How to make a major shift in a Redox potential: Ligand control of the oxidation state of dimolybdenum units

Article abstract of DOI:10.1021/ic025632w

A compound reported earlier (Polyhedron 1989, 8, 2339) as (Buⁿ₄N)₂H₂{Mo₂ [Mo(CO)₄(PhPO₂)₂]₂} has been reexamined. We find that the hydrogen atoms in this formula are not present. Therefore, the complex must be considered as having a central triply bonded Mo₂⁶⁺ unit, instead of a quadruply bonded Mo₂⁴⁺ unit. Our conclusion is based on a variety of experimental evidence, including X-ray crystal structures of four crystal forms, as well as the neutron crystal structure of one. This explains the relatively long Mo-Mo bond lengths found in the range 2.1874(7)-2.2225(7) A and the absence of a δ → δ* transition in the visible spectrum. From electrochemistry we also find that the diphosphonate ligand has such an exceptional ability to stabilize higher oxidation states that even common solvents such as CH₂CI₂ and C₂H₂OH readily oxidize the Mo₂⁴⁺ unit that is introduced from the Mo₂(O₂-CCH₃)₄ or [Mo₂(O₂CCH₃)₂ (NCCH₃)₆](BF₄)₂ employed in the preparation. The only chemically reversible wave at E_1/2 = -1.54 V vs Ag/AgCl corresponds to the reduction process Mo₂⁶⁺ → Mo₂⁵⁺.

Full text of DOI:10.1021/ic025632w

Inorg. Chem. 2002, 41, 4232−4238
How to Make a Major Shift in a Redox Potential: Ligand Control of the
Oxidation State of Dimolybdenum Units
,†
F. Albert Cotton,* Lee M. Daniels,† Chun Y. Liu,† Carlos A. Murillo,*,†,‡ Arthur J. Schultz,§ and
Xiaoping Wang†,§
Laboratory for Molecular Structure and Bonding, Department of Chemistry,
Texas A&M UniVersity, P.O. Box 30012, College Station, Texas 77842-3012, Department of
Chemistry, UniVersity of Costa Rica, Ciudad UniVersitaria, Costa Rica, and Intense Pulsed
Neutron Source, Argonne National Laboratory, Argonne, Illinois 60439
Received April 3, 2002
A compound reported earlier (Polyhedron 1989, 8, 2339) as (Bun4N)2H {Mo2[Mo(CO)4(PhPO ) ] } has been
2
2 2 2
reexamined. We find that the hydrogen atoms in this formula are not present. Therefore, the complex must be
considered as having a central triply bonded Mo26+ unit, instead of a quadruply bonded Mo24+ unit. Our conclusion
is based on a variety of experimental evidence, including X-ray crystal structures of four crystal forms, as well as
the neutron crystal structure of one. This explains the relatively long Mo−Mo bond lengths found in the range
2.1874(7)−2.2225(7) Å and the absence of a δ f δ* transition in the visible spectrum. From electrochemistry we
also find that the diphosphonate ligand has such an exceptional ability to stabilize higher oxidation states that even
common solvents such as CH Cl2 and C H OH readily oxidize the Mo24+ unit that is introduced from the Mo2(O -
2
2
5
2
CCH ) or [Mo2(O CCH ) (NCCH ) ](BF ) employed in the preparation. The only chemically reversible wave at
3 4
2
3 2
3 6
4 2
E
1/2 ) −1.54 V vs Ag/AgCl corresponds to the reduction process Mo26+ f Mo25+.
Introduction
(hpp ) the anion of 1,3,4,6,7,8-hexahydro-2H-pyrimido[1,2-
a]pyrimidine) has revealed that there is a smooth increase
in the Mo-Mo distance as the bond order decreases from 4
to 3.5 to 3, with the metal-metal separation varying from
Over a decade ago a compound was reported1 to which
an Mo24+ core was assigned. On this basis, it would have to
contain a quadruple bond. However, the reported Mo-Mo
bond length of 2.186(2) Å was described by the authors
themselves as “slightly longer than any other Mo-Mo
quadruple bond spanned by four bridging ligands”. Today,
with over 470 structures of compounds having (or believed
to have) quadruple bonds known,2 this reservation is even
more justifiable, given the fact that the bridging ligands are
of the 3-atom type that do not impose any twist on the Mo2X8
core. Thus, on the basis that in any group of data, the
genuineness of one point that is a distinct outlier should
always be suspect, reexamination of this compound seemed
worthwhile. Furthermore, recent preparation of authentically
2+
2.142(2) to 2.1722(9) Å for the Mo2(hpp)4 units.
Moreover, the very formulation of the compound,
(Bun N)2H2{Mo2[Mo(CO)4(PhPO2)2]2}, is also disquieting.
