Selective homogeneous hydrogenation of biogenic carboxylic acids with [Ru(TriPhos)H]+: A mechanistic study

Article abstract of DOI:10.1021/ja2034377

Selective hydrogenation of biogenic carboxylic acids is an important transformation for biorefinery concepts based on platform chemicals. We herein report a mechanistic study on the homogeneously ruthenium/phosphine catalyzed transformations of levulinic acid (LA) and itaconic acid (IA) to the corresponding lactones, diols, and cyclic ethers. A density functional theory (DFT) study was performed and corroborated with experimental data from catalytic processes and NMR investigations. For [Ru(TriPhos)H]⁺ as the catalytically active unit, a common mechanistic pathway for the reduction of the C=O functionality in aldehydes, ketones, lactones, and even free carboxylic acids could be identified. Hydride transfer from the Ru-H group to the carbonyl or carboxyl carbon is followed by protonation of the resulting Ru-O unit via σ-bond metathesis from a coordinated dihydrogen molecule. The energetic spans for the reduction of the different functional groups increase in the order aldehyde < ketone < lactone ≈ carboxylic acid. This reactivity pattern as well as the absolute values are in full agreement with experimentally observed activities and selectivities, forming a rational basis for further catalyst development.

Full text of DOI:10.1021/ja2034377

ARTICLE
pubs.acs.org/JACS
Selective Homogeneous Hydrogenation of Biogenic Carboxylic Acids
with [Ru(TriPhos)H] : A Mechanistic Study
+
†
†
†
,†
,†,‡
Frank M. A. Geilen, Barthel Engendahl, Markus H €o lscher, J €u rgen Klankermayer,* and Walter Leitner*
†
Institut f €u r Technische und Makromolekulare Chemie, RWTH Aachen University, Worringerweg 1, 52074 Aachen, Germany
‡
Max-Planck-Institut f €u r Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 M €u lheim a.d. Ruhr, Germany
S Supporting Information
b
ABSTRACT: Selective hydrogenation of biogenic carboxylic acids is an important
transformation for biorefinery concepts based on platform chemicals. We herein
report a mechanistic study on the homogeneously ruthenium/phosphine catalyzed
transformations of levulinic acid (LA) and itaconic acid (IA) to the corresponding
lactones, diols, and cyclic ethers. A density functional theory (DFT) study was
performed and corroborated with experimental data from catalytic processes and
+
NMR investigations. For [Ru(TriPhos)H] as the catalytically active unit, a
common mechanistic pathway for the reduction of the CdO functionality in
aldehydes, ketones, lactones, and even free carboxylic acids could be identified.
Hydride transfer from the RuꢀH group to the carbonyl or carboxyl carbon is followed by protonation of the resulting RuꢀO unit
via σ-bond metathesis from a coordinated dihydrogen molecule. The energetic spans for the reduction of the different functional
groups increase in the order aldehyde < ketone < lactone ≈ carboxylic acid. This reactivity pattern as well as the absolute values
are in full agreement with experimentally observed activities and selectivities, forming a rational basis for further catalyst
development.
’
INTRODUCTION
and selectivities toward the different functional groups in this
reaction network.
The selective catalytic conversion of biomass feedstock is an
Ruthenium/phosphine complexes have found application as
hydrogenation catalysts and especially complexes with the fa-
cially coordinating tridentate ligand TriPhos have been identified
to form highly active catalysts for the hydrogenation of esters,
important challenge for synthetic chemistry to build novel
sustainable supply chains for the production of transportation
1,2
fuels and industrial chemicals. Lignocellulose constitutes a
major fraction of the terrestrial biomass feedstock, and therefore
much scientific and industrial attention is paid on its conversion
to valuable products. A widely studied approach is the con-
version of the carbohydrate fraction of lignocellulose into a set of
defined platform molecules, which are used as key intermediates
for the synthesis of novel fuels and fuel additives. The
implementation of the concept of platform chemicals requires
a detailed understanding of the molecular principles regulating
their further transformations resulting in multistep reaction
cascades and synthetic networks.
A general theme in the synthetic pathways from carbohydrate
feedstocks is the reduction of the oxygen content via hydrogena-
tion and dehydration sequences. This is illustrated in Scheme 1
for the two potential platform chemicals levulinic acid (LA) and
itaconic acid (IA) yielding a set of isomeric lactones, diols, and
cyclic ethers. Recently, we have developed a catalytic system
7ꢀ16
amides, and even free carboxylic acids.
Bianchini and co-workers
3
,4
reported on the reactivity of Ru/TriPhos complexes toward various
1
7ꢀ19
substrates in hydrogenation reactions,
and studied its
20
complexes extensively by NMR and IR spectroscopy. They
were able to characterize Ru/TriPhos complexes containing a
dihydrogen molecule as ligand in solution by low-temperature
5,6
2
1ꢀ23
and/or high-pressure NMR techniques.
The presence of a
nonclassical hydrogen molecule also could be proved for Ir/TriPhos
and Rh/Triphos complexes. Cationic dihydrogen complexes of
ruthenium were shown to be active hydrogenation catalysts also
with mono- and bidentate phosphine ligands (such as PMe Ph or
2
24ꢀ26
dppe).
Despite the extensive work on the characterization and
catalytic application of cationic nonclassical ruthenium hydride com-
plexes, the mechanistic details of their role in complex hydrogenation
reactions remains largely unexplored. Recently, Chaplin and Dyson
proposed a mechanism for the hydrogenation of alkenes using a Ru/
consisting of a ruthenium precursor such as Ru(acac) (1), the ligand
TriPhos (TriPhos = 1,1,1-tris(diphenylphosphinomethyl)ethan, 2),
3
27
TriPhos complex comprising nonclassical hydrides, and Frediani
and co-workers used a deuterium labeling study to provide first
and an acidic additive (Chart 1), which allows the selective7
16
mechanistic insights for the hydrogenation of esters and lactones.
synthesis of each of the individual products in over 90% yield.
