Stereoelectronic factors that influence kinetic and thermodynamic sites of protonation of (η6-arene)molybdenum(phosphine)3 complexes

Article abstract of DOI:10.1021/om990281n

We have previously reported (J. Am. Chem. Soc. 1995, 117, 12639) the arene complex (η⁶-C₆H₆)Mo(TRIPOD) (1), where TRIPOD = 1,1,1-tris(diphenylphosphinomethyl)ethane, is protonated upon addition of 1 equiv of D⁺ to yield the metal-hydride [η⁶-C₆H₅D)Mo-(TRIPOD)H]⁺ (1H⁺-d₁) via exo addition of D⁺ to the arene ligand followed by migration to the metal of the endo proton of the putative η⁵-cyclohexadienyl complex [(η⁵-C₆H₆D)Mo-(TRIPOD)]⁺ (1?). The opposite isotopomer is obtained when (η⁶-C₆D₆)Mo(TRIPOD) (1-d₆) is employed to give [(η⁶-C₆D₅H)Mo(TRIPOD)D]⁺ (1D⁺-d₆). Our recent investigation of the electronic structure of 1 and its one-electron oxidized derivative [(η⁶-C₆H₆)Mo(TRIPOD)]⁺ (1⁺) suggested the mechanism of protonation is not driven entirely by electronic factors; steric factors are likely important for indirect protonation of 1. Consistent with that conclusion, we report herein that some other phosphine derivatives are protonated directly at the metal center. Thus, under the same reaction conditions the monodentate phosphine derivatives (η⁶-C₆D₆)Mo(PR₃)₃, where PR₃ = PPh₂Me (3-d₆), PPhMe₂ (4-d₆), and PMe₃ (5-d₆), are protonated directly at the metal 0%, 20%, and 40% of the time, respectively. The remainder of the protonations take place via the aforementioned indirect protonation of the arene ligand. Surprisingly, (η⁶-C₆D₆)Mo(TRIPHOS) (2-d₆), where TRIPHOS = bis(2-diphenylphosphinoethyl)phenylphosphine, reacts with H⁺ by direct protonation of the metal center 80% of the time to give a C₁-symmetric kinetic hydride [(η⁶-C₆D₆)Mo(TRIPHOS)H]⁺ (2_KH⁺-d₆). The other 20% of the time 2-d₆ reacts via the arene mechanism to give [(η⁶-C₆D₅H)Mo(TRIPHOS)D]⁺ (2_KD⁺-d₆). The two enantiomeric forms of 2_KH⁺ are observed to be in rapid equilibrium (k = 6.4 × 10³ s^-1 at -87°C). Compound 2_KH⁺ eventually isomerizes (k = 6.1 × 10^-4 s^-1 at -60°C) to the C_s-symmetric thermodynamic hydride [(η⁶-C₆H₆)Mo-(TRIPHOS)H]⁺ (2_TH⁺). We conclude from the crystal structures of 1-4, tracer studies, and electronic structure calculations of the five derivatives 1-5 presented herein that indirect protonation does not represent a general mechanism for this class of compounds. In addition to electronic factors that favor direct protonation of the metal center from an equatorial trajectory, subtle steric demands of the phosphine ligands play an important role in dictating whether a proton initially attacks the arene ligand or the metal of 1-5. The semiempirical PM3(tm) method may be used to map metal-based frontier orbitals of the complexes 1-5 onto plots of their surface electron densities, thereby revealing sterically accessible metal electron density. Such plots provide straightforward models that predict the mechanisms of protonation.

Full text of DOI:10.1021/om990281n

5004
Organometallics 1999, 18, 5004-5016
Ster eoelectr on ic F a ctor s Th a t In flu en ce Kin etic a n d
Th er m od yn a m ic Sites of P r oton a tion of
(η6-Ar en e)m olybd en u m (p h osp h in e)3 Com p lexes
Michael T. Ashby,* Victor S. Asirvatham, Angela S. Kowalski, and
Masood A. Khan
Department of Chemistry and Biochemistry, The University of Oklahoma,
620 Parrington Oval, Rm. 208, Norman, Oklahoma 73019
Received April 21, 1999
We have previously reported (J . Am. Chem. Soc. 1995, 117, 12639) the arene complex
(η6-C6H6)Mo(TRIPOD) (1), where TRIPOD ) 1,1,1-tris(diphenylphosphinomethyl)ethane, is
protonated upon addition of 1 equiv of D+ to yield the metal-hydride [(η6-C6H5D)Mo-
(TRIPOD)H]+ (1H+-d 1) via exo addition of D+ to the arene ligand followed by migration to
the metal of the endo proton of the putative η5-cyclohexadienyl complex [(η5-C6H6D)Mo-
(TRIPOD)]+ (1‡). The opposite isotopomer is obtained when (η6-C6D6)Mo(TRIPOD) (1-d 6) is
employed to give [(η6-C6D5H)Mo(TRIPOD)D]+ (1D+-d 6). Our recent investigation of the
electronic structure of 1 and its one-electron oxidized derivative [(η6-C6H6)Mo(TRIPOD)]+
(1+) suggested the mechanism of protonation is not driven entirely by electronic factors;
steric factors are likely important for indirect protonation of 1. Consistent with that
conclusion, we report herein that some other phosphine derivatives are protonated directly
at the metal center. Thus, under the same reaction conditions the monodentate phosphine
derivatives (η6-C6D6)Mo(PR3)3, where PR3 ) PPh2Me (3-d 6), PPhMe2 (4-d 6), and PMe3 (5-d 6),
are protonated directly at the metal 0%, 20%, and 40% of the time, respectively. The
remainder of the protonations take place via the aforementioned indirect protonation of the
arene ligand. Surprisingly, (η6-C6D6)Mo(TRIPHOS) (2-d 6), where TRIPHOS ) bis(2-
diphenylphosphinoethyl)phenylphosphine, reacts with H+ by direct protonation of the metal
center 80% of the time to give a C1-symmetric kinetic hydride [(η6-C6D6)Mo(TRIPHOS)H]+
(2KH+-d 6). The other 20% of the time 2-d 6 reacts via the arene mechanism to give [(η6-
C6D5H)Mo(TRIPHOS)D]+ (2KD+-d 6). The two enantiomeric forms of 2KH+ are observed to
be in rapid equilibrium (k ) 6.4 × 103 s-1 at -87 °C). Compound 2KH+ eventually isomerizes
(k ) 6.1 × 10-4 s-1 at -60 °C) to the Cs-symmetric thermodynamic hydride [(η6-C6H6)Mo-
(TRIPHOS)H]+ (2TH+). We conclude from the crystal structures of 1-4, tracer studies, and
electronic structure calculations of the five derivatives 1-5 presented herein that indirect
protonation does not represent a general mechanism for this class of compounds. In addition
to electronic factors that favor direct protonation of the metal center from an equatorial
trajectory, subtle steric demands of the phosphine ligands play an important role in dictating
whether a proton initially attacks the arene ligand or the metal of 1-5. The semiempirical
PM3(tm) method may be used to map metal-based frontier orbitals of the complexes 1-5
onto plots of their surface electron densities, thereby revealing sterically accessible metal
electron density. Such plots provide straightforward models that predict the mechanisms of
protonation.
In tr od u ction
processes that have been devised by Man2 such as the
hydrogenation of unsaturated organic substrates by
metal complexes that activate molecular hydrogen het-
erolytically.3 The issue of protonation of metal centers
versus ligand sites under thermodynamic conditions has
been raised before,4,5 and in a few cases rearrangement
of kinetic and thermodynamic protonation products has
Protonation of a metal, the simplest electrophilic
reaction, is a chemical modification that results in a
change in the formal oxidation state of the metal center
without altering its electron count. One impetus for
studying such reactions is that some biological pathways
likely involve protonation of metal centers of metallo-
proteins.1 For example, some isoforms of metallohydro-
genases have been optimized to reduce protons. Proto-
nation of metal complexes is also invoked in catalytic
(2) Gates, B. C. Catalytic Chemistry; J ohn Wiley: New York, 1991.
(3) J ames, B. R.; Ashby, M. T. In Inorganic Reactions and Methods;
Norman, A., Ed.; VCH: Weinheim, 1993; Vol. 16, pp 71-76.
(4) Kristja´nsdo´ttir, S. S.; Norton, J . R. In Transition Metal Hydrides;
Dedieu, A., Ed.; VCH: Weinheim, 1992; pp 309-359.
(5) Regioselective protonation of unsaturated hydrocarbon ligands
(but not arenes) has been reviewed: Henderson, R. A. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 946-967.
(1) Evans, D. J .; Henderson, R. A.; Smith, B. E. In Catalysis by
Nitrogenases and Synthetic Analogues; Reedijk, J ., Ed.; Marcel Dek-
ker: New York, 1993; pp 89-130.
10.1021/om990281n CCC: $18.00 © 1999 American Chemical Society
Publication on Web 11/02/1999

