrrole
(5.00 g, 40.6 mmol) and 1500 mL of benzene and was sparged with
nitrogen for 25 min. To this were added p-tolualdehyde (1.44 mL,
12.2 mmol), aqueous formaldehyde (2.76 mL of 37% solution, 34.7
mmol), and p-toluenesulfonic acid (150 mg, 0.79 mmol). The mixture
was stirred at 100 °C for 8 h. The reaction mixture was allowed to
cool, and the solution was bubbled with O2 for 12 h. The mixture was
concentrated under vacuum and the residue redissolved in chloroform
(300 mL). The resulting solution was washed once with 1 N aqueous
NaOH (200 mL) and twice with water (100 mL) and then concentrated
in vacuo. The crude reaction mixture was chromatographed with CH2-
Cl2 on a plug of alumina (15 cm × 5 cm) to remove polymeric
impurities. A second column was run using flash silica and CH2Cl2 to
remove bulk H2OEP (major product). H2TOEP was eluted from the
silica using a 5% ethyl acetate in CH2Cl2 system. The H2TOEP was
chromatographed a third time using conditions identical to those of
the second chromatography to remove any trace impurities; this yielded
368 mg (0.589 mmol, 5.8% yield) of H2TOEP. 1H NMR (C6D6): δ
10.43 (s, 2H, Hcis), 10.17 (s, 1H, Htrans), 4.31 (m, 12H, -CH2CH3),
3.05 (q, 4H, -CH2CH3), 2.15 (t, 18H, -CH2CH3), 1.42 (t, 6H,
-CH2CH3), 8.31 (d, 2H, Ho), 7.65 (d, 2H, Hm), 2.89 (s, 3H, aryl-CH3),
-2.72 (s, 1H, NH), -2.81 (s, 1H, NH). Mass spectrum: m/e calcd
625.5, found 625.5.
Preparation of [(TOEP)MoW(TOEP)] (2). The procedure was
identical to that used for the synthesis of 1. A 75 mg (0.120 mmol)
sample of H2TOEP yielded 10 mg (0.0065 mmol) of [(TOEP)MoW-
(TOEP)] (overall yield is 5.5%, based on porphyrin). UV-vis nm (log
ꢀ, C6H6): Soret 384 (4.82), 437 (4.49), 537 (3.56), 566 (3.59). 1H
NMR (ppm, C6D6): δ 9.19 (s, 2H, Mo Hcis), 9.41 (s, 1H, Mo Htrans),
4.25-4.50 (m, 6H, Mo -CH2CH3), 3.00-3.10 (q, 2H, Mo -CH2CH3),
3.73-3.99 (m, 6H, Mo -CH2CH3), 2.75-2.86 (q, 2H, Mo -CH2CH3),
1.65 (t, 18H, Mo -CH2CH3), 1.39 (t, 6H, -CH2CH3), 9.91 (d, 1H,
Mo Ho′), 7.95 (d, 1H, Mo Hm′), 6.98 (d, 1H, Mo Ho), 6.98 (d, 1H, Mo
Hm), 8.81 (s, 2H, W Hcis), 8.97 (s, 1H, W Htrans), 4.13-4.35 (m, 6H, W
-CH2CH3), 2.91-3.00 (q, 2H, W -CH2CH3), 3.61-3.85 (m, 6H, W
-CH2CH3), 2.65-2.74 (m, 2H, W -CH2CH3), 1.62 (t, 18H, W
-CH2CH3), 1.36 (t, 6H, W -CH2CH3), 9.86 (d, 1H, W Ho′), 7.95 (d,
1H, W Hm′), 6.98 (d, 1H, W Ho), 6.98 (d, 1H, W Hm), 2.59 (s, 3H,
aryl-CH3). Mass spectrum: m/e (cluster) calcd ) 1526.6, found 1527
(see the Supporting Information).
Results
1H NMR: Rapid Rotation Regime. To be consistent with
our previous study9 of the rotational barriers for 3 and 4, we
employed a single meso-tolyl substituent to break the charac-
teristic 4-fold symmetry of octaethylporphyrin and divide the
remaining meso-protons into two which are adjacent to the tolyl
substituent, Hcis, and one which is opposed, Htrans. The chemical
shift difference between these meso-protons results from
through-bond14 and through-space contributions from the meso-
tolyl substitutent. A characteristic 2:1 set of meso-proton
(10) Dr. Hilary Arnold Godwin, now an Associate Professor of Chemistry
at Northwestern University, first applied the Sessler et al. H2OEP preparation
to synthesis of meso-substituted H2OEP derivatives. See: Godwin, H. A.
Ph.D. Thesis, Stanford University, 1994.
(11) Sessler, J. L.; Mozaffari, A.; Johnson, M. R. Org. Synth. 1992, 70,
68-77.
(12) Collman, JP.; Barnes, C. E.; Woo, L. K. Proc. Natl. Acad. Sci. U.S.A.
1983, 80, 7684.
(13) Collman, J. P.; Garner, J. M.; Woo, L. K. J. Am. Chem. Soc. 1989,
111, 8141.

1458 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Collman et al.
