A dihydrogen bond between a bridging hydride and the NH proton of a coordinated dimethylamine: Solid state, solution and theoretical characterization

Article abstract of DOI:10.1016/j.jorganchem.2004.12.003

The addition of one equivalent of dimethylamine (DMA) to the 44 valence-electron triangular cluster anion [Re₃(μ₃-H) (μ-H)₃(CO)₉]^- (1) affords the novel unsaturated derivative [Re₃(μ-H)₄(CO) ₉(DMA)]^- (2, 46 valence electrons) which contains a dimethylamine molecule terminally coordinated to a cluster vertex. Theoretical calculations (DFT) reveal that in the more stable conformation the dimethylamine NH proton is directed towards the hydride bridging the opposite cluster edge in syn position, the close proximity of the ligands bound to the cluster surface allowing the formation of an unconventional N-H ? (μ-H)Re₂ hydrogen bond. The presence of this conformation in the solid state has been proven by an X-ray structural analysis of crystalline [PPh₄]2. Spectroscopic evidences (IR and NMR) indicate that the dihydrogen bond is maintained also in solution and, by the evaluation of the proton spin-lattice relaxation rates at variable temperature, a good estimate of the H ? H distance in solution has been determined.

Full text of DOI:10.1016/j.jorganchem.2004.12.003

Journal of Organometallic Chemistry 690 (2005) 2044–2051
www.elsevier.com/locate/jorganchem
A dihydrogen bond between a bridging hydride and the NH proton
of a coordinated dimethylamine: solid state, solution and
theoretical characterization
a,
b,
a
Monica Panigati *, Pierluigi Mercandelli *, Giuseppe DÕAlfonso ,
Tiziana Beringhelli , Angelo Sironi
a
b
a
Dipartimento di Chimica Inorganica, Metallorganica e Analitica, via Venezian 21, I-20133, Milano, Italy
Dipartimento di Chimica Strutturale e Stereochimica Inorganica, via Venezian 21, I-20133, Milano, Italy
b
Received 6 December 2004; accepted 7 December 2004
Available online 12 January 2005
Abstract
The addition of one equivalent of dimethylamine (DMA) to the 44 valence-electron triangular cluster anion [Re3(l3-H)
(l-H)3(CO)9]ꢀ (1) affords the novel unsaturated derivative [Re3(l-H)4(CO)9(DMA)]ꢀ (2, 46 valence electrons) which contains a dim-
ethylamine molecule terminally coordinated to a cluster vertex. Theoretical calculations (DFT) reveal that in the more stable
conformation the dimethylamine NH proton is directed towards the hydride bridging the opposite cluster edge in syn position,
the close proximity of the ligands bound to the cluster surface allowing the formation of an unconventional N–H ꢁ ꢁ ꢁ (l-H)Re2
hydrogen bond. The presence of this conformation in the solid state has been proven by an X-ray structural analysis of crystalline
[PPh4]2. Spectroscopic evidences (IR and NMR) indicate that the dihydrogen bond is maintained also in solution and, by the eval-
uation of the proton spin-lattice relaxation rates at variable temperature, a good estimate of the H ꢁ ꢁ ꢁ H distance in solution has
been determined.
ꢀ 2004 Elsevier B.V. All rights reserved.
Keywords: Cluster compounds; Density functional calculations; Hydride ligands; Hydrogen bonds; NMR spectroscopy; X-ray diffraction
1. Introduction
the classical X–H ꢁ ꢁ ꢁ Y hydrogen bonds (X and Y being
highly electronegative main group atoms).
Since it was first reported [1], the occurrence of weak
interactions between metal hydrides and hydrogen
atoms bonded to electronegative main group atoms (N
or O) has been the object of a great deal of research
[2]. The term ‘‘dihydrogen bond’’ (DHB) was coined
to distinguish this ‘‘unconventional’’ interaction from
Only a few examples of DHB have so far been re-
ported in cluster chemistry [3], and the proton acceptor
has usually been found to be a terminally coordinated
hydride. We have recently reported the first example
of a DHB involving a bridging hydride in the triangular
rhenium cluster anion [Re3(l-H)4(CO)9(HPz)]ꢀ, con-
taining a pyrazole ligand terminally coordinated to a
cluster vertex [4].
Here we present a second example of a N–H ꢁ ꢁ ꢁ (l-
H)Re2 dihydrogen bond in a triangular rhenium cluster,
involving a bridging hydride as a proton acceptor and
the NH proton of a coordinated amine. The character-
*
Corresponding
authors.
+390250314454 (P. Mercandelli).
Tel.:
+390250314447;
fax:
E-mail addresses: monica.panigati@unimi.it(M. Panigati), pierluigi.
mercandelli@unimi.it (P. Mercandelli).
