Hydrogen bonding and proton transfer to the trihydride complex [Cp*MoH3(dppe)]: IR, NMR, and theoretical investigations

FULL PAPER

DOI: 10.1002/ejic.200600093

Hydrogen Bonding and Proton Transfer to the Trihydride Complex

[Cp*MoH₃(dppe)]: IR, NMR, and Theoretical Investigations

Natalia V. Belkova,^[a]Pavel O. Revin,^[a]Maria Besora,^[b]Miguel Baya,^[c]Lina M. Epstein,^[a]

Agustí Lledós,*^[b]Rinaldo Poli,*^[c]Elena S. Shubina,*^[a]and Evgenii V. Vorontsov^[a]

Keywords: Dihydrogen bonding / Hydride ligands / Molybdenum / Proton-transfer mechanism / DFT calculations

The interaction between [Cp*MoH₃(dppe)] (dppe

=

[Cp*MoH₃(dppe)]–HFIP reaction. For all proton-transfer pro-

cesses investigated, the classical tetrahydrido cation forms

directly, without the observation of a nonclassical intermedi-

ate. DFT calculations have been carried out on the interac-

tion between TFE and HFIP and the model compound

[CpMoH₃(dpe)] (dpe = H₂PCH₂CH₂PH₂) both in the gas

phase and in CH₂Cl₂solvent with the polarizable continuum

Ph₂PCH₂CH₂PPh₂) and variety of proton donors has

a

been investigated by a combination of experiments and

DFT calculations. Weak proton donors [2-monofluoroethanol

(MFE) and trifluoroethanol (TFE)] allow the determination of

basicity factor (E_j

=

1.42 0.02) and thermodynamic

parameters for the hydrogen bond formation (∆H_HB

=

–4.9 0.2 and –6.1 0.3 kcalmol^–1; ∆S_HB= –15.7 0.7 and

–20.4 1 calmol^–1K^–1for MFE and TFE, respectively). For

TFE, a stable low-temperature proton-transfer equilibrium

(220–240 K) with the cationic classical tetrahydrido deriva-

tive [Cp*MoH₄(dppe)]⁺could be investigated independently

model and, to a more limited extent, on the full

[Cp*MoH₃(dppe)] system. A detailed comparison of the ob-

served and calculated frequency shifts for the M–H vi-

brations is presented. The calculations have explored the

relative energy and geometry of various configurations in-

volving either a hydride ligand or the metal as the principal

proton-accepting site. They have also probed two principal

proton-transfer pathways, leading to the unobserved non-

classical intermediate and to the observed classical product.

From these studies, it appears that a nonclassical intermedi-

ate may be obtained by a kinetically controlled proton trans-

fer to a hydride site, followed by an intramolecular re-

arrangement through a very low energy barrier. However, a

competitive low-energy pathway for direct proton transfer at

the metal site is also revealed by the calculations.

by UV/Vis (∆H°_PT

=

–2.8 0.4 kcalmol^–1and ∆S°_PT

=

–15 2 calmol^–1K^–1

)

and ¹H NMR (∆H°_PT

=

–2.7 0.5

kcalmol^–1and ∆S°_PT= –11 2 calmol^–1K^–1) spectroscopy.

Upon warming, however, the tetrahydride evolves by dihy-

drogen loss and formation of a hydride-free diamagnetic

product. Stronger proton donors [hexafluoroisopropanol

(HFIP), p-nitrophenol (PNP), perfluoro-tert-butyl alcohol

(PFTB), and HBF₄·OEt₂] lead to more extensive proton trans-

fer at lower donor/Mo ratios. A 1:1 proton-transfer stoichiom-

etry is indicated independently by a titration experiment

with UV/Vis monitoring for the [Cp*MoH₃(dppe)]–PNP reac-

tion, and by a stopped-flow kinetics investigation for the

Germany, 2006)

portance for catalysis and in biochemistry.^[1,2]The kinetic

preference for hydride versus metal protonation is now

quite firmly established,^[3–9]though exceptions have recently

been reported from studies carried out in our laborato-

ries.^[10,11]The proton transfer occurs via intermediates, for

which characteristic spectroscopic signatures have been es-

tablished, that contain hydrogen bonds between the proton

donor and the proton acceptor (the metal center or a hy-

dride ligand) (see Scheme 1).^[12,13]The term “nonclassical

Introduction

Proton-transfer processes to and from transition-metal

centers and hydride ligand sites continue to attract con-

siderable research interest because of their fundamental im-

[a] Nesmeyanov Institute of Organoelement Compounds, Russian

Academy of Sciences,

Vavilov Street 28, 119991 Moscow, Russia

E-mail: shu@ineos.ac.ru

[b] Departament de Química, Edifici Cn, Universitat Autónoma de hydrogen bonding” has been coined to address these inter-

Barcelona,

actions, while H-bonding specifically involving a hydride li-

08193 Bellaterra, Spain

gand has also been termed “dihydrogen bonding”.^[14]The

E-mail: agusti@klingon.uab.es

[c] Laboratoire de Chimie de Coordination, UPR CNRS 8241, lié

par convention à l’Université Paul Sabatier et à l’Institut

national Polytechnique de Toulouse,

relationship between the preferred thermodynamic site of

hydrogen bonding (i.e., the relative energy of species I and

IV) and the kinetic site of proton transfer (i.e., the relative

energy barrier leading to species II and V) is of interest to

us. We have recently presented combined experimental and

theoretical results showing that the preferred hydrogen-

205 Route de Narbonne, 31077 Toulouse Cedex, France

Fax: +33-561553003

E-mail: poli@lcc-toulouse.fr

Supporting information for this article is available on the

WWW under http://ww w .eurjic.org or from the author.

2192

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

Scheme 1.

bonding site for compound [Cp*Fe(dppe)H], that is, the hy-

dride site, corresponds to the site of preferred proton trans-

fer.^[15]In parallel work, some of us have studied the mecha-

nism of the hydride protonation, analyzing the influence of

medium polarity and concentration effects, both experimen-

tally and theoretically.^[16–18]We have demonstrated the in-

volvement of ion pairs formed by cationic dihydrogen com-

plexes and the anionic conjugate basis of the proton donor.

In these studies the stabilizing role of the solvent and the

homoconjugate anions [A···H···A]^–in the protonation pro-

cess has also been recognized.

Scheme 2.

In order to determine the general relationship between

the thermodynamics of hydrogen bonding to various pos-

sible sites and the energy barrier leading from each of these

sites to proton transfer, a greater number of case studies is

necessary, especially for polyhydride complexes where the

hydride ligands occupy inequivalent positions so that it is

also possible in principle to discriminate between different

hydride sites. For this reason, we have selected compound

[Cp*Mo(dppe)H₃], which was synthesized in one of our

laboratories a few years ago.^[19]In this system there are four

basic potential sites (the metal and the three hydride li-

gands) able to establish hydrogen bonds with a proton do-

nor. Moreover, several isomers could result from the pro-

ton-transfer process. The X-ray structure of the trihydride

compound was of sufficient quality to locate the hydride

ligands and showed that they occupy two different sites, as

shown in Scheme 2.^[20]The geometry is best described as an

ideal trigonal prism if the Cp* ligand is considered to oc-

cupy a single position in correspondence to its ring

centroid. However, the compound is highly fluxional and a

single hydride resonance, which is split into a binomial trip-

let by the two equivalent P nuclei, remained sharp down to

183 K.^[19]

analogous W derivative displays a pseudopentagonal bipy-

ramidal structure. As the two metal complexes show analo-

gous NMR properties, the same structure is also assumed

to be adopted by the Mo complex and is illustrated in

Scheme 2. Both Mo and W cationic tetrahydride complexes

are fluxional at room temperature, showing only one hy-

dride resonance and one phosphorus resonance, but the P

resonance decoalesces at low temperature, whereas the hy-

dride resonance does not. The protonation of

[Cp*Mo(dppe)H₃] in MeCN or the dissolution of

[Cp*MoH₄(dppe)]⁺in MeCN gives a variety of products

resulting from the substitution of H₂with MeCN. The

tungsten compound is stable under these conditions.^[19]

Preliminary studies of the hydrogen bonding and proton

transfer to the Cp*M(dppe)H₃(M = Mo, W) compounds

with weaker proton donors has been recently communi-

cated.^[21]We present here a full report of our investigation

on the Mo compound, whereas further details concerning

the W system, including a comparison of the two metal

complexes, will be reported in a forthcoming paper.^[22]

Protonation studies have previously been carried out

only with HBF₄, leading to the formal product of metal

Results

protonation,

the

classical

tetrahydride

complex

Experimental Studies

[Cp*MoH₄(dppe)]⁺.^[19]However, this product is unstable

and easily loses H₂. It could be isolated upon low-tempera-

Two solvents – THF or CH₂Cl₂– were used for spectro-

ture protonation in ether, where it precipitates. An X-ray scopic studies. Their relatively high polarity (ε = 7.3 for

structure was not obtained for the Mo complex, but the THF and 8.9 for CH₂Cl₂) helps maintain the ionic proton-

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2193

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

transfer products in solution. However, THF has itself pro-

The hydrogen bond enthalpies, ∆H_HB, were obtained by

nounced proton-acceptor properties and competes with the two independent methods (see Table 1). The first one is

hydride complex for hydrogen bonding to the proton donor. based on the empirical correlation outlined in Equation (1),

Therefore, much higher proton donor concentrations are originally established for classical H bonds and later ex-

needed in order to observe similar spectral changes to those tended to hydrogen bonds to metal centers and to hydride

in CH₂Cl₂.^[23]Unfortunately, compound [Cp*Mo(dppe)H₃] ligands as proton acceptors.^[12,13]The van’t Hoff method

slowly decomposes at room temperature in dichlorometh- uses the integrated IR intensities for the free alcohol ab-

ane, probably by attack of the solvent C atom by the hy- sorption band in the presence or absence of the hydride

dride ligand. The stability of the trihydride complex in complex as a function of temperature. It has the advantage

CH₂Cl₂is greatly enhanced at low temperatures, with no of also yielding the reaction entropy. There is quite good

visible change occurring at 200 K for at least 3 h. Thus IR agreement between the reaction enthalpies obtained from

and UV/Vis studies were carried out either in THF or in the two methods. It should be noted that the computed val-

CH₂Cl₂at low temperatures, or in CH₂Cl₂at higher tem- ues for the interaction with TFE (see later) are in good

peratures within a short timescale relative to the decompo- agreement with these experimental results.

sition reaction. The NMR studies were carried out at low

temperatures in CD₂Cl₂. In all cases, the extent of decom-

position was carefully monitored to insure the significance

of the results. The proton donors used in this study were

2-monofluoroethanol (MFE), 2,2,2-trifluoroethanol (TFE),

hexafluoroisopropanol (HFIP), perfluoro-tert-butyl alcohol

(PFTB), and p-nitrophenol (PNP).

