Selective hydrothermal reductions using geomimicry

Article abstract of DOI:10.1039/c9gc00636b

Reduction of carbon-carbon π-bonds has been demonstrated using iron powder as the reductant and simple powdered nickel as the catalyst in water as the solvent at 250 °C and the saturated water vapor pressure, 40 bars. Stereochemical, kinetic and electronic probes of the mechanism suggest reaction via a conventional Horiuti-Polyani process for hydrogenation at the nickel metal surface. Selective reduction of carbon-carbon π-bonds is observed in the presence of other functional groups. The reactions use benign and Earth-abundant reagents that are at low depletion risk and take place in water as the only solvent under conditions that are characteristic of many geochemical processes.

Full text of DOI:10.1039/c9gc00636b

View Article Online
View Journal
Green
Chemistry
Cutting-edge research for a greener sustainable future
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: C. Bockisch, E. D.
Lorance, G. Shaver, L. Williams, H. Hartnett, E. Shock and I. R. Gould, Green Chem., 2019, DOI:
10.1039/C9GC00636B.
Volume 20
This is an Accepted Manuscript, which has been through the
Number
May 2018
Pages 1919-2160
9
7
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Cutting-edge research for a greener sustainable future
rsc.li/greenchem
Accepted Manuscripts are published online shortly after 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
Information for Authors.
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 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
PAPER
Paul T. Anastas et al.
The Green ChemisTREE: 20 years after taking root with the 12
principles
rsc.li/greenchem

Page 1 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
Selective Hydrothermal Reductions Using Geomimicry
Christiana Bockisch,a Edward D. Lorance,b Garrett Shaver,a Lynda B. Williams,c Hilairy
E. Hartnett,*ac Everett L. Shock,*ac and Ian R. Gould*a
a School of Molecular Sciences, Arizona State University, Tempe, AZ 85287
b Department of Chemistry, Vanguard University, Costa Mesa, CA 92926
c School of Earth and Space Exploration, Arizona State University, Tempe, AZ 85287
Abstract
Reduction of carbon-carbon -bonds has been demonstrated using iron powder as the reductant
and simple powdered nickel as the catalyst in water as the solvent at 250°C and the saturated
water vapor pressure, 40 bars. Stereochemical, kinetic and electronic probes of the mechanism
suggest reaction via a conventional Horiuti-Polyani process for hydrogenation at the nickel metal
surface. Selective reduction of carbon-carbon -bonds is observed in the presence of other
functional groups. The reactions use benign and Earth-abundant reagents that are at low
depletion risk and take place in water as the only solvent under conditions that are characteristic
of many geochemical processes.
1

Green Chemistry
Page 2 of 38
View Article Online
DOI: 10.1039/C9GC00636B
Introduction
The vast majority of the Earth’s organic carbon is located within the crust, mainly in continental
margin sediments and sedimentary basins.1 The chemical reactions of this organic material are
involved in a wide range of important geo- and biogeochemical processes, from petroleum
generation to supporting subsurface microbial communities.2 Organic reactions under these
conditions are noteworthy in that they take place in water as the only solvent and use reagents
and catalysts that by definition are mainly benign and Earth-abundant.3,4 For these reasons,
organic reactions under conditions that mimic geochemical conditions, i.e. geomimicry, are
attracting increasing attention for green chemistry applications.5
Geochemically relevant temperatures are not extreme, usually well below the critical point of
water, i.e. ~150°C - ~250°C. Nevertheless, water at these temperatures and associated confining
pressures exhibits several characteristics that are beneficial to green chemistry applications. First,
the dielectric constant is considerably lower than at room temperature: at 250 ˚C, close to those
of methanol and acetone (~25), thus water at these temperatures is a good solvent for many
organic compounds.6 Second, the pKW of water at 250°C is 11 compared to the canonical 14 at
25°C, i.e., which means that the concentration of both hydronium and hydroxide ions are higher.7
Water at these temperatures thus has “built-in” potent homogeneous acid and base catalysts in
the form of H3O+ and -OH. The separation of products is also relatively easy under
hydrothermal conditions because upon cooling the water, non-polar organic products
will mainly separate from the water and simply float to the top. Hydrothermal conditions
support a wide range of organic transformations, many of which occur in water alone, without
the addition of any other reagents or catalysts,3,4,8 or they use only benign or Earth abundant
2