0022-328X/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.12.003

M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
2045
ization of this interaction has been supported both by
experimental evidence (IR, NMR and X-ray analysis)
and theoretical computations (DFT).
2. Results and discussion
It has already been shown that the addition of two-
electron donor molecules L to the ‘‘super-unsaturated’’
triangular cluster anion [Re3(l3-H)(l-H)3(CO)9]ꢀ (1,
44 valence electrons) [5] easily and selectively affords
unsaturated [Re3(l-H)4(CO)9(L)]ꢀ adducts (46 valence
electrons), in which the ligand L occupies an axial posi-
tion on the vertex of the cluster opposite to the Re-
(l-H)2Re basal interaction (see Scheme 1). On using
dimethylamine NHMe2 (DMA) as ligand L we have
now quantitatively obtained the novel cluster anion
[Re3(l-H)4(CO)9(DMA)]ꢀ (2).
Fig. 1. ORTEP drawing of the [Re3(l-H)4(CO)9(DMA)]ꢀ anion (2)
with partial labelling scheme. Thermal ellipsoids are drawn at the 30%
probability level. Hydrogen atoms were given arbitrary radii.
1
The H NMR spectrum (CDCl3, 298 K) shows three
hydridic resonances at d ꢀ 7.97, ꢀ9.34 and ꢀ9.69 ppm,
in 1:1:2 ratio, in agreement with the idealized Cs symme-
try of anion 2.
The rotational conformer of 2 depicted in Scheme 1,
in which the NH proton points towards the hydride
bridging the opposite cluster edge, has been found to
be present in crystalline [PPh4] 2 by an X-ray structural
analysis. DFT computations at the B3LYP/DZP level of
theory also indicate that the more stable rotamer for an-
ion 2 is that in which a hydrogen bond is formed. Spec-
troscopic investigations (IR and NMR) confirmed that
this conformation is also the preferred one in solution,
at least in solvents not able to act as proton acceptors.
interactions are neglected each metal attains a distorted
octahedral coordination.
Cluster 2 shows two long Re–Re distances (mean va-
˚
˚
lue 3.182 A) and one shorter interaction (2.799 A) corre-
sponding to a formal Re(l-H)2Re formal double bond;
a similar pattern has been observed in the related unsat-
urated clusters [Re3(l-H)4(CO)9L]ꢀ (L = carbonyl, pyri-
dine, triphenylphosphine, or pyrazole) [4,6].
As shown in Fig. 1, the H(N1) proton of the dimeth-
ylamino ligand bound to Re(1) forms a hydrogen bond
with the hydride H(b), bridging the opposite cluster edge
Re(2)–Re(3). The distance between H(N1) and H(b) is
2.1. Solid-state characterization of the anion [Re3(l-
H)4(CO)9(DMA)]ꢀ (2)
˚
2.239 A; this distance has been computed using idealised
positions for both the hydrogen atoms [7], even if they
were found in a difference Fourier map, in order to
avoid the systematic errors associated with X-ray dif-
fraction determined hydrogen atom positions. The
adoption of this conformation by anion 2 could be
attributed to an attempt of avoid unfavourable steric
interactions between the DMA methyl groups and the
diagonal carbonyl ligands bound to the Re(l-H)2Re
moiety. However, the presence of a manifest attractive
interaction can be inferred by the significant tilting of
the dimethylamino ligand towards the bridging hydride
The molecular structure of anion 2, as determined in
a crystal of its [PPh4]+ salt, is depicted in Fig. 1 with a
partial labelling scheme. A selection of bond parameters
is reported in Table 1.
Anion 2 possesses an overall idealised Cs symmetry
and contains an isosceles triangular metal core bearing
nine terminal carbonyls, three to each rhenium atom,
a dimethylamino ligand and four edge-bridging hydrido
ligands. When considering the bonding interaction
around the rhenium atoms, if the direct metal–metal
Scheme 1.

2046
M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
Table 1
Table 2
a
˚
˚
Selected bond distances (A) and angles (ꢁ) for the trinuclear complex 2
Relevant hydrogen bond parameters (A and ꢁ) for [PPh4]2
Re(1)–Re(2)
Re(1)–Re(3)
Re(2)–Re(3)
Re(1)–N(1)
Re(1)–C(11)
Re(1)–C(12)
Re(1)–C(13)
Re(2)–C(21)
Re(2)–C(22)
Re(2)–C(23)
3.1844(8) Re(3)–C(31)
3.1786(8) Re(3)–C(32)
2.7989(8) Re(3)–C(33)
1.914(13)
1.932(13)
1.897(13)
1.456(16)
1.472(17)
N(1)–H(N1)
1.025
2.239
2.332
3.032
3.197(12)
133.0
141.3
85.5
H(N1) ꢁ ꢁ ꢁ H(B)
H(N1) ꢁ ꢁ ꢁ O(220)
N(1) ꢁ ꢁ ꢁ H(B)
2.246(8)
N(1)–C(1)
1.889(12) N(1)–C(2)
1.915(12)
1.924(13)
N(1) ꢁ ꢁ ꢁ O(220)
N(1)–H(N1) ꢁ ꢁ ꢁ H(B)
N(1)–H(N1) ꢁ ꢁ ꢁ O(220)
H(B)–H(N1) ꢁ ꢁ ꢁ O(220)
a
1.898(12)
1.914(11)
1.910(12)
Primes refer to symmetry-related atoms (ꢀx, ꢀy, 1 ꢀ z).