Analysis of the ν_MHRegion in the IR Spectra

Analysis of the ν_OHRegion in the IR Spectra

The IR analysis in the ν_MHstretching vibration range

The interaction between [Cp*Mo(dppe)H₃] and proton

donors was first studied by IR in the ν_OHrange by using

MFE and TFE in CH₂Cl₂. The observed decrease of the

intensity (A) of the ν_OH(free)band and the appearance of a

low-frequency shifted broad ν_OHband for the hydrogen-

bonded OH group is unquestionable evidence for the for-

mation of hydrogen bonds between the alcohol OH proton

and the transition-metal hydride complex, although it does

not establish the nature of the hydrogen-bonding site. The

∆ν_OHband shift [∆ν_OH= ν_OH(free)– ν_OH(bonded)] increased

with the strength of the proton donors (see Figure 1 and

Table 1).

was carried out with the goal of learning about the nature

of the preferred hydrogen-bonding site. Previous studies on

monohydride complexes have demonstrated that the inter-

action with a hydride ligand has the effect of shifting the

M–H stretching vibration to lower frequencies, whereas the

opposite effect is associated to an interaction with the metal

site or other ligands.^[12,13]For polyhydride compounds,

however, the situation may not be so simple, as each vi-

brational normal mode results from the combination of

more M–H bond stretching vibrations, each of which may

be affected by hydrogen bonding in a different way. A pre-

vious combined experimental and theoretical study on the

dihydride complex Ru(PP₃)H₂[PP₃= P(CH₂CH₂PPh₂)₃]^[23]

was simplified by the fact that the two M–H bonds, being

trans to ligands that exert a very different trans effect, do

not mix extensively with each other, thus each of the two

bands is composed of an essentially pure stretching vi-

bration of a single M–H bond. The study of hydrogen

bonding implicating the Cp₂NbH₃complex,^[18]having C_2v

symmetry and one ν_MHvibrational band resulting from the

overlap of 2a₁+ b₁vibrations,^[24]revealed the formation of

dihydrogen-bonded complexes leading to band splitting

with a low-frequency shift for the hydrogen-bonded ν_MH

band. Although hydrogen bonding for other polyhydride

compounds has been investigated,^[11]no detailed analysis of

the normal modes and how these are affected by hydrogen

bonding have been reported. We intend to carry out such

Figure 1. IR spectra in the ν(R_FO–H) stretching region of MFE

(1) or TFE (2) in CH₂Cl₂in the presence of [Cp*Mo(dppe)H₃].

Table 1. Parameters of the dihydrogen bonding between [Cp*Mo(dppe)H₃] and MFE or TFE, and basicity factors.

[a]

[b]

ROH

ν_OH(free)

ν_OH(bonded)

∆ν

∆H_HB

∆S_HB

E_j

cm^–1

kcalmol^–1

calmol^–1K^–1

MFE

TFE

3612

3604

3368

3248

244

364

–4.9

–5.9

–4.9 0.2

–6.1 0.3

–15.7 0.7

1.44

1.41

–20.4

1

[a] Calculated by the empirical relationship of Equation (1). [b] From van’t Hoff’s method.

2194

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

an analysis here for the [Cp*Mo(dppe)H₃] complex, by a tetrahydride complex [Cp*Mo(dppe)H₄]⁺.^[19]New ¹H

combination of experimental and computational methods. NMR investigations carried out at low temperatures (200–

The trihydride complex displays wide ν_MHbands of com- 250 K) in CD₂Cl₂using either HBF₄·Et₂O or TFE show a

plex shape with the maxima at 1772 cm^–1in CH₂Cl₂and at direct transformation, without the detection of any interme-

1785 cm^–1in THF. Deconvolution of the spectrum in diate (notably nonclassical species). With the weaker proton

CH₂Cl₂, after subtraction of overtones because of the donor TFE, an equilibrium between the starting

phenyl groups, yields two overlapping bands at 1821 and [Cp*Mo(dppe)H₃] complex and the protonated product is

1772 cm^–1(∆ν_1/2= 57 and 54 cm^–1, correspondingly) (see established and a large excess is needed to achieve a high

Figure 2). A third, much weaker band is found at 1904 cm^–1degree of proton transfer.

(∆ν_1/2= 41 cm^–1). However, its parameters are less reliable

Well-established, free, and hydrogen-bonded complexes

because of the low intensity and superposition with one of exchange rapidly on the NMR timescale, yielding only one

the phenyl overtones. Therefore, it will not be used later for resonance. Typically, the effect of hydrogen bonding on

the comparison with the theoretical data on ν_MHin hydro- chemical shifts is small, especially in the case of polyhyd-

gen-bonded complexes. The intensities of the ν_MHbands rides, where the effect is averaged over the various hydride

increase as the temperature is decreased. The extinction co- ligands. The [Cp*Mo(dppe)H₄]⁺complex on one side, and

efficient of the strongest band increases from 49.0 to the rapidly equilibrating mixture of the free trihydride com-

60.3 Lmol^–1cm^–1upon cooling from 290 to 200 K in THF. plex and the hydrogen-bonded adduct [Cp*Mo(dppe)-

H₃···HOR_F] on the other side, give rise to triplet resonances

at about δ –3.7 (J_HP= 37.4 Hz) and –6.0 (J_HP= 48.0 Hz)

in the ¹H NMR spectrum at low temperatures (190–210 K).

The positions of these peaks change significantly with the

temperature, moving to about –3.5 and –5.7 at 300 K. The

chemical shift difference between these resonances de-

creases steadily from 2.32 ppm at 200 K to 2.22 ppm at

300 K. On the other hand, neither chemical shift, at a given

temperature, seems highly dependent on the nature of the

proton donor (TFE, HFIP, or HBF₄) or on its concentra-

tion. Furthermore, the J_HPvalues remain practically un-

changed over the entire temperature range. Therefore, the

chemical shift changes are probably unrelated to variations

in hydrogen-bonding equilibria. The relative amounts of the

Figure 2. IR spectrum in the ν(Mo–H) stretching region of

[Cp*Mo(dppe)H₃] (0.030 ) in CH₂Cl₂, after subtraction of the

phenyl overtones, showing the band decomposition.

two species are also afforded by a parallel monitoring of

the ³¹P{¹H} NMR spectrum (resonance at –75.0 ppm and

1

–92.8 ppm, respectively). As for the case of the H NMR

Upon interaction between the [Cp*Mo(dppe)H₃] com-

plex and MFE at 200 K in CH₂Cl₂the ν_MHband becomes

wider and shifts to lower frequencies by about 5 cm^–1. A

band decomposition analysis carried out after the subtrac-

tion of both dppe and MFE spectra shows that the two

major bands become wider and appear at different posi-

tions: 1822 and 1769 cm^–1(∆ν_1/2= 73 and 60 cm^–1, ∆ν =

+1 and –3 cm^–1, respectively). These changes can be attrib-

uted to the effect of hydrogen bonding on the MoH₃moi-

ety. The experiments in THF, as expected, necessitated the

use of the stronger proton donors at higher concentrations

in order to observe similar changes. For instance, using just

a 10-fold excess of TFE, the only observed change was a

small low-frequency band shift and intensity decrease upon

increasing the temperature, just like in the absence of

alcohol, without a shape change. This change is reversible

up to 270 K, demonstrating that no proton transfer occurs

with this alcohol in THF up to 270 K (cf. proton transfer

in CH₂Cl₂later). A deeper analysis of the M–H stretching

vibrations in the absence and presence of proton donors

will be carried out below in the light of the theoretical data.

resonances, the ³¹P{¹H} NMR chemical shifts do not

depend on the nature and concentration of the proton

donor.

The establishment of hydrogen bonding is more clearly

evidenced from the change of longitudinal relaxation time

(T₁) for the trihydride resonance (see Table 2). As expected,

this parameter decreases upon addition of proton donors,

in proportion to the strength of the donor.^[2,13]It is also

significant that the temperature corresponding to T_1min

increases slightly upon protonation from 220 to 230 K.

This phenomenon signals a decreased correlation time

(τ_C), that is, a slower tumbling motion, for the protonated

species, in agreement with its expected larger size. Thus,

slightly higher temperatures are needed to again raise τ_C

to the conditions required for the most efficient long-

2

itudinal relaxation (w₀τ_CϷ 1). The T_1minof the tetra-

hydride product is smaller relative to the parent complex,

but not in the range expected for dihydrogen complexes.^[25]

This decrease is expected, as the dipolar relaxation im-

plicates four hydride ligands instead of three. Note, how-

ever, that the tetrahydride resonance T₁value is also sen-

sitive, like that of the parent trihydride, to the nature of

NMR and IR Studies of Proton Transfer

As stated in the Introduction, the molybdenum trihyd- the proton donor used. A possible interpretation of this

ride complex can be protonated by HBF₄·Et₂O to give the difference is a stronger hydrogen bonding between one or

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2195

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

more hydride ligands with the stronger conjugate base spectroscopy because of the complex shape of the spectra,

CF₃CH₂O^–, thereby increasing the average H···H further complicated by the temperature dependence of the

distances.

extinction coefficients. The thermodynamic parameters of

this equilibrium were obtained from UV and NMR investi-

gations as shown below.

Table 2. Longitudinal relaxation time T_1min(ms) for [Cp*Mo-

(dppe)H₃] in CD₂Cl₂solution under different conditions.^[a]

T_min[K]

[Cp*Mo(dppe)H₃]

[Cp*Mo(dppe)H₄]⁺

Free

+HBF₄(1 equiv.)

+TFE (6 equiv.)

220

230^[b]

302

274

295

174

191

[a] At 400 MHz. [b] The resonances of both complexes have a mini-

mum T₁at the same temperature.

The complex [Cp*Mo(dppe)H₄]⁺was also characterized

by means of IR spectroscopy (see Figure 3). After the ad-

dition of 1.1 equiv. of HBF₄, new bands appeared in the

spectrum at 1818, 1839, and 1920 cm^–1(see spectrum b).

Formation of bands at higher frequency upon protonation

is spectroscopic evidence for the formation of a cationic

classical product. Treating [Cp*Mo(dppe)H₃] with 7 equiv.

of HFIP in THF at 200 K did not lead to significant proton

transfer, only a minor decrease of the initial 1785 cm^–1band

observed after the subtraction of the HFIP spectrum (see

spectrum c).

Figure 4. IR spectra in the ν(Mo–H) stretching region of

[Cp*Mo(dppe)H₃] (0.037 ) in CH₂Cl₂at 200 K. (a) Without TFE.

(b) With TFE (2 equiv.). (c) With TFE (5 equiv.). (d) With TFE

(10 equiv.).

Establishment of the Proton-Transfer Stoichiometry

The equilibrium resulting from the interaction between

[Cp*Mo(dppe)H₃] and p-nitrophenol (PNP; P_i= 1.27) was

investigated by UV/Vis spectroscopy. This was rendered

possible by the sensitivity of the proton donor chromo-

phore to hydrogen bonding and proton transfer,^[26]making

it possible to probe the nature of the species in equilibrium.

However, quantitative information on the concentration of

each species was not accessible. Spectra were recorded for

CH₂Cl₂solutions of PNP (0.001 ) in the presence of

[Cp*Mo(dppe)H₃] at different ratios from 1:0.1 to 1:2, in

the 200–250 K temperature range, where the observed spec-

tral changes were fully reversible. The spectra show wide

overlapping bands of both the phenol in its various forms

and the hydride complexes (both free and dihydrogen

bonded). The absence of free phenolate is signaled by the

absence of a band at 430 nm. A deconvolution analysis

Figure 3. IR spectra in the ν(Mo–H) stretching region of

[Cp*Mo(dppe)H₃] (0.084 ) in THF at 200 K. (a) Without proton

donor. (b) With HBF4 (1.1 equiv.). (c) With HFIP (7 equiv.) after

subtraction of the HFIP spectrum.

