Comparison of alkene hydrogenation in carbon nanoreactors of different diameters: Probing the effects of nanoscale confinement on ruthenium nanoparticle catalysis

Article abstract of DOI:10.1039/c7ta03691d

The catalytic properties of ruthenium nanoparticles (RuNPs) supported in carbon nanoreactors of different diameters-single walled carbon nanotubes (SWNTs, width of cavity 1.5 nm) and hollow graphitised nanofibers (GNFs, width of cavity 50-70 nm)-were evaluated using exploratory alkene hydrogenation reactions and compared to RuNPs adsorbed on the surface of SWNT or deposited on carbon black in commercially available Ru/C. Supercritical CO₂ is shown to be essential to enable efficient transport of reactants to the catalytic RuNPs, particularly for the very narrow RuNP@SWNT nanoreactors. Though the RuNPs in SWNT are observed to be highly active, they simultaneously reduce the accessible volume of very narrow SWNTs by 30-40% resulting in lower overall turnover numbers (TONs). In contrast, RuNPs confined in wider GNFs were completely accessible and demonstrated remarkable activity compared to unconfined RuNPs on the outer surface of SWNTs or carbon black. Control of the nanoscale environment around the catalytic RuNPs significantly enhances the stability of the catalyst and influences the local concentration of reactant molecules in close proximity to the RuNPs, illustrating the comparable importance of confinement to that of metal loading and size of NPs in the catalyst. Interestingly, extreme spatial confinement also appeared not to be the best strategy for controlling the selectivity of hydrogenations in a competitive reaction of norbornene and benzonorbornadiene, with wider RuNP@GNF nanoreactors displaying enhanced selectivity for the hydrogenation of the aromatic group containing alkene (benzonorbornadiene). This is attributed to the presence of nanoscale graphitic step-edges within the GNF making them an attractive alternative to the extremely narrow SWNT nanoreactors for preparative catalysis.

Full text of DOI:10.1039/c7ta03691d

View Article Online
View Journal
Journal of
Materials Chemistry A
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: M. Aygun, C. T.
Stoppiello, M. A. Lebedeva, E. F. Smith, M. del Carmen Gimenez-Lopez, A. N. Khlobystov and T. W.
Chamberlain, J. Mater. Chem. A, 2017, DOI: 10.1039/C7TA03691D.
This is an Accepted Manuscript, which has been through the
Volume
4 Number 1 7 January 2016 Pages 1–330
Royal Society of Chemistry peer review process and has been
accepted for publication.
Journal of
Materials Chemistry A
Accepted Manuscripts are published online shortly after
Materials for energy and sustainability
www.rsc.org/MaterialsA
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 2050-7488
PAPER
Kun Chang, Zhaorong Chang et al.
Bubble-template-assisted synthesis of hollow fullerene-like
MoS2 nanocages as a lithium ion battery anode material
rsc.li/materials-a

Page 1 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
Comparison of alkene hydrogenation in carbon nanoreactors of different
diameters: probing the effects of nanoscale confinement on ruthenium
nanoparticle catalysis
Mehtap Aygun,a Craig T. Stoppiello,a Maria. A. Lebedeva,a,b Emily F. Smith,c Maria del Carmen
Gimenez-Lopez,a Andrei N. Khlobystov.a,c and Thomas W. Chamberlaind*.
a School of Chemistry, University of Nottingham, University Park, Nottingham, UK, NG7 2RD.
b Department of Materials, University of Oxford, 16 Parks Road, Oxford, OX1 3PH, UK.
c Nanoscale & Microscale Research Centre, University of Nottingham, University Park, Nottingham,
UK, NG7 2RD.
d Institute of Process Research and Development, School of Chemistry, University of Leeds, Leeds, UK,
LS2 9JT.
*T.W.Chamberlain@leeds.ac.uk
Abstract
The catalytic properties of ruthenium nanoparticles (RuNPs) supported in carbon nanoreactors of different
diameters – single walled carbon nanotubes (SWNTs, width of cavity 1.5 nm) and hollow graphitised
nanofibers (GNFs, width of cavity 50-70 nm) – were evaluated using exploratory alkene hydrogenation
reactions and compared to RuNPs adsorbed on the surface of SWNT or deposited on carbon black in
commercially available Ru/C. Supercritical CO2 is shown to be essential to enable efficient transport of
reactants to the catalytic RuNPs, particularly for the very narrow RuNP@SWNT nanoreactors. Though the
RuNPs in SWNT are observed to be highly active, they simultaneously reduce the accessible volume of very
narrow SWNTs by 30-40 % resulting in lower overall turnover numbers (TONs). In contrast, RuNPs confined in
wider GNFs were completely accessible and demonstrated remarkable activity compared to unconfined RuNPs
on the outer surface of SWNTs or carbon black. Control of the nanoscale environment around the catalytic
RuNPs significantly enhances the stability of the catalyst and influences the local concentration of reactant
molecules in close proximity to the RuNPs, illustrating the comparable importance of confinement to that of
metal loading and size of NPs in the catalyst. Interestingly, extreme spatial confinement also appeared not to
be the best strategy for controlling the selectivity of hydrogenations in a competitive reaction of norbornene
and benzonorbornadiene, with wider RuNP@GNF nanoreactors displaying enhanced selectivity for the
hydrogenation of the aromatic group containing alkene (benzonorbornadiene). This is attributed to the
presence of nanoscale graphitic step-edges within the GNF making them an attractive alternative to the
extremely narrow SWNT nanoreactors for preparative catalysis.
1

