Reversible Displacement of Polyagostic Interactions in 16e Mn(CO)(R2PC2H4PR2)2} by H2,N2, and so2. Binding and Activation of η2-H2 trans to CO Is Nearly Invariant to C

Article abstract of DOI:10.1021/ic981263l

Electrophilic l6e [Mn(CO)(R₂PC₂}LPR₂)₂]⁺ complexes (R = Et, Ph) are synthesized by metathesis of MnBr-(CO)(R₂PC₂H₄PR₂)₂ with Na or Li salts of lo

Full text of DOI:10.1021/ic981263l

Inorg. Chem. 1999, 38, 1069-1084
1069
Reversible Displacement of Polyagostic Interactions in 16e [Mn(CO)(R2PC2H4PR2)2]+ by H2,
N2, and SO2. Binding and Activation of η2-H2 trans to CO Is Nearly Invariant to Changes
in Charge and cis Ligands
Wayne A. King, Brian L. Scott, Juergen Eckert, and Gregory J. Kubas*
Los Alamos National Laboratory, CST-18, MS-J514, Los Alamos, New Mexico 87545
ReceiVed October 29, 1998
Electrophilic 16e [Mn(CO)(R2PC2H4PR2)2]+ complexes (R ) Et, Ph) are synthesized by metathesis of MnBr-
(CO)(R2PC2H4PR2)2 with Na or Li salts of low-coordinating boron or gallium anions (e.g., [B{C6H3(3,5-CF3)2}4]-
or [Ga(C6F5)4] -). They contain weak polyagostic interactions that are reversibly displaced by H2, N2, and SO2
(which is a surprisingly weak ligand here). The agostic and H2 complexes, as well as the gallium anions including
the new species [{Ga(C6F5)3}2(µ-Cl)]-, have been characterized by NMR, IR, and X-ray crystallography. The
agostic Mn-H distances (e.g., 2.9 Å) are much longer than those found for the single agostic interactions in
Mo(CO)(diphosphine)2 and [Mn(CO)3(PCy3)2]+. The H-H and also the Mn-H distances have been determined
in the H2 complex by T1 measurements for both the H2 and HD isotopomers. IR data and C-O and M-C bond
lengths are used to gauge the π-acceptor strengths of ligands trans to the CO. The agostic C-H bonds are the
weakest ligands and also the weakest acceptors, but the H2 ligand is an excellent acceptor as strong as N2 and
ethylene. The variation of ν(CO) on increasing the basicity of the cis-phosphine (dppe versus depe) in trans-
M(CO)(diphosphine)2(L) is less than expected and far less than that on increasing the charge on the complex (M
) Mn+ versus Mo). The H-H bond lengths (0.87-0.90 Å) and J(HD) NMR couplings (32-34 Hz) in [Mn-
(CO)(R2PC2H4PR2)2(H2)]+ and other cationic H2 complexes with trans-CO are strikingly similar to their neutral
analogues and nearly invariant. Activation of H-H in the more electrophilic cationic systems occurs primarily
via increased σ donation from H2 as compared to the more electron-rich neutral analogues where back-bonding
dominates. The nature of the ligand trans to H2 (the strong acceptor CO here) controls the H-H distance more
so than all of the cis ligands combined, especially for cationic complexes.
Introduction
EBD but is closer than expected to that in Mo(CO)(R2PC2H4-
PR2)2(H2), where EBD is high and ED is lower.1b Dihydrogen is
an extremely versatile ligand and can bind to highly electrophilic
systems because of this amphoteric behavior. Importantly,
excessive back-bonding promotes X-H rupture, i.e., oxidative
addition (OA) but increasing σ donation from the X-H bond
to the metal cannot by itself give OA. We have recently focused
on developing highly electrophilic cationic fragments3 such as
[Mn(CO)(dppe)2]+, [Mn(CO)3(PCy3)2]+, [PtH(PPri3)2]+, and
[Re(CO)4(PCy3)]+ for systematic studies of the binding of H2,
silanes, and potentially, alkanes (dppe ) Ph2PC2H4PPh2). The
positive charge here favors η2 coordination over OA, and the
degree of activation of the H-H bond in these and other cationic
or dicationic H2 complexes is remarkably similar to that in
neutral analogues, as judged by H-H distance, J(HD), and other
parameters. For example, despite their increasing charge and
electrophilicity, the isoelectronic 16e fragments Mo(CO)-
(dppe)2,4 [Mn(CO)(dppe)2]+,3a and [M(CO)(diphosphine)2]2+ (M
The activation of H-H and other strong, inert σ bonds on
transition-metal centers is a foundation of catalysis and many
types of chemical/biochemical conversions. Metal-H2 binding
and by analogy X-H coordination are governed by both σ
donation from the X-H bond to the metal (ED) and π back-
bonding from metal to X-H σ* (EBD). The relative strength of
each bonding component has been quantified calculationally by
both Ziegler1 and Frenking2 and is dependent on the electronic
nature of the ancillary ligands. Donor ligands enhance EBD while
acceptors promote increased ED, and there is a fine balance
between the effects. For example, the M-H2 bonding energy
in electron-poor Mo(CO)5(H2) has a high ED and much smaller
(3) (a) King, W. A.; Luo, X.-L.; Scott, B. L.; Kubas, G. J.; Zilm, K. W.
J. Am. Chem. Soc. 1996, 118, 6782. (b) Butts, M. D.; Kubas, G. J.;
Scott, B. L. J. Am. Chem. Soc. 1996, 118, 11831. (c) Huhmann-
Vincent, J.; Scott, B. L.; Kubas, G. J. J. Am. Chem. Soc. 1998, 120,
in press. (d) Toupadakis, A. I.; Kubas, G. J.; King, W. A.; Scott, B.
L.; Huhmann-Vincent, J. Organometallics, submitted for publication.
(e) Hay, P. J.; Kubas, G. J., manuscript in preparation.
(4) (a) Kubas, G. J.; Burns, C. J.; Eckert, J.; Johnson, S.; Larson, A. C.;
Vergamini, P. J.; Unkefer, C. J.; Khalsa, G. R. K.; Jackson, S. A.;
Eisenstein, O. J. Am. Chem. Soc. 1993, 115, 569. (b) Sato, M.; Tatsumi,
T.; Kodama, T.; Hidai, M.; Uchida, T.; Uchida, Y. J. Am. Chem. Soc.
1978, 100, 447.
(1) (a) Li, J.; Dickson, R. M.; Ziegler, T. J. Am. Chem. Soc. 1995, 117,
11482. (b) Li, J.; Ziegler, T. Organometallics 1996, 15, 3844.
(2) (a) Dapprich, S.; Frenking, G. Angew. Chem., Int. Ed. Engl. 1995,
34, 354. (b) Ehlers, A. W.; Dapprich, S.; Vyboishchikov, S. F.;
Frenking, G. Organometallics 1996, 15, 105. (c) Frenking, G.; Pidum,
U. J. Chem. Soc., Dalton Trans. 1997, 1653. (d) Dapprich, S.;
Frenking, G. Organometallics 1996, 15, 4547.
10.1021/ic981263l CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/27/1999

1070 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
Table 1. Infrared and NMR Data
complex
ν(CO)a
δ 1Hb
δ
31P{1H}
MnBr(CO)(depe)2
1803
2.30, 1.86, 1.69, 1.19
1.59, 1.39, 0.92
1.8-0.7
74.2
81.4
80.3
75.8
83.5
83.4
83.3
72.6
[Mn(CO)(depe)2][BAr′4]
1853
[Mn(CO)(depe)2][Ga(C6F5)4]
[Mn(CO)2(depe)2][BAr′4]
[Mn(CO)(depe)2(H2)][BAr′4]
[Mn(CO)(depe)2(H2)][Ga(C6F5)4]
[Mn(CO)(depe)2(D2)][Ga(C6F5)4]
[Mn(CO)(depe)2(N2)][BAr′4]
MnBr(CO)(dppe)2
1888
1887
1.90, 1.66, 1.24
1.65, 1.22, 0.93, 0.71, -10.25
1.7-0.8, -10.23
1.5-0.6
1896
1821
1862 [1839]
1.68, 1.53, 1.39, 0.94
[Mn(CO)(dppe)2][BAr′4]
7.3-7.0, 6.16, 2.79
7.3-7.0, 6.23, 2.79
82.6
83.0
85.5
85.7
85.9
75.0
[Mn(CO)(dppe)2][Ga(C6F5)4]
[Mn(CO)(dppe)2(H2)][BAr′4]
[Mn(CO)(dppe)2(H2)][Ga(C6F5)4]
[Mn(CO)(dppe)2(D2)][Ga(C6F5)4]
[Mn(CO)(dppe)2(N2)][BAr′4]
[1896]
7.4-7.0, 2.52, 2.24, -7.23
7.4-7.0, 2.50, 2.24, -7.22
7.4-7.0, 2.50, 2.24
1911
7.3-7.1, 6.67, 2.67, 2.34
a In cm-1 for Nujol mulls except for values in brackets (for dichloromethane solution). b In ppm for fluorobenzene (depe complexes) or
dichloromethane (dppe complexes) solutions. All are multiplets except the broad high-field singlet for the H2 ligand. Anion resonances are not
listed (see Experimental Section). The field strength was 300 MHz, except for data where the PC2H4P bridge multiplet was resolved into several
multiplets (500 MHz).
) Fe, Ru, Os)5 all coordinate H2 reversibly trans to CO, with
H-H distances near 0.88 Å and J(HD) ) 32-34 Hz. The recent
phosphite analogues, [Mn(CO){P(OR)3}4]+, also display these
properties.6 The presence of the strong π-acceptor CO trans to
H2 is a crucial factor in these systems and greatly moderates
the lengthening of the H-H bond.
of MnBr(CO)5 in benzene in the presence of excess depe in a
procedure analogous to that previously reported for MnBr(CO)-
(dppe)2.7 1H NMR displays a multiplet at δ 1.1-2.3 resolvable
into four nonintegrable overlapping components (Table 1). The
presence of four resonances is consistent with Br trans to CO
and the depe ligands coordinated in a plane if it is assumed
that the chemical shift differences of the Et groups above and
below the plane are unresolved. 31P{1H} NMR shows a single
resonance at δ 74.2 also consistent with Br trans to CO. IR
spectra reveal a strong CO band at 1803 cm-1 comparable to
that for MnBr(CO)(dppe)2 (1821 cm-1 ). The lowering of the
CO stretch reflects increased π back-donation from Mn to CO
on substitution of dppe by the better donor depe.
As reported in a preliminary communication,3a the cationic
16e complex [Mn(CO)(R2PC2H4PR2)2]+ where R ) Ph was
readily prepared by metathesis of MnBr(CO)(R2PC2H4PR2)2
-
with salts of the low-coordinating anion B[C6H3(3,5-CF3)2]4
(BAr′4-), which is necessary to stabilize these highly reactive
cations. Herein we report in detail the complexes for both R )
Synthesis of Li[Ga(C6F5)4]. Li[Ga(C6F5)4] is prepared by
reaction of a slight excess of LiC6F5 with GaCl3 in hexane at
1
-78 °C. H and 19F NMR spectra are consistent with those
reported for [Bu4N][Ga(C6F5)4].8 Although elemental analysis
of Li[Ga(C6F5)4] yields inconsistent results owing to entrapment
of solvent and unidentified fluorocarbon impurities observed
by 19F NMR, the material can be used for subsequent reactions
without further purification. Pure samples of Li[Ga(C6F5)4] are
prepared as the etherate Li[Ga(C6F5)4]‚2Et2O by reaction of Ga-
(C6F5)3‚Et208,9 with 1 equiv of LiC6F5 in Et2O at -78 °C. Li-
[Ga(C6F5)4]‚2Et2O was not used as starting material in the
synthesis of [Mn(CO)(diphosphine)2]+ because Et2O can serve
as a weak ligand.
Synthesis of [Mn(CO)(depe)2][A] (A ) BAr′4, Ga(C6F5)4).
Stoichiometric reaction of NaBAr′4 with MnBr(CO)(depe)2 in
fluorobenzene conveniently yields [Mn(CO)(depe)2][BAr′4] as
a deep blue formally 16e complex analogous to the similarly
prepared [Mn(CO)(dppe)2][BAr′4].3a The bromide is metathe-
Et and R ) Ph and also the use of gallium-based anions as
useful alternates to boron-based anions. 10B is an interference
in neutron scattering studies of cationic H2 complexes and other
compounds, and development of non-boron-containing, low-
interacting counterions is important for such studies. The
unsaturated cations contain weak polyagostic C-H interactions
trans to CO that are displaced by H2, N2, and SO2, which is an
abnormally weak ligand here. Surprisingly, increasing the donor
ability of the phosphines by varying R from Ph to Et in [Mn-
(CO)(R2PC2H4PR2)2(H2)]+ has only a minor effect on the
activation of H2. This points to the important concept that the
influence of the cis-ligand set on H2 actiVation can be
inconsequential and is nearly always far less important than
that of the trans ligand. As will be shown, this seems to be
particularly true for cationic systems where back-bonding is
lower.
Results
Synthesis of MnBr(CO)(depe)2. MnBr(CO)(depe)2 is pre-
pared as a yellow, moderately air-sensitive solid by photolysis
sized off the complex as NaBr, giving a cationic system with a
(5) (a) Forde, C. E.; Landau, S. E.; Morris, R. H. J. Chem. Soc., Dalton
Trans. 1997, 1663. (b) Rocchini, E.; Mezzetti, A.; Rugger, H.;
Burckhardt, U.; Gramlich, V.; Zotto, A. D.; Martinuzzi, P.; Rigo, P.
Inorg. Chem. 1997, 36, 711.
(6) Albertin, G.; Antoniutti, S.; Bettiol, M.; Bordignon, E. Busatto, F.
Organometallics 1997, 16, 4959.
(7) Reimann, R. H.; Singleton, E. J. Organomet. Chem. 1972, 38, 113.
(8) Ludovici, K.; Tyrra, W.; Naumann, D. J. Organomet. Chem. 1992,
441, 363.
(9) Pohlmann, J. L. W.; Brinckmann, F. E. Z. Naturforsch., Teil B, 1965,
20, 5.