Page 3 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
reagents.9
Some hydrothermal reactions are unexpected, or even quite different from those described in
traditional organic chemistry textbooks. For example, the dehydration of alcohols to form
alkenes occurs quickly, without the need to add any reagents or catalysts, and with high chemical
yield, even though organic chemistry textbooks report that under conditions closer to ambient the
reaction proceeds in the reverse direction; i.e., water adds to alkenes to form alcohols.10 We
showed recently that selective organic oxidation reactions could be performed under
hydrothermal conditions using the extremely mild oxidizing agent copper(II).9 Hydrothermal
oxidation of a primary alcohol to an aldehyde was accomplished with Cu(II), with minimal
further oxidation to the carboxylic acid, even though water was the solvent.9 Oxidation reactions
are among the most frequently studied in green chemistry contexts;11 in contrast, organic
reductions have received somewhat less attention. There is considerable interest in green
reduction reactions; for example, in liquid fuel generation and deoxygenation,12 and conversion
of carbon dioxide into useful raw materials.13 Reduction by catalytic hydrogenation is widely
employed in traditional benchtop and industrial organic chemistry.14 Although well established,
the traditional methods for heterogeneous hydrogenation generally use rare and expensive metal
catalysts such as platinum and palladium that carry a high depletion risk.15 The traditional
methods also usually use molecular hydrogen as the reducing agent.14,16 Several alternate
methods for performing catalytic reduction have been described in the literature, most of which
have focused on replacing the expensive catalysts (see, for example, refs 17). However,
molecular hydrogen is not naturally abundant, and is usually produced industrially from fossil
fuels via steam reformation, a process that has significant energy cost and forms carbon dioxide
as a by-product.18 Hydrogen is also a flammable gas with storage and transportation constraints.
3

Green Chemistry
Page 4 of 38
View Article Online
DOI: 10.1039/C9GC00636B
Therefore, there has also been considerable interest in exploring other reducing agents as
alternatives to molecular hydrogen.19
Of particular interest in this regard is a recent communication in this journal by Schafer et al.,
who demonstrated selective reduction of organic functional groups using palladium on carbon as
the catalyst, but using metallic aluminum as the reducing agent instead of hydrogen gas, and
water as the solvent.20 Reduction of C=C, C=N, C=O and N=O bonds was accomplished in
addition to hydrogenolysis of C-O, C-N and C-X bonds. These reductions are interesting because
they eliminate the need for hydrogen gas, and use water as the solvent, although the catalyst is
still based on the expensive palladium metal.20
As part of a research program on hydrothermal organic reactions relevant to geochemical
processes,3,8-10 we have discovered several organic reduction reactions related to those reported
by Schafer et al.,20 that take place in hydrothermal water as the only solvent, that require only
simple powdered nickel as the catalyst and an even milder and less-expensive reducing agent,
metallic iron, avoiding the expensive palladium. Metallic iron has been extensively investigated
as a reagent for large-scale industrial processes and waste remediation,21 but reports of its use as
an organic reducing agent are rare.22 Here we describe experiments that probe the scope,
selectivity, and mechanism of this new system for reducing carbon-carbon -bonds.
Experimental
Materials
All chemical materials were obtained from Sigma-Aldrich, except 1,2-dimethylcyclohexene and
2-methyl-1-methylenecyclohexane, which were obtained from ChemSampCo. The 1,2-
dimethylcyclohexene was purified by microscale distillation. Metal powders were obtained as
4

Page 5 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
nickel nanopowder (two purchases of different lot numbers), and iron powder (micron sized).
Water for all experiments was 18.2 MΩ, obtained from a BarnsteadTM purification system.
Methods
Metal Powder Characterization. The surface area of the nickel nanopowders was measured
using the nitrogen adsorption Brunauer–Emmett–Teller method. The surface area of the first lot
of nickel purchased was found to be 0.94±0.01 m2/g, the second lot of nickel from the same
supplier was found to have a surface area of 3.52±0.02 m2/g.
Powder X-ray diffraction (XRD) was performed using a Siemens D5000 with a cobalt
anode, scanned from 20° to 90° (2θ). Co-Kα radiation (40 kV, 30 mA). Three full scans were
typically averaged to increase the signal/noise ratio. Where appropriate, samples were milled to
ca. 20 micron particle size before analysis. XRD patterns were analyzed using EVATM
interpretative software and the JCPDS database (MaterialsData.com). The nickel and iron
samples were also imaged using scanning electron microscopy (SEM). Each sample was
sputtered with gold to obtain 3-5 nm coating prior to imaging with an XL30 environmental
scanning electron microscope (ESEM) with field emission gun (FEG). SEM images of each
metal powder are provided in the Supplementary Information.
Hydrothermal Experiments. Hydrothermal reactions were performed in sealed quartz/fused
silica tubes (Technical Glass Products) that were approximately 20 cm long, with 0.6 cm external
diameter and 0.2 cm internal diameter. Each reaction was performed with 0.2 g of water and the
concentrations of the organic reactants were all 0.1 molal. In each experiment, 40μmol organic
per 0.02 m2 of Ni and 240μmol of Fe were used. To maintain a constant surface area, the mass of
5

