Rhenium-catalyzed transfer hydrogenation and deoxygenation of biomass-derived polyols to small and useful organics

Article abstract of DOI:10.1002/cssc.201200138

Hello MTO: Efficient and environmentally responsible conversion methods are needed to deal with excess glycerol production. The catalytic conversion of glycerol to allyl alcohol under neat conditions is catalyzed by methyltrioxorhenium (MTO). The yields o

Full text of DOI:10.1002/cssc.201200138

DOI: 10.1002/cssc.201200138
Rhenium-Catalyzed Transfer Hydrogenation and Deoxygenation of
Biomass-Derived Polyols to Small and Useful Organics
[
a]
Jing Yi, Shuo Liu, and Mahdi M. Abu-Omar*
The efficient utilization of biomass resources poses a grand
chloride. Relevant to this report, allyl alcohol can be prepared
[1]
[5]
challenge that is receiving extensive attention worldwide.
from the reaction of glycerol and formic acid. The reaction of
Glycerol, a byproduct of biodiesel production, has been in
oversupply since 1995, and its production rate continues to in-
polyols with formic acid has been advanced recently by Berg-
[6]
man and Ellman. They have also reported on the use of rhe-
nium carbonyl complexes to deoxygenate polyols with 3-octa-
[
2]
crease. Hence, efficient and environmentally responsible
methods for glycerol conversion are needed.
[7]
nol as the reductant in the presence of an acid additive. The
catalytic mechanism with rhenium carbonyl is undefined. Nich-
olas and co-workers have employed oxorhenium catalysts in
the deoxydehydration of vicinal diols with sulfite as the
In this Communication, we describe a catalytic method that
uses neat glycerol to make small and useful organics (SUO).
[
3]
The catalyst, methyltrioxorhenium (MTO), is tolerant to water
and air. The substrate, glycerol, participates in transfer hydro-
genation and deoxygenation to give volatile products, allyl al-
cohol, propanal, and acrolein, leaving the nonvolatile dihydrox-
yacetone (DHA) byproduct in the residue (Scheme 1). This reac-
[8]
oxygen atom acceptor. Our group has previously reported
[9]
H -driven deoxygenation of vicinal diols. Schlaf and Bullock
2
have pioneered the use of organometallic ruthenium catalysts
[10]
for the deoxygenation of alcohol functional groups.
Iron
oxide has been used in the de-
oxygenation of glycerol via de-
hydration and consecutive hy-
[11]
drogen transfer. There is also
interest in the selective oxidation
of glycerol to DHA. The Way-
mouth group has recently re-
ported on a palladium-catalyzed
oxidation of glycerol under am-
[12]
bient temperature.
In our research, MTO at the
low loading of 2 mol% relative
to glycerol was effective at
1658C in converting glycerol to
volatile products with allyl alco-
hol as the major product. The re-
action proceeds with impressive
efficiency (74% yield of volatiles
in 1 h), and ease of separation as
the by-product DHA is nonvola-
tile and remains in the reaction
Scheme 1. Rhenium-catalyzed deoxygenation reactions of glycerol, erythritol, and threitol to produce SUO.
tion can also be carried out in the presence of an alcohol as
a solvent and transfer hydrogenation agent. Other biomass-de-
rived polyols (erythritol and threitol) can be used to make
small organic molecules. Allyl alcohol and 2,5-dihydrofuran are
valuable building-block chemicals that are widely used in the
pharmaceutical and polymer industries.
flask. The volatile products are simply removed as they form
via direct distillation. The distilled fraction is composed of allyl
alcohol, propanal, and acrolein in the ratio of 1.0:0.22:0.15 (full
1
13
characterization by H, C NMR, and GC-MS). The reaction does
not require a solvent and is amenable to scaling up as the
products are removed readily at the reaction temperature. In
addition to MTO, NaReO , KReO , NH ReO , and (NH ) MoO
4
Several methods for the preparation of allyl alcohol have
4
4
4
4
4 2
[
4]
been documented. The traditional route is hydrolysis of allyl
were tested as catalysts under the same conditions (Table 1).
While KReO4 and NaReO4 gave low conversions and yields,
NH ReO4 gave comparable results to MTO when used at
[a] J. Yi, S. Liu, Prof. Dr. M. M. Abu-Omar
4
Department of Chemistry
Purdue University
West Lafayette, IN 47906 (USA)
Fax: (+1)765-494-0239
E-mail: mabuomar@purdue.edu
5 mol% loading. However, the ratio of the volatile products is
somewhat different from that observed with MTO, but still
with allyl alcohol as the major product. Molybdate did not cat-
alyze glycerol conversion even after 18 h at 1658C.