In the present paper, we describe a combined experimental and
theoretical study on a plausible molecular mechanism of this
system, which allows rationalizing the observed reactivities
Received: April 14, 2011
Published: July 25, 2011
r 2011 American Chemical Society
14349
dx.doi.org/10.1021/ja2034377
|J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
Scheme 1. Reaction Sequences for the Hydrogenation/Dehydration of Levulinic Acid (LA, Top) and Itaconic Acid (IA, Bottom),
Leading to the Corresponding Lactones, Diols, and Cyclic Ethers
Chart 1. Catalyst System for the Hydrogenation of LA
Table 1. Hydrogenation of LA and GVL with a Ruthenium
TriPhos Catalyst
a
yield (%)
b
b
entry substrate catalyst additive GVL 1,4-PDO
2-MTHF
a
1
2
3
4
5
6
LA
(1 + 2)
(1 + 2)
(1 + 2)
3
15
1
95
76
0
0
0
a
a
GVL
GVL
LA
3
3
95
3
c
4
22
90
1
73
3
c
GVL
GVL
4
7
c
4
0
96
a
3
Conditions: 10 mmol of LA, 0.1 mol % Ru(acac) (1), 0.2 mol % TriPhos
(
2), 1 mol % additive (3) where stated; reaction time, 18 h; hydrogen pres-
b
sure p(H ) = 100 bar; 160 °C. 1,4-PDO = 1,4-pentanediol, 2-MTHF = 2-
methyltetrahydrofuran. Conditions: 10 mmol LA, 0.1 mol % complex 4, 1
mol % additive (3) where stated; reaction time, 18 h; hydrogen pressure
2
c
’
RESULTS
p(H ) = 100 bar; 160 °C.
2
Experimental Background. During our work on levulinic
acid reduction we were able to identify a well-defined organo-
metallic species at the end of the hydrogenation reactions which
This was even more pronounced in the hydrogenation of GVL,
could be isolated from experiments with higher catalyst loadings
where the isolated complex 4 proved almost inactive in accord with
literature reports, as compared to 85% conversion with the in situ
7
32
and finally characterized as [(TriPhos)Ru(CO)(H) ] (4).
2
The formation of 4 can be explained by decarbonylation of trace
amounts of aldehydes comprised in the reaction system. Alde-
hyde groups can be formed as intermediates during lactone
reduction or from dehydrogenation of the product alcohols,
as was described by Cole-Hamilton for ethanol and by Hartwig
catalyst (Table 1, entries 2 and 5). When an acidic additive such as
the ionic liquid 3 (Chart 1) was present, the in situ catalyst and
complex 4 showed similar activities toward the hydrogenation of
GVL (Table 1, entries 3 and 6). Although the effect of the acidic
additive on the rate of hydrogenation of GVL might be associated
with acid catalyzed ring-opening of the lactone to form a more
active substrate (vide infra), the increased rate for levulinic acid in
the presence of acid suggests that complex 4 represents a resting
state which is converted into the active species by protonation
7
for 1,4-butanediol in the presence of rutheniumꢀphosphine
2
8,29
complexes.
The mechanistic aspects of these dehydrogena-
tion reactions have been studied very recently computationally
30,31
by B €u hl and Bolm.
To verify the formation of 4 under
33
hydrogenation conditions, the complex was synthesized inde-
pendently using a modified reaction protocol. Thus, Ru(acac)3
from LA and/or the acidic additive. Indeed monitoring the reaction
1
of complex 4 with LA under carefully controlled conditions by H
(
1) and TriPhos (2) were dissolved in propanal and stirred for 20
NMR at room temperature showed the clean formation of a mono-
hydride species (Scheme 2, Figure 1), which could be characterized as
complex 5 on the basis of mass spectroscopy (ESI, m/z = 841.6,
h at 150 °C and 120 bar of hydrogen. Analytically pure 4 was
isolated in 84% yield after precipitation with ethanol, making this
also a synthetically useful procedure for 4 and related com-
pounds. In addition to the liquid product propanol from catalytic
hydrogenation, ethane could be identified in the gas phase,
unequivocally demonstrating decarbonylation as the source of
the CO ligand.
Using complex 4 as catalyst for hydrogenation of LA to
γ-valerolactone (GVL), a slightly lower catalytic activity was ob-
served as compared to the in situ system (Table 1, entries 1 and 4).
34
100%) and heteronuclear NMR spectroscopy. The signal for the
hydride is a doublet of doublets of doublets at ꢀ5.89 ppm and shows
3
the characteristic J coupling with three different phosphorus atoms
PH
with a trans-coupling of 94.7 Hz and cis-couplings of 19.7 and 14.3
Hz, respectively. Taken together, these results strongly suggest that
+
the complex fragment [(TriPhos)RuH] defines the catalytically
active center for the hydrogenation steps of the sequences shown in
35
Scheme 1.
1
4350
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
Scheme 2. Reaction of the Isolated Complex 4 with Levulinic
Acid
conditions. This correlates with the fact that such conditions
favor cationic intermediates, although acid catalyzed ring-open-
ing of GVL could give a more active substrate than GVL itself and
remains an alternative possibility. The clear positive effects of
acid prompted us to chose the Triphosꢀrutheniumꢀhydride
+
fragment in its cationic form [(Triphos)RuH] as the common
active species under the optimized reaction conditions and to
investigate whether a unifying mechanism for all individual
transformations in this catalytic system can be identified on the
basis of this principle. We note, however, that this does not rule
out the possibility of a neutral cycle where the first reduction
steps would occur on neutral complexes such as 5. Indeed, the
direct hydride transfer in 5 to the coordinated carbonyl group has
a low-lying transition state with an activation barrier of 7.3 kcal/
mol. Further studies on the following steps and in particular on
the activation energy for acid hydrogenation via the neutral cycle
are currently under way.
Hydrogenation of Levulinic Acid, Step 1: Conversion to
γ-Valerolactone (Scheme 3, Figure 2). Starting from the
+
[
(Triphos)RuH] cation fragment, the structure LA-1 provides
a plausible entry into the catalytic cycle under strongly acidic
conditions and high hydrogen pressure. The hydrogenation of
LA is then initiated with the reduction of the keto group in LA-1
to give the hydroxy acid which cyclizes to GVL (Scheme 3). The
reduction is achieved by transfer of the classical hydride ligand to
the carbon atom of the coordinated carbonyl group via a typical
migratory insertion transition state LA-TS 1ꢀ2 (Chart 2). The
activation energy of this step amounts to 16.1 kcal/mol, while the
product LA-2 of the reaction is less stable than the reactants by
1
5.9 kcal/mol. The weak agostic interaction of the newly formed
CꢀH bond in LA-2 rearranges to a more stable conformation in
which both oxygen atoms of the substrate coordinate to the metal
center (LA-3).