Protonation of (η6-Arene)molybdenum(phosphine)3
Organometallics, Vol. 18, No. 24, 1999 5005
Ch a r t 1
been studied.6-8 We reported in a communication that
the arene complex (η6-C6H6)Mo(TRIPOD) (1) is proto-
nated via exo addition of H+ to the arene ligand to give
a transient η5-cyclohexadienyl complex [(η5-C6H7)Mo-
(TRIPOD)]+ (1q) that yields the metal-hydride (η6-
C6H6)Mo(TRIPOD)(H)]+ (1H+) by endo hydrogen atom
transfer to the metal (Figure 5).9 Prompted by our recent
study of the electronic structure of 1 that suggested
steric factors might be important,10 we now present a
full account of the earlier mechanistic investigation of
the protonation of 1 as well as related studies for the
derivatives (η6-C6H6)Mo(TRIPHOS) (2), (η6-C6H6)Mo-
(PPh2Me)3 (3), (η6-C6H6)Mo(PPhMe2)3 (4), and (η6-C6H6)-
Mo(PMe3)3 (5) (Chart 1), some of which react via direct
attack on the metal. A balance between steric factors
that are related more to the conformations of the
phosphine ligands than their substitution patterns and
the spatial nature of the metal-based frontier orbitals
is found to dictate the initial sites of protonation in these
complexes.
F igu r e 1. Thermal ellipsoid drawing (50% probability) of
(η6-C6H6)Mo(TRIPOD) (1) from the top and the side with
the labeling scheme.
Exp er im en ta l Section
Ch em ica ls a n d Solven ts. All operations were carried out
using Schlenk or glovebox techniques under argon or nitrogen.
Hydrocarbon solvents were distilled from sodium/benzophe-
none ketal, CH2Cl2 was distilled from CaH2, and methanol was
refluxed over Mg and distilled.11 All solvents were degassed
by three freeze-pump-thaw cycles before use. The phosphine
ligands were used as received from Aldrich and Strem Chemi-
cals. (η6-C6H6)2Mo12 and CDCl3F13 were synthesized according
to published procedures.
were performed by Midwest Microlabs, Indianapolis, IN.
Calculations were performed on a Silicon Graphics IRIS Indigo
2 Solid Impact.
Syn th esis of (η6-C6H6)Mo(TRIP OD) (1) a n d Its d 6 Iso-
top om er fr om a Melt Rea ction . (η6-C6H6)2Mo (0.34 g, 1.35
mmol) and TRIPOD (0.76 g, 1.22 mmol) were heated in a
sealed glass tube under vacuum at 160 °C for 48 h. The tube
was opened under nitrogen, the contents were extracted with
1:1 benzene/heptane (∼30 mL), and the extract was filtered.
The red-orange product crystallized from the solvent upon
cooling to 5 °C (0.62 g, 64%). 1H NMR (C6D6, 500 MHz, 20 °C):
δ 7.08 (m, 12 H, Ph) 6.96 (t, 6 H, J ) 7 Hz, Ph), 6.85 (t, 12 H,
J ) 7 Hz, Ph), 4.41 (m, 6 H, C6H6), 2.17 (m, 6 H, CH2), 1.16
(m, 3 H, CH3). 31P{1H} NMR (C6D6, 202 MHz, 20 °C): δ 46.59
(s). Anal. Calcd (found) for C47H45P3Mo (798.74): C, 70.68
(70.44), H, 5.68 (5.74). The isotopomer (η6-C6D6)Mo(TRIPOD)
(1-d 6) was synthesized using the same procedure with (C6D6)2-
In str u m en ts a n d Refer en ces. 1H, 2H, and 31P NMR
spectra were recorded on a Varian XL-500 spectrometer. The
NMR samples were prepared in tubes that had been glass-
blown onto Schlenk adapters. The solutions were freeze-
pump-thawed before the tubes were flame-sealed under
1
vacuum. H NMR spectra were referenced to residual solvent
peaks CDHCl2 (5.32 ppm), CHCl2F (7.47 ppm, d, J H-F ) 50
2
Hz), and C6D5H (7.24 ppm). H NMR spectra were referenced
to the corresponding 1H NMR chemical shifts. 31P NMR spectra
were referenced to external 85% H3PO4. Combustion analyses
1
Mo in 61% yield. H NMR (C6D6, 500 MHz, 20 °C) the same
as 1 except: δ 4.41 (m, 0.21 H, C6H6).
Gen er a l Syn th esis of (η6-C6H6)Mo(P )3 a n d Th eir d 6
Isotop om er s fr om Solu tion Rea ction s: 2 ((P )3 ) TRIP H-
OS), 3 (P ) P P h 2Me), 4 (P ) P P h Me2), 5 (P ) P Me3). The
procedure used to synthesize 2 is typical. (η6-C6H6)2Mo (0.20
g, 0.78 mmol) and TRIPHOS (0.40 g, 0.75 mmol) were dissolved
in benzene (10 mL). The flask was frozen, evacuated, and
sealed before heating in an oil bath at 160 °C for 24 h. The
flask was backfilled with Ar, methanol (30 mL) was layered
over the red-brown solution, and the flask was placed in a
refrigerator at 5 °C for 12 h to yield 2 (0.32 g, 59%) as red-
orange crystals. The isotopomers 2-d 6, 3-d 6, 4-d 6, and 5-d 6
were synthesized using a similar procedure from (η6-C6D6)2-
(6) Darensbourg, M. Y.; Liaw, W. F.; Riordan, C. G. J . Am. Chem.
Soc. 1989, 111, 8051.
(7) Gladfelter, W. L.; Stevens, R. E. J . Am. Chem. Soc. 1982, 104,
6454.
(8) Stevens, R. E.; Guettler, R. D.; Gladfelter, W. L. Inorg. Chem.
1990, 29, 451.
(9) Kowalski, A. S.; Ashby, M. T. J . Am. Chem. Soc. 1995, 117,
12639.
(10) Asirvatham, V. S.; Gruhn, N. E.; Lichtenberger, D. L.; Ashby,
M. T. Organometallics, submitted for publication.
(11) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 3rd ed.; Pergamon: Oxford, 1988.
(12) Silverthorn, W. E. Inorg. Synth. 1977, 17, 54.
(13) Siegel, J . S.; Anet, F. A. L. J . Org. Chem. 1988, 53, 2629.

5006 Organometallics, Vol. 18, No. 24, 1999
Ashby et al.
F igu r e 3. Thermal ellipsoid drawing (50% probability) of
(η6-C6H6)Mo(PPh2Me)3 (3) from the top and the side with
the labeling scheme.
F igu r e 2. Thermal ellipsoid drawing (50% probability) of
(η6-C6H6)Mo(TRIPHOS) (2) from the top and the side with
the labeling scheme.
for 2KH•+F 3CSO3- in CD2Cl2 were obtained by in situ synthesis
from 2 and F3CSO3H at -90 °C. The known compounds 3H+,
4H+, and 5H+ were characterized spectroscopically. All of the
isotopomers were characterized spectroscopically. The iso-
topomer distribution is given in Table 3. Static NMR spectra
are obtained for 1H+ and 2KH+ at low temperature, but
dynamic spectra are observed at elevated temperatures.
Mo and C6D6 as a solvent. The known compounds 2-5 and
their isotopomers were characterized spectroscopically.
For 2-d 6 (59% yield): 1H NMR (C6D6, 500 MHz, 20 °C) δ
7.95 (m, 2 H, Ph), 7.36 (m, 2 H, Ph), 7.12-7.24 (m, 11 H, Ph),
7.00 (m, 2 H, Ph), 6.84-6.94 (m, 8 H, Ph), 4.09 (m, 0.19 H,
C6D5H) 2.14 (m, 4 H, CH2) 1.68 (m, 2 H, CH2) 0.78 (m, 2 H,
CH2); 31P{1H} NMR (C6D6, 202 MHz, 20 °C) δ 104.8 (t, 1 P,
J P-P ) 17 Hz) 87.9 (d, 2 P).
For 1H•+P F 6 (56%): 1H NMR (CDCl2F, 500 MHz, -125
-
°C) δ 7.28 (m, 4 H, Ph), 7.20 (m, 4 H Ph), 7.12 (m, 6 H, Ph),
7.01 (m, 8 H, Ph), 6.90 (m, 4 H, Ph), 6.63 (m, 4 H, Ph), 5.30 (s,
6 H, C6H6), 2.31 (br s, 2H, CH2), 2.19, 2.65 (br m, 4 H,
diastereotopic CH2), 1.54 (br s, 3 H, CH3), -6.36 (td, 1 H, J PH
) 16 and 62 Hz, Mo-H); 31P{1H} NMR (CDCl2F, 202 MHz,
For 3-d 6 (61% yield): 1H NMR (C6D6, 500 MHz, 20 °C) δ
6.93-7.21 (m, 30 H, Ph), 3.88 (m, 0.23 H, C6D5H), 1.65 (s, 9
H, CH3); 31P{1H} NMR (C6D6, 202 MHz, 20 °C) δ 35.9 (s).
For 4-d 6 (30% yield): 1H NMR (C6D6, 500 MHz, 20 °C) δ
7.43 (m, 6 H, Ph), 7.11 (m, 6 H, Ph), 7.04 (m, 3 H, Ph), 3.86
(m, 0.15 H, C6D5H) 1.29 (s, 18 H, CH3); 31P{1H} NMR (C6D6,
202 MHz, 20 °C) δ 17.8 (s).
1
-125 °C) δ 43.2 (d, 2 P, J PP ) 62 Hz), 34.5 (t, 1 P); H NMR
(CD2Cl2, 500 MHz, 20 °C) δ 7.25 (t, 6 H, J ) 7 Hz, Ph), 7.11 (t,
12 H, J ) 7 Hz, Ph), 6.98 (m, 12 H, Ph), 5.24 (s, 6 H, C6H6),
2.41 (br s, 6 H, CH2), 1.57 (br s, 3 H, CH3), -6.98 (q, 1 H, J PH
) 45 Hz, Mo-H); 31P{1H} NMR (CD2Cl2, 202 MHz, 20 °C) δ
34.52 (s). Anal. Calcd (found) for C47H46P4F6Mo (944.71): C
For 5-d 6 (49% yield): 1H NMR (C6D6, 500 MHz, 20 °C) δ
3.65 (m, 0.27 H, C6D5H) 1.13 (s, 27 H, CH3); 31P{1H} NMR
(C6D6, 202 MHz, 20 °C) δ 3.4 (s).
59.76 (59.48), H 4.91 (4.95).
-
For 1D+-d 6‚F 3CSO3
:
1H NMR (CD2Cl2, 500 MHz, 20 °C)
Gen er a l P r oced u r e for th e Syn th esis of [(η6-C6H6)Mo-
(P )3H]P F 6 a n d Th eir d 6 Isotop om er s: 1H+ ((P )3 ) TRI-
P OD), 2TH+ ((P )3 ) TRIP HOS), 3H+ (P ) P P h 2Me), 4H+
(P ) P P h Me2), 5H+ (P ) P Me3). The procedure used to
δ 7.25 (t, 6 H, J ) 7 Hz, Ph), 7.11 (t, 12 H, J ) 7 Hz, Ph), 6.98
(m, 12 H, Ph), 5.30 (s, 1.17 H, C6D5H/C6D4H2), 2.31 (br s, 6 H,
CH2), 1.54 (br s, 3 H, CH3); 2H NMR (CD2Cl2, 500 MHz, 20
°C) δ 5.30 (s, 4.83 D, C6H6), -6.98 (q, 1 H, J PD ) 6 Hz, Mo-
synthesize 1H•+P F 6 is typical. Compound 1 (0.37 g, 0.46
-
D); 31P{1H} NMR (CD2Cl2, 202 MHz, 20 °C) δ 34.52 (t, J PD
)
mmol) was dissolved in CH2Cl2 (10 mL), and HPF6 (60% in
water, 0.5 mL) was added. After stirring 15 min, the solution
was extracted with water (2 × 10 mL) and the organic layer
was dried over CaCl2. Evaporation of the solvent yielded
1H•+P F 6- as a red solid that was recrystallized from CH2Cl2/
6 Hz).
-
:
1H NMR (CD2Cl2, 500 MHz, 20 °C) δ 7.67
For 2TH•+P F 6
(m, 4 H, Ph), 7.58 (m, 1 H, Ph), 7.55 (m, 4 H, Ph), 7.33-7.43
(m, 8 H, Ph), 7.19-7.29 (m, 8 H, Ph), 4.63 (s, 6 H, C6H6), 2.88
Et2O (0.24 g, 56%). Compound 2TH•+P F 6 has been reported
-
(14) George, T. A.; Hammud, H. H. J . Organomet. Chem., 1995, 503,
C1.
previously,14 but it was not fully characterized. The spectra

