Homogeneous catalytic hydrogenation of long-chain esters by an osmium pincer complex and its potential application in the direct conversion of triglycerides into fatty alcohols

Article abstract of DOI:10.1039/c2gc15960k

The osmium hydride complexes OsH₂(CO)[NH(CH₂P ⁱPr₂)₂] (1) and OsHCl(CO)[NH(CH ₂PⁱPr₂)₂] (2) were evaluated in the catalytic hydrogenation of hexyl octanoate and cis-3-hexenyl hexanoate to alcohols as model substrates for triglycerides. Both complexes achieve full conversion of the saturated ester at 220 °C and 800 psi pressure of hydrogen gas. In the presence of unsaturated substrates, the complexes hydrogenate CC bonds, but are subsequently ineffective in the reduction of the ester moiety. However complex 1 is capable of hydrogenating fully saturated triglycerides (i.e., hardened fats as obtained by separate initial hydrogenation of seed oils using either 1 or 2 or a standard heterogeneous hydrogenation catalyst) giving cetyl and stearyl alcohols as the main products.

Full text of DOI:10.1039/c2gc15960k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2012, 14, 1178
www.rsc.org/greenchem
PAPER
Homogeneous catalytic hydrogenation of long-chain esters by an osmium
pincer complex and its potential application in the direct conversion of
triglycerides into fatty alcohols†
Alberto Acosta-Ramirez,a Marcello Bertoli,b Dmitry G. Gusevb and Marcel Schlaf*a
Received 4th August 2011, Accepted 6th February 2012
DOI: 10.1039/c2gc15960k
The osmium hydride complexes OsH2(CO)[NH(CH2PiPr2)2] (1) and OsHCl(CO)[NH(CH2PiPr2)2] (2)
were evaluated in the catalytic hydrogenation of hexyl octanoate and cis-3-hexenyl hexanoate to alcohols
as model substrates for triglycerides. Both complexes achieve full conversion of the saturated ester at
220 °C and 800 psi pressure of hydrogen gas. In the presence of unsaturated substrates, the complexes
hydrogenate CvC bonds, but are subsequently ineffective in the reduction of the ester moiety. However
complex 1 is capable of hydrogenating fully saturated triglycerides (i.e., hardened fats as obtained by
separate initial hydrogenation of seed oils using either 1 or 2 or a standard heterogeneous hydrogenation
catalyst) giving cetyl and stearyl alcohols as the main products.
catalysts for primary alcohols to esters showing good air, moist-
Introduction
ure and thermal stability.23,24
Fatty alcohols are important technical products that find exten-
sive application in surfactants.1 Industrially, they are produced
either from fossil petrochemical resources through Aufbau-Reak-
tion and/or Oxo processes or by the hydrogenolysis of fatty acid
methyl esters (FAME – effectively bio-diesel in turn obtained by
methanol transesterification of triglycerides from seed or animal
oils/fats) using a copper-chromite based heterogeneous catalyst
that operates at high hydrogen pressures (3000–4500 psi ≈
200–300 atm) and high temperatures (250–300 °C).2–6 Due to
the drastic reaction conditions and high associated costs of this
process, substantial effort has been expended to find catalysts
that can convert fatty esters to corresponding long-chain alcohols
under milder conditions, which would lead to energy savings
both in the operation and design-build of plants carrying out this
process and therefore be potentially more environmentally
friendly. One strategy to achieve this goal is the use of a homo-
Applying the Principle of Microscopic Reversibility, we now
report the use of these complexes in the hydrogenation of esters
and their potential application to the direct conversion of trigly-
ceride seed oils into fatty alcohols, which is made possible
through high thermal stability of the osmium vs. the previously
reported ruthenium-based complexes. To the best of our knowl-
edge, no osmium-based homogeneous catalysts for the hydro-
genation of esters have been reported and also to date no
homogeneous catalysts have been used for the direct conversion
of triglycerides to fatty alcohols. This may be routed in the
intrinsic higher temperature stability of osmium vs. ruthenium
based catalyst systems.