4
The presence of the two hydrogen atoms is necessary to make
4+
the Mo2 unit Mo2 rather than Mo26+, but where are these
two hydrogen atoms located? The reported X-ray structure
gave no hint and the point was not discussed. On the other
1
hand, it was said that the H NMR spectrum “revealed the
presence of two protons (δ 2.79 in d6-acetone)”. It should
finally be noted that the compound in question was prepared
by what appeared to be a simple metathesis from Mo2(O2-
CCH3)4, consistent with there being no oxidation. How then
could oxidation have occurred? We have reinvestigated this
chemistry and resolved all these, and other inconsistencies.
+
+,2+
oxidized species2,3 Mo2(carboxylate)4 and Mo2(hpp)4
* To whom correspondence should be addressed. E-mail: cotton@
tamu.edu (F.A.C.); murillo@tamu.edu (C.A.M.).
† Texas A&M University.
Experimental Section
‡ Ciudad Universitaria.
§ Argonne National Laboratory.
Materials. All the solvents used were dried and distilled under
(1) Wong, E. H.; Valdez, C.; Gabe, E. V.; Lee, F. L. Polyhedron 1989,
8, 2339.
(2) Cotton, F. A.; Daniels, L. M.; Hillard, E. A.; Murillo, C. A. Inorg.
Chem. 2002, 41, 1639.
N2 by following standard procedures. All the chemicals were used
(3) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Timmons, D. J.;
Wilkinson, C. C. J. Am. Chem. Soc. 2002, in press.
4232 Inorganic Chemistry, Vol. 41, No. 16, 2002
10.1021/ic025632w CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/29/2002

How to Make a Major Shift in a Redox Potential
Table 1. X-ray Crystallographic Data for 1
1a
1b
1c ()1‚2CH2Cl2)
1d ()1‚3.43THF)
empirical formula
fw
space group
a, Å
b, Å
C64H92Mo4N2O16P4
1653.04
P21/n (No. 14)
12.9530(8)
15.785(1)
18.086(1)
90
98.782(1)
90
3654.6(4)
2
C64H92Mo4N2O16P4
1653.04
Pbca (No. 61)
16.865(1)
20.861(1)
20.882(1)
90
90
90
7346.7(9)
4
C66H96Cl4Mo4N2O16P4
1822.89
C
77.72H119.44Mo4N2O19.43P4
1900.35
P21/c (No.14)
14.214(1)
14.871(1)
22.103(2)
90
92.879(1)
90
4666(2)
2
P1h (No. 2)
12.181(1)
12.332(1)
14.399(1)
86.699(2)
72.836(2)
81.031(2)
2041.3(3)
1
c, Å
R, deg
â, deg
γ, deg
V, Å3
Z
T, K
213
213
213
213
λ, Å
0.71073
1.502
0.820
0.71073
1.495
0.816
0.71073
1.483
0.869
0.087 (0.246)
0.71073
1.353
0.655
d
calcd, g/cm3
µ, mm-1
R1a (wR2b)
0.033 (0.062)
0.062 (0.13)
0.079 (0.167)
a R1 ) ∑||Fo| - |Fc||/∑|Fo|. b wR2 ) [∑[w(Fo - Fc )2]/∑[w(Fo )2]]1/2
.
2
2
2
Table 2. Neutron Crystal Data and Structure Refinement Parameters
for 1a
as purchased; cis-tetracarbonylbis(piperidine)molybdenum4 and cis-
diacetatohexaacetonitriledimolybdenum tetrafluoroborate5 were pre-
pared according to literature methods.
empirical formula
fw
C64H92Mo4N2O16P4
1653.04
Physical Measurements. The infrared spectrum was recorded
space group
a, Å
b, Å
P21/n
1
on a Perkin-Elmer 16PC FT-IR spectrophotometer. H NMR and
12.952(2)
15.706(2)
17.948(3)
90
98.78(1)
90
31P NMR spectra were recorded on a Varian Unity 300 NMR
spectrometer; for 1H NMR, chemical shifts were referenced to TMS
(δ ) 0.00 ppm) or the signals of the deuterium solvents; for 31P
NMR, chemical shifts were referenced to 85% H3PO4. The
electronic spectrum of an acetonitrile solution of the complex (0.05
mM) was measured on a Cary 17 spectrophotometer. The cyclic
voltammogram of the complex (1.0 mM in THF) was recorded on
c, Å
R, deg
â, deg
γ, deg
V, Å3
Z
3608.3(9)
2
T, K
20.0(1)
1.522
d
calcd, g/cm3
a BAS 100 electrochemical analyzer with Bun NPF6 (0.1 M)
4
size, mm3
1.7 × 1.4 ×1.2
neutrons
time-of-flight Laue
1.306 + 0.885 λ
0.109
electrolyte, Pt working and auxiliary electrodes, a Ag/AgCl
reference electrode, and a scan rate of 100 mV/s.
