Catalytic Transformation of Levulinic Acid to 2-Methyltetrahydrofuran Using Ruthenium - N-Triphos Complexes

Article abstract of DOI:10.1021/cs502025t

A series of pre- or in situ-formed ruthenium complexes were assessed for the stepwise catalytic hydrogenation of levulinic acid (LA) to 2-methyltetrahydrofuran (2-MTHF) via γ-valerolactone (γVL) and 1,4-pentanediol (1,4-PDO). Two different catalytic systems based on the branched triphosphine ligands Triphos (CH₃C(CH₂PPh₂)₃) and N-triphos (N(CH₂PPh₂)₃) were investigated. The most active catalyst was the preformed ruthenium species [RuH₂(PPh₃){N(CH₂PPh₂)₃-κ³P}] (5), which gave near quantitative conversion of LA to 1,4-PDO when no acidic additives were present, and 87% 2-MTHF when used in conjunction with HN(Tf)₂. Various acidic additives were assessed to promote the final transformation of 1,4-PDO to 2-MTHF; however, only HN(Tf)₂ was found to be effective, and NH₄PF₆ and para-toluenesulfonic acid (p-TsOH) were found to be detrimental. Mechanistic investigations were carried out to explain the observed catalytic trends and importantly showed that PPh₃ dissociation from 5 resulted in its improved catalytic reactivity. The presence of acidic additives removes catalytically necessary hydride ligands and may also compete with the substrate for binding to the catalytic metal center, explaining why only an acid with a noncoordinating conjugate base was effective. Crystals suitable for X-ray diffraction experiments were grown for two complexes: [Ru(NCMe)₃{N(CH₂PPh₂)₃-κ³P}] (14) and [Ru₂(μ-Cl)₃{N(CH₂PPh₂)₃-κ³P}₂][BPh₄] (16). (Chemical Equation Presented).

Full text of DOI:10.1021/cs502025t

Research Article
pubs.acs.org/acscatalysis
Catalytic Transformation of Levulinic Acid to
2
‑Methyltetrahydrofuran Using Ruthenium−N‑Triphos Complexes
Andreas Phanopoulos, Andrew J. P. White, Nicholas J. Long,* and Philip W. Miller*
Department of Chemistry, Imperial College London, South Kensington, London, SW7 2AZ, United Kingdom
*
S Supporting Information
ABSTRACT: A series of pre- or in situ-formed ruthenium
complexes were assessed for the stepwise catalytic hydro-
genation of levulinic acid (LA) to 2-methyltetrahydrofuran (2-
MTHF) via γ-valerolactone (γVL) and 1,4-pentanediol (1,4-
PDO). Two different catalytic systems based on the branched
triphosphine ligands Triphos (CH C(CH PPh ) ) and N-
3
2
2 3
triphos (N(CH PPh ) ) were investigated. The most active
2
2 3
catalyst was the preformed ruthenium species [RuH (PPh )-
2
3
3
{
N(CH PPh ) -κ P}] (5), which gave near quantitative
2 2 3
conversion of LA to 1,4-PDO when no acidic additives were
present, and 87% 2-MTHF when used in conjunction with
HN(Tf) . Various acidic additives were assessed to promote
2
the final transformation of 1,4-PDO to 2-MTHF; however, only HN(Tf) was found to be effective, and NH PF and para-
2
4
6
toluenesulfonic acid (p-TsOH) were found to be detrimental. Mechanistic investigations were carried out to explain the observed
catalytic trends and importantly showed that PPh dissociation from 5 resulted in its improved catalytic reactivity. The presence
3
of acidic additives removes catalytically necessary hydride ligands and may also compete with the substrate for binding to the
catalytic metal center, explaining why only an acid with a noncoordinating conjugate base was effective. Crystals suitable for X-ray
3
diffraction experiments were grown for two complexes: [Ru(NCMe) {N(CH PPh ) -κ P}] (14) and [Ru (μ-Cl) {N-
3
2
2 3
2
3
3
(
CH PPh ) -κ P} ][BPh ] (16).
2 2 3 2 4
KEYWORDS: N-triphos, ruthenium, catalysis, levulinic acid, hydrogenation, triphos, 1,4-pentanediol, 2-methyltetrahydrofuran, biomass
1
e,8
INTRODUCTION
found uses as a biofuel after further esterification,
or as a
■
9
bioderived solvent (Scheme 1). This hydrogenation step has
been well-established, and many homogeneous and heteroge-
neous catalysts have been reported to efficiently and selectively
afford γVL.
The inherent stability of γVL makes its subsequent
The production of fuels and platform chemicals from materials
other than nonrenewable resources is central to achieving a
sustainable future. Lignocellulosic biomass represents the
largest source of biorenewable carbohydrates and, as such, is
a prime candidate for derivatization into other more valuable
compounds. Consequently, there has recently been a great deal
of academic and industrial interest in its conversion to other
9,10
6,11
transformation difficult and often requires harsh conditions.
Perhaps most interestingly, a tandem ring-opening/hydro-
genation pathway allows the formation of 1,4-pentanediol (1,4-
PDO), and a final dehydration affords 2-methyltetrahydrofuran
Biogenic diols can be used as
monomers to produce high-performance biodegradable poly-
1
−5
products.
Biomass-derived raw materials in general contain
excess functionality, and maintaining only the desired
functionality in the final product is key to their successful
implementation in upstream industrial and consumer items.
In general, these raw materials must undergo a partial
deoxygenation process, normally via dehydration reactions, to
afford well-defined, so-called platform chemicals. Levulinic acid
6
,12
(
2-MTHF) (Scheme 1).
1
b,6
6
,13
esters as well as fine-chemical intermediates.
To an even
greater extent than γVL, 2-MTHF has been championed as a
14−16
bioderived “green” solvent of the future.
Apart from its
sustainable production method, 2-MTHF has several benefits
over other ethereal solvents, such as a relatively high boiling
point, and immiscibility with water, facilitating workup and
(
LA) is one such platform chemical and can be readily
produced during acid treatment of sugar monomers. Initial
dehydration of many sugars, such as glucose, fructose, sucrose,
etc., affords the cyclic aldehyde 5-hydroxymethylfurfural (5-
14
recovery.
1
7
Recently, several reports have highlighted the use of a
tripodal phosphine ligand, the so-called “triphos” scaffold, for
HMF), a versatile platform chemical in its own right. Under
aqueous, acidic conditions, 5-HMF will undergo a ring-opening
step to afford levulinic and formic acid (Scheme 1). Starting
from a levulinic acid platform, the production of many desirable
molecules is possible. Under hydrogenation conditions, LA can
be initially transformed into γ-valerolactone (γVL), which has
Received: December 16, 2014
Revised: March 7, 2015
Published: March 9, 2015
©
2015 American Chemical Society
2500
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
Scheme 1. Reaction Sequence for the Transformation of Bioderived Sugar Monomers into Levulinic Acid via 5-
Hydroxymethylfurfural, And Subsequent Valorization via Sequential Hydrogenation and Dehydration Reactions To Ultimately
Afford 2-Methyltetrahydrofuran
Figure 1. Carbon- and nitrogen-centered triphosphine ligands and their complexes investigated in this work.
Table 1. Screening of Preformed and in-Situ-Generated Complexes for the Catalytic Hydrogenation of Levulinic Acid to 1,4-
a
Pentanediol.
d
yield, %
entry
1
catalyst
temperature, °C
pressure, bar
γVL
1,4-PDO
2-MTHF
ref
Ph
Triphos /[Ru(acac) ]
150
160
150
160
150
150
150
150
65
100
65
9
3
83
95
2
0
0
this work
12
3
b
Ph
2
Triphos /[Ru(acac) ]
3
3
3
85
22
60
53
1
0
this work
24
c
4
3
100
65
73
37
36
99
1
3
Ph
5
6
7
8
N-triphos /[Ru(acac) ]
2
this work
this work
this work
this work
3
4
5
6
65
<1
0
65
65
61
1
a
b
Conditions: 10 mmol LA, 20 mL THF, 0.5 mol % complex or 0.5 mol % [Ru(acac) ] and 1.0 mol % ligand, reaction time 25 h. 10 mmol LA, no
3
Ph
c
d
solvent, 0.1 mol % [Ru(acac) ], 0.2 mol % Triphos , reaction time 18 h. 10 mmol LA, no solvent, 0.1 mol % complex, reaction time 18 h. Yield
3
determined by GC analysis; full conversion was achieved in all cases.
the formation of catalysts for use in conventionally difficult
Herein, we report the use of a relatively unexplored family of
nitrogen-centered triphos ligands (Figure 1) for use as
homogeneous catalysts for the hydrogenation of levulinic
acid. The N-triphos ligand differs only from the parent triphos
ligand at the central bridgehead atom by replacement of a C−
4
,12,17−23
catalytic transformations.
The dominance of the
triphos ligand is likely to be due to the inherent stability
imparted via multidentate coordination of three phosphine
arms, preventing or at least slowing catalyst decomposition.
CH with a nitrogen atom. The main advantage of this N-
Ph
3
The eponymous Triphos ligand (1) (Figure 1) has been used
triphos ligand system over the carbon-centered analogue is its
in a multicomponent, ruthenium-based system that was found
ease of synthesis and accessibility to a much wider range of
to selectively produce γVL, 1,4-PDO, or 2-MTHF, depending
17a,25−29
phosphine arm substituents.
In this report, catalysts
12
on the conditions used. The use of acidic additives was found
to be critical for transforming 1,4-PDO to 2-MTHF, and
were generated in situ either from ruthenium(III) precursors or
2
9
from preformed ruthenium(II) complexes. Mechanistic
studies were undertaken to explain the observed product
distribution and catalytic activity of these N-triphos systems.