Ar-CH3
Table 1. 1H NMR Chemical Shifts for Molybdenum and Tungsten Porphyrin Dimers
a
compound
[Mo(OEP)]2
Hmeso
Ho′
Hm′
Ho
Hm
CH2CH3
4.32
CH2CH3
9.20
1.78
3.92
[Mo(TOEP)]2
9.46
9.29
8.42
10.07
8.06
7.00
6.89
4.63/4.13
3.09/2.88
4.11
1.76
1.44
1.67
2.61
[W(OEP)]2
3.76
4.22/3.78
2.95/2.63
[W(TOEP)]2
8.80
8.50
9.57
7.91
7.10
7.10
1.64
1.29
2.59
[(OEP)MoW(OEP)]
[(TOEP)MoW(TOEP)]
9.06 (Mo)
8.68 (W)
9.41/9.19 (Mo)
4.28/3.88 (Mo)
4.21/3.81 (W)
4.38/3.85 (Mo)
3.05/2.18 (Mo)
4.24/3.73 (W)
2.95/2.69 (W)
1.74 (Mo)
1.72 (W)
1.65 (Mo)
1.39 (Mo)
1.62 (W)
1.36 (W)
9.91 (Mo)
9.86 (W)
7.95 (Mo)
7.95 (W)
6.98 (Mo)
6.98 (W)
6.98 (Mo)
6.98 (W)
2.59 (Mo)
2.59 (W)
8.97/8.81 (W)
a Multiplets.
Comparison of analogous 1H NMR spectra for 3 and 4 (Table
1) demonstrates that each corresponding resonance is located
further downfield in the dimolybdenum species, 3. For this
reason we find it reasonable to assign the more downfield sets
in 2 to the molybdenum-porphyrin half of the dimer. Nonethe-
less, the distinctions between Mo(TOEP) and W(TOEP) reso-
nances are not indisputable.
1
Molecular models and H NMR experiments17 indicate that
the tolyl rings in 2, 3, and 4 are positioned roughly perpendicular
to the porphyrin plane as a result of steric interactions with the
1
adjacent ethyl groups. The dynamic behavior of the H NMR
is not complicated by tolyl atropisomerism. A meso-proton of
one porphyrin which is eclipsed with the opposing meso-tolyl
group is coplanar with the tolyl protons and is thus deshielded.
This deshielding is very sensitive to the angle of rotation, and
is neglible when the meso positions are staggered (ø ) 45°).10
A pronounced deshielding is only consistent with a conformation
in which the meso positions of each porphyrin are eclipsed.
1H NMR. Slow Rotation Regime. Peak broadening in the
1H NMR spectrum of 2 is significant immediately upon lowering
the temperature below 20 °C. Figure 2 illustrates the dynamic
behavior of the meso- and endo-o-tolyl bands in 2.
The Ho′ doublets at δ 9.81 ppm (W) and 9.86 ppm (Mo) in
Figure 1 are resolved into tw separate doublets each (W, δ 9.76,
9.90 ppm; Mo, δ 9.83, 10.0 ppm) in a ratio of 0.60:0.40 at low
temperature. This observation suggests that, at -70 °C, 2 exists
as a 1.5:1 ratio of two conformers: the dominant conformation
with Ho′ chemical shifts of δ 9.76 (W) and 9.83 (Mo) ppm, and
the lesser populated conformation with chemical shifts of δ 9.90
(W) and 10.0 (Mo) ppm. The room-temperature spectrum is
composed of time-averaged resonances derived from the chemi-
cal shift positions in these two conformations.
Analysis of the low-temperature meso-proton resonances is
also consistent with the existence of two rotamers in a ratio of
1.5:1. Examination of the 2D NOESY/EXCHSY spectrum of
2 at -78 °C (Figure 3) enables a confident assignment of these
two conformations as the anti and gauche eclipsed rotamers
(see Figure 4). The off-diagonal resonances in the NOESY/
EXCHSY spectrum of 2 may arise as a result of proton exchange
between magnetically inequivalent positions or as a result of
Figure 1. (a) meso-Proton regiochemistry in monosubstituted OEP
derivatives. (b) Aryl proton regiochemistry in metalloporphyrin dimers.
1
(c) H NMR of 2 (C6D6, 20 °C).
resonances exists for each half of the heterodimer, [(TOEP)-
1
MoW(TOEP)] (2). The room temperature H NMR spectrum
for 2 is shown in Figure 1.
Figure 1c also consists of bands readily assigned to the tolyl
groups and the â-ethyl substituents. The endo-tolyl protons (Ho′
and Hm′), which are directed across the metal-metal bond, are
much more strongly deshielded than the exo-protons (Ho and
Hm). Quantitative calculations of the diamagnetic anisotropies
(∆ø) of Mo2, W2, and MoW quadruple bonds manifest the
largest ∆ø values yet reported, over 1 order of magnitude greater
than the ∆ø values obtained for alkynes.15 As seen in Figure
1, the Ho′ resonances in 2 are shifted downfield by nearly 3
ppm relative to Ho.16
(14) (a) Johnson, A. W.; Oldfield, D. J. Chem. Soc., Chem. Commun.
1966, 794-798. (b) Bonnett, R.; Stephenson, G. F. J. Org. Chem. 1965,
30, 2791. (c) Callott, H. J.; Louati, A.; Gross, M. Tetrahedron Lett. 1980,
21, 3281. (d) Crossley, M. J.; King, L. G.; Pyke, S. M. Tetrahedron 1987,
43, 4569.
(15) (a) Chen, J. D.; Cotton, F. A.; Falvello, L. R. J. Am. Chem. Soc.
1990, 112, 1076. (b) Cotton, F. A.; James, C. A. Inorg. Chem. 1992, 31,
5298. (c) Cotton, F. A.; Eglin, J. L.; James, C. A. Inorg. Chem. 1993, 32,
681.