1
Figure 1. H NMR spectra of the formation of 5 (g; δ = ꢀ5.89 (ddd,
3
3
3
In the next step, one hydrogen from the dihydrogen ligand is
transferred as proton to the metal alkoxide group, while the other
forms a bond to the metal center renewing the ruthenium
hydride group (LA-4). The activation energy for this step is very
low (4.8 kcal/mol, LA-TS 3ꢀ4, Chart 2) reflecting the efficient
proton transfer in a four-membered transition state typical for
σ-bond metathesis. In contrast to oxidative addition/reductive
elimination, this pathway avoids the change of formal oxidation
states and major rearrangements in the coordination sphere as also
pointed out for example for the hydrogenolysis of rhodiumꢀ
1
H, JP ꢀH = 94.7 Hz, JP ꢀH = 19.7 Hz, JP ꢀH 14.3 Hz) ppm) from
ax
eq
eq
3
complex 4 (a; δ = ꢀ7.30 (dd, 2H, J
= 50.2 Hz, 18.3 Hz) ppm) in the
PꢀH
presence of LA. Spectra measured after t = 0 (a), 1 (b), 2 (c), 3 (d), 7 (e),
15 (f), 19 h (g).
The formation of cationic Ru complexes starting from neutral
rutheniumꢀdihydride complexes in the presence of TsOH or
NH PF under hydrogenation conditions was reported
4
6
3
3,36
previously.
A dynamic equilibrium between the neutral
40
41
dihydride trans-[(dppm) Ru(H) ] and the cationic hydridoꢀ
carboxylate groups. The energetic span for the overall reduction
as defined by the TOF-determining-intermediate (TDI) LA-1,
and the TOF-determining transition state (TDTS) LA-TS 3ꢀ4
amounts to 17.1 kcal/mol, indicating the reaction to be possible
under the given experimental conditions.
2
2
+
ꢀ
dihydrogen complex [(dppm) HRu(H )] (OR) in the pre-
2
2
sence of hexafluoroisopropyl alcohol was observed by NMR
3
7
spectroscopy. Furthermore, the group of Elsevier reported very
slow hydrogenation under neutral conditions, whereas the addi-
tion of HBF increased both the rate of the first step and the
Replacing the γ-hydroxypentanoic acid from the catalyst by
unreacted levulinic acid enables its lactonization to γ-valerolac-
tone via well-known standard acid catalyzed organic mechan-
isms, which were not calculated in this study. Alternatively,
cyclization could occur directly from LA-4, leading also back to
4
selectivity toward the diol in the reduction of dimethylphthalate
14
with the rutheniumꢀtriphos system. Comparable effects of acid
were obtained in the present study for the hydrogenation of
levulinic acid with the rutheniumꢀtriphos system. After 5 h at
1
60 °C and 100 bar hydrogen, 52% GVL, 43% pentanediol
LA-1 in the presence of LA and H . Both pathways offer plausible
2
(PDO), and 1% methyltetrahydrofuran MTHF were obtained in
routes for closing the catalytic cycle for this first hydrogenation.
Hydrogenation of LA, Step 2: Conversion of γ-Valerolac-
tone to 1,4-Pentanediol under Acidic Conditions (Scheme 4,
Figure 3). The next step in the hydrogenation sequence is
the reduction of the lactone GVL. The ruthenium catalyzed
hydrogenation of ester groups to the corresponding alcohols
is of great synthetic value adding additional generic interest in
the absence of additives, whereas the addition of aIL and NH PF6
4
as additives resulted in 91% MTHF and only 2% GVL. Product
inhibition in pathways based on neutral ruthenium hydrides may
38,39
be at least partly responsible for these differences.
It is thus evident from the experimental data of the present
study and related literature reports that the deep hydrogena-
tion beyond the ester level is greatly facilitated under acidic
4
2,43
this step.
1
4351
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
Scheme 3. Calculated Catalytic Cycle for the Hydrogenation of Levulinic Acid to γ-Valerolactone, Using the Cationic Complex
LA-1 as the Catalyst (Phosphine Ligand Omitted for Clarity)
Proton transfer from the dihydrogen ligand to the ruthe-
nium alkoxide leads to GVL-3 with an activation energy of
1
4.7 kcal/mol. Again the four-membered transition state
GVL-TS 2ꢀ3 provides a pathway for the hydrogenolysis of
the RuꢀO bond that is lower in energy than the hydride
transfer step. To proceed with the hydrogenation sequence,
the coordinated lactole (cyclic hemiacetale) has to convert to
its open hydroxy aldehyde form. Various pathways for this
rearrangement within the coordination sphere of the ruthe-
nium were investigated (paths IꢀIII). Although transition
states could be located in all cases, the energy barriers were
found to be significantly higher than for the hydride transfer
(
path I, 25.8 kcal/mol; path II, 30.7 kcal/mol; path III, 26.7
kcal/mol). The energetic span for path I defined by GVL-1
(
TDI) and GVL TS 3ꢀ5 (TDTS) amounts to 41.5 kcal/mol,
Figure 2. Energy profile for the hydrogenation of levulinic acid using
LA-1 as starting structure (ΔG in kilocalories per mole, relative to LA-1).
and similar unrealistically high values result for paths II and
III. Therefore, these routes can be ruled out to account for the
further transformation.
In analogy to LA, the primary hydrogenation product GVL can
In contrast, typical activation energies for the acid catalyzed
+
49,50
coordinate to the [(TriPhos)RuH] fragment via its carbonyl
opening of lactols are in the range of only 5 kcal/mol.
Thus,
group, leading to the starting complex GVL-1 under H pressure.
ring-opening of the free hemiacetal and recoordination of the
hydroxyl aldehyde to give GVL-6 provides a much lower energy
pathway and is considered here as a productive route in the
catalytic process. This assumption explains nicely the surprising
experimental observation that hydrogenation of pure GVL to
1,4-PDO is significantly slower under acid-free conditions than
the two-step conversion of LA (Table 1, entries 1 and 2).
Application of the acidic additive 3 in the hydrogenation of
GVL results in comparably high conversion as observed in the LA
conversion (Table 1, entry 3). This can be rationalized if the
barriers of paths IꢀIII become rate limiting for pure GVL,
2
The coordination via the carbonyl oxygen is the most prominent
44ꢀ46
binding mode for esters and lactones in literature.
Only very
few examples for coordination via the alkoxide or as a π-ligand
4
7,48
are known.
In accord with this, geometry optimizations
aiming at the localization of ether oxygen coordinated GVL did
not result in stable structures for GVL-1. Again, a migratory
insertion-type pathway for hydride transfer via transition state
GVL-TS 1ꢀ2 was found, leading to the lactolate complex GVL-2.