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1071
noncoordinating anion (A) and polyagostic interactions. Unlike
the dppe congener, the depe complex is extremely sensitive to
oxidation by air and CH2Cl2 and coordination by trace N2. In
fact, the high sensitivity of this system to oxidation makes study
by in situ generation in an NMR tube preferable to bulk
synthesis. A preparative scale reaction under Ar gives crystalline
[Mn(CO)(depe)2][BAr′4] containing a small amount of [Mn-
(CO)2(depe)2][BAr′4] (identified by 31P NMR) resulting from
oxidation by trace dioxygen and transfer of released CO to [Mn-
(CO)(depe)2]+. The identity of [Mn(CO)2(depe)2]+ is confirmed
by reaction of in situ generated agostic complex with CO in an
NMR tube reaction, as discussed below. Although isolated [Mn-
(CO)(depe)2]+ contains up to 15% dicarbonyl complex from
integration of 31P NMR spectra, a strong CO band is observable
at 1853 cm-1 which is slightly lower than that for the more
electrophilic dppe congener (1862 cm-1). The higher ν(CO) of
the agostic cation as compared to the parent bromide complex
is consistent with a more electron-deficient species. A CO band
observed at 1889 cm-1 is assigned to the dicarbonyl impurity,
as confirmed by independent synthesis (1888 cm-1).
the dark blue agostic complex, but not at -78 °C in frozen
fluorobenzene. The H2 complex is less oxygen sensitive than
the agostic complex and is isolated by precipitation with hexane
under H2 or even He, where little dissociation is observed
(although the light yellow microcrystalline complex will slowly
lose H2, which is more rapid in vacuo). The depe complex loses
H2 much less readily than the dppe complex owing to the greater
donor strength of depe and perhaps weaker agostic ethyl
interactions.
1H NMR for [Mn(CO)(depe)2(H2)][BAr′4] in C6D5F displayed
a multiplet at δ 1.7-0.7 for depe which was resolved at 500
MHz into four major overlapping multiplets. The broad H2
resonance, integrating to two protons, was located at δ -10.25
(19 Hz fwhm, 500 MHz), which displayed no observable 31P
coupling. A comparison to the δ -7.23 resonance for the dppe
analogue reveals greater shielding of H2 in the depe system,
again consistent with the greater donor properties of depe. The
HD complex, prepared by reacting in situ generated agostic
complex with HD gas, yielded a coupling constant of 33 Hz
indicative of a short H-H distance. The value for the dppe
analogue is only slightly less at 32 Hz. An estimated H-H
distance of 0.87-0.89 Å for the depe complex was calculated
from J(HD) using the correlations12 of Morris and Heinekey
and is similar to that for the dppe system as measured by solid-
state NMR (0.89(2) Å).3a Observation of a single 31P resonance
and four major 1H resonances for the depe ligand is consistent
with H2 trans to CO. On cooling the sample from 298 to 228
K, 1H NMR revealed small shifts for the depe ligands, the anion,
and the H2 ligand (to δ -10.33, with significant broadening to
72 Hz (fwhm, 500 MHz)). No resonances associated with an
oxidative addition product were observed. The 31P{1H} NMR
signal at 228 K was unshifted, and no other resonances were
All detailed NMR studies of the agostic complex were
1
performed on material generated in situ. H NMR in C6D5F
revealed a multiplet at δ 0.9-1.6 from the depe ligand and two
resonances at δ 8.32 and 7.65 typical for the BAr′4 anion.3a,10
The multiplet was resolved in the 500 MHz spectrum (298 K)
into three overlapping broad components at δ 1.59, 1.39, and
0.92. On cooling the sample to 228 K, the latter signals shifted
to δ 1.72, 0.99, and 0.22 and gained fine structure, but no signals
for the agostic interactions were observed. The resonances for
the anion shifted to δ 8.50 and 7.55. 31P{1H} NMR in C6D5F
at 298 K yielded a singlet at δ 81.4 that shifted slightly to δ
83.5 on cooling to 228 K. This is consistent with a highly
fluxional formally five-coordinate complex analogous to the
dppe congener.3a The existence of three major resonances in
the 1H NMR is also consistent with [Mn(CO)(depe)2]+ existing
in solution as a fluxional species. Variable temperature studies
in this system were limited by the relatively high freezing point
of fluorobenzene (-40 °C).
Addition of CO to a freshly generated sample of [Mn(CO)-
(depe)2]+ in fluorobenzene resulted in formation of a light
yellow solution of [Mn(CO)2(depe)2]+. 1H NMR (C6D6F)
revealed a multiplet at δ 1.9-1.2 that was resolved at 300 MHz
into three overlapping multiplets for the depe ligand (Table 1).
31P{H} NMR showed one resonance, and one CO stretch was
identified in the IR, consistent with trans CO as in [Mn(CO)2-
(dppe)2]PF6.11
[Mn(CO)(depe)2][Ga(C6F5)4] is generated in situ in an NMR
tube by the reaction of excess Li[Ga(C6F5)4] (the amounts
required vary according to purity) with MnBr(CO)(depe)2. The
properties and color of the [Ga(C6F5)4]- salt are identical to
those of the BAr′4- salt. Changing the anion has little effect on
the NMR spectroscopy of the cation, although small shifts in
the signals are seen (Table 1).
Synthesis of [Mn(CO)(depe)2(H2)][A]. The H2 complex is
cleanly generated by stoichiometric reaction of Na[BAr′4] with
MnBr(CO)(depe)2 in fluorobenzene under an atmosphere of H2.
The resulting yellow solution readily loses H2 in vacuo at room
temperature, as evidenced by the appearance of the color of
observed. IR data showed a strong CO band at 1887 cm-1
,
which is higher than that for the agostic complex and similar
to that for [Mn(CO)2(depe)2]+, indicating that the H2 is removing
electron density from the metal center through π back-bonding
like a CO ligand.
[Mn(CO)(depe)2(H2)][Ga(C6F5)4] is prepared by the reaction
of excess LiGa(C6F5)4 with MnBr(CO)(depe)2 under H2 in
fluorobenzene. The properties and NMR spectroscopy of the
boron and gallium salts are for the most part identical. The D2
complex is prepared under D2 and yields NMR spectra similar
to that for the H2 species except for the absence of the signal
for the H2 ligand.
Synthesis of [Mn(CO)(dppe)2][Ga(C6F5)4] and Its H2
Adduct. Deep blue microcrystals of [Mn(CO)(dppe)2][Ga-
(C6F5)4] are prepared by reaction of excess Li[Ga(C6F5)4] with
MnBr(CO)(dppe)2 under Ar in CH2Cl2, using procedures
identical to those for the preparation of the BAr′4 salt.3a 1H NMR
in CD2Cl2 revealed two multiplets at δ 7.3-7.0 and 6.23 for
the Ph groups and a single multiplet at δ 2.79 for the bridging
Et group of dppe. 31P{1H} NMR showed a singlet at δ 83.0.
These results, particularly the single multiplet for the PC2H4P
(10) (a) Heinekey, D. M.; Schomber, B. M.; Radzewich, C. E. J. Am. Chem.
Soc. 1994, 116, 4515. (b) Brookhart, M.; Grant, R. G.; Volpe, A. F.,
Jr. Organometallics 1992, 11, 3920. (c) Heinekey, D. M.; Radzewich,
C. E.; Voges, M. H.; Schomber, B. M. J. Am. Chem. Soc. 1997, 119,
4172.
(12) (a) Maltby, P. A.; Schlaf, M.; Steinbeck, M.; Lough, A. J.; Morris, R.
H.; Klooster, W. T.; Koetzle, T. F.; Srivastava, R. C. J. Am. Chem.
Soc. 1996, 118, 5396. (b) Luther, T. A.; Heinekey, D. M. Inorg. Chem.
1998, 37, 127.
(11) Garcia-Alonso, F. J.; Riera, V. Transition Met. Chem. 1985, 10, 19.

1072 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
bridge, are consistent with those found for the BAr′4 salt (Table
1), indicating that the cation is fluxional in solution.3a
to the spectral density function proposed by Morris must be
used.13a The constants in the above equations thus must be
multiplied by 0.25. Finally, on substituting the observed
relaxation values into eq 5 (R in units of s-1 and rHH in cm), a
value of 0.91(2) Å for the H-H distance is obtained that agrees
well with that determined by solid-state NMR (0.89(2) Å).3a
The H2 complex was prepared as above except under H2 in
CH2Cl2 using procedures similar to those for the preparation of
the BAr′4 salt.3a [Mn(CO)(dppe)2(H2)][Ga(C6F5)4] can also be
prepared by placing a CH2Cl2 solution of the agostic complex
under H2 and precipitating with hexane (direct reaction with
If the RH-ligand contributions are small, an equation for the
distance from Mn to the hydrogen atoms in the H2 ligand can
be derived using similar expressions and assumptions.
1
H2 in solid state fails). H NMR showed a muliplet at δ 7.4-
7.0 for the Ph groups and two multiplets at δ 2.50 and 2.24 for
the Et bridge of dppe. The dihydrogen 1H resonance was
observed at δ -7.22 (33 Hz fwhm at 300 Hz)). The NMR
signals for the [Ga(C6F5)4]- salt in CD2Cl2 are virtually identical
to those for the BAr′4 salt, indicating that the cation has an
octahedral geometry with CO trans to H2 and is unaffected by
the anion.
RHH(obs) ) RHH + RHMn
(6)
(7)
RHH(obs) ) (1.29 × 10-46)/rHH6 - (9.01 × 10-47)/rHMn
6
On substitution of rHH ) (1.21 × 10-46)/(RHH(obs) - RHD(obs)
)
6
The value of T1(min) for the H2 ligand in [Mn(CO)(dppe)2-
(H2)][Ga(C6F5)4] was determined to be 15.2(2) ms at 286 K at
300 K in CD2Cl2 and 116(8) ms for the HD analogue, which
are valuable measurements, as will be shown here. H-H
distances can be determined by T1 measurements, although the
effects of relaxation of the metal center on the H2 ligand must
be considered.13 Because 55Mn possesses a nuclear spin of 5/2,
the H-H distance measured in manganese-H2 complexes will
be systematically shortened by relaxation of the H2 ligand caused
by the metal center, RHMn. This is in addition to the effects of
relaxation caused by the ancillary ligands, RH-ligand. This is
expressed in eq 1 where R ) 1/T1.
into eq 7, eq 8 can be derived.
1.07RHD(obs) - 0.07RHH(obs) ) (9.01 × 10-47)/rHMn (8)
6
When the appropriate values are substituted into eq 8, a value
of 1.64(3) Å is obtained for the Mn-H distance, which is
reasonable when compared to the neutron diffraction value of
1.601(16) Å in MnH(CO)5.14 Often there is litle variation in
M-H distance in H2 and hydride complexes. This general
method is potentially useful to measure metal-dihydrogen
coordination geometries in other related systems. It is notewor-
thy that no correction term is needed here for determining the
M-H distances, which unlike H-H distances have not been
determined in solution (or in the solid state by methods other
than diffraction).
RHH(obs) ) RHH + RHMn + RHligand
(1)
The contribution to the relaxation of the hydrogen in the Mn-
HD complex can be similarly expressed in eq 2.
Synthesis of [Mn(CO)(depe)2(N2)][BAr′4]. The light yellow
N2 complex can be cleanly generated by stoichiometric reaction
of NaBAr′4 with MnBr(CO)(depe)2 in fluorobenzene under N2
and isolated as for the H2 complex. [Mn(CO)(depe)2(N2)][BAr′4]
will lose N2 readily when solutions or the solid are placed under
vacuum. Precautions similar to those used for the isolation of
the H2 complex must be observed to prevent degradation by
RHD(obs) ) RHD + RHMn + RHligand
(2)
Subtracting eq 2 from eq 1 will allow the effects of metal- and
ligand-based relaxation on the dihydrogen ligand to be canceled
out (eq 3).
RHH(obs) - RHD(obs) ) RHH - RHD
(3)
1
oxygen. H NMR in C6D5F displayed two resonances, δ 8.32
and δ 7.64, typical for the anion and a multiplet δ 1.7-0.9 which
was resolved into four major overlapping components at 500
MHz. 31P{H} NMR in C6D5F yielded only one resonance at δ
72.6 consistent with N2 trans to CO. The ν(CO) at 1896 cm-1
is higher in frequency than that for the agostic complex (1853
cm-1), indicating that the N2 ligand is removing electron density
from the metal by π back-bonding as for H2 and CO ligands. A
strong ν(NN) at 2146 cm-1 is lower in frequency than that for
the related dppe congener (2167 cm-1), which shows that the
depe system possesses greater π back-bonding ability than the
more electrophilic dppe system. These values are 77-96 cm-1
higher than those for the more electron-rich Mo species.4
min
min
Expressions for relating 1/T1 (HH) and 1/T1 (HD) values
to H-H and H-D distances at 300 MHz field strength can be
derived from the relationships in Desrosiers, et al.13b A similar
assumption that the rotational correlation coefficient Tc will be
approximately the same for both metal-H2 and -HD complexes
is made.13b A further assumption was made that the interatomic
distances rHH and rHD in the H2 and HD ligands are the same.
The values for RHH and RHD are calculated from the expressions
given in the Appendix and can be inserted into eq 3 to give eqs
4 and 5.
RHH(obs) - RHD(obs) ) (1.29 × 10-46)/rHH
-
6
The depe complex binds N2 completely at room temperature
under 1 atm of N2, which contrasts sharply to the dppe complex
where binding is only 37% complete.3a Several other cationic
complexes bind N2 weakly or not at all, including [Mn(CO)3-
(PCy3)2]+, where N2 coordination was not seen even at -58
°C.3d Only two 15N NMR resonances (δ -30.0, Mn-NN, and
δ -58.1, Mn-NN) were observed in C6D5F for a sample
enriched with 99.9% 15N2 (Isotech) by cycling a [Mn(CO)-
(depe)2(N2)][BAr′4] solution in C6D5F under vacuum and
backfilling with 15N2 several times. No 15N or 31P coupling was
6
(8.20 × 10-48)/rHD (4)
6
RHH(obs) - RHD(obs) ) (1.21 × 10-46)/rHH
(assuming rHH ≈ rHD) (5)
These relationships are derived for a nonrotating dihydrogen
ligand. For a rapidly rotating dihydrogen ligand, the correction
(13) (a) Bautista; M. T.; Earl, K. A.; Maltby, P. A.; Morris, R. H.;
Schweitzer, C. T.; Sella; A. J. Am. Chem. Soc. 1988, 110, 7031. (b)
Desrosiers, P. J.; Cai, L.; Lin, Z.; Richards, R.; Halpern, J. J. Am.
Chem. Soc. 1991, 113, 4173. (c) Crabtree, R. H. Angew. Chem., Int.
Ed. Engl. 1993, 32, 789.
(14) La Placa, S. J.; Hamilton, W. C.; Ibers, J. A. Inorg. Chem. 1969, 8,
1928.