In the context of a desirable greener process for the direct
hydrogenolysis of triglycerides to fatty alcohols the use of an
osmium-based catalyst deserves comment. The extreme toxicity
of osmium in its volatile tetroxide (OsO4) form, obtained directly
from the molten metal and oxygen gas has long been known25
and the material has even been considered as a chemical weapon
by terrorists.26 This toxicity is however believed to mainly be
associated with the powerful oxidizing properties of OsO4,
which is rapidly reduced by tissue and any other biological
geneous rather than
a
heterogeneous catalyst system.7–12
Impressive progress has been made in the homogeneously cata-
lyzed hydrogenation of esters to alcohols using metal–ligand
bifunctional ruthenium-based systems that tolerate different func-
tional groups in the substrate and employ mild conditions of
hydrogen gas pressure (750 psi ≈ 50 atm) and temperature
(25–140 °C).13–22 Recently, we reported that the osmium com-
plexes 1 and 2 (Chart 1) are versatile and robust dehydrogenation
aDepartment of Chemistry, University of Guelph, Guelph, Ontario N1G
2W1, Canada
bDepartment of Chemistry, Wilfrid Laurier University, Waterloo, Ontario
MS5 3H6, Canada
†Electronic supplementary information (ESI) available: Images of 2D
NMR spectra of the reaction of complex 1 with cis-3-hexenyl hexanoate
(3 pages). See DOI: 10.1039/c2gc15960k
Chart 1
1178 | Green Chem., 2012, 14, 1178–1188
This journal is © The Royal Society of Chemistry 2012

View Article Online
environment. Environmental and toxicological data of other
more reduced forms of osmium are relatively sparse,27,28 but
both osmium dioxide and osmium metal are considered to be
inert.29 Given that the use of osmium based catalysts proposed
here would take place under reducing conditions that cannot gen-
erate OsO4, the use of this metal should therefore not be an
issue, but would of course have to be reviewed in a full life-
cycle analysis covering all relevant dimensions of an actual
process design.
substrate was used, 57% conversion was observed when 1 was
used as catalyst (entry 18, Table 1) and 64% for catalyst 2.
Increasing the hydrogen pressure from 800 to 950 psi at 220 °C
did not result in a considerable change in the conversion of the
ester to the corresponding alcohols when the reactions were
carried out in toluene. However, when neat substrate was used, a
93% conversion was achieved after 24 h when using 1 (with
90%/85% yield of C6-OH/C8-OH and 2% and 3% yield of HH
and OO, respectively, entry 20) and 70% conversion when 2 was
used (with 61%/45% yield of C6-OH/C8-OH and 2% and 14%
yield of HH and OO, respectively, entry 21) as the catalyst.
In order to more closely mimic the unsaturation found in
typical triglyceride oils as obtained from corn, soy, palm, canola,
etc., cis-3-hexenyl hexanoate was employed as the second model
substrate (Scheme 1). Under the same reaction conditions as the
optimized reaction using hexyl octanoate as a substrate (800 psi
of hydrogen gas, 24 h at 220 °C, reactor charged in air), 100%
conversion was observed when 1 was used as the catalyst,
however only hexyl hexanoate was obtained as the main product
(97%) and only a small trace of 1-hexanol was formed (<3%).
An instant color change from colourless to deep brown occurred
in the solution when the catalyst was added. The same obser-
vation was made when the reactor was charged under inert
atmosphere with the reaction giving a marginally improved result
(100% conversion, 94% yield of hexyl hexanoate, 2% yield of 1-
hexanol and 4% of 1-hexanal). Similarly when 2 was employed
as the catalyst a 100% conversion and a quantitative yield of the
saturated ester was realized. When the catalytic reactions were
analyzed at different times and the hydrogen uptake of these
reactions monitored, it was found that the alkene hydrogenation
was complete after just 3 h at 220 °C and 800 psi of H2, but in
neither case did further ester hydrogenation to the alcohols take
place. However, when after complete hydrogenation of the
alkene a second equivalent of either 1 or 2 was added to the
reaction mixture and the reactor re-pressurized and re-heated,
67% hexyl hexanoate was converted to 1-hexanol. Also, by
visual inspection no further change or deepening of colour was
observed in the reaction upon the second catalyst addition.
Results and discussion
Hydrogenation of model systems
Hexyl octanoate (HO) was chosen as a GC-analyzable model
substrate for the hydrogenation of triglycerides. Reactions were
carried out under 800 psi hydrogen pressure (cold at time = 0) at
0.1 mol% catalyst load either in toluene solvent or neat substrate.
Table 1 summarizes the results of these reactions as a function of
temperature and gives the conversion of the HO substrate and
the yields of the expected products 1-hexanol (C6-OH) and 1-
octanol (C8-OH) and those of the redistribution and transesterifi-
cation products hexyl hexanoate (HH) and octyl octanoate (OO).