The marked difference between NH ReO and NaReO merit-
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201200138.
4
4
4
ed further investigation to understand the role of the cation.
ChemSusChem 0000, 00, 1 – 4
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

clohexanol, 1,3-propanediol, 1,2-propanediol) as solvents and
potential hydrogen transfer agents in our reaction. The yields
of volatile products from glycerol with MTO as the catalyst
after 1 h varied but in a narrow range of 50-55% (Supporting
Information). The exception was 1,3-propanediol, which gave
lower conversion and yield. 3-Octanol with NH ReO as a cata-
Table 1. Rhenium and molybdenum oxo catalysts for glycerol deoxyge-
nation.
[
a]
Catalyst
Amount
t
Conv. Yield of vola-
Product ratio (allylic
tile products alcohol/acrolein/-
%] propanal)
[
b]
[
mol%] [h] [%]
[
4
4
MTO
2
5
5
5
5
1
100
9
27
100
0
74
6
24
80
0
1:0.22:0.15
lyst afforded 74% of volatile products with the ratio allyl alco-
hol/acrolein/propanal=1.0:0.02:0.01. The reactions with alco-
hol solvents yielded ketone or aldehyde amounting to less
than would be expected from the amount of allyl alcohol pro-
duced. Therefore, even in 3-octanol as a solvent, transfer hy-
drogenation from glycerol was competitive with that from 3-
octanol (the sacrificial alcohol). In general, the use of alcohols
improved the ratio of allyl alcohol to the other two volatile
products: acrolein and propanal. Notably, a significant amount
of alkene from the sacrificial alcohol was also produced. This is
attributed to oxorhenium catalyzing alcohol dehydration under
NaReO
KReO
NH ReO
MoO
4
18
10
1
1:0.32:0.07
1:0.35:0.06
1:0.2:0.05
N/A
[
c]
4
4
4
(
[
NH
4
)
2
4
18
a] Neat at 1658C. [b] The maximum yield of volatile products is one-half
the amount of initial glycerol, since half of the glycerol serves as reduc-
tant giving 1,3-dihydroxyacetone (DHA). [c] Addition of pyridine does not
improve the yields.
[13]
Table 2. Effect of additives on the oxorhenium-catalyzed deoxygenation
our reaction conditions.
[
a]
of glycerol.
MTO-catalyzed transfer hydrogenation and deoxygenation
was investigated for other biomass-derived polyols, meso-er-
ythritol and meso-threitol (Scheme 1). The reaction was found
to be stereospecific. Erythritol gives a reasonably high yield of
dihydrofuran (58%), which is higher than previously reported
[
b]
Catalyst
MTO
Additive
t
Conv. Yield of vola- Product ratio (allylic
[h] [%] tile products alcohol/acrolein/-
[%]
propanal)
NaCl
KCl
HCl
2
1.5
1
1
1
1
2
1.5
100
N/A
100
100
100
100
N/A
100
74
61
78
78
96
93
42
91
1:0.1:0
[6a]
1:0.12:0
1:0.3:0.15
1:0.2:0.07
1:0.27:0
1:0.6:0
for the reaction of meso-erythritol with formic acid (39%).
meso-Threitol, on the other hand, gives low yields of deoxy-
genated volatile product and is converted to (3S,4S)-tetrahy-
drofuran-3,4-diol. Similarly, cis-1,2-cyclohexanediol is converted
under solvent free conditions with MTO to cyclohexene (91%
yield) while trans-1,2-cyclohexanediol did not react.
NH
NH
HCl
NaCl
KCl
4
Cl
[c]
NaReO
4
4
[
Cl
d]
[
e]
NH ReO
4
4
1:0.1:0
1:0.18:0.03
[a] Neat at 1658C. [b] 2 mol% catalyst with equimolar (2 mol%) additive
unless noted otherwise. [c] 1.5 mol% NH
NaCl.
4
Cl. [d] 1 mol% HCl. [e] 1.5 mol%
A number of experiments with different catalysts and additives
were investigated, and the results are summarized in Table 2.
For MTO, addition of sodium or chloride has no effect on the
amount of volatile products produced but appears to reduce
the ratio of propanal in the product mixture. Potassium, on the
other hand, appears to reduce the productivity of the catalyst
somewhat. The addition of ammonium salt or HCl exerts mod-
erate enhancement in the rate and somewhat reduced propa-
nal production. This observation is consistent with the ammo-
nium ion acting as an acid (proton) source at high tempera-
Glycerol-(OD) and d -glycerol-(OH) were used and their re-
3
5
3
action rates were compared to H -glycerol-(OH) to gain insight
5
3
into the mechanism and the rate-determining step. In these
studies, MTO was employed as the catalyst. The times it took
the reaction to reach completion for each of the labeled glyc-
erol substrates were compared to obtain kinetic isotope effects
(KIEs). Deuterium labeling on the alcohol groups in glycerol
showed no KIE while d -glycerol-(OH) took 3.5 h to reach com-
5
3
ture. Indeed, when NH Cl or HCl was used as an additive with
pletion (100% conversion), corresponding to a KIE of ca. 2.4.