+
consequently, a catalytic quantity of H was incorporated into
2
4
the proposed catalytic mechanism. A similarly tunable
heterogeneous system using highly dispersed copper in a
zirconia matrix was reported to effectively transform γVL into
RESULTS AND DISCUSSION
■
1
,4-PDO or 2-MTHF if the catalyst was prepared at lower
Conversion of Levulinic Acid to 1,4-PDO. The hydro-
genation of levulinic acid was investigated initially using in situ
6
calcination temperatures. Like other heterogeneous catalysts,
however, this catalyst requires very high temperatures (>200
catalysts formed from [Ru(acac) ] and the diphenylphosphino-
3
Ph
Ph
°C), decreasing its viability for industrial use.
containing ligands Triphos (1) and N-triphos (2). The
2
501
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
Scheme 2. Synthesis of Ruthenium Dihydride Carbonyl Complexes 3 And 4 Under Mild Conditions
Table 2. Screening of Preformed and in-Situ-Generated Complexes for the Catalytic Hydrogenation of Levulinic Acid to 2-
a
Methyltetrahydrofuran Using Acidic Additives
c
yield, %
entry
catalyst
additive
NH PF
pressure, bar
γVL
1,4-PDO
2-MTHF
ref
b
Ph
1
Triphos /[Ru(acac) ]
100
100
65
8
58
73
70
60
95
77
68
54
10
35
1
53
39
2
12
12
3
4
6
b
Ph
2
Triphos /[Ru(acac) ]
p-TsOH
3
Ph
3
4
5
6
7
8
9
Triphos /[Ru(acac) ]
NH PF
5
this work
this work
this work
this work
this work
this work
this work
this work
3
4
6
6
6
6
3
NH PF
65
2
0
4
Ph
N-triphos /[Ru(acac) ]
NH PF
65
35
<1
0
5
3
4
4
4
NH PF
65
0
4
p-TsOH
65
0
5
NH PF
65
8
<1
45
87
4
6
Ph
N-triphos /[Ru(acac) ]
HN(Tf)2
HN(Tf)2
65
0
3
10
5
65
<1
a
Conditions: 10 mmol LA, 20 mL THF, 0.5 mol % complex or 0.5 mol % [Ru(acac) ] and 1.0 mol % ligand, 5.0 mol % additive, 150 °C, reaction
3
b
Ph
c
time 25 h. 10 mmol LA, no solvent, 0.1 mol % [Ru(acac) ], 0.2 mol % Triphos , 1.0 mol % additive, 160 °C, reaction time 18 h. Yield determined
by GC; full conversion was achieved in all cases.
3
reaction conditions employed, 65 bar and 150 °C, were milder
than those of 100 bar and 160 °C previously reported by
Lietner et al. In line with the reactivity reported by Leitner et
used during this study (65 bar H , 150 °C), when 3 was
implemented as the catalytic species, formation of γVL was
2
1
2
predominant (Table 1, entry 3). This is in contrast to the
Ph
al., when the 1/[Ru(acac) ] system was used, good conversion
analogous N-triphos complex [RuH (CO){N(CH PPh ) -
3
2
2
2 3
3
to 1,4-PDO was observed (83%), with small amounts of γVL
κ P}] (4), where both the preformed complex (Table 1, entry
6), and in-situ-generated system (Table 1, entry 5) gave similar
product distributions. Despite previous reports showing 3 can
(
9%) and no 2-MTHF (Table 1, entry 1). When the ligand was
switched to the nitrogen-centered analogue 2, the product
distribution was considerably altered, with γVL being the major
product. Interestingly, trace amounts of 2-MTHF were also
observed in this reaction (Table 1, entry 5).
be isolated from catalytic reaction mixtures of 1/[Ru(acac) ]/
3
12
LA, under the conditions used during this study, it appears 3
is not produced from 1/[Ru(acac) ] mixtures, as evident from
3
Encouraged by these in situ catalytic results, we sought to
investigate various well-defined ruthenium complexes. Several
reports have highlighted carbonyl dihydride species as being
the different product distributions. Conversely, the similar
product distribution of 4 and 2/[Ru(acac) ] suggest that the
3
same catalytically active species is produced in both cases for
“
dormant” forms of the active catalyst during similar
the nitrogen-centered ligands.
1
2,24
reactions;
hence, such species could provide an excellent
Previous reactivity studies of 4 suggested that the carbonyl
ligand is robust and will remain coordinated to ruthenium, even
precatalyst. Recently, we reported the facile and high-yielding
synthetic routes to a series of ruthenium carbonyl dihydride
complexes, prepared from N-triphos ligands (Scheme 2). We
29
under acidic conditions. It was anticipated that by substituting
29
the CO ligand in 4 with a more labile PPh ligand, LA
3
also wished to make comparisons with the previously reported
hydrogenation would be more easily facilitated. This hypothesis
was based on the mechanism for the catalytic cycle proposed by
Leitner et al., which shows that the CO ligand on the complex
Ph
carbonyl dihydride complex featuring Triphos [RuH (CO)-
2
3
{
CH C(CH PPh ) -κ P}] (3); however, the reported synthesis
3
2
2 3
24
proved to be unsatisfactory, involving either several highly air-
is replaced by an oxygen bound carbonyl moiety of LA. The
triphenylphosphine dihydride complex [RuH (PPh ){N-
3
0,31
and moisture-sensitive steps
or harsh reaction conditions at
2
3
24
3
high temperatures and pressures. Complex 3, however, could
be easily prepared in high yields via the formation of the stable
intermediate carbonate complex, followed by a low-pressure
hydrogenation to give 3 (Scheme 2).
(CH PPh ) -κ P}] (5) was found to be much more active for
2 2 3
the hydrogenation of LA to 1,4-PDO, giving almost
quantitative conversion (Table 1, entry 7), thus substantiating
this hypothesis.
Complex 3 has previously been tested as a catalyst for the
To probe complex 5 further, the diphenylphosphine groups
were exchanged for more bulky and electron-donating
dicyclopentylphosphino groups: [RuH (PPh ){N-
hydrogenation of LA at 100 bar H and 160 °C (Table 1, entry
2
2
4
4
), and although it gave poorer yields of 1,4-PDO than the
2
3
3
related in-situ-generated system (1/[Ru(acac) ]) under the
(CH PCyp ) -κ P}] (6) (Table 1, entry 8). Disappointingly,
3
2
2 3
12
same conditions (Table 1, entry 2), the major product was
still found to be 1,4-PDO (73% and 95%, respectively).
Interestingly, at the lower pressure and temperature conditions
however, the catalysis did not improve and switched the
product distribution back to predominantly γVL. In addition, a
greater percentage of products from side reactions was
2
502
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
observed, consisting mainly of other alcohols. It is uncertain
whether these changes are the result of increased steric bulk
around the metal center, hindering coordination and activation
of the substrate, or due to the ruthenium’s being more electron-
rich as a result of the donating dialkylphosphines.
Complex 5 is therefore a promising precatalyst for the
conversion of LA to 1,4-PDO because it is both air- and
moisture-stable in the solid state. Despite the ability of 5 to
quantitatively convert LA to 1,4-PDO, further catalytic
dehydration of 1,4-PDO to 2-MTHF is also highly desirable.
This process requires acidic reaction conditions; hence,
catalytic testing was undertaken in the presence of three
different proton sources to investigate the formation of 2-
MTHF.
conditions. In the presence of NH PF , a shift to predom-
4
6
inantly γVL was observed when 4 was used as the precatalyst
(Table 2, entry 6). Here, apparently, different catalytically
relevant species are generated. Attempts to increase the yield of
1,4-PDO and push the reaction further along the hydro-
genation pathway by varying the amount of acid, pressure, or
solvent only proved detrimental (see Supporting Information
Table S2).
Increasing the acidity of reaction mixture by using p-TsOH
instead of NH PF again showed no improvement for the
4
6
production of 1,4-PDO or 2-MTHF (Table 2, entry 7). Similar
18
to the work reported by Cole-Hamilton et al., p-TsOH may
be coordinating to the ruthenium under the catalytic conditions
and, rather than acting as a reservoir for the catalytically active
+
Conversion of Levulinic Acid to 2-MTHF. Experiments
to convert LA to 2-MTHF were carried out using three
different acidic additives, NH PF , para-toluenesulfonic acid (p-
species, is removing the [Ru(triphos)] fragment from the
catalytic cycle (vide supra).
The use of complex 5 as the precatalytic species, which gave
almost quantitative conversion to 1,4-PDO in the absence of
acid, was similarly deactivated in the presence of NH PF as 4
4
6
TsOH) and bis(trifluoromethane)sulfonimide (HN(Tf) ), to
2
achieve the desired acidic conditions for this catalytic
4
6
transformation (Table 2). NH PF and p-TsOH have
(Table 2, entry 8). Clearly, the acidic additive is interacting with
the catalyst, resulting in significant catalytic inhibition beyond
γVL. Because this same inhibition was not previously observed
during similar reactions conducted in neat LA, it is possible
that the ability of acidic components to deactivate the catalyst is
solvent- or dilution-dependent. Indeed, changing the solvent
from THF to dioxane was found to alter the yield of γVL when
4 was used as the precatalyst (Table S2, entry 5).
4
6
previously been implemented by Leitner et al. and displayed
moderate to good yields for the production of 2- and 3-MTHF
from biomass-derived carboxylic acids (Table 2, entries 1 and
12
1
2
2
). It should be noted that previous experiments were
12
conducted under melt conditions (neat LA); however, to
ensure good incorporation of H in the present study, a
2
minimum volume of solution was required that necessitated the
use of solvent. This was due to limitations in the experimental
setup.