(16) The change in chemical shift, ∆δ, is related to ∆ø by ∆δ ) (∆ø/
4π)[(1 - 3 cos2θ)/3r3], where r is the distance of the test nucleus to the
source of anisotropy and θ is the angle between the r vector and the source
axis.
(17) A significant shielding of the methylene protons in the ethyl groups
adjacent to the tolyl substituent is observed (δ 3.00 and 2.83 vs 4.6-3.8
ppm for 2, Table 1), analogous to shifts in [10]paracyclophane: Agarwal,
A.; Barnes, J. A.; Fletcher, J. L.; McGlinchey, M. J.; Sayer, B. G. Can. J.
Chem. 1977, 55, 2575.

A Dynamic NMR and Resonance Raman Study
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1459
Figure 4. Eclipsed rotational isomers about a quadruple metal-metal
bond showing the meso-proton labeling scheme used in the text. The
two metalloporphyrin halves of the dimer are labeled separately for
each rotamer (anti on the left, gauche on the right). In this work X )
tolyl groups only.
are assigned as Mo H1 and Mo H3 according to the scheme in
Figure 4. This assignment is made on the basis that Mo-
(porphyrin) protons should be downfield of corresponding
W(porphyrin) protons (see Table 1) and protons eclipsed with
the opposing tolyl substituent should be downfield of protons
eclipsed with opposing meso-protons. The relative intensities
of Mo H1 and Mo H3 indicate that H1 belongs to the less
populated anti rotamer, and H3 to the more populated gauche
rotamer.20 Interestingly, W H1 and W H3 have slightly different
chemical shifts at -60 °C (see Figure 2), but accidentally
overlap in the limiting low-temperature spectrum. Nonetheless,
the separate cross-peaks labeled N3 and N4 in Figure 3 identify
their respective locations at δ 9.28 and 9.27 ppm.
Figure 2. Variable-temperature 1H NMR at 500 MHz of 2. Rates
The remaining off-diagonal resonances, labeled E1-E8 in
Figure 3, may be used to identify the chemical shift positions
of the remaining protons in Figure 4. As the anti rotamer
exchanges with the gauche, Mo H1 should exchange with Mo
H5. Figure 3 demonstrates a cross-peak (E3) which relates Mo
H1 to a third resonance buried underneath W H1 and W H3.
Thus, the chemical shift position of Mo H5 is determined to be
δ 9.27 ppm. Likewise, Mo H3 and Mo H4 should become
degenerate at a new chemical shift (Mo H2) as the gauche
rotamer returns to the anti. The NOESY/EXCHSY spectrum
demonstrates an exchange peak (E4) for Mo H3 with the
resonance at δ 8.99 ppm. Because Mo H4 and Mo H2 are both
Hcis and both eclipsed with meso-protons, they exhibit no
noticeable chemical shift difference and the cross-peak relating
Mo H4 to Mo H2 is located along the diagonal (E5).
determined by complete band shape analysis.
Similar reasoning enables assignment of the five W meso
bands. Cross-peak E6 relates W H1 to W H5, and cross-peak
E7 relates W H3 to W H2. Finally, W H4 is also exchanged
with W H2, but in contrast to Mo H4 and Mo H2, the tungsten
protons are slightly resolved at the low-temperature limit.
Nonetheless, the cross-peak relating W H4 and W H2, E8, is
still located near the diagonal. Cross-peaks E1 and E2 are the
result of Ho′ exchange between the two rotamers.
Figure 3. 500 MHz 2D NOESY/EXCHSY spectrum of 2 at -78 °C
(toluene-d8).
The pattern of meso-proton assignments at the low-temper-
ature limit (Figure 5) is solely consistent with population of
the two eclipsed rotamers as shown in Figure 4. The relative
intensities of each peak are also in good agreement with the
calculated values according to populations of 40% anti and 60%
gauche rotamers (Table 2). Further support for the eclipsed
ground-state assignments is provided by the close agreement
of the observed chemical shift difference between H1 and H5
(Mo, 0.45 ppm; W, 0.47 ppm) and the calculated maximum
deshielding expected for a tolyl group which is perfectly eclipsed
with a meso-proton (0.47 ppm).21 These two protons provide
through-space interactions (nuclear Overhauser effect) between
separate protons.18
There are 12 off-diagonal resonances observed in Figure 3.
Of these, four (labeled N1-N4) must be a result of NOE since
they relate an endo-tolyl Ho′ proton to a meso-proton. The
existence of NOE cross-peaks between endo-tolyl Ho′ protons
and meso-protons, but not between endo-tolyl Ho′ protons and
â-ethyl protons, is consistent with our claim that the two low-
temperature limit rotamers are eclipsed.19 The two farthest
downfield meso-proton resonances, at δ 9.71 and 9.52 ppm,
an ideal measure of the deshielding as both H1 and H5 are Htrans
,
(18) A more complete description of the NOESY pulse sequence is
available: States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson.
1982, 48, 286.
(19) The complete NOESY/EXCHSY spectrum is available as Supporting
Information.
(20) H1 and H3 are both eclipsed with opposing tolyl substituents;
however, H1 is trans to its own tolyl substituent and H3 is cis. Studies14 of
similar OEP.-X porphyrin complexes indicate that Htrans protons are located
downfield of Hcis protons.

1460 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Collman et al.
1
Figure 5. Assignment of the limiting low-temperature 500 MHz H
NMR spectrum of 2 according to the scheme in Figure 4.
Figure 6. Eyring plot for rotation about the metal-metal bond axis
in 2.