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1073
observed in the 15N NMR owing to the broadening from
coupling to the 5/2 spin of 55Mn. However, the observation of
only two 15N resonances indicates that N2 is exclusively
terminally bound in solution.
Table 2. Selected Bond Length (Å) and Angle (deg) Data for
[Mn(CO)(LL)2]+
bond (dppe)
LL ) dppe
LL ) depe
bond (depe)
Mn-P(1)
Mn-P(2)
Mn-P(3)
Mn-P(4)
Mn-CO
2.278(2)
2.280(2)
2.321(2)
2.335(2)
1.733(7)
1.173(8)
3.456(4)
3.589(4)
2.89(6)
2.275(3)
2.275(3)
2.265(3)
2.265(3)
1.72(2)
1.17(2)
3.44(1)
3.78(1)
3.42a
Mn-P(2)
Mn-P(2a)
Mn-P(1a)
Mn-P(1)
Mn-CO
Competition binding studies were carried out to examine the
relative binding ability of H2 versus N2. When samples of [Mn-
(CO)(depe)2][BAr′4] and [Mn(CO)(dppe)2][BAr′4] were placed
under a 1:1 mixture of H2 and N2, only the presence of the H2
complex was detected by 31P NMR for both systems, indicating
that H2 binds more strongly than N2. This is consistent with
the observed equilibrium binding properties of H2 and N2 in
the dppe system, where N2 is only 37% bound to Mn as
compared to 100% for H2. The result for the depe system could
possibly be strictly the result of kinetics. This was tested by
placing a sample of [Mn(CO)(depe)2(N2)]+ under an atmosphere
of H2 by evacuating He from the headspace over the solution
of the complex and backfilling with H2. Care was taken to
evacuate the He from the headspace over the sample only after
the C6D5F solvent was frozen to avoid dissociation of N2. On
warming and allowing the solution to stand under H2 for several
minutes, only the resonance for the H2 complex was observed
by 31P NMR. This direct replacement of the N2 ligand by H2
confirms that H2 is bound more strongly than N2 in the depe
system, i.e., a thermodynamic rather than kinetic control. This
result is in accord with thermodynamic data observed for the
Cr(CO)3(PCy3)2(L) system15 for L ) H2, N2 and also the general
observation that H2 is a better ligand than N2 on cationic
complexes.3d
C-O
C-O
Mn‚‚‚C(12)
Mn‚‚‚C(18)
Mn‚‚‚H(12)
Mn‚‚‚H(18)
Mn‚‚‚C(14)
Mn‚‚‚C(19)
Mn‚‚‚H(14)
Mn‚‚‚H(19)
2.98(6)
3.30a
angle (dppe)
LL ) dppe LL ) depe
angle (depe)
P(1)-Mn-P(4)
P(2)-Mn-P(3)
P(1)-Mn-P(3)
P(2)-Mn-P(4)
P(1)-Mn-P(2)
P(3)-Mn-P(4)
P(1)-Mn-CO
P(2)-Mn-CO
P(3)-Mn-CO
P(4)-Mn-CO
Mn-P(3)-C(31)
Mn-P(2)-C(13)
82.67(7)
82.90(7)
95.68(7)
98.02(7)
176.75(8)
166.61(7)
92.3(2)
90.8(2)
96.1(2)
97.2(2)
112.1(2)
115.1(2)
84.14(8)
84.14(8)
95.52(8)
95.52(8)
173.7(2)
174.0(2)
93.2(1)
93.2(1)
93.0(1)
93.0(1)
117.7(5)
113.3(5)
P(1a)-Mn-P(2a)
P(1)-Mn-P(2)
P(1a)-Mn-P(2)
P(1)-Mn-P(2a)
P(1)-Mn-P(1a)
P(2)-Mn-P(2a)
P(1a)-Mn-CO
P(1)-Mn-CO
P(2)-Mn-CO
P(2a)-Mn-CO
Mn-P(2)-C(18)
Mn-P(1)-C(14)
a Calculated from idealized position.
X-ray Structure of [Mn(CO)(depe)2][Ga(C6F5)4]. X-ray
diffraction studies of [Mn(CO)(depe)2]+ with the [Ga(C6F5)4]-
counterion were performed. Crystals were generated by syring-
ing fluorobenzene onto MnBr(CO)(depe)2 and Li[Ga(C6F5)4],
agitating the mixture vigorously, and layering hexane on top
of the resulting deep blue solution. The structure showed that
both the anion and cation lie on a 2-fold rotation axis which
runs through the Mn and both atoms of the carbonyl. The chiral
space group is consistent with the two chiral phosphorus centers.
The two chiral phosphorus atoms are trans, and the enantiomer
would contain the other two phosphorus atoms in analogous
chiral relationship. It is interesting that the enantiomer is resolved
in this structure and not in two dppe analogue structures
Synthesis of Mn(CO)(SO2)(dppe)2][BAr′4]. The reaction of
MnBr(CO)(dppe)2 and Na[BAr′4] in CH2Cl2 at -78 °C in the
presence of excess SO2 followed by warming to room temper-
ature gave an orange solution. The latter readily lost SO2 as
evidenced by the formation of the blue color of the agostic
complex and had to be maintained under an SO2 atmosphere.
An orange solid could be isolated by removing the solvent
completely and placing the residue under an atmosphere of SO2.
The proton NMR was similar to that of the H2 complex, and
31P{1H} NMR showed a singlet at δ 67.63. The solid loses SO2
under a He atmosphere and is partly disassociated in solution
at room temperature, as observed by 31P NMR which showed
a signal for the agostic complex at δ 82.14. The ratio of agostic
to SO2 complex was found to be 1:6. Full coordination can be
achieved by placing the solution under an atmosphere of SO2.
Trace impurities of O2 and SO3 in the commercial SO2 resulted
in partial oxidation of the agostic complex yielding a contami-
nant of the dicarbonyl, Mn(CO)2(dppe)2][BAr′4], as identified
by a 31P resonance at δ 78.35. The identity of this complex
was confirmed by separate generation in an NMR tube by
placing a sample of Mn(CO)(dppe)2][BAr′4] under CO in C6D5F
and observing a 31P{H} signal at δ 78.43. The level of impurity
of dicarbonyl in the material isolated from the bulk reaction
was determined to be 16% by 31P NMR of a sample under a
He atmosphere. Addition of SO2 to the NMR sample increases
the level of the dicarbonyl impurity. The instability of the SO2
complex hindered determination of infrared frequencies and
X-ray diffraction studies. It is conceivable that the SO2 is
coordinated here via oxygen atom(s) as in complexes with strong
main group and metal Lewis acids rather than through η1-S or
η2-S,O geometries.16
containing either the BAr′4 anion3a or the [Cl(Ga(C6F5)3)2]-
-
anion (unpublished work). Both of these structures are racemic
mixtures and have inversion elements between the molecules.
The coordination spheres about Mn for the depe and dppe
complexes are quite similar (Table 2, Figures 1 and 2). Both
systems consist of a distorted square pyramidal structure
containing remote Mn‚‚‚H-C interactions between the phos-
phine organo substituents and the metal center. Unfortunately,
unlike for the dppe congener, the hydrogens were not located
in [Mn(CO)(depe)2]+, and idealized positions for them were
calculated. For the dppe complex, two weak agostic interactions
were located,3a but for the depe analogue, the distances to four
C-H bonds are much longer, especially to the calculated
hydrogen positions (>3.3 Å). This indicates little if any agostic
interaction (see Discussion).
X-ray Structures of [Mn(CO)(depe)2(H2)][Ga(C6F5)4] and
[Mn(CO)(dppe)2(H2)] [Cl(Ga(C6F5)4)2]. X-ray quality crystals
of [Mn(CO)(depe)2(H2)][Ga(C6F5)4] were grown by layering
hexane over a fluorobenzene solution of the complex under H2.
Attempts to prepare crystals of the BAr′4 analogue resulted in
twinned crystals. The cation is an octahedral complex (Figure
3, Table 3) that contains an H2 ligand trans to CO similar to
the structure of the Mo analogue Mo(CO)(depe)2(H2).4a Electron
density for the H2 was found in appropriate locations but could
not be refined and was not included in the final refinement.
(15) Gonzalez, A. A.; Hoff, C. D. Inorg. Chem. 1989, 28, 4295.
(16) (a) Ryan, R. R.; Kubas, G. J.; Moody, D. C.; Eller, P. G. Struct.
Bonding (Berlin) 1981, 46, 47. (b) Kubas, G. J. Acc. Chem. Res. 1994,
27, 183.

1074 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
Figure 3. ORTEP diagram for the cation of [Mn(CO)(depe)2(H2)][Ga-
(C6F5)4] (50% probability ellipsoids).
Figure 1. ORTEP diagram for the cation of [Mn(CO)(depe)2][Ga-
(C6F5)4] (50% probability ellipsoids) and ball-and-stick drawing showing
idealized hydrogen positions with closest M‚‚‚H separations.
Figure 4. ORTEP diagram for the cation of [Mn(CO)(dppe)2(H2)]-
[Cl{Ga(C6F5)3}2] (50% probability ellipsoids).
Table 4. Selected Bond Length (Å) and Angle (deg) Data for
Mn(CO)(dppe)2(H2)][Cl(Ga(C6F5)3)2]
Mn-P(1)
Mn-P(2)
2.308(2)
2.324(2)
Mn-CO
1.839(12)
P(1)-Mn-P(2)
P(1a)-Mn-P(2)
P(1)-Mn-P(1a)
P(1)-Mn-CO
96.50(7)
83.50(7)
180.0
P(1a)-Mn-CO
P(2)-Mn-CO
P(2a)-Mn-CO
84.6(4)
87.7(4)
92.3(4)
95.4(4)
Figure 2. ORTEP diagram for the cation of [Mn(CO)(dppe)2][Ga-
(C6F5)4] (50% probability ellipsoids).
complex was prepared was synthesized from LiGa(C6F5)4 that
contained a chloride impurity as determined by a AgNO3 test.
Although 19F NMR revealed the presence of greater amounts
of fluorocarbon impurities as compared to other syntheses of
LiGa(C6F5)4, the major product was identified to be the anion
[Ga(C6F5)4]-. 13C NMR of this material indeed revealed only
the presence of [Ga(C6F5)4]-. However, the dinuclear chloride-
bridged anion, [Cl(Ga(C6F5)4)2]-, was present in sufficient
quantity to selectively crystallize with the [Mn(CO)(dppe)2-
(H2)]+ cation (Figure 4, Table 4). Electron density for the H2
was not located because of disorder between the H2 and trans-
CO ligands.
Table 3. Selected Bond Length (Å) and Angle (deg) Data for
[Mn(CO)(depe)2(H2)][Ga(C6F5)4]
Mn-P(1)
Mn-P(2)
Mn-P(3)
2.292(1)
2.291(2)
2.271(1)
Mn-P(4)
Mn-CO
C-O
2.269(1)
1.777(4)
1.139(5)
P(1)-Mn-P(2)
P(3)-Mn-P(4)
P(2)-Mn-P(3)
P(1)-Mn-P(4)
P(1)-Mn-P(3)
84.15(4)
85.40(4)
95.69(4)
94.78(4)
179.11(4)
P(2)-Mn-P(4)
P(1)-Mn-CO
P(2)-Mn-CO
P(3)-Mn-CO
P(4)-Mn-CO
178.29(5)
90.69(13)
92.51(14)
88.44(13)
88.83(15)
[Mn(CO)(dppe)2(H2)][BAr′4] gave twinned crystals. In an
attempt to grow crystals of the [Ga(C6F5)4]- salt by layering
hexane over a CH2Cl2 solution under H2, crystals of [Mn(CO)-
(dppe)2(H2)][Cl(Ga(C6F5)4)2] were isolated with a chloride-
bridged Ga anion. The agostic complex from which the H2
X-ray Structures of [Ga(C6F5)4]- and [Cl(Ga(C6F5)3)2]-.
Structural data for the anion of [Mn(CO)(depe)2(H2)][Ga(C6F5)4]
are given in Figure 5 and Table 5 and can be compared to that
for similar Ga compounds.17 The Ga in [Ga(C6F5)4]- is
tetrahedrally coordinated, although distorted from ideal geom-

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1075
Table 6. Selected Bond Length (Å) and Angle (deg) Data for the
[Cl{Ga(C6F5)3}2]- Anion in Mn(CO)(dppe)2(H2)][Cl{Ga(C6F5)3}2]
Ga-Cl(1)
Ga-C(31)
1.925(3)
2.002(8)
Ga-C(37)
Ga-C(43)
1.986(8)
2.003(8)
Cl(1)-Ga-C(31)
Cl(1)-Ga-C(37)
Cl(1)-Ga-C(43)
C(37)-Ga-C(31)
106.2(2)
104.9(3)
104.4(3)
121.6(3)
C(37)-Ga-C(43)
C(31)-Ga-C(43)
Ga-Cl(1)-Ga#2
112.2(3)
106.2(3)
138.9(5)
Table 7. Comparison of Metal-Agostic Distances
complex
M‚‚‚H,a Å
M‚‚‚C, Å
2.884(1)
Cr(CO)3(PCy3)2
2.240(1)
Mo(CO)(dppe)2
2.98(11)
2.20
2.27
2.89(6), 2.98(6)
2.01(9)
3.50(2)
3.007(4)
2.945(6)
Mo(CO)(dBuipe)2
W(CO)3(PCy3)2
[Mn(CO)(dppe)2]+
[Mn(CO)3(PCy3)2]+
[Re(CO)3(PCy3)2]+
[Ru(Ph)(CO)(PBut2Me)2]+
[IrH2(PBut2Ph)2]+
3.456(4), 3.589(4)
2.75(3)b
2.89(5)
Figure 5. ORTEP diagram for [Ga(C6F5)4]- (50% probability el-
lipsoids).
2.87, 2.88
2.81-2.94c
a Idealized M‚‚‚H if no esd’s given. b Average value for disordered
carbons. c For three independent molecules with two agostic interactions
each.
geometry around Ga is distorted from ideal symmetry. The
geometry about the Cl atom reveals that the Ga centers adopt
a bent geometry, which can be rationalized by repulsion of lone
pairs on the Cl. The Ga-C distances in [Cl(Ga(C6F5)3)2]- are
comparable to those in [Ga(C6F5)4]- and shorter than that in
[GaCl(CH3)3]- (2.381 Å)17a as a result of Ga being coordinated
to the electron-deficient Ga(C6F5)3 fragment. Although no other
X-ray diffraction data could be located for compounds of the
form [R3Ga-X-GaR3]-, IR and Raman data suggest a structure
with C2V symmetry for complexes where X is Cl or Br.18
Discussion
Polyagostic Interactions in [M(CO)(diphosphine)2]n+ Com-
plexes. The geometry of the [Mn(CO)(diphosphine)2]+ cations
is formally five-coordinate square pyramidal. The notable feature
is the presence of multiple remote agostic interactions with the
manganese center involving C-H groups located on the phenyl
or ethyl substituents of the phosphines. The dppe complex
Figure 6. ORTEP diagram for [Cl{Ga(C6F5)3}2]- (35% probability
ellipsoids).
Table 5. Selected Bond Length (Å) and Angle (deg) Data for the
[Ga(C6F5)4]- Anion in [Mn(CO)(depe)2(H2)][Ga(C6F5)4]
Ga-C(22)
Ga-C(28)
2.011(4)
2.014(6)
Ga-C(34)
Ga-C(40)
2.015(4)
2.005(4)
C(22)-Ga-C(40)
C(28)-Ga-C(40)
C(34)-Ga-C(40)
101.5(2)
113.1(2)
112.6(2)
C(22)-Ga-C(28)
C(34)-Ga-C(28)
C(22)-Ga-C(34)
115.7(2)
102.2(2)
112.1(2)
etry. This structure is similar to that for the anion in K[Ga-
(CH3)4], which is a distorted tetrahedron possessesing a Ga-C
distance of 2.31(3) Å with two reported C-Ga-C angles of
81(10)° and 125(10)°.17a The shortening of the Ga-C distances
in [Ga(C6F5)4]- is the result of the change from an electron
rich CH3 group to the electron-deficient C6F5 group. The
distances in [Ga(CH3)4]- are much more closely related to those
in the neutral Ga(C6H5)3 (Ga-C ) 1.946(7), 1.968(5)).17b
Structural data for the anion in [Mn(CO)(dppe)2(H2)][Cl(Ga-
(C6F5)3)2] are given in Figure 6 and Table 6. The Ga atoms in
[Cl(Ga(C6F5)3)2]- are tetrahedrally coordinated by three C6F5
groups and a bridging Cl. Metrical data indicate that the
contains two weak interactions, but the depe system appears to
have four C-H bonds from ethyl groups approaching the metal
center, albeit at much longer distances (Table 7) suggestive of
van der Waals contacts with perhaps only a small contribution
from agostic binding. The Mn systems contrast with the neutral
Mo(CO)(diphosphine)2 analogues, which contain only one
interaction from either an ortho phenyl C-H from dppe4b or
i
an isobutyl γ-C-H from Bu2PC2H4PiBu2 (dBuipe).19
(17) (a) Maire, J.; Kruerken, U.; Mirbach, M.; Petz, W.; Siebert, C. Gemelin
Handbook of Inorganic Chemistry Ga, Part 1; Kruerken, U., Mirbach,
M., Petz, W., Siebert, C., Eds.; Springer-Verlag: Berlin, 1987; pp
313-336. (b) Malone, J. F.; McDonald, W. S. J. Chem. Soc. A 1970,
3362.
(18) Wilson, I. L.; Dehnicke, K. J. Organomet. Chem. 1974, 67, 229.
(19) Luo, X. L.; Kubas, G. J.; Burns, C. J.; Butcher, R. J.; Bryan, J. C.
Inorg. Chem. 1995, 34, 6538.