Green Chemistry
Page 6 of 38
View Article Online
DOI: 10.1039/C9GC00636B
the nickel was adjusted to account for the different surface areas for the two lots of the powdered
nickel. After correcting for the different surface areas, there was no detectable difference
between the different lots of nickel.
To eliminate oxygen, all water, liquid reagents and solutions of soluble organics in water
were purged with UHP Argon for 1 hour before being added to the reaction tubes. The tubes
were purged with UHP argon for an additional 10 minutes before freezing in liquid nitrogen and
evacuating to 60-70 mTorr, followed by sealing with a hydrogen torch. Roughly 50% of the
volume in each tube at room temperature was sample, the rest was headspace. For reactions
longer than 2 hours, tubes were heated in a repurposed, preheated, gas chromatography oven. For
reactions shorter than 2 hours, a brass block, equipped with internal heating elements, was used,
because it allowed for much more rapid warm-up of the samples. The reactions were quenched at
the end of the desired time period by plunging the tubes into room temperature water. The
reactions were heated to a temperature of 250°C. Under these conditions the reaction pressure
can be estimated as the saturated water vapor pressure, i.e., 40 bar.7 This is only an estimate
since if any gases are formed, see further below, the pressure in the tubes will be higher than this
value.
At the end of the experiments, after heating and quenching, the tubes were frozen in liquid
nitrogen, and after opening, the organic material was extracted using 2 mL of dichloromethane
containing decane. The decane was used as an internal gas chromatography standard.
Quantitative gas chromatography was performed using a Bruker-Scion 456 gas chromatograph
equipped with a flame ionization detector. Products were quantified using calibration curves
determined using authentic standards. Benzene and cyclohexane were not separable by gas
chromatography, nor were the hexane isomers formed via reduction of 3-hexyne. Therefore,
6

Page 7 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
experiments starting with benzene and 3-hexyne were analyzed by 1H NMR using a Bruker 400
MHz instrument. The chemical yields for the reaction times indicated were characterized using
quantitative gas chromatography, by comparing the calibrated peak sizes for the products to the
calibrated peak size for the unreacted starting material. The mass reaction mass balances
determined this way were all greater than 90% for all reactions, i.e., side products comprised less
than 10% of the final reaction mixtures.
Molecular hydrogen generation was measured by opening a reaction tube in a 1 inch internal-
diameter Tygon tube capped with rubber septa. The volume of the Tygon tube with the glass
tube inside was determined by weighing the entire apparatus with and without water. The
reaction tubes were broken in the sealed Tygon tube, and 10 μL of the headspace was quickly
sampled using a gas-tight syringe and analyzed using a reducing compound photometer gas
chromatograph (RCP-GC), that was calibrated with a 50 ppm H2 standard with nitrogen balance
(Praxair). The volume of the Tygon tubing and the solubility constant for hydrogen in water at
25°C, determined using the revised Helgeson-Kirkham-Flowers equations of state,23 were used
to calculate the total mass of hydrogen generated in the reaction tubes.
Results and Discussion
Reduction of Cyclohexene
Cyclohexene was studied as an example alkene. Reaction of cyclohexene in water at 250°C with
nickel and iron results in 95% conversion of the alkene in 1 hour. The major product is
cyclohexane (90%) plus a small amount of benzene (10%), eq 1. Cyclohexane is presumably
7

Green Chemistry
Page 8 of 38
View Article Online
DOI: 10.1039/C9GC00636B
formed by hydrogenation of the alkene on the nickel surface (see below), with the hydrogen
produced from H2O upon oxidation of the iron. Further experiments were performed in order to
investigate the roles of the nickel and the iron.
Reaction of cyclohexene was compared in reactions with nickel alone, with iron alone and
with both metals present. With both metals at 250°C the cyclohexene conversion was found to be
55% after 5 minutes. Under the same conditions and same reaction time period with nickel alone,
the conversion was 25%. This experiment shows that nickel alone is capable of reducing the
cyclohexene, with nickel presumably acting as both the reducing agent and catalyst. With iron
alone for the same time period, cyclohexene conversion was negligible (<1%). These results
suggest that metallic nickel is required in order to catalyze reduction, and that neither metallic
iron or its oxidized forms are capable of catalyzing the reaction. Additional insight was obtained
from experiments in which the possible formation of molecular hydrogen was investigated using
thermodynamic computations and experiments.
To test for formation of molecular hydrogen, a sample tube was loaded with nickel, iron, and
water, as per a typical experiment, but no organic was added. The tube was heated at 250°C for 1
hour. No attempt was made to run the reaction to completion, since the reaction tubes are at risk
of breaking due to the increased pressure when no organic molecule is present to react with any
hydrogen produced. The headspace gas was sampled as described in the experimental section
and a total of 0.12 mg of molecular hydrogen was measured, corresponding to 2.8 x 10-4 molal
aqueous hydrogen at the experimental conditions. The reaction conditions evidently produce
8