Re(2)–Re(1)–Re(3)
Re(1)–Re(2)–Re(3)
Re(1)–Re(3)–Re(2)
Re(2)–Re(1)–N(1)
Re(2)–Re(1)–C(11)
52.190(17) Re(3)–Re(2)–C(21) 132.2(4)
63.799(15) Re(3)–Re(2)–C(22) 137.4(3)
action and should not be directly compared to the one
measured in solution by NMR or to the one computed
in vacuo (see below).
64.011(18) Re(3)–Re(2)–C(23)
94.6(4)
89.9(5)
91.3(6)
89.9(5)
98.9(4)
86.5(2)
90.3(4)
C(21)–Re(2)–C(22)
C(21)–Re(2)–C(23)
C(22)–Re(2)–C(23)
Re(1)–Re(3)–C(31)
Re(2)–Re(1)–C(12) 109.0(4)
Re(2)–Re(1)–C(13) 158.7(4)
2.2. Theoretical characterization of the anion [Re3(l-
H)4(CO)9(DMA)]ꢀ (2)
Re(3)–Re(1)–N(1)
Re(3)–Re(1)–C(11)
88.5(2)
88.5(3)
Re(1)–Re(3)–C(32) 108.7(4)
Re(1)–Re(3)–C(33) 159.2(4)
Re(2)–Re(3)–C(31) 133.3(4)
Re(2)–Re(3)–C(32) 134.2(4)
Re(3)–Re(1)–C(12) 160.7(4)
Re(3)–Re(1)–C(13) 106.7(4)
The possible existence of different rotamers of anion 2
has been investigated by means of a DFT computational
study. Computations have been done at the B3LYP/
DZP level of theory (see Section 4 for details). The rota-
tion of the dimethylamino ligand around the Re–N
bond has been studied by driving the (improper) torsion
angle x H(b) ꢁ ꢁ ꢁ Re(1)–N(1)–H(N1).
N(1)–Re(1)–C(11)
N(1)–Re(1)–C(12)
N(1)–Re(1)–C(13)
C(11)–Re(1)–C(12)
C(11)–Re(1)–C(13)
C(12)–Re(1)–C(13)
176.6(5)
95.3(4)
96.3(5)
87.0(5)
86.2(6)
91.7(5)
Re(2)–Re(3)–C(33)
C(31)–Re(3)–C(32)
C(31)–Re(3)–C(33)
C(32)–Re(3)–C(33)
Re(1)–N(1)–C(1)
Re(1)–N(1)–C(2)
C(1)–N(1)–C(2)
96.2(4)
91.9(5)
90.2(6)
89.5(6)
116.4(8)
115.0(8)
110.6(12)
Re(1)–Re(2)–C(21) 102.0(5)
Re(1)–Re(2)–C(22) 107.0(3)
Re(1)–Re(2)–C(23) 158.4(4)
Two rotamers can be identified, one corresponding to
the dihydrogen bonded species found in the solid state
(rotamer A, x = 0) and the other corresponding to a
180 rotation of the DMA ligand (rotamer B, see Fig.
3). Both conformations are found to be effective minima
at the level of theory employed, although rotamer B lies
on a very flat region of the potential energy surface.
A direct evaluation of the strength of an intra-molec-
ular hydrogen bond is hampered by the impossibility of
separating the two interacting fragments. A widely em-
ployed strategy to circumvent this problem consists in
comparing the stability of two different conformations
of the molecule in which the hydrogen bond is alterna-
tively present and absent. The energy difference between
the two rotamers A and B amounts to 5 kcal molꢀ1. This
value can only be taken as an upper limit for the
strength of the dihydrogen bond, since rotamer A is sta-
bilized with respect to B not only by the presence of this
interaction but also by steric factors. Indeed, the pres-
ence of unfavourable steric interactions between the
DMA methyl groups and the diagonal carbonyl ligands
in rotamer B is clearly proven by the significant tilting of
the coordination octahedra at the rhenium atoms and by
the sizeable lengthening of the Re–N bond distance
(the angle between the Re(1)–N(1) vector and the Re3
plane being 86.5(2)ꢁ).