Protonation of (η6-Arene)molybdenum(phosphine)3
Organometallics, Vol. 18, No. 24, 1999 5007
F igu r e 5. Kinetic products possible upon addition of D+
to (η6-C6H6)Mo(TRIPOD) (1) via (i) direct attack on the
metal (1 f 1D+-d 1) and (ii) indirect attack via a η5-
cyclohexadienyl ion (1 f 1q-d 1 f 1H+-d 1). The equilibria
1D+-d 1 h 1q-d 1′ h 1q-d 1′′ h 1H+-d 1 could explain the
incorporation of deuterium into the arene ligand even if
the initial protonation is at the metal (1 f 1D+-d 1), but
selective spin inversion transfer (SSIT) NMR studies rule
out such equilibria (see text). Only one arene hydrogen
atom is shown for the sake of clarity.
F igu r e 4. Thermal ellipsoid drawing (50% probability) of
(η6-C6H6)Mo(PPhMe2)3 (4) from the top and the side with
the labeling scheme.
(m, 2 H, CH2), 2.66 (m, 2 H, CH2), 2.12 (m, 2 H, CH2), 2.04 (m,
2 H, CH2), -6.06 (td, 1 H, J PH ) 4 and 49 Hz, Mo-H); 31P{1H}
NMR (CD2Cl2, 202 MHz, 20 °C) δ 110.08 (t, 1 P, J P-P) 32 Hz),
89.9 (d, 2 P), -144.0 (septet, 1 P, J P-F ) 709 Hz).
-
-
For 2TH+-d 6‚F 3CSO3 and 2TD+-d 6‚F 3CSO3
:
1H NMR
(CD2Cl2, 500 MHz, 20 °C) δ 7.67 (m, 4 H, Ph), 7.58 (m, 1 H,
Ph), 7.55 (m, 4 H, Ph), 7.33-7.43 (m, 8 H, Ph), 7.19-7.29 (m,
8 H, Ph), 4.63 (s, 0.31 H, C6D5H/C6D4H2), 2.88 (m, 2 H, CH2),
2.66 (m, 2 H, CH2), 2.12 (m, 2 H, CH2), 2.04 (m, 2 H, CH2),
-6.06 (td, 0.88 H, J PH ) 4 and 49 Hz, Mo-H); 2H NMR (CD2-
Cl2, 76.7 MHz, 20 °C) δ 4.63 (br m, 5.54 D, C6D5H/C6D4H2),
-6.06 (br m, 0.27 D, Mo-D); 31P{1H} NMR (CD2Cl2, 202 MHz,
20 °C) δ 110.3 (t, 0.8 P, J PP ) 32 Hz, (-CH2)2(Ph)P-H), 110.2
(br t, 0.2 P, J PP ) 32 Hz, (-CH2)2(Ph)P-D), 90.2 (d, 1.6 P,
4.36 (q, 0.91 H, J PH ) 4 Hz, C6D5H/C6D4H2), 1.73 (br s, 18 H,
2
CH3), -1.26 (q, 0.24 H, J PH ) 75 Hz, Mo-H); H NMR (CD2-
Cl2, 76.7 MHz, 20 °C) δ 4.36 (br s, 5.03 H, C6D6/C6D5H), -1.26
(q, 0.82 D, J PD ) 11 Hz, Mo-D); 31P{1H} NMR (CD2Cl2, 202
MHz, 20 °C) δ 11.6 (t, J PD ) 11 Hz, P-Mo-D), 11.5 (br s,
P-Mo-H).
For 5H+-d 6‚F 3CSO3- and 5D+-d 6‚F 3CSO3
-
:
1H NMR (CD2-
Cl2, 500 MHz, 20 °C) δ 4.63 (m, 0.87 H, C6D5H/C6D4H2) 2.32
(s, 27 H, CH3), -3.53 (q, 0.40 H, J PH ) 66 Hz, Mo-H); 2H NMR
(CD2Cl2, 76.7 MHz, 20 °C) δ 4.63 (m, 5.17 D, C6D6/C6D5H)
-3.53 (q, 0.56 D, J PD ) 10 Hz, Mo-D); 31P{1H} NMR (CD2Cl2,
202 MHz, 20 °C) δ 0.8 (br s).
(-CH2)(Ph)2P-H), 90.1 (m, 0.4 P, (-CH2)(Ph)2P-D).
For 2KH+‚F 3CSO3
-
:
1H NMR (CD2Cl2, 500 MHz, -90 °C) δ
8.02 (m, 2 H, Ph), 7.66 (m, 4 H, Ph), 7.34 (m, 2 H, Ph), 7.23
(m, 2 H, Ph), 7.16 (m, 4 H, Ph), 6.99 (br m, 11 H, Ph), 4.90 (s,
6 H, arene), 2.83 (m, 2 H, CH2), 2.48 (m, 2 H, CH2), 1.76 (m,
2 H, CH2), 1.55 (m, 2 H, CH2), -7.09 (dt, J PH ) 24 and 72 Hz,
Mo-H); 31P{1H} NMR (CDCl2F, 202 MHz, -125 °C) δ 99.9 (br
Kin etic P r od u ct Obta in ed fr om 1 a n d F3CSO3D a t Low
Tem p er a tu r e. A sample of 1 was added to an NMR tube
equipped with a Schlenk adapter, and the tube was evacuated.
The Freon CDCl2F was vacuum transferred into the tube, the
solution was frozen with liquid N2, and the tube was backfilled
with N2 gas. While still frozen, 1 equiv of F3CSO3D was added,
the tube was evacuated, and it was flame-sealed. The frozen
sample was transferred to an NMR probe that had been
precooled to -150 °C. The probe was allowed to warm slowly.
At -120 °C, the solution melted and a spectrum of 1 was
obtained. At approximately -85 °C,15 1H+-d 1 was obtained (no
1D+-d 1). No further change in the spectrum was observed as
the sample was warmed to 25 °C.
s, 1 P, -CH2P(Ph)CH2-), 76.9, 69.8 (br s, 2 P, -CH2PPh2).
-
For 3D+-d 6‚F 3CSO3
:
1H NMR (CD2Cl2, 500 MHz, 20 °C)
δ 7.42 (t, 6 H, J HH ) 6 Hz, Ph), 7.33 (t, 12 H, J HH ) 8 Hz, Ph),
7.05 (m, 12 H, Ph), 4.66 (q, 1.23 H, J PH ) 4 Hz, C6D5H/C6D4H2),
2
1.85 (br s, 9 H, CH3); H NMR (CD2Cl2, 76.7 MHz, 20 °C) δ
4.74 (br s, 4.77 D, C6D5H/C6D4H2), 0.06 (q, 1.00 D, J PD ) 11
Hz, Mo-D); 31P{1H} NMR (CD2Cl2, 202 MHz, 20 °C) δ 26.5 (t,
J PD ) 11 Hz).
For 4H+-d 6‚F 3CSO3- and 4D+-d 6‚F 3CSO3
:
1H NMR (CD2-
-
Cl2, 500 MHz, 20 °C) δ 7.43 (m, 9 H, Ph), 7.31 (m, 6 H, Ph),