radiation
data collection technique
µ(λ), cm-1
R(F)a
X-ray Structure Determinations. Suitable single crystals were
mounted on the tips of quartz fibers and then attached to a
goniometer head. X-ray diffraction data were collected on a Bruker-
SMART 1000 CCD area detector system at 213 K. All of the
structures were determined by direct methods and refined with
SHELXS6 and SHELXL-977 software programs, respectively. Non-
hydrogen atoms were refined with anisotropic displacement pa-
rameters, and hydrogen atoms were added in the calculated positions
and refined isotropically. A summary of crystallographic data for
the X-ray structures is given in Table 1.
Neutron Structure Determination. Neutron diffraction data
were collected from a single crystal of 1a at the Intense Pulsed
Neutron Source (IPNS), Argonne National Laboratory. Data col-
lection and analysis were performed according to published
procedures.8-10 The sample temperature was controlled at 20 K by
a Displex closed-cycle helium refrigerator (Air Products and
Chemicals, Inc., Model CS-202). An autoindexing algorithm11 was
used to obtain an initial orientation matrix from the peaks in one
Rw(F)b
0.070
∆Fmax, ∆Fmin
0.240, -0.257
a R(F) ) ∑||Fo| - |Fc||/∑|Fo|. b Rw(F) ){∑[w(Fo - Fc)2]/∑[w(Fo)2]}1/2
.
histogram. For intensity data collection, a total of 32 histograms
were measuredseach with a different diffractometer ø and φ angles
in order to cover a quadrant of reciprocal space. The counting time
for each histogram was 6 h and a total of 192 h (8 d) were used for
data collection. Bragg peaks were integrated in three dimensions
about their predicted locations and were corrected for the incident
neutron spectrum and the detector efficiency. Lorentz and wavelength-
dependent spherical absorption corrections were applied but sym-
metry-related reflections were not averaged (because of the
wavelength dependence of extinction). Non-hydrogen atomic
coordinates from the X-ray analysis were used as the starting
solution for the neutron structure. The GSAS software package12
was used for structural analysis. The refinement was based on F
using 5261 reflections with |Fo| > 2σ(Fo) and a minimum d-spacing
(4) Darensbourg, D. J.; Kump, R. L. Inorg. Chem. 1978, 17, 2680.
(5) (a) Cotton, F. A.; Reid, A. H., Jr.; Schwotzer, W. Inorg. Chem. 1985,
24, 3965. (b) Pimblett, G.; Garner, C. D.; Clegg, W. J. Chem. Soc.,
Dalton Trans. 1986, 1257.
(6) Sheldrick, G. M. SHELXS, program for crystal structure solution. In
Acta Crystallogr. 1990, A46, 467.
(7) Sheldrick, G. M. SHELXL97; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(8) Schultz, A. J.; Van Derveer, D. G.; Parker, D. W.; Baldwin, J. E.
Acta Crystallogr. 1990, C46, 276.
2
of 0.60 Å. Weights were assigned as w(Fo) ) [2Fo/(σ(Fo ) +
2 2
2
(0.002Fo ) )]2, where σ(Fo ) is the variance based on counting
statistics. Non-hydrogen atoms were refined isotropically. All
hydrogen atoms were refined with anisotropic atomic displacement
parameters at the final stage. A summary of parameters related to
the neutron diffraction analysis is presented in Table 2.
Preparation of cis-(CO)4Mo(PhPCl2)2. The ligand precursor
was prepared by modifying a published procedure4 as follows: To
(9) Schultz, A. J. Trans. Am. Cryst. Assoc. 1993, 29, 29.
(10) Figgis, B. N.; Sobolev, A. N.; Young, D. M.; Schultz, A. J.; Reynolds,
P. A. J. Am. Chem. Soc. 1998, 120, 8715.
(12) Larson, A. C.; Von Dreele R. B. GSAS General Structural Analysis
System; Los Alamos National Laboratory: Los Alamos, NM, 2000.