Page 9 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
molecular hydrogen in situ, but because the extent of conversion of the iron is not accurately
known the chemical yield of hydrogen cannot be determined.
Calculations based on thermodynamic parameters for nickel and iron in water were also
performed to gain insight into the thermodynamic stability of the metals under the experimental
conditions. The redox reactions considered, together with their equilibrium constants at 250°C
are given in Scheme 1.24-26 Magnetite, Fe3O4, and nickel oxide, NiO, were chosen as
representative metal oxidized species for this analysis. The revised Helgeson-Kirkham-Flowers
equation of state23 was used to generate an equilibrium nickel and iron stability field diagram as
a function of temperature and activity of aqueous hydrogen, Figure 1. For 240 µmol iron, and
allowing 100% conversion to magnetite and hydrogen, the thermodynamic calculations predict
oxidation of iron accompanied by the formation of 320 µmol of hydrogen. This hydrogen will
speciate between the aqueous and gas phases. Using the known equilibrium constant for this
speciation, Scheme 1, this results in 0.0039 molal hydrogen in the aqueous phase, indicated by
the closed circle in Figure 1. For a system of Fe, Ni, Fe3O4 and NiO with this quantity of
hydrogen, the thermodynamic equilibrium situation is characterized by complete essentially
oxidation of iron and no oxidation of nickel.
XRD analysis was performed on the solid products obtained after reaction of a mixture of
170mol of nickel and 450mol of iron, at 250°C for 1 hour, 40 hour and 68 hour reaction
periods in the absence of any organic. The solid products comprised a mixture that potentially
contained unreacted iron, unreacted nickel and the oxidized metal products. Nickel and iron have
overlapping peaks in XRD around 51°, Figure 2A. The solid products obtained from both the 1
hour and the 68 hour reactions had a broad peak in this region, Figure 2A, implying the presence
of unreacted nickel and/or iron in both samples. The earlier 1 hour reaction XRD pattern more
9

Green Chemistry
Page 10 of 38
View Article Online
DOI: 10.1039/C9GC00636B
Ni + H2O
NiO + H2
Log K250 = -3.79
3 Fe + 4 H2O
H2(aq)
Fe3O4 + 4H2
H2(g)
Log K250 = 2.36
Log K250 = 2.63
Scheme 1. Redox and other equations and their equilibrium constants at 250°C and 40 bar
used in the calculation of the iron/nickel in water stability field diagram of Figure 1.
10

Page 11 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
10
Fe + Ni
1
10-1
10-2
10-3
Fe3O4 + Ni
a[H2] (aq)
(molal)
0.0039 molal H2
250°C
10-4
10-5
Fe3O4 + NiO
200
225
250
275
300
325
350
Temperature (°C)
Fig 1. The stability fields of nickel, iron, iron oxide (Fe3O4) and nickel oxide (NiO), with respect
to temperature and activity of aqueous hydrogen. The calculated maximum concentration of
dissolved hydrogen for the hydrothermal experiments is indicated by the closed circle.
11

Green Chemistry
Page 12 of 38
View Article Online
DOI: 10.1039/C9GC00636B
A
B
Ni
Fe
68 hr
products
68 hr
products
1 hr
products
magnetite
Fe3O4
hematite
Fe2O3
51.0
51.5
52.0
52.5
53.0
53.5
54.0
20
30
40
50
60
70
80
90
angle 2 (degrees)
angle 2 (degrees)
Fig 2 (A) XRD patterns collected using Co-k radiation for (blue) the solid products of reaction
of nickel and iron for 68 hours at 250°C, (lighter grey) unreacted nickel powder, (darker grey)
unreacted iron powder, (red) the solid products of reaction of nickel and iron for 1 hour at 250°C.
The peaks are normalized to their maximum intensities. (B) XRD pattern for (upper blue) the
solid products of reaction of nickel and iron for 68 hours at 250°C, (middle black) a magnetite
(Fe3O4) standard, (lower brown) a hematite (Fe3O4) standard. The peaks indicated by diamond
symbols are associated with metallic nickel, the peaks indicated by circles are associated with
metallic iron.
12