Crystalline [PPh4] 2 presents additional interesting
structural features. A short inter-molecular N–H ꢁ ꢁ ꢁ O
hydrogen bonding interaction is observed between the
hydrogen atom of the dimethylamino ligand and a car-
bonyl ligand of a symmetry-related anion, resulting in
the formation of a dimer linked by two bifurcated
hydrogen bonds (see Fig. 2). The structural parameters
of this interaction are listed in Table 2.
The observed NH ꢁ ꢁ ꢁ H distance is somehow affected
by the presence of this additional inter-molecular inter-
˚
˚
(2.341 A) with respect to rotamer A (2.308 A). However,
the computed energy difference is in agreement with the
observation of only rotamer A in dichloromethane solu-
tion, while in solvents able to act as effective hydrogen
bond acceptors the possible presence of a significant
amount of rotamer B can be anticipated (see below).
Fig. 2. Hydrogen bonding interactions in the crystal structure of
[PPh4][Re3(l-H)4(CO)9(DMA)].

M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
2047
Fig. 3. B3LYP/DZP energy profile along the rotation of the dimethylamino ligand around the Re–N bond for anion 2. The computed structures
corresponding to the two minima, namely rotamer A and B, are also depicted.
˚
The presence of a dihydrogen bond in rotamer A is
further supported by some computed geometric and
vibrational data. In particular, the N–H distance com-
an NH ꢁ ꢁ ꢁ H distance of 2.228 A, in good agreement
with the measured value.
˚
puted for rotamer A (1.021 A) is stretched with respect
2.3. Solution characterization of the dimethylamine
conformation in anion 2
˚
to those computed for free DMA (1.016 A) and for the
non hydrogen bonded DMA moiety in rotamer B
˚
(1.017 A). Moreover, as in the solid state structure, in
A 2D NOESY experiment, performed at 273 K in
CD2Cl2, indicates the close proximity of the N–H pro-
ton to the bridging hydride Hb. First of all, the 2D
map (Fig. 4) allows the identification of the Hb reso-
nance as that of integrated intensity 1 which shows
nOe cross-peaks with both the NH proton and the
methyl groups of DMA. The same correlations are
exhibited also by the hydridic signal of integrated inten-
sity 2, due to the hydrides Hc bridging the lateral edges.
rotamer A the Re–N vector is clearly bent towards
the basal Re–Re edge, in agreement with the presence
of an attractive interaction between NH and Hb. The
m(N–H) harmonic stretching frequency computed for
rotamer A (3465 cmꢀ1) shows a significant shift to lower
wavenumbers with respect to the ones computed for
rotamer B (3515 cmꢀ1) and for the free DMA molecule
(3526 cmꢀ1). The computed IR intensities (ol/oq)2 are
very sensitive to the local charge on the NH proton:
the value found for rotamer A (57.0 a.u.) is indeed
much larger than the ones found for rotamer B (8.3
a.u.) or free DMA (0.8 a.u.). In order to make a com-
parison with a classical hydrogen bond, the geometry
of the dimethylamine dimer in its more stable Cs con-
formation has been optimised and a frequency compu-
tation has been carried out. The computed values [d(N–
H) = 1.022 A, m(N–H) = 3428 cmꢀ1, (ol/oq)2 = 225.8
˚
a.u.] indicates that the dihydrogen bond present in an-
ion 2 is weak compared to a classical N–H ꢁ ꢁ ꢁ N hydro-
gen bond.
In rotamer A, we compute an NH ꢁ ꢁ ꢁ H distance of
˚
2.134 A, in fair agreement with the experimental values
measured both in the solid state and in solution. In or-
der to compare the computed value to the better esti-
˚
mate in solution (2.24 A, from the evaluation of T1 at
variable temperature, see below) the effect of the libra-
tion of the DMA ligand must be taken into account.
The normal mode associated with the rotation around
the Re–N bond (29 cmꢀ1) makes the major contribution
to the lengthening of the H ꢁ ꢁ ꢁ H interaction. Employing
standard Boltzmann statistics [8], we compute at 193 K
1
Fig. 4. Selected region of H 2D NOESY experiment performed on a
CD2Cl2 solution of anion 2 (4.7 T, 273 K, sm = 0.3 s).

2048
M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
However, the basal hydride Hb has a much stronger
cross-peak with the NH signal than with the methyl res-
onance, the volume ratio between the (NH ꢁ ꢁ ꢁ Hb) and
(Me ꢁ ꢁ ꢁ Hb) cross-peaks being 5.0, in spite of the fact
that 6 protons contribute to the latter interaction (see
Table 3). On the contrary, the analogous ratio involving
the hydrides Hc bridging the lateral edges, was only 0.59.