The protonation process was studied more extensively yields three bands with maxima at 312, 351, and 395 nm.

with a wider range of fluorinated alcohols of different The first two bands are assigned to free PNP and to the

strength in dichloromethane by variable temperature IR. dihydrogen-bonded complex [Cp*(dppe)MoH₃]···HOC₆H₄-

An equilibrium between the hydrogen-bonded system and NO₂. Note that the 351 nm band is red-shifted not only

the tetrahydride complex was observed, which shifts toward relative to free PNP (∆λ = 39 nm), but also to the related

the protonation product for stronger alcohols and higher iron complex [Cp*(dppe)FeH]···HOC₆H₄NO₂(340 nm, ∆λ

alcohol/Mo ratios. For PFTB, the complete disappearance = 11 nm),^[27]signaling a stronger hydrogen bonding. This is

of the trihydride occurs in the presence of a twofold excess consistent with the higher basicity factor of [Cp*Mo-

at 200 K. The addition of 2 equiv. of TFE leads to only (dppe)H₃] (Table 1) versus that of [Cp*Fe(dppe)H] (1.35–

20% conversion, while a 10-fold excess is required to con- 1.38).^[27]The band at 395 nm is attributed to a hydrogen-

sume nearly all the hydride precursor (see Figure 4). For the bonded phenolate ion, [Cp*(dppe)MoH₄]⁺···[OAr]^–, be-

case of the weakest fluorinated alcohol MFE, a 50-fold ex- cause this is blue-shifted from the free phenolate band by

cess leads only to a moderate decrease of the MoH band 35 nm. Notably, this band is red-shifted in comparison to

intensity. The proton transfer with TFE at variable tem- the corresponding band previously attributed to [Cp*(dppe)-

peratures has given kinetic information (see later). Above Fe(H₂)]⁺···[ArOHOAr]^–.^[15]The assignment of this absorp-

250 K, a slow evolution of the spectroscopic properties was tion to the 1:1 hydrogen-bonded ion pair, rather than to a

observed, indicating instability for the tetrahydride product species containing a homoconjugated pair [ArOHOAr]^–,

(see below). A quantitative measurement of the equilibrium was confirmed by the titration experiment (see Figure 5).

position in a wide temperature range was not possible by IR Therefore, the proton-accepting strength of the Cp*(dppe)-

2196

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

MoH₃complex is sufficiently high to abstract a proton on [HFIP] was observed under the same (K_HB[HA] ϾϾ 1)

from a single molecule of PNP.

conditions. This difference is again consistent with the

higher proton-accepting ability of [Cp*Mo(dppe)H₃] versus

[Cp*Fe(dppe)H].

Thermodynamics of the Proton-Transfer Equilibrium

The equilibrium of proton transfer to [Cp*Mo(dppe)H₃]

was investigated by UV/Vis and ³¹P NMR spectroscopy in

the 200–240 K temperature range using TFE as proton do-

nors. The results of the UV/Vis study using TFE are shown

in Figure 7. The spectra of the initial trihydride (a) and final

tetrahydride (b) complexes have very different extinction co-

efficients at the λ_maxof the trihydride species (400 nm): 2128

and 245, respectively. The UV/Vis properties of the dihydro-

gen-bonded complexes are essentially indistinguishable

from those of the free hydride complex: the UV/Vis spectra

of [Cp*Mo(dppe)H₃] (0.02 ) in the presence of 15 equiv.

MFE do not differ significantly from those of pure trihyd-

ride. The same phenomenon was observed earlier for iron

hydride.^[15]Upon addition of TFE at 200 K, 54% of the

complex converts to the tetrahydride, leading to an absorp-

tion decrease. The intensity of the band increases upon

heating, showing that the equilibrium shifts towards the ini-

tial complex. These changes are perfectly reversible in the

200–240 K range.

Figure 5. Intensity changes at 390 nm vs. the PNP/Mo ratio.

The stoichiometry was also probed with a weaker proton

donor, that is HFIP (P_i= 1.05), by way of a kinetics ap-

proach, as HFIP and the species derived from its proton-

ation do not contain chromophores allowing their direct

detection. The reaction was studied in toluene at 293 K un-

der pseudo-first-order conditions with HFIP/MoH₃ratios

in the 20–80 range, using the stopped-flow technique. The

proton-transfer step required a few seconds to reach equi-

librium. On a much longer timescale, a slower decomposi-

tion reaction occurred, as indicated by NMR monitoring

(see below). The time evolution of the spectra could be

properly fitted on the basis of a first-order decay, giving an

observed rate constant that turned out to be independent

of the alcohol concentration (see Figure 6). The average

value for k_1obsis 10.1 0.2 s^–1. This result is consistent with

the involvement of a single HFIP molecule in the rate-de-

termining step (Scheme 3), because the hydrogen-bonding

pre-equilibrium is heavily shifted to the adduct under these

conditions [see Equation (2)]. Therefore, as K_HB[HA] ϾϾ

1, the expression simplifies to k_obs= (k₁+ k_–1). In the pre-

viously published kinetics study of the [Cp*Fe(dppe)H] +

HFIP system, on the other hand, a first-order dependence

Figure 7. UV/Vis study of the interaction between [Cp*Mo(dppe)-

H₃] and TFE (4 equiv.) in CH₂Cl₂. (a) Initial complex c = 0.02 .

(b) Tetrahydride c = 0.02 . (c) T = 200 K. (d) T = 240 K. The

other intermediate spectra were recorded at each 10 K step.

The temperature reversibility enabled the determination

of the equilibrium constant for the formation of

–

[Cp*Mo(dppe)H₄]⁺OCH₂CF₃. It was assumed that the

equilibrium involves only one TFE molecule, as experimen-

tally verified for the PNP and HFIP systems (see above,

Scheme 3). Taking into account the above-determined hy-

drogen-bond formation constant K_HB, the analysis of the

data of Figure 7 led to the calculation of the proton-trans-

Figure 6. Pseudo-first-order rate constants (k_1obs) for the proton

transfer from HFIP to [Cp*Mo(dppe)H₃].

Scheme 3.

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2197

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

fer equilibrium constant K_PTat each temperature. A van’t

Hoff analysis of these K_PTconstants yields ∆H°_PT

=

–2.8 0.4 kcalmol^–1and ∆S°_PT= –15 2 calmol^–1K^–1for

the proton-transfer process.

The proton-transfer equilibrium constants were also ob-

tained by ¹H NMR spectroscopy. Like for the UV/Vis inves-

tigation, the NMR technique only provides information on

the sum of the concentrations of rapidly equilibrating free

trihydride and its hydrogen-bonded adduct. Knowledge of

the K_HBallowed the calculation of the individual concen-

trations, from which the K_PTconstants could be calculated

at each temperature. The van’t Hoff analysis yields ∆H°_PT

= –2.7 0.5 kcalmol^–1and ∆S°_PT= –11 2 calmol^–1K^–1.

These values are identical to those established from the low-

temperature UV/Vis data within experimental error.

Scheme 4.

Decomposition of the Proton-Transfer Product: Hydrogen

Evolution

fast exchange could also average the resonances of free

TFE/CF₃CH₂O^–with that of any coordinated TFE. Nei-

ther does the OH resonance give any useful information, as

it also remains unsplit under all conditions because of rapid

exchange processes. In addition, the position of this reso-

nance is highly temperature- and concentration-dependent,

NMR monitoring of the [Cp*Mo(dppe)H₃]–TFE inter-

action at high temperatures (Ͼ280 K) revealed further

irreversible evolution. The hydride resonances of the start-

ing trihydride and the protonated tetrahydride complexes

disappeared completely within 2 h at room temperature. No

new hydride resonance appeared in the spectrum. On the

other hand, the formation of a new diamagnetic product

was indicated by the replacement of the Cp* resonances

(δ = 1.69 for the trihydride and 1.83 for the tetrahydride

complexes) with a new one at δ = 1.58 in the ¹H NMR and

by the appearance of a resonance at δ –69.5 in the ³¹P{¹H}

NMR spectrum. This indicates the selective formation of a

single, hydride-free diamagnetic product. This evolution is

accompanied by the appearance of a new sharp resonance

at δ = 4.61 in the ¹H NMR spectrum, which is characteristic

of free H₂. All this evidence points to reductive elimination

of two H₂molecules from the tetrahydride complex

[Cp*MoH₄(dppe)]⁺, the product being stabilized by coordi-

nation of the trifluoroethoxide anion.

even

before

any

dihydrogen

evolution

from

[Cp*MoH₄(dppe)]⁺takes place. All our attempts to isolate

the final product in the crystalline state have failed. The

spin states’ dichotomy for complex [Cp*Mo(dppe)-

(OCH₂CF₃)], as well as the thermodynamics of further

TFE coordination, has been investigated by DFT calcula-

tions (vide infra).

Computational Studies

The interaction of [Cp*Mo(dppe)H₃] with fluorinated

alcohols has been studied theoretically by using the model

complex [CpMo(dpe)H₃] (dpe = H₂PCH₂CH₂PH₂) and

TFE and HFIP as proton donors, both in gas phase and in

dichloromethane. Test calculations for the real complex

have also been performed. The computational study was

carried out with several objectives in mind: (i) explore the

thermodynamics of hydrogen bonding at different sites; (ii)

investigate the effect of hydrogen bonding at different sites

on the vibrational modes of the MoH₃moiety; (iii) investi-

gate the stability of the protonation products and the pos-

sible involvement of nonclassical intermediates during the

proton-transfer process; (iv) determine the protonation

pathway; and (v) rationalize the further evolution of the

protonation product.

Stable 16-electron compounds of type [(C₅R₅)MoL₂X]

(the X ligand has π-donating properties) have previously

been reported, for example [CpMo(CO)₂{As(tBu)₂}],^[28]

[Cp*Mo(PMe₃)(PHPh₂)(PPh₂)],^[29]

[Cp*Mo(dppe)Cl],^[30]

and [CpMo(PMe₃)₂(OH)].^[30,31]These precedents suggest

that the [Cp*Mo(dppe)(OCH₂CF₃)] molecule may be

stable. Interestingly, compounds [CpMo(PMe₃)₂(OH)] and

[Cp*Mo(dppe)Cl] have two unpaired electrons, whereas our

observed H₂elimination product is diamagnetic like com-

pound [Cp*Mo(PMe₃)(PHPh₂)(PPh₂)].^[29]It is also pos-

sible, however, that one additional TFE molecule adds to

the system to afford an 18-electron adduct, which could

even be further stabilized by intramolecular hydrogen bond-

The Free Trihydride Complex

ing (see Scheme 4).

The optimization of the free [CpMo(dpe)H₃] model

1

The H NMR spectrum does not remove this ambiguity, system gave a geometry that is in close correspondence

because only one δ_CHresonance for TFE is observed, at with that observed experimentally^[19,20]for the full

the same position as i²n free TFE (δ Ϸ 3.9). It is not split [Cp*Mo(dppe)H₃] system. For the purpose of the present

or shifted by the deprotonation process, possibly because of discussion, the most important parameters are those related

a rapid exchange between the free alcohol and its conjugate to the MoH₃moiety (see Figure 8). Taking the plane that

anion. The absence of a new resonance for a coordinated contains the two P atoms and the center of the Cp as a

TFE molecule represents only negative evidence, because reference, the three H ligands are disposed asymmetrically,

2198

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

two (H1 and H2) on one side and one (H3) on the other

side. The unique H3 atom shows the longest distance to the

Mo center (1.724 Å). The other two hydrides seem to oc-

cupy approximately symmetric positions relative to the

Mo–H vector to the unique H ligand (H3–Mo–H1 = 132°,

H3–Mo–H2 = 129°), but they in fact show quite different

Mo–H distances. One distance (to H2) is similar to that of

the unique hydride (1.717 Å), whereas the other one is

much shorter (1.682 Å). This asymmetry parallels a slight

asymmetry in the Cp*Mo(dppe) moiety (the two CNT–

Mo–P angles are different at 131.1 and 143.1°) and its ori-

gin is not clear. The three hydride ligands have quite dif-

ferent Mulliken charges: –0.017, –0.003, and –0.053 for H1,

H2, and H3, respectively. Thus H3 is clearly the most hyd-

ridic site. The distance between H1 and H2 (1.753 Å) indi-

cates that they do not directly interact with each other, be-

cause it is longer than the longest separation between hy-

drogen atoms in what have been described as “stretched di-

hydrogen ligands” {e.g., 1.49 Å in [OsH₅(PMe₂Ph)₃]⁺}.^[32]

In addition, a theoretical analysis of another stretched dihy-

drogen complex {i.e., [Os(H···H)(NH₂CH₂CH₂NH₂)₂-

(HCO₂)]⁺with a distance of 1.428 Å} indicates no critical

point connecting the two atoms.^[33]

Scheme 5.

To account for all the possible hydrogen-bonded minima

we optimized the adduct between TFE and the [CpMo-

(dpe)H₃] starting from the different regions around the

metal. In this way five stable hydrogen-bonded structures

were obtained, as shown in Figure 9. Table 3 shows the rel-

Figure 8. ORTEP view of the DFT optimized [CpMo(dpe)H₃]

molecule. The Cp- and P-bonded H atoms and ethylene backbone

have been omitted for clarity. The M–H normal modes are repre-

sented with their respected computed frequency and intensity

(10⁴Lmol^–1cm^–2) in parentheses.

