(OTf)2 as a homogeneous catalyst for the hydrogenation of biomass derived 2,5-hexanedione and 2,5-dimethyl-furan in aqueous acidic medium

Article abstract of DOI:10.1039/c7gc01956d

The complex [Ru(triphos)(CH₃CN)₃](OTf)₂ is an effective catalyst for the hydrogenation of 2,5-hexanedione and 2,5-dimethyl-furan in aqueous acidic medium at temperatures between 150 and 200 °C realizing up to 96% combined yields of 2,5-hexanediol and 2,5-dimethyl-tetrahydrofuran with the product distribution being sensitive to the amount of acid co-catalyst (HOTf) present. For the furan, the reaction pathway is through an acid-catalyzed hydrolysis to the dione rather than direct hydrogenation of the ring. The hydrogenation of the dione shows a first order dependence on hydrogen pressure as determined by direct hydrogen uptake rate measurements at temperature and pressure (1.38-6.90 MPa at 150 °C) and is postulated to operate through a heterolytic activation of hydrogen gas by [Ru(H)_x(triphos)(Y)_y]ⁿ⁺ (Y = solvent, water, counter ion) species formed in situ by loss and hydrogenation of the nitrile ligands. In water the catalyst is deactivated by dimerization to [Ru₂(μ-OH)₃(triphos)₂](OTf).

Full text of DOI:10.1039/c7gc01956d

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: E. Latifi, A.
Marchese, M. Hulls, D. V. Soldatov and M. Schlaf, Green Chem., 2017, DOI: 10.1039/C7GC01956D.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Volume 18 Number
7 7 April 2016 Pages 1821–2242
Green
Chemistry
Accepted Manuscripts are published online shortly after
Cutting-edge research for a greener sustainable future
www.rsc.org/greenchem
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1463-9262
CRITICAL REVIEW
G. Chatel et al.
Heterogeneous catalytic oxidation for lignin valorization into valuable
chemicals: what results? What limitations? What trends?
rsc.li/green-chem

Page 1 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
[Ru(triphos)(CH3CN)3](OTf)2 as a homogeneous catalyst for the hydrogenation of biomass
derived 2,5-hexanedione and 2,5-dimethyl-furan in aqueous acidic medium.
Elnaz Latifi, Austin Marchese, Margaret Hulls, Dmitriy V. Soldatov and Marcel Schlaf*
Department of Chemistry, University of Guelph
The Guelph-Waterloo Centre for Graduate Work in Chemistry (GWC)2
50 Stone Road East, Guelph, Ontario, Canada, N1G 2W1
mschlaf@uoguelph.ca
Abstract
The complex [Ru(triphos)(CH3CN)3](OTf)2 is an effective catalyst for the hydrogenation of 2,5-
hexanedione and 2,5-dimethyl-furan in aqueous acidic medium at temperatures between 150 and
200 °C realizing up to 96 % combined yields of 2,5-hexanediol and 2,5-dimethyl-tetrahydrofuran
with the product distribution being sensitive to the amount of acid co-catalyst (HOTf) present. For
the furan, the reaction pathway is through an acid-catalyzed hydrolysis to the dione rather than
direct hydrogenation of the ring. The hydrogenation of the dione shows a first order dependence
on hydrogen pressure as determined by direct hydrogen uptake rate measurements at temperature
and pressure (1.38 – 6.90 MPa at 150 °C) and is postulated to operate through a heterolytic
activation of hydrogen gas by [Ru(H)x(triphos)(Y)y]n+ (Y = solvent, water, counter ion) species
formed in situ by loss and hydrogenation of the nitrile ligands. In water the catalyst is deactivated
by dimerization to [Ru2(µ-OH)3(triphos)2](OTf).
1

Green Chemistry
Page 2 of 36
View Article Online
DOI: 10.1039/C7GC01956D
Introduction
The sugar-derivable furans furfural and 5-hydroxymethyl-furfural (5-HMF) are important
platform molecules for the production of chemicals from biomass.1-8 They are characterized by an
extremely high intrinsic reactivity based on their high density of organic functionalities, which
could be described as combining ether, cyclic bis-enol, aromatic, aldehyde and benzylic alcohol
functions, all in the same molecule. This opens up multiple pathways for an uncontrolled
decomposition and polymerization to molecularly poorly defined products generally referred to as
humins,9-11 which poses a substantial challenge to their conversion to value-added chemicals. In
this context, the humins themselves, which are also formed directly from C6 sugars such as
fructose (the precursor to 5-HMF as well as dihydropyrans as humin generating intermediates),12
could however also be considered to simply be another recalcitrant feedstock of complex
heterogeneous structure to be valorized.9 Using the hydrodeoxygenation (HDO) value chains
leading from furfural to 1,4-pentanediol and 2-methyl-tetrahydrofuran and 5-HMF to 2,5-
hexanediol and 2,5-dimethyl-tetrahydrofuran, respectively, as examples, Scheme 1 illustrates this
challenge and shows some of the conceivable reaction pathways leading to uncontrolled
polymerization via direct benzylic-type alkylation, aldol condensation or Diels-Alder reactions, or
any combination of the above. Similar challenges are present in alternative value chains, e.g., the
highly desirable conversion of the same substrates to the corresponding -diols 1,5-pentanediol
and 1,6-hexanediol that could find direct use in polyester and polyurethane (and with further
derivation) polyamide formulations.13-18
2

Page 3 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
The key for any such processes to be viable is that the individual hydrodeoxygenation (HDO)
steps, typically achieved by an iteration of acid-catalyzed loss (dehydration) or addition
(hydrolysis) of water followed by metal catalyzed hydrogenation or hydrogenolysis, must in each
case be kinetically competent over the various decomposition and polymerization pathways to
humins from 5-HMF (or directly from fructose) or furfural or the resinification of furfuryl alcohol
shown in Scheme 1. Furthermore, these non-productive pathways can not only lead to loss of
substrate, but also have the potential to result in an irreversible deactivation of the catalyst system.
This is true for both hetero- and homogeneously catalyzed HDO processes, with the former clearly
dominating the research endeavours in this field.19-23
As also indicated in Scheme 1, the propensity for humin formation is most pronounced for the
extremely reactive 5-HMF, while the intermediates and products of subsequent HDO steps show
– with the possible exception of the transient aldehyde or ketone hydrolysis products of the methyl-
furans – a decreasing functionality density and reactivity, ultimate leading to diols and
tetrahydrofurans. These products can be considered to be stable against polymerization in aqueous
medium.
3