Similar structures have been identified by NMR spectroscopy
as intermediates for lactone hydrogenation by the group of
3
9
+
Bergens. The energy barrier for GVL-TS 1ꢀ2 was determined
to be 21.6 kcal/mol, which is significantly higher than for the
corresponding step of the keto group hydrogenation in LA (16.1
kcal/mol). This is fully in line with the excellent selectivities that
can be experimentally achieved in the consecutive reactions
under optimized conditions. Even under identical conditions
whereas the ring opening is facilitated by H from LA or the
ionic liquid 3 and water from the condensation step in the case of
consecutive reaction.
The final steps of the hydrogenation leading to the conversion
of the hydroxy aldehyde in GVL-6 to the diol in GVL-9 are
essentially identical to the previously discussed CdO hydro-
genation sequences. The corresponding energy barriers are fairly
low with activation energies of 9.0 kcal/mol for GVL-TS 6ꢀ7
and 3.7 kcal/mol for GVL-TS 8ꢀ9. Ligand replacement of
(
160 °C, 100 bar) hydrogenation of LA to GVL is quantitative
within 2 h whereas GVL conversion takes ca. 18 h with the in situ
catalyst.
1
4352
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
Chart 2. Calculated Transition States of the LA Hydrogenation, LA-TS 1-2 Showing the Hydride Transfer to the Carbonyl
Carbon and LA-TS 3-4 Showing the σ-Bond Metathesis for the Protonation of the RuꢀO Bond
Scheme 4. Calculated Catalytic Cycle for the Hydrogenation of γ-Valerolactone to 1,4-Pentanediol under Acidic Conditions,
a
Using the Monocationic GVL-1 as the Catalyst
a
The Brønstedt acid catalyzed pathway from GVL-3 to GVL-5 was not calculated. The phosphine ligand was omitted for clarity.
7
1
,4-PDO by GVL closes the metal catalyzed cycle. Depending on
yields (Scheme 1). This transformation involves the catalytic
hydrogenation of a free carboxylic acid function as a first step, a
process that has very little precedent with homogeneous
the reaction conditions, the diol can undergo intramolecular
etherification to 2-MTHF via well-established acid catalysis.
Hydrogenation of Itaconic Acid: Conversion of Methyl-
succinic Acid to Methyl-γ-butyrolactone (Scheme 5,
Figure 4). The Ru/TriPhos catalytic systems was shown to allow
not only the hydrogenation of LA but also the conversion of IA
to the corresponding lactones, diol, and cyclic ether in excellent
7,16
catalysts.
We therefore investigated whether the same basic
mechanism at the [(TriPhos)RuH] fragment could be operative
for free carboxylic acid groups.
+
Since the mechanistic aspects of hydrogenation of CꢀC
double bonds with ruthenium phosphine complexes are
1
4353
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
Figure 3. Energy profile for the hydrogenation of γ-valerolactone to 1,4-pentanediol using GVL-1 as the starting structure in structural analogy to LA-1
ΔG in kilocalories per mole, relative to LA-1).
(
Scheme 5. Calculated Catalytic Cycle for the Hydrogenation of Itaconic Acid to Methyl-γ-butyrolactone Using the Monocationic
Complex IA-1 as the Catalyst (the Phosphine Ligand and the Charges of the Ru(II) Species Omitted for Clarity)
5
1
well-established, the mechanistic cycle was started with methyl-
succinic acid (MSA) as the primary hydrogenation product of IA.
In analogy to the substrates discussed above, the complex IA-1
could be located as a stable species on the energy surface and was
ketone reduction in LA (16.1 kcal/mol). This is fully in line with
the experimental conditions that require significantly higher
reaction temperatures for IA than for LA in the conversion to
7
the corresponding lactones.
52
used as a starting point. The hydride transfer to the carboxylic
carbon center occurs via transition state IA-TS 1ꢀ2 (Chart 3),
which is essentially identical to the pathway for the carbonyl
reduction in LA. The activation energy to reach IA-2 amounts to
The free coordination site at Ru in IA-2 is readily occupied by
the second carboxylic acid group to give IA-3, for which the
proton transfer to the ruthenium alkoxide requires only an
activation barrier of 8.1 kcal/mol via the σ-bond metathesis
transition state IA-TS 3ꢀ4. Dehydration of the resulting geminal
2
4.2 kcal/mol and is thus 8.1 kcal/mol more demanding than the
1
4354
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
The high selectivity for MGBL at lower reaction temperatures
results from substrate inhibition, based on greatly preferred
coordination of the diacid over the corresponding lactone.
This interpretation is supported by the concentration/time
profiles of the consecutive hydrogenation steps, which show,
an onset of the lactone reduction only after full conversion of the
diacid.
’
CONCLUSION
Ruthenium phosphine catalysts and in particular the ruthe-
nium/TriPhos system show a very promising potential for highly
selective hydrogenation/dehydration pathways in the conversion
of biogenic platform chemicals. Analyzing in detail a representa-
tive amount of individual steps of the resulting complex reaction
networks, the present study shows that this can be associated
with the ability to reduce the CdO functionality in aldehyde,
ketone, ester, and even free carboxylic acid groups. These
transformations can all be accomplished with a common me-
Figure 4. Energy profile for the hydrogenation of methylsuccinic acid
to methyl-γ-butyrolactone using IA-1 as the starting structure in analogy
to LA-1 (ΔG in kilocalories per mole, relative to IA-1).
+
chanistic principle in the case of [(TriPhos)RuH] as the active
Chart 3. Calculated Transition State IA-TS 1-2, Showing the
Hydride Transfer to the Carboxyl Carbon
center (Figure 5). Although a cycle involving neutral intermedi-
ates cannot be ruled out at this stage, the cationic cycle success-
fully explains the experimental activity trends.
In the common mechanistic scheme investigated here, the
reduction results from a hydride transfer on to the carbonyl or
carboxyl carbon via transition states typical for migratory inser-
tion. The subsequent hydrogenolysis of the metal-oxide unit
involves proton transfer via σ-bond metathesis from a coordi-
nated dihydrogen molecule, regenerating at the same time the
catalytically active RuꢀH unit. The interplay between the
classical and the nonclassical metal hydride coordination pro-
vides an overall pathway that does not require changes in the
40
formal oxidation state of the ruthenium center.
The heterolytic cleavage of the H2 molecule and the
associated proton transfer may be assisted by external basic
centers of substrates or solvents potentially reducing the
5
3
energy barrier for the hydrogenolysis steps even further.