Finally, the use of a strong acid featuring a highly
noncoordinating anion (HN(Tf) ) was evaluated. If the
2
In addition, the hydrogenation of amides to amines has been
investigated in collaborative work between the groups of
coordination of potentially ligating species formed in situ is
responsible for the catalyst inhibition, the removal of this
Ph
Leitner and Cole-Hamilton using ruthenium/Triphos systems
deactivation pathway should increase catalyst efficacy. Gratify-
18
Ph
with methanesulfonic acid (MSA) as an acidic additive. The
role of MSA in these reactions was studied in great detail, and it
was found that it acts to remove any hydride ligands as H2
before coordinating the ruthenium center. Various coordination
modes of MSA were found to exist in equilibrium as a mixture
of compounds featuring mono- or bidentate MSA as well as
bridged dimers. The authors suggest this equilibrium mixture
ingly, using a N-triphos /[Ru(acac) ] catalytic system in the
3
presence of 5 mol % HN(Tf) , 2-MTHF was afforded in a 45%
2
yield (Table 2, entry 9). If the product distribution of this
catalyst run is compared with the same catalyst system in the
absence of acidic components (Table 1, entry 5), similar yields
of γVL are observed in both cases. Without an acidic
component, the conversion of 1,4-PDO to 2-MTHF cannot
be achieved (because this is acid-catalyzed), and consequently,
little 2-MTHF was observed in the absence of acidic cocatalysts.
+
acts as a reservoir for the [Ru(triphos)] fragment, which is the
18
entry point for the catalytic cycle.
Using the same adapted conditions that gave excellent yields
of 1,4-PDO from LA, the addition of acidic components to the
reaction mixture was found to be detrimental to the catalytic
process in most cases. Initial comparison of the adapted
With the inclusion of HN(Tf) , it appears that any 1,4-PDO
that is generated is rapidly converted to 2-MTHF without the
acid hindering the catalytic activity of the ruthenium complex.
2
When complex 5 is used in conjunction with HN(Tf) as the
2
conditions used during this study with those previously
acidic cocatalyst (Table 2, entry 10), the majority of LA is
converted to 2-MTHF (87%), making this catalytic system
among the most active for this transformation, and under
relatively mild conditions compared with those previously
Ph
reported using a Triphos /[Ru(acac) ] catalytic system
3
showed that in the presence of NH PF , the catalytic cycle
4
6
was blocked, and almost no 2-MTHF was obtained (Table 2,
entry 3). In fact, the acid appeared to significantly alter the
catalyst and shift the major product from 1,4-PDO, as was seen
in the unacidified experiment (Table 1, entry 1), to γVL. Using
preformed catalyst 3, similar results to the in-situ-generated
species were obtained (Table 2, entry 4), suggesting that using
either precursor ultimately leads to the same catalytic active
species.
1
2,24
reported.
Similar to the in-situ-generated catalyst (Table 2,
entry 9), HN(Tf)2 does not hinder the activity of the
ruthenium species, allowing the transition metal component
to convert LA to 1,4-PDO, at which point the acidic
component can catalyze the final transformation of 1,4-PDO
to 2-MTHF. The different product distributions obtained when
various acidic additives are utilized suggest that this additive
must be not only highly acidic, but also highly noncoordinating.
The homogeneous nature of the catalyst was implied by
mercury poisoning experiments, which demonstrated that
Next, the N-triphosPh systems were considered. Unlike the
carbon-centered counterparts, a similar product distribution
Ph
was observed for the N-triphos /[Ru(acac) ] system regard-
3
32
less of whether acidic additives were used (Table 1, entry 5 and
elemental mercury did not alter the activity of the catalysts.
Table 2, entry 5). In the absence of acid, the use of either N-
In addition, upon completion of catalytic runs, if efforts were
taken to ensure no oxidative decomposition of any ruthenium
species present and the final reaction mixture was subjected to
Ph
triphos /[Ru(acac) ] or 4 as the catalytic system gave similar
3
product distributions, but this was not replicated under acidic
2
503
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
fractional distillation to remove catalytic products, well-defined
ruthenium complexes were always found to be present.
Qualitatively, the reaction mixture maintained a bright
orange/red color, indicative of ruthenium−N-triphos com-
plexes. If catalyst decomposition to ruthenium nanoparticles or
In the presence of NH PF in acetonitrile, both complexes 3
4
6
and 4 showed identical reactivity (Scheme 4), releasing
hydrogen and forming cationic, solvent-bound complexes 8
29
and 9, respectively. The reactivity of 9 with LA has been
previously reported by us, during an initial screening of the
reactivity of ruthenium complexes featuring 2, and shows LA
coordination can be realized via this route. Similar to the
structure of 7, LA coordination in 9 occurs with loss of a
“ruthenium black” had occurred, a significant darkening of the
solution would be expected. This color was maintained upon
storage of the final reaction mixture in air at −30 °C. Although
these ruthenium species could not be fully characterized, their
hydride ligand as H , with the second coordinating point
2
3
1
1
P{ H} NMR spectra indicated ligand coordination was
accessing the metal center through loss of the bound
3
preserved, and under some catalytic conditions, so were
ruthenium−hydride moieties.
acetonitrile, affording [Ru(CO){N(CH
2
PPh
] (11) (Scheme 4).
6
2
)
3
}-κ P{CH
3
C-
2
29
(O)(CH
)
C(O)O−κ O}][PF
2
2
Mechanistic Investigation. Under our catalytic condi-
tions, the data suggest four main conclusions: (i) in general,
systems involving N-triphos performed on a par with or better
than analogous systems with Triphos for the formation of 1,4-
PDO; (ii) the inclusion of NH PF or p-TsOH suppresses the
Complex 8 featuring the carbon-centered ligand 1 was found
to react in a manner identical to 9 in the presence of LA,
releasing the bound acetonitrile moiety and the remaining
hydride ligand to afford [Ru(CO){CH
C(CH
] (10) after LA
6
PPh ) }-
3
2 2 3
3
2
κ P{CH
C(O)(CH
)
C(O)O−κ O}][PF
4
6
3
2
2
overall catalytic activity, whereas (iii) inclusion of HN(Tf)2
appears to maintain similar catalytic activity and greatly
facilitates the conversion of 1,4-PDO to 2-MTHF; and (iv)
increasing the lability of ancillary ligands present on the
preformed catalyst precursors increases efficacy. A series of
stoichiometric reactions were performed between catalytic
species and levulinic acid under various conditions to help
explain and enforce some of these observations.
coordination. An NMR-scale reaction between 8 and 1.5
equiv of LA at room temperature showed the three signals
corresponding to 8 in its 31P{ H} NMR spectrum replaced by
1
1
those of 10 over approximately 10 h (Figure 2). The H NMR
spectra recorded over the first 10 h of the same reaction show
the disappearance of the doublet-of-triplets centered at −6.28
ppm, corresponding to the hydride ligand of 8 (Figure 3). The
formation of 10 as the product of the reaction between 8 and
LA was further confirmed by the presence of a mass peak at
869.1677 in the high-resolution mass spectrum.
The inclusion of NH PF was found to be highly detrimental
4
6
to catalysis when 3 or 4 was used as the precatalytic species.
Consequently, the reactivity of these complexes with LA was
studied both in the presence and absence of this proton donor.
Work already published by Leitner et al. has investigated the
reactivity of 3 and shown that it will react cleanly with LA at
Attempts to isolate and purify 8 resulted in decomposition to
3
a mixture of [Ru(CO)(NCMe)
{CH
3
C(CH
PPh
)
2
-κ P}] and
2
3
2
3
3
[Ru(NCMe) {CH C(CH PPh
)
-κ P}], among other impur-
3
3
2
2
ities. Consequently, reaction between 8 and LA was performed
using an impure sample, as can be observed in Figures 2 and 3,
both of which show additional resonances. The resonances
corresponding to 8 were identified through similarities to the
room temperature over several hours to form [RuH{CH C-
3
3
2
24
(
CH PPh ) }-κ P{CH C(O)(CH ) C(O)O−κ O}] (7).
During this reaction, the CO ligand is replaced by a neutral
2
2
3
3
2 2
29
donor carbonyl moiety of LA, and one hydride ligand of 3 will
previously well-characterized analogous 9. Nonetheless, the
impurities present are not thought to have interfered with the
conversion of 8 to 10, because their corresponding resonances
remained unaltered throughout the course of the reaction.
An important difference to note between complexes 7 and 11
is the presence of a bound hydride in 7 and none in 11.
Computational studies of the transformation of both LA to γVL
and γVL to 1,4-PDO on ruthenium start with the migratory
insertion of a bound hydride ligand into a coordinated carbonyl
react with the acid moiety of LA, releasing H and enabling
2
−
coordination of the resultant COO group (Scheme 3). This
affords a complex with LA bound in a bidentate coordination
mode, with one hydride ligand remaining.
Scheme 3. Reaction of Carbonyl Dihydride Complexes 3 and
4
with Levulinic Acid
10a,24
moiety of the target compound.
This suggests that for 11
to catalytically turn over, initiation steps that result in formation
of a Ru−H species must first occur, and this must proceed via
displacement of another ligand. This necessary “activation step”
may account for the decrease in catalytic activity when NH PF
4
6
is present in the reaction mixture using 4 as the precatalyst,
because 11 is likely to be formed in situ. Similar to mechanistic
reactions reported by Cole-Hamilton et al. using MSA as acidic
1
8
Conversely, an earlier study on 4 by our group shows
additive, the propensity of complexes featuring a [Ru-
+
markedly different reactivity. Complex 4, which is analogous to
except for the inclusion of a nitrogen atom at the apical
position of the triphosphine ligand, shows no direct reactivity
(triphos)] fragment to maintain a catalytic ruthenium−hydride
moiety is limited in the presence of acids.