Green Chemistry
Page 4 of 36
View Article Online
DOI: 10.1039/C7GC01956D
Scheme 1: Example reaction pathways and conceivable challenges to the conversion of biomass
derived furans to hydrodeoxygenated value-added products.
Using almost exclusively processes in which both the acid dehydration/hydrolysis and metal
hydrogenation/hydrogenolysis catalysts are heterogeneous, the production of both 2-methyl-furan
and 2,5-dimethyl-furan has been realized following the sequences hemi-cellulose  xylose 
furfural  furfuryl alcohol  2-methyl-furan or cellulose  glucose  5-HMF  2,5-
bis(hydroxymethyl)-furan  2,5-dimethyl-furan, respectively.24-44 Here the work by Vlachos et
al., that achieved very high yields and selectivities to furfuryl alcohol from furfural and 2,5-
dimethyl-furan from 5-HMF by employing transfer hydrogenation over Ru/C under mild
conditions, serves as a good example of how breaking the value-chain into separate optimized
4

Page 5 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
steps with isolation of the intermediates can help to avoid humin formation.45, 46 Starting with the
seminal work by Descotes et al.,47 the direct “single-pot” heterogeneous hydrogenation and
hydrogenolysis of either fructose/glucose or furfural/5-HMF to products beyond 2-methyl- and
2,5-dimethyl-furan, i.e., to tetrahydrofurans as well as diols and triols has also been extensively
studied.17, 48-69 In contrast, the explicit production of 2,5-hexanedione, i.e., the hydrolysis product
of the cyclic bis-enol 2,5-dimethyl-furan, from biomass has received less attention,70-75 even
though its hydrogenation product 2,5-hexanediol is arguably the most valuable product of the
example C6 value-chain shown in Scheme 1. 2,5-hexanediol could in principle serve as a polyester
or polyurethane component and its derivative 2,5-Dimethyl-2,5-hexanediol (DMHD) is an
intermediate in the preparation of pyrethroids and of 2,5-Dimethyl-2,5-bis(tert-butyl-peroxy)-
hexane to produce polyethylene copolymers and polyethylene rubbers (by BASF).76
As already stated, the vast majority of studies aimed at the HDO of sugars and sugar-derived furans
have employed heterogeneous hydrogenation and hydrogenolysis catalyst systems, while attempts
to use homogeneous catalyst systems based on transition metal complexes, that also tend to act
only as hydrogenation, but not hydrogenolysis catalysts, are comparatively much fewer.77-87
Recent examples of the successful application of homogeneous catalysts to the hydrogenation of
a biomass-derived furan substrates are the highly efficient solvent-free hydrogenation of furfural
to furfuryl alcohol at up to 160 °C by a ruthenium complex formed in situ from Ru(acac)3 and 1,4-
bis(diphenylphosphino)butane,88 and the metal-free hydrogenation of 2,5-dimethylfuran to 2,5-
dimethyltetrahydrofuran in 64 % yield by the frustrated Lewis pair formed by a boron triaryl
species and the THF solvent and product itself.89
5

Green Chemistry
Page 6 of 36
View Article Online
DOI: 10.1039/C7GC01956D
The HDO of highly polar and reactive sugar-derived furans forces the use of a polar liquid reaction
medium – most preferably water – rather than a reaction in the gas phase. Empirically,
temperatures > 150 °C are also required to trigger any dehydration or hydrolysis steps required.83
Under these conditions, both heterogeneous and homogeneous catalyst systems are subject to
specific challenges of catalyst activity and stability. For heterogeneous catalysts, fouling or coking
of the active surface can occur by substrate decomposition and polymerization ( humin
formation), which can in principle however be reversed by thermal and/or oxidative (“burn-off”)
re-activation. Other potential issues are the leaching of active metal and/or dissolution of the
catalyst support by the reaction medium, which through the generation of water as the necessary
reaction by-product of HDO reactions, will be aqueous in nature.23, 90
For homogeneous catalysts based on molecular dispersed transition metal complexes, the biggest
challenge is product separation and recovery and reuse of the catalyst,83, 91with successful recycling
being to date limited to the ruthenium complex catalyzed hydrogenation of levulinic acid to -
valerolactone.92, 93 Based on the required reaction conditions defined above, catalyst stability
against loss of ligand and reduction of the hydrogenation metal to bulk material in oxidation state
0, i.e., effectively decomposition to a heterogeneous system, are then the key challenges that have
to be resolved before recovery and reuse can be effected. It is therefore the opinion of the authors,
that the design, characterization and synthesis of homogeneous catalysts with unprecedented
thermal stability in aqueous (acidic) solution is the primary issue in the application of such systems
to the hydrodeoxygenation of sugar-derived substrates.
6

Page 7 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
Further hypothesizing that homogeneous catalyst systems may in specific reaction sequences show
advantages and result in different reactivity patterns and selectivities, we have developed a series
of homogeneous hydrogenation catalysts based on variations of a pyridine chelate ligand theme
coordinated to ruthenium(II). Exploiting the chelate effect,94 these complex catalysts do indeed
show unusually high, but still insufficient, stabilities at high-temperatures in aqueous acidic
medium and are applicable to the HDO of terminal diols and glycerol as well as the conversion of
2,5-dimethyl-furan and 2,5-hexanedione to 2,5-dimethyl-tetrahydrofuran (as the cis/trans isomeric
mixture), 2,5-hexandiol and hexane.82, 87, 95-100
Specifically,
the
catalyst
systems
[(4’-Ph-terpy)Ru(H2O)3](OTf)2
and
[(4'-Ph-
terpy)(quS)Ru(L)](OTf) (quS = 8-quinoline-thiol) can effect these transformations in moderate to
good yields at T ≥ 150 °C, but ultimately suffer from irreversible deactivation to the very stable
coordinatively saturated [Ru(4’-Ph-terpy)2](OTf)2 and inactive Ru(0), severely limiting their
recyclability.82, 87
Previous work by the groups of Elsevier,101, 102 Bianchini,103 Jun,104 Kilner,105 Dyson,106 Miller,
Cole-Hamilton, Klankermayer and Leitner107-111 as well as Beller,112-114 established the activity of
the [Ru(triphos)] (triphos = 1,1,1-tris(diphenylphosphinomethyl)ethane) fragment as a versatile
and very robust homogeneous hydrogenation catalyst for a broad range of substrates. Expanding
our investigations to chelating tridentate phosphine ligands, we here explore the applicability,
behaviour and limitations of this system for the conversions shown in Scheme 1 working
“backwards” along value-chain starting with 2,5-hexandione.
7