This possibility was not yet assessed in the present study,
because, in all mechanistic cycles investigated, the migratory
insertion-type hydride transfer could be associated with the
most energy demanding step. The activation barriers were
found to range from 3 to 9 kcal/mol for the aldehydes (X = H)
over 16 kcal/mol for the ketones (X = R) to 22ꢀ24 kcal/mol
for the lactones (X = OR) and acids (X = OH), respectively.
diol generates the significantly more stable aldehyde complex IA-
5
. The subsequent reduction of the aldehyde function follows
again the established sequence with activation energies which are
4
1
The energetic span, which ultimately determines the catalytic
turnover frequency for the overall reduction of the functional
groups, increases in the same order. These results are fully in
line with the observed experimental findings and allow ratio-
nalization of the optimized conditions for the selective trans-
formations. Therefore, these mechanistic insights may provide
guidelines also for future developments of efficient catalysts for
challenging hydrogenation reactions of carboxylic acids and
derivatives thereof.
lower than the ones described above for the reduction of the
q
aldehyde GVL-6 (ΔG = 2.7 kcal/mol for IA-TS 6ꢀ7 and 3.4
kcal/mol for IA-TS 8ꢀ9). The energetic spans reflect these
findings as well: Complex IA-1 and transition state IA-TS 3ꢀ4
are the TDI and the TDTS, respectively, and define an energetic
span of 31.0 kcal/mol for the carboxylic acid reduction, as
compared to 17.1 kcal/mol for the reduction of the keto group
in LA and only 7.0 kcal/mol for the reduction of the aldehyde
in GVL-6.
Once methyl-γ-butyrolactone (MGBL) is formed, its further
transformations can occur in full analogy to the isomeric GVL
substrate. The calculated activation energies for the initial
hydrogen transfer in MGBL hydrogenation are 19.8 kcal/mol
’ COMPUTATIONAL DETAILS
All calculations in this work were performed with the Gaussian09
5
4
program series. All structures were optimized using the dispersion
5
5
(
MGBL-TS 1ꢀ2, migratory insertion) and 18.3 kcal/mol
corrected B97-D functional together with the def2-SVP basis set for all
5
6ꢀ62
(MGBL-TS 2ꢀ3, σ-bond metathesis) resulting in a total energetic
elements and the associated ECP for ruthenium.
density fitting approximation was activated.
verified to be local minima or transition states by frequency calculation
showing 0 and 1 imaginary frequencies, respectively. Free enthalpies and
The automatic
All structures were
6
3,64
span of 29.7 kcal/mol to be overcome by the system during this
first lactone hydrogenation. This is in the same range as the ener-
getic span of the free carboxylic acid reduction (31.0 kcal/mol).
1
4355
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
Gibbs free energies reported include zero point energy corrections as
well as thermal corrections for T = 298.15 K and p = 1 atm. For many
structures IRC calculations were carried out additionally to prove the
localized transition state connecting the expected local minima. For final
evaluation of energies single-point calculations were carried out on the
B97-D/def2-SVP geometries using the B97-D functional with def2-
ordered from ABCR, Sigma-Alrich, or Fluka and used as received.
Solvents were purified by distillation under argon atmosphere and
subsequent storage over molecular sieves to remove traces of water.
Analytics. Conversion and selectivity of the catalytic reactions were
determined via GC using a Thermo Focus or a Siemens Sichromat
instrument, both equipped with flame ionizaiton detectors (FIDs). For
the levulinic acid system a Cp-Sil-Pona-CB column was used with
1-hexanol as internal standard. Peaks were assigned via GC-MS and pure
substance calibration. NMR spectra were recorded on a Bruker AV 400
5
6ꢀ62
TZVP
basis set and associated ECP for ruthenium using the
automatic density fitting approximation. Tables listing the energies
obtained are included in the Supporting Information as well as
Cartesian coordinates of all compounds. In the main text we refer to
1
spectrometer operating at 400.2 MHz for H, and a Bruker AV 600
q
1
ΔG and ΔG values obtained using the def2-TZVP energies together
spectrometer operating at 600.1 MHz for H. CDCl or CD Cl were
3
2
2
1
13
with zero-point energy (zpe) corrections as well as thermal correc-
tions from the lower level geometry optimizations unless noted
otherwise. Table 2 summarizes relative energies ΔGrel of all structures
used as the solvents. Chemical shifts of H and C NMR are reported
relative to tetramethylsilane (TMS) with solvent residual protons and
carbon atoms as internal standards, respectively. P chemical shifts are
reported relative to H PO .
3
1
q
and activation barriers ΔG for all transition states (all in kilocalories
3
4
per mole).
Hydrogenation of LA to 1,4-Pentanediol. Ru(acac)
3
(3.9 mg,
10 μmol), 1,1,1-tris(diphenylphosphinomethyl)ethane (12.5 mg, 20.0
μmol), and levulinic acid (1.16 g, 10.0 mmol) and where applicable N-
butyl-N -(4-sulfobutyl)imidazolium p-toluenesulfonate (43.3 mg, 10.0
’
EXPERIMENTAL SECTION
0
μmol) were placed in the glass liner of a 10 mL stainless steel reactor.
The high-pressure reactor was repeatedly evacuated and flushed with
argon and then pressurized with a predetermined amount of hydrogen
to adjust a pressure of 100 bar after heating up to the appropriate
reaction temperature. The mixture was stirred at the stated reaction
temperature for the given time. After the reaction the vessel was cooled
in an ice bath and slowly depressurized. The product mixture was
Safety Warning. Experiments with compressed gases must be
carried out only with appropriate equipment and under rigorous safety
precautions.
General Comments. If not stated otherwise, handling of chemicals
and manipulations were carried out under argon inert gas atmosphere
using standard Schlenk techniques. Ru(acac) was synthesized accord-
3
6
5
ing to a procedure described by Knowles et al. All chemicals were
1
analyzed by GC and H NMR.
Hydrogenation of GVL to 1,4-Pentanediol. Ru(acac)
3
(
(
3.9 mg, 10.0 μmol), 1,1,1-tris(diphenylphosphinomethyl)ethane
12.5 mg, 20.0 μmol), and where applicable N-butyl-N -(4-sulfobutyl)-
0
imidazolium p-toluenesulfonate (43.3 mg, 10.0 μmol) were placed in the
glass liner of a 10 mL stainless steel reactor. The high-pressure reactor was
repeatedlyevacuatedandflushedwithargon, thenGVL(1,05g, 10.0mmol)
was added, and the reactor was pressurized with a predetermined amount
of hydrogen to adjust to a pressure of 100 bar after heating to the
appropriate reaction temperature. The mixture was stirred at the
stated reaction temperature for the given time. After the reaction the
Figure 5. Schematic of production pathway.