Figure 9. Optimized geometries of [CpMo(dpe)H₃] and the

[CpMo(dpe)H₃···TFE] hydrogen-bonded adducts A–C.

The analysis of the M–H normal modes in the free com-

plex, also represented in Figure 8, will serve as a basis for

discussion of the changes induced by hydrogen bonding. A

comparison between the experimental spectrum for com-

pound [Cp*Mo(dppe)H₃] and the computed one for the

[CpMo(dpe)H₃] model shows a rather good agreement,

both in terms of the frequencies and the relative intensities.

As shown in Figure 8, the higher-energy normal mode (ν₁)

is essentially a pure stretching vibration of the hydride li-

gand H1, which shows the shortest (and therefore strongest)

bond. The other two vibrations are relatively close to each

other in frequency and are a mixture of the other two M–

H bond vibrations, the higher-frequency one (ν₂) being an

in-phase combination with the major contribution from the

Table 3. Relevant optimized geometrical parameters (distances in

Å; angles in °) for [CpMo(dpe)H₃···TFE] (A–C) adducts.^[a]

A

B1

B2

B3

C

Mo–H1^[b]1.678

1.687

1.684

1.682

1.680

(–0.002)

1.725

(–0.004) (+0.005) (+0.002) (0.000)

Mo–H2^[b]1.722

1.713 1.725 1.736

(+0.005) (–0.004) (+0.008) (+0.019) (+0.008)

Mo–H3^[b]1.734

1.724 1.720 1.717 1.731

(+0.010) (0.000) (–0.004) (–0.007) (+0.007)

0.977 0.980 0.977 0.979 0.979

(+0.012) (+0.015) (+0.012) (+0.014) (+0.014)

O–H^[c]

H···H1

H···H2

2.057

1.779

1.944

1.799

2.942

1.653

shorter bond to H2, and the lower-frequency one (ν₃) being H···H3

1.816

3.068

2.104

2.864

H···Mo

2.927

152.5

150.1

3.034

148.5

150.6

2.987

147.1

160.3

an out-of-phase combination with the major contribution

from the longer bond to H3.

O–H···H1

O–H···H2

O–H···H3 155.2

O–H···Mo 152.9

140.0

172.4

Hydrogen-Bonded Adducts with TFE: Structures and

Vibrational Modes

160.4

177.8

170.4

[a] For the location of H1, H2, and H3, see Figure 8. [b] Values in

parentheses are the changes relative to free [CpMo(dpe)H₃]. [c] Val-

ues in parentheses are the changes relative to free TFE.

[CpMo(dpe)H₃] has four potential hydrogen-bonding

sites: the three hydrides and the metal (Scheme 5).

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2199

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

evant optimized geometrical parameters. The computed fre- tance of the proton from the metal center in system B1 is

quencies and the normal modes for the M–H stretching also relatively short. In B3 the proton interacts only with

vibrations are depicted in Figure 10.

H2, with H2···H–O and Mo···H–O angles of 160.3° and

The adducts can be classified depending on the proton- 170.4°, respectively. Finally, system C is the only one where

acceptor site. System A is the only one where TFE interacts the most important interaction appears to involve the metal

with the unique hydride ligand H3. In systems B (B1–B3) center, as indicated by the shortest Mo···H distance and the

the interaction occurs at one or both of the other two hy- almost linear Mo···H–O bond. The metal–proton interac-

dride ligands, H1 and H2, that occupy pseudoequivalent tion is accompanied by a relatively long H···H3 interaction

positions. Systems B2 and B3 are topologically related, as (2.104 Å). This initial description of the hydrogen-bonded

they exhibit the same Mo–H···H–O moiety. They differ in adducts, based on the structural data, will be reinforced by

the relative orientation of the proton donor molecule, which the analysis of the changes in the vibrational frequencies

places the C–H bond in front of the second hydride in B3 and atomic charges of the MoH₃moiety on going from the

and in the opposite direction in B2. In system B1 the free complex to the TFE adduct.

alcohol molecule is placed with the same relative orienta-

The evolution of the Mulliken charges on the hydride

tion as in B2 but farther from the Cp ring. Both in B1 and atoms and on the Mo atom upon going from the free com-

B2 the proton is closer to H2, but also interacts with H1. plex to the TFE adducts also provides useful information.

Thus, these geometries may be better described as contain- The relevant charges are collected in Table 4. The O–H

ing a “bifurcated hydrogen bond”, similar to that described bond is lengthened in all complexes, in comparison to free

for the Cp₂NbH₃·HOR_Fadducts.^[18]In addition, the dis- TFE (0.965 Å). The magnitude of this elongation is variable

Figure 10. ORTEP view of the DFT optimized [CpMo(dpe)H₃···TFE] adducts, showing selected bonding parameters and a representation

of the normal modes with the computed frequency and intensity (A = 10⁴Lmol^–1cm^–2) in parentheses. The P-bonded H atoms and

ethylene backbone have been omitted for clarity.

2200

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

but is always rather large (0.012–0.015 Å). The positive shorter for H2, 1.799 Å). The charge slightly decreases on

charge of the proton increases by 0.027–0.047 units. In sys- Mo (∆q = +0.006). The trends of the H···H distances and

tem A TFE interacts only with the unique hydride ligand hydride charges agree with the evolution of the Mo–H bond

(H3). The negative charge of H3 increases from –0.053 to lengths: both Mo–H1 and Mo–H2 are lengthened (by

–0.128 (∆q = –0.075), whereas the charges of H1 and H2 +0.002 and +0.008 Å, respectively), whereas Mo–H3 is

remain essentially unchanged and that of Mo slightly de- shortened (–0.004 Å) (Table 3). The O–H vector bisects

creases (∆q = +0.009). The Mo–H3 bond is significantly the H1MoH2 angle, with an essentially coplanar

elongated (∆r_M–H= +0.010 Å), whereas the Mo–H1 and MoH1H2···HO fragment. System B2, therefore, appears to

Mo–H2 bonds are less affected (slight contraction for the be best described as having a bifurcated hydrogen bond

former and slight elongation for the latter). The composi- with H1 and H2. This situation reflects onto the vibrational

tion of the three M–H vibrational modes (Figure 10) re- modes as follows. The ν₁mode appears at a slightly higher

mains close to that observed in the free trihydride (Fig- frequency (+3 cm^–1), which is unexpected because the Mo–

ure 8). The ν₁mode is blue-shifted by +17 cm^–1, in agree- H1 bond is slightly lengthened by the hydrogen bond. The

ment with the slight contraction of the Mo–H1 bond; the ν₂mode exhibits a moderate low-frequency shift (–16 cm^–1;

ν₂mode does not shift significantly (the major contribution a compromise between the Mo–H2 bond weakening and

is from Mo–H2, which maintains essentially the same dis- the Mo–H3 bond strengthening), whereas the ν₃mode cor-

tance); and the ν₃mode is strongly red-shifted (∆ν₃

=

respondingly shows a moderate high-frequency shift

–21 cm^–1), correlating with the weakening of the Mo–H3 (+3 cm^–1) for the same reasons.

bond by the hydrogen bond.

For system B1, on the other hand, the charge on H1

remains unchanged, whereas it increases significantly on H2

(–0.073) and also on the Mo atom (–0.051), suggesting that

the metal center also participates directly in the hydrogen

bonding. The O–H···Mo distance is 0.1 Å shorter in B1

than in B2. A folding of the O–H vector along the H1H2

axis by 141.2° toward the metal, in the opposite direction

from the Cp ring, signals an interaction that forces the pro-

ton donor to approach the metal center. These changes sug-

gest the description of adduct B1 as featuring a bifurcated

hydrogen bond to H2 and Mo. Note that, whereas the

H···H distances would suggest that hydrogen bonding with

H2 is stronger in B1 than in B2 (1.779 Å vs. 1.799 Å), the

Mo–H2 distance shortens (–0.004 Å), rather than lengthens,

Table 4. Computed Mulliken charges for selected atoms in the

[CpMo(dpe)H₃] complex and its TFE adducts.^[a]

H(TFE) Mo

H1

H2

H3

[CpMo(dpe)H₃]

–0.316

–0.017

–0.003

–0.053

TFE

A

B1

B2

B3

C

0.326

0.353

0.373

0.368

0.364

0.373

–0.307

–0.367

–0.310

–0.296

–0.422

–0.015

–0.017

–0.046

–0.020

–0.010

–0.001

–0.076

–0.088

–0.123

0.002

–0.128

–0.047

–0.051

–0.049

–0.072

[a] For the location of H1, H2, and H3, see Figure 8.

The H1 and H2 hydrides are involved in the hydrogen in B1. In addition, the charge on H2 increases to a smaller

bond in systems B1–B3. For system B3, the major observed extent in B1 than in B2. These perturbations may be caused

effects of hydrogen bonding are a significant lengthening of by the Mo···HO interaction and suggest that conclusions

the Mo–H2 bond (+0.019 Å), a slight shortening of Mo– about the hydrogen bond nature in a complicated poly-

H3 (–0.007 Å), a dramatic increase for the H2 negative atomic system should not be based solely on the H···H dis-

charge (–0.120), and a slight decrease for that of Mo tances. The low-frequency shift for ν₁(–12 cm^–1) correlates

(+0.020). Therefore, these charge and bond length changes with the lengthening of the Mo–H1 bond and the strong

confirm the description of this adduct as involving a dihy- high-frequency shift for ν₂(+23 cm^–1) correlates with the

drogen bond to H2. The ν₁vibration is very weakly influ- Mo–H2 bond compression, whereas there is no clear corre-

enced by the dihydrogen bonding (∆ν₁= +4 cm^–1), whereas lation with the H···H distances. Note that system B1 exhib-

ν₂and ν₃are dramatically affected (Figure 10). In fact, hy- its mixing of the Mo–H1 and Mo–H2 vibrations in ν₁and

drogen bonding with H2 weakens the Mo–H2 bond to such ν₂, whereas ν₃is an essentially pure Mo–H3 vibration. The

a point that the ν₂mode (to which the Mo–H2 vibration latter experiences no shift relative to the free trihydride

contributes the most) now appears at a lower frequency compound, consistent with the invariance of the Mo–H3

than ν₃(∆ν₂= –36 cm^–1). Conversely, ν₃is shifted to a bond length.

higher frequency (∆ν₃= +17 cm^–1), in agreement with the

Finally, system C shows a dramatic increase of the metal

Mo–H3 bond shortening, and the band intensity is de- negative charge (∆q = –0.106), whereas the charge on H3

creased by a factor of two. An additional contribution to increases only slightly (∆q = –0.019). Together with the

the red-shift of ν₂and to the blue-shift of ν₃may be associ- structural features discussed above, these charge variations

ated with the modified mixing: in-phase for ν₂and out-of- suggest that the bonding in this adduct is dominated by the

phase for ν₃, that is, the opposite of the situation in the free metal site, and is unique amongst all the optimized adducts

trihydride.

