Acid-, water- and high-temperature-stable ruthenium complexes for the total catalytic deoxygenation of glycerol to propane

Article abstract of DOI:10.1002/chem.200901427

The ruthenium aqua complexes [Ru(H₂O)₂(bipy) ₂](OTf)₂, [cisRu(6,6′-Cl₂-bipy) ₂(OH₂O)₂)] (OTf)₂, [Ru(H ₂O)₂(phen)₂](OTf)₂, [Ru(H ₂O)₃(2,2′:6′,2″-terpy)](OTf) ₂ and [Ru(H₂O)₃(Phterpy)] (OTf)₂ (bipy = 2,2′-bipyridine; OTf^- = triflate; phen = phenanthroline; terpy = terpyridine; Phterpy = 4′-phenyl-2,2′: 6′,2″-terpyridine) are water- and acid-stable catalysts for the hydrogenation of aldehydes and ketones in sulfolane solution. In the presence of HOS(O)₂CF₃ (triflic acid) as a dehydration co-catalyst they directly convert 1,2-hexanediol to n-hexanol and hexane. The terpyridine complexes are stable and active as catalysts at temperatures ≥ 250°C and in either aqueous sulfolane solution or pure water convert glycerol into n-propanol and ultimately propane as the final reaction product in up to quantitative yield. For the terpy complexes the active catalyst is postulated to be a carbonyl species [(4′-R-2,2′:6′,2″-terpy)Ru(CO) (H₂O)₂]-(OTf)₂ (R = H, Ph) formed by the decarbonylation of aldehydes (hexanal for 1,2-hexanediol and 3-hydroxypropanal for glycerol) generated in the reaction mixture through acid-catalyzed dehydration. The structure of the dimeric complex [{(4′-phenyl-2,2′: 6′,2″-terpy)Ru(CO)}₂(μ-OCH₃) ₂](OTf)₂ has been determined by single crystal X-ray crystallography (Space group P1 (a = 8.2532(17); b = 12.858(3); C = 14.363(3) A; α = 64.38(3); β = 77.26(3); γ = 87.12(3)°, R = 4.36%).

Full text of DOI:10.1002/chem.200901427

DOI: 10.1002/chem.200901427
Acid-, Water- and High-Temperature-Stable Ruthenium Complexes for the
Total Catalytic Deoxygenation of Glycerol to Propane
Deeb Taher,[a] Michelle E. Thibault,[b] Domenico Di Mondo,[b] Michael Jennings,[c] and
Marcel Schlaf*[b]
Abstract: The ruthenium aqua com-
plexes [Ru(H2O)2A(bipy)2](OTf)2, [cis-
Ru(6,6’-Cl2-bipy)2A(OH2)2](OTf)2, [Ru-
(H2O)2A(phen)2](OTf)2,
(H2O)3(2,2’:6’,2’’-terpy)]
[Ru(H2O)3A(Phterpy)](OTf)2
diol to n-hexanol and hexane. The ter-
pyridine complexes are stable and
active as catalysts at temperatures
ꢁ2508C and in either aqueous sulfo-
lane solution or pure water convert
glycerol into n-propanol and ultimately
propane as the final reaction product
in up to quantitative yield. For the
terpy complexes the active catalyst is
postulated to be a carbonyl species
[(4’-R-2,2’:6’,2’’-terpy)Ru(CO)
ACHTUNGTRENNUNG
ACHTUGNRTNE(NUGN H2O)2]-
G
N
N
E
U
carbonylation of aldehydes (hexanal
for 1,2-hexanediol and 3-hydroxypro-
panal for glycerol) generated in the re-
action mixture through acid-catalyzed
dehydration. The structure of the di-
meric complex [{(4’-phenyl-2,2’:6’,2’’-
A
E
G
[Ru-
and
(bipy=
E
G
G
N
2,2’-bipyridine; OTfꢀ =triflate; phen=
phenanthroline; terpy= terpyridine;
Phterpy=4’-phenyl-2,2’:6’,2’’-terpyri-
dine) are water- and acid-stable cata-
lysts for the hydrogenation of alde-
hydes and ketones in sulfolane solu-
tion. In the presence of HOS(O)2CF3
(triflic acid) as a dehydration co-cata-
lyst they directly convert 1,2-hexane-
terpy)Ru(CO)}
2ACHTUNGTERNNUG(m-OCH3)2]ACHTUNGTNER(NUGN OTf)2 has
been determined by single crystal X-
ray crystallography (Space group P1
(a=8.2532(17); b=12.858(3); c=
14.363(3) ꢁ; a=64.38(3); b=77.26(3);
g = 87.12(3)8, R=4.36%).
¯
Keywords: biomass
· deoxygena-
tion · glycerol · homogeneous catal-
ysis · hydrogenation
Introduction
uses for glycerol, which have recently been expertly re-
viewed by several authors.[1–5] In this context, significant ad-
vances have been made into the deoxygenation of glycerol
to 1,2-propanediol by using heterogeneous catalysts.[6–17] 1,2-
Propanediol is an important anti-freeze agent and compo-
nent of coating, lubricant and cosmetic formulations and
Dow Chemical recently announced its intent to commercial-
ize this process using a proprietary heterogeneous cata-
lyst.[18] We recently discussed the possible origin of the typi-
cally observed selectivity for terminal glycerol deoxygena-
tion to 1,2-propanediol with heterogeneous catalysts.[19]
A challenging, but potentially economically very attrac-
tive, alternative would be the metal-catalyzed deoxygenation
of glycerol to 1,3-propanediol, a component of poly(pro-
The exponential growth of the global bio-Diesel industry
has lead to a large oversupply of glycerol with production
estimated to reach 1.2 million tons p.a. by 2010. This in turn
has triggered substantial efforts to find new and value added
[a] Prof. Dr. D. Taher
Chemistry Department
Tafila Technical University
P. O. Box: 179, Tafila 66110 (Jordan)
[b] M. E. Thibault, D. Di Mondo, Prof. M. Schlaf
The Guelph-Waterloo Centre for Graduate Work
in Chemistry (GWC)2, Department of Chemistry
University of Guelph
AHCTUNGTREGpNNUN ylene terephthalate) (PPT), marketed as Soronaꢀ and
Guelph, Ontario, N1G 2W1 (Canada)
Corterraꢀ by DuPont and Shell, respectively. Such a pro-
cess would compete with existing petrochemical and bio-
technological routes to this material: The Degussa-DuPont
process operates via hydration of acrolein to 3-hydroxypro-
pion aldehyde (3-HPA), which cannot be isolated, followed
by hydrogenation over a heterogeneous ruthenium catalyst,
whereas Shell produces 1,3-propanediol by hydroformyla-
tion of ethylene oxide and in situ hydrogenation of the same
[c] M. Jennings
Chemistry Department
University of Western Ontario
London, Ontario, Canada, N6A 5B7 (UK)
Supporting information for this article is available on the WWW
1
under http://dx.doi.org/10.1002/chem.200901427 and contains H and
13C NMR spectra for all complexes synthesized and an image of the
GC-MS trace and spectrum to establish the conversion of glycerol to
propane.
10132
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10132 – 10143

FULL PAPER
intermediate by using
a
homogeneous Co/Ru catalyst
mizing the secondary dehydrations of either the 3-HPA in-
termediate or the desired product 1,3-propanediol. As dem-
onstrated by Che this appears to be in principle possible
through the right combination of a) the type of solvent and
acid employed and b) the relative concentrations of acid,
water and substrate (which in turn releases water) in the re-
action mixture. However, to our knowledge the only quanti-
tative experimental data on the relative positions of the
three equilibria involved is given by Pressmann and
Lucas,[30] who determined the equilibrium constant for the
acid-catalyzed hydration of acrolein to 3-HPA in water, that
is, K’eq, over the temperature range 19.98 to 39.938C. From
system.[20,21] A new joint venture by DuPont and Tate &
Lyle uses a fermentation process based on glucose.[22–24]
Hawley and co-workers have suggested a very selective
and high-yielding glycerol dehydroxylation to 1,3-propane-
diol, but the reaction sequence requires the stoichiometric
use of tosyl chloride.[25] Sasaki and co-workers recently re-
ported the use of a mixed Pt/WO3/ZrO2 catalyst that yielded
24% of 1,3-propanediol, 13% 1,2-propanediol and 28% 1-
propanol at 1708C and 8 MPa in 1,3-dimethyl-2-imidazolini-
dinone solvent, representing a benchmark result for the use
of heterogeneous catalysts in the direct transformation of
glycerol to 1,3-propanediol.[26]
their data DH=ꢀ25 kJmolꢀ1 and DS=ꢀ0.9 J/
ACHTUNGTREN(NUNG Kꢃmol), that
Motivated by the seminal work of Che,[27] that had shown
that glycerol can be converted into mixtures of 1,2- and 1,3-
propanediol in up to ꢂ20% yield each using a combination
is, the reverse dehydration of 3-HPA to acrolein is disfa-
vored in aqueous solution at low temperature, but is likely
entropically highly favored at the much higher temperatures
required to overcome the activation barrier for the initial
dehydration of glycerol. The determination of the optimum
reaction conditions for the reaction cascade of Scheme 1
therefore depends on an empirical approach that varies the
defining parameters, in particular the acid- and water-con-
centration in the reaction mixture as attempted in the work
described here.
of H2WO4/[Rh(CO)2ACTHNUGTRNEUNG
(acac)] and by Braca et al.,[28] who re-
ported the conversion of glycerol into 1-propanol by using
HI/[Ru(CO)4I2] as catalysts under synthesis gas (H2/CO) in
water or amide solvents at 2008C, we postulated that the
combination of an acid catalyzed dehydration with an in situ
hydrogenation of 3-HPA by pyridine-type ligand supported
homogeneous catalysts may be well suited to achieve a se-
lective transformation of glycerol to 1,3- rather than 1,2-pro-
panediol.[29] The desired reaction pathway along with its po-
tential side reactions is shown in Scheme 1.
Catalyst selection and preparation: Following a set of explic-
itly formulated design criteria,[31] we have previously tested
the catalysts [{(Cp*)Ru(CO)2}2-
AHCTUNGTRENNUNG
(m-H)]OTf ,[32] [cis-Ru(6,6’-Cl2-
[31]
bipy)
2ACHUTNGTREN(NUG OH2)2]ACHTNUGTRENNUNG
arene)Ru(X)
(N\N)]
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
activity in the deoxygenation of
terminal diols and glycerol
under aqueous acidic condi-
tions. All three catalyst types
convert terminal diols to the
corresponding primary alcohols
via acid-catalyzed dehydration
to the corresponding aldehyde
followed by its hydrogenation
in situ. Under more forcing
conditions further dehydration
to the alkene and hydrogena-
Scheme 1. Potential pathways and side reactions for the conversion of glycerol to 1,3-propanediol.
From Scheme 1, a selective conversion of glycerol to 1,3-
propanediol hinges on an acid-catalyzed preferential loss of
the secondary rather than primary hydroxyl function of glyc-
erol via a secondary rather than primary carbocation-like
transition state leading to 3-HPA rather than acetol as the
key intermediate. The reaction requires the adjustment of
the three (de)hydration equilibria glycerol/3-HPA (Keq), 3-
HPA/acrolein (K’eq) and 1,3-diol/allyl alcohol (K’’eq) in order
to maximize the initial dehydration of glycerol, while mini-
tion to the alkane can take place.[31] The first reaction serves
as a model system that mimics the desired glycerol deoxyge-
nation pathway shown at the top of Scheme 1 and catalyst
activity for the hydrogenation of aldehydes and/or ketones
(generated by dehydration of internal vic-diols) therefore
constitutes a necessary, but not sufficient condition for the
deoxygenation of glycerol and/or higher sugar alcohols by a
given catalyst.[29] In contrast to the deoxygenation of termi-
nal diols, which we found to proceed at temperatures as low
Chem. Eur. J. 2009, 15, 10132 – 10143
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10133