Journal of Materials Chemistry A
Page 2 of 22
View Article Online
DOI: 10.1039/C7TA03691D
Keywords Nanoreactor, carbon nanotubes, confinement, hydrogenation, heterogeneous catalysis.
1. Introduction
Gaining control of chemical reactions in order to improve the yield of a particular product and/or change the
reaction pathway is a great challenge in chemical synthesis. One method commonly used to achieve this is to
perform reactions in sterically confined environments using materials such as zeolites, porous silica or alumina,
molecular cages and carbon nanomaterials to act as nanoscale reaction vessels or nanoreactors.1-9 All of these
systems have nanosized pores, holes or channels of different geometries and diameters which are capable of
accommodating reactants and affecting the distribution of products by imparting some form of steric effect on
the transition states or intermediates of reactions.
Carbon nanostructures (CNS) have recently become of great interest for use as nanoreactors in a variety of
different catalytic chemical reactions as they are robust, chemically inert, and available in a large range of
well-defined pore shapes and sizes. Most importantly, due to recent advances in production, CNS are now
readily available in large quantities at low cost which opens up the potential for application in large-scale
preparative syntheses for the first time.10-12
In addition to acting as reaction vessels and templates to the formation of specific products, CNS also make
the ideal support materials for metal nanoparticle (MNP) catalysts. By immobilizing catalytic NPs inside CNS it
is possible to combine all of the advantages of nanoreactors, offering control of the size and shape of the
reaction volume, with the inherent advantages of heterogeneous catalysis, i.e. enhanced stability and
recyclability of metal nanoparticles.13-17 As a result a significant number of studies have reported the
application of multi-walled carbon nanotubes (MWNTs) and hollow graphitised nanofibers (GNFs), with
internal channels ranging from 5-50 nm, as nanoscale reaction vessels and flow reactors for catalytic chemical
reactions. In which their commensurate hollow structure facilitates encapsulation of the metal nanoparticles
inside the nanotube and provides the perfect environment for reactions to occur within a strictly controlled
nanoscale volume.9, 17-22 Serp et al. performed the selective hydrogenation of cinnamaldehyde to cinnamyl
alcohol over a bimetallic Pt_Ru catalyst confined inside MWNTs and compared the metal catalyst activity with
both the same sized free standing MNPs and MNPs supported on the outside surface of MWNTs.17
A
significant increase in catalytic performance with higher turnover frequency and selectivity for cinnamyl
alcohol was observed for the catalyst inserted in MWNTs as a result of the confinement and enrichment of
reactant concentration inside the nanotubes due to stronger interactions between the molecules and the
internal surface of the carbon nanostructures. Pan et al. reported that Rh nanoparticles confined inside carbon
nanotubes substantially enhanced ethanol conversion compared with the same catalyst located outside of the
nanotubes.9 Moreover, platinum nanoparticles both inserted into GNFs and adsorbed on to the outside
surface were probed in the competitive hydrosilylation reaction of phenylacetylene by Solomonsz et al.,
demonstrating significant changes in the selectivity of reactions of aromatic and aliphatic molecules within the
nanoreactor.18
2