No other products were observed in the GC traces of the reaction
mixtures or the gas phase head space of the reactions (by micro-
GC) and no precipitates formed in the reaction solutions, i.e., the
mass balances of the reactions are complete as stated. At T >
175 °C both complexes are active as catalysts, but the chloro-
hydrido complex 2 requires the addition of a base activator
(NaOtBu), generating through loss of HCl the amido complex
[Os(H)(CO)(PNP)] 3 in situ.23 Under hydrogen gas, 3 gives rise
to the formation of 1. When used directly 1 is more active than
2, in particular at lower temperatures. E.g., at 175 °C after 24 h
and 800 psi of hydrogen pressure and toluene as a solvent (entry
5, Table 1) 46% conversion was observed when 1 was used as
catalyst, while 15% conversion occurred for 2 under the same
conditions (entry 7, Table 1). An improvement in the activity
was observed when the temperature and/or reaction time was
increased. The best conversion was observed when the reactions
were performed at 220 °C giving 87% conversion for 1 after
24 h (entry 13, Table 1) and 68% conversion for 2 under the
same conditions (entry 15, Table 1). Small amounts of hexyl
hexanoate (HH) and octyl octanoate (OO) were detected in the
GC traces of reaction mixtures run at 220 °C. The formation of
the latter side product is easily understood as a simple transester-
ification of the 1-octanol reaction product with the starting
material, while the formation of HH must be due to the fact that
these complexes also catalyze the acceptorless dehydrogenative
coupling of alcohols, as reported before.23 Both complexes show
an induction period during the catalytic reactions, as is evident
in the hydrogen uptake plots presented in Fig. 1. The reaction is
sensitive to the presence of an inert atmosphere. When the
reactor was charged under argon atmosphere (glove-box) and
then pressurized with hydrogen gas, the activity of complex 2
improves, reaching 82% conversion compared to 67% when the
reactor was charged in air (entries 17 and 15, respectively,
Table 1). The effect is less pronounced with complex 1, giving
87% conversion when charged in air vs. 93% under argon atmos-
phere (entries 13 and 16, respectively, Table 1). When neat
In order to elucidate the cause of the observed change in
colour and apparent reactivity of the osmium complexes upon
addition of the unsaturated ester to the catalyst solution a 1H, 13
C
and 31P NMR study was carried out. One equivalent of cis-3-
hexenyl hexanoate was added to a benzene-d6 solution of 1 and
then the reaction was followed by NMR spectroscopy. Under
these conditions, the color change in the solution is not as pro-
nounced as when a large excess of unsaturated ester is present as
is in the case of the catalytic reaction (substrate : catalyst ratio =
1000 : 1). The comparison of the 31P[1H] NMR spectrum before
and after addition of the ester (Fig. 2a vs. Fig. 2b) shows the
presence of a new species represented by the small new singlet
at 57.7 ppm, along with the singlet for the catalyst 1 at δ
1
64.3 ppm. In the H and 13C[1H] NMR spectra (Fig. 3b and 4b,
respectively) initially only the resonances due to the presence of
1 and cis-3-hexenyl hexanoate were observed. The NMR tube
was then heated to 100 °C (the highest temperature technically
available to us in this study) for 4 h during which time the reac-
tion turned yellow. The 31P[1H] NMR spectrum (Fig. 2c) then
showed not only the two resonances described above, but also a
new singlet at 82.2 ppm assigned to the known yellow-orange
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1178–1188 | 1179

View Article Online
amido complex [Os(H)(CO)(PNP)] (3) on the basis of our pre-
(Fig. 3d and 4c, respectively) the disappearance of the alkene
signals (5.25 and 5.45 ppm) revealed that most of the unsaturated
ester had reacted and also a new signal assignable to an aldehyde
proton is observed at 9.54 ppm. In the 13C[1H] spectrum
(Fig. 4c) several new resonances were observed in the chemical
shift region for carbonyls (one at 194 ppm for the CO of 3)
while the peaks for the alkene carbons were no longer observed.
The peaks centered at 67.8 and 27.1 ppm represent the osmium
amido complex associated CH2 moieties. This spectroscopic
information is consistent with the reaction of the hydrogen-
loaded catalyst 1 with the unsaturated ester yielding the amido
complex 3 and the saturated ester. Under catalytic conditions
(i.e., hydrogen atmosphere) 3 can then react with hydrogen gas
directly generating 1 and closing the catalytic cycle. Scheme 2
summarizes these reactions. The nature of the species rep-
resented by the peak at 57.7 ppm in 31P[1H] NMR however
remains unknown, as is its possible role in the deactivation of
the catalyst system for ester hydrogenation once exposed to
alkene functions. The presence of small amounts of aldehyde
suggests that a small amount of ester hydrogenation does in fact
take place in a competitive reaction in which the alkene hydro-
1
vious studies.23 The corresponding H spectrum shows the pres-
ence of two new multiplets at δ 3.15 and 2.81 (Fig. 3c)
assignable to the (NCH2) protons in this complex,31 along with
the resonances of both starting materials. After heating to 100 °C
for an additional 12 h, the 31P[1H] NMR spectrum (Fig. 2d),
showed a relative increase in the amount of the amido complex 3
signal and that of the unknown species at 57.7 ppm at the
expense of 1. In the 1H and 13C[1H] (J-MOD) NMR spectra
genation is much faster than that of the ester, i.e., kalkene ≫ kester
.