1076 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
long in all cases. Only a maximum total of 2e from the agostic
interactions need to be donated to these 16e fragments, so each
individual interaction represents donation of much less than 2e
from each C-H bond. These complexes thus accommodate
approach of available nearby electron density with the minimum
amount of distortion of their octahedral stereochemistry, even
if the metal can only receive small bits from several sources.
Although polyagostic interactions have been observed in
several systems,22,23 those involving phosphine C-H bonds are
still rare. A 4-fold interaction (two M-P-ortho-phenyl-H and
two M-tert-butyl-H) exists in the formally 14e complexes
M(PPhBut2)2 (M ) Pt, Pd; M‚‚‚H estimated to be 2.5-2.6 Å).22a
A more recent and relevant example is 14e [Ru(Ph)(CO)(PBut2-
Me)2]+ with two agostic Ru‚‚‚H-C interactions originating from
one tert-Bu substituent on each phosphine.22b The preference
for donation from tert-Bu over Me substituents is attributed to
better steric approach of the former (giving a five-membered
ring) than the latter (four-membered). The Ru-C distances are
quite short, 2.87 and 2.88 Å, indicative of strong agostic
interactions23 to the more highly unsaturated 14e fragment (the
polyagostic bonding to Mn involves less-unsaturated 16e
fragments). Related 14e [IrH2(PBut2Ph)2]+ also shows two strong
interactions, but an ortho-metalated analogue gives only one.22e,f
16e [IrH2(PCy2Ph)3]+ gives one agostic interaction, but the less
bulky PPri2Ph analogue shows only very distant M‚‚‚C (e.g.,
3.46 Å) as for the Mn-diphosphine system. Steric factors are
thus clearly of major importance in agostic phosphine systems,
as also shown computationally.22e,f The proximity (or unstrained
approachability) of C-H bonds to the metal is obviously critical
to the strength and/or number of interactions.
The very long, tenuous polyagostic Mn interactions most
likely are a result of the reduced Mn-P-C-C-H ring size as
compared to the Mo system, which prevents full interaction of
the C-H bond (for the dppe system, the average Mn-P is 2.303
Å versus 2.452 Å for the larger Mo coordination sphere). For
both [Mn(CO)(diphosphine)2]+ complexes and also the Mo-dppe
analogue, the M-P-C bond angles in the chelating five-
membered agostic ring were not severely contracted compared
to the analogous angles for the atoms not involved in the agostic
bonding. This is unlike the case for the less constrained
monodentate phosphine system M(CO)3(PCy3)2 (M ) Cr,20 W,21
Re+10a and Mn+10a) where the metal can more strongly attract
a single C-H, and Cr-P-C is 98.8° compared to ca. 120° for
the equivalent angles not involved in the agostic interaction.
Also, the P atom is able to bend away from normal position
here (P-Cr-P ) 160.2°) to facilitate closer approach of the
agostic C-H. The Mn‚‚‚H distances are in fact much longer
Relative Electrophilicity of Cationic Versus Neutral Frag-
ments and Correlation of CO Stretching Frequencies and
Bond Lengths with π-Acceptor Strengths of trans Ligands.
To assess the activation of H2 and other π-acceptor ligands on
a series of related complexes, IR frequencies and metrical data
for the CO ligand trans to the ligand of interest are correlated
in Table 8. The variations include metal, ligand, and charge on
the complex. Clearly the Mn(I) cations have the highest ν(CO),
about 100 cm-1 higher than the corresponding Mo(0) species,
demonstrating the much higher electrophilicity of cations and
the consequent reduced back-bonding to CO that raises ν(CO).
The recently reported dicationic complex [Fe(H2)(CO)(dppe)2]2+
(2.89(6) and 2.98(6) Å for dppe; 3.30 and 3.42 Å calculated
for each Mn‚‚‚Hâ and each Mn‚‚‚HR for depe) than that found
for the single CH interaction in [Mn(CO)3(PCy3)2]+ (2.01(9)
Å). Table 7 compares these distances for a variety of agostic
systems. The Mo‚‚‚H distance is also long, 2.98(11) Å, in Mo-
(CO)(dppe)2,4b although it must be kept in mind that Mo is a
larger second-row metal. On the other hand, for Mo(CO)(dBui-
pe)2, the calculated Mo‚‚‚H distance is much shorter, 2.20 Å,
and the Mo-P-C angle in the chelate ring (six-membered here)
is somewhat contracted (113.0° relative to the other Mo-P-C
angles of the same type (122.8° avg)). It is not clear why the
interaction is significantly stronger here than in the other Mo
and Mn diphosphine systems, although the six-membered ring
may bring the C-H bond closer to the metal with less overall
bond strain than for a five-membered ring. The agostic hydrogen
lies more directly trans to the CO (H‚‚‚Mo-C ) 166.0°) than
for Mo(CO)(dppe)2 or the Mn species, where the H‚‚‚Mo-C
angles are much more acute. This angle is even closer to linear
(173.1°) in Cr(CO)3(PCy3)2.
has a far higher ν(CO), 2006 cm-1 5a
, than any complex in Table
8. This value is 110 cm-1 higher yet than that for the
monocationic Mn+ congener, emphasizing that the charge on
the complex is the dominant factor here, much more than the
basicity of the cis-phosphine coligands. In the H2 and N2
adducts, ν(CO) for the depe and dBuipe systems are only ca.
10-30 cm-1 lower than for the dppe species. This is less than
may have been expected for substituting four trialkylphosphine
ligands by four diarylalkylphosphine donors, but it is consistent
with the effect in MnBr(CO)(diphosphine)2. The difference is
even less in the cationic Mn species than in the Mo complexes
(15 vs 32 cm-1 for N2 adducts), reflecting the leveling effect
The M‚‚‚C distances also reflect the relative strengths of the
agostic interactions. For Mo(CO)(dBuipe)2 it is 3.007(4) Å,
much shorter than that in the dppe congener (3.50(2) Å) and
those in the Mn cations but significantly longer than in the
majority of cases. For [Mn(CO)(dppe)2]+ the distances are
3.456(4) and 3.589(4) Å, and for the depe analogue they are
3.44(1) for each M‚‚‚CR and 3.78(1) Å for each M‚‚‚Câ, very
(22) (a) Otsuka, S.; Yoshida, T.; Matsumoto, M.; Nakatsu, K. J. Am. Chem.
Soc. 1976, 98, 588. (b) Huang, D.; Streib, W. E.; Eisenstein, O.;
Caulton, K. G. Angew. Chem., Int. Ed. Engl. 1997, 36, 2004. (c) Cole,
J. M.; Gibson, V. C.; Howard, J. A. K.; McIntyre, G. J.; Walker, G.
L. P. Chem. Commun. 1998, 1829. (d) Evans, W. J.; Anwander, R.;
Ziller, J. W.; Khan, S. I. Inorg. Chem. 1995, 34, 5927. (e) Cooper, A.
C.; Clot, E.; Huffman, J. C.; Streib, W. E.; Maseras, F.; Eisenstein,
O.; Caulton, K. G. J. Am. Chem. Soc. 1999, 121, 97. (f) Ujaque, G.;
Cooper, A. C.; Maseras, F.; Eisenstein, O.; Caulton, K. G. J. Am.
Chem. Soc. 1998, 120, 361.
(20) Zhang, K.; Gonzalez, A. A.; Mukerjee, S. L.; Chou, S.-J.; Hoff, C.
D.; Kubat-Martin, K. A.; Barnhart, D.; Kubas, G. J. J. Am. Chem.
Soc. 1991, 113, 9170.
(21) Wasserman, H. J.; Kubas, G. J.; Ryan, R. R. J. Am. Chem. Soc. 1986,
108, 2294.
(23) Braga, D.; Grepioni, F.; Biradha, K.; Desiraju, G. R. J. Chem. Soc.,
Dalton Trans. 1996, 3925.

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1077
Table 8. Comparison of ν(CO) and CO Bond Lengths in Adducts of d6 Fragments with Ligands of Varying π-Acceptor Strength Trans to CO
metal fragment
ligand
ν(CO), cm-1
C-O, Å
M-C, Å
ref
[Mn(CO)(dppe)2]+
agostic
H2
N2
“agostic”
H2
N2
1839
1896
1911
1853
1887
1896
1888
1709
1723
1814
1809
1724
1777
1725
1797
1843
1835
1834
1865
1888
1870
1889
1971
1961
1.173(8)
1.16(2)
1.733(7)
1.839(12)
3a
3a
3a
this work
this work
this work
this work
43
4
4a
4b
19
4a
46
21
42b, 47
21
30c
48
49
[Mn(CO)(depe)2]+
Mo(CO)(dppe)2
1.17(2)
1.139(5)
1.72(2)
1.777(4)
CO
NH3
agostic
H2
1.192(12)
1.179(6)
1.127(20)
1.190(5)
1.903(9)
1.933(5)
1.973(16)
1.886(4)
N2
Mo(CO)(dBuipe)2
W(CO)3(PCy3)2
agostic
N2
H2Oa
agostic
H2
1.21(1)
1.180(7)
1.17(1)
1.875(8)
1.902(6)
1.99(1)
N2
C2H4
CO
1.159(5)
1.13(1)b
1.097(19)
1.094(13)
1.09(1)
1.977(4)
2.05(1)b
2.055(17)
1.958(11)
2.04(1)
2.006e
mpz+c
Id
W(CO)3(PPri3)2
Mo(CO)3(PPri3)2
W(CO)5
50
51
2b, 52
2b, 53
2b
SO2
H2
N2
2.013e
NO+
2.178e
a Water protons are hydrogen bonded to the axial CO ligand and also a lattice THF. b Average value. c Ligand is methylpyrazinium cation;
complex is [W(mpz)(CO)3(PCy3)2][PF6]. d Complex is a 17-e W(I) radical; W-C distance may not correlate with those for the W(0) species.
e Calculated.
of the positive charge. The agostic species show a reversal of
this trend for Mn and little difference in ν(CO) for Mo,
indicating that the strength and/or number of agostic interactions
may be a factor (the Mn-depe complex has virtually no agostic
contribution).
σ donors such as NH3. In contrast, the H2 ligand is an excellent
acceptor as strong as N2 and ethylene, and all of the complexes
with these ligands have similar ν(CO) values much higher than
those for the corresponding agostic complexes. As can be seen
from Table 8, the C-O bond length always decreases and the
M-C bond length increases substantially when acceptor ligands
replace the trans agostic interactions. This effect, like the
increased ν(CO), is attributable primarily to reduced M f CO
back-bonding because of increased competition for dπ-electron
density. Theoretical calculations of the bond lengths in M(CO)5L
(M ) Cr, Mo, W) also led to a similar correlation that strong
acceptors L increase the M-COtrans length whereas poor
acceptors decrease it.2b The M(CO)3(PR3)2(L) complexes show
the greatest range of trans-ligand effects. The longest C-O
distance, 1.21(1) Å, and shortest W-C length is for the pure
σ-donating H2O adduct, although there is intermolecular hy-
drogen bonding between CO and H2O that might weaken the
C-O bond. The ν(CO) value is 1725 cm-1, the lowest by far
in the W(L)(CO)3(PCy3)2 series. At the other end of the spectrum
are complexes with powerful π acceptors such as CO, SO2, and
mpz+ (methylpyrazinium), which have C-O in the 1.09-1.13
Å range and ν(CO) at 1865-1889 cm-1. Also in this group is
WI(CO)3(PPri3)2, which is a WI radical with a less electron-
rich 17e metal center. The H2, N2, and ethylene ligands give
intermediate parameters, indicating they are not quite as strongly
π-accepting as CO and SO2.
The recently reported6 H2 complexes with phosphite donors,
[Mn(H2)(CO)P4][BPh4] (P ) P(OEt)3, PPh(OEt)2) have higher
ν(CO) values (1926 cm -1) than any of the complexes in Table
8, indicating that they are significantly more electron poor than
the corresponding bidentate phosphine complexes where ν(CO)
is 30 cm-1 lower. The complexes resulting from dissociation
of H2 from [Mn(H2)(CO)P4][BPh4] are claimed to have agostic
interactions on the basis of 31P NMR evidence. However, unlike
our phosphine complexes, there is a glaring disparity in the
reported ν(CO) values in the solid state for the P(OEt)3 (1872
cm-1) and PPh(OEt)2 (1915 cm-1) species. Furthermore, the
small difference between the latter value and that for the
corresponding H2 complex (1926 cm-1) is far less than those
for other systems with agostic versus H2 ligands. Finally, the
yellow color of the isolated products of H2 removal conflicts
with the characteristic intense blue color of the agostic Mn and
Mo diphosphine systems. Because the BPh4 anion is more likely
to interact with the metal than BAr′4, and a coordinating solvent
(ethanol) was used in isolating the solids, it is conceivable that
the solid phosphite complexes have anion or solvent interactions
rather than agostic interactions. The three 31P NMR signals
reported for {Mn(H2)(CO)[P(OEt3)4}[BPh4] in CH2Cl2 solution
at 183 K might result from a mixture of agostic and solvento
forms (CH2Cl2 coordinates3c to Re(CO)4(PR3)+). It should be
noted that for [Mn(CO)(dppe)2][BAr′4], only one 31P signal was
observed down to 198 K in CD2Cl2 because of the high
fluxionality of the agostic interactions.3a
Inelastic Neutron Scattering Studies of [Mn(CO)-
(diphosphine)2(H2)]+. Considerable insight into metal-H2
interactions has been obtained by studying the barrier to rotation
of coordinated H2 by inelastic neutron scattering (INS) tech-
niques.24 This barrier arises primarily from the variation in the
back-bonding interaction between dπ(M) and the σ*(H2) orbitals
as the H2 ligand is rotated around the σ bond to the metal.
The IR data and also C-O and M-C bond lengths can also
be used to gauge the π-acceptor strengths of ligands trans to
the CO because of the well-known trans effect. The agostic
C-H bonds are the weakest ligands and also the weakest
acceptors, and the influence on ν(CO) is similar to that for pure
(24) (a) Eckert, J.; Kubas, G. J. J. Phys. Chem. 1993, 97, 2378. (b) Taylor,
A. D.; Wood, E. J.; Goldstone, J. A.; Eckert, J. Nucl. Instrum. Methods
Phys. Res. 1984, 221, 408. (c) Eckert, J. Spectrochim. Acta 1992, 48A,
363. (d) Eckert, J. Physica 1986, 136B, 150. (e) Albinati, A.; Klooster,