5008 Organometallics, Vol. 18, No. 24, 1999
Ashby et al.
Selective Sp in In ver sion Tr a n sfer (SSIT) Exp er im en t.
Ch a r t 2
A sample of 1H•+P F 6 was added to an NMR tube equipped
-
with a Schlenk adapter, and the tube was evacuated. The
Freon CDCl2F was vacuum transferred into the tube, and it
was flame-sealed. Because of the large difference in frequency
between the arene and hydride resonances, the standard spin
inversion transfer (SIT) pulse sequence16-19 was impractical.
An alternative pulse sequence was devised. Selective spin
inversion transfer (SSIT) data were collected using the two-
pulse sequence d1, 2π/2, τm, π/2, aq. The transmitter offset
frequency was centered on the arene resonance. The relaxation
delay d1 was set equal to 5 times the longest T1 to allow
complete longitudinal relaxation between pulses. The 180°
pulse was of sufficiently low power (and long duration) that
surrounding resonances were unaffected. The mixing time τm
was varied between 0.001 s and 5T1. A 90° pulse was applied
at full power prior to FID acquisition. Data analysis for the
SSIT pulse sequence is the same as the SIT pulse sequence.20
A SSIT experiment at 25 °C indicated that no spin exchange
takes place on the time scale of nuclear relaxation (T1
1 s).
≈
Mea su r em en t of th e Isom er iza tion Ra te of 1H+. The
31P NMR spectrum of 1H+ exhibits Cs symmetry at -125 °C
with a 2:1 ratio of phosphorus resonances at 43.2 and 34.5
ppm, respectively (vide supra). As the sample is heated, the
two 31P signals broaden and merge. At the coalescence tem-
perature of -70 °C and ∆δ of 202.36 Hz, the rate constant
was determined using the equation for a two-site exchange
with a 2:1 population:21
Ch a r t 3
2π∆δ
k203K
)
≈ 3.01∆δ ) 5.3 × 103 s-1
(1)
(2)
ez + e-z
1
x
z ) ln{3 + 2 2}
6
The corresponding activation barrier was estimated at the
coalescence temperature:22,23
∆Gq ) -RTc[ln(k/Tc) + ln(h/K)] ) 35 kJ mol-1
(3)
Mea su r em en t of th e Ep im er iza tion Ra te of 2KH+. The
kinetic hydride 2KH+ was synthesized in situ at -90 °C in CD2-
Cl2 from 2 and F3CSO3H using the aforementioned process.
The 31P resonances that correspond to the terminal P donors
of the TRIPHOS ligand of 2KH+ are nearly static at -97 °C
(69.8 and 76.9 ppm). Under fast exchange conditions, the
nuclei of the terminal P donors are rendered magnetically
equivalent and the spectrum becomes consistent with overall
Cs symmetry. At the coalescence temperature of -87 °C and
∆δ of 202.36 Hz, the rate constant was determined using the
standard equation:24
where R ) 8.314 J mol-1 K-1 is the gas constant, Tc is the
coalescence temperature, h ) 6.626 × 10-34 J s is Planck’s
constant, and K ) 1.381 × 10-23 J K-1 is Boltzmann’s con-
stant.
Gen er a l P r oced u r e for Low -Tem p er a tu r e NMR Stu d -
ies of 2KH+. A crystalline sample of 2 was placed as a solid in
a N2-purged NMR tube equipped with a Schlenk adapter.
Solvent was added, and the resulting solution was frozen with
liquid N2. The acid solution was syringed (using a gastight
Hamilton syringe) onto the frozen matrix under a positive N2
gas flow, and then the valve was closed, vacuum was applied
to the heterogeneous frozen mixture, and the tube was flame-
sealed. The tube was transferred quickly from the liquid N2
bath to a n-pentane/liquid N2 slush bath at -100 °C for
transport to the spectrometer. The probe of the spectrometer
was precooled to -90 °C before insertion of the NMR tube.
3
-1
x
k186K ) 2π(∆δ)/ 2 ) 6.4 × 10 s
(4)
The corresponding activation barrier was estimated by eq 3
to be 31 kJ mol-1 at the coalescence temperature.
Mea su r em en t of th e Isom er iza tion Ra te of 2KH+ to
2TH+. A solution of 2 (40 mM) and F3CSO3H (40 mM) was
prepared at -90 °C using the aforementioned procedure. After
the NMR sample was warmed to -60 °C in the spectrometer,
three species are observed: the parent compound 2, the kinetic
monohydride 2KH+, the thermodynamic monohydride 2TH+,
and small amounts of a kinetic dihydride of Cs symmetry,
(15) It has been noted that CF3SO3H freezes very rapidly in CH2-
Cl2 at -95 °C and is not dissolved: Rothfuss, H.; Gusev, D. G.; Caulton,
K. G. Inorg. Chem. 1995, 34, 2894. The protonation of 1 is probably
very fast at low temperature, and warming to -85 °C is necessary to
dissolve the acid.
(16) Bellon, S. F.; Chen, D.; J ohnston, E. R. J . Magn. Reson. 1987,
73, 168.
(17) Alger, J . R.; Prestegard, J . H. J . Magn. Reson. 1977, 27, 137.
(18) Kuchel, R. W.; Chapman, B. E. J . Theor. Biol. 1983, 105, 569.
(19) Robinson, G.; Kuchel, P. W.; Chapman, B. E. J . Magn. Reson.
1985, 63, 314.
(20) Ashby, M. T.; Govindan, G. N.; Grafton, A. K. J . Am. Chem.
Soc. 1994, 116, 4801.
2+
2+
2KH2 (Charts 1-3). A thermodynamic dihydride 2TH2
(Chart 3) that is observed at higher acid concentrations is not
observed under these conditions. The distributions of these
species as a function of time are illustrated in Figure 6. Given
the fact that we have reacted an equimolar quantity of 2 and
H+, assuming the equilibrium between 2KH+ and 2KH2 is
2+
(21) Shanan-Atidi, H.; Bar-Eli, K. H. J . Phys. Chem. 1970, 74, 961.
(22) Allerhand, A.; Gutowsky, H. S.; J onas, J .; Meinzer, R. A. J . Am.
Chem. Soc. 1966, 88, 5.
(24) Green, M. L. H.; Wong, L.-L.; Sella, A. Organometallics 1992,
11, 2660.
(23) Faller, J . W. Adv. Organomet. Chem. 1977, 16, 211-239.

Protonation of (η6-Arene)molybdenum(phosphine)3
Organometallics, Vol. 18, No. 24, 1999 5009
F igu r e 6. Distribution of 2 (O), 2KH+ (×), 2TH+ (]), and
2+
2KH2 (0) as a function of time when 1 equiv of H+ is
added to a 40 mmol solution of 2 in CD2Cl2 at -60 °C (left).
The plot of ln[2KH+]/time (+) that was used to calculate
the effective rate constant of kobs ) 5.0 × 10-4 s-1 for
interconversion of 2KH+ to 2TH+ at -60 °C (right). All of
the data for the plot on the right are illustrated, but only
every fifth point is illustrated for the plots on the left.
rapid (as is suggested by the time dependence of the concen-
tration curves), and the rate of formation of the product 2TH+
depends only on the concentration of 2KH+:
2 + 2KH22+ yk1z 2 2KH+
k-1
k2
2KH+ 98 2TH+
The rate is given by25
d[2TH+]
dt
k2
K
x
1
[2KH+]
(5)
F igu r e 7. Proposed reaction pathways for the intercon-
version of 2, 2KH+, and 2TH+. The rate constant for
epimerization of 2KH+ was measured at the coalescence
temperature of -87 °C to be k ) 6.4 × 103 s-1 (∆Gq = 31
kJ mol-1). The rate constant for isomerization of 2KH+ to
2TH+ was determined to be k ) 6.1 × 10-4 s-1 (∆Gq = 65
kJ mol-1) at -60 °C.
)
2 +
K
x
1
where the equilibrium constant K1 may be determined by
integration of the components at t ) 0:
[2KH+]2
K1 )
) 77
(6)
2+
[2][2KH2
]
Table 2, the computed geometries are compared with the
metric data for 1-4 that were determined by X-ray crystal-
lography. The computed geometry of 5 is also given in Table
2. The CPK models of Figure 8 may be compared with the
thermal ellipsoid figures for 1-4 in Figures 1-4. Figure 8 also
illustrates surface electron density plots of the (η6-C6H6)Mo-
(P)3 derivatives 1-5 onto which is mapped the surface electron
density of the three highest occupied molecular orbitals (the
metal t2g-type orbitals) as calculated using the PM3(tm)
method. The isodensity surfaces (0.002 e/au3) have been chosen
to reflect van der Waals radii. The colors range from red (no
frontier orbital character) to blue (5 × 10-5 e/au3 frontier
orbital character). For some derivatives, especially 4, several
stationary-state comformers of similar energies were encoun-
tered. The ones illustrated in Figure 8 most closely resemble
the conformations that were observed in the solid state. All of
the computed stationary-state conformers gave mapped sur-
face electron density plots that were comparable (with respect
to penetrating metal density) to those that are illustrated in
Figure 8. Electron structure calculations that were performed
on 1-4 using the coordinates that were determined by X-ray
crystallography also gave mapped surface electron density
plots that were comparable to those of Figure 8.
From the data of Figure 6, kobs ) 5.0 × 10-4 s-1, yielding k2 )
6.1 × 10-4 s-1, which gives an activation barrier of ∆Gq ) 65
kJ mol-1 at -60 °C.
Tr a cer Stu d ies In volvin g 1-5. All of the tracer experi-
ments that are summarized in Table 3 were carried out at
room temperature (22 °C). Each experiment was carried out
twice, once in CD2Cl2 (to obtain 1H NMR spectra) and once in
CH2Cl2 (to obtain 2H NMR spectra). Solutions of the d6
derivatives of 1-5 (ca. 100 mM) were transferred to NMR
tubes in a glovebox. Approximately 1 equiv of F3CSO3H was
added via a microliter syringe to the NMR tube, the tube was
flame-sealed, and the NMR spectra were collected. NMR
spectra collected the following day revealed no significant
changes.
Electr on ic Str u ctu r e a n d Geom etr y Ca lcu la tion s. The
molecular mechanics and molecular orbital calculations on 1-5
were performed with SPARTAN 5.0.26 The geometries of 1-5
were optimized using the Merck molecular force field (MMFF),
then refined using the semiempirical method PM3(tm). In
(25) Schmid, R.; Sapunov, V. N. Non-formal Kinetics; Verlag Che-
mie: Weinheim, 1982.
(26) Spartan, version 5.0; Wavefunction, Inc.: 18401 Von Karman
Avenue, Suite 370, Irvine, CA 92612.
X-r a y Cr ysta llogr a p h y. The conditions that were used to
collect X-ray data for 1-4 are summarized in Table 1. The