Both NH PF and p-TsOH form potentially coordinating
3
4
6
29
+
−
with LA, even when heated for extended periods. This was
surprising to us, considering the structural similarity of the two
complexes. Efforts are currently underway within our group to
understand this unexpected stability; however, long-range
nitrogen−metal interactions (>3.4 Å) may be partially
responsible (see Supporting Information). Similar interactions
species upon release of H : namely, NH3 and p-TsO ,
respectively. In a closed system that does not allow release of
these in-situ-generated species, it is likely that they will compete
for coordination to the metal center with the substrate. Indeed,
this is observed for MSA, which is structurally similar to p-
1
8
TsOH. When 3 is reacted with NH PF in THF, a mixture of
4
6
Ph
have previously been observed in molybdenum−N-triphos
products is observed in which either THF (12) or NH (13)
coordinates to the empty site generated by loss of a hydride
3
3
3
complexes.
2
504
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
Scheme 4. Reaction of Carbonyl Dihydride Complexes 3 and 4 with NH PF in Acetonitrile and Complex 4 with NH PF in
4
6
4
6
THF, Plus Subsequent Reaction of the Generated Species (8, 9, 12, and 13) with Levulinic Acid
Figure 2. Stacked 31P{ H} NMR spectra of reaction between 8 and levulinic acid over 10 h at room temperature (162 MHz, acetone-d ).
1
6
ligand (Scheme 4). Dissolving this reaction mixture in
acetonitrile will convert 12 to 9, demonstrating the greater
coordinating ability of acetonitrile over THF, but 13 remains
unaffected, suggesting this species is fairly robust. A mass
spectrum of this acetonitrile solution shows peaks correspond-
ing to both 9 and 13.
ruthenium center reduced the catalytic activity of this species,
changing the major product obtained from 1,4-PDO to γVL.
There is a propensity for complexes 3 and 4 to maintain a
coordinated CO ligand, and these complexes were found to
underperform compared with 5, which features a more labile
PPh group. Unlike 4, which displayed no reactivity toward
3
29
Acids that generate relatively noncoordinating anions, such
LA, when 5 was reacted directly with LA, free PPh was
3
as HN(Tf) , will not compete for metal binding in the same
instantly released at room temperature (◆), and several new
species were formed as observed in the 31P{ H} NMR spectra
(Figure 4). The proportions of these new species changed over
the course of 2 weeks at room temperature and after further
heating for several hours at 85 °C. Unfortunately, upon
working up this reaction solution, only the starting material
(complex 5) was isolated pure, precluding the full character-
2
1
way as NH PF or p-TsOH, allowing substrate binding to occur
4
6
more readily. As was observed during the catalytic experiments
incorporating acidic additives (Table 2), good conversion of LA
to 2-MTHF was observed only when acids with non-
coordinating conjugate bases were utilized. Moreover, the
coordination of in-situ-generated conjugate bases to the
2
505
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
1
Figure 3. Stacked H NMR spectra showing the hydride region of reaction between 8 and levulinic acid over 10 h at room temperature (400 MHz,
acetone-d6).
Figure 4. Stacked 31P{ H} NMR spectra of reaction between 5 and levulinic acid over 13 days at room temperature (162 MHz, C D ). The
1
6
6
diamonds (◆) signify the instant release of free PPh3.
2
506
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
ization of new species formed during the reaction. Other
purification methods gave a mixture of products comprising
those observed during the reaction as well as trace amounts of
the carbonyl-containing 4, which presumably formed from the
24
decarbonylation of LA. Despite being unable to satisfactorily
characterize the new species, the increased reactivity of 5 over 4
toward LA is evidently due to increased lability of the
coordinated ligands due to the immediate loss of PPh3.
The reactivity of 5 with LA and H in the presence of
2
NH PF was investigated, and it also displayed much lower
4
6
catalytic activity. In contrast to the carbonyl-containing
complexes 3 and 4, when 5 was reacted with NH PF in
4
6
acetonitrile, all of the ligands except for the triphosphine were
substituted for solvent, affording [Ru(NCMe) {N(CH PPh ) -
3
2
2 3
3
−
κ P}][PF ] (14) with two PF units acting as noncoordinat-
6
2
6
ing counterions (Scheme 5). This species could be isolated in a
Scheme 5. Synthesis of [Ru(NCMe) {N(CH PPh ) -
3
2
2 3
3
κ P}][PF ] (14) via Two Different Routes
6
2
Figure 5. Crystal structure of one (14-A) of the two independent
complexes present in the crystals of 14 (50% probability ellipsoids).
analysis were obtained first by anion exchange of the chloride
counterion to tetraphenylborate (16), and grown by slow
diffusion of diethyl ether into a methanol solution of 16. The
solid state structure of 16 showed the cation to be the expected
diruthenium species with three chloride bridges (Figure 6).
Both ruthenium centers have distorted octahedral coordination
geometries (Table 4) with the tridentate phosphine ligands
occupying facial sites; the P−Ru−P bite angles are in the ranges
8
6.56(3)−90.73(3)° and 87.96(3)−89.65(3)° at Ru1 and Ru2
7
2% yield when 10 equiv of NH PF was used. Interestingly, 14
respectively). Surprisingly, all three chloride bridges are
noticeably asymmetric, with the bond to Ru1 being significantly
shorter than that to Ru2 in each case; the Ru−P distances do
not show any discernible pattern. The Ru1···N1 separation of
3.444(2) Å is likewise shorter than its Ru2···N51 counterpart
(3.512(2) Å), with associated torsional twists about Ru···N
vectors of ∼43 and 51°, respectively. The reason for these
differing geometries at Ru1 and Ru2 is not readily apparent.
Although it is fairly resistant to chemical transformation,
4
6
was also exclusively formed when 5 and NH PF were reacted
4
6
in a 1:1 molar ratio and could be isolated in a 49% yield,
indicating reaction between the second hydride and NH4+ to
produce H occurs very rapidly. Crystals suitable for X-ray
2
diffraction analysis were grown by vapor diffusion of diethyl
ether into a concentrated acetonitrile solution of 14 overnight.
The crystal structure of 14 was found to contain two
independent cations (14-A and 14-B, Figures 5 and S22,
respectively). The two cations have similar conformations, and
in each case, the ruthenium center has a distorted octahedral
coordination geometry (Table 3). The tridentate phosphine
ligand binds in a facial fashion with P−Ru−P bite angles in the
ranges 86.70(4)−91.90(4)° and 87.81(4)−89.73(4)° for
complexes 14-A and 14-B, respectively. Despite very similar
Ru1···N1 separations of 3.468(3) and 3.480(3) Å for 14-A and
treating 15 with 4 equiv of AgPF in acetonitrile under reflux
6
for several hours affords analytically pure 14 after recrystalliza-
tion from acetonitrile/diethyl ether, with concomitant precip-
itation of AgCl from the reaction mixture (Scheme 5).
Despite superficially appearing to be reactive due to the
potentially labile acetonitrile ligands, 14 is, in fact, robust and
stable to both moisture and air in the solid state. Reaction of 14
with 4 bar H at room temperature showed almost no reactivity
over a period of 6 days at room temperature, and even
extended heating at 50 °C for 7 days resulted in only several
new small resonances in the P{ H} NMR spectrum and no
1
4-B, respectively, the torsion angles about the Ru1···N1
2
vectors differ by ∼5°, being ∼58 and 53°, respectively. Both
complexes display bond lengths and angles that fall within the
34
31
1
standard range of similar complexes.
1
To produce tangible amounts of 14 for subsequent reactivity
studies, it was essential to establish an alternative synthetic
procedure for this complex that did not involve 5 as an
intermediate. Accordingly, the dimeric species [Ru (μ-Cl) {N-
new hydride signals by H NMR spectroscopy, suggesting very
slow decomposition rather than formation of a new hydride-
containing ruthenium species. The addition of LA to 14 proved
similarly unreactive, with almost no reactivity after heating at 50
°C for 7 days. The rapid formation of 14 from 5 suggests that a
solvent-bound complex similar to 14 may have formed under
the catalytic conditions when 5 is used as a catalytic precursor
in the presence of acid. In addition, the stability of 14 in the
2
3
3
(
[
CH PPh ) -κ P} ][Cl] (15) was synthesized from 2 and
RuCl (DMSO) ]. This air- and moisture-stable complex
2 2 3 2
2
4
precipitated as a yellow powder from toluene and was easily
isolated by filtration and subsequent washing to afford an
analytically pure sample. Crystals suitable for X-ray diffraction
presence of H and LA suggests that similar species may persist
2
2
507
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
Table 3. Selected Bond Lengths (Å) and Angles (deg) for the Two Independent Complexes (14-A and 14-B) Present in Crystals
of 14
A
B
A
B
Ru1−P3
Ru1−P5
Ru1−P7
Ru1···N1
2.3192(11)
2.3129(10)
2.3271(10)
3.468(3)
2.3149(11)
2.3210(11)
2.3182(11)
3.480(3)
Ru1−N44
Ru1−N47
Ru1−N50
2.107(3)
2.098(3)
2.093(3)
2.105(3)
2.106(3)
2.101(4)
P3−Ru1−P5
P3−Ru1−P7
P3−Ru1−N44
P3−Ru1−N47
P3−Ru1−N50
P5−Ru1−P7
P5−Ru1−N44
P5−Ru1−N47
91.90(4)
89.12(4)
P5−Ru1−N50
P7−Ru1−N44
P7−Ru1−N47
P7−Ru1−N50
N44−Ru1−N47
N44−Ru1−N50
N47−Ru1−N50
87.82(9)
91.84(9)
86.96(10)
88.35(9)
86.70(4)
87.81(4)
174.91(9)
90.27(9)
171.71(9)
90.14(9)
97.13(9)
99.53(10)
171.31(10)
83.25(12)
84.18(13)
84.02(13)
173.93(9)
85.07(13)
83.64(12)
86.54(13)
98.15(10)
88.37(3)
100.16(10)
89.73(4)
92.93(9)
98.21(9)
174.19(10)
170.68(10)
Table 4. Selected Bond Lengths (Å) and Angles (deg) for 16
Ru1−Cl1
Ru1−Cl2
Ru1−Cl3
Ru1−P3
Ru1−P5
Ru1−P7
Ru1···N1
Ru1···Ru2
Cl1−Ru1−Cl2
Cl1−Ru1−Cl3
Cl1−Ru1−P3
Cl1−Ru1−P5
Cl1−Ru1−P7
Cl2−Ru1−Cl3
Cl2−Ru1−P3
Cl2−Ru1−P5
Cl2−Ru1−P7
Cl3−Ru1−P3
Cl3−Ru1−P5
Cl3−Ru1−P7
P5−Ru1−P3
P3−Ru1−P7
P5−Ru1−P7
Ru1−Cl1−Ru2
Ru1−Cl2−Ru2
2.4559(6)
2.4808(7)
2.4803(7)
2.3017(7)
2.2802(7)
2.2712(7)
3.444(2)
3.4486(3)
77.58(2)
77.70(2)
177.99(3)
93.65(2)
95.21(2)
77.51(2)
101.30(2)
169.03(2)
96.57(2)
100.46(2)
94.31(2)
171.53(2)
87.26(3)
86.56(3)
90.73(3)
88.81(2)
87.13(2)
Ru2−Cl1
Ru2−Cl2
Ru2−Cl3
Ru2−P53
Ru2−P55
Ru2−P57
Ru2···N51
2.4724(6)
2.5231(7)
2.4954(7)
2.2701(7)
2.2779(7)
2.2852(7)
3.512(2)
Cl1−Ru2−Cl2
Cl1−Ru2−Cl3
P53−Ru2−Cl1
P55−Ru2−Cl1
P57−Ru2−Cl1
Cl3−Ru2−Cl2
P53−Ru2−Cl2
P55−Ru2−Cl2
P57−Ru2−Cl2
P53−Ru2−Cl3
P55−Ru2−Cl3
P57−Ru2−Cl3
P53−Ru2−P55
P53−Ru2−P57
P55−Ru2−P57
Ru1−Cl3−Ru2
76.50(2)
77.11(2)
166.61(3)
105.32(2)
89.15(2)
76.46(2)
90.95(2)
167.04(3)
104.95(2)
104.76(3)
91.32(2)
165.54(3)
87.96(3)
89.65(3)
87.97(3)
87.75(2)
3
Figure 6. Crystal structure of [Ru (μ-Cl) {N(CH PPh ) -κ P} ]-
2
3
2
2
3
2
[
BPh ] (16) (50% probability ellipsoids).