M. Schlaf et al.
as 1108C, control experiments established that the Brønsted
acid-catalyzed initial dehydration of glycerol requires tem-
peratures in excess of 1508C.[19] The catalysts tested by us to
date thus fail to convert glycerol, because they decompose
due to loss of their ligand support framework and reduction
to ruthenium metal at temperatures above 1258C as indicat-
ed by the formation of a black precipitate in the reaction
mixtures. An example for this limitation is the complex [cis-
ease of preparation from commercially available compo-
nents by (or for the new complex 4 analogous to) well estab-
lished synthetic protocols,[40–43] their stability in air and their
solubility and compatibility with water, which is intrinsic to
the aquo complexes themselves rather than a function of the
supporting ligand framework, as, for example, realized with
sulfonated triphenyl phosphines.[44]
Ru(6,6’-Cl2-bipy)2ACHTUNGTRENNUNG(OH2)2]HCATUNGTRENN(UGN OTf)2 (1Cl), which had originally
been developed as a water-soluble hydrogenation catalyst by
Lau and co-workers.[33,34] They postulated that the cis config-
uration of the complex - locked in this configuration by the
steric interactions between the ortho chloride substituents -
would result in an active ionic hydrogenation catalyst[35]
with two adjacent labile coordination sites. However, an ex-
amination of the extensive literature on the parent unsubsti-
Results
Catalyst screening against carbonyl substrates: To establish
the principle viability of complexes 1 to 4 as catalysts for
the deoxygenation of polyalcohols they were first screened
as hydrogenation catalysts for a series of 20 representative
carbonyl compounds employing conditions similar to those
anticipated in actual polyalcohol deoxygenation reactions
and using a 24-well parallel reactor with 4 of the wells serv-
ing as controls for cross-contamination (none was observed).
Reaction conditions were 9:1 sulfolane:water solvent at
1758C for 3 h under 7.5 MPa hydrogen pressure and a sub-
strate concentration of 500 mmolLꢀ1 with 0.5 mol% catalyst
load. Sulfolane was chosen as the solvent owing to its high
boiling point (2858C), high dielectric constant (e=43), mis-
cibility with water, allowing it to dissolve both polyalcohols
and the cationic catalysts, and its high chemical stability
under the reducing as well as aqueous acidic conditions nec-
essary for the deoxygenation of polyalcohols. All four cata-
lysts are stable against decomposition under these condi-
tions positioning them among of the most robust homogene-
ous hydrogenation catalysts ever reported. The catalysts
give >90% conversion of propanal and hexanal to 1-propa-
nol and n-hexanol, respectively and also convert a variety of
ketones to the corresponding secondary alcohols, albeit in
lower yields that scale inversely with the steric accessibility
of the carbonyl function in the substrate. Owing to secon-
dary dehydrations and hydrogenations as well as aldol con-
densation, oligomerization or polymerization, which are
driven by the high solvation energy of water in sulfolane,
very complex reaction mixtures with poorly defined mass
balances resulted for most of the ketones under these condi-
tions.[45]
tuted bipy [Ru
ACHTUNGTRENNUNG(H2O)2ACHUTNRTGE(GNNUN bipy)2]ACTHNUGTREN(GUNN OTf)2 (1) and phen [Ru-
A
N
ACHTUNGTRENNUNG
configuration represents the thermodynamically more stable
isomer in these systems and that the isomerization to the
trans-configuration requires photochemical activation.[36,37]
We therefore hypothesized that the parent bipy and phen
complexes 1 and 2 should also be active hydrogenation cata-
lysts under aqueous acidic conditions. Remarkably, and in
spite of being among best understood and most investigated
ruthenium complexes known with hundreds of papers pub-
lished, these complexes have, to our knowledge, never been
tested as hydrogenation catalysts. Owing to the lower steric
demand of the non-substituted ligands they should also be
less susceptible to ligand loss at higher temperatures, that is,
result in a thermally more robust catalyst system viable at
Tꢁ1508C, while at the same time allowing easier access of
the substrate to the metal centre. A further extension of this
concept are the complexes [Ru
ACHTNUGTNER(NUNG H2O)3ACHTUNRTGEGNUN(N terpy)]ACHTUNGRTEN(NUGN OTf)2 (3) and
[Ru(H2O)3A(Phterpy)](OTf)2 (4), (terpy=terpyridine; Phter-
A
N
ACHTUNGTRENNUNG
py=4’-phenyl-2,2’:6’,2’’-terpy) in which the tris-chelating
meridional coordination mode of the (4’-phenyl)-2,2’:6’,2’’-
terpy ligand[38,39] should result in a higher complex formation
constant (vs. the bis-chelating bipy or phen ligands) and
hence even higher temperature stability, while providing
three adjacent labile mer-oriented aqua ligands that can
easily be displaced by hydrogen and/or substrate leading to
a potentially more active catalyst with sterically easily acces-
sible coordination sites. The structures of these catalysts are
shown in Figure 1. Common to all four complexes is their
Terminal diol deoxygenation: Having established the princi-
pal activity of the catalysts 1–4 as carbonyl hydrogenation
catalysts we tested them in the deoxygenation of the 1,2-
hexanediol model system that mimics the desired deoxyge-
nation of the terminal diol unit in a sugar polyalcohol
(Scheme 1). 1,2-Hexanediol rather than 1,2-propanediol was
chosen as the glycerol model, owing to the lower volatility
of the potential deoxygenated products 1-hexanol, 1-hexene
and n-hexane vs 1-propanol, propene and propane enhanc-
ing the reliability of the quantitative analysis of the reaction
mixtures by GC against an internal standard. Table 1 sum-
marizes the results of the deoxygenation reactions per-
formed with this substrate comparing the observed product
distribution for the four newly tested catalysts 1–4 with that
Figure 1. Structures of the catalysts.
10134
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10132 – 10143