Table 2. Calculated and Experimental Peak Integrations for the
Low T Spectrum of 2
barriers for the same molecule, and for these reasons its use is
not recommended by the authors.
chemical
shift
calculated peak actual peak
assignment rotamers
integrationa
integration
Initial broadening of the toluene-d8 1H NMR spectrum of 2
9.69
9.49
9.24
Mo H1
Mo H3
Mo H5
W H1
W H3
Mo H2
Mo H4
W H5
W H2
W H4
anti
0.40
0.60
1.60
0.37
0.52
1.55
is seen at -80 ( 5 °C. This value of TI results in a ∆G q
)
rot
gauche
gauche
anti
gauche
2 × anti
gauche
gauche
2 × anti
gauche
10.7 ( 0.3 kcal/mol.22
Complete Band Shape Analysis. The meso, Ho′, and Hm′
protons of 2 provide eight separate spin systems (Hcis, Htrans
,
8.98
1.40
1.41
Ho′, and Hm′ for each metal) that may be analyzed through
complete band shape analysis (CBA). As each spin system is
characterized by a separate set of spectral parameters, a multiple
verification of rate results from the best fit of all simulations at
a given rate. The rates corresponding to each variable-
temperature spectrum have been listed in Figure 2.23 The barrier
8.80
8.70
0.60
1.40
0.52
1.43
a anti ) 0.40, gauche ) 0.60, as determined from the endo-tolyl
resonances.
for interconversion of the anti and gauche rotamers (∆G q
)
rot
10.6 ( 0.1 kcal/mol), determined from an Eyring plot24 of the
best-fit rates versus 1/T (Figure 6), is in excellent agreement
with the low-temperature limiting spectrum analysis. The very
low value of ∆S qrot renders this barrier temperature independent,
but H1 is eclipsed with the tolyl group while H5 is not. This
agreement is very good evidence in support of an eclipsed
geometry between the meso-protons and the tolyl substituents.
The 1.5:1 ratio of gauche:anti eclipsed rotamers in 2 is nearly
identical to the ratios previously determined for 3 and 4.9
Deviation from the purely statistical ratio of 2:1 is consistent
with a slight thermodynamic preference (e 2 kcal/mol) for the
anti conformation. In all three cases the syn rotamer, with tolyl
substituents eclipsed, is not populated to any detectable extent.
Exclusion of the syn rotamer is not surprising since models have
shown the tolyl protons to extend across the metal-metal bond
to within 0.15 Å of its center. Such an orientation would result
in severe tolyl-tolyl repulsive interactions and accounts for the
absence of the syn rotamer in all three metalloporphyrin dimers
studied.
δ-Bond Rotational Barrier. Low-Temperature Limiting
Spectrum Method. A first estimate of the rotational barrier
for 2 is possible through identification of the temperature, TI,
when line broadening of the low-temperature limiting spectrum
just becomes perceptible. This method has been presented by
Faller22 as an alternative to coalescence point analysis (CPA).
CPA is dependent on the chemical shift difference (Hz) between
sites in the absence of exchange, and thus varies with the field
strength of the spectrometer. Not only are a variety of
spectrometer field strengths now in use, but the chemical shift
differences between sites for meso-protons and endo-tolyl
protons at the low-temperature limit are not the same. Coa-
lescence point analysis would yield several different rotational
within experimental error: ∆G q ) ∆H q
.
rot
rot
Vibrational Spectroscopy. The three meso-tolyl octaethyl-
metalloporphyrins (2, 3, and 4) which have been investigated
by dynamic NMR10 (vide supra) have also been characterized
by resonance Raman spectroscopy. In each case enhancement
of a mode assigned as the metal-metal stretch was effected
with UV-excitation (363.8 nm) into the blue side of the dimer
Soret band (see Figure 8). Assignment of the νMM modes was
made on the basis of similar vibrational studies of other
quadruply bonded group VIB dimers,25 depolarization ratios,
and subtraction of the spectrum obtained from the corresponding
metalloporphyrin monomer.26 Enhancement of the metal-metal
stretch in this fashion is consistent with the enhancement
mechanism which has been proposed for axial-ligand vibrational
modes in resonance with the Soret band of many other
metalloporphyrins. A change in the internuclear distance of
the metal-metal bond could occur either via overlap of the a2u
(23) Rates were calculated with the program DNMR 5, obtained from
Dr. Richard Counts at the Quantum Chemical Program Exchange, Indiana
University, and adapted for use on a desktop PC. Transverse relaxation
times, T2, were obtained at each temperature from the line widths in [(OEP)-
MoW(OEP)], which undergoes the same dynamic exchange but between
sites which are magnetically equivalent.
(24) The formula -ln(k/T) ) -ln(ΚkB/h) + ∆G q/RT is the Eyring
equation, where Κ is the transmission coefficient, k the rate, kB the
Boltzmann constant, and h Planck’s constant. See: Sandstrom, J. Dynamic
NMR Spectroscopy; Academic Press: New York, 1982.
(21) meso-Proton deshielding due to the ring current of an aryl substituent
as a function of the angle of rotation about the metal-metal bond was
calculated with the Waugh-Fessenden classical model of ring current
magnetic anisotropy as developed by Johnson-Bovey: Bovey, F. A.
Nuclear Magnetic Resonance; Academic Press: New York, 1988; p 108.
(22) Faller, J. In Encyclopedia of Inorganic Chemistry; King, R. B., Ed.,
Wiley: England, 1994; pp 3914-3933.
(25) (a) Cotton, F. A.; Walton, R. A. Multiple Bonds Between Metal
Atoms; Wiley: New York, 1993; pp 735-739. (b) Tait, C. D.; Garner, J.