4
NaReO it behaved similarly to NH ReO : high yields of volatile
Furthermore, the product distributions for glycerol-(OD) and
4
4
4
3
products with selectivity for allyl alcohol. The NH Cl additive
d -glycerol-(OH) were comparable giving allyl alcohol/acrolein/
4
5
3
exhibited higher selectivity than HCl. Interestingly, addition of
NaCl to NH ReO reduced the catalyst’s productivity while KCl
propanal=1:0.26:0.07. In both deuterium-labeled substrates,
the amount of propanal was less than that observed with H5-
4
4
showed no adverse effect. Although our additive studies re-
sults cannot be fully rationalized at this time, they point to the
need for proton or an activating cation to be associated with
the perrhenate anion to make it catalytically viable. Similar ob-
servations with different cations have been noted recently for
glycerol-(OH) (cf. first entry in Table 1). We cannot offer a ra-
3
tionalization at this point for this difference in the products
distribution. Nevertheless, the KIE observed for d -glycerol-
5
(OH) is reproducible and is decisively more than experimental
3
error constituting a primary KIE. In other words, the CꢀH/D
bond of glycerol is involved in the rate-determining step.
Another important mechanistic question is the pathway by
which acrolein and propanal are produced. It is feasible that
[
8]
the sulfite driven deoxygenation of glycols.
[7]
Since 3-octanol has been used as a reducing agent, we
tested its use and other high boiling alcohols (1-heptanol, 1-cy-
&2
&
www.chemsuschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 4
ÝÝ
These are not the final page numbers!

order dependence on [glycerol] at lower concentrations, such
experiments have been difficult as the reaction is significantly
retarded in inert solvents. Notwithstanding these obstacles,
the first-order kinetic profiles for allyl alcohol formation and
the KIE are consistent with the formation of a Re diolate (Ia)
followed by transfer hydrogenation as detailed in Scheme 2.
Intermediate Ia acts as a bifunctional catalyst with the oxo
ligand as a hydride acceptor and the alkoxide as a proton ac-
ceptor.
k1
MTO þ glycerol G H ½Iaꢂ
ð1Þ
ð2Þ
kꢀ1
k2
½
I ꢂ þ glycerol ƒ!allyl alhohol þ MTO
a
2
d½allylalcoholꢂ
k k ½Reꢂ ½glycerolꢂ
1
2
T
¼
ffi
dt
k
ꢀ
þ k ½glycerolꢂ þ k ½glycerolꢂ
1
1
2
ð3Þ
Figure 1. Reaction profile for glycerol conversion to allylic alcohol, acrolein,
and propanal. Conditions: 2 mol% MTO at 1658C.
k k ½Reꢂ ½glycerolꢂ
1
2
T
k þ k
1
2
[
15]
[8]
acrolein results from the oxidation of allyl alcohol. Three ex-
periments were conducted to address these questions. When
the reaction was run under Ar instead of open air, the amount
of acrolein and propanal did not change. A control reaction of
allyl alcohol and MTO did not
Both Espenson and Nicholas have observed rhenium di-
olate with glycols. Also in our hands the reaction of MTO with
glycerol after heating for a brief 50 s followed by rapid cooling
and addition of CDCl or [D ]DMSO shows rhenium diolate for-
3
6
produce acrolein. The reaction
with glycerol was followed by
quantifying yields at different
time intervals. A typical reaction
profile is shown in Figure 1.
It is evident that acrolein and
propanal are formed in parallel
to allyl alcohol. The rate of for-
mation of allyl alcohol follows
first-order kinetics and the ob-
served rate constant depends on
MTO (at 215ꢁ58C, ky =2.3ꢁ
ꢀ
3
ꢀ1
1
3
0
s
for 1 mol% MTO and
ꢀ
3
ꢀ1
.5ꢁ10
s
for 2 mol% MTO).
the experimen-
Therefore,
tally determined rate law is
d[allyl alcohol]/dt=ꢀd[glycerol]/
dt=k[glycerol][MTO] . Based on
T
previous studies by the groups
[
14]
[15]
of Andrews, Espenson,
and
[
8]
Nicholas,
and our results
herein, we propose allyl alcohol
formation from rhenium diolate
via hydride transfer from
a second molecule of glycerol
according to the kinetic Scheme
described by Equations (1) and
(2). The corresponding stea-
dy-state rate law is given in
Equation (3). In the limit of
(
k +k )[glycerol]@k , the rate
1 2 ꢀ1
simplifies to that observed ex-
perimentally. While it would be
Scheme 2. Proposed mechanism for the formation of allyl alcohol (major product) and minor products (propanal
useful to demonstrate second- and acrolein) from the conversion of glycerol by MTO.