High-Temperature Stable Ruthenium Complexes
FULL PAPER
Table 1. Deoxygenation of 1,2-hexanediol by [Ru
(1), cis-[Ru(6,6’-Cl2-bipy)2A(OH2)2](OTf)2 (1Cl), [Ru(OH2)
(2), [Ru(OH2)3A(terpy)](OTf)2 (3) and [Ru(OH2)3(4’-phenyl-2,2’:6’,2’’-
terpy)]
(OTf)2 (4) as a function of temperature.[a]
G
2A
trol experiments without the addition of the homogeneous
catalyst gave negligible conversions (<1%) of the diol sub-
strate.
N
G
E
(phen)2]
ACHTUNGTRENNUNG
G
E
G
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
The data in Table 1 reveal the anticipated correlation of
temperature stability of the catalysts with decreasing ligand
steric demand and increasing chelation. The chloride-substi-
tuted catalyst 1Cl decomposes at temperatures above
1258C, the unsubstituted catalysts 1 and 2 show decomposi-
tion as evidenced by the formation of a black ruthenium
metal precipitate in the reaction mixture at 150 and 1758C,
respectively, whereas the tris-chelate catalysts 3 and 4 are
active at up to 2508C—for catalyst 4 without any evidence
of decomposition, that is, depending on the reaction temper-
ature (see below) clear orange to deep purple-red solution
are obtained at the end of the reaction. The relative temper-
ature stability of the catalysts also directly scales with the
minimum temperature at which they become active. Cata-
lyst 1, 1Cl and 2 are active at 1258C, whereas 3 and 4 re-
quire 150 or 1758C to show conversion. The maximum yield
of 1-hexanol at the 24 h time point was 26% observed with
catalyst 3 at 1508C and the maximum yields of the total de-
oxygenation product hexane quantifiable from the sulfolane
solution is 5% with catalysts 3 and 4 at temperatures
>1758C. For catalysts 3 and 4 these numerical results how-
ever systematically underestimate the actual catalyst activity,
as a hexane layer separates from the sulfolane phase above
these concentrations, that is, ꢂ5% (here equivalent
25mmolL)ꢀ1 represents the solubility limit of hexane in sul-
folane. The identity of the hexane was confirmed by GC
Catalyst
T [8C]
1
1Cl
2
3
4
Recovery
Yield of identified products [%][b]
125
150
175
200
225
250
1,2-hexanediol
1-hexanol
hexane
48
3
0
68
23
2
94
1
0
>95[c] >95[c]
0
0
0
0
hexane layer separates no
no
no
no
no
1,2-hexanediol
1-hexanol
hexane
0
17
0
n/d[d]
n/d
n/d
no
4
14
0
0
26
3
93[c]
0
0
hexane layer separates no
no
no
no
1,2-hexanediol
1-hexanol
hexane
0[d] n/d
1
0
0
11
0
0
9
77
5
0
n/d
n/d
n/d
5[e]
hexane layer separates no
no
yes
no
1,2-hexanediol
1-hexanol
hexane
0[d] n/d
0
2.0 n/d
0[d]
4
2
0[d]
0
21
n/d
0
6[e]
5[e]
yes
hexane layer separates no
n/d
no
yes
1,2-hexanediol
1-hexanol
hexane
n/d
n/d
n/d
n/d
n/d
n/d
n/d
0[d]
0
3
n/d
n/d
n/d
n/d
0
0
5[e]
yes
hexane layer separates n/d
no
1,2-hexanediol
1-hexanol
hexane
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
n/d
0
0
5[e]
yes
hexane layer separates n/d
1
and GC-MS analysis, and H NMR spectroscopy. As the in-
[a] Reaction conditions: 500 mmolLꢀ1 substrate in sulfolane,
100 mmolLꢀ1 dimethylsulfone as internal standard, 0.5 mol% catalyst, 4
equivalent of HOTf w.r.t. Ru, 4.82 MPa of H2(g), 24 h. [b] By quant. GC
and GC-MS of the sulfolane reaction mixture. [c] No activity. [d] Catalyst
decomposition observed, n/d=not determined. [e] Hexane content of the
sulfolane phase. A hexane layer separates from the sulfolane solution,
that is, 5% represents the solubility limit of hexane in sulfolane—not the
total hexane yield—see main text.
ternal standard is not present in the hexane layer and some
of the hexane is lost to the gas phase upon venting the
cooled reactor upon completion of the reaction the amount
of hexane formed cannot be quantified by GC. Attempts to
cleanly separate the phases and determine the amount of
hexane formed by weighing also proved to be unsatisfactory.
However, the 1-hexanol concentration vs time profiles as a
function of temperature obtained from reactions catalyzed
by 4 illustrated in Figure 2 show a transient build-up of a
much higher concentration of 1-hexanol (up to 40% yield)
than that found after 24 h, which led us to conclude that cat-
alysts 3 and 4 do in fact catalyze a quantitative deoxygena-
tion of 1,2-hexanediol to hexane through the pathway 1,2-
of the previously reported Lau-system 1Cl[31,33,34] and estab-
lishing the temperature limits of the catalysts. All reactions
were performed in a 50 mL autoclave for 24 h under
4.82 MPa hydrogen pressure (at ambient temperature at t=
0) with 500 mmolLꢀ1 solutions of 1,2-hexanediol in sulfo-
lane, 100 mmolLꢀ1 dimethylsulfone as an internal GC stan-
dard, 0.5 mol% catalyst load and 4 equivalents of HOTf per
ruthenium as the acid dehydration catalyst. HOTf was
chosen as the acid, owing to its high hydrolytic stability,
non-oxidizing nature and the low coordination tendency of
the triflate counterion. The latter is essential to avoid coor-
dinative inhibition of the catalytic cycle in which hydrogen
competes with all other possible ligands (substrate, water,
solvent, counterions) present. The yields given in Table 1 re-
flect the concentration of the sulfolane soluble components
of the reaction mixture after 24 h and were determined by
quantitative GC with multi-level calibration against authen-
tic solutions of 1,2-hexanediol, 1-hexanol and the internal
standard dimethylsulfone (100 mmolLꢀ1) in sulfolane. Con-
Figure 2. 24 h time-profile for the concentration of 1-hexanol in the de-
AHCTUNGTREGUNoNN xygenation of 1,2-hexanediol catalyzed by 4.
Chem. Eur. J. 2009, 15, 10132 – 10143
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10135

M. Schlaf et al.
hexanediol!hexanal!1-hexanol!hexene!hexane. This is
also supported by the presence of traces of hexene in the re-
action mixture, but complete absence of other GC identifia-
ble condensation products (e.g. dihexyl ether, 1,3-dioxolanes
or 1,4-dioxanes)[19] in the GC traces of these reactions. Con-
trol experiments under identical conditions but using 1-hex-
anol as the substrate yielded mixtures of hexane and hexene
in the presence and mixture of dihexyl ethers and hexene in
the absence of metal catalyst.
troscopy, MALDI-TOF MS analysis, and single crystal X-
ray crystallography. Figure 3 shows an ORTEP plot of the
structure of the cation.[48] Complex 5 crystallizes in the tri-
In contrast the equally low mass balances observed with
the catalysts 1, 1Cl and 2 that can only operate at much
lower temperatures are caused by the formation of higher
molecular weight (poly)ether condensates that are not vola-
tile enough for GC analysis. No hexane is formed with these
catalysts and at low conversions the GC traces of these reac-
tion show the presence of 2-pentyl-4-butyl-1,3-dioxolane and
cis/trans-2,5- and cis/trans-2,6-1,4-dioxolane and 1,1-, 1,2-
and 2,2-hexanediol ethers that were also identified in our
earlier work, but not quantified in this study.[19,31]
From these experiments the terpy systems 3 and 4 emerge
as some of the most robust homogeneous hydrogenation
catalysts ever reported[46] and further investigations focused
on these, in particular on the 4’-phenyl-2,2’:6’,2’’-terpy based
catalyst 4, as this ligand is substantially easier to prepare
than the parent terpy ligand.[39] Initially, solutions of the tri-
Figure 3. ORTEP plot of the cation of [{(4’-phenyl-2,2’:6’,2’’-
terpy)Ru(CO)}2ACHTUNGRTENNUG(m-OCH3)2]ACHTUTGNREN(NUNG OTf)2 (5) as determined by single crystal X-
ray crystallography. Ellipsoids are drawn at 50% probability level. Hy-
drogen atoms, triflate counterions and methanol solvent contained in the
unit cell are omitted for clarity.
¯
clinic space group P1 (a=8.2532(17); b=12.858(3); c=
14.363(3)8; a=64.38(3); b=77.26(3); g = 87.12(3) ꢁ, R=
4.36%). The characteristic bond length about the ruthenium
ꢀ
ꢀ
ꢀ
aqua complexes 3 and 4 in (aqueous) sulfolane are deep
centre are (in ꢁ) Ru N11=2.090, Ru N22=1.972, Ru
ꢀ
ꢀ
red-purple and air-stable. Samples taken during or after re-
action at temperatures up to 2008C appear yellow-orange,
but reactions using 4 at temperatures >2008C revert to a
clear red-deep purple at the end of the reaction with no visi-
ble decomposition or formation of a precipitate. This cata-
lyst solution can be reused without apparent loss of activity,
that is, after reconstitution of the 1,2-hexanediol concentra-
tion to 500 mmolLꢀ1 and heating to T>2008C the recycled
solution again achieves complete conversion of the substrate
to hexane.
N23=2.094, Ru O1=2.131 and Ru C3=1.835. The charac-
teristic bond angles in the planar core of the dimer are (in
deg) N11-Ru-N22=79.24, N22-Ru-N23=79.53, N11-Ru-
N23=158.77, O1-Ru-O1i =76.39, Ru-O1-Rui =103.81, C3-
Ru-O1=173.81, O1-Ru-N22=166.77 and N22-Ru-C3=
95.74 resulting in a distorted octahedral coordination envi-
ronment for the ruthenium.
The IR spectrum of the dimer 5 in the solid state (KBr)
shows two n(CO) bands at 1975 and 2079 cmꢀ1 correspond-
ing to the strong symmetry allowed Bu and the weak sym-
metry forbidden Ag band in the C2h point group of the
cation dimer, whereas in MeOH only a single n(CO) band
at 1967 cmꢀ1 is observed, suggesting that in solution and
with participation of traces of water present in the MeOH
the dimer dissociates into a mononuclear complex [(4’-
Nature of the active catalyst: The distinct temperature de-
pendent color changes of 4 observed in the course of the
catalytic reactions prompted us to investigate this system in
more detail. Starting with the hypothesis that a reaction of
the pro-catalyst [Ru
G
G
phenyl-2,2’:6’,2’’-terpy)Ru(CO)ACHTUNGERTN(GNUN H2O)ACHTUNTRGEG(NNUN OCH3)]CAHTNUGTREN(GUNN OTf) (6) or
(4) with the actual hydrogenation substrate hexanal (formed
by the acid-catalyzed dehydration of 1,2-hexanediol[19]—see
analogous reaction for glycerol in Scheme 1) causes the
color change, the triaqua complex 4 was reacted with excess
hexanal in deuterated methanol at 1558C (sealed NMR
tube) for three days and the reaction monitored by
1H NMR. The same deep purple to orange–yellow color
change as in the catalytic reactions in sulfolane along with a
complete change of the appearance of the aromatic region
of the NMR spectrum was observed.[47] Scale-up of this reac-
tion followed by slow evaporation of the methanol solvent
allowed the isolation of the methoxy-bridged centro-sym-
metric dimeric carbonyl complex [{(4’-phenyl-2,2’:6’,2’’-
in the presence of triflic acid and water as present in the
actual reaction mixtures into [(4’-phenyl-2,2’:6’,2’’-
terpy)Ru(CO)ACHTUNRTGENNUG(H2O)2]ACHTUTGNREN(NGUN OTf)2 (7) as the actual catalyst rest-
ing state in aqueous solution. The dissociation reaction is il-
lustrated in Scheme 2 showing one of several conceivable
stereochemical arrangements of the aquo, methoxy and car-
bonyl ligands in 6 and 7. For comparison the n(CO) band in
the related complexes [(2,2’:6’,2’’-terpy)Ru(CO)(X)2], X=
Cl, Br, whose X-ray structure has also been determined,
appear at 1948 and 1944 cmꢀ1 (Nujol).[49]
When complex 5 is used as the pro-catalyst using the
same reaction conditions (0.5 mol% ruthenium) as listed in
the entries for 200–2508C in Table 1 the same result as for
complex 4 is observed, that is, complete conversion of 1,2-
hexanediol to 1-hexanol and hexane with phase separation
terpy)Ru(CO)}
2ACHTUNGTRENNUNG(m-OCH3)2]ACHTUGNTREN(NUGN OTf)2 (5) as an orange solid,
1
which was fully characterized by H, 13C NMR, and IR spec-
10136
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10132 – 10143