This mirrors the results of the catalytic reaction of the unsatu-
rated ester discussed above in which 2% of 1-hexanol and 4% of
1-hexanal were formed. A deactivation of the catalyst system
against esters in the presence of alkenes may be routed in the
complete conversion of the catalyst into a presumably inactive
species represented by the peak at 57.7 ppm in the 31P[1H] NMR
or conceivably a secondary reaction of the amido complex 3
with the alkene function of the unsaturated ester at higher temp-
eratures. E.g., a hydroformylation of the double bond to a
branched aldehyde leading to loss of the carbonyl function from
the osmium centre could occur, which in turn may render the
Fig. 1 Hydrogen uptake for the catalytic hydrogenation of hexyl
octanoate in toluene at different temperatures using (a) 1 and (b) 2 as cat-
alyst (presented as the pressure drop in the reactor at temperature).30
Table 1 Hydrogenation of hexyl octanoate to 1-hexanol and 1-octanol using osmium complexes 1 and 2a
Entry
Catalyst
Solvent
Temp (°C)
Time (h)
Conversion (%)
C6-OH/C8-OH (%)
HHd/OOe (%)
1
2
3
4
5
6
7
8
1
1
2
1
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Neat
130
130
130
175
175
175
175
200
200
200
200
220
220
220
220
220
220
220
220
220
220
8
0
9
0
4
46
0
15
44
69
16
46
63
87
49
68
88
82
57
64
93
70
—
9/10
—
4/4
45/47
—
—
—
—
—
—
—
—
—
—
—
—
—
3/4
—
1/4
3/4
2/5
0/6
2/17
2/3
2/14
24
24
8
24
8
24
8
24
8
24
8
24
8
24
24
24
24
24
24
24
1
2b
2b
1
15/16
44/43
62/72
16/17
45/47
63/62
77/82
48/50
62/62
70/82
72/77
54/46
55/31
90/85
61/45
9
1
10
11
12
13
14
15
16
17
18
19
20
21
2b
2b
1
1
2b
2b
1c
2b,c
1
2b
1f
2b,f
Neat
Neat
Neat
a Reaction conditions: 0.1 mol% of osmium catalyst in 25 mL total volume. Total amount of substrate in the reactions in toluene solvent was 37 mmol.
= 800 psi. Conversion and yield were determined by quantitative GC against authentic samples. Cyclohexane was used as internal standard.
P
H2
b NaOtBu was used as the co-catalyst. c The reactor was charged under argon atmosphere. d Hexyl hexanoate. e Octyl octanoate. f P = 950 psi.
H2
1180 | Green Chem., 2012, 14, 1178–1188
This journal is © The Royal Society of Chemistry 2012

View Article Online
Scheme 1
Fig. 2 31P[1H] NMR spectra in benzene-d6 of: (a) complex 1. (b) Reaction of 1 and cis-3-hexenyl hexanoate at room temperature. (c) Same reaction
after 4 h at 100 °C. (d) Same reaction after another 12 h at 100 °C.
Fig. 3 1H NMR spectra in benzene-d6 of: (a) complex 1. (b) Reaction of 1 and cis-3-hexenyl hexanoate at room temperature. (c) Same reaction after
4 h at 100 °C. (d) Same reaction after 16 h at 100 °C.
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1178–1188 | 1181

View Article Online
Fig. 4 13C[1H] NMR spectra in benzene-d6 of: (a) complex 1. (b) Reaction of 1 and cis-3-hexenyl hexanoate at room temperature. (c) Same reaction
after 12 h at 100 °C.
Scheme 2
resulting species inactive for ester hydrogenations. As the study
of the system in solution at temperatures relevant to the catalytic
ester hydrogenation (T ≥ 175 °C – cf. Fig. 1 and Table 1) is at
present beyond our technical capabilities, we are however not
able to provide experimental evidence for this hypothesis.