ChemSusChem 0000, 00, 1 – 4
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
3
&
ÞÞ
These are not the final page numbers!

1
mation by H NMR (Supporting Information). Furthermore,
ture in the reaction mixture was measured by an immersed ther-
mometer. The system was heated using a preheated oil bath set at
heating the diolate in [D ]DMSO to elevated temperature pro-
6
1
1
758C. However, the reaction solution temperature reached only
658C. 10 min later, the first drop of volatiles was collected. After
duces allylic alcohol. Another point of interest is to investigate
the effect of water on the reaction as it is a by-product of di-
olate formation. When the glycerol reaction was conducted in
the presence of 4 ꢂ molecular sieves, the volatile products
yield was increased to 99% with enhanced selectivity for allylic
alcohol (allylic alcohol/acrolein/propanal=1:0.06:0). The reac-
tion time, however, in the presence of molecular sieves was
longer (2.5 versus 1 h to reach 100% conversion). All of these
observations are consistent with the proposed reaction
mechanism.
ca. 1 h the production of volatiles ceased as no more condensates
were visible in the distillation column, and the reaction was as-
sumed to be complete. The volatile products (2.82 g, 74%) were
collected and analyzed without further purification.
Acknowledgements
Funding for this research was provided by DOE-BES, grant DE-
FG-02-06ER15794.The authors thank Dr. Bruce R. Cooper at the
Purdue University Metabolite Profiling Facility for assistance with
mass spectrometry analysis, and Dr. Kothanda R. Pichaandi for
useful discussion.
Bullock showed that 1,2-propanediol can produce propa-
[
16]
nal.
We hypothesize that propanal and acrolein are pro-
duced from side reactions involving intermediate I to give 1,2-
a
propanediol and 1,3-propanediol. Indeed, control experiments
with these diols afforded 2-ethyl-4-methyl-1,3-dioxolane (re-
sulting from the condensation of propanal and 1,2-propane-
diol) and 2-vinyl-1,3-dioxane (resulting from the condensation
of acrolein and 1,3-propanediol). In the case of the glycerol re-
action 1,2- and 1,3-propanediol do not accumulate to give the
condensation products. Furthermore, any small condensation
products from aldehyde and glycerol remain in the residue as
non-volatiles.
Keywords: biomass
hydrogen transfer · methyltrioxorhenium
·
catalysis
·
hydrodeoxygenation
·
[1] a) M. Paster, J. L. Pellegrino, T. M. Carole, Industrial Bioproducts: Today
and Tomorrow, 2003. Available at http://www1.eere.energy.gov/bio-
mass/products.html (accessed May 2012); b) Biomass Energy Center,
available at at http://ww.biomassenergycentre.org.uk (accessed May
2012); c) D. R. Dodds, R. A. Gross, Science 2007, 318, 1250–1251; d) R. C.
Brown, Biorenewable Resources: Engineering New Products from Agricul-
ture, Iowa State Press, Ames, 2003, p. 240.
In summary, MTO and NH ReO catalyze the conversion of
4
4
glycerol to allyl alcohol under neat conditions at 1658C. Propa-
nal and acrolein are the major volatile products. Dihydroxyace-
tone is the oxidation product and it remains in the reaction
flask because of its low vapor pressure. Kinetic isotope effect
in conjunction with control experiments provides insights with
regard to the reaction mechanism with MTO. Allyl alcohol is
formed via transfer hydrogenation from glycerol to a rhenium-
diolate complex followed by rapid alkene extrusion from the
resulting rhenium(V) diolate, a reaction with ample literature
[
2] a) E. K. Wilson, Chem. Eng. News 2002, 80, 46–49; b) R. Christoph, B.
Schmidt, U. Steinberner, W. Dilla, R. Karinen, Glycerol, in Ullmann’s Ency-
clopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2006, pp. 1–16.
[3] For a recent Review on MTO, see: F. E. Kꢃhn, A. Scherbaum, W. A. Herr-
mann, J. Organomet. Chem. 2004, 689, 4149–4164.