Green Chemistry
Page 8 of 36
View Article Online
DOI: 10.1039/C7GC01956D
Experimental
General. All reagents and solvents were purchased from commercial sources. 2,5-hexanedione,
2,5-dimethyl-furan and 1,4-dioxane were passed through a short plug of neutral Al2O3 (Brockmann
Activity I) immediately before use to remove any impurities, peroxides or stabilizers present that
could lead to false negatives in the catalytic reactions. All water used was HPLC grade.
Trifluoromethanesulfonic (HOTf, triflic) acid was stored under argon atmosphere and measured
into reaction mixtures using a micro-liter syringe. Catalyst syntheses were performed under argon
atmosphere using standard Schlenk-line techniques. [Ru2(µ-Cl)3(triphos)2]Cl and
[Ru(triphos)(CH3CN)3](OTf)2 (1) were obtained following the published procedure by Venanzi
and co-workers in yields comparable to those previously reported.115 Pale yellow single crystals
of 1 suitable for X-ray analysis were obtained by slow diffusion of diethylether into a methanol
solution of 1. In the same manner X-ray grade crystals of [Ru2(µ-OH)3(triphos)2](OTf) (2) were
obtained from an orange amorphous solid isolated from the catalytic reactions. All NMR spectra
were obtained on 300 MHz, 400 MHz, or 600 MHz spectrometers and calibrated to the residual
solvent signal. GC-MS analyses were performed on a Varian Saturn 2000 GC/MS using a 30 m
Stabilwax-da (acid-deactivated polyethylene glycol) column running in CI (CH3CN) mode.
Reaction products were identified by comparison to the retention times and MS fragmentation
patterns of authentic samples and database matches. GC analyses were performed on a Varian
3800 with FID detectors using the same column as the GC-MS. Quantification was carried out
using internal standard calibration against 100 mmol L-1 dimethyl sulfone (DMS) in a three level
calibration against authentic samples of 2,5-hexanedione, 2,5-hexanediol, 2,5-dimethyl-furan and
cis/trans-2,5-dimethyl-tetrahydrofuran (grouping and reporting the cis/trans-isomers of the latter
8

Page 9 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
as one product). Other products formed were quantified using the response factors of these
compounds based on the concept of the effective carbon number.116 All hydrogenation experiments
employed industrial grade H2 gas (99.995 %). All preparative hydrogenation experiments were
carried out in an Autoclave Engineers (AE) Mini-reactor with a 50 mL 316 stainless steel (316SS)
reactor vessel and impeller. At a total reaction solution volume of 25 mL the reactor has a gas-
phase headspace of 50 mL (unused reactor body space and enclosed and pressurized magnet-drive
stirring assembly) ensuring an adequate and not stoichiometrically limiting supply of hydrogen.
Unless otherwise specified (cf. control experiments) reactor vessel and impeller were thoroughly
cleaned and polished after every run by lathe at 600 rpm with 3M abrasive pads and sand-blasting,
respectively. Regular control reactions without addition of catalyst showed only marginal
conversion (< 5 %) of substrate to hydrogenated products (see main text) caused by the background
activity of the reactor walls.117 Hydrogenation experiments aimed at determining the reaction order
in hydrogen gas were carried out in the University of Guelph’s dedicated High-Pressure
Hydrogenation Facility using a 300 mL Autoclave Engineers Hastelloy™ (C-276) reactor
connected to a large high-pressure hydrogen reservoir via a mass flow controller and fitted with
an automatic pressure and PID temperature control system. The system allows the direct
measurement of hydrogen consumption at a constant (within ± 7 kPa ≡ ± 1 psi) pressure and thus
provides a direct differential rate measurement at a given temperature (ambient to 723 K), pressure
(0.69 – 10.3 MPa ≡ 100 - 1500 psi) and stirring rate. In these reactions a total reaction volume of
50 mL was employed providing for a pressurized headspace gas volume of > 250 mL (see further
discussion on mass flow limitations in main text). A single batch of catalyst 1 (X-ray grade
9

Green Chemistry
Page 10 of 36
View Article Online
DOI: 10.1039/C7GC01956D
crystals) was used for the kinetic study. Control reactions in the absence of catalyst showed no
background activity of the Hastelloy™ reactor material.
Representative procedure for a preparative catalytic hydrogenation reaction.
In a 25 mL volumetric flask were combined catalyst 1 (0.0286 g, 0.025mmol) and ~ 10 mL of
water (for 2,5-hexanedione substrate) or a 5:1 mixture of 1,4-dioxane/water (for the 2,5-dimethyl
furan substrate) to form a pale yellow suspension (water) or solution (1,4-dioxane/water).
Methylsulfone (0.235 g, 2.5 mmol, GC internal standard), 2,5-hexanedione (2.853 g, 25 mmol) or
2,5-dimethyl-furan (2.4105 g, 25 mmol) and (where applicable) 1-5 equivalents HOTf with respect
to ruthenium content were added and the volume made up to 25 mL with water or the 5:1 mixture
of 1,4-dioxane/water. The solution was then thoroughly mixed by vigorous manual shaking of the
flask for 60 seconds, placed in an Autoclave Engineers MiniReactor and purged three times with
hydrogen. The reactor was twice pressurized to 5.5 MPa (800 psi) with H2(g) and vented and then
re-pressurized to this working pressure, sealed and heated to the set temperature (150-225 °C),
which in all cases was reached in ≤ 20min. The reaction was stirred (500 rpm) at temperature for
16 h and then cooled to ambient. The reaction products were analyzed and quantified by GC/GC-
FID.
Procedure for the determination of the reaction order of hydrogen gas in the hydrogenation
of 2,5-hexanedione to 2,5-hexanediol.
In a 50 mL volumetric flask 2,5-hexanedione (5.706 g, 50 mmol) was dissolved in a 5:1 mixture
of 1,4-dioxane/water. The solution was then thoroughly mixed by vigorous manual shaking of the
flask for 60 seconds and placed into the 300 mL reactor vessel and an open glass (5 mm NMR)
10