In order to overcome the incompatibility of the osmium cata-
lysts 1 and 2 with alkenes, the conversion of unsaturated esters
to saturated alcohols was subsequently carried out in two separ-
ate steps. First the hydrogenation of the CvC bond in the 3-
hexenyl hexanoate model substrate was carried out using Pd–C
as catalyst and only the subsequent reduction of the resulting
saturated hexyl hexanoate using 1 as a catalyst. As anticipated
the hydrogenation of the alkene with Pd–C proceeds at much
milder conditions. After 60 min at 70 °C and 600 psi pressure of
H2 total conversion of the unsaturated substrate to hexyl hexano-
ate is achieved. After the removal of the heterogeneous catalyst
by filtration, catalyst 1 was added and the second hydrogenation
was performed at the previously established reaction conditions
(220 °C for 24 h and 800 psi of H2) resulting in 67% conversion
with a 64% yield of 1-hexanol and 3% yield of n-hexane.32
When the Pd catalyst was not removed from the reaction
mixture, the only product observed was n-hexane with full con-
version of the substrate.
Direct hydrogenation of triglycerides
The encouraging results described above, prompted us to carry
out a preliminary study using the same methodology in a direct
1182 | Green Chem., 2012, 14, 1178–1188
This journal is © The Royal Society of Chemistry 2012

View Article Online
conversion of triglycerides (i.e., canola, soybean, corn or palm
oil) to the corresponding fatty alcohols and glycerol (Scheme 3).
The oils consist of ∼80–90% unsaturated (mainly oleic and
linoleic acids) and 10–20% saturated (mainly palmitic and
stearic acids) acid glycerol esters.33 The experiments were
carried out analogous to the reactions performed with the cis-3-
hexenyl hexanoate model substrate. Solutions of 10 mL seed oil
(a mixture of canola/soybean oil, commercial brand, ρ ≈ 0.92 g
mL−1) in 15 mL of toluene were used to perform the exper-
iments. The alkene functionalities in the seed oil were hydrogen-
ated at 800 psi of H2 for 4.5 h at 100 °C, using Pd–C as the
catalyst. Complete conversion to the saturated form of the oil
evident from the margarine-like physical appearance of the reac-
tion mixture upon cooling to ambient temperature after the first
hydrogenation.
The heterogeneous catalyst was then removed by hot filtration
of the melted reaction mixture. The second hydrogenation was
carried out using complex 1 as the catalyst. In order to estimate
the amount of catalyst needed, it was assumed that the triglycer-
ide was composed of pure stearic acid ester (MW = 956 g
mol−1), i.e., the reaction was assumed to contain ∼9.5 mmol of
triglyceride. Using the solid residue containing the saturated tri-
glyceride and toluene the hydrogenation reaction was performed
using 0.1 mol% or 0.3 mol% of 1, 950 psi of H2 and 220 °C for
24 h consuming ∼200 or ∼300 psi of hydrogen gas, respectively.
Applying the virial theorem for a non-ideal behaviour of hydro-
gen gas, neglecting the vapour pressures of the solvent, substrate
and products and scaling to the 6 mol equivalents of hydrogen
1
was confirmed by H and 13C[1H] NMR spectroscopy (Fig. 5
and 6a, respectively), in which the alkene proton peaks at
5.9 ppm are no longer present after the hydrogenation and the
shifting of the characteristic glycerol backbone proton signals
(6.8–4.4 ppm) to higher fields. The saturation of the oil was also
required for
a full conversion of the substrate to the
Scheme 3
Fig. 5 1H NMR (CDCl3) spectra of (a) a toluene solution of unsaturated seed oil and (b) a toluene solution of seed oil after the hydrogenation of the
alkenes.
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1178–1188 | 1183

View Article Online
Fig. 6 13C[1H] NMR spectra of toluene solutions of (a) the fully saturated seed oil (CDCl3) and (b) the final product of the ester hydrogenation reac-
tion catalyzed by 1 (DMSO-d6).
corresponding fatty alcohols the observed pressure drops reflect
conversions of ∼60% for the 200 psi pressure drop and ∼90%
for the 300 psi pressure drop.34 The 1H NMR spectrum
(recorded in DMSO-d6 at 70 °C) of the mixture after the hydro-
genation reaction (Fig. 7a) no longer showed the well defined
signals for the glycerol protons of the saturated triglyceride at
4.4, 4.6 (CH2–O) and 5.6 (CH–O) ppm (Fig. 5b), but instead a
group of new signals between 3 and 4.3 ppm interpreted as
originating from a complex blend of mono- and diglycerides,
with the larger broad multiplet centered at ∼3.45 ppm assigned
to the presence of CH2–O moieties of the fatty alcohols formed.
They correspond to a group of new peaks between 63 and
73 ppm also assigned to the fatty alcohols in the 13C[1H] NMR
spectrum (J-MOD) of the reaction mixture (Fig. 6b),35 which
also shows small amounts of remaining ester carbonyl groups
due to the presence of the mono- and diglycerides. The GC-MS
traces of a toluene extract of the reaction mixture showed cetyl
and stearyl alcohols as the main volatile products verified against
the traces and fragmentation patterns of authentic samples
(Fig. 8).