High-Temperature Stable Ruthenium Complexes
FULL PAPER
Scheme 2. Dissociation of dimer complex 5 into monomeric complexes 6 and/or 7.
of hexane from the sulfolane reaction mixture is observed.
The color of the initially orange–yellow reaction mixture at
the end of the reaction is purple, matching that of the reac-
tions performed with 4 at these temperatures. When 1-hexa-
nol is used as the hydrogenation substrate no color change
to orange–yellow takes place, but slow hydrogenation to
hexane is still observed along with the formation of di-n-
hexylether, which is not present in reaction mixtures starting
from 1,2-hexanediol.
glycerol in sulfolane/water) and 10–50% (v/v) of water con-
tent.
In no instance was 1,3-propanediol or 1,2-propanediol de-
tected in these reactions at any time. Instead only the
double deoxygenation product 1-propanol could be quanti-
fied, whose yields as a function of water (in% v/v) and acid
(in mol equivalents with respect to ruthenium) content for
catalysts 3 and 4 are graphically represented in Figure 4 and
Figure 5, respectively. For catalyst 3 yields of up to ꢂ5% of
1-propanol are present in the reaction mixture after 2 h at
low water content (10%), increasing to a maximum of
ꢂ14% at 24 h and 40% water content with variable
From these observations we postulate, that in the orange–
yellow solutions formed by using 3 or 4 as the procatalysts,
the active hydrogenation catalysts are in fact the carbonyl
complexes [(L)Ru(CO)
2,2’:6’,2’’-terpy or ultimately the corresponding transient hy-
dride complexes [(L)Ru(CO)(H2O)(H)](OTf) or possibly
ACHUTGTNRENG(NU H2O)2]ACHTUNGTERN(NUGN OTf)2; L=terpy, 4’-Phenyl-
A
ACHTUNGTRENNUNG
[(L)Ru(CO)(H)2] resulting from the heterolytic activation
of hydrogen gas.[50] The carbonyl complexes [(L)Ru(CO)-
ACHTUNGTRENNUNG(H2O)2]ACHTUNGTRENNUNG(OTf)2 are formed in situ through decarbonylation
of the actual hydrogenation substrate hexanal. The (cata-
lytic) decarbonylation of aldehydes at ruthenium centers has
multiple precedents in the literature.[51–54] The color change
back to purple after complete conversion of the substrate to
1-hexanol and hexanes at temperatures> 2008C then points
to a thermal loss of CO(g) from the complexes [(L)Ru(CO)-
A
ACHTUNGTRENNUNG(OTf)2 regenerating the triaqua complexes [(L)Ru-
A
ACHTUNGTRENNUNG
medium.
Deoxygenation of glycerol:
In sulfolane solution: After establishing the tolerance of
pro-catalysts 3 and 4 to the temperatures required for the
initial dehydration of glycerol in acidic sulfolane solution we
investigated the deoxygenation of glycerol by these catalysts
as a function of acid and water content of the sulfolane reac-
tion mixture at t=0 h using a 24ꢃ2 mL well parallel reactor
at the upper temperature limits of catalyst 3, that is, 1758C
avoiding any decomposition (cf. Table 1) and the upper tem-
perature limit of the reactor set-up, that is, 2008C for cata-
lyst 4, well below the decomposition temperature of this cat-
alyst. With four of the individually cooled wells containing
sulfolane only and thus serving as controls for cross-contam-
ination (none was observed), the reactor allows the simulta-
neous screening of a 4ꢃ5 array of reaction mixtures ranging
in triflic acid content from 4, 8, 12 to 16 equivalents with re-
spect to ruthenium catalyst (identical conditions as for 1,2-
hexanediol, that is, 0.5 mol% in a 500 mmolLꢀ1 solution of
Figure 4. Yield of 1-propanol in the deoxygenation of glycerol with 3 in
sulfolane at 1758C, 7.58 MPa H2(g) and variable amount of water and
HOTf added as indicated. Top: t=2 h; bottom: t=24 h.
Figure 5. Yield of 1-propanol in the deoxygenation of glycerol with 4 in
sulfolane at 2008C, 7.58 MPa H2(g) after 2 h and variable amount of
water and HOTf added as indicated.
Chem. Eur. J. 2009, 15, 10132 – 10143
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10137

M. Schlaf et al.
amounts of glycerol remaining in the reaction mixture (0%
glycerol at 10% water content to ꢂ50% glycerol at 50%
water content). No clear trends or quantitative correlations
emerge between the acid and water content of the reaction
mixtures and the amount of propanol observed, reflecting
the complexity of the interconnected (de)hydration equili-
bria discussed earlier (Scheme 1). Catalyst 4 shows similar
behavior, that is, there is no simple correlation of the 1-
propanol content with the acid or water content of the reac-
tion mixture at the 2 h time point. After 24 h the GC traces
of the reaction mixtures show only very small amounts of 1-
propanol, acrolein and pairs of peaks assigned to traces of
the isomeric mixtures of cis/trans-2-(2’-hydroxy-ethyl)-4-(hy-
droxymethyl)-1,3-dioxolane (by GC-MS) formed by acetali-
zation of glycerol with 3-HPA. In total these amount to
<25 mmolL,ꢀ1 that is, <5% of the initial glycerol concen-
tration. After 24 h no glycerol was detectable for the reac-
tions containing 10 and 20 water% and 4, 8, 12 and 16
equivalents of HOTf acid, that is, complete conversion of
glycerol had taken place. In light of our earlier result with
the 1,2-hexanediol model substrate, we therefore postulated
that the very low mass balances of these reactions as ob-
served by quantitative GC of the liquid reaction mixtures
are the consequence of the formation of the gaseous total
deoxygenation product propane.
In NMP solution: In an attempt to attenuate the overall
acidity of the reaction solution the solvent was switched to
the slightly basic N-methyl-pyrrolidin-2-one (NMP)[56] while
keeping all other parameters for the glycerol deoxygenation
reactions constant. Regardless of the acid/water content (4-
16 equiv/10-50% using the parallel reactor as before) of the
reaction solution no deoxygenation of glycerol was observed
in this solvent at 2008C and after 24 h the concentration of
glycerol almost equaled the initial concentration. Black pre-
cipitate was observed in the reaction mixture indicating cat-
alyst decomposition. Using a single well 50 mL reactor,
which allows an increase of the reaction temperature to
2508C, and with 4 equiv of HOTf and 4.82 MPa of H2(g)
pressure (cold) the reaction yielded 4% 1-propanol, 12% of
cis/trans-2-(2’-hydroxy-ethyl)-4-(hydroxymethyl)-1,3-dioxo-
lane (equivalent to 24% glycerol) and 68% glycerol recov-
ery along with catalyst decomposition for a total mass bal-
ance of 98%, that is, no higher molecular weight conden-
sates or gaseous products were formed. As in the reactions
carried out in sulfolane the reaction mixture underwent the
color changes from purple to orange (sample taken after
1 h) and back to purple over the course of the 24 h reaction
run again suggesting transient formation of a carbonyl com-
plex [(4’-phenyl-2,2’:6’,2’’-terpy)Ru(CO)ACHTNUTRGENNUG(H2O)2]ACHTUGNTERN(NUGN OTf)2 as
the active catalyst.
To be able to prove the production of propane by probing
the gas headspace of these reactions (technically not possi-
ble nor meaningful with the parallel reactor) the reaction of
catalyst 4 at a water content of 10% and with 8 equivalents
of acid was repeated with 100 mL of reaction solution in a
600 mL Parr reactor at 2008C and 7.58 MPa H2(g) pressure
for 24 h. GC/GC-MS analysis of gas samples obtained from
the 500 mL headspace volume of this reaction after ꢂ3 h
showed a dominant peak of propane and traces of acrolein,
propanal, propene and 1-propanol (identified by comparison
with authentic gas/vapor samples).[55] Gas samples obtained
after 24 h showed pure propane. GC analysis of the reaction
solution after 24 h did not show any peaks other than those
of the solvent sulfolane and internal standard dimethylsul-
fone (DMS) and MALDI-TOF MS analysis of the same re-
action solution in both positive and negative ionization
mode showed the presence of triflic acid and various frag-
ments of the ruthenium catalyst, but equally did not reveal
the presence of any non-volatile glycerol condensates or
polyethers and no precipitate was observed in the reaction
mixture. From this we conclude that under the reaction con-
ditions employed, catalyst 4 quantitatively converts glycerol
into propane.
Also carried out were reactions designed to mimic the
conditions and results reported in the patent by Che, which
claims production of substantial amounts of both 1,2-pro-
panediol and 1,3-propanediol in NMP solvent using a com-
bined [Rh(CO)2ACTHUNTRGNE(UNG acac)]/H2WO4 catalyst under 31.7 MPa of
syn-gas atmosphere.[27] Deoxygenation reactions were car-
ried out with catalyst 4 at 225 and 2508C using 6.20 MPa of
H2(g) pressure (cold) and at the same catalyst, substrate and
acid concentrations that gave the optimum yields of 1,3-pro-
panediol reported in the patent.[57] Samples were taken at 0,
1, 2, 4, 8, and 24 h. The reaction mixtures are initially purple
with yellow suspended H2WO4 (that is, the tungstic acid is
not soluble in NMP) and again turn orange after 1 h of heat-
ing under hydrogen. As in the other NMP reactions a black
precipitate appears indicating catalyst decomposition. The
samples taken at the 1, 2, 4 and 8 h time points showed no
deoxygenation products. After 24 h in both reactions ꢂ5%
of 1-propanol, ꢂ2.5% of 1,3-propanediol, ꢂ5% 1,2-pro-
panediol and ꢂ5% cis/trans-2-(2’-hydroxy-ethyl)-4-(hydrox-
ymethyl)-1,3-dioxolane had formed as verified by GC and
GC-MS against authentic samples. 36% of the glycerol re-
mained along with other condensation products that could
not be identified by GC-MS. Running the reaction using our
own standard concentrations[58] at 2258C yielded ꢂ4% of 1-
propanol with 95% of glycerol recovery. Again a black pre-
cipitate was formed indicating decomposition of the catalyst.
In further control experiments using catalyst 4 and 1,3-
propanediol instead of glycerol, but otherwise identical con-
ditions, transient 1-propanol yields of up ꢂ8% are observed
in samples drawn from the reactor at 3 h time intervals after
3, 6, and 9 h reaction time, but after 24 h the same result as
glycerol is obtained, that is, any 1,3-propanediol that may
have formed would under the reaction conditions required
for the initial dehydration of glycerol also be converted into
1-propanol and ultimately propane.
In water: The requirement for an acid-catalyzed dehydration
of glycerol makes the use of pure water as the reaction
medium counterintuitive, as the loss of water to 3-HPA or
acrolein occurs against the equilibria (Keq and K’eq in
Scheme 1). However, in terms of overall reaction medium
10138
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10132 – 10143