M.; Collman, J. P.; Sattelberger, A. P.; Woodruff, W. H. J. Am. Chem.
Soc. 1989, 111, 9072-9077.
(26) RR spectra were also obtained for the monomers Mo(TOEP)(O)
and W(TOEP)(O)(Cl). The modes assigned as νMM in the dimers were absent
from the monomer spectra.

A Dynamic NMR and Resonance Raman Study
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1461
Table 3. Calculated Structural Parameters from Observed M-M
Stretching Frequencies
νM-M
,
k,
bond
cm-1 mdyn/Å D, Åa order
ref
[Mo(OEP)]2
[Mo(TOEP)]2
[(OEP)MoW(OEP)]
[(TOEP)MoW(TOEP)] 278
310
310
279
2.72 2.33
2.72 2.33
2.89 2.30/2.40 4.0 this work
2.87 2.30/2.40 4.0 this work
4.0 this work
4.0 this work
[W(OEP)]2
[Ru(OEP)]2
[Ru(OEP)]2
[Ru(OEP)]2
[Os(OEP)]2
275
285
301
310
233
254
266
4.08 2.24
2.42 2.40
2.70 2.33
2.86 2.30
3.04 2.37
3.46 2.31
3.96 2.25
4.0 this work
2.0 32
2.5 32
+
2+
3.0 32
2.0 25b
2.5 25b
3.0 25b
+
[Os(OEP)]2
[Os(OEP)]2
2+
a The bond lengths (D) have been estimated from an empirical
relationship (ref 28), which gives distinct formulas for second and third
row homodimers. Both results have been presented for the MoW
heterodimers.
Table 3 summarizes the structural parameters calculated from
the metal-metal bond vibrational data for several metallopor-
phyrin dimers. Force constants were obtained from the diatomic
oscillator approximation,27 and metal-metal distances were
estimated using an empirical correlation between force constants
and metal-metal bond lengths.28
Spectra of [(OEP)MoW(OEP)] (1) and [Mo(OEP)]2 (5) were
also obtained to determine the validity of the diatomic ap-
proximation wherein ligand effects (such as mass) are neglected
in the calculation of metal-metal bond force constants. Figure
7 shows the low-frequency region of the RR spectra for 1, 2, 3,
4, and 5. Our results indicate that the substantial difference in
ligand mass between the OEP and TOEP porphyrin dimers has
a relatively small effect on the metal-metal bond stretching
frequency. The shift for the two Mo-Mo stretches is at most
1-2 cm-1, and the shift for the Mo-W stretches is at most
2-3 cm-1, both to lower frequency for the TOEP dimers.29
Several other lines of evidence are also consistent with our
assumption that the metal-metal stretch is essentially M2
localized. The metal-nitrogen force constant is very weak
relative to the metal-metal force constant; thus, mixing of
metal-nitrogen stretching or metal-metal-nitrogen bending
coordinates with the metal-metal stretch coordinates is limited.
The same may not be said for analogous M2X4(PR3)4 dimers,
since in those molecules the metal-halide force constant is
comparable in magnitude with the metal-metal force constant.
To estimate the exact percentage of localized diatomic character
exhibited by the metal-metal stretches in M2(OEP)2 dimers,
we have performed a simple normal coordinate analysis (see
the Supporting Information) with the structural parameters from
our single-crystal X-ray diffraction study of [Ru(OEP)]2.30 The
results from this study indicate that only minor systematic
corrections of a 10% or less reduction in the force constant need
to be applied to the diatomic oscillator approximation used to
describe metal-metal bonding between second and third row
metals in porphyrin systems. These studies and the small mass
Figure 7. Low-frequency RR spectra upon excitation at 363.8 nm for
1, 2, 3, 4, and 5. Samples were prepared in toluene-d8 solution and
flame sealed. Solvent resonances were not subtracted from the raw data,
but are given for comparison. All spectra are baseline corrected. Sample
1
stability was confirmed by H NMR analysis following excitation.
Figure 8. Absorption spectrum for 2 in benzene. All group VIB dimer
electronic absorption spectra are qualitatively identical.
(27) k ) (3.55 × 1017)µν2, where k ) force constant (mdyn/Å), µ )
reduced mass of the two metal atoms (g), and ν ) vibrational frequency
(cm-1).
(28) Miskowski, V. M.; Dallinger, R. F.; Christoph, G. G.; Morris, D.
E.; Spies, G. H.; Woodruff, W. H. Inorg. Chem. 1987, 26, 2127-2132.
The appropriate equation for elements Rb-Xe is DMM ) 1.83 + 1.51
exp(-k/2.48), where D is in Å and k is in mdyn/Å. For elements Cs-Rn,
porphyrin orbital with the metal-metal σ-bond, or via overlap
of the porphyrin 3eg* orbital with the metal-metal π-bonds.
The change in electron density on the porphyrin that ac-
companies the ring-centered π f π* Soret transition may thus
result in a change in electron density and hence bond length of
the metal-metal bonded axial ligand.
DMM ) 2.01 + 1.31 exp(-k/2.36)
.
(29) Quality spectral data were not obtained for W2OEP2, probably due
to its very poor solubility, so a third comparison is not available.
(30) Collman, J. P.; Barnes, C. E.; Swepston, P. N.; Ibers, J. A. J. Am.
Chem. Soc. 1984, 106, 3500-3510.

1462 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Collman et al.
effect seen upon substitution of TOEP for OEP provide good
support for the trend observed in Table 3. The Raman data
indicate that the force constants increase in the order Mo-Mo
< Mo-W < W-W, implying a trend in overall metal-metal
bond strength that places the heterodimer intermediate between
the two homodimers, but significantly weaker than their
calculated average.