(11) Jacobson, R. A. J. Appl. Crystallogr. 1986, 19, 283.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4233

Cotton et al.
a suspension of cis-tetracarbonylbis(piperidine)molybdenum (1.52
g, 4.02 mmol) in 20 mL of CH2Cl2 was added dropwise dichlo-
rophenylphosphine (1.63 g, 9.11 mmol) with vigorous stirring. A
yellow solution developed as the yellow solid of the starting material
disappeared. The solution was then heated to reflux for about 15
min until the color turned to orange red. After removal of the solvent
under reduced pressure, 5 mL of tetrahydrofuran was added by
syringe. The mixture was stirred for 10 min, then filtered to remove
a white crystalline solid. Removal of the solvent from the filtrate
under vacuum afforded an oily, orange, crude product, cis-
(CO)4Mo(PhPCl2)2, which was used for the next reaction step
without further purification.
by layering a CH2Cl2 solution of 1 with hexanes. These were
separated manually. The crystals were mainly 1a (red brown), some
1b (brown), and a few 1c (pale brown). Orange-red crystals of 1d
were obtained by slow diffusion of hexanes into a THF solution of
1. Crystals of 1a and 1b contained unsolvated molecules while 1c
contained two interstitial CH2Cl2 molecules per formula weight
(1c ≈ 1‚2CH2Cl2). 1d crystallized with 3.43 THF molecules per
formula weight (1d ≈ 1‚3.43THF); two of these are weakly
coordinated in axial positions and the remaining 1.43 are disordered
and occupying crystal interstices. Large crystals of 1a, used for
neutron diffraction, crystallized after slow diffusion of ethyl ether
into an CH3CH2OH/CH3CN (5:1, v:v) solution of 1.
Preparation of (Bun N)H3[(CO)4Mo(PhPO2)2]. The ligand was
4
prepared by modifying a published procedure13 as follows: The
material prepared above was dissolved in 15 mL of CHCl3. With
vigorous stirring, an aqueous solution of NaOH (1.50 g, 37.6 mmol,
in 50 mL of H2O) was added dropwise over a period of 1 h. The
layers were separated. After the acidity of the aqueous solution
was adjusted to a pH of 8, the mixture was filtered; the pH was
readjusted to 7. Again with vigorous stirring, tetrabutylammonium
bromide (12.0 g, 37.2 mmol) was added in small portions. A white
precipitate formed immediately. The solid was separated by
filtration, washed with water and ether, and then dried under
Results and Discussion
As noted in the Introduction, the Mo-Mo distance of
2.186(2) Å in the anion of 1 was recognized from the
beginning to be surprisingly long for a quadruple bond.1 A
recent survey of all Mo&Mo bond lengths reinforced that
perception,2 especially since there was bridging by three-
atom groups, which usually tends to shorten rather than
lengthen metal-to-metal bonds. In view of the fact that the
reported Mo-Mo distance of 2.186(2) Å would not be
surprising for a triple bond,3 we set out to see whether that
might be the case. To accomplish this, there were three major
tasks:
(1) To show that there is no structural basis for postulating
the presence of two hydrogen atoms per formula unit. To
prove a negative is always difficult, though not always
impossible. One approach is to make the positive so difficult
to accept that it is, effectively, disproved.
1
vacuum. Yield: 1.25 g (42.6%). H NMR δ (ppm in acetone-d6):
7.80 (m, 4H, phenyl), 7.32 (t, 6H, phenyl), 3.48 (t, 8H, -CH2-),
1.85 (p, 8H, -CH2-), 1.47 (h, 8H, -CH2-), 1.01 (t, 12H, -CH3).
Preparation of (Bun N)2Mo2[Mo(CO)4(PhPO2)2]2, 1. Method
4
A. Purple [Mo2(O2CCH3)2(NCCH3)6](BF4)2 (0.146 g, 0.200 mmol)
was mixed with colorless (Bun N)H3[(CO)4Mo(PhPO2)2] (0.312 g,
4
0.425 mmol) in 20 mL of acetonitrile and the mixture was stirred
for 30 min. The resulting dark brown solution was transferred to
another flask containing 0.450 g of anhydrous K2CO3. After 4 h
the color of the solution had turned to brown. The solvent was
then removed under vacuum, giving a red-brown residue. The
residue was extracted with 15 mL of dichloromethane. After
filtration, the solution was layered with hexanes, affording 0.178 g
(53.8%, based on the ligand) of red-brown crystals after 4 days.
(2) To examine other types of physical data to show that
they are consistent with the presence of a triple bond, but
not with the presence of a quadruple bond.
(3) To explain how the postulated quadruple bond became
a triple bond.
Method B. To a mixture of (Bun N)H3[(CO)4Mo(PhPO2)2] (0.240
4
1. Structural Results. The compound originally reported
has been prepared in the same crystalline form (1b) and in
two others, 1a and 1c. Also we have prepared a fourth
crystalline variant, 1d, in which there are two axially
coordinated THF molecules. All four structures have been
solved and refined with X-ray diffraction data, the structures
being shown in Figures 1, 2, 3, and 4, respectively, and the
structure of 1a has also been refined with neutron diffraction
data (Figure 5). The salient feature of all these results is that
all of them are very similar to that reported earlier. For
example, bond distances and angles in the forms 1a, 1b, and
1c are the same within three standard deviations. For 1d there
are differences due to the presence of axial THF molecules.