4
in the catalytic reaction mixture and may represent a significant
pathway for potentially complete catalyst deactivation, or at the
least temporary removal from the catalytic cycle. The
identification of complexes 8−14 and their observed high
stability and low reactivity suggest that these species are being
formed under catalytic conditions and are responsible for the
low catalyst activity when NH PF is included as a catalytic
used during this study. Increasing the overall lability of catalyst
by substitution of the carbonyl ligand in 4 with a PPh ligand
3
afforded a complex that, when used as a catalyst precursor,
allowed almost quantitative conversion of LA to 1,4-PDO
under relatively mild conditions.
4
6
additive.
The use of acidic additives was found to be detrimental to
catalysis in two cases (NH PF and p-TsOH), whether a
4
6
CONCLUDING REMARKS
■
preformed complex was used or an in-situ-generated system.
The only system to maintain similar catalytic activity under
The sequential catalytic hydrogenation of the biomass-derived
compound levulinic acid (LA) to γ-valerolactone (γVL); 1,4-
pentanediol (1,4-PDO); and, ultimately, 2-methyltetrahydro-
furan (2-MTHF) using a series of N-triphos−ruthenium
complexes as catalytic precursors has been investigated.
Ph
both non- and acidified conditions was N-triphos /[Ru-
(acac) ]. Changes to the reaction condition,s including varying
3
the amount of acidic additive, higher pressures, or different
solvents, did not lead to enhanced generation of 2-MTHF but,
in most cases, led to increased side-product formation. Using
the strong, noncoordinating acid HN(Tf)2 was found to
considerably enhance conversion of LA to 2-MTHF, in the best
case leading to yields of 87%. This is likely due to the
noncoordinating nature of the conjugate base, which will not
compete with the substrate for binding to the metal center.
Previously, the ruthenium dihydride complex [RuH (CO)-
2
3
{
CH C(CH PPh ) -κ P}] (3) had been identified as a dormant
3
2
2 3
form of the active catalyst for the hydrogenation of LA, and the
analogous complex featuring the nitrogen-centered triphos-
phine ligand 2, [RuH (CO){N(CH PPh ) -κ P}] (4), was
3
2
2
2 3
found to perform better under the relatively mild conditions
2
508
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
1
3
1
Mechanistic investigations were conducted to try to identify
possible species that would be present and potentially
persistent under the catalytic conditions. Carbon-centered
solvent for C{ H} NMR spectroscopy, and an external
3
31
1
H PO4 standard for P{ H} NMR spectroscopy. Pseudo
doublets, triplets, and doublets-of-triplets that occur as a result
of identical J value coupling to two or more chemically
nonequivalent nuclei are assigned as dd or ddd and are
recognized by the inclusion of only one or two J-coupling
values. 13C{ H} NMR spectra were assigned with the aid of
DEPT-135, HSQC, and HMBC correlation experiments. Mass
spectrometry analyses were conducted by the Mass Spectrom-
etry Service, Imperial College London. Elemental analyses were
carried out by Mr. Stephen Boyer of the School of Human
Sciences, London Metropolitan University. X-ray diffraction
analyses were carried out by Dr. Andrew White of the
Department of Chemistry at Imperial College London. Details
of single-crystal X-ray diffraction analysis can be found in the
Supporting Information.
Ph
Triphos was found to show activity different from that of
Ph
associated N-triphos systems, as demonstrated by the
29
reactivity between 3 or 4 and LA, and may account for the
differing product distributions observed when these complexes
are used catalytically. Reaction of 3 and 4 with 1 or more
equivalents of NH PF in MeCN solvent resulted in loss of a
1
4
6
single hydride ligand, lost as H , and coordination of
2
acetonitrile solvent. These intermediary complexes were
found to react with LA through dissociation of the remaining
hydride moiety and the bound solvent molecule, allowing LA to
coordinate in a bidentate fashion. Importantly it seems, upon
LA coordination under acidic conditions, no hydride ligands
remain coordinated. Because migratory insertion of a hydride
has been suggested as the first step in the catalytic
General Catalytic Procedure. Method A. A solution of
levulinic acid (1.16 g, 10.0 mmol); [Ru(acac) ] (19.8 mg, 0.05
24
hydrogenation of LA, the lack of hydrides under acidic
conditions is likely to hinder catalysis.
3
mmol, 0.5 mol %); and triphosphine ligand (1.0 mol %); and,
where applicable, acidic additive (5.0 mol %) in THF (20 mL)
was syringed into a prepurged high pressure reactor under a
flow of nitrogen. The atmosphere was changed to hydrogen by
repeatedly pressurizing to 5 bar and depressurizing (×3), before
the pressure was raised to 50 bar and the reactor was sealed.
The reactor was heated to 150 °C and stirred, causing the
internal pressure to rise to 65 bar once the reaction
temperature was reached. The reaction mixture was stirred at
this temperature and pressure for 25 h before the vessel was
cooled in an ice bath and slowly depressurized. The product
mixture was analyzed by GC and H and P{ H} NMR
spectroscopy.
Method B. Identical to method A, except a preformed
ruthenium catalyst precursor was utilized (0.5 mol %) instead
of a [Ru(acac) ]/triphosphine ligand mixture. Analysis of each
run was done by GC and H and P{ H} NMR spectroscopy.
Mercury Poisoning Experiment. A solution of levulinic acid
(1.16 g, 10.0 mmol), [Ru(acac) ] (19.8 mg, 0.05 mmol, 0.5 mol
%) and N-triphos (61.2 mg, 0.10 mmol, 1.0 mol %) in THF
(20 mL) was added via syringe into a prepurged high-pressure
reactor under a flow of nitrogen. The atmosphere was changed
to hydrogen by repeatedly pressurizing to 5 bar and
depressurizing (×3) before the pressure was raised to 50 bar
and the reactor was sealed. The reactor was heated to 150 °C
and stirred, causing the internal pressure to rise to 65 bar. The
reaction mixture was stirred at this temperature and pressure
for 2 h before the vessel was cooled in an ice bath and slowly
depressuzied. A sample was submitted for GC analysis (Table
S1, entry 1). The reaction mixture was transferred to a Schlenk
flask, and mercury (200 mg, 1.00 mmol) was added. After
stirring for 2 h at room temperature, the mercury was separated
from the solution. The solution was transferred back into the
The inclusion of more-labile ancillary ligands to the
complexes being used as precatalytic species resulted in better
catalytic activity, with 5 giving almost quantitative conversion of
LA to 1,4-PDO. This increased reactivity is evident from the
reaction between 5 and LA, which forms several unidentified
products; however, it importantly shows instantaneous release
of PPh . This indicates that the lability of this ligand is integral
3
for its catalytic activity. Upon treatment of 5 with NH PF , all
4
6
ancillary ligands are lost and substituted for acetonitrile solvent,
forming the highly stable dicationic complex 12. Heating this
1
31
1
complex in the presence of LA or 4 bar H gave almost no
2
reaction over 7 days and, consequently, may represent a
temporary or permanent deactivation pathway if present in the
catalytic reaction mixture, which is likely in the presence of
NH PF .
4
6
3
1
31
1
EXPERIMENTAL SECTION
Safety Warning. Experiments with compressed gases must
be carried out with appropriate equipment and under rigorous
safety precautions.
■
3
Ph
General Considerations. All preparations were carried out
using standard Schlenk line techniques under an inert
atmosphere of N2 unless otherwise stated. Solvents were
dried over standard drying agents and stored over 3 Å
molecular sieves. All starting materials were of reagent grade
and purchased from either Sigma-Aldrich Chemical Co. or
VWR International and used without further purification.