As reported in Table 3, the distances between the
hydrido ligands (as computed by DFT) are virtually
insensitive to the conformation of anion 2. The param-
eterisation of the nOe volumes with respect to the dis-
tances Ha ꢁ ꢁ ꢁ Hb and Ha ꢁ ꢁ ꢁ Hc is therefore possible
without making any assumption as to the presence in
solution of a specific rotamer. The nOe derived distances
for the (NH,Me) ꢁ ꢁ ꢁ (Hb, Hc) interactions are reported
in Table 3, along with the values computed for rotamers
A and B employing the DFT optimised geometries and
averaging the equivalent interproton vectors assuming a
fast rotation of the methyl groups [9]. These values
unambiguously confirm the (almost) exclusive presence
of rotamer A in dichloromethane solution.
The values of T1 for both Hb and Hc reach their mini-
mum value at approximately 193 K. The increase of
the relaxation rates of Hb and Hc with respect to Ha
(RH–H) is clearly attributable to the dipolar contribution
of the NH and methyl protons to the relaxation of these
hydrides. From the 2D NOESY experiment it is clear
that for the syn hydride Hb the main contribution comes
from the NH proton. Therefore, estimating (at the min-
imum of the temperature course of T1, where the sc va-
lue is known) the difference between the relaxation rates
of the syn (Hb) and anti (Ha) hydride, it is possible to
Further experimental evidence of the proton–hydride
interaction is usually provided by the shortening of the
longitudinal relaxation times T1 of the hydride, due to
the dipolar contribution of the close hydrogen atom
[10]. We therefore measured the proton longitudinal
relaxation time T1 of all the hydrides and the NH proton
in 2, in CD2Cl2 solution, in the temperature range 175–
273 K (see Table 4). Their logarithmic plot versus 1/T
(shown in Fig. 5) displays the typical ‘‘v’’ shape. The
T1 values measured for Hb and Hc are always shorter
than those measured for the hydride Ha, bridging the
basal edge in the opposite side of the cluster plane.
Fig. 5. Temperature dependence of the longitudinal relaxation times
of the hydridic and NH resonances in anion 2 (4.7 T, CD2Cl2): } Ha,
n Hb, n Hc and ꢂ NH.
Table 3
Theoretical (DFT) and experimental (nOe) averaged interproton distances for anion 2
Ha ꢁ ꢁ ꢁ Hb
1
Ha ꢁ ꢁ ꢁ Hc
2
NH ꢁ ꢁ ꢁ Hb
1
NH ꢁ ꢁ ꢁ Hc
2
Me ꢁ ꢁ ꢁ Hb
6
Me ꢁ ꢁ ꢁ Hc
12
Multiplicity
Averaged DFT distance (rotamer A)
Averaged DFT distance (rotamer B)
Relative nOe volumes (CD2Cl2)
Derived distance (CD2Cl2)
2.448
2.442
0.35
2.729
2.755
0.30
2.132
4.591
0.85
2.1
2.563
3.449
0.59
2.5
3.615
2.967
0.17
3.7
2.846
2.703
1.00
3.1
Relative nOe volumes (acetone-d6)
Derived distance (acetone-d6)
0.35
0.38
0.30
2.5
0.27
2.8
0.34
3.3
1.00
3.1
Table 4
Temperature dependence of T1 values (ms) for the hydrides and the NH proton of anion 2 in CD2Cl2
T [K]
T1 (Ha)
T1 (Hb)
T1(Hc)
T1(NH)
175
179
183
193
203
223
253
273
143.4
138.1
138.7
146.3
167.3
250.1
462.7
662
127.2
120.2
118.3
119.7
134.3
191.9
355.0
493
116.5
115.1
114.7
118.9
133.2
197.9
375.0
525
149.5
129.1
121.4
114.0
121.4
168.4
307.3
427

M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
2049
calculate the distance rHꢀH between NH and Hb accord-
ing to Eq. (1) [11]. The resulting value is 2.24(2) A [12],
in good agreement with the values obtained from the
nOe data, the X-ray data and the theoretically derived
values,
average values computed for rotamers A and B. This
is in agreement with the presence in acetone-d6 of a sig-
nificant amount of both conformers, rotamer B now
being stabilized by a conventional hydrogen bond with
the oxygen atom of the solvent.
˚
RH–H ¼ ðT1Þꢀ1
H–H
ꢀ
ꢁ
l20c4Hꢀh2IðI þ 1Þ
40p2r6H–H
1
4
3. Conclusions
¼
sc
þ
:
ð1Þ
1 þ x02sc2 1 þ 4x02sc2
The structure of anion 2 has been characterized by
NMR spectroscopy, X-ray analysis and DFT computa-
tions. The experimental and theoretical data have clearly
shown that, both in the solid state and in solution, in the
preferred conformation the NH proton of the coordi-
nated amine is directed towards the hydride bridging
the opposite cluster edge.
In regard to the hydrides bridging the lateral edges
(Hc), the 2D NOESY experiment shows that they
strongly interact with both the NH and the methyl
groups. Therefore, being impossible to separate the con-
tributions of the two different groups of atoms to the
relaxation rate of these hydrides, no hydride to ligand
distance could be calculated.