Page 3 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
The latest studies reveal that the size and shape of the nanotube channel enhances the stability and
selectivity of the confined NP catalysts allowing strict control of the nanoparticle size, functionality and
reactivity by providing stabilization to the NPs, preventing aggregation into larger particles and bulk metal.23-26
Less is known about the use of narrower SWNTs and double-walled carbon nanotubes (DWNTs), with an
internal mean diameter of ~1-2 nm, however it has been shown that they enable the formation of small, highly
stable metal nanoparticles.25 The advantageous nature of MNPs@SWNT catalysts, in which the cavity
dimensions are commensurate with the size of small organic reactant molecules, is that they impart significant
steric influence on reaction pathways thus exhibiting greater effects on the products of reactions than wider
MWNTs. We recently carried out hydrogenation reactions of alkenes catalyzed by RuNPs confined within
extremely narrow SWNTs and observed very high yields of the products as a result of an enrichment of
reactant species inside the nanonoreactor due to confinement effects.27 Narrow DWNTs (nanotube diameter
of 1-1.5 nm) were utilised to stabilise Pd-V bimetallic nanoparticles and the resultant catalyst was shown to be
more active for benzene hydroxylation than the same NPs in wider, and hence less confined MWNT (nanotube
diameter = 4-8 nm).28 Similar enhancements in activity as a result of extreme nanotube confinement have also
been reported for sub-nanometre titania NPs in DWNTs for propylene epoxidation29 and Re NPs in DWNT for
benzene hydroxylation.30
There are also a few examples of non-catalysed chemical reactions within SWNT, such as a study by Miners et
al. who reported the effect of SWNT diameter on the selectivity of N-phenylacetamide bromination within
SWNT and showed that the inner cavity of the nanotube changed the regioselectivity and activity of a
bromination reaction.10, 31-32 These studies demonstrate that the extreme confinement imposed by the unique
reaction environment of SWNT-based nanoreactors can significantly alter the selectively and rate of chemical
reactions. In addition the SWNT support changes the chemical and physical properties of the NP catalyst,
allowing strict control of the NP functionality and reactivity as well as NP size.27
Thus, confinement of catalytic processes in very narrow carbon nanoreactors has been shown to control the
size and shape selectivity of hydrogenation reactions. In this study we explore the effects of such extreme
confinement on the mass transport of reactants and products into and out of SWNT-based catalytic
nanoreactors and compare this to GNFs, which are wider, and therefore more accessible to reactants and
crucially, in contrast to MWNTS, have structured internal channels which can also influence chemical
reactions.
2. Experimental
SWNTs, GNFs and C60 were purchased from Helix Material Solutions (Arc-discharge, USA), Pyrograf Products
Inc (PR19, chemical vapor deposition, USA) and SES Research (USA), respectively. All other reagents and
solvents were purchased from Sigma-Aldrich (UK) and used without further purification. All of the glassware
required to perform the experiments was thoroughly cleaned with ‘aqua regia’ (concentrated hydrochloric and
nitric acids (3:1)) and rinsed with deionized water prior to use.
3