Page 11 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
tube reaching several centimeters above the liquid level and containing catalyst 1 (0.0572 g, 0.050
mmol) placed into the reactor. (See ESI for images of this set-up). This isolates the catalyst from
the solution during pressurizing and heating. The reactor was twice pressurized to 5.5 MPa (800
psi) with H2(g) and vented and then re-pressurized to the desired working pressure heated to 423
K (150 °C) using a PID controller to prevent temperature overshoot. Once this temperature had
been reached (~ 60 min.) the mechanical stirrer was started (500 rpm), which breaks the glass tube
inside the reactor releasing the catalyst into solution. This marks time = 0 for the hydrogen uptake
rate measurements. Hydrogen uptake at constant temperature 423 K (150 °C) and pressure (200,
400, 600, 800 and 1000 psi ≡ 1.38, 2.76, 4.14, 5.12 MPa) was followed by data-logging from the
mass flow controller at 1 s time intervals until it ceased, i.e., over 1-3.5 hours depending on
hydrogen pressure.
Results
Catalyst synthesis and characterization. The complex [Ru(triphos)(CH3CN)3](OTf)2 (1) was
selected as the catalyst, as it is easily prepared via [Ru2(µ-Cl)3(triphos)2]Cl (2) following the
protocol by Venanzi et al.115 We anticipated that it a) should be soluble in polar aqueous media, b)
should generate
a
catalytically active ruthenium hydride species of the type
[Ru(H)x(triphos)(Y)y]n+ (Y = solvent, water, counter ion) by loss of the labile CH3CN ligands
followed by heterolytic activation of H2(g) to Ru-H and HY,106, 111 and c) would allow more precise
control of the amount of coordinated ruthenium present rather than rely on in situ coordination of
the triphos ligand to Ru(acac)3 or similar.
11

Green Chemistry
Page 12 of 36
View Article Online
DOI: 10.1039/C7GC01956D
A control reaction in which 1 was dissolved in CH3CN and heated to 423 K under 5.5 MPa H2(g)
resulted in the quantitative conversion of the solvent to ethylamine, limited only by the amount of
H2(g) supplied. This established that at T  423 K any ethylamine generated by this process does
– even when present in large amounts – not coordinatively inhibit the catalyst. This is as expected
based on the studies by Bianchini and Beller that employed the [Ru(triphos)] system as a catalyst
for the hydrogenation of nitriles.103, 112
In order to obtain samples of high purity of 1 for catalysis and kinetic studies, it was recrystallized
by diffusion of Et2O into a methanolic solution yielding pale yellow crystals of sufficient quality
for X-ray analysis. Figure 1 shows the structure of 1.
Figure 1: ORTEP of the cation of [(triphos)Ru(CH3CN)3](OTf)2 (1) at the 50 % probability level
oriented to show the 3-fold symmetry of the complex. (Triflate counter ion and hydrogens omitted
for clarity).
12

Page 13 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
Hydrogenation of 2,5-hexanedione by 1 in water.
Table 1 summarizes the results of a temperature dependence study of the activity of 1 for the
hydrogenation of 2,5-hexanedione in water. While 1 is only marginally soluble in water at ambient
temperature and the concentration employed, quantitative conversion of the substrate to 2,5-
hexanediol is realized under 5.5 MPa H2(g) at 150 and 175 °C (entries 1 and 2) giving a turn over
number (TON) of 2000 (due 2 equivalents of H2(g) consumed per mol of substrate) and turn over
frequency of 125 h-1 as the lower limits for these reaction conditions and time (16 h). No catalyst
activity was observed at 125 °C or lower, while at higher temperatures (200 and 225 °C, entries 3
and 4) relatively small, but increasing amounts of substrate decomposition (determined by the
mass balance deficiency against the GC internal standard – last column in Table 1) are observed
along with 11 % of dehydration ring-closure to cis/trans-2,5-dimethyl-tetrahydrofuran at the
highest temperature (225 °C) attempted. Carrying out a reaction with a 10-fold increase in substrate
concentration under otherwise identical conditions (Table 1, entry 5) also gives full conversion
yielding a viscous solution of ~ 2:1 2,5-dimethyl-tetrahydrofuran and 2,5-hexanediol and
establishes a catalyst turn over number (TON) of 20,000 and TOF of 1250 h-1 as the lower limits
in this case.
The reactions generate orange solutions along with a small amount of an orange coating of an
amorphous solid on the reactor body that readily dissolves in MeOH. Running a reaction without
cleaning this coating from the reactor body gives only marginal conversion, i.e., this material is
inactive as a catalyst (entry 6), instead giving large amount of substrate decomposition (39 %) and
leaving unreacted substrate (49 %). Carrying out a control reaction after cleaning the reactor and
13

Green Chemistry
Page 14 of 36
View Article Online
DOI: 10.1039/C7GC01956D
without addition of 1 gives a very similar result with the marginal conversion of substrate to diol
attributed to the background activity of the 316SS reactor body material.117 In contrast, twice
replenishing the filtered reaction solution obtained from a reaction at 150 °C (entry 1) with the
same amount of substrate (entries 8 and 9), again results in full conversion to 2,5-hexanediol. This
establishes that under these conditions the observed catalytic hydrogenation activity does in fact
reside in the homogeneous solution and not in the reactor body or heterogeneous catalyst
decomposition products such as the orange solid.
Table 1: Hydrogenation of 2,5-hexanedione by [Ru(triphos)(CH3CN)3](OTf)2 (1) in water.a
Entry
T
Mass
Balance
Deficiencyc
[ °C]
150
175
200
225
150
200
200
150
150
[%]b
0
[%]b
0
[%]b
100
100
95
[%]b
[%]b
0
1
2
0
0
2
2
0
7
9
0
0
0
0
0
3
0
0
3
4
0
11
66
0
76
11
0
5d
6e
7f
8 g
9 h
0
34
49
45
0
5
39
39
0
0
7
0
100
0
0
100
0
aReaction conditions: 2,5-Hexanedione [1000 mmol/L] in water, 5.5 MPa (800 psi) H2 (g), dimethylsulfone (ISTD) [100 mmol/L], catalyst load [1
c
mmol/L = 0.1% w.r.t. substrate], reaction time = 16 h. bBy quant. GC-FID against internal standard; ±1%; Mass Balance Deficiency due to
decomposition of the substrate and/or products to insoluble and non-volatile polymers and solids (humins) not quantifiable by GC. dReaction carried
out at 10,000 mmol/L substrate concentration, i.e., 0.01 % catalyst load. eControl reaction without catalyst and in presence of an orange precipitate
f
g
partially coating the reactor. Control reaction without catalyst after cleaning reactor. First re-use of solution of reaction 1 with same amount of
substrate re-added. hSecond re-use of solution of reaction 7 with same amount of substrate re-added.
14