1184 | Green Chem., 2012, 14, 1178–1188
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 7 1H NMR spectra (70° C, DMSO-d6) of toluene solution reaction mixtures of the ester hydrogenations using (a) 0.1 mol% of 1 catalytic hydro-
genation reaction and (b) 0.3 mol% of 1.
carbon double bonds present in the triglyceride using a standard
alkene hydrogenation catalyst, complex 1 is capable of directly
converting seed oils (canola/soybean) into fatty alcohols at temp-
eratures and pressures significantly lower than those required for
the heterogeneous catalysts typically employed, but at higher
temperature than those tolerated by the previously described
ruthenium-based homogeneous ester hydrogenation catalysts, for
which however no triglyceride hydrogenation/hydrogenolysis
has been demonstrated to date.13,14 Whether or not the osmium-
based systems will be viable for an actual technical application
Conclusion
The osmium complexes OsH2(CO)[NH(CH2PiPr2)2] (1) and
OsHCl(CO)[NH(CH2PiPr2)2] (2) are active catalysts in the
hydrogenation of esters with the latter requiring the addition of
base as an activator. Both complexes are also active in the hydro-
genation of alkenes, although, when they are used as catalyst in
the reduction of unsaturated esters once the alkene has been
hydrogenated, no subsequent reduction of the carbonyl moiety
was observed. With a separate initial saturation of any carbon–
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1178–1188 | 1185

View Article Online
Fig. 8 GC-MS traces of the reaction mixture of ester hydrogenation reaction using (a) 0.1 mol% of 1 and (b) 0.3 mol% of 1 as the catalyst.
will in large part depend on their recyclability, which was not
investigated in this first study. The high temperature stability of
the catalysts suggests that this may possibly be achieved by
removal of the products through vacuum distillation.
number.36 GC-MS analyses of the reaction solutions were con-
ducted using a Varian 3800 GC/Saturn 2000 MS using column
and GC methods as for the GC analyses. Control reactions
carried out in toluene solutions in the absence of catalysts 1 and
2 gave no hydrogenated products.
Experimental
Materials
General
Toluene and cyclohexane (dry solvent, Fisherbrand) was
degassed and dried using 3 Å molecular sieves in an LC Tech-
nology Solutions Solvent Purification System. CDCl3, C6D6, and
DMSO-d6 were purchased from Cambridge Isotope Labs. Hexyl
hexanoate was purchased from SAFC Supply Solutions and cis-
3-hexenyl hexanoate was purchased from TCI America. Both
esters were passed through a neutral alumina column and
degassed through a series of freeze–pump–thaw cycles. Seed oil
(a mixture of canola/soybean oil according to the label of a com-
mercial brand obtained from a local supermarket) was used as
received. OsH2(CO)[NH(CH2PiPr2)2] (1) and OsHCl(CO)[NH
(CH2PiPr2)2] (2) were synthesized according to the literature.23,24
Pd–C (5% Pd) was purchased from Aldrich.
NMR spectra were recorded on a 400 MHz instrument. 1H NMR
chemical shifts are reported in ppm versus residual protons in
deuterated solvents as follows: 7.27 ppm CDCl3, 7.16 ppm
C6D6, 2.49 ppm DMSO-d6. 13C[1H] NMR chemical shifts are
reported in ppm versus residual 13C in the solvent: 77.0 ppm
CDCl3, 128.4 ppm C6D6, 39.5 ppm DMSO-d6. 31P NMR shift
are reported versus 85% H3PO4 as an external standard. Indus-
trial grade hydrogen gas was used for all hydrogenation exper-
iments. An Autoclave Engineers (AE) Minireactor with a 50 mL
stainless steel (316SS) reactor vessel was employed for the all
hydrogenation reactions. Where indicated the reactor body was
charged inside an inert-gas dry-box (Argon) and sealed before
exposure to the atmosphere. GC analyses were performed on a
Varian 3800 using a 30 m/0.25 mm/0.025 mm medium polarity
DB-1701 column. Quantitative analysis of the 1-hexanol, 1-
octanol, hexyl octanoate and cis-3-hexenyl hexanoate exper-
iments was performed through a three-level calibration against
authentic samples using cyclohexane as the internal standard
(100 mmol L−1). Quantification for all other substrates was
approximated using the GC-FID response factor for 1-hexanol or
1-octanol, respectively, corrected for the effective carbon
Representative procedure for the hydrogenation of esters
In the glove box a 25.0 mL volumetric flask was charged with
hexyl octanoate (8.6 g, 37 mmol), cyclohexane (210 mg,
2.5 mmol), and the catalyst (0.037 mmol), using toluene as a
solvent. When 2 was used as the catalyst NaOtBu (9 mg,
0.09 mmol) was also added. Then, the solution was transferred
1186 | Green Chem., 2012, 14, 1178–1188
This journal is © The Royal Society of Chemistry 2012

View Article Online
to a vial, sealed and removed from the glove box. The reactor
vessel was charged with this toluene solution. The reactor was
then flushed out 3 times with H2 gas (approx. 600 psi) and then
pressurized to the working pressure with H2 gas, and allowed to
equilibrate for 2 min. 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 tempera-
ture via the sample tube, which was first flushed with 0.5 mL of
the reaction mixture to ensure cross-contamination from an
earlier sample did not occur. After 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 analysis.