Journal of Materials Chemistry A
Page 4 of 22
View Article Online
DOI: 10.1039/C7TA03691D
Catalyst preparation:
SWNTs were annealed at 380 °C for 20 minutes to open their termini and remove any residual amorphous
carbon from the internal cavities; a 20-30 % weight loss was observed prior to use. Average SWNT length after
thermal treatment was reported previously to be 2.43 ± 0.85 µm.33 GNFs were thermally annealed at 450 °C
for 1 hour prior to use. The average length after thermal treatment was measured by TEM to be 15.34 ± 12.10
µm, (see Electronic Supplementary Information file (ESI) Figure S-2).
RuNPs@SWNT and RuNPs@GNF - The metal carbonyl precursor, Ru3(CO)12 (1.05 mg/0.21 mg, masses
equivalent to the wt.% of Ru metal required for SWNT and GNF respectively) were combined with freshly
opened SWNTs (10 mg) or GNFs (10 mg) in a quartz ampoule, and sealed under vacuum (10-6 bar) heated at
140 °C for 3 days. After 3 days, the sample inside the quartz ampoule was cooled by immersing in an ice bath.
The sample was then removed from the ampoule, sonicated in tetrahydrofuran (10 mL) for 15 minutes, then
filtered through a PTFE membrane filter (pore size 0.2 μm) and repetitively washed with tetrahydrofuran (3 x
10 mL). After washing, the sample was sealed in a quartz ampoule under an argon atmosphere and heated at
600 °C for 2 hours to decompose the metal carbonyl into the desired pure metal nanoparticles.
RuNPs/(C60@SWNT) - Freshly opened SWNT (10 mg) and C60 (20 mg) were sealed under vacuum (10-6 bar) in a
quartz ampoule and heated at 500 °C for 2 days. After 2 days, the sample was removed from the ampoule and
sonicated in toluene (10 mL) for 15 minutes. The sample was then filtered through a nylon membrane filter
(pore size 0.2 μm) and repetitively washed with toluene (3 x 10 mL). After washing, the C60@SWNT (10 mg)
and Ru3(CO)12 (0.65 mg, a mass equivalent to the wt.% of Ru metal required) were combined in a quartz
ampoule and sealed under an argon atmosphere and heated at 600 °C for 2 hours decompose the metal
carbonyl into the desired pure metal nanoparticles.
HRTEM analysis was performed on a JEOL 2100 Field emission gun microscope with an information limit of
0.12 nm at 100 kV or 200kV. High resolution scanning transmission microcpy (HRSTEM) images were acquired
using the JEOL digital STEM system. Samples (RuNPs@SWNT, RuNPs/C60@SWNT, RuNPs@GNF and Ru/C) were
prepared for TEM analysis by dispersing the materials in HPLC grade iso-propanol using ultra-sonication, then
drop casting the resultant suspension onto a lacey carbon film coated copper grid. 1H NMR spectra were
recorded using a Bruker DPX300 NMR spectrometer. 1H NMR spectra were taken in CDCl3 and were referenced
to residual trimethysilane (TMS) (0 ppm) and reported as follows: chemical shift, multiplicity (s = singlet, d =
doublet, t = triplet, dd = doublet of doublet, m = multiplet). TGA analysis was performed on a TA Instruments
TGA-SDTQ600 analyser. Samples for TGA analyses were heated in an inert atmosphere up to 1000 °C with a
heating rate of 10 °C/min. The powder X-ray diffraction (XRD) patterns were obtained using a PANanalytical
X’Pert PRO diffrctometer equipped witha Cu-Ka radiation source operating at 40 kV and 40 mA, with 0.05252°
step size and 5925.18 second step time. Surface area analysis was performed using the Brunauer–Emmett–
Teller (BET) method based on adsorption data in the relative pressure (P/Po) range 0.02 to 0.22 by measuring
nitrogen sorption isotherms of the samples (50 mg) at -196 °C on a Micromeritics ASAP 2020 sorptometer. X-
ray photoelectron spectrocopy (XPS) samples were analysed using the Kratos AXIS ULTRA with a
4