Page 15 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
Recrystallization of the orange solid from MeOH/Et2O enabled single-crystal X-ray analysis that
identified it as [Ru2(µ-OH)3(triphos)2](OTf) (2) and established a deactivation pathway of catalyst
1 in aqueous solution under these conditions by dimerization via three bridging hydroxyl
ligands.118 Figure 2 shows the structure of 2. The binuclear complex has each Ru(II) center in
distorted octahedral coordination by three O-atoms from bridging hydroxyl anions and three P-
atoms of the chelating ligand. The Ru-O and Ru-P distances are within 2.13-2.16 and 2.25-2.28 Å,
respectively. The O-Ru-O, P-Ru-P and O-Ru-P angles range from 73.1-74.6, 85.8-88.2 and 93.4-
105.3o, respectively. The Ru-Ru distance is 3.09 Å. The cation has approximate symmetry of a
three-fold rotation axis. As a whole, the cation is similar to other previously reported binuclear
Ru(II) complexes with three hydroxyl bridges, e.g., [Ru2(PMe3)6(OH)3](BF4)119,
[Ru2(PMe3)2(OH)3](BPh4),120 and [Ru2(PMe3)2(OH)3](phenolate)(phenol) and others.121
15

Green Chemistry
Page 16 of 36
View Article Online
DOI: 10.1039/C7GC01956D
Figure 2: ORTEPs of [Ru2(µ-OH)3(triphos)2](OTf) (2) at the 50 % probability level in side and
axial view showing the position of the three bridging hydroxyl functions. (Oxygens in red. Triflate
counter ion, carbon-bound hydrogens and solvent molecules omitted for clarity).
Hydrogenation of 2,5-dimethyl-furan in 5:1 1,4-dioxane/water
Taking one step back in the value chain of Scheme 1, towards 5-HMF as one of the ultimately
desirable starting materials, the hydrogenation of 2,5-dimethyl-furan by 1 in water was attempted.
Scheme 2 summarizes the overall reaction cascade with the observed side- and follow-up reactions
already described in our previous study using [(4’-Ph-terpy)Ru(H2O)3](OTf)2 as the catalyst.82
16

Page 17 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
Scheme 2: Reaction cascade and observed intermediates and reaction products for the hydro-
deoxygenation of 2,5-hexanedione and 2,5-dimethyl-furan in aqueous acidic medium.82
However, compared to our earlier results with [(4’-Ph-terpy)Ru(H2O)3](OTf)2 as the catalyst in
which the same reaction gave up to 65 % conversion,82 the bi-phasic emulsions formed did not
give any hydrogenated products, but resulted in complete rapid decomposition and polymerization
of the substrate to brown sticky resins in the reactor, i.e., quantitative formation of humins,
precluding the use of pure water as the reaction medium in this case. We attribute this to the much
lower solubility of 1 in pure water compared to [(4’-Ph-terpy)Ru(H2O)3](OTf)2, which as an aqua
complex is freely soluble. In contrast, in the biphasic emulsion 1 more readily dissolves in the
furan substrate phase where interaction of the catalyst with the undiluted furan then results in the
observed decomposition to humins, which can be caused either by the generation of free HOTf by
the catalyst or conceivably the metal centre directly acting as a Lewis acid catalyst. Together both
1 and 2,5-dimethyl-furan do however freely dissolve in a variety organic solvents. However, no
17

Green Chemistry
Page 18 of 36
View Article Online
DOI: 10.1039/C7GC01956D
catalytic activity against 2,5-dimethyl-furan was observed in toluene or other non-polar solvents,
even when biphasic aqueous mixtures of these solvents in the presence of phase-transfer catalysts
were employed (e.g., sodium dodecyl benzoic acid or NaBPh4), i.e., while 2,5-dimethyl-furan is
insoluble in water, its hydrolysis to 2,5-hexandione as the actual hydrogenation substrate is a
prerequisite for its conversion by the homogeneous catalyst system. This reaction pathway is also
supported by the observation of 2,5-hexanediol in reactions starting from 2,5-dimethyl-furan and
by control reactions from our previous study,82 in which no hydrolysis of 2,5-dimethyl-
tetrahydrofuran to 2,5-hexanediol was observed even under much more extreme reaction
conditions (225 °C with added acid in water). This – by necessity – means that any 2,5-hexanediol
formed from 2,5-dimethyl-furan can only originate from 2,5-hexandione generated by hydrolysis
of 2,5-dimethyl-furan and not from the hydrolysis of 2,5-dimethyl-tetrahydrofuran generated by
the direct hydrogenation of 2,5-dimethyl-furan. Following our previous reports,82, 87, 122 1,4-
dioxane was therefore again selected as the reaction medium on the basis of its anticipated
inertness and similarity to the possible ultimate reaction product cis/trans-2,5-dimethyl-
tetrahydrofuran and its full miscibility with water. The reaction medium was therefore modified
to the azeotropic 5:1 mixture of 1,4-dioxane/water, in which both the substrate and catalyst 1 are
fully soluble at ambient temperature. Control reactions, in which a 1 mol/L solution of 2,5-
dimethyl-furan in this solvent mixture was heated to 125 or 150 °C in the presence of 0.1 mol %
(= 1 mmol/L) [H3O+][OTf],123 but absence of 1 does in fact lead to hydrolysis of the furan to 2,5-
hexanedione, but is dominated by substrate decomposition to humins. Figure 3 shows the evolution
of the composition of this mixture over 7 hours at 150 °C. The azeotropic reaction medium also in
principle allows for a simple separation of products, solvent and catalyst by distillation. At a lower
18