In addition to the Soret, the electronic absorption spectra of
2, 3, and 4 exhibit an intense band at 430-440 nm (Figure 8).
Although the metal-metal stretching band is strongly enhanced
with excitation at 363.8 nm, it is not seen upon excitation at
432, 439, 457, 476, or 488 nm (Raman data not shown). These
excitation wavelengths span the electronic transition at 438 nm
and were performed to understand the scattering mechanism
which gives rise to the observed metal-metal stretch. The most
well-supported hypothesis is that the 438 nm band is a ligand-
to-metal charge transfer from the porphyrin HOMO (3eg*) to
the δ* (b1u) orbital, but we have not confirmed the assignment
by identification of an associated MCD A term.
The porphyrin skeletal vibrations observed in the frequency
range 1470-1610 cm-1 in the meso-tolyl-substituted OEP
dimers 2, 3, and 4 are similar upon excitation into either
absorption band. Vibrational assignments of the porphyrin
skeletal modes were inferred from the depolarization ratios and
from comparisons with the vibrational spectra of Ni(OEP)31 and
[Ru(OEP)]2.32 The Ni(OEP) monomer is a very useful standard
for comparison since a complete normal coordinate analysis has
been correlated with the spectra,33 and the vibrational numbering
scheme for OEP vibrations was developed for this compound.
The similarity of metalloporphyrin dimer spectra with metal-
loporphyrin monomer spectra is consistent with previous studies.
The vibrational peak positions of [Ru(OEP)]2 are almost
identical to those in Ru(OEP)(CO),32 and even stacked silicon
phthalocyanine vibrations are not shifted significantly from the
silicon phthalocyanine monomer.34
Figure 9 shows the high-frequency region of the RR spectra
for 2, 3, and 4 obtained with excitation at 432 nm. The extreme
complexity of these molecules has precluded us from performing
complete, high-frequency region normal-mode analyses and thus
being able to assign each of the observed porphyrin skeletal
modes according to the scheme in ref 33. However, the overall
qualitative picture which emerges upon inspection of Figure 9
demonstrates a striking similarity between the high-frequency
spectra of Ni(OEP) and [W(TOEP)]2. In contrast, the dimo-
lybdenum spectrum is significantly more complicated and
exhibits several distinct new features, while the molybdenum-
tungsten spectrum appears as a superposition of the homodimer
spectra. These features are consistent with a greater “bending-
back” distortion of molybdenum-porphyrin entities, while
tungsten-porphyrins resemble the four-coordinate, planar Ni-
(OEP). The mixed-metal dimer appears to exhibit both features
simultaneously.
Figure 9. High-frequency RR spectra obtained for 2, 3, and 4 with
excitation at 432 nm into an electronic charge-transfer band. Solvent
was toluene-d8 for each sample.
ing will be a simple function of distance to the tolyl group.
Figure 1 clearly demonstrates that the W Hcis and W Htrans
protons are much closer in their time-averaged chemical shifts
than the corresponding Mo Hcis and Mo Htrans. We offer this
evidence as further support of our contention that the mixed-
metal dimer is simultaneously composed of one drastically
distorted porphyrin (which is therefore unable to impart as large
an anisotropy to the opposed W Hcis and W Htrans) and one nearly
planar porphyrin (Figure 10).
1
The H NMR spectra of 2, 3, and 4 are also consistent with
this structural difference between molybdenum- and tungsten-
porphyrin complexes. The chemical shift difference of Hcis and
Htrans is dominated by the through-space magnetic anisotropy
of the opposing tolyl group. Assuming exchange between
eclipsed ground state conformations, the magnitude of deshield-
Discussion
We have estimated the δ-bond strengths of isostructural
Mo2,9 W2,9 and MoW porphyrin dimers by measuring the
barrier to rotation about the metal-metal bond. Historically,
δ-bond strengths have been correlated with δ f δ* electronic
transition energies.35 Such studies imply stronger δ-bond
strengths for the dimolybdenum verses ditungsten quadruple
(31) Kitigawa, T.; Abe, M.; Ogoshi, H. J. Chem. Phys. 1978, 69, 4516-
4525.
(32) Tait, C.Drew; Garner, J. M.; Collman, J. P.; Sattelberger, A. P.;
Woodruff, W. H. J. Am. Chem. Soc. 1989, 111, 7806-7811.
(33) Abe, M.; Kitagawa, T.; Kyogoku, Y. J. Chem. Phys. 1978, 69,
4526-4534.
(34) Simic-Glavaski, B.; Tanaka, A. A.; Kenney, M. E.; Yeager, E. J.
Electroanal. Chem. 1987, 229, 285-296.
(35) Cotton, F. A.; Eglin, B. H.; James, C. A. Inorg. Chem. 1993, 32,
2104-2106.

A Dynamic NMR and Resonance Raman Study
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1463
Figure 11. Definition of the δ-bond rotational barrier, ∆G qrot, used
herein.
Figure 10. Structural representations of 2, 3, and 4 as implied by RR
and H NMR results.
1
Table 4. ∆G q in 2, 3, and 4 as Determined by Complete Band
rot
bonds,36 in striking contrast to our results. However, recent
reports37 have indicated that several factors must be considered
in the analysis of δ f δ* electronic transition energies. Such
factors include the amount of metal-ligand δ-bonding,38 the
degree to which configuration interaction is involved in the
ground state,39 and mixing of the δ f δ* transition with metal-
to-ligand charge transfer (MLCT) bands.40 These studies
conclude that simple correlations should not be made between
δ-bond strength and δ f δ* energies.