1078 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
Table 10. J(HD) Coupling Constants and H-H Distances for H2
Table 9. Tunneling Frequency, ωt (cm -1), Observed and
Calculated Rotational Transitions, τ (cm-1), Barrier Heights, Vn
(kcal/mol), and δ(Mn(H2)), i.e., Wag or Rock, for Dihydrogen in
[Mn(CO)(η2-H2)L2]+, L ) dppe, depea
Complexes of Various Fragments with Trans-CO Ligands
metal fragment
Cr(CO)3(PPri3)2
Cr(CO)3(PCy3)2
Mo(CO)3(PCy3)2
Mo(CO)(dppe)2
J(HD), Hz
H-H, Åa
ref
35
0.84-0.85
40
40
41
0.85(1)b
L
0.87b
dppe, B ) 41
obs calcd
depe, B ) 43
34
0.88(1)b
4a, 41
0.85-0.87
0.92-0.94
0.89(1)b
obs
calcd
Mo(CO)(dBzpe)2
30
55
W(CO)3(PPri3)2
33.5
41,42
c
V2 ) 1.22
V4 ) -0.11
V2 ) 1.17
V4 ) -0.15
0.86-0.88
W(CO)3(PCy3)2c
0.89(1)b
41
3a
ωt
τ
4(1)
188, 218
332
3.3
186, 223
329
4.2
[Mn(CO)(dppe)2]+
32
0.89(2)b
178, 212
315
178, 219
322
0.89-0.90
0.87-0.89
0.89-0.90
0.88-0.89
0.87-0.89
0.87-0.89
0.85-0.87
0.85-0.87
0.89-0.90
0.87-0.89
0.89-0.90
0.92-0.94
0.90-0.92
0.87-0.89
0.87-0.89
0.86-0.88
0.87-0.88
0.85-0.87
0.89-0.90
[Mn(CO)(depe)2]+
33
this work
6
[Mn(CO)(P(OEt)3)4]+
[Mn(CO)(PPh(OEt)2)4]+
Mn(CO)2(P(OEt)3)3]+
[Mn(CO)3(PCy3)2]+
[Re(CO)4(PPh3)]+
32
δ(Mn(H2))
δ(Mn(H2))
422
580
390
535
32.5
33
33
6
6
a Uncertainties in the frequencies are approximately 2% of their
value.
3d
33.9
33.8
32
3c
[Re(CO)4(PCy3)]+
3c
[Re(CO)3(PCy3)2]+
[Re(CO)3(PPri3)2]+
[Re(CO)3(PPh3)2]+
[Re(CO)3{P(OEt)3}2]+
[Re(CO)2(PMe2Ph)3]+c
[Re(CO)2{P(OEt)3}3]+
[Re(CO){P(OEt)3}4]+c
[ReH2(CO)(PMe3)3]+c
[Fe(CO)(dppe)2]2+
10c
10c
10c
54
Therefore, the barrier height, derived from measurements of the
rotational transitions, provides an indication of the relative
strength of the back-bonding interaction. INS measurements of
the rotational and vibrational transitions of the H2 ligand for
both the dppe and depe congeners were performed to determine
the rotational barriers and relate them to other data concerning
the nature of the M-H2 bonding in these compounds.
33
32
30
31
44
33
54
33
54
33.6
33.1
34
45
5a
[Ru(CO)(dppp)2]2+
[Os(CO)(dppp)2]2+
5b
The rotational tunneling lines for [Mn(CO)(H2)(dppe)2]+ were
extremely broad and poorly defined in the same manner as was
recently reported24e for [RuH(H2)(dppe)2]+ but unlike those
observed for most other H2 complexes that do not have the fairly
symmetric diphosphine ligands. The reason for this observation
must lie in the fact that the Mn center is nearly coplanar with
the four P atoms for both of the present complexes, with the
result that the rotational potential wells have very flat bottoms.
At low temperatures, this would give rise to an inhomogeneous
broadening of the rotational tunneling lines because the equi-
librium position for this potential is not well defined in the
presence of the large zero-point vibrational amplitude of the
H2 ligand. The observed rotational tunneling band position, as
well as the torsional and some vibrational frequencies, are
presented in Table 9.
Barrier heights were computed in the same manner previously
reported24 with values of the rotational constant B (calculated
here from the NMR-derived H-H distances) held fixed at 41
and 43 cm-1 for the dppe and depe complexes, respectively.
The barrier heights derived in this fashion, along with the
calculated values of the rotational transitions, are also given in
Table 9. The results for the rotational transitions and associated
barrier heights for both compounds are found to be rather
similar, in accord with the fact that the degree of H-H bond
activation as measured by d(HH) is similar. Differences in the
frequencies of the rocking and wagging modes are more
pronounced. This may in part be related to differences in mixing
of these vibrations with skeletal motions that would be expected
to differ for the dppe and depe ligands because of the
substantially different masses of phenyl and ethyl groups.
It should be noted that the rotational barriers for the dppe
and depe complexes are nearly equal, ∼1.2 kcal/mol, which
indicates that the influence of the cis ligands (e.g., raising or
lowering back-donation) is relatively unimportant here, as
discussed below. The barriers are less than that for the related
compound [FeH(H2)(dppe)2]+,24f but greater than that in Mo-
32
5b
a Except where noted, calculated from and bracketed by the empirical
relationships, rHH ) 1.42-0.0167J(HD) (ref 12a) and rHH ) 1.44-
0.0168J(HD) (ref 12b). b Measured by solid-state NMR (ref 41).Neutron
diffraction distances for Mo(CO)(dppe)2 are about 0.07 Å lower because
H2 rotation foreshortens the H-H bond (ref 4). c In equilibrium with
dihydride tautomer in solution.
(CO)(H2)(dppe)2 (0.5-0.6 kcal/mol).4a In addition to the degree
of back-bonding present, the barrier is strongly influenced by
distortions of the MP4 skeleton in diphosphine complexes, as
has been demonstrated by calculations by Eisenstein.4a The
barrier would essentially be zero if the ligands cis to H2 were
perfectly 4-fold symmetric. However, as distortions of the
phosphines away from octahedral positions increase, so do the
barriers to H2 rotation, which depend on differences in back-
bonding betweeen H2 orientations. From electronic arguments
alone, the decreased back-donation in the Mn cation should have
given a lower barrier than that for the Mo complex. However,
the Mn complexes are also structurally less distorted than the
Mo analogues, so it appears that the higher barriers observed
for the cationic complexes could be a result of other factors,
perhaps influence by the anion. Rotational barriers have indeed
been shown to be very sensitive to the environment near the
H2, including even variations in lattice solvent.4a The similarity
of the barrier for the Mn-dppe and -depe complexes thus could
be a result of steric influences rather than electronic consider-
ations, but it is nonetheless consistent with the low influence
of the cis ligands on the properties of H2 binding here.
Assessment of the Activation of Dihydrogen in [Mn(CO)-
(diphosphine)2(H2)]+ and Related Complexes: Trans Ligand
Effects Dominate. Molecular H2 coordination chemistry affords
a grand opportunity to study how metals activate and oxidatively
add H2 as a function of metal-ligand sets and overall charge
of the complex. Conversely, H2 coordination can be used to
probe site-specific electronics on a wide variety of metal
fragments. When the H-H bond lengths and J(HD) NMR
couplings for series of neutral, cationic, and dicationic group
6-8 H2 complexes with trans-CO ligands are compared (Table
10), a surprising consistency is observed despite the large
variation in electrophilicity of the metal. Because M-H2
W. T.; Koetzle, T. F.; Fortin, J. B.; Ricci, J. S.; Eckert, J.; Fong, T.
P.; Lough, A. J.; Morris, R. H.; Golombek, A. P. Inorg. Chim. Acta
1997, 259, 351. (f) Eckert, J.; Blank, H.; Bautista, M. T.; Morris, R.
H. Inorg. Chem. 1990, 29, 747.

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1079
bonding is governed by both the Lewis acid strength of the metal
for σ bonding (donation from H2) and the π back-bonding ability
of the metal (H2 as an acceptor), the observed H-H bond length
results from a balance of these two bonding components. Thus,
the activation of coordinated H2 in the electropositive systems
is occurring primarily via increased σ donation from H2 as
compared to the electron-rich neutral systems where back-
donation dominates. As stated in the Introduction, theoretical
studies1,2 indicate that EBD is about twice as strong as ED in
Mo(CO)(PH3)4(H2), but this is reversed for electron-poor Mo-
(CO)5(H2).1b The nature of the metal is of minor importance
compared to the effects of changing the ligand set or charge.
Importantly, the sum of the bonding energies of the two
components remains nearly constant in a particular system such
as M(CO)n(PH3)5-n(H2), and for highly electrophilic metal
centers, the loss in back-bonding is almost completely offset by
increased σ donation from H2 to the electron-poor metal center.
The extremely electron-poor dication [Fe(CO)(dppe)2(H2)]2+
actually binds H2 more strongly than the neutral complexes that
are stabilized by back-bonding (unlike the Mn+ and Mo0
analogues, the Fe2+ complex is stable in Vacuo5a). Although
theoretical calculations for such cationic H2 systems have not
yet been reported, preliminary ab initio results for [Mn(CO)-
(diphosphine)2(H2)]+ versus the neutral Mo analogue indicate
this to be true.3e Computations25 for CO coordination in
[M(CO)6]n (M ) group 4-9 metal; n ) -2 to +3) show that
σ donation from CO increases as positive charge increases, just
as for H2. The first CO dissociation energy is in fact higher for
the cationic [M(CO)6]n complexes, which clearly would not have
been expected on the basis of back-bonding arguments alone.25b
no change when the H2 ligand in [Mn(CO)(depe)2(H2)]+ is
replaced by CO (1887 versus 1888 cm-1).
Significantly, H2 binds molecularly to Mo(CO)(dppe)2 but
undergoes oxidative addition on more electron-rich Mo(CO)-
(depe)2 (and also on W(CO)(dppe)228 because of the increased
back-bonding ability of W over Mo).4a However, for the cationic
Mn species, H2 binds molecularly to both the dppe and depe
congeners. The positive charge disfavors H-H bond rupture
The consistency in H-H activation on both the cationic and
neutral fragments is nonetheless astonishing, especially in terms
of J(HD) values which occur in the narrow range of 32-34 Hz
for 19 of the 23 complexes listed in Table 10. This suggests a
limit to the activation of the H-H bond by complexes that bind
H2 principally through the σ interaction. This should be expected
since the 3-center 2e interaction of a vacant d orbital with the
filled H2 σ orbital limits depletion of the electron density
between the two hydrogen atoms as compared to population of
the H2 σ* orbital, which ultimately can lead to total rupture of
the H-H bond. Most of the H-H bond lengths for the cationic
systems lie in the relatively narrow range of 0.86-0.90 Å,
compared to the neutral systems. Surprisingly, J(HD) is slightly
lower, and the H-H bond distance is slightly longer in [Mn-
(H2)(CO)(dppe)2]+ than in the depe analogue. This same
counterintuitive pattern has also been seen in the group 8 [RuH-
(H2)(PP)2]+ system where J(HD) is 32 Hz when PP is either
dppe or depe, despite electrochemical properties and the pKa
of η2-H2 showing increased electron richness at the metal for
depe.29 On the other hand, in the series [Re(H2)(CO)n-
{P(OEt)3}5-n]+, J(HD) decreases from 33 to 30 Hz as n
increases from one to three (increasing electrophilicity), which
is an even more emphatic reversal of the expected trend. A few
other cationic systems show little variation in H-H activation
with changes in relative donor/acceptor ability of the cis-ligand
sets, and it has been previously noted that “unstretched” H2
complexes (H-H < 0.9 Å) do not show very large changes in
J(HD) on ligand variation.13c It is thus now quite clear that the
nature of the cis ligands has little systematic effect in these
cations, especially when compared to neutral complexes at the
brink of oxidative addition such as Mo(CO)(H2)(PP)2. The H-H
bond in the latter system is so close to cleaving (the lowest
J(HD) observed for any complex with trans CO is 30 Hz for
PP ) dBzpe) that cis-ligand variation does give a predictable,
meaningful trend.
similar to the calculated distance in triangulo H3+, 0.87 Å,26
a
good model for H2 activation strictly through σ interactions only.
Stable d0 M-H2 complexes have not been isolated because
back-donation cannot take place here. However, cationic H2
complexes still must retain some degree of back-bonding
because they are isolatable and the CO stretching frequencies
for [Mn(CO)(diphosphine)2(H2)]+ are significantly higher than
those for the corresponding agostic complexes. The H-H bond
is a much stronger π acceptor compared to the C-H bond in
agostic interactions and metal-alkane complexes because orbital
mismatch for M(η2-CH) greatly diminishes back-donation.27
According to Ziegler’s calculations,1b a moderate amount of
back-bonding to H2 is present even in Mo(CO)5(H2), indicating
that H2 competes well with CO for back-bonding. Experimental
evidence for the latter is the fact that ν(CO) undergoes almost
(25) (a) Ehlers, A. W.; Ruiz-Morales, Y.; Baerends, E. J.; Ziegler, T. Inorg.
Chem. 1997, 36, 5031. (b) Szilagyi, R. K.; Frenking, G. Organome-
tallics 1997, 16, 4807.
(26) (a) Pang T. Chem. Phys. Lett. 1994, 228, 555. (b) Farizon, M.; Farizon-
Mazuy, B.; de Castro Faria, N. V.; Chermette, H. Chem. Phys. Lett.
1991, 177, 451.
(28) Ishida, T.; Mizobe, Y.; Hidai, M. Chem. Lett. 1989, 11, 2077.
(29) (a) Bautista, M. T.; Cappellani, E. P.; Drouin, S. D.; Morris, R. H.;
Schweitzer, C. T.; Sella, A.; Zubkowski, J. J. Am. Chem. Soc. 1991,
113, 4876. (b) Cappellani, E. P.; Drouin, S. D.; Jia, G.; Maltby, P.
A.; Morris, R. H.; Schweitzer, C. T. J. Am. Chem. Soc. 1994, 116,
3375.
(27) Saillard, J.-Y.; Hoffmann, R. J. Am. Chem. Soc. 1984, 106, 2006.