5010 Organometallics, Vol. 18, No. 24, 1999
Ta ble 1. Cr ysta l Da ta for Com p ou n d s 1-4
Ashby et al.
1
2
3
4
formula
fw
C
47H45MoP3
C40H39MoP3
708.56
C45H45MoP3
774.66
C30H39MoP3
588.46
798.68
space group
cryst syst
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
T, K
P21/n
P21/n
R3
P21/n
monoclinic
9.636(4)
20.277(5)
19.670(3)
90
93.05(2)
90
3838(2)
203(2)
monoclinic
9.1830(10)
17.385(2)
21.413(3)
90
101.830(10)
90
3345.9(7)
188(2)
rhombohedral
10.7703(5)
monoclinic
9.4393(8)
9.4066(9)
32.346(4)
90
95.518(7)
90
2858.7(5)
173(2)
110.241(3)
933.33(8)
173(2)
Z
D
4
4
1
4
calc, g cm-3
1.382
0.500
1.407
0.564
1.378
0.512
1.367
0.644
µ, mm-1
final R indicesa
[I > 2σ(I)]
R1 ) 0.0379
wR2 ) 0.0910
R1 ) 0.0594
wR2 ) 0.1147
1.095
R1 ) 0.0451
wR2 ) 0.1095
R1 ) 0.0614
wR2 ) 0.1206
1.045
R1 ) 0.0224
wR2 ) 0.0587
R1 ) 0.0225
wR2 ) 0.0595
1.045
R1 ) 0.0297
wR2 )0.0730
R1 ) 0.0369
wR2 ) 0.0788
1.054
R indices [all data
used in refinement]
GOF
R1 ) ∑|Fo| - |Fc||/∑|Fo|; wR2 ) [∑[w(Fo - Fc2)]/∑[w(Fo2)2]]1/2; w ) 1/∑[σ2(Fo)2 + (aP)2 + (bP)2 + (cP)2], P ) (Fo)2[1/3 + [2*(Fo)2/3]] for
a
2
Fo g 0 (otherwise zero). GOF ) [∑[w(Fo - Fc2)]/(n - m)]1/2, where n ) no. of reflections observed and m ) no. of parameters.
2
2
Ta ble 2. Selected In ter a tom ic Dista n ces (Å) a n d An gles (d eg) Deter m in ed by X-r a y Diffr a ction
(Com p ou n d s 1-4) a n d Com p u ta tion a lly Usin g th e P M3(tm ) Meth od (Com p ou n d s 1-5)
(η6-C6H6)Mo(TRIPOD) (η6-C6H6)Mo(TRIPHOS) (η6-C6H6)Mo(PPh2Me)3 (η6-C6H6)Mo(PPhMe2)3 (η6-C6H6)Mo(PMe3)3
1
2
3
4
5
XRD
PM3(tm)
XRD
PM3(tm)
XRD
PM3(tm)
XRD
PM3(tm)
PM3(tm)
C1
C3
C1
C1
C3
C3
C1
C3
C3
Mo-P1
Mo-P2
Mo-P3
Mo-Bzcent
2.412(1)
2.389
2.3687(8)
2.4045(8)
2.3926(8)
1.789(4)
2.400
2.383
2.369
1.886
2.4438(5)
2.407
2.4431(7)
2.4431(6)
2.4428(7)
1.781(3)
2.389
2.386
2.3866(9)
2.3843(9)
1.816(3)
a
1.883
85.76
1.788(2)
92.37(2)
1.903
93.44
1.886
93.97
1.878
93.30
P1-Mo-P2
P1-Mo-P3
P2-Mo-P3
Bzcent-Mo-P1 129.2(1)
Bzcent-Mo-P2 128.4(1)
Bzcent-Mo-P3 131.0(1)
84.81(3)
83.50(3)
83.13(3)
78.91(3)
79.80(3)
92.74(3)
131.8(1)
132.3(1)
124.0(1)
79.11
79.16
99.05
133.66
128.01
122.75
92.32(2)
92.25(2)
90.49(2)
123.6(1)
123.6(1)
124.9(1)
128.21
123.6(1)
122.79
122.43
122.90
a
Bzcent is the centroid of a η6-arene ligand.
X-ray data for 1 were collected with an Enraf-Nonius CAD-4
diffractometer, and the data for 2-4 were collected on a
Siemens P4 diffractometer. Monochromated Mo KR radiation
(λ ) 0.710 73 Å) was employed. The data were corrected for
Lorentz and polarization effects. An empirical absorption
correction based on ψ scans was applied to 2-4. All of the
structures were determined and refined using the SHELXTL
(Siemens) system. The structures of 1 and 2 were solved by
direct methods. The structures of 3 and 4 were solved by the
heavy atom method. All structures were refined by full-matrix
least-squares on F2 using all reflections. All of the hydrogen
atoms were included in the refinements with idealized param-
eters. For all the hydrogen atoms except the methyl atoms,
the temperature factors were assigned to be 1.2 times the U(eq)
value of the carbon atom on which they ride; for the methyl
hydrogens the values were 1.5 times that of the U(eq) values
of the corresponding carbon atoms. For 2, two conformations
of the TRIPHOS ligand are observed, which were modeled by
large thermal parameters for C(37) and C(38) and two occupa-
tion sites (50:50) for the arene rings C7-C12 (C7A-C12A) and
C19-C24 (C19A-C24A). Only one conformation is depicted
in Figure 2. A figure that depicts both conformations is
available in the Supporting Information. For 3, best refinement
was achieved in the noncentric space group R3 as a racemic
twin with a final BASF parameter of 0.26(2), suggesting a 26:
74 mixture. Other details concerning the structure refinements
are summarized in Table 1. Selected metric data for 1-4 are
compared in Table 2. Tables of atomic coordinates and
equivalent isotropic displacement parameters, anisotropic
displacement parameters, hydrogen coordinates and isotropic
parameters, and bond lengths and angles are available as
Supporting Information.
Resu lts
Syn th esis a n d Solu tion P r op er ties of 1-5. Com-
plex 1 (Chart 1) is most effectively synthesized in a melt
reaction containing (η6-C6H6)2Mo and TRIPOD. How-
ever, the derivatives 2-5 are more readily synthesized
if a solvent is employed. Compounds 1-5 and their
isotopomers were characterized by NMR and mass
spectrometry and, in the case of new compounds, by
combustion analysis. 1H, 2H (when appropriate), and 31P
NMR were measured. It is noteworthy that signals for
the 1H (and 2H for isotopomers) nuclei of the arene
ligands and the 31P nuclei were very broad (and some-
times not observed) for 1 and 3-5 in dichloromethane
solvent, though the 1H resonances of the rest of the
phosphine ligands were observed. The observed reso-
nances exhibited expected chemical shifts. In contrast,
benzene solutions of the same samples exhibited all of
the expected resonances. We believe this peculiar sol-
vent effect is due to trace paramagnetic impurities that

Protonation of (η6-Arene)molybdenum(phosphine)3
Organometallics, Vol. 18, No. 24, 1999 5011
are soluble in dichloromethane but not soluble in
benzene. We have previously characterized the para-
magnetic one-electron-oxidation product of 1,10 and we
suspect the likely impurity is [(η6-C6H6)Mo(P)3]+. Self-
exchange between (η6-C6H6)Mo(P)3 and [(η6-C6H6)Mo-
(P)3]+ could give rise to broadened resonances, especially
in the first coordination sphere of the metal center.27
We have noted on occasion that small amounts of a
yellow precipitate (the color of 1+) forms when benzene
solutions of 1 are allowed to stand, even under an inert
atmosphere. Since the solutions were prepared from
crystalline samples with particular attention to exclu-
sion of moisture and oxygen, the paramagnetic impurity
may be the result of partial oxidation by the solvent or
some type of disproportionation. A similar problem was
observed in other polar solvents such as THF. Interest-
ingly, we encountered no problem obtaining complete
NMR spectra of 2 in dichloromethane. To minimize
formation of the impurity, dichloromethane solutions of
1-5 were prepared fresh and used immediately for the
studies that are described herein. It is noteworthy that
the corresponding protonated species 1H+-5H+ do not
exhibit stability problems in dichloromethane, and all
of their NMR resonances were observed. This is consis-
tent with the expected greater stabilities of the proto-
nated species with respect to oxidation.
Molecu la r Str u ctu r es of 1-4. Since rationalization
of the regiochemistry of protonation will depend on an
understanding of the molecular structures of 1-5 (vide
infra), crystal structures of 1-4 were determined by
X-ray crystallography. Efforts to obtain diffraction qual-
ity crystals of 5 failed. The crystal structure parameters
for 1-4 are summarized in Table 1. The molecular
structures of 1-4 are compared in Table 2. Thermal
ellipsoid drawings of 1-4 and the labeling schemes are
presented in Figures 1-4. Other crystallographic infor-
mation is available in the Supporting Information. Com-
pound 3 exhibits crystallographically imposed C3 sym-
metry, whereas 1 and 4 are approximately C3 symmetric.
The molecular structure of 2 has no crystallographic
symmetry, although small conformational changes re-
veal its Cs molecular symmetry. The unremarkable
metric data of Table 2 compare to that previously
reported for other (η6-arene)Mo(P)3 complexes such as
(η6-C6H5(PPhMe))Mo(PPh2Me)3,28 [(µ2-η6-C6H5(PMe2))-
Mo(PPhMe2)2]2,29 and (η6-C6H5(PPhMe))Mo(PPh2Me)-
(Ph2PCH(Me)CH2PPh2).30 The molecular structures of
1-5 that were computed using the PM3(tm) method are
compared in Table 2 with those that were determined
by X-ray crystallography for 1-4. Figure 8 (left) depicts
the computed structures of 1-5 as space-filled models.
Of particular interest with respect to the present study
are the conformations observed for 1-5 (vide infra). The
computed conformations (Figure 8, left) resemble the
experimental conformations (Figures 1-4).
F igu r e 8. CPK models (left) and surface electron density
plots (right) of (η6-C6H6)Mo(P)3: 1 ((P)3 ) TRIPOD), 2 ((P)3
) TRIPHOS), 3 (P ) PPh2Me), 4 (P ) PPhMe2), and 5 (P
) PMe3). The isodensity surfaces (0.002 e/au3) of the
electron density plots have been chosen to reflect the van
der Waals radii of the CPK models. The surface electron
density of the three highest occupied molecular orbitals (the
metal t2g-type orbitals) as calculated using the PM3(tm)
method have been mapped onto the isodensity surfaces.
The colors range from red (no frontier orbital character)
to blue (5 × 10-5 e/au3 frontier orbital character). The
geometries are optimized by the PM3(tm) method (although
similar figures are obtained using the crystal structures
of 1-4). Note that two perspectives of 2 are presented. The
CPK view from the “side” suggests the metal center is
sterically accessible, and the electron density plot shows
metal density is available at the surface. The view from
the “front” demonstrates available surface metal density,
but high steric congestion.
Electr on ic Str u ctu r es of 1-5. We have recently
reported a detailed combined theoretical and experi-
(27) Bertini, I.; Luchinat, C. In Physical Methods for Chemists;
Drago, R. S., Ed.; Saunders College Publishing: Ft. Worth, TX, 1992;
pp 500-552.
(28) Mason, R.; Thomas, K. M.; Heath, G. A. J . Organomet. Chem.
1975, 90, 195.
(29) Cotton, F. A.; Luck, R. L.; Morris, R. H. Organometallics 1989,
8, 1282.
(30) Morris, R. H.; Sawyer, J . F.; Sella, A. Acta Crystallogr., Sect. C
1988, 44, 23.