1H NMR spectroscopy provide clear evidence of the
close approach of the NH proton to the Hb hydride.
Short distances between a proton and a hydride are usu-
ally taken as proof of the presence of a dihydrogen
bond. However, as previously stated, in the present case
the dimethylamine conformation might be mainly
dictated by the steric hindrance exerted by the diagonal
carbonyl ligands, rather then by an attractive proton–
hydride interaction.
DFT calculations and experimental results indicate
that this conformation, beside minimizing steric repul-
sions between the ligands coordinated to the cluster
core, is furthermore stabilized by a weak intra-molecular
N–H ꢁ ꢁ ꢁ (l-H)Re2 dihydrogen bond, whose estimated
enthalpy is as low as 2.2 kcal molꢀ1
.
In contrast to what is observed for the analogous
cluster anion containing an axial pyrazole ligand, anion
2 is very stable at room temperature in dichloromethane
solution. Indeed, the weak acidity of the proton donor,
reflected by the very low DHꢁ value, and the bridging
coordination of the proton acceptor, that reduces its
thermodynamic and kinetic ‘‘hydricity’’, do not promote
the protonation pathway leading to H2 evolution.
Evidence that the short NH ꢁ ꢁ ꢁ (l-H)Re2 distance in
2 is not due to steric constraints only but also to the
presence of a true dihydrogen bond is provided by IR
studies. The presence of a hydrogen bond usually causes
a red-shift and a strong broadening of the stretching
absorption of the X–H hydrogen bond donor. In this
case, the m(NH) IR band of the coordinated amine of
2 appears at 3262 cmꢀ1, approximately a 100 cmꢀ1 shift
to lower wavenumbers with respect to the literature va-
lue for pure DMA in a solvent not able to act as proton
acceptor (3361 cmꢀ1 in n-hexane) [13]. An hydrogen
bonding interaction is therefore shown to be present in
2, even if it is rather weak. From the IR data it is possi-
ble to calculate the enthalpy associated with the hydro-
gen bond formation, through the empirical Iogansen
equation (Eq. (2)) [14]. This correlates the change in po-
sition of the stretching band of the X–H bond (Dm) with
the hydrogen bond enthalpy. In this case the measured
4. Experimental
The reactions were performed under N2, in solvents
dried and deoxygenated by standards methods. Dimeth-
ylamine solution (2 M in THF, Aldrich) was used as
received. The starting compound [PPh4][Re3(l3-H)(l-
H)3(CO)9] was prepared as described in the literature
[15]. The NMR spectra were acquired on Bruker
AC200 and DRX300 spectrometers and the IR spectra
on a Bruker Vector 22FT instrument.
4.1. Synthesis of [PPh4][Re3(l-
H)4(CO)9(DMA)]([PPh4]2)
Dm value corresponds to DHꢁ = ꢀ2.2 kcal molꢀ1
,
ꢀDHꢃ ¼ 18Dm=ðDm þ 720Þ:
ð2Þ
A sample (18.6 mg, 0.0161 mmols) of [PPh4][Re3(l3-
H)(l-H)3(CO)9] was dissolved at room temperature in
freshly distilled CH2Cl2 (5 mL). Addition of dimethyl-
amine (2 M in THF, 10 lL, 0.020 mmols) caused a
change in the solution colour. IR (CH2Cl2):
m(CO) = 2034 m, 1999vs, 1912vs cmꢀ1. Addition of n-
hexane gave a yellow precipitate that was dried under
vacuum (15.3 mg, 0.0125 mmols, isolated yields 78%).
Slow diffusion of n-hexane in a CH2Cl2 solution of 2
Further evidence that the proton–hydride interaction
significantly contributes to the preferred dimethylamine
rotamer present in CH2Cl2 is provided by the observa-
tion that in a solvent able to act as proton acceptor, dif-
ferent rotamers are present. A 2D NOESY experiment
performed at 273 K in acetone-d6 shows that Hb has a
stronger cross-peak with the Me hydrogen atoms than
with NH (see Table 3). The nOe derived distances for
the (NH,Me) ꢁ ꢁ ꢁ (Hb,Hc) interactions lie between the
1
afforded crystals suitable for X-ray analysis. H NMR

2050
M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
(CD2Cl2, 300 K) d = 2.79 (d, 6, CH3, JHH = 6.3 Hz),
1.82 (septuplet, 1, NH), ꢀ7.92 (s, 1, Ha), ꢀ9.26 (s, 1,
Hb), ꢀ9.70 (s, 2, Hc).
www.ccdc.cam.ac.uk/conts/retrieving.html or from the
Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK (fax: (+44)1223-336-
033; e-mail: deposit@ccdc.cam.ac.uk).