Page 19 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
1,4-dioxane/water ratio the furan is not fully soluble and phase separation is observed. The amount
of water in this mixture thus represents its optimized maximum concentration and enables the
hydrolysis of the furan to 2,5-hexanedione while also serving to supress condensation reactions to
humins by aldol condensation (cf. pathway b, Scheme 1).
Figure 3: Control experiment for the hydrolysis of 2,5-dimethyl-furan to 2,5-hexanedione in 5:1
1,4-dioxane/water at 150 °C in the presence of 0.1 mol % HOTf.
19

Green Chemistry
Page 20 of 36
View Article Online
DOI: 10.1039/C7GC01956D
As the results of a temperature dependence study summarized in entries 1-5 of Table 2 show, in
the presence of 1 much less humin formation and instead conversion to 2,5-hexanediol and 2,5-
dimethyltetrathydrofuran occurs in moderate to good yields reaching an optimum at 200 °C with
65 % yield of hydrogenated products.
In the absence of added acid co-catalyst, reusing the solution from this reaction by replenishing
with the same amount of substrate gives substantially lower conversions due to catalyst
deactivation by formation of 2, which in the dioxane/water mixture remains dissolved forming
deep orange solutions.
With the addition of increasing amounts of HOTf (1 to 5 equivalents with respect to 1) as a
hydrolysis co-catalyst, almost complete conversion (up to 96 %, entry 9) to hydrogenated products
is achievable at 150 °C, with a TON/TOF of only 120/7.5 h-1 to 2,5-hexanediol, but 1800/112.5 to
2,5-dimethyl-tetrahydrofuran, i.e., in the more acidic solution, ring closure to cis/trans-2,5-
dimethyl-tetrahydrofuran becomes the increasingly dominant product by acid-catalyzed ring-
closure of the diol (Scheme 2), which is consistently generated as a ~ 1:1.25 trans:cis isomeric
mixture from all reactions.124 In the presence of 5 equivalents of acid, clear orange solutions are
still obtained and reusing the solution (Entry 10, Table 2) still realizes 80 % of hydrogenated
products. Repeating the reaction of entry 9 at higher hydrogen pressure for shorter times (entry 11,
13.8 MPa for 3 h) enhances the yield of the more desirable 2,5-hexanediol, but also gives a higher
amount of humin formation, which is however an artifact of the required reactor set-up at these
higher pressures (see Discussion section).
20

Page 21 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
Table 2: Hydrogenation of 2,5-dimethyl-furan by [Ru(triphos)(CH3CN)3](OTf)2 (1) in 5 : 1 1,4-dioxane/water.a
Mass
Balance
Deficiencyc
Entry
T
HOTf
[ °C] [eqv.
to Ru]
[%]b
[%]b
[%]b
[%]b
[%]b
[%]b
1
2
150
175
200
200
225
150
150
150
150
150
150
0
0
0
0
0
1
3
4
5
5
5
83
31
16
55
21
73
4
2
2
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
6
0
0
0
0
0
13
13
19
20
23
19
17
14
4
35
45
20
36
2
21
20
5
3
4d
5
19
0
6
7
54
22
6
25
64
90
41
48
8
0
9
0
10e
11f
7
39
40
13
12
0
aReaction conditions: 2,5-dimethyl-furan [1000 mmol/L] in 5:1 1,4-dioxane : water, 5.5 MPa (800 psi) H2 (g), dimethylsulfone (ISTD) [100 mmol/L],
catalyst load [1 mmol/L = 0.1% w.r.t. substrate], reaction time = 16 h. bBy quant. GC-FID against internal standard; ±1%; cMass Balance Deficiency
due to decomposition of the substrate and/or products to insoluble and non-volatile polymers and solids (humins) not quantifiable by GC. dRe-use
e
of solution of reaction 3 with same amount of substrate re-added. Re-use of solution of reaction 9 with same amount of substrate re-added).
fReaction time 3 h at 13.8 MPa (2000 psi) H2(g).
Kinetic studies
In order to gain some insight into the mechanism and limiting factors of the reactions catalyzed by
1 in aqueous media, the reaction order of hydrogen gas in the hydrogenation of 2,5-hexanedione
at 150 °C (473 K) was determined by directly measuring the rate of hydrogen uptake at five
constant pressures ranging from 1.32 MPa (200 psi) to 6.90 MPa (1000 psi). To ensure that the
data obtained is not skewed by solubility limitations of the catalyst, the reactions were conducted
in 5:1 1,4-dioxane/water, in which 1 is completely soluble at ambient temperature at the
concentrations employed (identical to those used in the preparative reactions described above, cf.
Tables 1 and 2). To further ensure that the measured rates of hydrogen uptake do not simply reflect
21

Green Chemistry
Page 22 of 36
View Article Online
DOI: 10.1039/C7GC01956D
the diffusion limit of hydrogen gas into the reaction solution, we explicitly calculated the actual
amount of hydrogen present in 50 mL of water at temperature and pressure using the data and
models provided by Baranenko and Kirov.125,126 The calculation establishes that even at the lowest
constant pressure employed (1.38 MPa = 200 psi) the molar ratio of [H2]/[Ru] is never less than
10. This calculation actually provides only a lower limit of hydrogen solubility and availability in
our reaction solutions, as the solubility of hydrogen is comparatively lower in water than in less
polar organic solvents or mixtures of these with water. The more limited data and models provided
by Brunner suggest that hydrogen solubility is in fact substantially higher in the 1,4-dioxane/water
mixture than in pure water.127 Taking further into account that the literature data refer to static gas
dissolution equilibria, while the reactions were stirred at 500 rpm, which should substantially
enhance gas diffusion into the reaction solution, we have high confidence that the hydrogen uptake
rates reported in Table 3, which represent a direct differential reaction rate measurement, do not
reflect any mass transport limitations in the reactor, but are in fact the true hydrogen consumption
rates by the catalytic system under the chosen reaction conditions. In addition, this approach has
the advantage that rate data is obtained without disturbing the reaction equilibria of the reaction
by drawing samples for quantitative analysis, e.g., by GC, and determining reaction order by
plotting concentration changes over time.
22