1080 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
Despite the tradeoff in σ donation and π back-donation in
M-H2 bonding, some structure-bonding aspects of H2 coordina-
tion remain unaccountable. The theoretical model for Mo(CO)-
(dppe)2(H2) features high back-bonding, yet the actual complex
has one of the highest J(HD) values and shortest H-H distances
known. This is even more puzzling when compared to the
cationic Mn congeners that actually have somewhat lower
J(HD), longer H-H distances, and no proclivity to cleave the
H2 on increasing the donor strength of the cis-phosphines. Other
types of H2 complexes with different ligand sets, including
cationic species such as [OsH(H2)(PP)2]+,29 show much lower
J(HD) in the 11-26 Hz range. The obvious explanation is that
the nature of the ligand trans to H2, which is the strong
π-acceptor CO for all complexes in Tables 8 and 10, has a
powerful leveling effect that may be underestimated in theoreti-
cal analyses. The trans effect is of critical importance2d,30 in H2
binding as in all of coordination chemistry, and indeed a
comprehensive survey (Table 10) shows that J(HD) < 30 Hz
is unknown for η2-HD trans to CO. As a corollary, H-H
distances rarely are obserVed to be much longer than 0.9 Å in
complexes with CO trans to H2, regardless of ligand set or
oVerall charge. However, complexes with mild σ-donor ligands
trans to H2 or π-donors such as chloride have greatly elongated
H-H bonds (0.96-1.34 Å) and J(HD) from 9 to 28 Hz because
of increased back-bonding to H2. A good comparison is between
the second-row group 6 and 7 congeners, Mo(CO)(H2)(dppe)2
(J(HD) ) 34 Hz; H-H ) 0.88 Å) and TcCl(H2)(dppe)2 where
H-H elongates 0.2 Å to 1.08 Å.31 [RuCl(H2)(dppe)2]+ has a
J(HD) of 26 Hz, much lower than that for congeners with trans
CO or hydride,32 and analogues with alkyl diphosphines have
an even lower value of 16 Hz, showing strong, unmitigated H2
activation. If the trans ligand is a strong σ-donor such as hydride,
there is a potent trans labilizing effect that reduces σ donation
from H2, which once again weakens M-H2 binding strength
and contracts the H-H distance (J(HD) generally >28 Hz).
Certain electron-rich neutral complexes with hydride trans to
Comparative Binding of Other Ligands to [Mn(CO)-
(dppe)2]+. In regard to silane coordination on group 6 systems,33
[Mn(CO)(dppe)2][BAr′4] binds SiH4 much more weakly than
Mo(CO)(depe)2, which gives a tautomeric equilibrium between
the σ complex and the oxidative addition product.33a A Mn-
SiH4 complex could not even be observed, because the H2
always present as an impurity in SiH4 coordinated more strongly.
Organosilanes also did not give stable complexes, even at -70
°C. Coordination of silanes to first row group 6 and 7 metals
thus appears to be much weaker than to second and third row
metals as PhSiH3 did not bind at all to M(CO)3(PR3)2 for Cr
and Mn+ but underwent oxidative addition for W.3d,33c This may
be a result of increased steric congestion at the smaller first
row metal when bulky phosphines are present. In general, silane
binding and activation is much more variable with respect to
the metal fragment than that for H2, which coordinates in a much
more steady fashion throughout first to third row metals and,
for the most part, oxidatively adds more predictably.33c
Poor bases such as Et2O and CH2Cl2 do not bind to the Mn-
diphos complexes, indicating that they are not as electrophilic
as [Re(CO)4(PCy3)]+ and [PtH(PPri3)2]+, which do.3b,c Surpris-
ingly, the normally strong amphoteric ligand SO2, which
irreversibly binds to the neutral Mo analogues and most
transition-metal fragments,14a coordinates weakly and reversibly
to [Mn(CO)(dppe)2]+, as found for [Mn(CO)3(PCy3)2]+.3d As
recently concluded, there is a clear trend that electrophilic
cationic systems favor H2 binding over N2 and even stronger π
acceptors such as SO2.3d Although H2 is generally considered
to be a “weak” ligand, the highly amphoteric nature of H2
bonding to transition metals makes H2 much more versatile than
N2, olefins, silanes, and virtually any other ligand. Dihydrogen
is thus able to stably coordinate or oxidatively add to a wider
array of transition-metal fragments (except d0 systems, which
cannot back-donate), particularly cationic species, than even
most “strong” ligands.
30b
H2 such as trans-IrCl2H(H2)(PR3)2 actually bind H2 more
weakly than many of the highly electrophilic cationic systems.
Conclusions
The electrophilic 16e [Mn(CO)(PP)2]+ complexes (PP )
diphosphine) are stabilized by low-coordinating boron or gallium
anions and contain weak polyagostic interactions reversibly
displaceable by H2 and N2. The large new gallium anion [(Ga-
(C6F5)3)2(µ-Cl)]- is reported, which along with [Ga(C6F5)4]-
is a useful counterion for neutron scattering studies. The Mn‚
‚‚H-C distances are much longer than those found for the single
interactions in Mo(CO)(PP)2 and [Mn(CO)3(PCy3)2]+ and are
more characteristic of van der Waals contacts, especially for
the depe complex. The H-H and also the Mn-H distances have
been determined in [Mn(H2)(CO)(dppe)2]+ in solution by NMR
T1 measurements on both the H2 and HD isotopomers. This
method is potentially useful for other H2 complexes, especially
for metal centers with high nuclear spin. IR data and C-O and
M-C bond lengths are used to gauge the π-acceptor strengths
of ligands trans to the CO. The agostic C-H bonds are the
weakest “ligands” and also the weakest acceptors, with ν(CO)
similar to that for pure σ donors such as NH3. In contrast, H2,
N2, and ethylene are excellent acceptors. The C-O bond length
always decreases, and the M-C bond length increases when
these acceptor ligands replace the agostic interactions. The
variation of ν(CO) on increasing the basicity of the cis-
For cis-IrCl2H(H2)(PR3)2, J(HD) decreases dramatically to 12
Hz and ab initio calculations (R ) H) show a spectacular
increase in H-H distance from 0.81 to 1.4 Å (1.11 Å
experimental for R ) i-Pr) on going from the trans to the cis
isomer.30b Although the hydride cis to H2 in the latter has some
influence (cis effect), the influence of the trans ligand on H2
actiVation generally far exceeds that of the entire cis-ligand
set. This is true particularly in cations with trans CO where
back-bonding effects are lower. This enormous dependence on
fragment stereochemistry can be expected to extend to other σ
complexes, including alkane complexes where back-donation
to C-H is much lower.
(30) (a) Schlaf, M.; Lough, A. J.; Maltby, P. A.; Morris, R. H. Organo-
metallics 1996, 15, 2270, and references therein. (b) Albinati, A.;
Bakhmutov, V. I.; Caulton, K. G.; Clot, E.; Eckert, J.; Eisenstein, O.;
Gusev, D. G.; Grushin, V. V.; Hauger, B. E.; Klooster, W. T.; Koetzle,
T. F.; McMullan, R. K.; O’Loughlin, T. J.; Pelissier, M.; Ricci, J. S.;
Sigalas, M. P.; Vymenits, A. B. J. Am. Chem. Soc. 1993, 115, 7300.
(31) Burrell, A. K.; Bryan, J. C.; Kubas, G. J. J. Am. Chem. Soc. 1994,
116, 1575.
(33) (a) Luo, X.-L.; Kubas, G. J.; Burns, C. J.; Bryan, J. C.; Unkefer, C.
J. J. Am. Chem. Soc. 1995, 117, 1159. (b) Luo, X.-L.; Kubas, G. J.;
Bryan, J. C.; Burns, C. J.; Unkefer, C. J. J. Am. Chem. Soc. 1994,
116, 10312. (c) Butts, M. D.; Kubas, G. J.; Luo, X.-L.; Bryan, J. C.
Inorg. Chem. 1997, 36, 3341.
(32) Chin, B.; Lough, A. J.; Morris, R. H.; Schweitzer, C.; D′Agostino, C.
Inorg. Chem. 1994, 33, 6278.

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1081
added slowly to the LiC6F5 over 30 min. The reaction was allowed to
slowly warm over a period of 16 h to room temperature. The hexane
was filtered off the solids, which were then extracted with 250 mL of
toluene. The extraction was repeated with 200 mL of toluene, all of
which was then removed in vacuo. The solids were taken up in 75 mL
of CH2Cl2 and then precipitated with 120 mL of hexane. The volume
was reduced 75%, and then 90 mL of hexane was added to complete
the precipitation. The solution was filtered off, and the product was
dried in vacuo. Yield: 6.033 g. A second extraction was performed on
the reaction residues yielding 4.500 g of product. Overall yield: 44%.
The reaction product gave a positive Li flame test and a negative AgCl
test. Although the product did not give consistent elemental analysis,
19F and 13C NMR spectroscopy in THF-d8 of the anion was consistent
with that previously reported for [Bu4N][Ga(C6F5)4].8 The product was
found by 19F NMR to contain minor amounts of impurities that did
not affect its further use. 13C (THF-d8): δ 150.0 (JC-F ) 237.8 Hz,
ortho); 140.8 (JCF ) 249.2 Hz, meta); 137.3 (JCF ) 236.5, para); 121.5.
19F (THF-d8): -123.1 (d, 8F, ortho), -160.5 (t, 4F, para), -165.4 (t,
8F, meta).
Synthesis of Li[Ga(C6F5)4]‚2Et2O. C6F5Br was dried over P2O5 and
vacuum transferred prior to use. Under an Ar atmosphere, C6F5Br was
dissolved in 150 mL of Et20 and cooled to -78C. The reaction mixture
was then charged with 4.2 mL of 1.6 M n-BuLi and allowed to stir for
10 min. Ga(C6F5)3‚Et20 in 100 mL of Et2O was added to the solution,
and then the reaction was allowed to warm to room temperature and
stirred over 20 h. The volume of the solution was reduced in vacuo
75%. The solution was filtered and then cooled to -78 °C. The colorless
precipitate was filtered off and stirred with hot hexane, forming an oil
that solidified on cooling. The hexane was decanted off the solids, which
were dried in vacuo. The product gave a positive flame test for Li.
Anal. Calcd for C32H20F20GaLi4O2: Ga, 7.83%. Found: Ga, 7.77%.
1H NMR (THF-d8): 3.36 (q, 4H, CH2); 1.10 (t, 6H, CH3). 13C (THF-
d8): 149.8 (JCF ) 232.2 Hz, ortho); 140.8 (JCF ) 249.8 Hz, meta);
137.2 (JCF ) 271.5 Hz, para); 121.5. 19F (C6D6 (1 mL) + THF (0.3
mL)): -123.0 (d, 8F, ortho); -159.3(t, 4F, para); -164.6 (t, 8F, meta).
In Situ Generation of [Mn(CO)(depe)2][BAr′4]. A J. Young NMR
tube was charged with MnBr(CO)(depe)2 (0.011 g, 0.019 mmol) and
Na[BAr′4] (0.017 g, 0.019 mmol) in a He-filled Vacuum Atmospheres
drybox. The NMR tube was quickly charged with C6D5F and closed
off. On shaking, a deep blue solution was formed. The yield of [Mn-
(CO)(depe)2]+ was essentially quantitative by 31P NMR. 1H NMR
(C6D5F): δ 8.32 (s, 8 H, C6H3(CF3)2, ortho); 7.65 (s, 4 H, C6H3(CF3)2,
para); 1.6-0.9 (m, 48H, depe). 31P{1H} NMR (C6D5F): δ 81.4 (s).
Large-Scale Synthesis of [Mn(CO)(depe)2][BAr′4]. The reaction
was run under conditions identical to those for [Mn(CO)(depe)2(H2)]-
[BAr′4] below but under an atmosphere of Ar. On completion of the
reaction, a deep blue solution had formed. During filtration and vacuum
transfer of the hexane onto the reaction solution, a slight yellowing of
the reaction mixture was detected. From 0.515 g (0.895 mmol) of MnBr-
(CO)(depe)2 and 0.918 g (1.03 mmol) of Na[BAr′4], 1.180 g of product
was isolated. NMR revealed that the product contained a mixture of
[Mn(CO)(depe)2]+ and[Mn(CO)2(depe)2]+ cations by 31P NMR. Integra-
tion of the respective 31P NMR signals showed the presence of 15%
[Mn(CO)2(depe)2][BAr′4]. The latter was identified by reacting an in
situ generated sample of [Mn(CO)(depe)2]+ with CO (see below). The
dicarbonyl complex presumably resulted from minor decomposition
of the agostic complex by trace O2 during filtration and vacuum transfer
steps.
phosphines in trans-M(CO)(PP)2(L) for L ) H2 and N2 is far
less than that for increasing charge (M ) Mn+ vs Mo).
There is a fine balance between H2 f M donation and M f
H2 back-donation in the activation of the H-H bond. In highly
electrophilic cationic systems, increased σ donation compensates
for the reduction of back-bonding prevalent in more electron-
rich neutral analogues. Thus, the H-H bond lengths and J(HD)
of [Mn(CO)(PP)2(H2)]+ and other cationic complexes with H2
trans to CO are strikingly similar to their neutral analogues and
nearly invariant. The ligand trans to H2 controls the H-H
distance more so than all of the cis ligands combined and
placement of a strong acceptor such as CO trans to an open
coordination site in a cationic group 5-10 system Virtually
guarantees formation of a M(η2-X-H) σ complex regardless
of the metal or cis ligands. This would not have been expected
from the accepted principle that increased electron richness at
the metal lengthens the H-H bond, and thus the viewpoints on
M-H2 bonding and activation continue to evolve. In this
context, hydrogenase enzymes that catalyze interconversion of
H2 and protons contain strong π-acceptor ligands such as CO
and CN rarely seen in biological systems.34 The molecular
binding and activation of H2 could thus be elegantly controlled
at the binding site merely by cis versus trans disposition of the
acceptor ligands.
Experimental Section
MnBr(CO)5, GaCl3, C6F5Br, and Li(n-butyl) were purchased from
Aldrich Chemical Co., and depe from Strem Chemical Co. MnBr(CO)-
3a
(dppe)2 and Na[BAr′4]10b were prepared according to literature
methods. C6H6 was distilled from Na/K alloy and hexane, and Et2O
from Na/benzophenone under Ar. UHP grade gases were used for all
procedures. CO and SO2 were purchased from Matheson and used
without further purification. Solids were manipulated in a Vacuum
Atmospheres drybox under a He atmosphere. Hexane used for LiGa-
(C6F5)4 preparation was scrubbed of olefins with H2SO4 and then
distilled under Ar from Na/K alloy. CH2Cl2 was distilled from P2O5 or
CaCl2. MnBr(CO)(depe)2, LiGa(C6F5)4, and Li[Ga(C6F5)4]‚2Et2O were
prepared using standard Schlenk techniques. Cationic metal complexes
were prepared using standard high vacuum techniques in CH2Cl2 and
C6H5F solvents freshly vacuum transferred from P2O5 or CaH2. Hexanes
used in preparing cationic systems were vacuum transferred from Na/K
alloy. NMR measurements were performed on Bruker WM300 MHz,
Varian Unity 300 MHz, and Bruker AMX500 MHz instruments. FTIR
measurements were performed on a Biorad FTIR Spectrometer.
Synthesis of MnBr(CO)(depe)2. MnBr(CO)5 (1.432 g, 5.209 mmol)
and depe (2.617 g, 12.69 mmol) were added to 200 mL of benzene.
The orange reaction mixture was irradiated for 3 h under a N2 or Ar
flush with a medium-pressure Hg lamp using a quartz water-jacketed
immersion well. On completion, the yellow solution was filtered under
Ar. The volume of the benzene was reduced, and the product was
precipitated with hexane, isolated by filtration, and dried in vacuo (yield,
55%). Further purification can be afforded by precipitation from toluene
with anhydrous ethanol. Anal. Calcd for C21H48BrOP4Mn: C, 43.8%;
1
H, 8.41%. Found: C, 43.8%; 8.50%. H NMR (CD2Cl2): δ 2.3-1.1
(m, 48H, depe). 31P{1H} NMR (121.42 MHz, CD2Cl2): δ 74.2. IR
(Nujol mull): ν(CO) 1803 cm-1
.
In Situ Generation of [Mn(CO)(depe)2][Ga(C6F5)4]. A J. Young
NMR tube was charged with MnBr(CO)(depe)2 (0.019 g, 0.01 mmol)
and Li[Ga(C6F5)4] (0.030 g, 0.040 mmol) in a He-filled Vacuum
Atmospheres drybox. The NMR tube was quickly charged with C6D5F
and closed off. On shaking, a deep blue solution was formed. The yield
of the cation was quantitative by 31P NMR. 1H NMR (C6D5F): δ 1.8-
0.7 (m, 48H, depe).31P{1H} NMR (121.42 MHz, C6D5F): δ 80.3 (s).
Synthesis of [Mn(CO)(depe)2(H2)][BAr′4]. Fluorobenzene (10-
15 mL) was vacuum transferred onto MnBr(CO)(depe)2 (0.551 g, 0.957
mmol) and Na[BAr′4] (0.902 g, 1.02 mmol) at -78 °C. The reaction
was then placed under a positive pressure of H2 and allowed to warm
to room temperature. As the solution thawed and the reaction began
stirring, the color changed from yellow to blue green. After reaction
Synthesis of Li[Ga(C6F5)4]. GaCl3 was sublimed, and C6F5Br was
dried over P2O5 and vacuum transferred prior to use. Trace olefins were
removed from hexane by treatment with H2SO4 and distillation from
Na/K alloy (a critical step in the synthesis). Under an Ar atmosphere,
C6F5Br (32.03 g, 0.1297 mol) was added to 500 mL of hexane and
cooled to -78 °C. The reaction mixture was then charged with 52 mL
of 2.5 M n-BuLi. A white precipitate formed on addition. GaCl3 (5.70
g, 0.0324 mol) was dissolved in 100 mL of hexane. The GaCl3 was
(34) See, for example: Pavlov, M.; Siegbahn, P. E. M.; Blomberg, M. R.
A.; Crabtree, R. H. J. Am. Chem. Soc. 1998, 120, 848, and references
therein.