This strongly supports that idea that we are dealing with the
same compound as that originally reported.
g, 0.327 mmol), Mo2(O2CCH3)4 (0.065 g, 0.152 mmol), and Bun -
4
NBH4 (0.620 g, 2.41 mmol) was added 15 mL of tetrahydrofuran.
The resulting solution was stirred overnight. Addition of 30 mL of
toluene to the reaction solution produced a brown precipitate that
was washed with 5 mL of hexanes. After addition of 5 mL of
ethanol, the solid dissolved and then a brick-red crystalline solid
formed very quickly. The product was separated by filtration,
washed with a small amount of ethanol, and dried under vacuum,
yielding 0.120 g (47.8%).
Spectroscopic Characterization of 1. IR (cm-1, KBr): 3055
(m, sharp, H-C stretching, phenyl), 2968 (H-C stretching,
methylene), 2869.5 (H-C stretching, methyl), 2000, 1881, 1846
(s, broad, C-O stretching, carbonyl). 1H NMR δ (ppm in acetone-
d6): 8.12 (d, 8H, phenyl), 7.35 (t, 12H, phenyl), 3.34 (t, 16H,
-CH2-), 1.73 (p, 16H, -CH2-), 1.35 (h, 16H, -CH2-), 0.94 (t,
1
24H, -CH3). H NMR δ (ppm in CD2Cl2): 8.06 (d, 8H, phenyl),
Since hydrogen atoms cannot be reliably determined from
X-ray diffraction data we turned to neutron diffraction,
checking very carefully for possible signs of such atoms.
Special attention was given to the region around the oxygen
atoms of the cis-Mo(CO)4(PhPO2)2n- entity. Chemically this
would be the most likely place for them to be. Due to the
7.43 (t, 12H, phenyl), 2.81 (t, 16H, -CH2-), 1.38 (p, 16H,
-CH2-), 1.21 (h, 16H, -CH2-), 0.90(t, 24H, -CH3). 31P NMR
δ (ppm in acetone-d6): 219.3 (s). 31P NMR δ (ppm in CD2Cl2):
222.73 (s). UV-visible spectrum λmax (nm, in acetonitrile): 373
(ꢀ ) 4.8 × 103 M-1 cm-1).
Growth of Crystals for Structural Analysis. Three crystal
forms, 1a, 1b, and 1c, were obtained from the same Schlenk tube
1
negative neutron scattering length of H, hydrogen atoms
appear as negative troughs in neutron Fourier maps. For 1a,
(13) Wong, E. H.; Turnbull, M. M.; Hutchinson, K. D.; Valdez, C.; Gabe,
E. J.; Lee, F. L.; Le Page, Y. J. Am. Chem. Soc. 1988, 110, 8422.
we found that the largest negative residual in the difference
4234 Inorganic Chemistry, Vol. 41, No. 16, 2002

How to Make a Major Shift in a Redox Potential
Figure 1. The dianion in (Bun N)2Mo2[Mo(CO)4(PhPO2)2]2, 1a, drawn
4
Figure 3. The dianion {Mo2[Mo(CO)4(PhPO2)2]2}2- in 1‚2CH2Cl2, 1c.
The two halves of the anion are related by a crystallographic inversion
center. Displacement ellipsoids are drawn at the 50% probability level. The
Mo(1)-Mo(1A) distance is 2.1925(6) Å.
with displacement ellipsoids at the 50% probability level. There is an
inversion center at the midpoint of the triply bonded Mo2 unit. The Mo-
(1)-Mo(1A) distance is 2.1902(3) Å and the four independent Mo-
Ophosphonate distances are 1.993(1), 1.998(1), 2.002(1), and 2.014(1) Å.
Figure 4. The dianion in (Bun N)2Mo2[Mo(CO)4(PhPO2)2]2‚3.43THF, 1d.
Figure 2. The dianion {Mo2[Mo(CO)4(PhPO2)2]2}2- in 1b. An inversion
center is located at the midpoint of the triply bonded dimetal unit.
Displacement ellipsoids are shown at the 50% probability level. The Mo-
(1)-Mo(1A) distance is 2.1874(7) Å.
4
There are two weakly coordinated THF molecules occupying axial positions
of the Mo26+ unit (Mo‚‚‚O(9) ) 2.657 Å) which cause a small but significant
elongation of the dimolybdenum triple bond (Mo(1)-Mo(1A) ) 2.2225-
(7) Å). Displacement ellipsoids are given at the 40% probability level.