[
4] L. Krꢄhling, J. Krey, G. Jakobson, J. Grolig, L. Miksche, Allyl Compounds,
in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim,
2002, pp. 447–465.
[5] O. Kamm, C. S. Marvel, Org. Synth. 1941, Coll. Vol. 1, 42–44; 1921, 1, 15.
6] a) E. Arceo, P. Marsden, R. G. Bergman, J. A. Ellman, Chem. Commun.
009, 3357–3359; b) E. Arceo, J. A. Ellman, R. G. Bergman, ChemSu-
sChem 2010, 3, 811–813.
[
2
[
14,15]
precedence.
perrhenate salts point to the requirement of an activating
proton or cation,” which can be hypothesized to afford
Furthermore, our investigations with different
[7] E. Arceo, J. A. Ellman, R. G. Bergman, J. Am. Chem. Soc. 2010, 132,
11408–11409.
[8] I. Ahmad, G. Chapman, K. M. Nicholas, Organometallics 2011, 30, 2810–
“
+
+
ZO-ReO (where Z=H or K ) as the viable catalytic core. The
3
2818.
use of high-boiling-point alcohol facilitates the reaction, and
the alcohol can serve as a hydrogen transfer agent. Neverthe-
less, the use of alcohol does not drastically increase the yields
of allylic alcohol as transfer hydrogenation from glycerol itself
remains kinetically competitive. This reaction can be employed
successfully with other biomass-derived polyols such as meso-
erythritol. The cation dependence of the reaction cannot be
explained based on the data in hand. It is not evident why
NH ReO4 is an effective catalyst but not Na or KReO . This
[
9] J. E. Ziegler, M. J. Zdilla, A. J. Evans, M. M. Abu-Omar, Inorg. Chem. 2009,
48, 9998–10000.
[10] a) M. E. Thibault, D. V. DiMondo, M. Jennings, P. V. Abdelnur, M. N. Eber-
lin, M. Schlaf, Green Chem. 2011, 13, 357–366; b) P. Ghosh, P. J. Fagan,
W. J. Marshall, E. Hauptman, R. M. Bullock, Inorg. Chem. 2009, 48, 6490–
6500; c) D. Di Mondo, D. Ashok, F. Waldie, N. Schrier, M. Morrison, M.
Schlaf, ACS Catal. 2011, 1, 355–364.
[11] Y. Liu, H. Tuysuz, C. J. Jia, M. Schwickardi, R. Rinaldi, A. H. Lu, W. Schmidt,
F. Schuth, Chem. Commun. 2010, 46, 1238–1240.
12] R. M. Painter, D. M. Pearson, R. M. Waymouth, Angew. Chem. 2010, 122,
646–9649; Angew. Chem. Int. Ed. 2010, 49, 9456–9459.
[
4
4
9
question and detailed kinetic studies of other substrates are
under investigation.
[13] T. J. Korstanje, J. T. B. H. Jastrzebski, R. J. M. K. Gebbink, ChemSusChem
2010, 3, 695–697.
[14] G. K. Cook, M. A. Andrews, J. Am. Chem. Soc. 1996, 118, 9448–9449.
[15] Z. Zou, J. H. Espenson, J. Org. Chem. 1996, 61, 324–328.
[16] M. Schlaf, P. Ghosh, P. J. Fagan, E. Hauptman, R. M. Bullock, Angew.
Chem. 2001, 113, 4005–4008.
Experimental Section
For a typical reaction, a 15 mL two-neck round-bottomed flask was
charged with 55 mmol (5.10 g) of glycerol and 1.1 mmol (274 mg)
MTO. The flask was connected to a distillation set including ther-
mometer, distillation column, and a collecting flask. The tempera-
Received: February 23, 2012
Revised: March 21, 2012
Published online on && &&, 0000
&4
&
www.chemsuschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 4
ÝÝ
These are not the final page numbers!

COMMUNICATIONS
Hello MTO: Efficient and environmen-
tally responsible conversion methods
are needed to deal with excess glycerol
production. The catalytic conversion of
glycerol to allyl alcohol under neat con-
ditions is catalyzed by methyltrioxorhe-
nium (MTO). The yields of volatile prod-
ucts and selectivity for allyl alcohol are
very good. Mechanistic studies reveal
transfer hydrogenation via bifunctional
catalysis.
J. Yi, S. Liu, M. M. Abu-Omar*
& – &&
&
Rhenium-Catalyzed Transfer
Hydrogenation and Deoxygenation of
Biomass-Derived Polyols to Small and
Useful Organics
ChemSusChem 0000, 00, 1 – 4
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
&
5
&
ÞÞ
These are not the final page numbers!