References
1 K. Noweck and W. Grafahrend, in Ullmann’s Encyclopedia of Industrial
Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2000.
2 A. Behr, A. Westfechtel and J. P. Gomes, Chem. Eng. Technol., 2008, 31,
700–714.
3 U. R. Kreutzer, J. Am. Oil Chem. Soc., 1984, 61, 343–348.
4 H. Buchold, Chem. Eng., 1983, 90, 42–43.
5 T. Voeste and H. Buchold, J. Am. Oil Chem. Soc., 1984, 61, 350–352.
6 T. Turek and D. L. Trimm, Catal. Rev. Sci. Eng., 1994, 36, 645–683.
7 R. A. Grey, G. P. Pez and A. Wallo, J. Am. Chem. Soc., 1981, 103, 7536–
7542.
8 U. Matteoli, G. Menchi, M. Bianchi and F. Piacenti, J. Mol. Catal., 1988,
44, 347–355.
9 H. T. Teunissen and C. J. Elsevier, Chem. Commun., 1998, 1367–1368.
10 K. Nomura, H. Ogura and Y. Imanishi, J. Mol. Catal. A: Chem., 2002,
178, 105–114.
11 M. C. van Engelen, H. T. Teunissen, J. G. de Vries and C. J. Elsevier, J.
Mol. Catal. A: Chem., 2003, 206, 185–192.
12 B. Boardman, M. J. Hanton, H. van Rensburg and R. P. Tooze, Chem.
Commun., 2006, 2289–2291.
13 J. Zhang, G. Leitus, Y. Ben-David and D. Milstein, Angew. Chem., Int.
Ed., 2006, 45, 1113–1115.
Representative procedure for hydrogenation of seed oil
(canola/soybean oil mixture)
14 L. A. Saudan, C. M. Saudan, C. Debieux and P. Wyss, Angew. Chem.,
Int. Ed., 2007, 46, 7473–7476.
15 L. Saudan, P. Dupau, J.-J. Riedhauser, P. Wyss and S. A. Firmenich,
Switz., WO2006106483A1, 2006.
16 L. Saudan, P. Dupau, J.-J. Riedhauser, P. Wyss and S. A. Firmenich,
Switz., WO2006106484A1, 2006.
17 L. Saudan, C. Saudan and S. A. Firmenich, Switz., WO2008065588A1,
2008.
18 C. Saudan, L. Saudan and S. A. Firmenich, Switz., WO2010038209A1,
2010.
19 C. Saudan, L. Saudan and S. A. Firmenich, Switz., WO2010061350A1,
2010.
20 L. Saudan, P. Dupau, J.-J. Riedhauser, P. Wyss and S. A. Firmenich,
Switz., US20100280273A1, 2010.
21 M. Ito, T. Ootsuka, R. Watari, A. Shiibashi, A. Himizu and T. Ikariya, J.
Am. Chem. Soc., 2011, 133, 4240–4242.
22 W. N. O. Wylie, A. J. Lough and R. H. Morris, Chem. Commun., 2010,
46, 8240–8242.
23 M. Bertoli, A. Choualeb, A. J. Lough, B. Moore, D. Spasyuk and D.
G. Gusev, Organometallics, 2011, 3479–3482.
24 M. Bertoli, A. Choualeb, D. G. Gusev, A. J. Lough, Q. Major and
B. Moore, Dalton Trans., 2011, 40, 8941–8949.
25 F. R. Brunot, J. Ind. Hyg., 1933, 15, 136–143.
26 I. Makarovsky, G. Markel, A. Hoffman, O. Schein, A. Finkelstien,
T. Brosh-Nissimov, Z. Tashma, T. Dushnitsky and A. Eisenkraft, Isr. Med.
Assoc. J., 2007, 9, 750–752.
27 C. Chen, P. N. Sedwick and M. Sharma, Proc. Natl. Acad. Sci. U. S. A.,
2009, 106, 7724–7728.