Fourier map for the neutron refinement is only ca. 20% of
the height of a hydrogen atom. A plot of the Mo2O8 core
and surroundings shown in Figure 6 clearly establishes the
absence of additional peaks. Thus, there is no evidence for
the two alleged hydrogen atoms. Without these two hydrogen
distances were 2.46(1) and 2.52(1) Å.14 A comparison of
the molecular dimensions of 1 with those of the quadruply
bonded K4Mo2(SO4)4‚2H2O,15 the monooxidized analogue
K3Mo2(SO4)4‚3.5H2O,15 and that of Cs2[Mo2(HPO4)4(H2O)2]
reveals an interesting pattern as shown in Table 4. Clearly,
the Mo-Mo distance increases as the bond order decreases
from 4 to 3.5 to 3 and the electronic configuration changes
from σ2π4δ2 to σ2π4δ to σ2π4, respectively. An opposite trend
is observed for the Mo-Ooxoanion distances: these decrease
as the oxidation state increases.
In summary, the values of molecular dimensions of
compound 1 in all its forms appear to be consistent with the
presence of a triple bond, but not a quadruple bond.
2. Other Characterization Data. We may begin with the
analytical data reported not for 1b but for the so-called
atoms, the central Mo2 unit must be classified as Mo26+ and
4+
not Mo2
.
Selected bond distances, given in Table 3, show that the
Mo-Mo distances from X-ray data at 213 K for the three
nonaxially ligated crystalline forms (1a, 1b, and 1c) vary
within the narrow range of 2.1874(7)-2.19255(6) Å (com-
pare to 2.186(2) Å reported earlier1). The distance determined
from neutron data for 1a, collected at the much lower
temperature of 20 K, 2.178(8) Å, is ca. 1% shorter. For 1d
there are two axially coordinated THF molecules, and the
Mo-OTHF distance is 2.657(3) Å. This induces a small but
significant (1%) lengthening of the Mo-Mo distance to
2.2225(7) Å. Interestingly, for [Mo2(HPO4)4(H2O)2]2- the
MotMo distance was 2.223(2) Å and the Mo-Owater
(14) Bino, A.; Cotton, F. A. Inorg. Chem. 1979, 18, 3562.
(15) Cotton, F. A.; Frenz, B. A.; Pedersen, E.; Webb, T. R. Inorg. Chem.
1975, 14, 391.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4235

Cotton et al.
Figure 5. Perspective view of the anion in (Bun N)2Mo2[Mo(CO)4-
4
(PhPO2)2]2, 1a, as determined from neutron diffraction data. Atoms are
drawn at their 80% thermal probability level. The Mo(1)-Mo(1A) distance
is 2.178(8) Å.
sodium analogue (Bun N)2Na2{Mo2[Mo(CO)4(PhPO2)2]2}.1
4
To the level of the accuracy of the experimental data, it is
impossible to distinguish from the C, H, N elemental analyses
between this formula, the same formula with H2 in place of
Na2, or the same formula with neither H2 nor Na2.
(a) 1H NMR Spectrum. The presence of H2 in the formula
was reported to be confirmed by a peak at δ 2.79 ppm in
acetone-d6 corresponding to two protons. We have been
unable to reproduce the result, but we have noticed that there
is a signal at 2.85 ppm when water is present in the
1
deuterated acetone. The rest of our H NMR spectrum in
acetone is completely consistent with that reported for the
supposedly (Bun N)2H2{Mo2[Mo(CO)4(PhPO2)2]2} formula.
4
Figure 6. Plots of neutron Fourier maps of the two mutually perpendicular
planes defined by the two metal atoms and the oxygen atoms of trans
phosphonate groups showing the absence of H atoms in the dianion of 1.
For contrast, see the plot of one of the PC6H5 groups, given in Supporting
Information, which clearly shows each of the 5 hydrogen atoms of the phenyl
group.
The spectrum is equally consistent when run in CD2Cl2, in
which case a signal due to any water present appears at ca.
1.54 ppm. When the two deuterated solvents are not properly
dried, the amount of water present varies from run to run,
depending on the batch used. In the original report, water
may well have been, by chance, close to the concentration
required to suggest that two additional hydrogen atoms were
present in the compound.