5012 Organometallics, Vol. 18, No. 24, 1999
Ashby et al.
mental investigation of the electronic structures of 1 and
1+. We concluded that the three highest occupied
that the initial attack is exo. Referring to Figure 5, exo
addition would correspond to 1 f 1q-d 1 f 1H+-d 1. If
the hydride ligand exchanges rapidly with the hydro-
gens of the arene ligand (1D+-d 1 h 1q-d 1′ h 1q-d 1′′ h
1H+-d 1) and one ignores isotope effects, statistics dictate
a 1:6 ratio of 1D+-d 1:1H+-d 1. Thus, the initial intermo-
lecular attack could be at the metal (1 f 1D+-d 1),
followed by intramolecular endo protonation of the arene
(1D+-d 1 f 1q-d 1′), a hydrogen shift (1q-d 1′ f 1q-d 1′′),
and finally transfer of the endo proton to the metal
(1q-d 1′′ f 1H+-d 1). Although it is problematic to dif-
ferentiate between these two pathways with the tracer
experiments, we have qualitative kinetic information
that rules out the 1 f 1D+-d 1 f 1q-d 1′ f 1q-d 1′′ f
1H+-d 1 pathway. We have prepared equimolar solutions
of 1 and F3CSO3D at -150 °C in CDCl2F and found
protonation does not occur until the solution is warmed
to ca. -85 °C.15 The first product that is observed in
these experiments is 1H+-d 1. Selective spin inversion
transfer (SSIT) NMR experiments at 25 °C indicate that
no spin exchange takes place between the arene ligand
hydrogens and the metal hydride on the time scale of
nuclear relaxation (T1 ≈ 1 s). Extrapolating to -85 °C,
the rate of formation of 1H+-d 1 must be much faster
than the rate of intramolecular chemical exchange
(1D+-d 1 h 1q-d 1′ h 1q-d 1′′ h 1H+-d 1).
Com p a r a tive Regioch em istr y of P r oton a tion of
1-5. When 1-5 were reacted with D+, we obtained
evidence in some cases that the isotope tracer was
exchanging with adventitious sources of H+ (probably
in the form of moisture) before reacting with the metal
complexes. In such cases mass balance was poor and
not all of the deuterium tracer was accounted for in the
metal-hydride complexes. To avoid this problem, all of
the tracer studies we report herein involve reaction of
the η6-C6D6 derivatives of 1-5 with H+. These experi-
ments were performed at least twice, once in CD2Cl2 so
1H NMR could be run, and once in CH2Cl2, so 2H NMR
could be run. In some cases (especially 1 and 2) the
experiments were repeated dozens of times using the
same or slightly different conditions (i.e., different
solvents, concentrations, and temperatures) to ensure
reproducibility and generality of the conclusions herein.
In the case of 1, direct protonation of the metal center
was never indicated. For 2, attack at the arene ligand
and the metal was always observed with partitioning
of ca. 4:1.
frontier orbitals are largely metal in character (the t2g
-
type orbitals) and very close in energy. In fact, various
computational methods (semiempirical, ab initio HF and
DFT) predict different ordering of these orbitals de-
pending on the model and the basis set employed.
However, all of the computational methods agree the
three orbitals are close in energy and that two of the
three frontier orbitals are largely metal in character,
but two bear significant η6-arene character. Figure 8
(right) illustrates the three highest occupied frontier
orbitals of 1-5 (HOMO, HOMO-1, and HOMO-2)
mapped onto plots of total electron density using
computed geometries. Thus, Figure 8 (right) illustrates
the sterically accessible metal electron density. Impor-
tantly, very similar plots were obtained when the crystal
structures of 1-4 were employed in the calculation,
albeit with different isodensity surface parameters to
account for the shorter metal-ligand distances that are
observed experimentally as compared to the computed
distances (Table 2). Also, very similar plots for 1 (i.e.,
no surface-accessible metal density) were observed for
PM3(tm), HF6-31G*, and DFT-BP* calculations.10 Since
the qualitative nature of Figure 8 (left) does not appear
to be sensitive to interatomic distances, the conforma-
tions that are predicted using PM3(tm), HF, MP2, and
DFT are comparable, and we only intend to discuss the
qualitative features of Figure 8; we only discuss the
results of the PM3(tm) calculations here.
Tr acer Stu dies In volvin g P r oton ation of 1. Treat-
ment of 1 with 1 equiv of F3CSO3D in dichloromethane
gave 1H+-d 1 (Chart 2). The PF6 salt of 1H+-d 1 was
-
characterized by 1H, 2H, and 31P NMR spectroscopy and,
after deprotonation with NaOMe, by DIP and FAB MS.
The NMR data reveal the presence of a Mo-H (not a
Mo-D) bond. The parent peaks of the deprotonated
product 1-d 1 were one m/z unit higher than those of the
undeuterated compound. The latter experiment sug-
gests that the arene ligand is involved in the protonation
of the metal center, but it does not demonstrate that
the hydride ligand originates from the arene, for it could
originate from another source of protium, e.g., from the
solvent. Also, as discussed above, a curious feature of
the various (η6-arene)Mo(P)3 derivatives is that the
arene resonances are extremely broad in dichloromethane
(the solvent employed for the protonation reactions),
making accurate integration difficult. Addressing the
concern that the D+ tracer may undergo exchange with
adventitious sources of H+ (e.g., moisture) before react-
ing with the metal complex, a complementary experi-
ment was performed. The perdeuterioarene derivative
1-d 6, which was synthesized from (η6-C6D6)2Mo, was
treated with 1 equiv of F3CSO3H in dichloromethane
to yield [(η6-C6D5H)Mo(TRIPOD)D]+, 1H+-d 6. The PF6
salt of 1H+-d 6 was characterized by 1H, 2H, and 31P
NMR spectroscopy and, after deprotonation with NaOMe
to give 1-d 5, by DIP and FAB MS. The protonation
experiment employing 1-d 6 demonstrates that the hy-
dride ligand originates from the arene ligand.
Discu ssion
E lect r op h ilic Ar om a t ic Su b st it u t ion of Ar en e
Liga n d s. Complexation of arenes has a marked effect
on their reaction properties.31,32 Since the metal is net
electron withdrawing, binding arenes to metals has the
effect of activating them toward nucleophilic attack
while deactivating them with respect to electrophilic
attack. We are aware of only three types of electrophilic
aromatic substitution reactions that occur with metal-
bound arenes: (1) protonation, (2) Friedel-Crafts
acetylation,33-38 and (3) addition of iminium salts.39
Sp in -La belin g NMR Exp er im en ts To Ru le Ou t
Alter n a tive Mech a n ism s. While the origin of the
hydride ligand is demonstrated by the tracer studies,
these experiments do not unambiguously demonstrate
(31) Silverthorn, W. E. Adv. Organomet. Chem. 1975, 13, 47-137.
(32) Hegedus, L. S. Transition Metals in the Synthesis of Complex
Organic Molecules; University Science Books: Mill Valley, CA, 1994.
(33) Riemschneider, R.; Becker, O.; Franz, K. Monatsch. Chem. 1959,
90, 571.