Ph
Triphos (1) was purchased from Sigma-Aldrich Chemical
Co., and used without further purification. Ligand 2, N-
Cyp
triphos , and complexes 4−6 and [RuCl (DMSO) ] were
2
4
2
4,29,35
prepared as previously reported.
Catalytic reactions were
performed in a high-pressure adapted Autoclave Engineers 100
mL Mini Reactor with a Universal Reactor Controller Series
control unit. Conversion and selectivity of the catalytic
reactions were determined via gas chromatography using a
Hewlitt Packard Gas Chromatograph 5890 Series II with a
J&W Scientific Agilent 122-5032 DB-5 column and equipped
with a flame ionization detector (FID) and dodecane as internal
standard. Peaks were assigned via GC/MS and pure substance
high-pressure reactor and repressurized to 50 bar H and
2
heated to 150 °C for 4 h, after which the reactor was cooled in
an ice bath and slowly depressurized. A sample was analyzed by
GC (Table S1, entry 2). A control experiment was conducted
using the same procedure without addition of mercury (Table
S1, entries 3 and 4).
Synthesis of [RuH (CO){CH C(CH PPh ) -κ P}] (3). A
solution of [Ru(CO )(CO){CH C(CH PPh ) -κ P}] (621 mg,
0.76 mmol) in THF (20 mL) was injected into a high-pressure
reactor under a flow of nitrogen. The atmosphere was changed
to hydrogen and pressurized to 15 bar at room temperature
before the mixture was heated to 100 °C (increasing the
3
2
3
2
2 3
1
13
1
31
1
3
calibration. H, C{ H}, and P{ H} NMR spectra were
recorded on Bruker AV-400, AV-500, or DRX-400 spectrom-
eters at 294 K unless otherwise stated. Chemical shifts are
reported in parts per million using the residual proton
3 3 2 2 3
1
impurities in the solvents for H NMR spectroscopy, the
2
509
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
3
1
1
1
internal pressure to ∼20 bar) and stirred for 2 h. After the
mixture was cooled to room temperature, the gas was carefully
vented, and the atmosphere was changed to nitrogen. The
solution was transferred from the reactor to a Schlenk flask,
filtered, and diluted with ethanol (15 mL), and the solution was
concentrated to ∼5 mL, causing precipitation of an orange
powder. Additional ethanol (15 mL) was added to fully
precipitate the product that was isolated by cannula filtration,
washed with ethanol (3 × 3 mL) and dried in vacuo. Complex
P{ H} NMR (THF-d , 162 MHz) δ: −144.0 (septet, 1P, J
8
PF
2
2
2
= 706 Hz), −13.0 (ddd, 1P, J = 206 Hz, J = 26 Hz, J
=
PP
PP
PP
3
1
1
24 Hz), 7.0−8.0 (m, 1P), 25.5−26.0 (m, 1P). 13: P{ H}
1
NMR (THF-d , 162 MHz) δ: −144.0 (septet, 1P, J = 706
8
PF
2
2
Hz), −8.5 (dd, 1P, J = 28 Hz, J = 23 Hz), 7.0−8.0 (m,
PP
PP
2
2
1P), 26.0 (dd, 1P, J = 34 Hz, J = 23 Hz). HRMS (ES): m/
PP
PP
102
3
+
z calcd. for C H N OP Ru ([M − PF ] ) 759.1397, found
4
0
40
2
6
756.7177. Attempts to purify and isolate these species resulted
in decomposition; however, direct reaction of a mixture of 12
and 13 with levulinic acid resulted in formation of [Ru(CO)-
3
was isolated as an orange powder (379 mg, 0.50 mmol, 66%)
24
3
2
with characterization identical to that previously reported.
{N(CH PPh ) }-κ P{CH C(O)(CH ) C(O)O-κ O}][PF ]
(11) with characterization identical to what has previously been
reported.
2 2 3 3 2 2 6
NMR-Scale Reaction of [RuH (CO){CH C(CH PPh ) -
2
3
2
2 3
3
29
κ P}] (3), NH PF and Levulinic Acid. To a Schlenk flask
4
6
3
3
was added [RuH (CO){CH C(CH PPh ) -κ P}] (31.7 mg,
NMR-Scale Reaction of [RuH (PPh ){N(CH PPh ) -κ P}]
2
3
2
2 3
2
3
2
2 3
0.042 mmol) and suspended in acetonitrile (2 mL). A solution
(5) and Levulinic Acid. To a Young’s tap NMR tube was
added [RuH (PPh ){N(CH PPh ) -κ P}] (16.1 mg, 0.017
3
of NH PF (6.8 mg, 0.042 mmol) in acetonitrile (1 mL) was
4
6
2
3
2
2 3
added via cannula, and the mixture was stirred at room
temperature for 2 h, turning homogeneous. The solvent was
removed in vacuo, affording a yellow powder that was dried in
vacuo for 15 min, washed with hexane (3 × 2 mL), and again
dried in vacuo for 15 min. The powder was then dissolved in
mmol) dissolved in C D (0.7 mL); initial NMR analysis
were performed. In a separate Schlenk flask was prepared a
6 6
solution of levulinic acid (56.3 mg, 0.48 mmol) in C D (2
6
6
mL)., and 0.1 mL (0.024 mmol LA, 1.5 equiv) was added to the
NMR tube via syringe. The NMR tube was sealed and analyzed
by NMR spectroscopy regularly over the course of 2 weeks at
room temperature. The NMR tube was then heated to 85 °C
for a total of 35 h, with analysis by NMR spectroscopy at room
temperature periodically. Removal of the solvent in vacuo and
subsequent redissolution of the resultant yellow powder in
methanol (2 mL) gave crystals of pure [RuH (PPh ){N-
acetone-d (0.3 mL) and transferred in its entirety to a Young’s
6
tap NMR tube via cannula, which was washed through with
additional acetone-d (0.3 mL). Initial NMR analysis was
6
3
performed on [RuH(CO)(NCMe){CH C(CH PPh ) -κ P}]-
3
2
2 3
1
[
1
(
8
PF ] (8): H NMR (acetone-d , 400 MHz) δ: −6.28 (ddd,
6
6
2
2
31
1
H, J = 81 Hz, J = 16 Hz, Ru−H). P{ H} NMR
HP
HP
2
3
1
3
acetone-d , 162 MHz) δ: −144.0 (septet, 1P, J = 706 Hz),
(CH PPh ) -κ P}] when left to stand overnight. The super-
6
PF
2 2 3
2
2
.0−8.5 (m, 1P), 24.5 (dd, 1P, J = 43 Hz, J = 26 Hz), 40.0
natant was saved, and the solvent was removed in vacuo. NMR
PP
PP
2
2
(
dd, 1P, J = 43 Hz, J = 26 Hz). In a separate Schlenk flask
analysis of the resultant residue in C D6 displayed an
PP
PP
6
was prepared a solution of levulinic acid (7.5 mg, 0.065 mmol,
.5 equiv) in acetone-d (0.3 mL), and this was added to the
uncharacterizable mixture of products, including [RuH (CO)-
2
3
1
{N(CH PPh ) -κ P}] (4).
6
2
2 3
3
NMR tube via syringe. The NMR tube was sealed, stirred for
Synthesis of [Ru(NCMe) {N(CH PPh ) -κ P}][PF ] (12).
3 2 2 3 6 2
1
31
1
3
0 s using a vortex stirrer, and analyzed by H and P{ H}
Method A. To a Schlenk flask was added [RuH (PPh ){N-
2 3
3
NMR spectroscopy every hour for 10 h. Quantitative
(CH PPh ) -κ P}] (82.1 mg, 0.084 mmol) and NH PF (24.6
2 2 3 4 6
3
conversion of 8 to [Ru(CO){CH C(CH PPh ) }-κ P{CH C-
mg, 0.15 mmol), and the flask was evacuated and backfilled
with nitrogen (×3). Acetonitrile (2 mL) was added, and the
mixture was stirred at room temperature for 20 h. The solvent
was removed in vacuo, and the resultant powder was washed
with diethyl ether (3 × 2 mL) and dried in vacuo. Redissolving
in acetonitrile and slow diffusion of diethyl ether into the
solution by vapor diffusion afforded crystals overnight, which
were isolated by cannula filter, washed with diethyl ether (3 × 2
mL), and dried in vacuo. The combined supernatant and
washings afforded a second batch of crystals that were isolated
and washed as the first batch were. Both batches gave
analytically pure white crystals of 12 that were suitable for X-
3
2
2 3
3
2
1
(
(
2
7
3
O)(CH ) C(O)O-κ O}][PF ] (10) was observed. H NMR
2 2 6
Triphos
3
acetone-d , 400 MHz) δ: 1.96−1.99 (m, 3H, CH
),
=
6
3
Triphos
3
HH
LA
.03−2.08 (m, 6H, CH
), 2.24 (s, CH ), 2.65 (t, J
2
3 HH
LA
LA
Hz, CH2 ), 2.80 (t, J = 6 Hz, CH2 ), 6.99−7.76 (m,
0H, Ph). C{ H} NMR (acetone-d , 127 MHz) δ: 28.1 (s,
), 37.5 (s, CH2 ), 38.2 (s, CH2 ),
), 129.2 (d, J = 11 Hz, CH ), 129.8 (dt, J
4 Hz, JCP = 5.5 Hz, CH ), 131.2 (s, CH ), 131.4 (s, CH ),
31.6 (t, JCP = 5 Hz, CH ), 132.03 (s, CH ), 174.0 (s, CO ),
). P{ H} NMR (acetone-
d , 162 MHz) δ: −144.0 (septet, 1P, J = 706 Hz), 5.0 (dd,
P, J = 35 Hz), 35.0 (dd, 1P, J = 35 Hz). HRMS (ES): m/
1
3
1
6
LA
Triphos
LA
LA
CH3 ), 31.5 (m, CH
2
Triphos
Ph
3
1
1
1
9.4 (s, C
=
CP
CP
Ph
Ph
Ph
Ph
Ph
LA
LA
Triphos 31
1
91.62 (s, CO ), 206.9 (s, CO
1
6
PF
2
2
1
1
ray diffraction experiments (61.6 mg, 0.055 mmol, 72%). H
PP
PP
1
3
02
+
z calcd. for C H O P Ru ([M − PF ] ) 869.1652, found
NMR (CD
Cl , 400 MHz) δ: 2.32 (S, 9H, CH ), 4.23 (s, 6H,
2 2 3
47
46
4
6
869.1677.