q
Table 2. Relative Energies ΔGrel of All Structures and Activation Barriers ΔG for All Transition States (All in Kilocalories per
Mole)
q
q
q
structure
ΔG
ΔG
structure
GVL-6
ΔG
ΔG
structure
ΔG
ΔG
a
b
Scheme 3/Figure 2
5.8
14.8
13.4
14.2
17.8
3.9
Scheme 5/Figure 5
LA-1
0.0
16.1
15.8
12.3
17.1
1.6
GVL-TS 6ꢀ7
GVL-7
9.0
3.7
IA-1
0.0
24.2
23.9
22.8
31.0
16.7
7.7
LA-TS 1ꢀ2
LA-2
16.1
4.8
IA-TS 1ꢀ2
IA-2
24.2
8.1
GVL-8
LA-3
GVL-TS 8ꢀ9
GVL-9
IA-3
LA-TS 3ꢀ4
LA-4
IA-TS 2ꢀ4
IA-4
a
Scheme 4/Figure 3, path II
IA-5
a
0
Scheme 4/Figure 3, path I
ꢀ5.3
GVL-3
8.4
IA-6
13.0
15.7
13.1
14.1
17.6
ꢀ0.1
ꢀ2.7
0
GVL-1
GVL-TS 3 -4
GVL-4
39.1
5.8
30.7
26.7
IA-TS 6ꢀ7
IA-7
2.7
3.4
GVL-TS 1ꢀ2
GVL-2
16.3
9.9
21.6
14.7
25.8
IA-8
a
GVL-TS 2ꢀ3
GVL-3
24.6
10.4
36.2
7.4
Scheme 4/Figure 3, path III
IA-TS 8ꢀ9
IA-9
00
GVL-3
10.9
0
0
0
GVL-TS 3ꢀ5
GVL-5
GVL-TS 3 -6
37.6
8.1
MGBL-1
0
GVL-6
MGBL-TS 1ꢀ2
MGBL-2
17.2
8.7
19.8
18.3
MGBL-TS 2ꢀ3
27.0
a
b
Relative to LA-1. Relative to IA-1.
1
4356
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
ARTICLE
vessel was cooled in an ice bath and slowly depressurized. The
’ AUTHOR INFORMATION
1
product mixture was analyzed by GC and H NMR.
Hydrogenation of LA or GVL with Complex 4 as the
Corresponding Author
leitner@itmc.rwth-aachen.de; klankermayer@itmc.rwth-aachen.
de
0
Catalyst. Where applicable, N-butyl-N -(4-sulfobutyl)imidazolium
p-toluenesulfonate (4.3 mg, 5.0 μmol) was placed in the glass liner of a
10 mL stainless steel reactor. The high-pressure reactor was repeatedly
evacuated and flushed with argon; then complex 4 (3.9 mg, 5.0 μmol)
dissolved in degassed LA (0.51 g, 5.0 mmol) or GVL (0.5 g, 5.0 mmol)
was transferred via syringe. The reactor was pressurized with a predeter-
mined amount of hydrogen to adjust a pressure of 100 bar after heating to
the appropriate reaction temperature. The mixture was stirred at the stated
reaction temperature for the given time. After the reaction the vessel was
cooled in an ice bath and slowly depressurized. The product mixture was
analyzed by GC.
’
ACKNOWLEDGMENT
This work was supported by the Cluster of Excellence “Tailor
Made Fuels from Biomass” (TMFB), which is funded by the
Excellence Initiative of the German federal and state govern-
ments to promote science and research at German universities.
We are grateful for generous allocation of computer time by the
Centre for Computing and Communication (RZ) of RWTH
Aachen University. We thank Mrs. Ines Bachmann-Remy for the
acquisition of the 600 MHz NMR spectra.
Synthesis of Carbonyldihydrido(1,1,1-tris(diphenylphos-
phinomethyl)ethane)ruthenium(II) (4). A glass liner of a 10 mL
stainless steel high-pressure reactor was loaded with Ru(acac) (1; 100
3
mg, 0.25 mmol, 1 equiv) and TriPhos (2; 174.9 mg, 0.27 mmol,
’
REFERENCES
(1) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.;
1.1 equiv) and the liner was equipped with a magnetic stirrer bar and
transferred into the corresponding reactor. The atmosphere was chan-
ged to argon, and propanal (2.5 mL) was added. The reactor was
pressurized with 120 bar of hydrogen and stirred for 20 h at 150 °C. The
reactor was cooled to 0 °C, and the pressure was released into a rubber
balloon. The product mixture was transferred into a Schlenk flask and
ethanol (1 mL) was added. The solution was filtered off, and the solid
was washed with ethanol (1 mL) twice to yield the product (4) as light
yellow powder (166.3 mg, 0.21 mmol, 84%). RuC H O MW: 785.14
Cairney, J.; Eckert, C. A.; Frederick, W. J., Jr.; Hallett, J. P.; Leak, D. J.;
Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.; Tschaplinski, T.
Science 2006, 311, 484.
(
2) Kamm, B.; Kamm, M.; Gruber, P. R.; Kromus, S. In Biorefineries
—
Industrial Processes and Products. Status Quo and Future Directions;
Kamm, B., Gruber, P. R., Kamm, M., Eds.; Wiley-VCH: Weinheim,
Germany, 2006; Vol. 1, p 3.
(3) Corma, A.; Iborra, S.; Velty, A. Chem. Rev. 2007, 107, 2411.
1
5 21
1
3
g/mol. H NMR (600 MHz, CD Cl , 25 °C): δ 7.68 (t, 8 H, J = 7.8
Hz, arom), 7.32 (t, 4 H, J = 8.7 Hz, arom), 7.17 (t, 2 H, J = 7.2 Hz,
2
2
HH
(
4) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106,
3
3
HH
HH
4044.
arom), 7.06ꢀ7.14 (m, 8 H, H-12, arom), 7.02 (m, 8 H, arom), 2.31
(
5) Top Value-Added Chemicals from Biomass ꢀ Results of Screening
3
2
PH HH 2
(
dd, 2 H, J = 15.1 Hz, J = 6.5 Hz, CH equatorial), 2.20 (dd, 2 H,
for Potential Candidates from Sugars and Synthesis Gas; Werpy, T.;
Petersen, G., Eds.; U.S. Department of Energy: Washington, DC, 2004,
Vol. 1.
(6) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539.
(7) Geilen, F. M. A.; Engendahl, B.; Harwardt, A.; Marquardt, W.;
Klankermayer, J.; Leitner, W. Angew. Chem., Int. Ed. 2010, 49, 5510.
(8) Van Engelen, M. C.; Teunissen, H. T.; de Vries, J. G.; Elsevier, C.