Page 13 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
closely resembles that of iron, which is expected since there is more iron than nickel in starting
mixture, Figure 2A. After 68 hours, however, the XRD pattern more closely resembles that of
nickel, Figure 2A (complete XRD patterns are provided as Supplementary Information). This is
consistent with more rapid consumption (oxidation) of iron compared to nickel under the
experimental conditions over 68 hours. New peaks are found in the XRD patterns of the reaction
products, Figure 2B. Compared to the combined iron/nickel peak at 51°, the intensities of these
new peaks are higher for the 68 hour reaction than for the 1 hour reaction, consistent with more
oxidation of the metals to generate more product at the longer reaction time. The new peaks are a
close match to those of authentic magnetite and not to hematite, Figure 2B. Analysis of samples
that were milled to 20 m using EVA and the JCPDS database identified magnetite as the only
XRD detectable product (see Supplementary Information). Although the chemical yield of
magnetite cannot be determined from these experiments, these results strongly suggest oxidation
of iron under the experimental conditions to form molecular hydrogen and magnetite.
XRD analysis of a corresponding reaction of iron alone at 250°C for 70 hours (with no
nickel) gave product peaks identical to those from the nickel/iron experiments that were assigned
to magnetite. Reaction of nickel alone for 70 hours (with no iron) gave peaks corresponding to
unreacted nickel only, with no other XRD detectable reaction products on this timescale.
The cyclohexene reduction experiments, the measurements of hydrogen production, the XRD
product analyses and the thermodynamic calculations in Figure 1 together are consistent with
reaction in which metallic iron reacts with water to produce hydrogen and magnetite as reduction
and oxidation products, respectively, with catalytic reaction of molecular hydrogen with the
alkene occurring at the surface of metallic nickel. Nickel does not appear to oxidize under the
experimental conditions, consistent with the predictions of the thermodynamic calculations
13

Green Chemistry
Page 14 of 38
View Article Online
DOI: 10.1039/C9GC00636B
(Figure 1). As described in more detail below, catalytic activity is maintained even over very
long reaction time periods, which is also consistent with preservation of metallic nickel under the
reaction conditions.
Small amounts of an oxidized product benzene were also observed in the cyclohexene
reduction experiments. This is not entirely unexpected since the benzene is observed to form
only at very early reaction times, i.e., presumably before build-up of appreciable hydrogen
concentration via iron oxidation. Many hydrogenation catalysts are also dehydrogenation
catalysts.27
Reaction Stereochemistry
Reduction of cyclohexene does not provide information about the detailed mechanism of the
hydrogen addition reaction. Stereochemistry is frequently used as a probe for the mechanisms of
surface catalyzed reactions;16 therefore, 1,2-dimethylcyclohexene was studied since the reduction
products in this case are different if addition occurs to the same face (which gives cis-1,2-
dimethylcyclohexane as the product) or different faces (which gives trans-1,2-
dimethylcyclohexane as the product) of the double bond. Nickel and nickel alloys have
previously been shown to exhibit preferential formation of cis-isomers in the reduction of 1,2-
dimethylcycloalkenes at temperatures closer to ambient, consistent with syn-addition of H2.28
1,2-Dimethylcyclohexene was therefore studied as a stereochemical probe for hydrothermal
reduction. Reaction of 1,2-dimethylcyclohexene was performed with both nickel and iron at
14

Page 15 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
Scheme 2 Products from hydrothermal reduction of 1,2-dimethylcyclohexene with iron and
nickel at 250°C after 30 min compared to corresponding reduction at room temperature (25°C)
with hydrogen and Pd/C. The alkene isomer 2-methylene-1-methylcyclohexane is observed at
short reaction times under both sets of conditions.
15

Green Chemistry
Page 16 of 38
View Article Online
DOI: 10.1039/C9GC00636B
250°C. The reaction was quenched after 30 minutes in order to identify the primary reaction
products. Cis- and trans-1,2-dimethylcyclohexane were formed in a ratio of 58:42 under these
conditions, Scheme 2. This suggest that the reaction occurs primarily by syn-addition, but the
preference for the cis-isomer is only slight, suggesting other possible competing processes. The
trans-isomer cannot be a secondary product formed by isomerization of the cis-isomer. Reaction
of cis-1,2-dimethylcyclohexane under the same conditions results in less than 1% isomerization
to the trans-isomer on a timescale of 30 minutes.
Reduction of 1,2-dimethylcyclohexene was also performed using a conventional Pd/C
catalyst with molecular hydrogen in methanol at room temperature. In this case, 1,2-
dimethylcyclohexanes were formed in a cis-/trans- ratio of 25:75, i.e., the major product under
these conditions is the trans-isomer and the product of syn-addition is minor. With Pd/C and
molecular hydrogen at room temperature, a sample of the products was also taken at a very early
reaction time, 5 seconds, after which time 4% conversion of the alkene was observed. The
products observed at this very short time were the cis- and trans-1,2-dimethylcyclohexanes
formed in the same 25:75 ratio as at higher alkene conversions. However, the major product at
this short reaction time, 80% of the product mixture, was an isomer of the starting alkene, 2-
methylene-1-methylcyclohexane, Scheme 2. This strongly suggests that alkene isomerization
competes with hydrogenation. Isomerization of 1,2-dimethylcyclohexene in the presence of
palladium has previously been reported by Nishimura et al.29 Importantly, syn-addition of H2 to
this exocyclic alkene would be expected to form the more stable trans-1,2-dimethylcyclohexane.
The reliability of 1,2-dimethylcyclohexane as a stereochemical probe in the Pd/C reduction is
thus compromised by the formation of alkene isomers.
16