Measurement of the rotational barrier for quadruply bonded
complexes is an accurate, direct method for estimation of the
δ-bond strength. The σ and π contributions to the quadruple
bond are cylindrically symmetrical, and hence exhibit no angular
preferences for overlap. However, the δ-bond achieves maxi-
mum overlap in an eclipsed conformation, and is completely
nonbonding when the dxy orbitals are staggered at 45°. Since
exchange between the anti and gauche eclipsed conformations
of metalloporphyrin dimers such as 2, 3, and 4 requires a
staggered transition state (see Figure 11), the rotational barrier,
∆G qrot, is an accurate estimation of ∆Gδ in the absence of ligand
steric contributions (∆G qrot ) ∆Gδ - ∆Gs). Previous rotational
barrier studies41 of [Mo(OEP-X)]2 with linear substituents (X
) formyl, isocyanate, and NH2) all indicate a barrier to rotation
which is within experimental error of that observed for [Mo-
(TOEP)]2. Insensitivity of these rotational barriers to the
incorporation of meso-aryl substituents allows an estimation of
∆Gs (e0.5 kcal/mol).42
Shape Analysis
dimer
∆G q
rot
[(TOEP)MoW(TOEP)]
[Mo(TOEP)]2
[W(TOEP)]2
10.6 ( 0.1
10.8 ( 0.1
12.9 ( 0.1
likewise suggest that the ditungsten δ-bond is also stronger than
either dimolybdenum or molybdenum-tungsten (Table 4).
Manning and Trogler were the first to propose that the increased
radial distribution of 5d vs 4d orbitals (a consequence of
relativistic effects) compensates for longer bond lengths and
increased internuclear repulsions inherent to third row metal
dimers.39
As shown in Table 3, empirical equations derived from
vibrational data predict that the W2 and Os2 bonds are shorter
than the metal-metal bonds in isoelectronic Mo2 and Ru2
congeners. Many pairs of homologous Mo2 and W2 dimers have
been structurally characterized, including the tetraphenylpor-
43
phyrin dimers ([Mo(TPP)]2 and [W(TPP)]2),44 and in every
case the W2 bond length is about 0.10 Å longer than the
corresponding Mo2 bond length.25 Although the empirical
equations have proved extremely accurate with the ligand
systems from which they were derived, the equations may
simply not be applicable to porphyrin dimers. Calculated force
constants for the W2 and Os2 porphyrin dimers are 25-35%
larger than the corresponding values for Mo2 and Ru2 dimers.
The magnitude of this difference is by far the largest yet
observed; indeed, the ligand systems used to generate the
empirical equations exhibit force constant changes of only
4-8% upon substitution from Mo-Mo to W-W.25 However,
most of these ligand systems bridge the metal-metal bond, and
thus may be expected to greatly affect the strength of the metal-
metal bond through their own steric restraints and electronic
preferences. The large dependence of force constant on metal
identity in our systems may well be a result of the unbridged,
unrestrained nature of the metal-metal bond. That the metal-
metal stretches in metalloporphyrin dimers are essentially M2
localized is also supported by a detailed analysis given as
Supporting Information. Normal-mode calculations therein
indicate that the contribution of metal-metal motion to the
potential energy distribution (PED) of the vibrations assigned
Comparisons may now be made between the overall qua-
druple bond strengths for Mo2, W2, and MoW as well as
between δ-bond strengths for the same molecules. Metal-metal
quadruple bond force constants indicate that the ditungsten
species, 5, contains the strongest quadruple bond (Table 3). This
result is consistent with our rotational barrier studies, which
(36) Cotton, F. A.; Extine, M. W.; Felthouse, T. R.; Kolthammer, B. W.
S.; Lay, D. G. J. Am. Chem. Soc. 1981, 103, 4040-4045.
(37) Hopkins, M. D.; Gray, H. B.; Miskowski, V. M. Polyhedron 1987,
6, 705-714.
(38) Sattelberger, A. P.; Fackler, J. P. J. Am. Chem. Soc. 1977, 99, 1258-
1259.
(39) Manning, M. C.; Trogler, W. C. J. Am. Chem. Soc. 1983, 105,
5311-5320.
(40) Hopkins, M. D.; Schaefer, W. P.; Bronikowski, M. J.; Woodruff,
W. H.; Miskowski, V. M.; Dallinger, R. F.; Gray, H. B. J. Am. Chem. Soc.
1987, 109, 408-416.
(41) Collman, J. P.; Woo, L. K. Proc. Natl. Acad. Sci. U.S.A. 1984, 81,
2592-2596.
(43) For [Mo(TPP)]2, the Mo-Mo separation is 2.24 Å: Yang, C.-H.;
Dzugan, S.; Goedken, V. L. J. Chem. Soc., Chem. Commun. 1986, 1313-
1315.
(44) For [W(TPP)]2, the W-W separation is 2.33 Å: Kim, J. C.; Goedken,
V. L.; Lee, B. M. Polyhedron 1996, 15, 57-62.
(42) The steric component of the barrier must be less than the difference
in measured barriers of OEP-X derivatives. These barriers are conveniently
tabulated in ref 41.

1464 J. Am. Chem. Soc., Vol. 120, No. 7, 1998
Collman et al.