High-Temperature Stable Ruthenium Complexes
FULL PAPER
acidity discussed above pure water falls in between sulfolane
or sulfolane/water mixtures and the basic NMP solvent,
which prompted us to also investigate pure water as the re-
action solvent and again using complex 4 as the procatalyst.
We also hypothesized that the use of pure water as the reac-
tion medium may inhibit the secondary dehydration of 3-
HPA to acrolein, thus favoring the formation of the desired
1,3-propanediol. As the results listed in Tables 2 and 3 show,
glycerol is in fact efficiently deoxygenated in pure water
using either HOTf or H2WO4 as the acid catalyst, where the
latter is not soluble, but as in NMP forms a suspension, but
again no 1,2- or 1,3-propanediol was formed.
H2(g) consumed (2 mol equivalents/propanol and 3 mol
equivalents/propane generated) and any propane formed
the observed pressure drops measured after cooling the re-
actor back down to ambient temperature (295 K) give an
approximate yield of propane. For the reactions where pro-
pane was formed, the observed pressure drops matched the
expected theoretical value within ꢂ10% and the presence
of propane was again verified by GC-MS of the gas head-
space of the reactor. As with the reaction is sulfolane and
NMP, MALDI-TOF MS analysis in both positive and nega-
tive ionization mode of the liquid phase did not show any
non-volatile condensation products resulting in the quantita-
tive mass balances listed in Tables 2 and 3 and further af-
firming the results obtained in sulfolane presented above.
With HOTf as the acid co-catalyst (Table 2) the reaction
was carried out at 200, 225 and 2508C. At the two lower
temperatures both the conversion of glycerol and yields of
1-propanol and propane scale with the amount of acid
added, whereas at 2508C complete conversion of the sub-
strate is achieved in 24 h in all cases. The amounts of 1-
propanol detected in these reaction mixtures vary, but show
a decreasing trend with increasing acid content. The small
amounts of 2-propanol formed are attributed to Markovni-
kov rehydration of the intermediate propene. The solutions
All reactions listed in Tables 2 and 3 were carried out in a
50 mL Autoclave Engineers mini-reactor. In an attempt to
directly quantify the amount propane formed in these reac-
Table 2. Deoxygenation of glycerol by [Ru
terpy)](OTf)2 (4) in pure water as a function of HOTf content and tem-
perature.[a]
ACHTUGNTERN(NUNG OH2)3(4’-phenyl-2,2’:6’,2’’-
ACHTUNGTRENNUNG
T
Mol equiv Glycerol[b] 1-Propanol[b] 2-Propanol[b] n-Propane[c]
[oC] HOTf/Ru [%]
[%]
[%]
[%]
200
200
200
16
32
64
92
82
76
49
59
26
12
5
0
0
0
0
8
12
14
19
19
28
35
35
35
22
17
23
2
0
0
0
0
0
0
0
0
0
0
2
3
0
1
0
0
6
10
32
22
46
53
60
65
78
81
74
98
93
100
200 100
of the [RuACHTUGNRTNEN(UNG Phterpy)ACHTNUTREGNG(NNU OH2)3]ACHTNGURTEN(NUNG OTf)2 (4) catalyst in water were
225
225
225
16
32
64
initially purple, but became colorless after complete conver-
sion of the substrate with formation of an orange-brown
precipitate. The IR spectrum (KBr) of this solid shows one
band at n˜ =1612 cmꢀ1, its H1 NMR spectrum in [D4]MeOH
only resonances assignable to the 4-Ph-terpy ligand and the
MALDI-TOF mass spectrum only fragments assignable to
{Run(4-Ph-terpy)m} (n, m=1, 2) species, that is, the precipi-
tate is not composed of polymeric organic species formed
form the glycerol substrate, but originates from a reaction of
the catalyst to a water insoluble ruthenium complex of un-
known structure.
225 100
250
250
250
250
250
250
4
8
12
16
32
64
0
0
0
6
0
250 100
[a] Reaction conditions: 500 mmolLꢀ1 substrate in water, 100 mmLꢀ1 di-
methylsulfone as internal standard, 0.5 mol% catalyst, 4.82 MPa of
H2(g), 24 h. [b] By quant. GC and GC-MS. [c] By GC-MS, mass balance
inferred and by pressure drop.
In light of the trace of 1,3-propanediol observed in the re-
actions using H2WO4 in NMP and recognizing the possibly
unique properties of H2WO4 as a solid surface-active acid in
a selective dehydration of glycerol,[26,27] we then carried out
a second series of experiments using the combination of pro-
catalyst 4 with H2WO4 in water, the results of which are
summarized in Table 3. As with the reactions in NMP a
black precipitate formed with use of H2WO4 as the acid and
again no 1,2- or 1,3-propanediol was detected in any of the
reaction mixtures and the yields of 1-propanol and propane
are comparable to those obtained with the homogeneous
acid HOTf. Notable is the build-up of 1-propanol and prod-
uct distribution with 8 equivalent of H2WO4 observed after
2 h of reaction time that resembles the reaction profile ob-
tained in the deoxygenation of the 1,2-hexanediol model
system (Figure 2) and the fact that glycerol conversion
occurs even without the addition of acid. This is due to the
formation of free HOTf in the reaction mixture through the
heterolytic activation of H2(g) by the catalyst into protons
and a hydride complex, a standard feature observed with
ionic hydrogenation catalysts.[32,59,60]
Table 3. Deoxygenation of glycerol by [Ru
terpy)](OTf)2 (4) in pure water as a function of H2WO4 acid and temper-
ature.[a]
ACHTUGNTERN(NUNG OH2)3(4’-phenyl-2,2’:6’,2’’-
ACHTUNGTRENNUNG
T
Mol equiv Glycerol[c] 1-Propanol[c] 2-Propanol[c] n-Propane[c]
[oC] H2WO4/Ru [%]
[%]
[%]
[%]
250
250
250
250
0
1
4
8
36
7
0
0
53
35
39
30
24
26
0
2
5
0
0
29
52
65
76
21
250 8[d]
[a] Reaction conditions: 500 mmolLꢀ1 substrate in water, 100 mmLꢀ1 di-
methylsulfone as internal standard, 0.5 mol% catalyst, 5.5 MPa of H2(g),
24 h. [b] By quant. GC and GC-MS. [c] By GC-MS, mass balance inferred
and by pressure drop. [d] 2 h.
tions, the reactor was fitted with a pressure sensor allowing
tracking of the total pressure as a function of the reaction
chemistry occurring. Assuming ideal gas behavior for the
Chem. Eur. J. 2009, 15, 10132 – 10143
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10139

M. Schlaf et al.
Discussion
step. Thus our results suggest that it may not be possible to
realize an effective catalyzed synthesis of 1,3-propanediol
from glycerol using catalyst systems in which both the acid
and metal catalyst employed are homogeneous. In this sce-
nario the relative hydrogenation activity of the metal cata-
lyst has no impact on the product distribution, which is in-
stead entirely governed by the selectivity of the dehydration
reactions. The formation of 1,3-propanediol in the presence
of H2WO4 as observed by Che and ourselves then logically
suggests that the surface of this heterogeneous acid must
have specific interactions with the glycerol substrate leading
to the desired chemo- and regioselectivity of dehydration to
3-HPA rather than acetol as observed with other solid acids
(see the results reported by Sasaki and coworkers, who also
used H2WO4).[26]
Regardless of the wide range of acid/water concentrations
employed no 1,3-propanediol could be detected in any of
the homogeneous sulfolane or water reaction mixtures. A
marginal amount of both 1,2- and 1,3-propanediol along
with 1-propanol was produced in the NMP/H2WO4 reaction
mixtures, which, however, led to catalyst decomposition. In
analogy to the known selectivity of heterogeneous catalyst
systems for the conversion of glycerol to 1,2-propanediol
through dehydration to acetol,[19] the appearance of 1,2-pro-
panediol in the latter must be a function of the heterogene-
ous nature of the tungstic acid in NMP medium. The pres-
ence of 1,3-propanediol also suggests that the heterogeneous
H2WO4 acid is less active in catalyzing the secondary dehy-
dration reaction of 3-HPA to acrolein.
In the hydrogenation reactions, hydride complexes must
be formed from the aqua-complexes 1–4 under the reaction
conditions in order to arrive at a catalytically active system.
However our inability to synthesize and/or isolate any such
complexes precluded any meaningful mechanistic studies. In
analogy to the mechanism proposed by Lau and co-workers
for 1Cl[33,34] the mechanism of the catalytic hydrogenation of
the aldehydes and alkenes formed in the reaction mixtures
is very likely ionic in nature,[35] that is, relies on the hetero-
lytic activation of hydrogen gas on the relatively electron
poor ruthenium centre making the overall catalytic cycle
compatible with the necessarily aqueous acidic reaction en-
vironment. The actual hydrogenation then occurs either by
direct transfer of a hydride ligand from the ruthenium
centre to a protonated substrate without actual coordination
to the substrate to metal centre as observed for the related
In contrast, in purely homogeneous phase with HOTf as
the acid catalyst and under the reaction conditions em-
ployed (that is, Tꢁ1758C in sulfolane or water solvent), the
actual hydrogenation substrate is not 3-HPA, but likely
acrolein formed by the rapid double dehydration of glycerol.
With reference to Scheme 1 this means that the activation
barrier for the initial dehydration of glycerol (associated
with k1) must be higher than that for the second dehydration
to the a,b-unsaturated acrolein (associated with k2) and -
since control reactions in sulfolane solvent with 1,3-propane-
diol as the substrate yielded the same result as with glycerol
- also higher than that for any other subsequent acid-cata-
lyzed dehydration, that is, DGꢀ(k1)@ DGꢀ(k2), DGꢀ(k3) and
DGꢀ(k4) and for our catalysts probably also DGꢀ(k2)!
ꢀA
ACHTUNGTRENNUNGDG CHTUNGTRENNUNG(kdiol). As all metal catalyzed hydrogenations of the re-
sulting aldehydes to alcohols or alkenes to alkanes are ther-
modynamically favored, this then explains the formation of
1-propanol and ultimately propane as the only detectable
products under homogeneous acidic conditions. Figure 6
qualitatively summarizes the resulting reaction cascade, in
which the initial glycerol dehydration is the rate determining
(terpy)]+ and CO2 by Ishitani[61] and
complex [RuHACTHNGUTRENNU(G bipy)ACHTNUGTRENNUNG
Creutz[62] or by insertion of the carbonyl or alkene substrate
ꢀ
into a Ru H bond followed by release of the substrate
ꢀ
through protonation of the resulting Ru-O or Ru C bond as
proposed by Lau for 1Cl. The latter could be considered a
reverse ionic hydrogenation mechanism.
Figure 6. Reaction cascade for the deoxygenation of glycerol to propane under acidic conditions in homogeneous phase.
10140
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10132 – 10143