Page 17 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
Hydrothermal reduction of 1,2-dimethylcyclohexene with nickel and iron at 250°C was also
sampled at a reaction time with low conversion (30 minutes, 10% conversion). Similar to the
Pd/C reduction, 2-methyl-1-methylenecyclohexane was also observed as the major early-time
product in addition to the 1,2-dimethylcyclohexane isomers, suggesting alkene isomerization
occurs under these conditions as well. In addition, small quantities of alcohols and other alkene
isomers (9% of the total product mixture) were observed, presumably formed by reversible
addition of water to an alkene. Dehydration of these alcohols occurs rapidly under the reaction
conditions,10 and alcohols are not observed at greater alkene conversions, nevertheless, hydration
followed by dehydration represents another potential mechanism for isomerization. We conclude
that alkene isomerization either directly on the nickel surface, or by hydration/dehydration,
competes with syn-addition of hydrogen to alkenes under the reaction conditions, and that these
isomerization reactions significantly reduce the stereospecificity of the reaction.
Alkyne Reduction
To further explore the scope of the reduction reactions, and in an attempt to find a better
stereochemical probe of the reaction stereochemistry, two different alkynes were reduced.
Reaction of diphenylacetylene (DPA) at 250°C with nickel and iron gave a mixture of cis- and
trans-stilbene in an 80:20 ratio at all reaction times, from 5 minutes to 60 minutes. This
observation is consistent with mainly syn-addition of H2 since addition to the same face of the
alkyne forms the less stable cis-isomer. The trans-stilbene product is likely to be a primary
product rather than a secondary product formed via isomerization of the cis-stilbene, since
reaction of cis-stilbene under the same conditions for 5 minutes resulted in less than 1%
isomerization to the trans-isomer. As expected from the cyclohexene reduction
17

Green Chemistry
Page 18 of 38
View Article Online
DOI: 10.1039/C9GC00636B
a>c
k+H
0.032 min-1
a>t
c
k+H
0.0074 min-1
k+H
0.094 min-1
Fe/Ni
H2O 250°C
t
k+H
0.11 min-1
Scheme 3 Reaction pathways and pseudo-first order rate constants used to fit the time-
dependent data for hydrothermal reduction (addition of hydrogen) of diphenylacetylene (a) with
iron and nickel at 250°C, to form cis-stilbene (c) and trans-stilbene (t), and their follow-up
reductions, see Fig 3. The uncertainties in the rate constants are estimated to be ±15%, based
on repeated measurements.
18

Page 19 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
100
80
Mole %
60
40
20
0
0
10
20
30
40
50
60
Time (mins)
Fig 3 Time dependence of the relative concentrations or reactant and products for hydrothermal
reduction of diphenylacetylene with iron and nickel at 250°C. The solid lines represent the best
fits to the data according to the kinetic scheme and rate constants shown in Scheme 3.
19

Green Chemistry
Page 20 of 38
View Article Online
DOI: 10.1039/C9GC00636B
experiments, the stilbene products of this reaction are eventually further reduced to bibenzyl after
1 hour, at which point essentially 100% conversion from DPA to bibenzyl is achieved. In order
to determine the relative reactivities of the C-C double and triple bond -systems we conducted a
detailed kinetic analysis of time-dependent data.
Concentration versus time data are shown in Figure 3 for DPA, the two stilbene isomers and
the final product biphenyl. The time-dependence of these four species was fitted to the kinetic
scheme summarized in Scheme 3 using the COPASI software package.30 The best global kinetic
fit to the data is shown in Figure 3 and the corresponding values for the adjustable parameter
pseudo-first order rate constants are summarized in Scheme 3. To determine the best fit values of
the rate constants, minimization of error was performed using both the Levenberg/Marquardt and
simulated high-temperature annealing algorithms. Error minimization was also performed using
different starting values for the rate constants.
The overall rate constant for reduction of the alkyne, given as the sum of k+Ha>c and k+Ha>t, is
roughly one order of magnitude smaller than the rate constants for reduction of the cis- and
t
trans- alkene isomers, k+Hc and k+H , respectively. Note that reduction of the alkyne is slower than
reduction of the alkenes, even though alkyne reduction is generally more exothermic than alkene
reduction.31 This suggests that the difference in reactivity between the alkynes and alkenes is a
consequence of kinetic control of reactivity under the experimental conditions. Differences in
rates of reaction of alkenes and alkynes in various addition reactions are often observed under
conditions closer to ambient, and are attributed to the kinetic stability of the -bond electrons in
alkynes compared to alkenes. This is generally assumed to be a consequence of a shorter overall
bond between the sp-hybridized carbons of the alkyne.31 Slower hydrogenation of alkynes
compared to alkenes on a metal catalyst has been observed previously, and assigned to
20

Page 21 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
Scheme 4. Products observed for the hydrothermal reduction of 3-hexyne with iron and nickel
at 250°C after 5 minutes.
21