CH ), 7.15−7.28 (m, 24H, CH), 7.30−7.41 (m, 6H, CH).
2
3
13
1
NMR-Scale Reaction of [RuH (CO){N(CH PPh ) -κ P}]
C{ H} NMR (CD Cl , 101 MHz) δ: 4.18 (s, CH ), 50.8−
2
2
2 3
2
2
3
(
4) and NH PF in THF. To a Young’s tap NMR tube was
51.1 (m, CH ), 129.4−129.5 (m, CH), 131.4 (s, CH), 132.3−
4
6
2
3
31 1
added [RuH (CO){N(CH PPh ) -κ P}] (33.6 mg, 0.0452
132.4 (m, CH). P{ H} NMR (CD Cl , 162 MHz) δ: 9.0 (s,
2
2
2 3
2 2
1
19
mmol) and NH PF (7.7 mg, 0.0472 mmol) dissolved in
3P, PPh ), −144.0 (heptet, 2P, J = 712 Hz, PF ). F NMR
4
6
2
PF
6
1
THF-d (0.7 mL). The NMR tube was shaken for 1 min using a
(CD Cl , 381 MHz) δ: −72.4 (d, 12F, J = 712 Hz, PF ).
FP
102 2+
8
2
2
6
1
31
1
vortex stirrer and analyzed by H and P{ H} NMR
spectroscopy over the course of 3 days at room temperature.
HRMS (ES): m/z calcd. for C H N P Ru ([M − 2PF ] )
4
5
45
4
3
6
418.0950, found 418.0950. Anal. Calcd for C H N P F Ru
45
45
4 5 12
The formation of two species was observed: [RuH(CO)-
(found): C, 48.01 (47.89); H, 4.03 (3.94); N, 4.98 (4.88).
3
(
(
THF){N(CH PPh ) -κ P}] (12) and [RuH(CO)(NH ){N-
Method B. To a Schlenk flask was added [Ru (μ-Cl) {N-
2
2 3
3
2
3
3
3
CH PPh ) -κ P}] (13). After 3 days, the solvent and volatiles
(CH PPh ) -κ P} ] (146 mg, 0.11 mmol), AgPF (114 mg,
2
2 3
2 2 3 2 6
were removed in vacuo, and the resultant powder was
0.45 mmol, 4.2 equiv), and acetonitrile (5 mL). The suspension
was stirred and heated to 70 °C for 15 h and remained a yellow
suspension throughout. After cooling to room temperature and
1
31
1
redissolved in CD CN and analyzed by H and P{ H}
3
NMR spectroscopy and high-resolution mass spectrometry. 12:
2
510
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
3
allowing to settle, the supernatant was isolated from the powder
was added [Ru(NCMe) {N(CH PPh ) -κ P}][PF ] (24.6 mg,
3 2 2 3 6 2
via cannula filter, and the powder was washed with acetonitrile
0.022 mmol) and partially dissolved in CD Cl (0.6 mL) using
2 2
1
31
1
19
(
2 mL) which was added to the saved supernatant. This
a vortex stirrer. Initial H, P{ H}, and F NMR analyses were
carried out. A solution of levulinic acid (5.3 mg) in CD Cl (0.3
solution was concentrated in vacuo to ∼1.5 mL, and the
product crystallized by the slow addition of diethyl ether via
vapor diffusion. Analytically pure colorless crystals with
characterization identical to those obtained via method A
2
2
mL) was prepared in a separate Schlenk flask, and 0.22 mL (3.9
mg levulinic acid, 0.034 mmol, 1.5 equiv) was added to the
solution of 12 via syringe at room temperature. The solution
1
31
1
19
(
100 mg, 0.089 mmol, 83%) grew overnight.
was analyzed by H, P{ H}, and F NMR spectroscopy
periodically for 9 days at room temperature and 7 days at 50
°C. Almost no reactivity was observed.
3
Synthesis of [Ru (μ-Cl) {N(CH PPh ) -κ P} ][Cl] (13). To
2
3
2
2 3
2
a Schlenk flask was added [RuCl (DMSO) ] (480 mg, 1.00
2
4
mmol), 2 (611 mg, 1.00 mmol), and toluene (10 mL), and the
mixture was heated to 100 °C for 12 h. Upon heating, the
reaction mixture changed to a deep orange color and formed a
yellow precipitate. The yellow precipitate was isolated by
filtration and washed first with toluene (3 × 15 mL) and then
diethyl ether (3 × 15 mL). The light yellow solid was dried in
ASSOCIATED CONTENT
Supporting Information
The following file is available free of charge on the ACS
Publications website at DOI: 10.1021/cs502025t.
■
*
S
1
Gas chromatrograms of representative catalytic runs;
tables detailing additional catalytic runs including
vacuo (760 mg, 0.49 mmol, 97%). H NMR (CD Cl , 400
2
2
3
MHz) δ: 4.09 (br s, 12H, CH ), 6.88 (t, 24H, J = 8 Hz,
2
HH
31
1
Ph‑ortho
3
Ph‑para
mercury poisoning experiments;1 P{ H} NMR spectra
CH
), 7.19 (t, 12H, J = 8 Hz, CH
), 7.35 (m, 24H,
HH
31 1
CHPh‑meta). C{ H} NMR (CD Cl , 101 MHz) δ: 53.7 (s,
13
1
of complexes 8, 10, and 13; H and P{ H} NMR
spectra of complexes 14−16. Additional crystallographic
data, including 50% probability ellipsoid ORTEP diagram
of 14-B. Preliminary DFT studies (PDF)
2
2
Ph‑ortho
Ph‑para
CH ), 128.0 (s, CH
), 129.9 (s, CH
), 133.8
2
Ph‑meta
31
1
(
CH
). P{ H} NMR (CD Cl , 162 MHz) δ: 18.0 (s).
HRMS (ES): m/z calcd. for C H N P Cl Ru ([M − Cl] )
2
2
1
02
+
7
8
72
2
6
3
2
1
531.1274, found 1531.1310. Anal. Calcd for
C H N P Cl Ru .CH Cl (found): C, 57.43 (57.76); H,
AUTHOR INFORMATION
Corresponding Authors
E-mail: n.long@imperial.ac.uk.
E-mail: philip.miller@imperial.ac.uk.
78
72
2
6
4
2
2
2
■
4
.51 (4.67); N, 1.70 (1.62).
3
Synthesis of [Ru (μ-Cl) {N(CH PPh ) -κ P} ][BPh ] (14).
2
3
2
2 3
2
4
*
*
To a solution of 13 (200 mg, 0.13 mmol) in dichloromethane
5 mL) was slowly added a solution of NaBPh (44.5 mg, 0.13
(
4
Notes
mmol) in acetonitrile (0.5 mL), and the mixture was stirred at
room temperature for 2 h. A light colored precipitate of NaCl
formed over this time. The yellow filtrate was collected by
filtering through a small pad of Celite. Crystals suitable for X-
ray diffraction experiments were obtained overnight by slow
diffusion of diethyl ether into this solution and were collected
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Imperial College London for funding, via the
Frankland Chair endowment. Johson Matthey plc are also
thanked for the loan of the precious-metal salts used in this
work. We are also grateful to Imperial College/EPSRC for a
small equipment grant that enabled the purchase of high-
pressure equipment.
as yellow crystals after isolation and washing with diethyl ether
1
(
4
2
CH
3
(
1
1
146 mg, 0.079 mmol, 62%). H NMR (CD Cl , 400 MHz) δ:
2
2
3
BPh‑para
.07 (br s, 12H, CH ), 6.85 (t, J = 1 Hz, CH
), 6.87 (t,
= 7 Hz,
HH
2
HH
3
Ph‑ortho
3
8H, J = 7 Hz, CH
), 7.01 (t, 8H, J
HH
BPh‑ortho
3
Ph‑para
), 7.18 (t, 12H, J = 7 Hz, CH
), 7.34 (m,
HH
Ph‑meta
BPh‑meta
). 13C{ H} NMR
1
2H, overlapping CH
and CH
REFERENCES
■
4
BPh‑para
CD Cl , 101 MHz) δ: 53.7 (s, CH ), 122.2 (s, CH
),
),
2
2
2
(1) (a) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106,
044−4098. (b) Chheda, J. N.; Huber, G. W.; Dumesic, J. A. Angew.
BPh‑ortho
Ph‑meta
Ph‑ortho
Ph‑para
26.1 (s, CH
), 128.0 (s, CH
), 136.5 (s, CH
), 130.0 (s, CH
BPh‑meta
31
1
33.8 (s, CH
). P{ H} NMR
Chem., Int. Ed. 2007, 46, 7164−7183. (c) Corma, A.; Iborra, S.; Velty,
A. Chem. Rev. 2007, 107, 2411−2502. (d) Marshall, A.-L.; Alaimo, P. J.
Chem.Eur. J. 2010, 16, 4970−4980. (e) Takagaki, A.; Nishimura, S.;
Ebitani, K. Catal. Surv. Asia 2012, 16, 164−182.