High-Temperature Stable Ruthenium Complexes
FULL PAPER
AHCTNURTGEN[GNU cis-RuACHTUNGTREN(NUGN bipy)2ACHTNUEGTRGN(NNU OH2)2]AHTCUGNRTNE(NUNG OTf)2 (1): [cis-RuCl2ACTHNUTGENRUGN(bipy)2]·2H2O (0.522 g,
Conclusions
1.003 mmol) was stirred in degassed H2O (50 mL) under Ar. AgOTf
(0.527 g, 2.050 mmol) was added, and the mixture was stirred at 708C for
1 h to give an orange-brown solution with a white precipitate. Upon cool-
ing, the mixture was filtered through a plug of Celite to remove the
AgCl. The H2O was removed at 508C on the vacuum line, and the red
solid was vacuum-dried overnight. Yield: 0.6451 g (86%). 1H NMR
(300 MHz, D2O): d=6.98 (t, J=6.5 Hz, 1H), 7.65 (m, 2H), 7.78 (t, J=
6.5 Hz, 1H), 8.14 (t, J=7.5 Hz, 1H), 8.26 (d, J=8.1 Hz, 1H), 8.47 (d, J=
8.4 Hz, 1H), 9.26 ppm (d, J=5.4 Hz, 1H); 13C NMR (75 MHz, D2O): d=
123.2 (CH), 123.4 (CH), 125.6 (CH), 126.8 (CH), 135.7 (CH), 137.4
(CH), 151.5 (CH), 154.4 (CH), 158.3 (C), 160.5 ppm (C); elemental anal-
ysis calcd (%) for C22H20F6N4O8RuS2: C 35.34; H 2.70; N 7.49; found: C
35.18; H 2.58; N 7.44.
In combination with HOTf as a Brønsted acid co-catalyst
the complexes [Ru
G
N
ACHTUNGRTEN(NGNU OTf)2 (3) and [Ru-
A
(OTf)2 (4) form an air-,
water-, acid- and high-temperature (> 2508C) stable catalyst
system for the total deoxygenation of glycerol to propane in
sulfolane or aqueous medium. Under these reaction condi-
tions this is the result of the high propensity of the initial de-
hydration and hydrogenation products 3-hydroxy-propional-
dehyde (3-HPA) and 1,3-propanediol to undergo secondary
Brønstedt acid-catalyzed dehydration reactions to acrolein
or allyl alcohols, both of which are hydrogenated to n-prop-
anol, followed by a third dehydration to propene and hydro-
genation to the final product propane, which under the re-
ducing reaction conditions (hydrogen atmosphere) consti-
tutes the thermodynamic sink of the system.
The high temperature stability of the 4’-phenyl-2,2’:6’,2’’-
terpy complexes suggests that these catalysts may also be
applicable to deoxygenation reactions of higher sugars and
sugar alcohols to high-value added products, for example,
the conversion of erythritol to THF,[63,64] xylitol to hydroxy-
methyl tetrahydrofuran, d-fructose to 2-hydroxymethyl-6-
methyl-tetrahydrofuran and sorbitol to other deoxygenated
species. Experiments towards the realization of these goals
using these catalysts are currently under way in our labora-
tories.
[RuACHTNURTGENN(UG CO3)ACHTUNGTREN(NUGN phen)2]: The preparation of this intermediate follows the pro-
tocol given by Johnson et al. and Bonneson et al.[37,66] As no NMR or IR
data has previously been reported for these complexes they are given
here for completeness. [cis-RuCl2ACHTNUGRTNEG(UN phen)2]·2H2O (1.080 g, 1.899 mmol)
was refluxed in degassed H2O (30 mL) for 15 min, and then Na2CO3
(3.173 g, 0.0299 mol) was added. The mixture was refluxed for 2 h,
cooled in the fridge, filtered, and the dark purple microcrystalline solid
washed with H2O and ether, then dried. Yield: 0.958 g (97%). 1H NMR
(300 MHz, CD3OD): d=7.27 (dd, J=5.5, 7.9 Hz, 4H), 7.77 (d, J=5.1 Hz,
2H), 8.10 (d, J=8.7 Hz, 2H), 8.22 (m, 8H), 8.71 (d, J=8.1 Hz, 2H),
9.69 ppm (d, J=4.8 Hz, 2H); 13C NMR (75 MHz, CD3OD): d=125.5
(CH), 126.7 (CH), 128.7 (CH), 128.8 (CH), 131.8 (C), 134.0 (CH), 135.6
(CH), 150.5 (C), 152.2 (C), 153.6 (CH), 155.2 ppm (CH); IR (KBr): n
O) 1560 cmꢀ1
[cis-Ru(OH2) (0.503 g,
ACHTUNGTRENNUNG(C=
.
A
R
(phen)2]
E
(2):[37]
[Ru(CO3)
N
0.964 mmol) was suspended in degassed H2O (10 mL)in a foil-wrapped
flask. HOTf (32 mL, 1.5m in H2O) was added. Bubbles were seen, the so-
lution turned red, the solid dissolved, and then reprecipitated. The mix-
ture was stirred for 0.5 h at room temperature, then the volume was re-
duced by half. The solid was filtered, rinsed 3 times with 1 mL portions
of H2O, then 3 times with 5 mL portions of ether, and vacuum dried.
Yield: 0.632 g (82%) dark red-orange powder. 1H NMR (300 MHz,
D2O): d=7.18 (dd, J=5.4 and 8.1 Hz, 2H), 7.77 (d, J=5.1 Hz, 2H), 8.00
(d, J=9.0 Hz, 2H), 8.17 (m, 6H), 8.72 (d, J=7.8 Hz, 2H), 9.68 ppm (d,
J=4.8 Hz, 2H); 13C NMR (75 MHz, D2O): d=119.6 (q, JCF =315 Hz,
CF3SO3ꢀ), 124.2 (CH), 125.6 (CH), 127.6 (CH), 127.7 (CH), 130.3 (C),
130.5 (C), 134.9 (CH), 136.5 (CH), 149.2 (C), 151.1 (C), 152.8 (CH),
155.3 ppm (CH); elemental analysis calcd (%) for C26H22F6N4O8RuS2: C
39.15; H 2.78, N 7.03; found: C 39.30; H 2.75; N 7.08.
Experimental Section
General: All manipulations were performed under an atmosphere of
argon employing standard Schlenk-line techniques or within a dry-box
and using freshly distilled organic solvents or degassed water. Commer-
cially obtained reagents were used as received. High-pressure hydrogena-
tion reactions were conducted with 25 mL reaction solutions in an Auto-
clave Engineers (AE) Mini-Reactor with a 50 mL stainless steel (316 SS)
reactor vessel or with 2 mL reaction solutions in a 24ꢃ4 mL well HEL
CAT24 parallel reactor using standard 2 mL borosilicate glass tubes leav-
ing four wells as blanks to check for cross-contamination (none was ob-
served). All high-pressure experiments employed industrial grade hydro-
gen gas. GC analyses were carried out by using a Varian 3800 using
either a 30 m DB-1701 column, for 1,2-hexanediol and carbonyls, or a
30 m polar WAX column, for glycerol and all other substrates. Quantifi-
cation was achieved through a three-level calibration against authentic
samples, employing dimethylsulfone (DMS) as an internal standard. GC-
MS analysis were carried out by using a Varian Saturn 2000 GC/MS em-
ploying either a 30 m DB-1701 or a 30 m polar WAX column running in
default EI mode. Sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4 (TSP)
was used as a reference compound for NMR spectrum obtained in D2O.
All other NMR shifts were referenced to TMS using the residual solvent
signals. IR spectra were recorded by using either a BOMEM or a NICO-
LET 4700 FT-IR employing solution cells with 0.1 mm CaF2 or as KBr
pellets Elemental analyses were performed by M-H-W Laboratories,
[Ru
gassed H2O (25 mL). AgOTf (2.044 g, 7.955 mmol) was dissolved in de-
gassed H2O (10 mL) and added to the [RuCl3A(terpy)] suspension. The
ACHTNUGTERN(UNNG OH2)3ACHTUNRTGEGNUN(N terpy)]ACHTUNGTREN(NUGN OTf)2 (3): (1.112 g, 2.523 mmol) was suspended in de-
CTHUNGTRENNUNG
flask was covered in foil and stirred 1 h at 508C. The green solution was
cooled and filtered through a plug of Celite to remove the AgCl. Zn dust
(4.12 g, 0.063 mol) was added to the filtrate, which instantly changed to a
maroon color. The solution was stirred 15 min at room temperature, then
filtered through a plug of Celite. The H2O was removed using the
vacuum line and the solid vacuum dried overnight. Yield: 1.225 g (71%)
dark purple powder. 1H NMR (300 MHz, D2O): d=7.72 (t, J=6.3 Hz,
3H), 8.02 (t, J=7.8 Hz, 2H), 8.31 (d, J=7.8 Hz, 2H), 8.38 (d, J=7.8 Hz,
2H), 9.12 ppm (d, J=4.5 Hz, 2H); 13C NMR (75 MHz, D2O): d=119.6
(q, JCF =315 Hz, CF3SO3ꢀ), 121.8 (CH), 122.9 (CH), 127.5 (CH), 131.9
(CH), 137.9 (CH), 153.0 (CH), 160.3 (C), 163.9 ppm (C); UV/vis (H2O):
lmax(e) =533 (1350), 483 (1260), 363 nm (1430 Lmolꢀ1 cmꢀ1). This proce-
dure follows the one given by Adeyemi et al., who prepared the corre-
sponding perchlorate complex, but could not isolate this salt as a solid.[43]
As the ZnACHTUNRGTNE(UNG OTf)2 formed cannot be separated from the highly soluble
Phoenix, AZ. cis-RuCl
61% and 81% yield, respectively via ruthenium blue followed by reduc-
tion of the [cis-RuCl2A(bipy)2]Cl and [cis-RuCl2A(phen)2]Cl formed with
3A
SnCl2.[41] RuCl (terpy) was prepared as reported by Sullivan et al.[42]
RuCl3(4’-phenyl-2,2’:6’,2’’-terpy) was prepared as described by Constable
2A
2ACHTUNGTRENNUNG(bipy)2 and cis-RuCl2ACHTUTGNREN(NUNG phen)2 were prepared in
complex we were unable to obtain meaningful elemental analysis for this
material. Instead MALDI-TOF mass spectrometry analysis was per-
formed. m/z calcd for C15H17N3O3102Ru: 389.03, found 567.23 [Ru-
C
CHTUNGTRENNUNG
CHTUNGTRENNUNG
A
(terpy)]+, 351.16 [Ru (terpy)]+, 334.13
ACHTNUGETNR(NUGN OH2)2ACHUTGTNRENNUG ACHTUNGETRNN(GUN OH2)ACTHNUGTRENNGUN
and Cargill Thomson.[65] cis-[Ru(6,6’-Cl2-bipy)
CHUTGNRETNUN(G OH2)2]ACHUTNGTREN(NUNG OTf)2 (1Cl) was
AHCTUNGTRENNUNG
prepared as previously reported by Lau and co-workers[33,34] and our-
ACHTUNGTRENNUNG
selves.[31]
Chem. Eur. J. 2009, 15, 10132 – 10143
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10141