Protonation of (η6-Arene)molybdenum(phosphine)3
Organometallics, Vol. 18, No. 24, 1999 5013
While there are many examples of the first two types
of reactions, only one example of addition of iminium
salts has been reported, the reaction of [(η4-C6H6)Mn-
(CO)3]- with [R2CdNR2]+ to give (η5-C6H6CR2NR2)Mn-
(CO)3.39 In this case the metal is anionic, the arene
ligand is bound η4, and the electrophile is cationic, which
are factors that are undoubtedly important. Other
electrophiles (besides of course protons) apparently do
not react with [(η4-C6H6)Mn(CO)3]-.40 We have made
similar observations for 1. Indeed, 1 reacts with [CH2d
NMe2]I (Eschenmoser’s salt) to give [(η6-C6H5(CH2-
NMe2))Mo(TRIPOD)(H)]I,41 which has been character-
ized in situ by its distinctive NMR spectrum; however,
the product has proven difficult to isolate. Compound 1
does not react with other electrophiles such as Et3OBF4
and F3CSO3Me. Instead, hydrolysis products from the
electrophiles simply protonate 1 to give 1H+. Alkyl
halides do react with 1 under neat conditions at elevated
temperatures in the absence of Lewis acid catalysts to
give mixtures of 1H+ and [(η6-C6H5R)Mo(TRIPOD)-
(H)]+.41 We suspect these reactions are the result of
initial single electron transfer (SET). Interestingly, we
were unable to achieve Friedel-Crafts acetylation of 1
with CH3COCl/AlCl3. Protonation, it would seem, is the
only clean electrophilic reaction that occurs for 1 and
other (η6-arene)Mo(P)3 complexes. Before turning our
attention to the mechanism of protonation of these
complexes, we will first discuss the protonation of η6-
arene complexes and make a distinction between such
reactions and those for η2- and η4-arene complexes.
P r oton a tion of η6-Ar en e Meta l Com p lexes. Pro-
tonation reactions of η6-arene metal complexes were first
studied by Wilkinson et al. in 1962.42 In that study a
series of arene-Cr complexes in a BF3‚H2O/CF3COOH
mixture were shown to have protonated metal centers
Olah’s work, McGlinchey has suggested that the pro-
tonated metal centers of [(η6-C6Et6)M(CO)3(H)]+ exhibit
an unprecedented static C3v geometry with the hydride
ligands oriented along the symmetry axis, enforced by
the steric demands of the arene ligand. Molecular orbital
calculations at various levels suggest such a geometry
is disfavored electronically.10,50 Furthermore, their ar-
gument is based upon the simplicity of NMR spectra
seen at low temperature, but we have observed dynamic
spectra for 1H+ under similar conditions. Finally, it
should be pointed out that phosphine ligands are more
basic and poorer π-acceptors than carbonyls;10 therefore,
substitution of one or more of the carbonyl ligands of
(η6-arene)M(CO)3 with phosphine ligands can produce
highly electron-rich metal complexes that are readily
protonated.14,51-55 However, phosphine substitution does
not have a marked effect on the ordering or character
of the frontier orbitals of three-legged piano-stool com-
plexes.10
P r oton a tion of η2- a n d η4-Ar en e Meta l Com -
p lexes. Electrophilic reactions with arene complexes
which are η2 and η4 bound are also known. Cooper et
al. have protonated [(η4-C6H6)Cr(CO)3]2- and [(η4-C6H6)-
Mn(CO)3]- complexes to form cyclohexadienyl species,
and in the former case protonation was shown to be
metal mediated (i.e., protonation of the metal followed
by proton transfer to the endo face of the arene, precisely
the opposite mechanism compared to that observed for
1).56,57 This example clearly illustrates the significance
of steric and electronic factors (vide infra). Dihapto-
coordinated aromatic hydrocarbons have been proto-
nated to form the cyclohexadienyl cations.58 In a recent
study, Harman et al. have protonated [Os(NH3)5(η2-
arene)](OTf)2 to form the [Os(NH3)5(η3-areneH)](OTf)3
complexes in which areneH is an η3-bound cyclohexa-
dienyl ligand.59 Given the electron-withdrawing nature
of the metal, it seems likely that electrophilic attack on
η2- and η4-arene metal complexes takes place at the
uncoordinated double bond(s).
1
characterized by high-field H NMR resonances. In the
intervening years, derivatives of (η6-arene)Cr(CO)3 have
been thoroughly studied due in part to the ease of their
syntheses.43-49 Particularly relevant to the present
study is an investigation by McGlinchey et al. of the
protonation of (η6-C6Et6)M(CO)3 (M ) Cr and W) using
13C NMR and computational methods.50 In contrast to
Ster eoelectr on ic F a ctor s Th a t Gover n th e Site
of Atta ck by H+ on 1-5. Protonation of 1-5 in aprotic,
nonpolar media such as dichloromethane presumably
involves delivery of H+ as a tight ion pair with its
conjugate base (X-). It is reasonable to expect that the
nature of the base might influence the mechanism.
Thus, a strong base would involve less polarization of
the H-X bond and a bulky X- group might discourage
direct attack by HX on the metal center. While these
issues may be important in determining selectivity, the
(34) Ercoli, R.; Calderazzo, F.; Mantica, E. Chim. Ind. (Milan) 1959,
41, 404.
(35) Gracey, D. E. F.; J ackson, W. R.; J ennings, W. R. J . Chem. Soc.,
Chem. Commun. 1968, 366.
(36) Herberich, G. E.; Fischer, E. O. Chem. Ber. 1962, 95, 2803.
(37) J ackson, W. R.; J ennings, W. B. J . Chem. Soc., Chem. Commun.
1966, 824.
(38) J ackson, W. R.; J ennings, W. B. J . Chem. Soc. B 1969, 1221.
(39) Park, S.-H. K.; Geib, S. J .; Cooper, N. J . J . Am. Chem. Soc.
1997, 119, 8365.
(40) Thompson, R. L. Ph.D. Thesis, University of Pittsburgh, 1993.
(41) Asirvatham, V. S.; Ashby, M. T. Unpublished results.
(42) Davison, A.; McFarlane, W.; Pratt, L.; Wilkinson, G. J . Chem.
Soc. 1962, 3653.
(43) Olah, G. A.; Yu, S. H. J . Org. Chem. 1976, 41, 717.
(44) Kursanov, D. N.; Setkina, V. N.; Petrovskii, P. V.; Zdanovich,
V. I.; Baranetskaya, N. K.; Rubin, I. D. J . Organomet. Chem. 1972,
37, 339.
(45) Setkina, V. N.; Baranetskaya, N. K.; Ginzburg, A. G.; Zdanovich,
V. I.; Nefedova, M. N.; Kursanov, D. N. J . Organomet. Chem. 1973,
61, 287.
(46) Losilkina, V. I.; Petrovskii, P. V.; Baranetskaya, N. K.; Setkina,
V. N. Bull. Acad. Sci. USSR, Div. Chem. Sci. 1984, 33, 723.
(47) Rosca, S.; Chiraleu, F.; Rosca, S. I. Rev. Roumaine Chim. 1991,
36, 693.
(51) Green, M. L. H.; Mitchard, L. C.; Silverthorn, W. E. J . Chem.
Soc. A 1971, 2929.
(52) Green, M. L. H.; Mitchard, L. C.; Silverthorn, W. E. J . Chem.
Soc., Dalton Trans. 1974, 1361.
(53) Green, M. L. H.; Hughes, A. K.; Lincoln, P.; Martin-Polo, J . J .;
Mountford, P.; Sella, A.; Wong, L.; Bandy, J . A.; Banks, T. W.; Prout,
K.; Watkin, D. J . J . Chem. Soc., Dalton Trans. 1992, 2063.
(54) Losilkina, V. I.; Petrovskii, P. V.; Baranetskaya, N. K.; Setkina,
V. N.; Kursanov, D. N. Bull. Acad. Sci. USSR, Div. Chem. Sci. 1983,
32, 2537.
(55) Losilkina, V. I.; Baranetskaya, N. K.; Tolkunova, V. S.; Krylova,
A. I.; Vil’chevskaya, V. D.; Setkina, V. N. J . Gen. Chem. USSR (Engl.
Transl.) 1987, 57, 753.
(56) Leong, V. S.; Cooper, N. J . J . Am. Chem. Soc. 1988, 110, 2644.
(57) Thompson, R. L.; Lee, S.; Rheingold, A. L.; Cooper, N. J .
Organometallics 1991, 10, 1657.
(58) Winemiller, M. D.; Kopach, M. E.; Harman, W. D. J . Am. Chem.
Soc. 1997, 119, 2096.
(59) Winemiller, M. D.; Kelsch, B. A.; Sabat, M.; Harman, W. D.
Organometallics 1997, 16, 3672.
(48) Losilkina, V. I.; Estekhina, M. N.; Baranetskaya, N. K.; Setkina,
V. N. J . Organomet. Chem. 1986, 299, 187.
(49) Lillya, C. P.; Sahatjian, R. A. Inorg. Chem. 1972, 11, 889.
(50) Mailvaganam, B.; Sayer, B. G.; McGlinchey, M. J . Organome-
tallics 1990, 9, 1089.

5014 Organometallics, Vol. 18, No. 24, 1999
Ashby et al.
the direction of the η6-arene ligand. On the other hand,
no such conformational restrictions exist for the PPh2-
Me derivative 3; yet essentially the same conformations
about the M-P bonds of 1 and 3 are observed in their
crystal structures, and calculations further suggest the
conformation of 3 that is observed in the solid state is
thermodynamically preferred. Accordingly, 1 and 3
exhibit the same reactivity pattern (100% attack on the
arene ligand). Although the reactivity pattern of 2 may
seem peculiar, space-filling models based upon the
Ta ble 3. P a r tition in g of Meta l ver su s Ar en e
Atta ck by H+ on (η6-C6D6)Mo(P )3 u p on Rea ction
w ith 1 equ iv of F 3CSO3H a t 22 °C As Mea su r ed by
1H a n d 2H NMRa
1H NMR
2H NMR
metal
attack
(%)
arene
attack
(%)
metal
attack
(%)
arene
attack
(%)
cmpd, ligand
1, TRIPOD
2, TRIPHOS
3, (PPh2Me)3
4, (PPhMe2)3
5, (PMe3)3
0
88
0
22
37
100
12
100
78
0
72
0
16
39
100
28
100
84
crystal structure of 2 suggest that ready access to the
-
63
61
side of the molecule (between Mo-P(Ph)(CH2
) and
2
Mo-P(Ph)2(CH2-) bonds) is possible (Figure 8, left), and
this is confirmed in the surface electron density plots
of Figure 8 (right). Semiempirical PM3(tm) calculations
nicely predict essentially the conformations of 1-5 that
are observed experimentally. Furthermore, mapping the
occupied metal-containing frontier orbitals onto plots of
the calculated surface electron density for 1-5 renders
convenient models that predict the preferred sites of
electrophilic attack (Figure 8, right). The electron
density plots of 1 and 3 reveal no accessible metal
electron density (except vis-a`-vis their arene ligands).
In contrast, electron density plots for 2, 4, and 5 suggest
that direct access to metal electron density is possible
for these species. Although these plots suggest that
direct attack on the metal centers is possible, they do
not predict the relative reactivities of 2, 4, and 5.
However, the CPK models clearly reveal increased steric
accessibility to the metal centers in the order 4 < 5 <
2, which happens to be the order of reactivity at the
metal centers that is observed experimentally (as evi-
denced by the relative partitioning of attack at the metal
versus attack at the arene ligand). Of considerable note
is the plot for 2 that predicts attack at the metal to give
the kinetic hydride 2KH+ and not the thermodynamic
hydride 2TH+. Although even more surface frontier
orbital electron density appears available between the
two (-CH2)P(Ph)2 groups (Figure 8, right), that site is
sterically shielded by the cleft formed by aryl substit-
uents (Figure 8, left).
a
Note: these results were obtained from separate experiments
in CD2Cl2 for the 1H NMR and CH2Cl2 for the 2H NMR. The
percentages are derived from integrations of the arene and hydride
ligand resonances taking into account the ratio of the C6D6/C6D5H
isotopomers that were reacted with H+. It is assumed that H+
attack on the C6D5H isotopomer takes place without a KIE (i.e.,
D+ will be transferred to the metal 5/6 of the time and H+ will be
transferred 1/6 of the time).
purpose of this study has been to explore the influence
of the ancillary ligands. Furthermore, we have obtained
very similar results for the tracer studies when F3-
CSO3H, F3CCO2H, HPF6 (aq), HBF4‚(OEt2), and HCl
(aq) are employed.
Our recent study of the electronic structures of 1 and
1+ illustrated that two of the three filled metal-based
frontier orbitals have substantial arene character.10
Thus, while protonation of the metal center is thermo-
dynamically preferred, an incoming electrophile that is
prevented from attacking the metal center directly by
its own steric demands or those imposed by the ligands
about the metal can still access metal electron density
that is delocalized onto the η6-arene ligand. Indeed,
protonation as it turns out is a fine probe of subtle
differences between the sterics of different ancillary
ligands. Such issues are particularly important for
phosphine ligands which are frequently used as chiral
ancillary ligands in asymmetric reactions that are
catalyzed by metals.60-63 Selectivity of such catalytic
systems is largely dictated by the sterics of the ancillary
ligands. Perhaps the most interesting aspect of the
selectivity that is observed in the protonation of 1-5 is
the fact that attack on the metal is not dictated strictly
by the substitution patterns of the phosphine ligands.
Thus, the derivative that is most readily attacked at
the metal (2, the TRIPHOS derivative) arguably bears
one of the more bulky phosphine ligands, whereas the
derivative that bears presumably less bulky trialkyl-
phosphine ligands (5, the PMe3 derivative) still exhibits
substantial amount of attack at the arene ligand.
It is clear that the conformations of unsymmetrically
substituted phosphines about their M-P and P-C
bonds play an important role in partitioning electro-
philic attack on the metal versus the arene ligand. The
conformations about the M-P bonds of some of the
ligands are enforced by chelation. For example, the aryl
substituents of the TRIPOD derivative 1 are forced in
Dyn a m ic P r op er ties of 1H+ a n d 2H+. Complex
1H+ exhibits Cs symmetry in solution and in the solid
state.9 Thus, the phosphine ligands are rendered chemi-
cally inequivalent. This is evidenced in the low-temper-
ature 31P NMR of 1H+ that reveals a 2:1 ratio of
phosphine resonances. A rate constant of 5.3 × 103 s-1
and activation barrier of ∆Gq ) 35 kJ mol-1 was
measured for stereoisomerization of 1H+ at the coales-
cence temperature of -70 °C. We have measured the
same barrier for 1H+-d 7, suggesting no kinetic isotope
effect for the stereoisomerization. The lack of an isotope
1
effect and the preservation of H/31P coupling during
isomerizations indicate an intramolecular process. There
does not appear to be room in the pocket of the tripodal
phosphine ligand of 1H+ to accommodate the hydride
ligand, so the barrier to isomerization is apparently
associated with navigation of the hydride ligand be-
tween the Mo-P and Mo-arene groups. The rate
constant of 6.4 × 103 s-1 and comparable activation
barrier of ∆Gq ) 31 kJ mol-1 for interconversion of the
two enantiomers of 2KH+ at the coalescence tempera-
ture of -87 °C indicate a similar process. However, the
isomerization of 2KH+ to 2TH+ with a rate constant of
(60) Asymmetric Catalysis; Bosnich, B., Ed.; Martinus Nijhoff:
Dordrecht, 1986.
(61) Kagan, H. B. In Comprehensive Organometallic Chemistry;
Wilkinson, G., Stone, F. G. A., Eds.; Pergamon: Oxford, 1982; Vol. 8,
pp 463-498.
(62) Noyori, R. Science 1990, 248, 1194.
(63) Ojima, I. Pure Appl. Chem. 1984, 56, 99.