Page 5 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
mono-chromated Al kα X-ray source (1486.6eV) operated at 10 mA emission current and 12 kV anode
potential (120 W).
Benzonorbornadiene (2a) synthesis by Diels Alder reaction
Benzonorbornadiene was synthesized according to the previously reported procedure.34 1,2-Dibromobenzene
o
(5.00 g, 21.45 mmol) and cyclopentadiene (1.42 g, 21.45 mmol) were stirred in toluene (13 mL) at 0 C under
Ar. n-BuLi (12 mL, 1.78M in hexanes, 21.45 mmol) was added to this solution dropwise over 30 min during
which the reaction solution became first yellow then cloudy white. After an additional 10 min at 0 oC the
mixture was allowed to warm to room temperature, stirred overnight and treated with H2O (20 mL) and
extracted with hexanes (3 × 30 mL). The organic layer was dried over MgSO4, filtered, and concentrated to
obtain yellow oil. The product was purified by chromatography on silica gel eluting with hexanes to provide a
clear and colourless oil (1.6 g, 0.11 mmol, 65 %).
Benzonorbornadiene (1,4-dihydro-1,4-methano-naphthalene) (2a); 1H NMR (300 MHz, 297 K, CDCl3, δ, ppm):
7.24-7.21 (m, 2 H), 6.95-6.92 (m, 2H), 6.80 (m, 2 H), 3.90 (m, 2 H), 2.34-2.30 (m, 1H), 2.26-2.23 (m, 1H).
Hydrogenation reactions using lab glassware:
All reactions were performed in a pyrex pressure tube (10 bar) with a stirring bar. RuNPs@carbon nanoreactor
(in each case an amount equivalent 0.0017 mmol% of Ru in the reaction mixture) was suspended in
cyclooctane (1 mL) in the bottom of the tube and the resulting mixture was saturated with H2 by bubbling a
mixture of 10 % H2 and 90 % Argon gas (1 bar) through the solution at 25 °C for 30 minutes. The alkene (0.05
mL) was then added to the resulting H2 saturated solution. The tube was sealed and the resultant suspension
was heated at 110 °C for 24 hour. After this, the reaction was stopped and the mixture was cooled down to
room temperature. The resultant material was analyzed by 1H NMR spectroscopy.
Hydrogenation reactions using a high pressure scCO2 batch reactor:
In general, RuNPs@carbon nanoreactor (0.0017 mmol% of Ru) and alkene substrate (1 mmol) were put into a
high pressure reactor (10 mL). The reactor volume was degassed thoroughly with H2 for 30 min. Then, the
reactor was sealed and pressurized with H2 (10 bar) and heated to 110 °C by adding CO2 (100 bar) and left for
24 hour. At the end of the reaction, the reactor was cooled and depressurized. The resultant material was
analyzed by 1H NMR spectroscopy. All catalytic reactions were performed in duplicate, and the yields given are
averages of the two experiments.
Norbornane (1b); 1H NMR (300 MHz, 297 K, CDCl3, δ, ppm): 2.21 (m, 2 H), 1.50 (m, 4H), 1.18 (m, 6 H).
1,4-Methano-1,2,3,4-tetrahydronaphthalene (2b); 1H NMR (300 MHz, 297 K, CDCl3, δ, ppm): 7.30-7.10 (m, 4 H,
ArH), 3.50-3.35 (m, 2H, CH), 2.10-1.50 and 1.50-1.10 (m, 6H, CH2).
The use of high pressure scCO2 batch reactor:
5

Journal of Materials Chemistry A
Page 6 of 22
View Article Online
DOI: 10.1039/C7TA03691D
In general, the autoclave reactor was loaded with RuNPs@SWNT and norbornene substrate (1 mmol) and an
O-ring was placed to seal the reactor. The reactor volume was degassed thoroughly with H2 for 30 min by
opening the outlet valve. Then, the reactor was sealed and pressurized with H2 (10 bar) and heated initially to
40 °C in order to add the CO2 (50 bar) as a supercritical phase using a pickle pump and then heated slowly to
110 °C to make sure that overall pressure is 100 bar by adding more scCO2 at this temperature. The reaction
was then left for 24 hour and cooled under room temperature to depressurize slowly by opening outlet valve.
3. Result and Discussion
In our study, a simple and efficient approach for catalyst synthesis in the inner cavity of carbon nanotubes is
used to create very active Ru nanoparticles within an extremely constrained carbon nanoreactor
environment.35 In this approach, Ru3(CO)12 molecules inserted from the vapour phase into the freshly opened
SWNT are decomposed to form Ru nanoparticles within the nanotube (RuNPs@SWNT). High resolution
transmission electron microscopy (HRTEM) imaging of the resultant hybrid nanomaterial RuNPs@SWNT
reveals the shape, size and location of Ru nanoparticles inside the nanoreactor. HRTEM confirmed that the
RuNPs were located mostly inside the SWNT, where the nanotube sidewall stabilises and templates the
formation of NPs resulting in small, well defined particles with a narrow size distribution, (dNP = 0.74 ± 0.18 nm,
Table 1). A combination of thermogravimetric analysis (TGA) and inductively coupled plasma optical emission
spectrometry (ICP-OES) was used to determine the precise loading of Ru in RuNPs@SWNT showing the
material to be 1.6 % Ru by weight. This represents the maxiumum loading of Ru nanoparticles into SWNT with
the remaining Ru material deposited and subsequently washed from the outer walls of the nanoutbes prior to
nanoparticle formation, an essential step to ensure no unconfined nanoparticles are present to take part in the
reaction (see supporting information file Figures S-13 and Table S-1). However, TEM reveals a small portion
(<10 %) of NPs are located on the outside of the nanotubes where they do not benefit from the same
templating effect and therefore are larger and less uniform (dNP = 2.49 ± 0.85 nm (Figure 1a and Figure S-1 in
Electronic Supplementary Material). Energy dispersive X-ray spectroscopy (EDX), XPS, and lattice spacing
analysis of HRTEM images and powder XRD are consistent with a hexagonally close packed structure of
metallic the RuNPs (see Figure 1 and ESI for full details).36
6