A–C in this respect. It is interesting to observe that not only

For system B2, the negative charge increases signifi- is the Mo–H3 bond elongated in this adduct (+0.007), but

cantly, relative to the free trihydride complex, on both H1 the Mo–H2 bond is also (+0.008 Å), whereas the Mo–H1

(∆q = –0.029) and especially H2 (∆q = –0.085), correlating bond is slightly shortened. There is no obvious explanation

well with the H···H distances (longer for H1, 1.944 Å; for this phenomenon. The evolution of the normal modes

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2201

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

is again in line with the observed changes of Mo–H bond

The calculated enthalpies, taking the BSSE-corrected en-

lengths: ν₁shows a small high-frequency shift (+4 cm^–1), ergies, are slightly decreased, but the ordering B Ͼ C Ͼ A

whereas ν₂and ν₃exhibit large low-frequency shifts (–33 is preserved. The interaction enthalpy was also estimated

and –20 cm^–1, respectively).

on the basis of Iogansen’s empirical relationships [Equa-

tion (1)^[34]and Equation (3)^[35]] using the computed ν(OH)

stretching vibration frequencies or ν(OH) band intensities

A, which are also reported in Table 5. Remarkably, the ∆H

computed in the different ways are very similar. Indeed, the

highest values obtained (about 6 kcalmol^–1) are close to the

experimental hydrogen-bond enthalpies. The ν(OH) bands

are found at lower frequency {∆ν(OH) = 227–292 cm^–1},

Hydrogen-Bonded Adducts with TFE: Relative Stability

The relative stability of these systems was estimated in

different ways (see Table 5). Comparing the energy changes

associated to the hydrogen bond formation, the stronger in-

teractions are those involving the two equivalent hydride

ligands. Structures B show the most exothermic hydrogen

bond. This behavior can be traced back to the establish-

ment of two dihydrogen bonds in B1 and B2 or only one

strong dihydrogen bond in B3.^[18]Although the primary in-

teraction is shorter in system B3, the relative stability of the

three systems is not too different. A fast exchange of the

ROH molecule between the three B sites can be expected

from the structural features and low energy differences

found between the three isomers. The interactions at the

unique hydride site (A) and at the metal site (C) have similar

energies, around 2 kcalmol^–1less exothermic than for the B

structures. A fast exchange between the A and C sites can

be also expected. The gas-phase complexation energies span

a range between 9 and 12 kcalmol^–1, similar to those calcu-

lated for the same kind of interactions in other transition-

metal hydride systems,^[15,17,18]and notably higher than the

experimental enthalpies. Previous results concerning hydro-

gen bonding in transition-metal systems showed that the

Basis Set Superposition Error (BSSE) can be very impor-

tant in this type of system.^[18]BSSE-corrected bonding en-

ergies have been calculated, showing that the BSSE ac-

counts for 40–50% of the interaction energy. However, after

correction all the adducts are still stable, with interaction

energies between 5 and 8 kcalmol^–1. It is worth mentioning

that a negative and significant interaction energy remains

at the metal site after the BSSE correction, in contrast with

what we have found in a recent study of the hydrogen bond-

ing to the [Cp*Fe(dppe)H].^[15]This result demonstrates the

basicity of the molybdenum center in this compound and

points towards competition between the hydride and the

metal sites for the proton.

and with

a

dramatically increased intensity [from

in free TFE to (4.99–

0.38×10⁴Lmol^–1cm^–2

7.77)×10⁴Lmol^–1cm^–2in the H-bonded adducts], in good

agreement with the experimental observations. Both corre-

lations work well for medium-strength hydrogen bonds of

organic bases and appear to be successfully applicable also

to organometallic complexes.^[12,13]Note that Equation (3)

is considered as universal (valid for different types of H

complexes in solution or in the gas phase) and more precise.

∆H = 2.9∆A^1/2= 2.9(A^1/2

– A^1/2

)

free

(3)

bonded

PCM calculations give the solvation free energy of a spe-

cies, and its partition into enthalpic and entropic parts is

not possible.^[36]From the ∆G_solvof free [CpMo(dpe)H₃],

free TFE, and their adduct, it is possible to get a rough

estimation of the ∆G of the hydrogen-bond formation in

solution, which does not take into account the internal con-

tributions to the entropy. The ∆G obtained in this way for

adducts A–C in CH₂Cl₂are collected in Table 5. Although

the ∆G(CH₂Cl₂) are less exothermic than the ∆H in gas

phase, for all the adducts ∆G(CH₂Cl₂) is negative, pointing

out the stability of the hydrogen-bonded species in solution.

Moreover, the ordering B Ͼ C Ͼ A is maintained.

Protonation Products

The calculations on the protonation product, the system

having the [CpMo(dpe)H₄]⁺stoichiometry, reveal the exis-

tence of two stable local minima for the classical tetrahy-

dride complex and another two for a dihydrogen–dihydride

tautomer, [CpMo(dpe)(H₂)H₂]⁺. They can be envisaged as

the products of the proton transfer to the different proton-

acceptor sites revealed by the hydrogen-bonded adducts.

The optimized structures are shown in Figure 11. Relevant

geometrical parameters are available as Supporting Infor-

mation.

Table 5. Calculated parameters of the hydrogen bonding between

[CpMo(dpe)H₃] and TFE.

A

B1

B2

B3

C

The most stable isomer is the tetrahydride TETRA1. This

structure presents a rather symmetrical arrangement, with

two hydrides at each side of the plane defined by the two

phosphorus atoms and the center of the Cp ring. The tetra-

hydride nature of this complex is evident from the H···H

distances, the shortest one being 1.799 Å. There are two sets

of Mo–H distances, longer (1.708 and 1.706 Å) for the two

most distant hydrides, and shorter (1.671 and 1.674 Å) for

the closest ones. The geometry of TETRA1 agrees with the

experimental structure of [Cp*WH₄(dppe)]⁺and its descrip-

tion as a distorted pseudopentagonal bipyramid (see

Scheme 2).^[19]The four H ligands are practically in the same

∆E [kcalmol^–1

∆E(BSSE)^[a][kcalmol^–1

∆H(BSSE)^[b][kcalmol^–1

ν(OH) [cm^–1

A (10^–4) [Lmol^–1cm^–2

∆H(∆ν)^[c][kcalmol^–1

∆H(∆A)^[d][kcalmol^–1

∆G(CH₂Cl₂)^[e][kcalmol^–1

]

–9.6

–5.1

–3.5

3617

(4.99)

–4.3

–4.7

–1.6

–11.3

–7.9

–6.0

3553

(7.25)

–5.2

–5.8

–3.1

–11.2

–7.8

–6.1

3605

(7.15)

–4.5

–6.0

–3.2

–11.5

–5.9

3552

(7.77)

–5.2

–6.3

–3.4

–10.4

–5.9

–4.3

3576

(6.23)

–4.9

–5.5

–2.8

]

[a] Complexation energy corrected by the basis set superposition

error. [b] Complexation enthalpy, taking the BSSE-corrected en-

ergy. [c] Application of Equation (1),^[34]using the computed

∆ν(OH). [d] Application of Equation (3),^[35]using the computed in-

tensities A. [e] Complexation free energy in dichloromethane.

2202

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

The tetrahydride TETRA1 remains the most stable species

in CH₂Cl₂, with TETRA2, DIH1, and DIH2 being placed

6.6, 3.3, and 5.0 kcalmol^–1above TETRA1.

To assess how the chosen model affects the results, we

have studied the protonation products in the full system,

performing full QM calculations on the actual

[Cp*Mo(dppe)H₄]⁺complex. The four stable structures

found for the model [CpMo(dpe)H₄]⁺system have also

been located in the full system, with very similar arrange-

ments of the MoH₄unit to the model systems. A general

view of the optimized geometries is shown in the Support-

ing Information. As can be appreciated in Table S1, the

structural changes induced by the modeling of the real li-

gands are minor.

The relative energies of the different isomers of the full

system are comparable to those of the model system. The

tetrahydride TETRA1 is still the most stable species, with

TETRA2, DIH1, and DIH2 placed 7.0, 4.4, and

6.6 kcalmol^–1above TETRA1. Thus, the calculations indi-

Figure 11. Views of the optimized geometries for isomeric struc-

tures having the [CpMo(dpe)H₄]⁺stoichiometry.

plane (the fourth H is only 0.010 Å away) and one P atom cate that the classical tetrahydride product TETRA1 is

is only 0.035 Å away from the least-squares H₄plane. The strongly favored relative to all possible dihydrogen struc-

five angles between adjacent bonds in the pseudopentago- tures. This result agrees with the experiment, as the tetrahy-

nal plane are relatively similar and close to that expected dride is the only observed protonation product. The pres-

for the equatorial ligands in a pentagonal bipyramid (72°). ence of dihydrogen–dihydride species at low relative ener-

TETRA1 can be considered the product of proton transfer gies suggests that they may be intermediates along the pro-

to the metal site, although it may also result from the oxi- ton-transfer pathway.

dative addition of a coordinated dihydrogen ligand. Both

possibilities will be considered in the next two sections.

The simplifications introduced in the modeling of the

system have very little influence in the relative energies of

The second tetrahydride minimum, TETRA2, has three the protonation products. To further check the differences

hydrides on one side of the P–Cp_centroid–P plane and one between the model and the full systems we have computed

on the opposite side. The Mo–H distances for the three hy- the gas phase proton affinity (PA), that is the energy change

drides on the same side are 1.666, 1.710, and 1.668 Å, the associated to Equation (4)

longest one being that to the central hydride. The shortest

([Mo]-H₃) + H⁺Ǟ ([Mo]-H₄)⁺

(4)

H···H distance is 1.602 Å (H1···H2), in agreement with the

tetrahydride nature of this isomer. The Mo–H distance of

the unique hydride H3 is 1.691 Å. TETRA2 could be re-

garded as the protonation product at the metal site close to

hydrides H1 and H2. TETRA2 is considerably less stable

than TETRA1, lying 8.0 kcalmol^–1higher.

Two dihydrogen–dihydride structures have been charac-

terized as local minima (DIH1 and DIH2). Their geome-

tries are comparable to that of the parent complex

[CpMo(dpe)H₃], but with a η²-H₂ligand in place of a hy-

dride. DIH1 is the most stable dihydrogen–dihydride iso-

mer, only 3.7 kcalmol^–1above TETRA1. The presence of a

dihydrogen ligand in the coordination sphere of the metal

where [Mo] stands for CpMo(dpe) or Cp*Mo(dppe) and

TETRA1 has been considered the protonation product. The

model system presents a slightly lower proton affinity

(249.9 kcalmol^–1) than the full system (271.0 kcalmol^–1),

but the difference amounts to only 8%. As all four proton-

ated structures have very similar relative energies in the full

and model systems, we can conclude that all the proton-

acceptor sites of the MoH₃unit are a little more basic in

the full system than in the model one.

Proton-Transfer Reaction Profiles with HFIP

We have also carried out a study of the [CpMoH₃(dpe)·

is apparent from the H–H distance of 0.896 Å between the HFIP] adduct, analogous to that illustrated above for the

adjacent H3 and H4 atoms. This structure is reached by [CpMoH₃(dpe)·TFE] adduct. The results are not detailed

protonation of the unique hydride H3 in the initial trihyd- for the sake of brevity, but the main outcome is that five

ride. The second dihydrogen–dihydride isomer (DIH2) is at- hydrogen-bonded structures comparable to the A–C ones

tained by proton transfer to the H1 or H2 ligands. The H– reported for TFE (Figure 9) have been located, which will

H distance in the dihydrogen ligand (0.851 Å) is shorter be correspondingly labeled AЈ, B1Ј, B2Ј, B3Ј, and CЈ. They

than in DIH1, indicating a weaker metal–H₂interaction. In follow the same stability ordering BЈ Ͼ CЈ Ͼ AЈ, with

agreement with this, DIH2 is less stable than DIH1, slightly increased interaction energies, caused by the

5.1 kcalmol^–1above TETRA1.

stronger acidity of HFIP compared with TFE. For instance,

We have also considered the relative stability of the four the ∆H(BSSE) are –3.9, –6.3, –5.7, –6.8, and –4.4 kcalmol^–1

[CpMo(dpe)H₄]⁺isomers in dichloromethane. There is little for the HFIP adducts AЈ, B1Ј, B2Ј, B3Ј, and CЈ, respec-

change on going from the gas phase to the CH₂Cl₂medium. tively.

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2203

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

When one imagines the reaction trajectory leading to

proton transfer, it can be predicted that system CЈ may lead

to the classical tetrahydride complex directly, whereas sys-

tems AЈ–BЈ could yield dihydrogen intermediates. The most

stable protonation product (TETRA1) formally arises from

proton transfer to CЈ, whereas the most stable dihydrogen

complex (DIH1) may be viewed as formally arising from

proton transfer from AЈ to the H3 hydride site. For this

reason we have studied both proton-transfer pathways. In

addition, the close vicinity of the protonation sites in the

AЈ and CЈ structures will allow a study of the direct compe-

tition for the proton between a hydride and the metal sites

and a discussion of whether a direct proton transfer to the

metal is possible in this case.