M. Schlaf et al.
(75 mL) for 16 h. The red solution was cooled and filtered. The solid was
rinsed three times with EtOH (10 mL), then ether, and vacuum-dried.
Yield: 0.980 g (98%) dark red-brown powder. This paramagnetic materi-
al was used in the next step without further purification or characteriza-
tion.
not occur. At 24 h the reactor heating was turned off and the reactor
placed in an ice bath for 30 min to condense any volatile products. The
reactor was vented, opened, and a final sample taken for GC. The re-
mainder of the reaction solution was transferred to a large vial for stor-
age. Glycerol deoxygenations were carried out in an analogous manner.
Representative procedure for a glycerol deoxygenation experiment using
the 24-well parallel reactor: The catalyst (5 mm, 0.5 mol% of substrate
concentration) was weighed into a 50.0 mL volumetric flask. The flask
was filled to the mark with glycerol stock solution (1000 mm glycerol,
200 mm dimethyl sulfone, in sulfolane) and the solution mixed well. Acid
stock solutions were prepared by adding the desired volume of HOTf
(18, 35, 53, 71 mL for 4, 8, 12, 16 equivalents with respect to catalyst) to a
10.0 mL volumetric flask and filling to the mark with the catalyst solu-
tion. Water (0.15, 0.30, 0.45, 0.60, 0.75 mL for 10, 20, 30, 40, 50% H2O)
and sulfolane (0.60, 0.45, 0.30, 0.15, 0.00 mL for 10, 20, 30, 40, 50% H2O)
were measured into 2 mL glass test tubes, followed by 0.75 mL of the
acid stock solution. Final [glycerol]=500 mm, [dimethyl sulfone]=
100 mm, [catalyst]=2.5 mm, [HOTf]=10, 20, 30, 40 mm in the sample test
tubes. Dimethyl sulfone stock solution (100 mm in sulfolane, 1.0 mL) was
added to the four blank test tubes. A 2ꢃ2 mm stir bar was added to each
test tube, but to ensure complete homogeneity each tube was thoroughly
mixed using a vortex mixer. A 0.5 mL sample was taken for initial GC
analysis from all but the blank samples. The tubes were loaded into the
parallel reactor and the reactor sealed. The reactor was evacuated for
2 min using a water aspirator, pressurized to 7.58 MPa with H2 gas, and
allowed to equilibrate for 2 min. The evacuation/pressurization cycle was
repeated twice more. The reactor was placed in a glass wool-lined alumi-
num heating block on a hotplate, magnetic stirring was set to the maxi-
mum, and the reactor was heated to the reaction temperature. Timing
started once the reactor reached the set operating temperature. At the
end of the reaction, heating was stopped and the reactor placed in an ice
bath for 30 min, followed by a dry ice/acetone bath for 5 min to condense
and freeze any volatile products. The reactor was vented, warmed to
room temperature, opened, and the remainder of the solutions in the test
tubes transferred to GC vials for analysis.
[Ru
G
E
(4):
[RuCl3(4’-phenyl-
2,2’:6’,2’’-terpy)] (0.880 g, 1.703 mmol) was suspended in degassed H2O
(30 mL). AgOTf (1.320 g, 5.14 mmol) was dissolved in degassed H2O
(15 mL) and added to the Ru suspension. The flask was covered in foil
and stirred for 1 h at 808C. The dark green solution was filtered through
a plug of Celite and Zn dust (2.7 g, 24.5 mmol) was added to the filtrate.
The instantly purple solution was stirred for 15 min and then filtered
through Celite. The H2O was removed in vacuo. Yield: 0.800 g (68%)
dark purple powder. 1H NMR (300 MHz, D2O): d=7.41 (d, J=7.2 Hz,
1H), 7.52 (t, J=7.5 Hz, 2H), 7.67 (t, J=6.6 Hz, 2H), 7.86 (d, J=7.5 Hz,
2H), 7.96 (t, J=7.7 Hz, 2H), 8.39 (d, J=8.1 Hz, 2H), 8.51 (s, 2H),
9.06 ppm (d, J=5.1 Hz, 2H); 13C NMR (75 MHz, D2O): d=118.7 (CH),
119.6 (q, JCF =315 Hz, CF3SO3ꢀ), 122.8 (CH), 126.7 (CH), 127.5 (CH),
129.3 (CH), 129.8 (CH), 135.8 (C), 137.8 (CH), 143.2 (C), 153.1 (CH),
160.0
(H2O):
lmax(e)=533 nm
(1151 Lmolꢀ1 cmꢀ1). This is analogous to the one given by Adeyemi
et al., who prepared the corresponding terpy perchlorate complex, but
also could not isolate this salt as a solid.[43] As the Zn
ACTHUNTRGNE(UGN OTf)2 formed
cannot be separated from the highly soluble complex we were unable to
obtain meaningful elemental analysis for this material. Instead MALDI-
TOF mass spectrometry analysis was performed. m/z calc’d for
C21H15N3O3102Ru: 465.42, found 719.21 [Ru(4’-phenyl-2,2’:6’,2’’-ter-
py)2Ru]+, 464.08 [Ru
(OH2)2(4’-phenyl-2,2’:6’,2’’-terpy)]+ 427.12 [Ru
terpy)]+, 411.13 [Ru(4’-phenyl-2,2’:6’,2’’-terpy)]+.
[(4’-phenyl-2,2’:6’,2’’-terpy)Ru(CO)(m-OCH3]2A(OTf)2 (5): [Ru
phenyl-2,2’:6’,2’’-terpy)](OTf)2 (25 mg) was dissolved in deuterated meth-
G
ACHTUNGTRENNUNG
G
ACHTUGNERTN(NUNG OH2)3(4’-
anol (0.5 mL) and hexanal (100 mL) was added to that solution. Three
freeze-pump-thaw cycles were performed on the NMR tube to remove
the air and prior to charging the NMR tube with Ar gas. The NMR tube
was heated for 3 days at 1558C, leading to a change in color from purple
to orange and the formation of a small quantity of black solid. After fil-
tration, crystals suitable for single-crystal X-ray analysis were obtained
from this solution by slow evaporation. Scale-up of this procedure in
normal methanol and precipitation with diethylether yields the complex
as an orange–yellow powder in 75% yield. 1H NMR (300 MHz,
CD3OD): d=7.59 (m, 3H), 7.79 (m, 2H), 8.03 (dd, J1 =7.9 Hz, J2 =
8.4 Hz, 2H), 8.27 (dt, J1 =8.0 Hz, J2 =7.75 Hz, J3 =1.74 Hz, 2H), 8.82 (s,
2H), 8.70 (d, J=8.0 Hz, 2H), 8.98 ppm (d, J=5.72 Hz, 2H); 13C NMR
(75 MHz, CD3OD): d=122.8 (CH), 126.0 (CH), 128.9 (CH), 129.6 (CH),
130.7 (CH), 132.0 (CH), 137.5 (C), 141.6 (CH), 153.7 (C), 154.9 (CH),
159.4 (C), 160.17 ppm (C); IR
n(CO)=1967 cmꢀ1; UV/vis (H2O): lmax(e)=399 nm (1191 Lmolꢀ1 cmꢀ1);
MALDI-TOF mass spectrometry: [(4’-phenyl-2,2’:6’,2’’-terpy)2Ru2A(m-
OCH3)]+ =C23H33N6O1102Ru2 =853,
found 855.22; [(4’-phenyl-
terpy)Ru(CO)
(OTf)]+ =C22H15F3N3O4102Ru1S1 =588, found 588.05; [(4’-
phenyl-terpy)Ru
(OTf)]+ =C21H15F3N3O3102Ru1S1 =560, found 560.06.
Acknowledgements
The authors thank Dr. Garry Hanan, Universitꢄ de Montreal, Montreal,
QC for helpful discussions regarding the synthesis of terpy ligands. This
work was supported by the Natural Science and Engineering Research
Council of Canada (NSERC), the Canadian Foundation for Innovation
(CFI), Natural Resources Canada, (NRCAN), the Ontario Ministry of
Agriculture, Food and Rural Affairs (OMAFRA), the BioCap Canada
Foundation and DuPont. Deeb Taher thanks the Deanship of Research,
Tafila Technical University for financial assistance for travel to Canada.
CTHUNGTRENNUNG
[1] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C. Della Pina,
Angew. Chem. Int. Ed. 2007, 46, 4434–4440.
Representative procedure for a 1,2-hexanediol deoxygenation experi-
ment: 1,2-Hexanediol stock solution (500 mm 1,2-hexanediol, 100 mm di-
methyl sulfone, in sulfolane) was dispensed into a 25.0 mL volumetric
flask, HOTf (22 mL, 4 equivalents with respect to catalyst) was added,
and the solution mixed well. The catalyst (2.5 mm, 0.5 mol% of substrate
concentration) was weighed into a small vial. The stock solution and cat-
alyst were combined in the AE minireactor vessel and stirred in the
sealed reactor for several minutes before opening the reactor, removing
0.5 mL for initial GC analysis, and resealing the reactor. The reactor was
evacuated for 2 min using a water aspirator, pressurized to 4.82 MPa with
H2 gas, and allowed to equilibrate for 2 min. The evacuation/pressuriza-
tion cycle was repeated twice more. Stirring was set at about 200 rpm,
and the reactor was heated to the reaction temperature. Samples were
taken at 1, 2, 4, and 8 h from reaching the set operating temperature
through the sample tube, which was first flushed with 0.5 mL of the reac-
tion mixture to ensure cross-contamination from an earlier sample did
[2] C. H. C. Zhou, J. N. Beltramini, Y. X. Fan, G. Q. M. Lu, Chem. Soc.
Rev. 2008, 37, 527–549.
[3] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Chim. Oggi
2008, 26, 32.
[4] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 2008, 108, 5253–5277.
[5] R. R. Soares, D. A. Simonetti, J. A. Dumesic, Angew. Chem. 2006,
118, 4086–4089; Angew. Chem. Int. Ed. 2006, 45, 3982–3985.
[6] D. G. Lahr, B. H. Shanks, Ind. Eng. Chem. Res. 2003, 42, 5467–5472.
[7] J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel, C.
Rosier, Green Chem. 2004, 6, 359–361.
[8] D. G. Lahr, B. H. Shanks, J. Catal. 2005, 232, 386–394.
[9] A. Perosa, P. Tundo, Ind. Eng. Chem. Res. 2005, 44, 8535–8537.
[10] M. A. Dasari, P. P. Kiatsimkul, W. R. Sutterlin, G. J. Suppes, Appl.
Catal. A 2005, 281, 225–231.
[11] C. W. Chin, M. A. Dasari, G. J. Suppes, W. R. Sutterlin, AIChE J.
2006, 52, 3543–3548.
10142
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 10132 – 10143