1082 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
with H2 over the solution, the color became light yellow. After 45 min,
the solution was filtered, and volume was reduced by half. On reduction
in volume, the solution became green, but on backfilling with H2, the
solution became yellow again. The solution was cooled to -78 °C,
and then hexane was vacuum transferred onto the frozen yellow
fluorobenzene solution. Care must be taken to not vacuum transfer
hexane onto the blue agostic intermediate, as trace amounts of O2
present in the system will react with it. The product mixture was placed
under H2 and allowed to warm to room temperature. On warming and
mixing of the solvents, a light yellow precipitate formed. The product
was isolated by filtration and dried in vacuo for 1 h. During this time,
the solid became light blue as H2 was released. The solid was then
placed under H2, and the resulting yellow solid was isolated and stored
In Situ Generation of [Mn(CO)2(depe)2][BAr′4]. A J. Young NMR
tube was charged with MnBr(CO)(depe)2 (0.009 g, 0.01 mmol) and
Na[B(C6H3(3,5-CF3)2)4] (0.015 g, 0.017 mmol) in a He-filled Vacuum
Atmospheres drybox. The NMR tube was quickly charged with C6D5F
and closed off. On shaking, a deep blue solution was formed. The NMR
tube was attached to a vacuum line and the solvent was cooled to -196
°C. The He atmosphere was removed, and the tube was backfilled with
CO, resulting in a light yellow solution. Yield: 100% by 31P NMR.
1H(C6D5F): δ 8.63 (s, 8 H, C6H3(CF3)2, ortho); 7.96 (s, 4 H, C6H3-
(CF3)2, para); δ 1.9-1.2 (m, 48H, depe). 31P{1H} NMR (121.42 MHz,
CD2Cl2): δ 75.8 (s).
Generation of [Mn(CO)2(depe)2][BAr′4]. A 0.04 g solid sample
of material from the large scale synthesis of [Mn(CO)(depe)2][BAr′4]
was placed under an atmosphere of CO until the blue color of the agostic
complex had bleached completely, yielding 0.04 g of light yellow solid.
1
under H2. Yield: 87.3%. H NMR (C6D6F): δ 8.31 (s, 8 H, C6H3-
(CF3)2, ortho); 7.63 (s, 4 H, C6H3(CF3)2, para); 1.7-0.7 (m, 48H, depe);
-10.25 (s, 19 Hz fwhm at 500 MHz, 2H, Mn-(H2)). 31P{1H} NMR
IR: ν(CO) 1888 cm-1
.
(C6D6F): δ 83.5 (s). IR (Nujol mull): ν(CO) 1896 cm-1
.
Competition Study of Binding H2 and N2 to [Mn(CO)(depe)2]-
[BAr′4] and [Mn(CO)(dppe)2][BAr′4]. A typical procedure was
performed as follows. A 1-L ballast was evacuated to 5 × 10-5 mmHg.
At 77 K, the ballast was charged to 307.5 mmHg with N2 and then
charged to 615.6 mmHg with H2. The ballast was allowed to warm
until the pressure had risen to 820.0 mmHg. A J. Young NMR tube
containing in situ generated [Mn(CO)(depe)2][BAr′4] in C6D5F cooled
to 195 K was evacuated, exposed to the gas from the ballast, and
warmed to room temperature. The sample was allowed to equilibrate
with the gas mixture on completion of the reaction, as evidenced by
the bleaching of the blue color of the agostic complexes. The NMR
tube was then closed off, and the solution was analyzed by NMR as
described in the Results Section.
Synthesis of Mn(CO)(SO2)(dppe)2][BAr′4]. CH2Cl2 (15 mL) was
vacuum transferred onto MnBr(CO)(dppe)2 (0.315 g, 0.328 mmol) and
Na[BAr′4] (0.307 g, 0.346 mmol) at -78 °C. The reaction vessel was
backfilled with SO2 and allowed to warm to room temperature under
positive pressure of SO2. The reaction was allowed to stir for 30 min.
The solution color initially changed from yellow orange to blue, and
as the SO2 reacted, it became orange. The solution was filtered, and
the complex began to lose SO2, as evidenced by a return of the blue
color of agostic complex. The solution was then maintained under an
SO2 atmosphere. Attempts to precipitate the complex from CH2Cl2 with
15 mL of hexane, added by vacuum transfer onto the CH2Cl2 solution
cooled to -78 °C, failed. The CH2Cl2 was removed in vacuo, and the
resulting oil was stirred with 15 mL of hexane under an atmosphere of
SO2 until the product solidified. The hexane solution was filtered off,
and the solid was dried in vacuo. The resulting solid lost SO2 and was
placed under an atmosphere of SO2 prior to isolation (yield: 0.549 g).
The product was stored under SO2. 1H NMR (CD2Cl2): δ 7.74 (s, 8 H,
C6H3(CF3)2, ortho); 7.52 (s, 4 H, C6H3(CF3)2, para); 7.48-6.82 (m, 40H,
Ph); 3.12 (m, 4H, PCH2CH2P); 2.66 (m, 4H, PCH2CH2P). 31P{H}(CD2-
Cl2): δ 67.63.
Synthesis of [Mn(CO)(depe)2(N2)][BAr′4]. The complex was
prepared under identical conditions to the above H2 complex but under
an atmosphere of N2 using MnBr(CO)(depe)2 (0.411 g, 0.714 mmol)
and Na[BAr′4] (0.902 g, 1.02 mmol). The complex was produced in
86% yield and was stored under N2. The bound N2 possesses a similar
1
lability to that of bound H2. H NMR (C6D6F): δ 8.32 (s, 8H, C6H3-
(CF3)2, ortho), 7.64 (s, 4 C6H3(CF3)2, para); δ 1.7-0.9 (m, 48 H, depe).
IR (Nujol mull, cm-1): ν(CO), 1896; υ(NN), 2146.
Synthesis of [Mn(CO)(depe)2(H2)][Ga(C6F5)4]. The compound was
prepared under identical conditions to [Mn(CO)(depe)2(H2)][BAr′4] from
MnBr(CO)(depe)2 (2.010, 2.094 mmol) and Li[Ga(C6F5)4] (2.507 g,
1
3.366 mmol) in 84% yield. H NMR (C6D5F): δ 1.7-0.8 (m, 48H,
depe), -10.23 (s, 25 Hz fwhm at 500 MHz, 2H, Mn-H2). 31P{1H}
NMR (C6D5F): δ 83.4 (s).
Synthesis of [Mn(CO)(depe)2(D2)][Ga(C6F5)4]. The compound was
prepared under conditions identical to the above with the exception
that the reaction was run under D2. Using MnBr(CO)(depe)2 (1.508,
1.571 mmol) and Li[Ga(C6F5)4] (1.898 g, 2.548 mmol), the complex
1
was produced in 84% yield. H NMR (C6D5F): δ 1.5-0.6 (m, 48H,
depe). 31P{1H} NMR (C6D5F): δ 83.3 (s).
Synthesis of [Mn(CO)(dppe)2][Ga(C6F5)4]. Toluene (30 mL) was
vacuum transferred onto MnBr(CO)(dppe)2 (2.012 g, 2.096 mmol) and
Li[Ga(C6F5)4] (2.855 g, 3.833 mmol) at -78 °C. The reaction was
warmed to room temperature and stirred for 30 min, yielding a blue
solution. The toluene was removed in vacuo, and the residue was
extracted with CH2Cl2. The blue solution was filtered, and the product
was crystallized by reduction of volume of the CH2Cl2 and addition of
hexane by vacuum transfer at -20 to -30 °C. A blue oil formed and
then solidified. The resulting light blue supernatant was decanted off,
1
and the product was dried in vacuo for 3 h. Yield: 96%. H NMR
(CD2Cl2): δ 7.3-7.1 (m, 24 H, C6H5); 6.23 (m, 16 H, C6H5); 2.79 (m,
8 H, PCH2CH2P). 31P{1H} NMR (CD2Cl2): δ 83.0 (s).
X-ray Structure Determination of [Mn(CO)(depe)2][Ga(C6F5)4].
Crystallographic data are summarized in Table 11. A blue rectangular
block was mounted in a capillary and flame sealed. The capillary was
then placed on a Bruker P4/CCD/PC diffractometer and cooled to 210
K using a Bruker LT-2 temperature device. The data were collected
using a sealed, graphite monochromatized Mo KR X-ray source. The
lattice was determined using 44 reflections. A hemisphere of data was
collected using a combination of æ and ω scans, with 30 s frame
exposures and 0.3° frame widths. Data collection, initial indexing, and
cell refinement were handled using SMART software, and frame
integration and final cell parameter calculation were carried out using
SAINT software.35 The final cell parameters were determined using a
least-squares fit to 2354 reflections. The data were corrected for
absorption using SADABS program.36 Decay of reflection intensity was
not observed.
Synthesis of [Mn(CO)(dppe)2(H2)][Ga(C6F5)4]. Method A. The
compound was prepared under conditions identical to those for [Mn-
(CO)(depe)2(H2)][BAr′4] with the following modifications. CH2Cl2 was
used as the reaction solvent, and MnBr(CO)(dppe)2 (0.958, 0.998 mmol)
and Li[Ga(C6F5)4] (1.195 g, 1.604 mmol) were the reactants. The final
product was crystallized from a mixture of 5 mL of CH2Cl2 and 50
mL of hexane (yield: 1.721 g). 31P NMR was found to be clean, but
the product contains a slight amount of unreacted Li[Ga(C6F5)4] which
was present in excess. 1H NMR (CD2Cl2): δ 7.4-7.0 (m, 40H, C6H5);
2.50 (m, 4H, PCH2CH2P); 2.24 (m, 4H, PCH2CH2P); -7.22 (s, 2H, 33
Hz fwhm at 300 MHz, Mn-H2). 31P{1H} NMR (CD2Cl2): δ 85.9 (s).
Method B. CH2Cl2 (20 mL) was vacuum transferred onto [Mn(CO)-
(dppe)2][Ga(C6F5)4] (2.007 g, 1.240 mmol). The resulting solution was
placed under an atmosphere of H2 and quickly turned from blue to
yellow. [Mn(CO)(dppe)2(H2)][Ga(C6F5)4] precipitated out partially from
solution and was isolated by reduction of solution volume at -20 to
-30 °C and precipitation with hexane. Yield: 95.9%.
(35) (a) SMART Version 4.210, 1996, Bruker Analytical X-ray Systems,
Inc., 6300 Enterprise Lane, Madison, Wisconson 53719. (b) SAINT
Version 4.05, 1996, Bruker Analytical X-ray Systems, Inc., Madison,
Wisconson 53719.
(36) SADABS, first release, G. M. Sheldrick, University of Gottingen,
Germany.
Synthesis of [Mn(CO)(dppe)2(D2)][Ga(C6F5)4]. The compound was
prepared by Methods A and B above except under an atmosphere of
D2. NMR resonances were identical except for the absence of the signal
at δ -7.22 due to Mn-H2.