Table 3. Selected Bond Lengths (Å) for 1a
neutronb
X-rayc
1a
1a
1b
1c
1d
(b) 31P NMR Spectrum. We have confirmed the original
report that the compound displays only a single 31P reso-
nance. If two additional hydrogen atoms were present, the
only reasonable place for them to be would be on two of
the PhPO2 oxygen atoms. However, this would mean that
not all phosphorus atoms could be equivalent. To rule out
the possibility that these protons might have been removed
MotMo
2.178(8) 2.1902(3) 2.1874(7) 2.1925(6) 2.2225(7)
Mo-Ophosphonate 2.002[1] 2.002[8] 2.004[3] 2.006[3] 2.011[3]
Mo-OTHF
Mo0-Cd
Mo0-P
2.657(3)
2.017[2] 2.016[6] 2.015[5] 2.016[5] 2.013[6]
2.515[6] 2.508[8] 2.515[1] 2.488[1] 2.524[2]
a Numbers in square brackets are the standard deviations for average
distances. b 20 K. c 213 K. d Mo0 represents the molybdenum atom in the
phosphonate ligand.
and solvated by acetone molecules, we also recorded the 31
P
spectrum in CH2Cl2, but here again, only one 31P signal was
seen.
near this energy. There is, however, a well-defined band at
380 nm (ca. 26 000 cm-1, which is much too high to be a
δ f δ* transition) that can be assigned to a π f δ transition
according to the analysis of the spectrum of [Mo2(HPO4)4]2-
given by Hopkins et al.17 Thus, the electronic spectrum
furnishes evidence that the δ orbital is empty.
(c) Electronic Absorption Spectrum. The spectrum is
shown in Figure 7. A characteristic δ f δ* transition that
is diagnostic for the Mo24+ core might be expected to occur
approximately where it does in [Mo2(SO4)4]4-, which is to
say around 520 nm.16 Clearly, there is no absorption at or
(17) Hopkins, M. D.; Miskowski, V. M.; Gray, H. B. J. Am. Chem. Soc.
1986, 108, 959.
(16) Bowen, A. R.; Taube, H. Inorg. Chem. 1974, 13, 2245.
4236 Inorganic Chemistry, Vol. 41, No. 16, 2002

How to Make a Major Shift in a Redox Potential
Table 4. A Comparison of the Molecular Dimensions of Ionic Mo2 Species Bound to Oxoanionsa,b
n+
compd
Mo-Mo
Mo-Ooxoanion
Mo‚‚‚Lax
2.591(4)
2.550(4)
2.545(3)
2.52(1), 2.46(1)
2.910(1)
Mo-Mo bond order
ref
K4Mo2(SO4)4‚2H2O
2.111(1)
2.167(1)
2.162(1)
2.223(2)
2.232(1)
2.1902(3)
2.1874(7)
2.1925(6)
2.2225(7)
2.139[4]
2.067[3]
2.062[3]
2.013[10]
2.003[4]
2.002[8]
2.004[3]
2.006[3]
2.011[3]
4
15
15
15
14
14
K3Mo2(SO4)4‚3.5H2Oc
3.5
3.5
3.0
3.0
Cs2Mo2(HPO4)4‚2H2O
(pyH)3[Mo2(HPO4)4Cl]
1a
1b
1c
1d
this work
this work
this work
this work
2.657(3)
a Distances in Å. b Numbers in square brackets are the standard deviations for average distances. c Two independent molecules.
Figure 8. The cyclovoltammogram of 1 in THF, with the potential in
volts, referenced to Ag/AgCl, and showing the only chemically reversible
wave at E1/2 ) -1.54 V corresponding to the Mo26+/Mo25+ couple. Under
these conditions, the E1/2 for the ferrocene/ferrocenium couple appeared at
0.620 V.
Figure 7. Electronic spectrum of 1 in acetonitrile showing the absence of
the δ f δ* transition, expected for a quadruply bonded Mo24+ unit but not
4+
expected for a triply bonded Mo2 species, in the region around 520 nm.
The band at 380 nm corresponds to a π f δ transition (see text).
3. How Is the Triple Bond Formed? If it is now accepted
that there is overwhelming experimental evidence in favor
of 1 containing an Mo-Mo triple bond not a quadruple bond,
then the conversion of the Mo&Mo in the starting material,
e.g., Mo2(O2CH3)4, to an MotMo bond in the anion of 1
must be accounted for to close the story. At no point in the
preparation of the various crystalline forms of 1 was any
reagent present that would be ordinarily identified as an
oxidant. Also, all possible precautions were taken to exclude
oxygen.
The answer to the question is that the necessary oxidants
are the solvents, CH2Cl2 or C2H5OH. The ability of all
chloroaliphatic compounds to act as oxidizing agents is well-
known.3,18 Ethanol can react acidicly to generate hydrogen,
as it will do with any sufficiently electropositive coreactant,
e.g., Na. An electrochemical study of the anion of 1 in
tetrahydrofuran (Figure 8) showed that there is only one
chemically reversible one-electron wave at the very negative
solvents it must be impossible to isolate the reported 4-
anion (containing Mo24+), but only the 2- anion (containing
Mo26+). It appears that the key feature for the stabilization
of the highly oxidized Mo2 unit by cis-[Mo2(CO)4(Ph-
PO2)2]4- anions is the large amount of negative charge, 8-,
6+
distributed over the 8 oxygen atoms of the 2 anions.