Figure 12. MO diagrams for homonuclear and heteronuclear metal-metal quadruple bonds.
as νMM are on the order of 80-90%. The nearly complete
isolation of the metal-metal stretch in these molecules may be
understood as a consequence of the relatively large mass of the
metal atoms, relatively strong metal-metal bond, and relatively
weak metal-porphyrin stretching and bending interactions.
Despite the preponderance of evidence suggesting that third
row metals form significantly stronger bonds than second row
metals, in neither the rotational barrier nor the vibrational studies
is the dimolybdenum bond significantly strengthened by sub-
stitution of a single tungsten atom. Table 3 manifests a small
increase in quadruple bond strength for MoW relative to the
Mo2 species; however, Table 4 demonstrates the presence of a
slightly weaker MoW δ-bond.
Molecular orbital theory predicts that the bond strength, ∆E,
of a metal-metal bond will be dominated by two factors. The
interaction energy will be made larger as the atomic orbitals
achieve more geometric overlap via incorporation of more
diffuse orbitals. However, ∆E is a maximum when the initial
energies of the interacting orbitals are identical. Previous
heterodimer studies8 have not directly addressed how the
strength of a heterodimeric metal-metal bond is affected by
the difference in atomic orbital energy levels. Greater elec-
tronegativity of MoII relative to WII lowers the energy levels of
its atomic d-orbitals (Figure 12). Molecular orbital theory
predicts that mismatch of the atomic orbital energies will have
a deleterious effect on the interaction energy for formation of
the heteronuclear bond.
The face-to-face nature of dxy-dxy overlap does not make
effective use of the radially diffuse third row dxy orbital. Thus,
the MoW δ-bond is slightly weakened from the homonuclear
Mo2 congener upon incorporation of the lower energy W 5dxy
orbital. This slight weakening occurs despite the increased
radial distribution of the 5d dxy orbital. In contrast, the σ- and
π-bonds are composed of atomic d-orbitals which have their
distribution maxima directed toward each other, as opposed to
the face-to-face orientation required for δ-bond formation.
Direct orientation of the dz and dxz, dyz orbitals enables a more
2
efficient utilization of the more diffuse 5d orbitals for σ- and
π-bonding, hence, the force constant45 for the overall quadruple
bond in MoW is increased from its value in Mo2. The same
result is seen in the resonance Raman studies of MoW(Cl)4-
(PMe3)4 and Mo2(Cl)4(PMe3)4.7
Repeated attempts to obtain X-ray suitable single crystals of
the quadruple bonded metalloporphyrin dimers have been
unsuccessful. Such a study may be expected to reveal a bond
length for MoW which is slightly increased from the average
lengths of Mo2 and W2. Although not available for the
porphyrin dimers, Morris has reported the crystal structure for
7
MoW(Cl)4(PMe3)4 and Cotton has determined analogous
46
structures of Mo2(Cl)4(PMe3)4 and W2(Cl)4(PMe3)4.46 The
M-M bond distances increase in the order Mo2 (2.130 (0) Å)
< MoW (2.2092 (7) Å) < W2 (2.262 (1) Å). Indeed, the MoW
bond length is approximately 0.015 Å longer than the average
of the homodimers, consistent with our comparison of the three
νMM force constants for 2, 3, and 4.
Conclusions
The rotational barrier (∆G q ) 10.6 ( 0.1 kcal/mol) and
rot
metal-metal bond force constant (k ) 2.87 mdyn/Å) have been
reported for the mixed-metal dimer [(TOEP)MoW(TOEP)]. The
results indicate the existence of competing effects on the
determination of heterometallic bond strengths. Atomic orbital
energy mismatch due to a difference in metal atom electrone-
gativities causes a net loss of Mo dxy-W dxy interaction energy
(45) Estimates indicate that the δ-bond contributes only about 10% of
the overall bond strength; therefore, trends in the quadruple bond force
constants will be primarily indicative of changes in the σ- and π-bond
strengths.
(46) Cotton, F. A.; Extine, M. W.; Felthouse, T. R.; Kolthammer, B.
W.; Lay, D. G. J. Am. Chem. Soc. 1981, 103, 4040.

A Dynamic NMR and Resonance Raman Study
J. Am. Chem. Soc., Vol. 120, No. 7, 1998 1465
and thus a δ-bond which is slightly weaker in MoW than in
either Mo2 or W2. However, the incorporation of radially diffuse
W 5d orbitals increases the geometric overlap between the
Acknowledgment. We thank the NSF (Grant CHE 9123187-
A4) for financial support. Special thanks to the mass spectros-
copy facility at UCSF and Dr. Doris Hung at Northwestern for
mass spectra of 1 and 2 and to Pharmacyclics for their generous
gift of diethylpyrrole used in the syntheses of H2OEP and H2-
TOEP.
2
respective Mo and W dz and dxz, dyz orbitals which form
heterometallic σ- and π-bonds. The result is a metal-metal
bond force constant which is slightly higher for MoW than Mo2,
but still weaker than the average of Mo2 and W2. We now
propose that the unbridged heteronuclear MoW quadruple bond
does not exhibit bonding which may simply be described as an
average of that observed in the homodimers. Our results
indicate that heteronuclear metal-metal bonds are slightly
weaker than the average of the homodimers due to a decrease
in effective orbital mixing resulting from the difference in
constituent atomic orbital energies.
Supporting Information Available: Complete 2D NOESY/
EXCHSY spectrum, calculated and experimental mass spectra
for (OEP)MoW(OEP)PF6 and (TOEP)MoW(TOEP), and normal
coordinate analyses for M2(porphyrin)2 dimers (56 pages). See
any current masthead page for ordering information and Web
access instructions.
JA972130M