Green Chemistry
Page 22 of 38
View Article Online
DOI: 10.1039/C9GC00636B
differences in the adsorption of the two functional groups to the surface of the catalyst.32
Reduction of 3-hexyne was also studied further to explore the stereochemistry of the
hydrothermal hydrogenations, Scheme 4. The smaller ethyl substituents on the 3-hexyne
compared to the phenyl rings on DPA could result in changes in the adsorption of the alkyne to
the catalytic surface compared to DPA, which in turn could influence the rate of reaction or the
reaction stereochemistry. Indeed, reduction of 3-hexyne was observed to be much faster than
reduction of DPA. Conversion of 3-hexyne into products was >95% complete after only 5
minutes compared to ~13% conversion of DPA after the same time period. The reaction mixture
after this time period consisted of 72% alkenes and 27% hexane, Scheme 3. Of the alkenes, 68%
was cis-3-hexene, 6% trans-3-hexene, and 26% was trans-2-hexene. The primary 3-hexene
products were formed in a cis-/trans- ratio of 92:8, which is even larger than that found for
reduction of DPA. These results suggest that hydrogen addition is syn-, and that the preference
for syn-addition depends to some extent on molecular structure.
Electron Donating and Withdrawing Substituent Effects
The mechanistic studies described so far are consistent with a conventional Horiuti-Polyani
mechanism for hydrogenation at a metal surface.33 The reaction conditions, however are
reminiscent of several dissolving metal reductions, in which the hydrogen atoms are produced by
a metal that formally provides the electron and a Brønsted acid that formally provides a proton.34
In dissolving metal reductions the electron and proton transfer processes are often sequential.
Specifically, electron transfer is often a first step, forming a radical anion that is subsequently
protonated. Alternatively, the substrate could be protonated first, and then accept an electron. At
the reaction temperature, the pKw is ca. 11, compared to ca. 14 at ambient, i.e., the water
22

Page 23 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
autoionizes more efficiently.7 Therefore, the hydronium concentration is higher, and together
with the higher thermal energy, proton transfer reactions can occur that would not otherwise be
expected in water at ambient temperature and pressure (see, for example, ref 10). Although
molecular hydrogen is clearly formed when iron is heated to 250°C, in the presence of an organic
substrate an electron/proton transfer mechanism at the metal surface may compete with
molecular hydrogen formation and contribute to the overall reduction process.
The electron/proton transfer mechanisms common in dissolving metal reductions have ionic
intermediates.34 Evidence for ionic intermediates in a mechanism can be obtained from studies of
the influence of electron donating and withdrawing substituents. For example, in a recent study
of the reduction of styrenes using a homogeneous iron-based catalyst, styrenes with electron
withdrawing substituents were found to reduce much slower than those with donating groups. A
Hammett plot of the kinetic data was consistent with a proposed mechanism involving rate-
determining hydride transfer to form an intermediate that has a formal negative charge on the
benzylic carbon.35 We looked for similar substituent effects on the rates of the present
reductions.
Hydrothermal reduction of styrene was performed as described above and compared to the
corresponding reduction of p-methoxystyrene and m-trifluoromethylstyrene. Reaction of all three
styrenes resulted in formation of the corresponding ethylbenzene as the only detectable product
(92-99% chemical yield), Figure 4. Reaction was complete within 15 minutes for each styrene.
The kinetics of the reactions were determined from concentration versus time data, Figure 4. For
m-trifluoromethylstyrene, a steady-state was reached at 93% conversion whereas for the other
styrenes no starting material could be detected using our method of analysis. That the m-
trifluoromethylstyrene reaction reached a steady-state may be a consequence of the strong
23

Green Chemistry
Page 24 of 38
View Article Online
DOI: 10.1039/C9GC00636B
100
80
k+H = 0.44 min-1
k+H = 0.75 min-1
Mole %
60
40
20
0
MeO
MeO
k+H = 0.65 min-1
k-H = 0.07 min-1
CF3
CF3
0
5
10
15
20
25
30
35
Time (mins)
Fig 4. Kinetics of addition of hydrothermal reduction of three styrenes. The curves are best fits
to the experimental data (symbols) according to pseudo-first order kinetics. The pseudo-first
order rate constants for reduction (addition of hydrogen) are k+H. For m-trifluoromethylstyrene a
steady-state can be observed, therefore the rate constant for dehydrogenation (removal of
hydrogen), k-H, can also be determined. The uncertainties in the rate constants are estimated to
be ±15%, based on repeated measurements.
24