High-Temperature Stable Ruthenium Complexes
FULL PAPER
[12] Y. Kusunoki, T. Miyazawa, K. Kunimori, K. Tomishige, Catal.
Commun. 2005, 6, 645–649.
[42] B. P. Sullivan, J. M. Calvert, T. J. Meyer, Inorg. Chem. 1980, 19,
1404–1407.
[43] S. A. Adeyemi, A. Dovletoglou, A. R. Guadalupe, T. J. Meyer,
Inorg. Chem. 1992, 31, 1375–1383.
[13] T. Miyazawa, Y. Kusunoki, K. Kunimori, K. Tomishige, J. Catal.
2006, 240, 213–221.
[44] B. Cornils, W. A. Herrmann, Applied Homogeneous Catalysis with
Organometallic Compounds, VCH, Weinheim, 1996, pp. 87–88.
[45] Detailed results of this screen are given in the Supporting Informa-
tion.
[14] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A
2007, 329, 30–35.
[15] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A
2007, 318, 244–251.
[46] H. Brunner in Homogeneous Hydrogenation, Vol. 1 (Eds.: B. Cor-
nils, W. A. Herrmann), Wiley-VCH, Weinheim, 2002, pp. 195–212.
[47] See the Supporting Information for images of the NMR spectra of
all complexes described.
[16] J. Feng, H. Fu, J. Wang, R. Li, H. Chen, X. Li, Catal. Commun.
2008, 9, 1458–1464.
[17] A. Alhanash, E. F. Kozhevnikova, I. V. Kozhevnikov, Catal. Lett.
2008, 120, 307–311.
[48] CCDC 733652 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
[49] G. B. Deacon, J. M. Patrick, B. W. Skelton, N. C. Thomas, A. H.
White, Aust. J. Chem. 1984, 37, 929–945.
[50] In our hands attempts to synthesize and isolate these hydride com-
plexes were unsuccessful.
[51] R. H. Prince, K. A. Raspin, J. Chem. Soc. A 1969, 612–618.
[52] G. Domazetis, B. Tarpey, D. Dolphin, B. R. James, Chem. Commun.
1980, 939–940.
[53] Y.-Z. Chen, W. C. Chan, C. P. Lau, H. S. Chu, H. L. Lee, G. Jia, Or-
ganometallics 1997, 16, 1241–1246.
[54] J. N. Coalter, J. C. Huffman, K. G. Caulton, Organometallics 2000,
19, 3569–3578.
[55] See the Supporting Information for an image of the GC-MS trace
and spectrum obtained.
[18] M. McCoy, Chem. Eng. News 2007, 85, 7.
[19] R. R. Dykeman, K. L. Luska, M. E. Thibault, M. D. Jones, M.
Schlaf, M. Khanfar, N. J. Taylor, J. F. Britten, L. Harrington, J. Mol.
Catal. A 2007, 277, 233–251.
[20] J. F. Knifton, T. G. James, K. D. Allen, P. R. Weider, J. B. Brown,
L. H. Slaugh, US 6.903.044, 2005.
[21] T. Haas, B. Jaeger, R. Weber, S. F. Mitchell, C. F. King, Appl. Catal.
A 2005, 280, 83–88.
[22] S. Vollenweider, G. Grassi, I. Konig, Z. Puhan, J. Agric. Food Chem.
2003, 51, 3287–3293.
[23] S. Vollenweider, C. Lacroix, Appl. Microbiol. Biotechnol. 2004, 64,
16–27, and references therein.
[24] Y. Doleyres, P. Beck, S. Vollenweider, C. Lacroix, Appl. Microbiol.
Biotechnol. 2005, 68, 467–474.
[25] K. Y. Wang, M. C. Hawley, S. J. DeAthos, Ind. Eng. Chem. Res.
2003, 42, 2913–2923.
[26] T. Kurosaka, H. Maruyama, I. Naribayashi, Y. Sasaki, Catal.
Commun. 2008, 9, 1360–1363.
[27] T. Che (Celanese Corporation), US 4.642.394, 1987.
[28] G. Braca, A. M. R. Galletti, G. Sbrana, J. Organomet. Chem. 1991,
417, 41–49.
[29] M. Schlaf, J. Chem. Soc. Dalton Trans. 2006, 4645–4653.
[30] D. Pressman, H. J. Lucas, J. Am. Chem. Soc. 1942, 64, 1953–1957.
[31] Z. Xie, M. Schlaf, J. Mol. Catal. A 2005, 229, 151–158.
[32] M. Schlaf, P. Gosh, P. J. Fagan, E. Hauptman, R. M. Bullock, Angew.
Chem. 2001, 113, 4005–4008; Angew. Chem. Int. Ed. 2001, 40, 3887–
3890.
[56] The pKa of protonated NMP (aqueous scale) is approximately ꢀ0.5,
that is, substantially higher than that of H3O+ (ꢀ1.74), which due to
solvent leveling is the strongest acid present in the sulfolane/water
mixtures. pKa values from J. March, Advanced Organic Chemistry,
6th ed., Wiley-Interscience, New York, 2006.
[57] 3500 mmolLꢀ1glycerol, 0.35 mol% catalyst, 5 equiv of H2WO4 in
25 mL NMP and 80 mmolLꢀ1 DMS as an ISTD.
[58] 500 mmolLꢀ1 of glycerol, 100 mmolLꢀ1 DMS, 0.5 mol% ruthenium
catalysts and 8 equivalents of H2WO4 (with respect to ruthenium).
[59] R. M. Bullock, Angew. Chem. 2007, 119, 7504–7507; Angew. Chem.
Int. Ed. 2007, 46, 7360–7363.
[33] C. P. Lau, L. Cheng, Inorg. Chim. Acta 1992, 195, 133–134.
[34] C.-P. Lau, L. Cheng, J. Mol. Catal. 1993, 84, 39–50.
[35] R. M. Bullock, Chem. Eur. J. 2004, 10, 2366–2374.
[36] B. Durham, S. R. Wilson, D. J. Hodgson, T. J. Meyer, J. Am. Chem.
Soc. 1980, 102, 600–607.
[60] M. Schlaf, P. Ghosh, P. J. Fagan, E. Hauptman, R. M. Bullock, Adv.
Synth. Catal. 2009, 351, 789–800.
[61] H. Konno, A. Kobayashi, K. Sakamoto, F. Fagalde, N. E. Katz, H.
Saitoh, O. Ishitani, Inorg. Chim. Acta 2000, 299, 155–163.
[62] C. Creutz, M. H. Chou, J. Am. Chem. Soc. 2007, 129, 10108–10109.
[63] L. E. Manzer, US 6.593.481 B1, 2003.
[37] P. Bonneson, J. L. Walsh, W. T. Pennington, A. W. Cordes, B.
Durham, Inorg. Chem. 1983, 22, 1761–1765.
[64] L. E. Manzer, US 7.019.155, 2006.
[38] D. L. Jameson, L. E. Guise, Tetrahedron Lett. 1991, 32, 1999–2002.
[39] J. Wang, G. Hanan, Synlett 2005, 1251–1254.
[40] P. A. Lay, A. M. Sargeson, H. Taube, Inorg. Synth. 1986, 24, 291–
299.
[65] E. C. Constable, A. M. W. C. Thomson, New J. Chem. 1992, 16, 855–
867.
[66] E. C. Johnson, B. P. Sullivan, D. J. Salmon, S. A. Adeyemi, T. J.
Meyer, Inorg. Chem. 1978, 17, 2211–2215.
[41] T. Togano, N. Nagao, M. Tsuchida, H. Kumakura, K. Hisamatsu,
F. S. Howell, M. Mukaida, Inorg. Chim. Acta 1992, 195, 221–225.
Received: May 27, 2009
Published online: August 19, 2009
Chem. Eur. J. 2009, 15, 10132 – 10143
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10143