Starting from the hydrogen-bonded adducts CЈ and AЈ,

the final products of the proton transfer should be the ion

pairs made up by the TETRA1 (metal site protonation) or

DIH1 (hydride site protonation) cations and the

[(CF₃)₂HCO]^–([X]^–) anion. However, whereas the optimiza-

tion of the TETRA1·[X] ion pair was successful, attempts to

optimize the DIH1·[X] ion pair failed. Several optimizations

starting from different input geometries were tried but led

systematically to the same initial hydrogen-bonded adduct.

The optimized geometries of hydrogen-bonded complexes

and the ion pair are available in the Supporting Infor-

mation.

As no stable dihydrogen–conjugate base ion pair was

found, we have not pursued the proton-transfer pathway at

the hydride site. We have further explored the protonation

at the metal. Initially the gas-phase reaction profile leading

from CЈ to TETRA1·[X] has been investigated by using the

O–H distance of the HFIP donor as the pivot parameter

along the reaction coordinate, all other geometrical param-

eters being optimized. Then, the maximum of this curve has

been used as the starting point to locate the transition state,

which practically coincides with this maximum. The profile

in the dichloromethane solvent was obtained from single-

point calculations on each point of the gas-phase profile

with the solvent PCM. The energy of the starting hydrogen-

bonded adduct CЈ was taken as the zero of energy. The

energetic profiles are depicted in Figure 12.

The gas-phase energy barrier to form the ion pair at the

metal site is 24.7 kcalmol^–1. The CH₂Cl₂solvent slightly

favors the kinetics of the proton transfer, decreasing the en-

ergy barrier to 22.8 kcalmol^–1. In the transition state the

O–H bond is completely broken (O–H = 1.581 Å) and the

Mo–H bond is almost formed (Mo–H = 1.795 Å). The dis-

tance between the proton that is being transferred and the

contiguous H3 hydride (1.795 Å) is too long to consider the

formation of a dihydrogen ligand. Thus, even in this

crowded system there is a pathway for protonating the

metal, without involvement of the hydride ligands.

Figure 12. Potential energy curves for the proton transfer from

HFIP to the Mo trihydride at the metal site. Plain curve and

squares: in the gas phase; dashed line and triangles: in dichloro-

methane solution. The O–H length of the transferring proton has

been taken as the reaction coordinate.

ated species.^[15,17,18]Thus, we have also studied the proton

transfer to complex [CpMo(dpe)H₃] with two HFIP mole-

cules.^[15,27]The final products of the proton transfer are the

ion pairs made up by the TETRA1 or DIH1 cations and

the [(CF₃)₂HCO···HOCH(CF₃)₂]^–([XHX]^–) homoconjugate

anion. With the inclusion of the second AH molecule, both

ion pairs give stable local minima and it is possible to study

the proton-transfer process both at the metal and hydride

sites. The corresponding starting points are the hydrogen-

bonded adducts CЈ·HFIP and AЈ·HFIP, with a second

HFIP molecule joined by a O···H hydrogen bond to the

first HFIP molecule. The optimized geometries are available

in the Supporting Information.

The thermodynamic viability of the proton transfer in-

volving a second alcohol molecule [Equation (5)] is very dif-

ferent from the situation shown above for a single HFIP

molecule. Now the reaction is clearly exothermic, by

–5.2 kcalmol^–1for the CЈ+HFIP/TETRA1·[XHX] couple

and –3.2 kcalmol^–1for the AЈ+HFIP/DIH1·[XHX] one.

∆G(CH₂Cl₂) values, although considerably decreased (–1.6

and –0.7 kcalmol^–1for TETRA1 and DIH1, respectively),

suggest the thermodynamic feasibility of the reaction in

dichloromethane and indicate that the equilibrium is dis-

placed toward the right. The presence of a strong hydrogen

bond in the homoconjugated pair has a substantial impact

on the overall reaction energy.

[CpMoH₃(dpe)···HOCH(CF₃)₂] + (CF₃)₂CHOH Ǟ

[CpMo(dpe)“H₄”]⁺·[(CF₃)₂CHO···H···O–CH(CF₃)₂]^–(5)

Both reaction profiles, leading from AЈ·HFIP to

DIH1·[XHX] and from CЈ·HFIP to TETRA1·[XHX], have

There is experimental evidence that the participation of been investigated, both in the gas phase and in dichloro-

a second molecule of the proton donor might be crucial for methane solvent, taking the O–H distance of the transfer-

the proton transfer to transition-metal hydride by means ring proton as the reaction coordinate. The energy of the

of the so-called homoconjugate pairing effect.^[8,16,27]In our starting hydrogen-bonded adduct was taken as the zero of

previous theoretical studies, the second proton donor mole- energy in each case. The energetic profiles are depicted in

cule was used to locate the gas-phase minima of the proton- Figure 13.

2204

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

strong [A···H···A]^–hydrogen bond in the anion of the ion

pair can be inferred from the shortening of the O···O

distance (2.439 and 2.396 Å in [DIH1]⁺·[XHX]^–and

[TETRA1]⁺·[XHX]^–, respectively). The O··H distances

(1.359 and 1.241 Å) and O–H–O angles (178.5 and 178.1°)

also agree with a proton shared by two [X]^–units. The acid-

ity of the coordinated dihydrogen makes the homoconju-

gate anion interact more strongly with the H of the dihydro-

gen than with that of the tetrahydride (M–H···O distances

of 2.467 and 2.493 Å in [DIH1]⁺·[XHX]^–and

[TETRA1]⁺·[XHX]^–, respectively). In opposition, the

O···H···O interaction in the homoconjugate anion is

stronger in [TETRA1]⁺···[XHX]^–than in [DIH1]⁺···[AHA]^–,

according to both O···O and O···H distances.

Dihydrogen–Dihydride Ǟ Tetrahydride Interconversion in

the Ion Pair

To solve the dichotomy between easy protonation at the

hydride site and no detection of dihydrogen intermediates,

we have studied the interconversion from the [DIH1]⁺···

[XHX]^–to the [TETRA1]⁺···[XHX]^–ion pair. As in previous

studies of dihydrogen Ǟ dihydride interconversions,^[37]we

have chosen the H–H distance as a reaction coordinate for

the H–H bond breaking. This distance varies from 0.884 Å

in the dihydrogen structure to 1.812 Å in the tetrahydride.

For each fixed value of the reaction coordinate all the other

geometrical parameters, including those of the homoconju-

gate anion, were optimized. The profiles in dichlorometh-

ane solvent were obtained from single-point calculations on

each point of the gas-phase profiles with the solvent PCM.

The energy of the dihydrogen ion pair was taken as the zero

of energy. The energetic profiles in gas phase and CH₂Cl₂

are depicted in Figure 14.

Figure 13. Potential energy curves for the proton transfer from two

HFIP molecules to the Mo trihydride: (a) at the metal site; (b) at

the hydride site. Plain curves and squares: in the gas phase; dashed

lines and triangles: in dichloromethane solution. The O–H length

of the transferring proton has been taken as the reaction coordi-

nate.

The gas-phase barriers to form the ion pair at the hydride

and the metal site are 11.2 and 12.9 kcalmol^–1, respectively.

As already found with one HFIP molecule, the CH₂Cl₂sol-

vent favors the kinetics of the proton transfer, decreasing

the energy barriers to 9.4 and 10.2 kcalmol^–1, respectively.

The ion pairs resulting from the protonation at the hydride

([DIH1]⁺·[XHX]^–) and metal site ([TETRA1]⁺·[XHX]^–) are

found to be 6.3 and 4.8 kcalmol^–1, respectively, above the

hydrogen-bonded complexes in the gas phase, and 2.7 above

and 0.3 kcalmol^–1below, in dichloromethane. The solvent

also favors the thermodynamics of the proton transfer, sta-

bilizing the charged species to a greater extent than the ini-

tial neutral hydrogen-bonded ones. The role played by the

homoconjugate pairing in the proton-transfer process can

be appreciated following the evolution of the two O–H dis-

tances of the HFIP dimer throughout the protonation pro-

cess. In the hydrogen-bonded initial species AЈ·HFIP and

CЈ·HFIP there is a normal hydrogen bond between the oxy-

gen atom of the first HFIP molecule and the proton of the

second one. The O···O and O···H distances are 2.718 and

1.761 Å in AЈ·HFIP and 2.703 and 1.745 Å in CЈ·HFIP. The

O–H bond of the second HFIP molecule is only slightly

Figure 14. Potential energy curve for the interconversion from the

([DIH1]⁺·[XHX]^–) to the ([TETRA1]⁺·[XHX]^–) ion pairs. Plain

curves and squares: in the gas phase; dashed lines and triangles:

in dichloromethane solution. The H–H length of the coordinated

dihydrogen has been taken as the reaction coordinate.

The calculations illustrate that the dihydrogen–dihydride

lengthened by the hydrogen bond interaction (O–H = 0.980 Ǟ tetrahydride rearrangement in the ion pairs easily takes

and 0.981 Å in AЈ·HFIP and CЈ·HFIP, respectively; the place. The energy barrier is only 1.8 kcalmol^–1in gas phase

O–H distance in free HFIP is 0.968 Å). The presence of a and 1.3 kcalmol^–1in CH₂Cl₂. The homoconjugate anion

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2205

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

does not appreciably change its position along the intercon-

version, as proved by the evolution of the H···O distance

(from 1.359 Å in the dihydrogen structure to 1.241 Å in the

tetrahydride). The process is exoergic, ending up in a tetra-

hydride ion pair 1.5 kcalmol^–1more stable than the initial

dihydrogen in the gas phase (–1.7 kcalmol^–1in CH₂Cl₂).

The solvent influence in this process is very small because

it does not involve a charge change.

Discussion

It has been shown in previous spectroscopic^[14,38]and

theoretical^[15,17,39]studies of monohydride compounds that

the MH···HX interaction weakens the MH bond (low-fre-

quency shift for the M–H stretching vibration), whereas the

HM···HX interaction strengthens it (high-frequency shift).

The present contribution reports the first detailed analysis

of the effect of hydrogen bonding on the M–H stretching

vibrations for a polyhydride compound where the individ-

Dihydrogen Evolution

We have theoretically studied the successive elimination ual M–H stretching components are heavily mixed into the

of two H₂molecules in the model [CpMo(dpe)H₃]–TFE normal modes. The first outcome of this investigation is

system. The results are graphically assembled in Figure 15 that the variation of the vibrational frequencies can be pre-

and discussed only briefly here. More details can be found dicted, in most cases, by extension of the previously estab-

in the Supporting Information. The significant difference lished trends for monohydride complexes:^[12,13]normal

between the results in the gas phase and in CH₂Cl₂can be modes whose main contribution is from M–H vibrations

attributed to the solvation of the free TFE molecules (note where the hydride is directly implicated in hydrogen bonds

that the gap is about twice greater when both molecules as a proton acceptor experience a low-frequency shift. Con-

have been consumed, relative to the intermediate situations versely, normal modes whose main contribution is from M–

where only one has been consumed). The results show that H vibrations where the hydride is not directly implicated in

the first H₂elimination, affording the [CpMo(dpe)- hydrogen bonds (either the metal center or other hydride

H₂(OR_F)] intermediate (unobserved for the real system), is ligands act as the proton-acceptor site) experience a low-

less favorable than the process leading to loss of two H₂frequency shift. However, the contribution of various M–H

molecules. Note also that the H₂elimination process will stretching vibrations, some of which may be proton ac-

further benefit from an entropic drive. Furthermore, the ceptors in hydrogen bonds while others may not, in the nor-

subsequent addition of a second alcohol molecule, to afford mal modes calls for a more detailed and in-depth analysis.

the 18-electron adduct, results in a significant stabilization.