J. J. Mol. Catal. A: Chem. 2003, 206, 185.
3
2
3
JPH = 15.1 Hz, J = 6.5 Hz, CH equatorial), 2.13 (d, 2 H, J = 8.2 Hz,
HH
2
PH
3
CH axial), 1.52 (s, 3 H, CH ), ꢀ7.30 ppm. (dd, 2 H, J = 18.3 Hz, 50.2
2
3
PH
31
3
Hz, RuꢀH). P NMR (242.9 MHz, CD Cl , 25 °C): δ 34.5 (t, 1 P, J
=
2
2
PP
3
13
3
1.2 Hz, P-axial), 26.7 ppm (d, 2 P, JPP = 31.2 Hz, P-equatorial). C NMR
(
1
150.1 MHz, CD Cl
28.4, 127.8, 127.5, 38.8, 38.4, 35.0, 33.5 ppm.
2
2
, 25 °C): δ 209.5, 139.8, 138.0, 132.4, 131.1, 128.6,
Reaction of Carbonyldihydrido(1,1,1-tris-(diphenylphos-
phinomethyl)ethane)ruthenium(II) (4) with Levulinic Acid.
(
9) Comprehensive Asymmetric Catalysis, Jacobsen, E. N.; Pfaltz,
A.; Yamamoto, H., Eds. Springer, Heidelberg, Germany, 1999.
(
(
In an NMR tube complex 4 (15.5 mg, 21 μmol) was dissolved in CD Cl
2
2
10) Hara, Y.; Wada, K. Chem. Lett. 1991, 553.
11) Mehdi, H.; F ꢀa bos, V.; Tuba, R.; Bodor, A.; Mika, L. T.; Horv ꢀa th,
(0.7 mL). Levulinic acid (4.0 mg, 34 μmol) was added, and the mixture
was shaken for 30 s. The color of the solution changed from light yellow
1
I. T. Top. Catal. 2008, 48, 49.
to orange within the first 10 min. After 15 min the first H NMR was
1
(12) Ohkuma, T.; Noyori, R. In Handbook of Homogeneous Hydro-
genation; de Vries, J. G., Elsevier, C. J., Eds.; Wiley-VCH: Weinheim,
Germany, 2007; Vol. 4, p 1105.
(
(
(
Commun. 2007, 3154.
(16) Rosi, L.; Frediani, M.; Frediani, P. J. Organomet. Chem. 2010,
695, 1314.
recorded and for the next 18 h H NMR were recorded hourly. m/z:
1
8
41.6 (calc: 842.2). H NMR (600 MHz, CD
2
Cl
2
, 25 °C): δ 7.96 (m, 2
H, arom), 7.80 (m, 2 H, arom) 7.56 (m, 2 H, arom), 7.32 (m, 1 H, arom),
13) Teunissen, H. T.; Elsevier, C. J. Chem. Commun. 1997, 667.
14) Teunissen, H. T.; Elsevier, C. J. Chem. Commun. 1998, 1367.
15) Nunez, A. A.; Eastham, G. R.; Cole-Hamilton, D. J. Chem.
7
1
.36 (m, 2 H, arom), 7.25 (m, 2 H, arom), 7.08 (m, 1 H, arom), 6.99 (m,
H, arom), 6.78 (m, 2 H, arom), 2.72 (m, 2 H, LA-CH ), 2.62 (m, 2 H,
2
LA-CH
2
), 2.48 (m, 1 H, TriPhos-CH
2
2
), 2.46 (m, 1 H, TriPhos-CH ),
2
.38 (m, 1 H, TriPhos-CH ), 2.37 (m, 1 H, TriPhos-CH ), 2.29 (m, 1 H,
2
2
TriPhos-CH ), 2.25 (s, 3 H, LA-CH ), 2.23 (m, 1 H, TriPhos-CH ),
2
3
2
2
.16 (m, 1 H, TriPhos-CH
2
), 1.59 (m, 3 H, TriPhos-CH
3
), ꢀ5.89
(17) Bianchini, C.; Meli, A.; Moneti, S.; Vizza, F. Organometallics
1998, 17, 2636.
(18) Bianchini, C.; Meli, A.; Moneti, S.; Oberhauser, W.; Vizza, F.;
3
3
3
ppm (ddd, 1 H,
JPaxH = 94.7 Hz, JPeqH = 19.7 Hz, JP H = 14.3
eq
31
Hz, RuꢀH). P NMR (242.9 MHz, CD Cl , 25 °C): δ 47.8 (dd, 1
2
2
3
3
3
Herrera, V.; Fuentes, A.; Sanchez-Delgado, R. A. J. Am. Chem. Soc. 1999,
P, JPP = 40.2 Hz, JPP = 19.0 Hz, P-equatorial), 19.0 (dd, 1 P, JPP = 40.2
3
3
3
1
21, 7071.
19) Bianchini, C.; Meli, A.; Peruzzini, M.; Vizza, F.; Zanobini, F.
Coord. Chem. Rev. 1992, 120, 193.
20) Belkova, N. V.; Bakhmutova, E. V.; Shubina, E. S.; Bianchini, C.;
Peruzzini, M.; Bakhmutov, V. I.; Epstein, L. M. Eur. J. Inorg. Chem. 2000,
Hz, JPP = 28.9 Hz, P-equatorial), 4.4 ppm (dd, 1 P, JPP = 28.9 Hz, JPP
9.0 Hz, P-axial).
=
(
1
(
’
ASSOCIATED CONTENT
Supporting Information. Tables listing all calculated
2163.
S
b
(21) Bianchini, C.; Moneti, S.; Peruzzini, M.; Vizza, F. Inorg. Chem.
energies in hartrees and Cartesian coordinates for all structures
and the text giving the complete ref 54. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
997, 36, 5818.
(22) Bakhmutov, V. I.; Bianchini, C.; Peruzzini, M.; Vizza, F.;
Vorontsov, E. V. Inorg. Chem. 2000, 39, 1655.
1
4357
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358

Journal of the American Chemical Society
23) Bakhmutov, V. I.; Bakhmutova, E. V.; Belkova, N. V.; Bianchini,
ARTICLE
(
functions, as only marginal differentiation of the two regioisomers is
7
C.; Epstein, L. M.; Masi, D.; Peruzzini, M.; Shubina, E. S.; Vorontsov,
observed experimentally.
E. V.; Zanobini, F. Can. J. Chem. 2001, 75, 479.
(53) Dobereiner, G. E.; Nova, A.; Schley, N. D.; Hazari, N.;
Miller, S. J.; Eisenstein, O.; Crabtree, R. J. Am. Chem. Soc. 2011, 133,
7547.