Page 23 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
Table 3: Measured rates of hydrogen uptake in the hydrogenation of 2,5-hexanedione by
[Ru(triphos)(CH3CN)3](OTf)2 (1) in 5:1 1,4-dioxane/water at 150 °C (423 K) as a function of hydrogen pressure.
Entry
p(H2)b
(295 K)
[psi]
p(H2)b
p(H2)c
p(H2)c
Rate of
H2
Rate of
H2
uptake
[mol/s]
7.868
p(H2)
log [p(H2)]
normalized
Rate of H2 uptake
normalized
log [Rate of H2
uptake]
(295 K) (422 K) (422 K)
[MPa]
normalized
[psi]
[MPa]
uptake
[dimensionless]f
[dimensionless]f
[dimensionless]f
normalized
[scc/s]d
[dimensionless]f
1
2
3
4
5
200
400
600
800
1000
1.38
2.76
4.14
5.52
6.90
303
531
2.089
3.661
5.488
7.315
8.963
0.0897
0.1633
0.2873
0.3264
0.4435
1.000
1.752
2.627
3.502
4.290
0.0000
0.2437
0.4195
0.5443
0.6325
1.000
1.821
3.203
3.639
4.944
0.0000
0.2602
0.5055
0.5610
0.6941
14.32
25.20
28.63
38.90
796
1061
1300
aReaction conditions: 50 mL solution of 2,5-hexanedione [1000 mmol/L] in 5:1 1,4-dioxane : water, catalyst load [1 mmol/L = 0.1% w.r.t. substrate].
bH2(g) reactor fill pressure at ambient condition (293 K). cPressure in reactor reached at 423 K – held constant at this temperature to ± 1 psi. dssc/s
= standard cubic centimeter of hydrogen uptake per second as measured by calibrated mass flow controller at NIST NTP (T = 293.15 K, p = 101.325
kPa). eAssuming ideal gas behaviour for H2(g). fBy division by data from 1st entry.
The rates of hydrogen uptake at a given pressure listed in Table 3 were constant over the range of
10-90 % of substrate conversion (total amount 50 mmol) and the total uptake of hydrogen gas after
completion of the reaction (after 0.75-3.5 h depending on pressure) closely matched the expected
amounts (100 mmol for complete and selective conversion to 2,5-hexanediol). As the data in Table
3 shows the measured rates of hydrogen uptake scale with the applied pressure. Determination of
the reaction order of hydrogen follows Equation (1).128
푚푚표푙푠 표푓퐻
2
푟푎푡푒 =
= 푘′ × [퐻2]푎
(1)
푡푖푚푒
푘′ = [푐푎푡푎푙푦푠푡]푏 × [푠푢푏푠푡푟푎푡푒]푐 × …
As the actual (and constant) hydrogen concentrations in the solutions are not exactly known,
pressures and rates were normalized to dimensionless numbers by division of the data for Entry 1
in Table 3. Direct or log/log plots of these normalized rate vs the normalized pressures both give
slopes close to unity with acceptable correlation coefficients (Figure 4) indicating a 1st order in
H2(g).
23

Green Chemistry
Page 24 of 36
View Article Online
DOI: 10.1039/C7GC01956D
Figure 4: Direct and log/log plots of the normalized rates vs normalized pressures.
Discussion
Previous mechanistic studies – both by NMR and in silico106, 111, 129 – suggest that the [Ru(triphos)]
fragment can heterolytically activate H2(g) into an active Ru-H species and a proton, i.e., activation
of H2(g) by 1 can generate at least one equivalent of acid. Starting from 1 such a Ru-H species
must be generated by dissociation of (at least) one of the CH3CN ligands from 1. The proton is
then taken up either by ethylamine generated by hydrogenation of the nitrile ligands in 1 or – in
the presence of excess acid (cf. entries 6-10, Table 2) – water as the strongest available base in the
aqueous medium. A [Ru(triphos)(H)n(L)m](2-n)+ (n, m = 1 or 2, L = labile ligand: H2O, solvent,
+
substrate, product) species and either H3CCH2NH3 or hydronium together could then act as a ionic
hydrogenation catalyst 130
From the reaction of entry 9, Table 1, the lower TON and turn over frequency (TOF) limits for 1
in this reaction are 20,000 and 1,250 h-1, but possibly are substantially higher, which was however
not further explored in this study. The reaction order for H2(g) (cf. Table 3 and Figure 4) suggests
that in the hydrogenation of 2,5-hexanedione by 1 in aqueous medium, the activation of hydrogen
24

Page 25 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
by a [Ru(triphos)] fragment is the rate-determining step. As the only species in the system capable
of activating hydrogen gas is such a [Ru(triphos)] fragment, this then suggests that the reaction is
also (at least) 1st order in ruthenium catalyst and operates through a Ru-H species of unknown
structure.131
Comparison of the results of the hydrogenation reactions of 2,5-hexanedione vs 2,5-dimethyl-furan
for reactions under identical conditions (entries 1-3 in Tables 1 and 2) shows that this kinetic model
clearly does not apply to the furan substrate. Instead (and with reference to Figure 3), the hydrolytic
ring opening to the dione as the actual hydrogenation substrate must be the rate determining step,
which then in turn is dependent on the relative concentration of acid catalyst and water, i.e., HDO
reactions targeting furans should – substrate solubility permitting – best be conducted in pure
water. However, as already stated above, in the 1,4-dioxane/water azeotrope, the relative
concentration of water is at its maximum for a homogeneous solution of 2,5-dimethyl-furan, i.e.,
for this substrate this parameter cannot be further optimized in this solvent system. The limited
solubility of 1 in pure water then represents a further limitation of this system. The amount of acid
added in the hydrolysis control reaction illustrated in Figure 3 represents the maximum
concentration of H3O+ that could be generated by 1, assuming formation of a single hydride ligand
through the heterolytic activation of hydrogen gas. The fact that in the absence of added acid at
temperatures from 150-225 °C (Table 2, entries 1-5) 1 gives much less humin formation than free
acid at a ruthenium equivalent concentration (0.1 mol % of substrate = 1 equivalent/Ru) suggests
that either not all of 1 participates in the activation of hydrogen gas, necessarily resulting in a lower
overall acid concentration, or – more likely – that a continuous hydrogenation of 2,5-hexanedione
prevents humin formation via the aldol condensation pathway of the dione substrate (cf. Scheme
25