Secondly, the existence, in the present case, of additional

At least part of this stabilization may be attributed to the interactions with other neighboring proton-accepting sites

intramolecular hydrogen bond. Thus, the calculations point may be the major cause of the complicated changes found,

to the identity of the final product as a solvated species. It for example the changes of certain Mo–H bond lengths that

is also of interest to mention that the 16-electron [CpMo- are found to be opposite to expectations based on H···H

(dpe)(OR_F)] intermediate prefers to adopt a singlet ground distances (e.g., shortening of Mo–H2 in B1, lengthening of

state (the triplet is higher in energy by 6.5 and 7.3 kcalmol^–1Mo–H2 in C). An interaction involving exclusively the

in the gas phase and in solution, respectively). On the other metal center, without involvement of at least secondary in-

hand, a parallel calculation on [CpMo(PMe₃)₂(OH)] shows teractions with hydride ligands, does not occur for this com-

that the triplet is 2.3 kcalmol^–1lower than the singlet, in pound. It can be easily imagined, by extrapolation, that a

agreement with the experimental observation or paramag- direct hydrogen bond with a metal site may be difficult in

netism for this molecule.^[30,31]

general, if not impossible, when the system has a relatively

Figure 15. Computed energy profile (solid lines: gas phase; dashed lines: CH₂Cl₂solution) in kcalmol^–1for the H₂elimination from

complex [CpMo(dpe)H₃] (R_FOH = TFE). Full details are reported in the Supporting Information.

2206

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

crowded coordination sphere and especially when more = 1.27) interaction shows a breakpoint at a 1:1 ratio and a

than one hydride ligand is present. Also, the same circum- visible band consistent with the formation of the

stances may favor the establishment of simultaneous inter- [Cp*(dppe)MoH₄]⁺···[OAr]^–ion pair. On the other hand,

actions with more than one hydride ligand.

the kinetic study of the Cp*MoH₃(dppe)–HFIP (P_i= 1.05)

The computational study allows rather good repro- interaction illustrates the need for only one HFIP molecule

duction of the experimentally determined hydrogen bond to reach the transition state for proton transfer. This situa-

enthalpy for the [Cp*Mo(dppe)H₃]·TFE interaction by use tion markedly differs from that previously established for

of the [CpMo(dpe)H₃] model system. The calculations also the protonation of the less basic hydride ligand in complex

indicate that the most stable hydrogen-bonded adducts be- [Cp*Fe(dppe)H] (E_j= 1.35 0.02), where a second molecule

tween [CpMo(dpe)H₃] and TFE are complexes B, where the is necessary.^[27]Interestingly, the formation of hydrogen-

proton donor approaches the complex from the side of the bonded ion pairs is not only evident from the UV/Vis study

1

molecule containing hydrides H1 and H2. Structure B3, of the Cp*MoH₃(dppe)–PNP system, but also from the H

which yields the lowest energy minimum, is also the model NMR spectroscopic data through the different T₁values

structure that best reproduces the experimentally observed obtained for the TFE and HBF₄protonation products

frequency shifts: the ν₂band experiences a large low-fre- (Table 2), although the latter result cannot distinguish be-

quency shift and appears 3 cm^–1lower than ν₃mode in free tween the simple (X^–) and the homoconjugate (XHX^–)

trihydride, whereas the ν₃band experiences a large high- anion.

frequency shift, ending up near the position of the ν₂mode

The low-temperature NMR study does not show any evi-

in free trihydride (compare Figure 8 and Figure 10). The dence for the formation of a dihydrogen intermediate,

composition of the normal modes is in agreement with this whereas the low-temperature IR study suggests that the pre-

interpretation (the low-frequency band in B3 has the major ferred hydrogen-bonding pathway involves the hydride li-

component from the Mo–H2 vibration, whereas the low- gand. Thus, an intermediate dihydrogen complex along the

frequency band in the free trihydride is mostly a Mo–H3 proton-transfer pathway may be expected. The computa-

vibration). The other structural models of the hydrogen- tional investigation suggests that, both in gas phase and in

bonded adduct (A, B1, B2, and C) yield calculated normal dichloromethane, the tetrahydride ion pair resulting from

mode frequency patterns in greater discrepancy with the ex- the protonation at the metal site is more stable than the

periment. In conclusion, it seems that the proton donor ap- dihydrogen–dihydride ion pair arising from the hydride pro-

proaches the complex from the side of the molecule con- tonation, in agreement with the experiment. Both proton-

taining the hydrides H1 and H2, and establishes a hydrogen transfer pathways leading to hydride and metal protonation

bond with the hydride H2.

show low barriers, in agreement with the experimental evi-

One interesting issue for proton-transfer reactions to hy- dence of a very fast process. However, these values are

dride molecules is the number of proton donor molecules rather similar (9.4 and 10.2 kcalmol^–1for proton transfer

that are involved in proton transfer. Although the involve- from the HFIP dimer in CH₂Cl₂), indicating that in this

ment of only one molecule should be sufficient in principle, molybdenum trihydride there is no clear kinetic preference

it has been shown in several cases that the protonation for the protonation at the hydride site. This theoretical re-

product contains the homoconjugate anion, therefore im- sult contrasts with the commonly admitted kinetic prefer-

plicating two proton donor molecules in the reaction stoi- ence for the protonation at the hydride site, and with a re-

chiometry.^[8,16,17]It was also shown by kinetics investi- cent theoretical study of the protonation of CpFe(dpe)H

gations that a second proton donor molecule is necessary with two HFIP molecules.^[15]

to trigger proton transfer from fluorinated alcohols to the

If the nonclassical intermediate indeed forms during the

hydride complex [Cp*Fe(dppe)H].^[15,27]On the other hand, proton-transfer process, its activation barrier to the re-

when the stronger acid CF₃COOH was used with the same arrangement leading to the classical tetrahydride product

hydride system, even a single acid molecule was sufficient must be very low. Indeed, a very low barrier for the re-

to transfer the proton. Theoretical calculations have also arrangement of the model compounds ([DIH1]⁺···[XHX]^–

shown that the involvement of a second proton donor (HX) to [TETRA1]⁺···[XHX]^–where HX = HFIP) is shown by

molecule strengthens hydrogen bonding of the first HA the calculations. Note that this rearrangement does not ne-

molecule with the proton acceptor in the MH···HX···HX cessitate any ion pair dissociation, as the anion does not

adduct.

appreciably change its position along the interconversion.

The basicity of the hydride ligand of Cp*MoH₃(dppe)

To summarize, the tetrahydride product can originate

(E_j= 1.42 0.02) is higher than that of trihydride complexes either from the direct protonation at the metal site or from

[18]

studied previously [compare with E_jof 0.93 for Cp₂NbH₃

the fast isomerization of the dihydrogen intermediate and

and 0.94 for Cp*RuH₃(PCy₃)]^[40]and is among the highest the available experimental and theoretical results do not al-

reported in the literature.^[13]Consequently, it can be proton- low us to clearly distinguish these two possibilities. The in-

ated by such a weak acid as TFE. Two separate experiments volvement of a transient dihydrogen species in the forma-

on the proton-transfer process indicate that complex tion of trihydride by protonation of a transition-metal dihy-

Cp*MoH₃(dppe) is able to accept a proton from a single dride has been reported in a mechanistic study of the pro-

molecule of proton donor, at least from donors as weak as tonation of Cp₂MH₂(M = Mo, W) by an excess of anhy-

HFIP. The UV/Vis study of the Cp*MoH₃(dppe)–PNP (P_idrous HCl.^[41]

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2207

A. Lledós, R. Poli, E. S. Shubina et al.

FULL PAPER

UV measurements were carried out by use of a cryostat (Carl Zeiss

Jena) in the 200–290 K temperature range. The accuracy of the

temperature adjustment was 0.5 K. The reagents were mixed at

low temperature and the cold mixtures were subsequently transfer-

red into the precooled cryostat.

Conclusions

The unambiguous determination of the hydrogen-bond-

ing site for a polyhydride compound by NMR and IR spec-

troscopic techniques is not an easy task. The high fluxion-

ality of polyhydride complexes on the NMR timescale ren-

ders the application of this method less straightforward,

while the use of vibrational spectroscopy is complicated by

the normal mode structure and weak intensity of the M–H

stretching bands. The theoretical analysis of the vibrational

modes, which was performed here for the first time for a

transition-metal polyhydride compound in the absence and

presence of hydrogen bonding, shows a different behavior

for each M–H vibrational normal mode. Nevertheless, the

evolution of the vibrational frequencies follows in most

cases the trends previously established for monohydride

complexes: a low-frequency shift for modes that primarily

involve the M–H vibrations where the H ligand is impli-

cated in the hydrogen bond and a high-frequency shift for

modes that primarily involve M–H vibrations where the H

ligand is not implicated in the hydrogen bond (hydrogen

bonding to other hydride ligands or to the metal atom).

According to the calculations the most stable hydrogen-

bonded adducts are those where the proton donor ap-

proaches the Cp*MoH₃(dppe) complex from the side of the

molecule containing the hydrides H1 and H2. Of these, the

one giving good agreement with the experimental IR obser-

vations features a hydrogen bond with H2.

Dihydrogen bond formation precedes the proton transfer

to form the cationic tetrahydride [Cp*(dppe)MoH₄]⁺com-

plex, stabilized by hydrogen bonding to the proton donor

conjugate base. This is reversible at low temperatures (be-

low 250 K) when using p-nitrophenol or TFE. The precur-

sor Cp*(dppe)MoH₃complex is sufficiently basic to accept

the proton from a single molecule of these two proton donors.

The experiments do not show any direct evidence for the

formation of a dihydrogen intermediate, whereas the pre-

ferred hydrogen-bonding pathway, according to the compu-

tational study, appears to involve the hydride ligand. The

intermediacy of a dihydrogen complex along the proton-

transfer pathway is possible, as suggested by the computa-

tions. However, the calculated low barrier for its isomeriza-

tion into the classical tetrahydride complex would rational-

ize our inability to observe it experimentally.

NMR Investigations: NMR studies were carried out in standard 5-

mm NMR tubes containing solutions of the complexes in CD₂Cl₂.

¹H and ³¹P NMR spectroscopic data were collected with a Bruker

AMX 400 spectrometer operating at 400.13 and 161.98 MHz,

respectively. The conventional inversion-recovery-pulse method

(180-τ-90) was used to determine the variable-temperature longitu-

dinal relaxation time T₁. Low-temperature experiments were car-

ried out in the 180–260 K temperature range using a TV-3000

Bruker temperature control unit. The accuracy and stability of tem-

perature was 1 K. All mixings between the alcohols and the hy-

dride complexes were performed at low temperature.

Stopped-Flow Investigations: The stopped-flow kinetic runs were

carried out at 293 K with a Bio-Logic SF20 apparatus coupled to

a J&M TIDAS MMS-UV/Vis diode-array UV/Vis spectrophotom-

eter using a cuvette with a 0.15-cm pathlength. The data were col-

lected within the first 10 s, yielding reproducible results. Data

analyses were carried out by using the SPECFIT3 global analysis

package of programs.^[42]At least three runs were averaged to yield

the reported results.