Reversible Displacement of Polyagostic Interactions
Inorganic Chemistry, Vol. 38, No. 6, 1999 1083
Table 11. Summary of Crystallographic Data
[Mn(CO)(depe)2]+
[Ga(C6F5)4]-
[Mn(CO)(depe)2(H2)]+
[Mn(CO)(dppe)2(H2)]+
[Ga(C6F5)4]-
[Cl{Ga(C6F5)3}2]-
formula
fw
T, °C
C45H48OP4F20GaMn
1233.37
-163
0.71073
orthorombic
P21212
blue
C45H50OP4F20GaMn
1235.38
-76
C89H50OP4ClF30Ga2Mn
2059.00
-135
0.71073
monoclinic
C2/c
λ, Å
0.71073
triclinic
crystal system
space group
color
P1h
yellow
yellow
a, Å
b, Å
c, Å
17.528(4)
18.192(4)
8.169(2)
12.987(2)
14.942(2)
15.622(2)
105.184(6)
109.207(8)
103.632(7)
1.585
10.224(4)
32.73(3)
24.310(9)
90
97.69(3)
90
1.696
8062(8)
4
R, deg
â, deg
γ, deg
F
calcd, g cm-3
1.573
2604.8(10)
2
V, Å3
2584.8(6)
2
Z
µ, mm-1
R [I > 2σ(I)]a
Rw[I > 2σ(I)]b
0.988
1.050
0.0738
0.0522
0.0798
0.1442c
0.1462d
0.1884e
a R ) ∑||Fo| - |Fc||/∑|Fo|. b Rw ) [∑[w(Fo - Fc2)2]/∑[w(Fo )2]]1/2
.
c w ) 1/[σ2(Fo ) + (0.0672*P)2]. d w ) 1/[σ2(Fo ) + (0.0947*P)2]. e w )
2
2
2
2
1/[σ2(Fo ) + (0.1064*P)2 + 46.6639*P].
2
The structure was solved in space group P21212 using Patterson and
difference Fourier techniques. The initial solution revealed the man-
ganese, gallium, and phosphorus atom positions. The remaining atomic
positions were determined from subsequent Fourier synthesis. All
hydrogen atom positions were fixed using the HFIX command in
SHELXTL PC.37 The C-H distances were fixed at 0.96 Å for methyl
and 0.97 Å for methylene. The hydrogen atoms were refined using a
riding model with their isotropic temperature factors set to 1.5 (methyl)
or 1.2 (methylene) times the equivalent isotropic U of the carbon atom
they were bound to. Attempts to locate the agostic hydrogen atom
positions were not successful. The final refinement included anisotropic
temperature factors on all non-hydrogen atoms. Structure solution and
graphics were performed using SHELXTL PC. SHELXL-93 was used
for structure refinement and creation of publication tables.38
X-ray Structure Determination of [Mn(CO)(depe)2(H2)][Ga-
(C6F5)4]. Crystallographic data are summarized in Table 11. A yellow
rectangular block was mounted in a capillary, flame sealed under H2,
and then placed under a liquid nitrogen cold stream on a Siemens P4/
PC diffractometer. The radiation used was graphite monachromatized
Mo KR radiation. The lattice parameters were optimized from a least-
squares calculation on 25 carefully centered reflections of high Bragg
angle. The data were collected using ω scans with a 0.78° scan range.
Three check reflections monitored every 97 reflections showed no
systematic variation of intensities. Lattice determination and data
collection were carried out using XSCANS Version 2.1b software. All
data reduction were performed using SHELXTL PC Version 4.2/360
software. The structure refinement was performed using SHELX-93
software.39
majority of all other atom positions. Subsequent Fourier synthesis gave
all remaining non-hydrogen atom positions. The methyl and methylene
hydrogen atoms were fixed in positions of ideal geometry, with C-H
distances of 0.96 and 0.97 Å, respectively. The hydrogen atoms were
refined using the riding model in the HFIX facility in SHELX-93, with
their isotropic temperature factors fixed at 1.2 (methylene) or 1.5
(methyl) times the equivalent isotropic U of the carbon atom they were
bonded to. Attempts to locate and refine the hydrogen molecule were
not successful. The final refinement included anisotropic thermal
parameters on all non-hydrogen atoms.
X-ray Structure Determination of [Mn(CO)(dppe)2(H2)][Cl(Ga-
(C6F5)3)2]. Crystallographic data are summarized in Table 11. A yellow
rectangular plate was mounted on a thin glass fiber using silicone grease.
The crystal, which was mounted directly from the reaction vessel under
a hydrogen gas flow, was then immediately placed under a liquid
nitrogen stream on a Siemens P4/PC diffractometer. The radiation used
was graphite monochromatized Mo KR radiation. The lattice parameters
were optimized from least-squares calculation on 25 carefully centered
reflections of high Bragg angle. Three check reflections monitored every
97 reflections showed no systematic variation of intensities. Lattice
determination and data reduction, including Lorentz polarization
corrections and structure solution and graphics, were performed using
SHELXTL PC Version 4.2/360 software. The structure refinement was
performed using SHELX-93 software.39 The data were not corrected
for absorption due to the low absorption coefficient.
Space groups Cc and C2/c were suggested by systematic absences.
The structure was initially solved in Cc using direct methods to locate
the gallium, chlorine, manganese, and phosphorus atoms. The remaining
The structure was solved in space group P1h using Patterson and
difference Fourier techniques. This solution yielded the metal and the
(45) Gusev, D. G.; Nietlispach, D.; Eremenko, I. L.; Berke, H. Inorg. Chem.
1993, 32, 3628.
(37) SHELXL PC Version 4.2/360, 1994, Bruker Analytical X-ray Systems,
Inc., Madison, Wisconson 53719.
(46) Kubas, G. J.; Burns, C. J.; Khalsa, G. R. K.; Van Der Sluys, L. S.;
Kiss, G.; Hoff, C. D. Organometallics 1992, 11, 3390.
(47) Kubas, G. J.; Ryan, R. R.; Vergamini, P. J., unpublished results.
(48) Van der Sluys, L. S.; Huffman, J. C.; Kubas, G. J., unpublished results.
(49) Bruns, W.; Hausen, H.-D.; Kaim, W.; Schulz, A. J. Organomet. Chem.
1993, 444, 121.
(50) Lang, R. F.; Ju, T. D.; Kiss, G.; Hoff, C. D.; Bryan, J. C.; Kubas, G.
J. J. Am. Chem. Soc. 1994, 116, 7917.
(51) Kubas, G. J.; Jarvinen, G. D.; Ryan, R. R. J. Am. Chem. Soc. 1983,
105, 1883.
(52) Andrea, R. R.; Vuurman, M. A.; Stufkens, D. J.; Oskam, A. Recl.
TraV. Chim. Pays-Bas 1986, 105, 372.
(53) Burdett, J. K.; Downs, A. J.; Gaskill, G. P.; Graham, M. A.; Turner,
J. J.; Turner, R. F. Inorg. Chem. 1978, 17, 523.
(38) SHELX-93, G. M. Sheldrick, University of Gottingen, Germany.
(39) XSCANS and SHELXTL PC are products of Siemens Analytical X-ray
Instruments, Inc., 300 Enterprise Lane, Madison, Wisconson 53719.
SHELX-93 is a program for crystal structure refinment written by G.
M. Sheldrick, 1993, University of Gottingen, Germany.
(40) Kubas, G. J.; Nelson, J. E.; Bryan, J. C.; Eckert, J.; Wisnieski, L.;
Zilm, K. Inorg. Chem. 1994, 33, 2954.
(41) Zilm, K. W.; Millar, J. M. AdV. Magn. Opt. Reson. 1990, 15, 163.
(42) (a) Kubas, G. J.; Ryan, R. R.; Swanson, B. I.; Vergamini, P. J.;
Wasserman, H. J. J. Am. Chem. Soc. 1984, 106, 451. (b) Kubas, G.
J.; Unkefer, C. J.; Swanson, B. I.; Fukushima, E. J. Am. Chem. Soc.
1986, 108, 7000.
(43) Tatsumi, T.; Tominaga, H.; Hidai, M.; Uchida, Y. J. Organomet. Chem.
1980, 199, 63.
(44) Luo, X.-L.; Michos, D.; Crabtree, R. H. Organometallics 1992, 11,
237.
(54) Albertin, G.; Antoniutti, S.; Garcia-Fontan, S.; Carballo, R.; Padoan,
F. J. Chem. Soc., Dalton Trans. 1998, 2071.
(55) Luo, X.-L.; Kubas, G. J.; Burns, C. J.; Eckert, J. Inorg. Chem. 1994,
33, 5219.

1084 Inorganic Chemistry, Vol. 38, No. 6, 1999
King et al.
2
τc/(1 + τc ωH2) ) 2.39 × 10-10
non-hydrogen atoms appeared in subsequent difference maps. The Cc
refinement, with all non-hydrogen atoms, failed to converge at this
step at R1 ) 9%. Moreover, many atoms were nonpositive definite. A
twin refinement in Cc corresponded to a perfect racemic mixture with
R1 ) 9% but did not converge. At this juncture, the refinement was
converted to C2/c with the carbon and oxygen atoms of the carbonyl
fixed at one-half occupancy. The refinement proceeded well and
converged with no nonpositive definite atoms. The six-membered rings
of the dppe ligands were refined as rigid bodies, with ring carbon-
carbon distances fixed at 1.39 Å. No hydrogen atoms corresponding
to a bound H2 were found in a high angle difference map. All hydrogen
atoms were fixed and refined in positions of ideal geometry. The C-H
distances were fixed at 0.93 Å (ethyl) and 0.97 Å (phenyl). All hydrogen
atoms were refined using the riding model in the HFIX facility in
SHELX-93 and had their isotropic temperature factors fixed at 1.2 times
the equivalent isotropic U of the atom they were bonded to. The final
refinement included anisotropic thermal parameters on all non-hydrogen
atoms.
4τc/(1 + 4τc ωH2) ) 5.18 × 10-10
2
6
RHH ) (1.29 × 10-46)/rHH
2
τc/(1 + τc ω-2) ) 2.60 × 10-10
2
3τc/(1 + τc ωH2) ) 7.18 × 10-10
6τc/(1 + τc ω+2) ) 1.31 × 10-9
RHD ) (8.20 × 10-48)/rHD
2
6
2γH γMn2(h/2π)2
2
2
RHMn
)
S(S + 1){τc/(1 + τc ω-2) + 3τc/(1 +
15rHMn
2
2
τc ω H2) + 6τc/(1 + τc ω+2)}
γMn ) 6.598 × 103 rads-1gauss-1
Ineleastic Neutron Scattering Studies. High-frequency data ob-
tained at 15 K on the FDS instrument23b at the Manuel Lujan Jr. Neutron
Scattering Center of Los Alamos National Laboratory provided
vibrational data, including transitions to the excited librational states
(“torsions”) of the H2 ligand. Low-frequency data could only be
collected (T ) 5 K) on the dppe compound using the cold neutron
time-of-flight spectrometer IN5 at the Institut Laue-Langevin, Grenoble
(France) and on the Mibemol spectrometer of the Laboratoire Le´on
Brillouin, Saclay (France). These data yield the tunnel splitting of the
librational ground state of the coordinated H2.23c Approximately 1-2
g samples, sealed under a partial H2 atmosphere in quartz sample
holders, were used for these experiments. The vibrational spectra were
obtained with the aid of a spectral difference technique23d that utilizes
two samples, one with H2 ligands and the other with D2, to subtract
vibrations not associated with the H2 ligand.
S ) 5/2
ωMn ) 2π(74.1 × 106) ) 4.66 × 108 rad s-1
ω- ) ωH - ωMn ) 2.35 × 109 rad s-1
ω+ ) ωH + ωMn ) 1.41 × 109 rad s-1
2
2γH γMn2(h/2π)2/15 ) 4.61 × 10-39
2
τc/(1 + τc ω-2) ) 2.73 × 10-10
2
3τc/(1 + τc ωH2) ) 7.18 × 10-10
2
6τc/(1 + τc ω+2) ) 1.24 × 10-9
This work was supported by the Department of Energy, Office of
Basic Energy Sciences, Chemical Sciences Division. The authors also
thank Jean Huhmann-Vincent for organosilane reactions. This work
has also benefitted from the use of facilities at the Manuel Lujan Jr.
Neutron Scattering Center, a National User Facility funded as such by
the Department of Energy, Office of Basic Energy Sciences.
6
RHMn ) (9.01 × 10-47)/rHMn
Derivation of equation for Mn-H distance:
Appendix. Derivation of equations for determining H-H and Mn-H
distance by T1 measurements. In all cases R ) 1/T1.
RHH(obs) ) RHH + RHMn + RHligand RHligand ≈ 0
RHHObs ) RHH + RHMn
3γH4(h/2π)2
RHH(obs) ) (1.29 × 10-46)/rHH6 + (9.01 × 10-47)/rHMn
6
2
2
RHH
)
{τc/(1 + τc ωH2) + 4τc/(1 + 4τc ωH2)}
6
10rHH
rHH6 (1.21 × 10-46)/(RHH(obs) - RHD(obs)
)
2γH2γ2(h/2π)2
S(S + 1){τc/(1 + τc2ω2) +
RHH(obs) ) (1.29 × 10-46)/(1.21 × 10-46) (RHH(obs) - RHD(obs)) +
RHD
)
6
15rHD
6
2
2
(9.01 × 10-47)/rHMn
3τc/(1 + τc ω H2) + 6τ c/(1 + τc ω + 2)}
6
RHH(obs) ) 1.07 (RHH(obs) - RHD(obs)) + (9.01 × 10-47)/rHMn
RHH(obs) - 1.07(RHH(obs) - RHD(obs)) ) (9.01 × 10-47)/rHMn
γH ) 2.675 × 104 rads-1 gauss-1
γD ) 4.107 × 103 rads-1 gauss-1
τc ) 0.63/ω H ωH ) 2π(300 × 106) ) 1.88 × 109 rad s-1
6
RHH(obs) - 1.07RHH(obs) + 1.07RHD(obs) ) (9.01 × 10-47)/rHMn
6
6
1.07RHD(obs) - 0.07RHH(obs) ) (9.01 × 10-47)/rHMn
τc ) 3.34 × 10-10 rad-1
s
S ) 1
R must be in s-1 and distance r is in cm. NMR constants were taken
from ref 13b. τc is assumed to be the same in all equations, an
approximation also made in ref 13b.
ωD ) 2π(46.2 × 106) ) 2.90 × 108 rad s-1
ω- ) ωH - ωD ) 2.17 × 109 rad s-1
ω+ ) ωH + ωD ) 1.59 × 109 rad s-1
3γH4(h/2π)2/10 1.71 × 10-37
Supporting Information Available: X-ray data for complexes 1-3
(crystal information, atomic coordinates, bond lengths and angles,
anisotropic displacement factors, hydrogen coordinates, packing dia-
grams, labeled ORTEPS). This material is available free of charge via
the Internet at http://pubs.acs.org.
2γH2γD2(h/2π)2/15 ) 1.79 × 10-39
IC981263L