(
CD Cl , 162 MHz) δ: 18.0 (s). HRMS (ES): m/z calcd. for
2 2
101
104
+
C H N P Cl Ru Ru ([M − BPh ] ) 1532.1297, found
1
6
78
72
2
6
3
4
532.1685. Anal. Calcd for C102H BCl N P Ru (found): C,
92 3 2 6 2
(
2) (a) Zakzeski, J.; Bruijnincx, P. C. A.; Jongerius, A. L.;
2.83 (62.95); H, 4.85 (4.74); N, 1.42 (1.59).
NMR-Scale Reaction of [Ru(NCMe) {N(CH PPh ) -κ P}]-
3
Weckhuysen, B. M. Chem. Rev. 2010, 110, 3552−3599. (b) Pandey,
3
2
2 3
M. P.; Kim, C. S. Chem. Eng. Technol. 2011, 34, 29−41.
[
[
PF ] (12) and H . To a high-pressure NMR tube was added
Ru(NCMe) {N(CH PPh ) -κ P}][PF ] (22.5 mg, 0.020
3 2 2 3 6 2
6 2 2
(
3) Hartwig, J. F.; Sergeev, A. G. Science 2011, 332, 439−443.
3
(4) vom Stein, T.; Weigand, T.; Merkens, C.; Klankermayer, J.;
Leitner, W. ChemCatChem 2013, 5, 439−441.
(5) Balaraman, E.; Fogler, E.; Milstein, D. Chem. Commun. 2012, 48,
1111−1113.
mmol) and partially dissolved in CD Cl (0.5 mL) using a
2
2
1
31
1
19
vortex stirrer. Initial H, P{ H}, and F NMR analyses were
carried out. The atmosphere in the NMR tube was changed to
hydrogen by repeatedly evacuating and backfilling (×3) while
the solution was frozen in liquid nitrogen, before the NMR
tube was pressurized to 4 bar and allowed to warm to room
(6) Du, X.-L.; Bi, Q.-Y.; Liu, Y.-M.; Cao, Y.; He, H.-Y.; Fan, K.-N.
Green Chem. 2012, 14, 935−939.
7) van Putten, R.-J.; van der Waal, J. C.; de Jong, E.; Rasrendra, C.
(
1
31
1
B.; Heeres, H. J.; de Vries, J. G. Chem. Rev. 2013, 113, 1499−1597.
(8) Zhou, C.-H.; Xia, X.; Lin, C.-X.; Tong, D.-S.; Beltramini, J. Chem.
Soc. Rev. 2011, 40, 5588−5617.
temperature. The solution was analyzed by H, P{ H}, and
1
9
F NMR spectroscopy periodically for 6 days at room
temperature and 7 days at 50 °C. Almost no reactivity was
(
9) (a) Wettstein, S. G.; Alonso, D. M.; Chong, Y.; Dumesic, J. A.
observed.
Energy Environ. Sci. 2012, 5, 8199−8203. (b) Qi, L.; Mui, Y. F.; Lo, S.
W.; Lui, M. Y.; Akien, G. R.; Horvat
1477.
3
NMR-Scale Reaction of [Ru(NCMe) {N(CH PPh ) -κ P}]-
́
h, I. ACS Catal. 2014, 4, 1470−
3
2
2 3
[
PF ] (12) and Levulinic Acid. To a Young’s tap NMR tube
6 2
2
511
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512

ACS Catalysis
Research Article
(
10) (a) Assary, R. S.; Curtiss, L. A. Chem. Phys. Lett. 2012, 541, 21−
6. (b) Tukacs, J. M.; Kiraly, D.; Stradi, A.; Novodarszki, G.; Eke, Z.;
Dibo, G.; Kegl, T.; Mika, L. T. Green Chem. 2012, 14, 2057−2065.
c) Michel, C.; Zaffran, J.; Ruppert, A. M.; Matras-Michalska, J.;
Jędrzejczyk, M.; Grams, J.; Sautet, P. Chem. Commun. 2014, 50,
2450−12453. (d) Ding, D.; Wang, J.; Xi, J.; Liu, X.; Lu, G.; Wang, Y.
Green Chem. 2014, 16, 3846−3853.
11) Serrano-Ruiz, J. C.; West, R. M.; Dumesic, J. A. Annu. Rev.
Chem. Biomol. Eng. 2010, 1, 79−100.
12) Geilen, F. M. A.; Engendahl, B.; Harwardt, A.; Marquardt, W.;
Klankermayer, J.; Leitner, W. Angew. Chem., Int. Ed. 2010, 49, 5510−
514.
13) Pagliaro, M.; Ciriminna, R.; Kimura, H.; Rossi, M.; Pina, C. D.
2
́
́
́
́
(
1
(
(
5
(
Angew. Chem., Int. Ed. 2007, 46, 4434−4440.
(
(
14) Aycock, D. F. Org. Process Res. Dev. 2007, 11, 156−159.
́
15) Pace, B.; Hoyos, P.; Fernandez, M.; Sinisterra, J. V.; Alcantara,
́
A. R. Green Chem. 2010, 12, 1380−1382.
16) Pace, V.; Hoyos, P.; Castoldi, L.; de Maria, P. D.; Alcan
R. ChemSusChem 2012, 5, 1369−1379.
17) (a) Hanton, M. J.; Tin, S.; Boardman, B. J.; Miller, P. J. Mol.
Catal. A: Chem. 2011, 346, 70−78. (b) vom Stein, T.; Meuresch, M.;
Limper, D.; Schmitz, M.; Holscher; Coetzee, J.; Cole-Hamilton, D. J.;
Klankermayer, J.; Leitner, W. J. Am. Chem. Soc. 2014, 136, 13217−
3225.
18) Coetzee, J.; Dodds, D. L.; Klankermayer, J.; Brosinski, S.;
Leitner, W.; Slawin, A. M. Z.; Cole-Hamilton, D. J. Chem.Eur. J.
013, 19, 11039−11050.
19) Mellone, I.; Peruzzini, M.; Rosi, L.; Mellmann, D.; Junge, H.;
Beller, M.; Gonsalvi, L. Dalton Trans. 2013, 42, 2495−2501.
20) Savourey, S.; Lefevre, G.; Berthet, J.-C.; Cantat, T. Chem.
Commun. 2014, 50, 14033−14036.
21) (a) Beydoun, K.; vom Stein, T.; Klankermayer, J.; Leitner, W.
(
́
tara, A.
(
̈
1
(
2
(
(
̀
(
Angew. Chem., Int. Ed. 2013, 52, 9554−9557. (b) Beydoun, K.;
Ghattas, G.; Thenert, K.; Klankermayer, J.; Leitner, W. Angew. Chem.,
Int. Ed. 2014, 53, 11010−11014.
(
22) (a) Wesselbaum, S.; vom Stein, T.; Klankermayer, J.; Leitner, W.
Angew. Chem., Int. Ed. 2012, 51, 7499−7502. (b) Wesselbaum, S.;
Moha, V.; Meuresch, M.; Brosinski, S.; Thenert, K. M.; Kothe, J.; vom
Stein, T.; Englert, U.; Ho
Chem. Sci. 2015, 6, 693−704.
23) Savourey, S.; Lefevre; Berthet, J.-C.; Thuer
Cantat, T. Angew. Chem., Int. Ed. 2014, 53, 10466−10470.
24) Geilen, F. M. A.; Engendahl, B.; Holscher, M.; Klankermayer, J.;
̈
lscher, M.; Klankermayer, J.; Leitner, W.
(
̀
́
y, P.; Genre, C.;
(
̈
Leitner, W. J. Am. Chem. Soc. 2011, 133, 14349−14358.
(
(
25) Mar
̈
kl, G.; Jin, G. Y. Tetrahedron Lett. 1981, 22, 1105−1108.
26) Miller, P. W.; White, A. J. P. J. Organomet. Chem. 2010, 695,
1
(
138−1145.
27) Fillol, J. L.; Kruckenberg, A.; Scherl, P.; Wadepohl, H.; Gade, L.
H. Chem.Eur. J. 2011, 17, 14047−14062.
28) Scherl, P.; Kruckenbery, A.; Mader, S.; Wadepohl, H.; Gade, L.
H. Organometallics 2012, 31, 7024−7027.
29) Phanopoulos, A.; Brown, N. J.; White, A. J. P.; Long, N. J.;
Miller, P. W. Inorg. Chem. 2014, 53, 3742−3752.
30) Rhodes, L. F.; Venanzi, L. M. Inorg. Chem. 1987, 26, 2692−
695.
31) Bakhmutov, V. L.; Bakmutova, E. V.; Belkova, N. V.; Bianchini,
(
(
(
2
(
C.; Epstein, L. M.; Masi, D.; Peruzzini, M.; Shubina, E. S.; Voronstov,
E. V.; Zanobini, F. Can. J. Chem. 2001, 79, 479−489.
(
32) (a) Widegren, J. A.; Finke, F. G. J. Mol. Catal. A: Chem. 2003,
1
98, 317−341. (b) Whitesides, G. M.; Hackett, M.; Brainard, R. L.;
Lavalleye, J.-P. P. M.; Sowinski, A. F.; Izumi, A. N.; Moore, S. S.;
Brown, D. W.; Staudt, E. M. Organometallics 1985, 4, 1819−1830.
(
33) Walter, O.; Huttner, G.; Kern, R. Z. Naturforsch. 1996, 51b,
9
(
22−928.
34) (a) Landgrafe, C.; Sheldrick, W. S.; Sudfeld, M. Eur. J. Inorg.
̈
Chem. 1998, 407−414. (b) Bianchini, C.; Dal Santo, V.; Meli, A.;
Oberhauser, W.; Psaro, R.; Vizza, F. Organometallics 2000, 19, 2433−
2444.
(
35) Bratsos, J.; Alessio, E. Inorg. Synth. 2010, 35, 148−152.
2
512
DOI: 10.1021/cs502025t
ACS Catal. 2015, 5, 2500−2512