Green Chemistry
Page 26 of 36
View Article Online
DOI: 10.1039/C7GC01956D
1), or both. Two significant trends emerge from the acid concentration series reactions represented
by entries 1 and 6-9 in Table 2: in the presence of metal catalyst higher acid concentrations
(unexpectedly) give lower amounts of humin and (expectedly) higher amounts of ring closure THF
product. These trends are visualized in Figure 5.
Figure 5: Product distribution in the hydrogenation of 2,5-dimethyl-furan by 1 at 423 K at 5.5
MPa H2(g) in 5:1 1,4-dioxane/water as a function of the equivalents of HOTf acid co-catalyst
added. (2,5-DMF = 2,5-dimethylfuran; 2,5-HDO = 2,5-hexanedione; 2,5-DMTHF = 2,5-
dimethyltetrahydrofuran; MBD = Mass Balance Deficiency due to humin formation).
Higher acid concentration thus accelerates hydrolysis, limiting humin formation directly from the
furan. Suppression of humin formation by acid catalyzed aldol condensations of the hydrolysis
product 2,5-hexanedione (by pathway b) in Scheme 1) is then mainly contingent on the kinetic
26

Page 27 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
competency of the metal-catalyzed hydrogenation reactions to the diol. A limit on the maximum
viable substrate concentration (not explored here) would however be reached if the condensation
reactions a) or b), both of which must be 2nd order in either dione or furan, would become faster
than hydrogenation.
Beyond its role in the hydrolysis ring-opening of the furan to the actual dione hydrogenation
substrate, the HOTf acid co-catalyst also appears to prevent deactivation of the catalyst by
formation of 2. It could do so by protonating one or several of the -OH bridges in the dinuclear
species (2) that does not possess a free coordination site for H2(g) activation, generating species
with labile aquo ligands. In any of these the labile aquo ligand can be displaced by an 2-H2
ligand,132 ultimately leading to formation of an active hydride species. It is then further conceivable
that any of the resulting and still mono- or bis-hydroxy bridged dimeric species could also be active
as hydrogenation catalysts, which – at least in the absence of acid – would lead formally to
fractional reaction orders larger than unity for 1, which is undesirable. Therefore an optimized
amount of acid co-catalyst could not only impact the furan hydrolysis to the dione hydrogenation
substrate, but also helps to maximize the efficient use of ruthenium present in the system.
According to our kinetic study even lower degrees of humin formation should in principle be
achievable by applying both higher acid co-catalyst concentrations and higher H2(g) pressures, as
this should enhance the kinetic competency of both the hydrolysis and the hydrogenation steps.
However, as the reaction of entry 11 of Table 2 shows, which resulted in a higher degree of humin
formation than the same reaction at lower pressure, this approach is in the batch reactors employed
met with some technical limitations. In order to operate at the higher pressure of 13.9 MPa (2,000
27

Green Chemistry
Page 28 of 36
View Article Online
DOI: 10.1039/C7GC01956D
psi, cold), this reaction had to be conducted in the much larger 300 mL Hastelloy reactor with
higher mass and heat capacity, which – in order to avoid temperature overshoots that destroy the
catalyst – requires a heat up time to the catalyst activation temperature of 423 K that is about 3
times longer (20 vs 60 min.) than for the 50 mL mini-reactor employed in all other reactions listed
in Tables 1-3. During this heat-up time hydrolysis of the furan substrate and decomposition either
directly or via acid-catalyzed aldol condensation of the dione already occurs before the catalyst
activates to transform the substrate to the stable diol or THF product. (In contrast, this is not a
problem in the kinetic studies on 2,5-hexanedione, which were conducted in the absence of acid
co-catalyst). As this reaction also ceased to consume hydrogen after 3 hours, it was stopped at this
time resulting in a higher yield of the diol vs the THF product indicating that acid-catalyzed ring-
closure to the ether is comparatively slow in the 1,4-dioxane/water azeotrope. Both observations
illustrate that awareness of chemical engineering parameters reaching beyond any intrinsic
chemical limitations is important when considering transformations of biomass-derived substrates
with high self-reactivity, in particular when running in batch reactors as in the present case.
Conclusions
The factors that limit a successful HDO of the biomass derived substrates targeted here are specific
to each actual substrate in the value chain, as the different steps in the reaction cascades have very
different rate-limiting features. For the actual hydrogenation substrate 2,5-hexanedione they are
dependent on hydrogen pressure and by necessary extension metal catalyst concentration, while
for the furan substrate, which cannot be directly hydrogenated by the homogeneous catalyst, the
relative concentrations of water and acid co-catalyst that govern the hydrolysis step are decisive.
28

Page 29 of 36
Green Chemistry
View Article Online
DOI: 10.1039/C7GC01956D
This constitutes a unique challenge, as in homogeneously catalyzed batch reactions the entire
reaction cascade takes place simultaneously, with the role of the catalyst in the system presented
here solely being the hydrogenation of 2,5-hexandione, while the product selectivities (diol vs
THF) are dependent on the multiple (de-)hydration ring-opening/closing equilibria in the reaction
cascade shown in Scheme 2. For both substrates, conducting the reactions in pure water would
therefore be advantageous, as this would maximize the rate of furan hydrolysis to the actual dione
hydrogenation substrate, suppress the aldol condensation pathway of the latter and maximize the
yield of the – arguably – most valuable 2,5-hexanediol by inhibiting ring-closure to 2,5-
dimethyltetrahydrofuran. The inactivity of the catalyst against furans thus presents an opportunity
that differentiates this homogeneous system from heterogeneous systems that directly yield the
tetrahydrofuran rather than diol products from biomass derived furans. The ultimately desirable
extension of the experiments described here to targets further “back” in the value chain will then
also require the ability to conduct reactions in pure water at optimized minimal acid co-catalyst
concentrations in order to supress humin formation anticipated to be much more severe with these
higher reactivity substrates. In light of the now known deactivation pathway of 1 and the
insolubility of 2 in water, realizing an increased catalyst solubility and activity in water – again at
optimized minimal acid co-catalyst concentrations – by modification of its structure in addition to
a further enhanced thermal catalyst stability will thus be a key issue to improve on the current
system to make it directly applicable to the conversion of the more reactive and more water-soluble
substrates furfural or furfuryl alcohol and 5-HMF or 2,5-(bis-hydoxymethyl)-furan to diols rather
than tetrahydrofurans.
29