28 S. Rauch, B. Peucker-Ehrenbrink, M. E. Kylander, D. J. Weiss,
A. Martinez-Cortizas, D. Heslop, C. Olid, T. M. Mighall and H.
F. Hemond, Environ. Sci. Technol., 2010, 44, 881–887.
29 I. C. Smith, B. L. Carson and T. L. Ferguson, Environ. Health Perspect.,
1974, 8, 201–213.
In the glove box a 25.0 mL volumetric flask was charged with
10 mL of seed oil, using toluene as a solvent. Then, the solution
was transferred to a vial, sealed and removed from the glove
box. The mini-reactor vessel was charged with this toluene sol-
ution and 200 mg of Pd–C. was added. The reactor was then
pressurized to 800 psi with H2 gas, and allowed to equilibrate
for 2 min. Stirring was set at about 200 rpm and the reactor was
heated to 100 °C for 4 h. After 24 h the reactor heating was
turned off and the reactor placed in an ice bath for 30 min. The
reactor was then vented, opened, and a sample of the waxy solid
was taken for NMR analysis. The Pd heterogeneous catalyst was
removed by hot filtration using celite and then the osmium cata-
lyst 1 (6 or 18 mg = 0.1 or 0.3 mol%), using toluene as a solvent
(10 mL) was added to the white solid mixture reaction. The
reactor was then flushed out 3 times with H2 gas (approx.
600 psi) and then pressurized to 950 psi with H2 gas, and
allowed to equilibrate for 2 min. Stirring was set at about 200
rpm and the reactor was heated at 220 °C. After 24 h the reactor
heating was turned off and the reactor placed in an ice bath for
30 min. The reactor was vented, opened, and a final sample
taken for GC and NMR analysis. The presence of fatty alcohols
was confirmed by comparison with retention time and mass
spectrum of authentic samples.
30 The fact that the hydrogen uptake data presented here do not directly cor-
respond to the stoichiometries of the quantitative yields as determined by
GC (see Table 1) reflects the strongly non-linear response of the non-
ideal hydrogen/toluene gas phase composition in the reactor headspace as
a function of total pressure at the reaction temperature, which lies above
the boiling point of toluene. E.g., a 6-fold increase in conversion as a
function of time (entries 4 and 5, Table 1) does not translate to the same
amount of pressure drop in the reactor, which shows an only 3-fold
increase in apparent uptake. This is because the mole fraction of hydro-
gen in the gas phase decreases exponentially with decreasing pressure.
For a quantitative treatment of this phenomenon see J. J. Simmick, H.
M. Sebastian, H.-M. Lin and K.-C. Chao, J. Chem. Eng. Data, 1978, 4,
339–340.
Reactivity of 1 with cis-3-hexenyl hexanoate
In the glove box 25 mg of 1 (0.047 mmol) were dissolved in
C6D6 in a Teflon sealed NMR tube. Then, cis-3-hexenyl hexano-
ate was added (9 mg, 0.047 mmol). The solution turns slightly
brown and NMR spectra were recorded. Then, the reaction was
heated to 100 °C using an oil bath. The reaction was monitored
after 4 and 16 h of heating by NMR.
31 The hydride signal of the amido complex 3 occurs at −21.4 ppm and is
not shown in Fig. 3 for reasons of scale. Images of 13C JMOD and 2D
1H NMR spectra used to make the peak assignments are given in the
ESI.†.
32 When the removal of the heterogeneous catalyst was performed under Ar
atmosphere, the conversion of hexyl hexanoate reached 71%.
Acknowledgements
The authors thank the Natural Science and Engineering Research
Council (NSERC) of Canada and the Ontario Ministry of Agri-
culture, Food and Rural Affairs (OMAFRA) for funding.
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2012, 14, 1178–1188 | 1187

View Article Online
33 A. Thomas, in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-
VCH Verlag GmbH & Co. KGaA, 2000.
contrast to the data presented in Fig. 1, in this case the pressure drops do
directly reflect the stoichiometric hydrogen uptake.
34 Pressure drops refer to the initial and final pressure in the reactor at room
temperature before and after the reaction, i.e., at temperatures well below
the boiling points of the liquid components of the reaction mixture. In
35 M. Jie, C. C. Lam, M. K. Pasha, K. L. Stefenov and I. Marekov, J. Am.
Oil Chem. Soc., 1996, 73, 1011–1017.
36 J. T. Scanlon and D. E. Willis, J. Chromatogr. Sci., 1985, 23, 333–340.
1188 | Green Chem., 2012, 14, 1178–1188
This journal is © The Royal Society of Chemistry 2012