(54) Frisch, M. J. et al. Gaussian 09, Revision A.02; Gaussian:
Wallingford, CT, 2009.
(
24) Lough, A. J.; Morris, R. H.; Ricciuto, L.; Schleis, T. Inorg. Chim.
Acta 1998, 270, 238.
25) Albinati, A.; Klooster, W. T.; Koetzle, T. F.; Fortin, J. B.; Ricci,
(
J. S.; Eckert, J.; Fong, T. P.; Lough, A. J.; Morris, R. H.; Golombek, A. P.
Inorg. Chim. Acta 1997, 259, 351.
(55) Grimme, S. J. Comput. Chem. 2006, 27, 1211.
(56) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297.
(57) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97,
2571.
(58) Schaefer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
5829.
(26) Grocott, S. C.; Skelton, B. W.; White, A. H. Aust. J. Chem. 1983,
36, 259.
(
(
27) Chaplin, A. B.; Dyson, P. J. Inorg. Chem. 2008, 47, 381.
28) Morton, D.; Cole-Hamilton, D. J.; Utuk, I. D.; Paneque-Sosa,
M.; Lopez-Poveda, M. J. Chem. Soc., Dalton Trans. 1989, 489.
(
(
(
29) Zhao, J.; Hartwig, J. F. Organometallics 2005, 24, 2441.
30) Sieffert, N.; B €u hl, M. J. Am. Chem. Soc. 2010, 132, 8056.
31) Johansson, A. J.; Zuidema, E.; Bolm, C. Chem.—Eur. J. 2010,
(59) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Theor.
Chem. Acc. 1997, 97, 119.
(60) Weigend, F.; Furche, F.; Ahlrichs, R. J. Chem. Phys. 2003,
119, 12753.
1
6, 13487.
(
32) (a) Kilner, M.; Tyers, D. V.; Crabtree, S. P.; Wood, M. A. Davy
(61) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Process Technology Ltd., London, U.S. Patent Application 2005/0234269
A1, 2005. (b) Wood, M.; Crabtree, S. P.; Tyers, D. V.; WO05/051875, 2005.
Theor. Chim. Acta 1990, 77, 123.
(62) Metz, B.; Stoll, H.; Dolg, M. J. Chem. Phys. 2000, 113, 2563.
(63) Dunlap, B. I. J. Chem. Phys. 1983, 78, 3140.
(64) Dunlap, B. I. J. Mol. Struct. (THEOCHEM) 2000, 529, 37.
(65) Knowles, T. S.; Howells, M. E.; Howlin, B. J.; Smith, G. W.;
Amodio, C. A. Polyhedron 1994, 13, 2197.
(
33) Hara, Y.; Kusaka, H.; Inagaki, H.; Takahashi, K.; Wada, K.
J. Catal. 2000, 194, 188.
34) When the acidic ionic liquid 3 was used as a proton source,
(
formation of a similar monohydridic fragment was observed. Compared
to levulinic acid, the reaction was fast at room temperature and gas
formation was observed.
(
2
35) The possibility of one PPh group of TriPhos temporarily
opening up an additional coordination site has been demonstrated by
Horv ꢀa th in rhodium-catalyzed hydroformylation: Kiss, G.; Horv ꢀa th, I. T.
Organometallics 1991, 10, 3798 This may provide additional pathways
+
+
for example on the basis of [(P∩P2)Ru(CO)H] or [(P∩P2)RuH]
fragments. All transformations in the present study could be accom-
plished on the basis of three coordination sites in facial proximity.
(
(
36) Rappert, T.; Yamamoto, A. Organometallics 1994, 13, 4984.
37) Ayllon, J. A.; Gervaux, C.; Sabo-Etienne, S.; Chaudret, B.
Organometallics 1997, 16, 2000.
(38) Zhao, J.; Hartwig, J. F. Organometallics 2005, 24, 2441.
(39) Takebayashi, S.; Bergens, S. H. Organometallics 2009, 28, 2349.
(40) (a) Hutschka, F.; Dedieu, A.; Leitner, W. Angew. Chem., Int. Ed.
Engl. 1995, 34, 1742. (b) Hutschka, F.; Dedieu, A.; Eichberger, M.;
Fornika, R.; Leitner, W. J. Am. Chem. Soc. 1997, 119, 4432.
(41) (a) Kozuch, S.; Shaik, S. J. Am. Chem. Soc. 2006, 128, 3355.
(
(
b) Uhe, A.; Kozuch, S.; Shaik, S. J. Comput. Chem. 2011, 32, 978.
c) Kozuch, S.; Shaik, S. Acc. Chem. Res. 2011, 44, 101.
(
42) Saudan, L. A.; Saudan, C. M.; Debieux, C.; Wyss, P. Angew.
Chem., Int. Ed. 2007, 46, 7473.
43) (a) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew.
(
Chem., Int. Ed. 2006, 45, 1113. (b) Zhang, J.; Leitus, G.; Ben-David, Y.;
Milstein, D. J. Am. Chem. Soc. 2009, 127, 12429.
(
44) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Angew. Chem., Int.
Ed. Engl. 1990, 29, 256.
45) Evans, D. A.; Rovis, T.; Kozlowski, M. C.; Tedrow, J. S. J. Am.
Chem. Soc. 1999, 121, 1994.
(
(
46) Daley, C. J. A.; Bergens, S. H. J. Am. Chem. Soc. 2002, 124, 3680.
(47) Kapoor, P.; Pathak, A.; Kapoor, R.; Venugopalan, P.; Corbella,
M.; Rodríguez, M.; Robles, J.; Llobet, A. Inorg. Chem. 2002, 41, 6153.
48) Meiere, S. H.; Ding, F.; Friedman, L. A.; Sabat, M.; Harman,
W. D. J. Am. Chem. Soc. 2002, 124, 13506.
49) Meng, Q.; Zhang, C.; Huang, M.-B. Can. J. Chem. 2009, 87,
610.
50) Lebedev, A.; Leite, L.; Fleisher, M.; Stonkus, V. Chem. Hetero-
cycl. Compd. 2000, 36, 775.
51) Handbook of Homogeneous Hydrogenation; de Vries, J. G.,
Elsevier, C. J., Eds.; Wiley-VCH: Weinheim, Germany, 2007.
52) The sodium carboxylates of MSA and IA could not be hydro-
(
(
1
(
(
(
genated with the present catalytic system. This experimental finding is in
line with the free acid rather than the carboxylate group as starting point.
The cycle was studied for only one of the two very similar carboxylic acid
1
4358
dx.doi.org/10.1021/ja2034377 |J. Am. Chem. Soc. 2011, 133, 14349–14358