Final Observations
This work has made it evident that the quadridentate anion
cis-[Mo(CO)4(PhPO2)2]4- has a great capacity to stabilize
the Mo26+ core relative to Mo24+. This remarkable property
(shared, no doubt, by the [Cr(CO)4(PhPO2)2]4- and [W(CO)4-
(PhPO2)2]4- ions) was not anticipated. It is very interesting
that the only other ligand to share this property, as recently
reported,3,19 is the hpp- ligand. The two are shown side by
side for comparison as I and II, which raises the question:
what do they have in common? We propose that they have
nothing in common and achieve their effect by different
means.
potential of -1.54 V vs Ag/AgCl that corresponds to a
5+
reduction process that has been assigned as Mo26+ f Mo2
.
The guanidinate anion in hpp- is a very strong base, both
Clearly the oxidation to Mo26+ is highly favorable. Further-
more, at such a potential CH2Cl2 and C2H5OH are not stable
in the σ sense and in the π sense.20 It therefore stabilizes the
6+
4+
more highly charged Mo2 core by transferring negative
against reduction. Consequently, the oxidation of the Mo2
charge to it. The tetranegative anion I stabilizes the more
highly charged Mo26+ by offsetting positive charge electro-
statically with its own large negative charge.
units in the starting materials takes place. Thus, from these
(18) (a) Canich, J. A. M.; Cotton, F. A.; Dunbar, K. R.; Falvello, L. R.
Inorg. Chem. 1988, 27, 804. (b) Save´ant, J.-M. Acc. Chem. Res. 1993,
26, 455. (c) Pause, L.; Robert, M.; Save´ant, J.-M. J. Am. Chem. Soc.
2000, 122, 9829. (d) Save´ant, J.-M. J. Am. Chem. Soc. 1987, 109,
6788.
(19) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.; Timmons, D. J. Chem.
Commun. 1997, 1449.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4237

Cotton et al.
occur, thereby lessening the capacity of the phosphate ligand
to stabilize the higher oxidation state. The [Ru2(SO4)4(H2O)2]2-
ion is another example22 of the ability of a highly charged
set of bridging ligands to stabilize the core Ru26+, which is
usually unstable relative to the Ru25+ and Ru24+ cores. This
same dichotomy can sometimes be found in mononuclear
chemistry, where, for example, formally hexavalent tungsten
can be stabilized by six electron-donating methyl groups in
W(CH3)6 or six electron-withdrawing fluorine atoms in WF6.
Further discussion will be provided in an upcoming publica-
tion.3
Acknowledgment. We thank the National Science Foun-
dation for financial support. The work at Argonne National
Laboratory was supported by the Department of Energy,
Basic Energy Sciences-Materials Sciences, under Contract
No. W-31-109-ENG-38.
This is not the first example of these alternative ways of
6+
stabilizing a highly charged central ion. For the Mo2 ion
itself there are other examples. The hpp- anion, II, stabilizes
6+
the Mo2 core in the paddlewheel complexes Mo2(hpp)4-
XY (X ) Y ) Cl; X ) Cl, Y ) BF4; X ) Y ) BF4) by the
2-
π/σ basicity effect3,19 while the HPO4 ion allows the
Supporting Information Available: X-ray crystallographic files
in CIF format for four crystalline forms 1a, 1b, 1c, and 1d and a
CIF file with neutron diffraction data for 1a and a PDF file of the
plot of the neutron Fourier map of one of the PC6H5 groups showing
each of the 5 hydrogen atoms of the phenyl group. This material is
available free of charge via the Internet at http://pubs.acs.org.
formation of the stable [Mo2(HPO4)4(H2O)2]4-,3-,2- ions.21
However, it should be mentioned that there is an important
difference between the present phosphonate ligand and the
phosphate anions, for which reversible waves were reported
at -0.67 and 0.25 V vs SCE.21 The potentials for the latter
were measured in phosphoric acid where protonation must
IC025632W
(20) (a) Novak, I.; Wei, X.; Chin, W. S. J. Phys. Chem. A 2001, 105, 1783.
(b) Schwesinger, R. Chimia 1985, 39, 269.
(21) Chang, I.-J.; Nocera, D. G. J. Am. Chem. Soc. 1987, 109, 4901.
(22) Cotton, F. A.; Datta, T.; Labella, L.; Shang, M. Inorg. Chim. Acta
1993, 203, 55.
4238 Inorganic Chemistry, Vol. 41, No. 16, 2002