4.2. Variable temperature T1 and two-dimensional
NOESY experiment on 2 in CD2Cl2
4.4. Computational study
A solution of [PPh4]2 (21.1 mg, 0.0172 mmols) in
CD2Cl2 (0.5 mL) in an NMR tube was degassed through
repeated freeze-thaw cycles. 1H NMR variable tempera-
ture spectra were acquired on a 4.7 T instrument from
175 to 273 K, and at each temperature the 1H T1 values
were obtained by a three-parameter fit of the intensities
of the signals, in spectra recorded with the standard
nonselective inversion recovery pulse sequence, with 14
variable delays. The results are reported in Table 4.
All the calculations were done using an empirically
parameterised density functional theory (DFT) method
incorporating BeckeÕs three-parameter hybrid functional
along with the Lee–Yang–Parr correlation functional
(B3LYP) [20]. A basis set incorporating the ‘‘small core’’
relativistic effective core potentials (ECP) of Hay and
Wadt was used for the rhenium atoms along with valence
double-f functions [21] augmented with an energy-opti-
mised set of 5p functions [22], yielding a final contraction
of (341/341/21), similar but different from the basis set
LanL2DZ included in GAUSSIAN 03 [23]. The standard
valence double-f basis set with a single polarization func-
tion 6-31G(d,p) [24] was used for all remaining atoms.
Geometries were optimised in redundant internal coordi-
nates [25], employing the GDIIS algorithm [26], until the
maximum (root-mean-square) force was less than
0.00045 (0.00030) a.u. The energy profile shown in Fig.
3 was computed freezing the (improper) torsion angle
x to the values from 20ꢁ to 160ꢁ in steps of 20ꢁ and allow-
ing all the remaining internal coordinates to relax. Calcu-
lations were done without the imposition of any
symmetry except for the two proposed minima (rotamer
A and B) which were found to possess Cs symmetry.
Analytic frequency computations were done for the
two proposed minima (rotamer A and B), for free dim-
ethylamine and for the dimethylamine dimer. A complete
list of coordinates and frequencies can be requested of
the authors (P.M.). All the computations were per-
formed with GAUSSIAN 03 [23].
1
The same sample was also used for the H two-dimen-
sional NOESY phase-sensitive experiment shown in
Fig. 4. It was performed at 273 K, on a 4.7 T field, with
s
m = 0.3 s, 8 FIDs, sweep width = 4347 Hz, 1 K data
points for 256 experiments. Shifted sine-bell functions
were applied in both dimensions before the Fourier
transform and after zero-filling to 1 K in F1.
4.3. X-ray diffraction structural analysis
Crystal data for [PPh4]2: [C24H20P]+[C11H11NO9-
Re3]ꢀ = C35H31NO9PRe3, Mr = 1199.18, monoclinic,
space group P21/c (No. 14), a = 14.344(4), b =
˚
17.417(4), c = 15.488(4) A, b = 91.18(1)ꢁ, V =
3868.5(17) A , Z = 4, T = 295(2) K, graphite-monochro-
3
˚
˚
mated Mo Ka radiation (k = 0. 71073 A), qcalcd = 2.059
g
cmꢀ3
,
F(000) = 2240, colourless crystal 0.24 ·
0.18 · 0.12 mm3, l(Mo Ka) = 9.455 mmꢀ1, empirical
absorption correction (SADABS [16], 28470 symmetry
equivalent reflections, effective data to parameters ratio:
9.3), minimum/maximum transmission factors 0.166/
0.322, Bruker SMART diffractometer,
x
scans
(Dx = 0.3ꢁ), 2400 frames each at 20 s exposure keeping
the detector at 5.0 cm from the sample,
References
[1] J.C. Lee, A.L. Rheingold, B. Muler, P.S. Pregosin, R.H. Crabtree,
Chem. Commun. (1994) 1021.
1.8 6 h 6 25.1ꢁ,
index
ranges
h = ꢀ17 ! 17,
k = ꢀ20 ! 20, l = ꢀ18 ! 18, 45369 reflections of which
6854 unique (Rint = 0.0297, Rr = 0.0187), 5865 reflec-
tions with I > 2r(I), intensity decay 4%, solution by di-
rect methods (SIR97 [17]) and subsequent Fourier
synthesis, anisotropic full-matrix least-squares on F2
using all reflections (SHELX97 [18]), hydrogen atoms
placed in idealised position [19], phenyl groups refined
as variable metric regular hexagons, data/para-
meters 6854/401, S(F2) = 1.046, R(F, I > 2r(I)) =
0.0424, wR(F2, all data) = 0.1145, weighting scheme
[2] L.M. Epstein, E.S. Shubina, Coord. Chem. Rev. 231 (2002) 165.