Protonation of (η6-Arene)molybdenum(phosphine)3
Organometallics, Vol. 18, No. 24, 1999 5015
6.1 × 10-4 s-1 and an activation barrier of ∆Gq ) 65 kJ
mol-1 at -60 °C appears quite different. We attribute
the higher barrier of the stereoisomerization of 2KH+
to 2TH+ compared with stereoisomerization of 2KH+ to
attack by protons on the metal is favored over attack
on the arene ligand, and equatorial attack on (η6-arene)-
Mo(P)3 is preferred over axial attack:
steric factors. The CPK model of 2 clearly reveals a gap
-
between the arene ligand and the M-P(Ph)(CH2
)
2
bond, but no such gap exists between the arene ligand
and the M-P(Ph)2(CH2-) bonds.
Qu a n tifica tion of Ster ic Effects. The relationship
between structure and reactivity64 and the concept of
steric hindrance65 have been recognized for more than
a century. In the interim there have been may attempts
to quantify steric effects. The approaches may be
classified as (1) direct physical measurement via metric
data, (2) inference of relative size from trends in
reactivity, and (3) the use of molecular mechanics and/
or molecular orbital models. Perhaps the most widely
applied measure of sterics in organometallic chemistry
(and an example of the first approach) is the “cone
angle” (θ) introduced by Tolman as a means of describ-
ing the steric requirements of phosphine ligands.66,67
Rules were later developed to deal with nonaxial sub-
stituents and unsymmetrically substituted phosphines.
Thus, the values of θ for PMe3, PPhMe2, and PPh2Me
are 118°, 122°, and 136°, respectively. This is the same
trend that we observe in the reactivities of 3-5 toward
protons. Although Tolman’s original definition of θ was
based on a “hard sphere” model, it is well recognized
that steric and electron factors are both important,68 and
these factors are frequently inextricably commingled.
Since Tolman’s first effort to quantify the steric require-
ments of phosphine, there have been many modifica-
tions to Tolman’s original parameters. For example,
Stahl and Ernst69 corrected θ based on experimental
bond dissociation enthalpies in bis(2,4-dimethylpenta-
dienyl)titanium complexes, and Mosbo et al. made
related corrections using computed geometries and
heats of formation.70 Both of these studies reported the
same trend for PMe3 and PPhMe2, but PPh2Me was not
considered. The relative ordering of monodentate phos-
phine ligands according to their steric requirements is
relatively straightforward, and many reactivity trends
can be understood using such models. Although the
steric demands for bidentate phosphine ligands have
been modeled by the Tolman method,66,67,71,72 and such
treatments have proven valuable in predicting reactivity
trends,73 to our knowledge no such analysis of tridentate
phosphine ligands has been attempted. The present
study in effect makes use of all three approaches (vide
supra) to evaluate the steric environments about the
metal centers of 1-5. We conclude for the following
reasons that, in the absence of dominating steric effects,
1. Derivatives 1 and 3 exhibit comparable orientations
about their Mo-P bonds and, therefore, similar equato-
rial steric environments. The axial approach is blocked
by the TRIPOD backbone in the case of 1 and compara-
tively open in the case of 3 (cf. 5). Nonetheless, 1 and 3
are both attacked 100% at the arene, which indicates a
preference for attack at the arene ligand rather than
axial attack in the cases where equatorial approach is
hindered. 2. The reactivity trends of 3-5 follow the
steric nature of the substituents in the equatorial plane
((Ph)2 vs (Ph)(Me) vs (Me)2). 3. Unlike the C3 symmetry
of 4 that orients one Ph substituent along each of the
three possible equatorial trajectories, the C1 symmetry
of 2 allows the -CH2P(Ph)CH2- moiety to block ap-
proach from only one side, leaving the other side
exposed (cf. Figure 2):
The axial approach is relatively unencumbered for 2,
once again suggesting an electronic preference for
equatorial attack. Accordingly, both steric and electronic
factors play important roles in determining the kinetic
sites of protonation of 1-5.
Con clu sion s
We conclude from the crystal structures of 1-4, tracer
studies, and electronic structure calculations for the five
derivatives 1-5 presented herein that the steric de-
mands of the phosphine ligands play an important role
in dictating whether a proton attacks the arene ligand
or the metal of 1-5. A priori prediction of the relative
steric requirements of the phosphine ligands that have
been employed in this study is a challenge (cf. the
TRIPHOS and PMe3 derivatives). However, mapping
the metal-based frontier orbitals of the complexes 1-5
onto plots of their surface electron density (thereby
revealing sterically accessible metal electron density)
calculated using the semiempirical PM3(tm) method
and geometries derived from both the crystal structure
coordinates (for 1-4) and geometries that were opti-
mized using the PM3(tm) method (for 1-5) provides a
straightforward model that predicts the mechanisms of
attack by protons.
(64) Hofmann, A. W. Chem. Ber. 1872, 5, 704.
(65) Meyer, V. Chem. Ber. 1894, 27, 510.
(66) White, D.; Coville, N. J . Adv. Organomet. Chem. 1994, 36, 95-
158.
(67) Tolman, C. A. Chem. Rev. 1977, 77, 313.
(68) For an excellent example of ancillary ligands influencing the
relative importance of electronic (i.e., pKb) and steric (i.e., θ) factors
see: Geno, M. K.; Halpern, J . J . Am. Chem. Soc. 1987, 109, 1238.
(69) Stahl, L.; Ernst, R. D. J . Am. Chem. Soc. 1987, 109, 5673.
(70) DeSanto, J . T.; Mosbo, J . A.; Storhoff, B. N.; Bock, P. L.; Bloss,
R. E. Inorg. Chem. 1980, 19, 3086.
(71) Casey, C. P.; Whiteker, G. T. Isr. J . Chem. 1990, 30, 299.
(72) Casey, C. P.; Whiteker, G. T.; Campana, C. F.; Powell, D. R.
Inorg. Chem. 1990, 29, 3376.
(73) Casey, C. P.; Whiteker, G. T.; Melville, M. G.; Petrovich, L. M.;
Gavney, J . A., J r.; Powell, D. R. J . Am. Chem. Soc. 1992, 114, 5535.

5016 Organometallics, Vol. 18, No. 24, 1999
Ashby et al.
and equivalent isotropic displacement parameters, anisotropic
displacement parameters, hydrogen coordinates and isotropic
parameters, bond lengths and angles for 1-4, and representa-
tive spectra. X-ray crystallographic files for 1-4, in CIF
formats, are available through the Internet only. This material
is available free of charge via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. The financial support of the
National Science Foundation (CHE-9612869) and the
Oklahoma Center for the Advancement of Science and
Technology (OCAST HR98-078) is gratefully acknowl-
edged.
Su p p or tin g In for m a tion Ava ila ble: Thermal ellipsoid
drawing of 2 showing its disorder; tables of atomic coordinates
OM990281N