Computational Details: Calculations were performed with the

Gaussian98^[43]package at the DFT/B3LYP level.^[44–46]Effective

core potentials (ECP) were used to represent the innermost elec-

trons of the molybdenum atom as well as the electron core of phos-

phorus atoms.^[47,48]The basis set for the Mo and P atoms was that

associated with the pseudopotential,^[47,48]with a standard double-

ζ LANL2DZ contraction,^[43]supplemented in the case of P with a

set of d-polarization functions.^[49]The carbon and hydrogen atoms

of the transition-metal complexes that are not bonded to the metal

atom, together with the atoms of proton donor molecules (C, F,

H) that are not involved in hydrogen bonds, were described with a

6-31G basis set.^[50]The carbon and hydrogen atoms directly

bonded to the metal and the proton donor molecules’ hydrogen

and oxygen atoms involved in hydrogen bonding were described

with a 6-31G(d,p) set of basis functions.^[51]Geometry optimiza-

tions were carried out without symmetry restrictions and the re-

sulting structures were characterized as minima by the real value

of all 3N-6 vibrational frequencies. Solvent effects were taken into

account by means of polarized continuum model (PCM) calcula-

tions,^[52,53]using standard options.^[43]The solvation free energies

were computed in dichloromethane (ε = 8.93) at the gas-phase opti-

mized geometries. The gas-phase complexation energies were cor-

rected from the BSSE according to the counterpoise method of

Boys and Bernardi.^[54]

Experimental Section

Supporting Information (for details see the footnote on the first

page of this article): Table of optimized geometrical parameters for

[CpMo(dpe)H₃] and [CpMo(dpe)H₄]⁺complexes, views of opti-

mized [Cp*Mo(dppe)H₃], [Cp*Mo(dppe)H₄]⁺, a variety of dihydro-

gen-bonded adducts, and a description of computational results on

the H₂elimination process from [CpMo(dpe)H₄]⁺(7 pages).

General: All manipulations were carried out under argon by stan-

dard Schlenk techniques. The [Cp*Mo(dppe)H₃] hydride was syn-

thesized according to the literature.^[19]All solvents used (CH₂Cl₂,

THF, toluene) were dried using appropriate agents and freshly dis-

tilled prior to use.

IR and UV/Vis Investigations: IR measurements were performed

on “Specord M82” and Infralum FTIR-81 (Lumex) spectrometers

using CaF₂cells (0.12 cm path length). In order to limit the signifi-

cance of self-association phenomena, a concentration of the proton

donor between 0.005 and 0.01  was used for the ν(OH) measure-

ments. The UV measurements were performed on Specord M-40

and Varian Cary 5 spectrophotometers. Low-temperature IR and

Acknowledgments

We thank the European Commission through the HYDROCHEM

program (contract HPRN-CT-2002-00176) for support of this

work. Additional bilateral support (PICS, France–Russia) and

national support from the CNRS (France), the RFBR

2208

www.eurjic.org

Eur. J. Inorg. Chem. 2006, 2192–2209

Hydrogen Bonding and Proton Transfer to [Cp*MoH₃(dppe)]

FULL PAPER

[25] R. H. Crabtree, D. G. Hamilton, Adv. Organomet. Chem. 1988,

28, 299–338.

[26] E. I. Gutsul, N. V. Belkova, G. M. Babakhina, L. M. Epstein,

E. S. Shubina, C. Bianchini, M. Peruzzini, F. Zanobini, Russ.

Chem. Bull. 2003, 52, 1204–1206.

[27] N. V. Belkova, P. O. Revin, L. M. Epstein, E. V. Vorontsov, V. I.

Bakhmutov, E. S. Shubina, E. Collange, R. Poli, J. Am. Chem.

Soc. 2003, 125, 11106–11115.

[28] M. Luksza, S. Kimmer, W. Malisch, Angew. Chem. Int. Ed.

Engl. 1983, 22, 416–417.

(05-03-22001, 05-03-32415), and the Division of Chemistry and

Material Sciences of RAS (Russia), and from MEC (project

CTQ2005-09000-CO2-01; Spain) is also gratefully acknowledged.

A.L. thanks the Generalitat de Catalunya for a Distinció per a la

Promocio de la Recerca Universitària. N.B. thanks the Russian Sci-

ence Support Foundation for an individual grant. M.B. thanks the

Spanish Ministerio de Educación y Ciencia for a post-doctoral fel-

lowship.

[29] R. T. Baker, J. C. Calabrese, R. L. Harlow, I. D. Williams, Or-

ganometallics 1993, 12, 830–841.

[30] R. Poli, E. A. Quadrelli, New J. Chem. 1998, 22, 435–450.

[31] J. C. Fettinger, H.-B. Kraatz, R. Poli, E. A. Quadrelli, Chem.

Commun. 1997, 889–890.

[32] T. J. Johnson, A. Albinati, T. F. Koetzle, J. Ricci, O. Eisenstein,

J. C. Huffman, K. G. Caulton, Inorg. Chem. 1994, 33, 4966–

4976.

[1] M. Peruzzini, R. Poli, Recent Advances in Hydride Chemistry,

Elsevier, Amsterdam, 2001.

[2] V. I. Bakhmutov, Eur. J. Inorg. Chem. 2005, 245–255.

[3] G. Parkin, J. E. Bercaw, J. Chem. Soc., Chem. Commun. 1989,

255–257.

[4] M. S. Chinn, D. M. Heinekey, J. Am. Chem. Soc. 1990, 112,

5166–5175.

[5] G. Jia, A. J. Lough, R. H. Morris, Organometallics 1992, 11,

161–171.

[6] P. Hamon, L. Toupet, J.-R. Hamon, C. Lapinte, Organometal-

lics 1992, 11, 1429–1431.

[7] J. R. Hamon, P. Hamon, L. Toupet, K. Costuas, J. Y. Saillard,

C. R. Chim. 2002, 5, 89–98.

[8] E. T. Papish, F. C. Rix, N. Spetseris, J. R. Norton, R. D. Wil- [37] J. Tomàs, A. Lledós, Y. Jean, Organometallics 1998, 17, 4932–

liams, J. Am. Chem. Soc. 2000, 122, 12235–12242.

[9] E. T. Papish, M. P. Magee, J. R. Norton, in Recent Advances

in Hydride Chemistry (Eds.: M. Peruzzini, R. Poli), Elsevier,

Amsterdam, 2001, pp. 39–74.

[10] E. S. Shubina, A. N. Krylov, N. V. Belkova, L. M. Epstein,

A. P. Borisov, V. D. Mahaev, J. Organomet. Chem. 1995, 493,

275–277.

[11] A. Albinati, V. I. Bakhmutov, N. V. Belkova, C. Bianchini, I.

De Los Rios, L. Epstein, E. I. Gutsul, L. Marvelli, M. Peruz-

zini, R. Rossi, E. S. Shubina, E. V. Vorontsov, F. Zanobini, Eur.

J. Inorg. Chem. 2002, 1530–1539.

[33] F. Maseras, A. Lledós, M. Costas, J. M. Poblet, Organometal-

lics 1996, 15, 2947–2953.

[34] A. V. Iogansen, Theor. Exp. Chem. 1971, 7, 312–317.

[35] A. V. Iogansen, Spectrochim. Acta, Part A 1999, 55, 1585–1612.

[36] M. H. Baik, R. A. Friesner, J. Phys. Chem. A 2002, 106, 7407–

7412.

4939.

[38] L. M. Epstein, N. V. Belkova, E. I. Gutsul, E. S. Shubina, Pol.

J. Chem. 2003, 77, 1371–1383.

[39] G. Orlova, S. Scheiner, J. Phys. Chem. A 1998, 102, 260–269.

[40] N. V. Belkova, L. Epstein, E. S. Shubina, unpublished observa-

tions.

[41] R. A. Henderson, K. E. Oglieve, J. Chem. Soc., Dalton Trans.

1993, 3431–3439.

[42] R. A. Binstead, A. D. Zuberbühler, B. Jung, Specfit/32 v.

3.0.36, Spectrum Software Associates, Chapel Hill, NC, 2004.

[43] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,

M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Mont-

gomery Jr, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.

Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas,

J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.

Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson,

P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rab-

uck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Or-

tiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Pis-

korz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T.

Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gon-

zalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,

M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon,

E. S. Replogle, J. A. Pople, Gaussian 98, Revision A.9,

Gaussian, Inc., Pittsburgh PA, 1998.

[12] L. M. Epstein, N. V. Belkova, E. S. Shubina, in Recent Ad-

vances in Hydride Chemistry (Eds.: M. Peruzzini, R. Poli), El-

sevier, Amsterdam, 2001, pp. 391–418.

[13] L. M. Epstein, E. S. Shubina, Coord. Chem. Rev. 2002, 231,

165–181.

[14] N. V. Belkova, E. S. Shubina, L. M. Epstein, Acc. Chem. Res.

2005, 38, 624–631.

[15] N. V. Belkova, E. Collange, P. Dub, L. M. Epstein, D. A. Le-

menovskii, A. Lledós, O. Maresca, F. Maseras, R. Poli, P. O.

Revin, E. S. Shubina, E. V. Vorontsov, Chem. Eur. J. 2005, 11,

873–888.

[16] N. V. Belkova, A. V. Ionidis, L. M. Epstein, E. S. Shubina, S.

Gruendemann, N. S. Golubev, H. H. Limbach, Eur. J. Inorg.

Chem. 2001, 1753–1761.

[17] N. Belkova, M. Besora, L. Epstein, A. Lledós, F. Maseras, E.

Shubina, J. Am. Chem. Soc. 2003, 125, 7715–7725.

[44] C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. B: Condens. Mat-

ter 1988, 37, 785–789.

[18] E. Bakhmutova, V. Bakhmutov, N. Belkova, M. Besora, L. Ep- [45] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.

stein, A. Lledós, G. Nikonov, E. Shubina, J. Tomas, E. Vo-

rontsov, Chem. Eur. J. 2004, 10, 661–671.

[46] P. Stephens, F. Devlin, C. Chabalowski, M. Frisch, J. Phys.

Chem. 1994, 98, 11623–11627.

[19] B. Pleune, R. Poli, J. C. Fettinger, Organometallics 1997, 16,

[47] W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284–298.

[48] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.

[49] A. Höllwarth, M. Bohme, S. Dapprich, A. Ehlers, A. Gobbi,

V. Jonas, K. Kohler, R. Stegmann, A. Veldkamp, G. Frenking,

Chem. Phys. Lett. 1993, 208, 237–240.

1581–1594.

[20] J. C. Fettinger, B. Pleune, R. Poli, J. Am. Chem. Soc. 1996, 118,

4906–4907.

[21] J. Andrieu, N. V. Belkova, M. Besora, E. Collange, L. M. Ep-

stein, A. Lledós, R. Poli, P. O. Revin, E. S. Shubina, E. V. Vo-

rontsov, Russ. Chem. Bull. 2003, 52, 2679–2682.

[22] J. Andrieu, N. V. Belkova, M. Besora, E. Collange, L. M. Ep-

stein, A. Lledós, R. Poli, P. O. Revin, E. S. Shubina, E. V. Vo-

rontsov, paper to be published.

[23] E. Gutsul, N. Belkova, M. Sverdlov, L. Epstein, E. Shubina,

V. Bakhmutov, T. Gribanova, R. Minyaev, C. Bianchini, M.

Peruzzini, F. Zanobini, Chem. Eur. J. 2003, 9, 2219–2228.

[24] R. B. Girling, P. Grebenik, R. N. Perutz, Inorg. Chem. 1986,

25, 31–36.

[50] W. Hehre, R. Ditchfie, J. Pople, J. Chem. Phys. 1972, 56, 2257–

2261.

[51] P. Harihara, J. Pople, Theor. Chim. Acta 1973, 28, 213–222.

[52] J. Tomasi, M. Persico, Chem. Rev. 1994, 94, 2027–2094.

[53] C. Amovilli, V. Barone, R. Cammi, E. Cancès, M. Cossi, B.

Menucci, C. S. Pomelli, J. Tomasi, Adv. Quantum Chem. 1998,

32, 227–261.

[54] S. F. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553–566.

Received: February 2, 2006

Published Online: March 31, 2006

Eur. J. Inorg. Chem. 2006, 2192–2209

www.eurjic.org

2209

Article Doi

Hydrogen bonding and proton transfer to the trihydride complex [Cp*MoH3(dppe)]: IR, NMR, and theoretical investigations

DOI: 10.1002/ejic.200600093

Source and publish data:

Authors:

Article abstract of DOI:10.1002/ejic.200600093

Full text of DOI:10.1002/ejic.200600093

Products guided by the article

R&D Labs maybe for 919272-70-1

Relevant to this article

Hot Product