[3] (a) S. Aime, R. Gobetto, E. Valls, Organometallics 16 (1997)
5140;
´
(b) S. Aime, M. Ferriz, R. Gobetto, Organometallics 18 (1999)
2030;
(c) S. Aime, E. Diana, R. Gobetto, M. Milanesio, E. Valls, D.
Viterbo, Organometallics 21 (2002) 50.
[4] T. Beringhelli, G. DÕAlfonso, M. Panigati, P. Mercandelli, A.
Sironi, Chem. Eur. J. 8 (2002) 5340.
[5] (a) T. Beringhelli, G. DÕAlfonso, Chem. Commun. (1994) 2631;
(b) T. Beringhelli, G. DÕAlfonso, M. Panigati, F. Porta, P.
Mercandelli, M. Moret, A. Sironi, Organometallics 17 (1998)
3282.
2
w ¼ 1=½r2ðF o2Þ þ ð0:085PÞ þ 26Pꢄ, where P ¼ ðF 2oþ
2F 2cÞ=3, maximum/minimum residual electron density
[6] (a) T. Beringhelli, G. Ciani, G. DÕAlfonso, H. Molinari, A.
Sironi, Inorg. Chem. 24 (1985) 2666;
2.778/ꢀ2.037 e Aꢀ3. CCDC-253668 ([PPh4]2) contains
˚
the supplementary crystallographic data for this paper.
This data can be obtained free of charge via
(b) G. Ciani, G. DÕAlfonso, M. Freni, P. Romiti, A. Sironi, A.
Albinati, J. Organomet. Chem. 136 (1977) C49;

M. Panigati et al. / Journal of Organometallic Chemistry 690 (2005) 2044–2051
2051
(c) T. Beringhelli, G. DÕAlfonso, M. Freni, G. Ciani, A. Sironi,
Dalton Trans. (1986) 2691.
[14] A.V. Iogansen, Hydrogen Bond, Nauka, Moscow, 1981, cited in;
N.V. Belkova, E.S. Shubina, A.V. Ionidis, L.M. Epstein, H.
Jacobsen, A. Messmer, H. Berke, Inorg. Chem. 36 (1997)
1522.
[7] The hydrido ligand was placed in a potential energy optimised
position using the program HYDEX: A.G. Orpen, Dalton Trans.
(1980) 2509. The N–H bond distance was fixed to the average
[15] T. Beringhelli, G. DÕAlfonso, M.G. Garavaglia, Dalton Trans.
(1996) 1771.
˚
value found for aliphatic amine by neutron diffraction (1.025 A):
CSD (version 5.24, November 2002). F.H. Allen, Acta Crystal-
logr., Sect. B: Struct. Sci. 58 (2002) 380.
[16] G.M. Sheldrick, SADABS, Universita¨t Go¨ttingen, Germany, 1996.
[17] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C.
Giacovazzo, A. Guagliardi, A.G.G. Moliterni, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 32 (1999) 115.
[8] The
averaged
distance
has
been
computed
as
Ædæ = ꢂd(q)eꢀE(q)/RTdq/ꢂeꢀE(q)/RTdq in which E(q) = 1/2kq2.
[9] To average the equivalent interproton distances the TroppÕs
equation has been employed. D. Neuhaus, M.P. Williamson,
The Nuclear Overhauser Effect in Structural and Conforma-
tional Analysis, second ed., Wiley-VCH, New York, 2000, pp.
167–178.
[18] G.M. Sheldrick, SHELX97, Universita¨t Go¨ttingen, Germany, 1997.
[19] A.G. Orpen, Dalton Trans. (1980) 2509.
[20] (a) A.D.J. Becke, Chem. Phys. 98 (1993) 5648;
(b) C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[21] P.J. Hay, W.R.J. Wadt, Chem. Phys. 82 (1985) 299.
[22] M. Couty, M.B. Hall, J. Comput. Chem. 17 (1996) 1359.
[23] GAUSSIAN 03 (Revision B.04), Gaussian Inc., Pittsburgh, PA,
2003.
[10] S. Aime, E. Diana, R. Gobetto, M. Milanesio, E. Valls, D.
Viterbo, Organometallics 21 (2002) 50, and refs therein.
[11] A. Abragam, The Principles of Nuclear Magnetism, Oxford
University Press, Oxford, UK, 1973 (Chapter 8).
[24] W.J. Hehre, L. Radom, P.v.R. Schleyer, J.A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, New York, 1986.
[25] C. Peng, P.Y. Ayala, H.B. Schlegel, M.J. Frisch, J. Comput.
Chem. 17 (1996) 49.
[12] The standard deviation on the computed distance were obtained
by propagation of errors starting from the T1 values 146.3(10) and
119.7(4) ms for Ha and Hb at 193 K.
´ ´
[26] P. Csaszar, P. Pulay, J. Mol. Struct. 114 (1984) 31.
[13] H. Wolff, G. Gamer, J. Phys. Chem. 76 (1972) 871.