Page 25 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
electron withdrawing substituent stabilizing the vinyl -bond. The pseudo-first order rate
constants for addition of hydrogen to the styrenes to form the ethylbenzenes, k+H, and the
pseudo-first order rate constant for the reverse reaction for m-trifluoromethylstyrene, k-H, were
determined by fitting the data according to first order kinetics, and are summarized in Figure 4.
Both the electron donating and withdrawing groups increase the rate of reduction compared
to the parent styrene, k+H, although we note the effects are small: factors of 1.5 and 1.7 for the
withdrawing and donating substituents, respectively. This argues strongly against an ionic
intermediate playing any substantial role in the reaction, since the substituent effects would be
expected to be much larger and in opposite directions if that were the case. At the very least, the
formations of ionic intermediates are not connected to the rate determining steps in the reactions.
If benzylic radical intermediates were to be formed in a rate-determining steps then the
substituent effects would be expected to be even smaller due to the absence of charge. However,
the data are also not consistent with formation of radical intermediates, since it has been shown
that a m-trifluoromethyl substituent destabilizes a benzyl radical,36 in which case reduction
should be slower for the m-trifluoromethylstyrene, and we observe the opposite effect. In
addition, no coupling or disproportionation products could be detected in any of the reduction
reactions, arguing against the formation of freely diffusing radicals.
The stereochemical and electronic probes of the mechanism are consistent with reaction via
conventional metal-catalyzed formation of carbon-hydrogen bonds at the metal surface after
dissociative adsorption of hydrogen and the substrate to the metal surface.33,37 In one detailed
kinetic study of hydrogenation of styrene on palladium particles the authors concluded that
insertion of the first hydrogen atom was an important contributing factor to the overall rate of
reaction.38 Information on the relative rates of reduction of the various structures included in this
25

Green Chemistry
Page 26 of 38
View Article Online
DOI: 10.1039/C9GC00636B
work is summarized in Table 1, as the first half lives of the reactions under the same conditions.
These half-lives vary over roughly an order of magnitude. In principle, it should be possible to
find conditions in which one functional group could be reduced in the presence of another, if
their relative rates of reaction differ by one order of magnitude. However, the observed behavior
of the functional groups shows that the situation is not so simple. The alkyne DPA is the slowest
to react, slower than any of the alkenes, but the alkyne 3-hexyne reacts much more quickly,
faster than many of the alkenes. The rates of hydrogenation of -bonds on metal surfaces are
determined by a complex interplay between rates of chemisorption, desorption, and the rate of
hydrogen atom transfer to the alkene in the chemisorbed state.33,37 It is not obvious how to
predict relative rates of reactions a priori and thus predicting conditions for selective reduction
of the -bonds in these kinds of structures is not possible at this time.
Reaction Selectivity
Reduction of styrene shows that reduction of simple carbon-carbon double bonds can be
accomplished in the presence of an aromatic ring. Benzene rings can, however, be reduced all the
way to cyclohexane under hydrothermal conditions on much longer timescales, Scheme 5. At
250°C and 40 bars, complete reduction to cyclohexane requires 9 days. A pseudo-first order
reaction rate constant of 2.4  10-4 min-1 can be determined from time-dependent concentration
data collected over this time period (data not shown), Scheme 5. Slow reduction of benzene is
not surprising due to well-known low-reactivity of the aromatic ring, and because the reaction
presumably proceeds via a cyclohexadiene, which itself is a known hydrogen atom donor that
releases molecular hydrogen to reform benzene.39 Cyclohexadiene or cyclohexene cannot be
detected by gas
26

Page 27 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
Table 1. First half-lives for reduction at 250°C and 40 bar in water in the presence of iron and
nickela
Substrate
t1/2 (min)
1.6 ± 0.2
1.0 ± 0.2
0.9 ± 0.2
CF3
MeO
4 ± 2.0
< 5
6.3 ± 1.0
7.5 ± 1.0
18 ± 1.5
a The first half-lives and their uncertainties are derived from the rate constants obtained from
analysis of time-dependent concentration data, see Figs 2 and 3 and Scheme 3. For cyclohexene,
the half-life estimate is based on one experiment take to 55% conversion, the uncertainty in this
case is based on analysis of the reproducibility of repeated measurements. For 3-hexyne the half-
life of <5 mins is based on experiments that show >95% conversion in this time period.
27

Green Chemistry
Page 28 of 38
View Article Online
DOI: 10.1039/C9GC00636B
Scheme 5. Proposed pathway for hydrothermal hydrogenation of benzene; note the diene and
alkene structures cannot be isolated. Conversion of benzene into cyclohexane is observed only
after 9 days. The uncertainty in the pseudo-first order reaction rate constant, k+H, is estimated to
be ±15%.
28

Page 29 of 38
Green Chemistry
View Article Online
DOI: 10.1039/C9GC00636B
Table 2. Selective hydrothermal reductions with iron and nickel at 250°C
Reaction
Time
Conversion
Selectivity
(%)
Starting Compound
Major Product
(%)
1 hour
100
95
15 min
93
82
97
O
O
70h
70
OH
OH
29