Page 7 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
7

Journal of Materials Chemistry A
Page 8 of 22
View Article Online
DOI: 10.1039/C7TA03691D
Figure 1. a-c) HRTEM images of RuNPs@SWNT loaded with 1.6 % of Ru by wt. Black and white arrows indicate
RuNPs inside and outside of the SWNT respectively, and enlarged region (c) showing the Ru [002] lattice
spacing (d = 0.21 nm) of an individual, hcp structured, metallic NP which corresponds to the 2θ peak centred at
42.2° in the XRD (see ESI). d) HRTEM image of RuNPs/C60@SWNT loaded with 3.1 % of Ru by wt. after filling the
SWNT with C60, and e), HRTEM image of the C60@SWNT support material showing the internal channels of the
SWNT are completely full of C60. Enlarged region (inset) shows only single C60@SWNT structure in which a 0.3
nm van der Waals gap can be observed between C60 molecules and the SWNT wall confirming that there is no
space for reactants or Ru NP precursors to fit. White arrows show RuNPs located on the outside of the SWNT
and the black arrows show C60 molecules. f) The size distribution of RuNPs located inside the SWNT highlights
the narrow diameter distribution of the RuNPs (blue), and g), the size distribution of RuNPs on the outside of
the SWNT (red) is observed to be greater than the confined NPs inside and comparable to the NPs located on
the outside of the SWNT in RuNPs@SWNT. Scale bars: 2 nm in all images.
Table 1. Nanoparticle sizing data and theoretical active Ru surface area for all carbon nanoreactor supported RuNP
catalysts.
Average size of
RuNPs (nm)
Theoretical active Ru
surface area (m2/g of
catalyst)a
Catalyst
0.74 ± 0.18
2.56 ± 0.62
3.58 ± 1.14
6.63 ± 2.45
RuNPs@SWNT
RuNPs/C60@SWNT
RuNPs@GNF
Ru/Cc
7.46
4.31
1.00
2.69
a The active Ru surface area of the RuNPs catalysts was calculated assuming all of the NPs to be the average
diameter measured by TEM (see SI file for full details).
The RuNPs@SWNT catalyst was then investigated using exploratory alkene hydrogenation reactions and as the
RuNPs are located predominantly within the nanotube channel all reactions are assumed to be performed
under the effects of extreme confinement. When selecting an appropriate reactant it is important that the van
der Waals size of the reactant molecules is smaller than the 1.5 nm diameter of the SWNT channel. As the
hydrogenation of cycloalkenes and butanal in continuous flow reactors has recently been demonstrated in
nanotubes, in this study we selected
a bicyclic alkene norbornene (1a) and a tricyclic alkene
benzonorbornadiene (2a), both of which have non-planar angular shapes, and in the case of 2a an additional
aromatic ring increasing the steric bulk as compared to 1a (Scheme 1).27
The RuNPs@SWNT catalyst was initially tested in the hydrogenation reaction of norbornene using
molecular hydrogen in a conventional organic solvent, cyclooctane, at atmospheric pressure using laboratory
glassware. Very low catalytic activity was observed for RuNP@SWNT under these conditions (a 10 % yield of
8

Page 9 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
norbornane in 24 h), and is attributed to the restricted space in the narrow nanotube channels hindering
access of the solvent/reactants to the confined metal catalyst. Typically, the use of supercritical CO2 (scCO2)
can eliminate such mass transfer problems in nano- or microporous structures,27 therefore, a high pressure
scCO2 batch reactor system was utilized in which the excellent diffusivity and mass transfer properties of scCO2
are exploited to efficiently deliver the reagents to the RuNP catalyst surface within the narrow nanoreactors
(Figure 2).
Figure 2. A schematic diagram of the scCO2 hydrogenation batch reactor.
Test hydrogenation reactions were performed using the scCO2 batch reactor in the presence of RuNPs@SWNT,
and norbornene showed a higher TON (total number of product molecules formed per available Ru active site,
see supporting information for details of how this was calculated),27 than the larger benzonorbornadiene
under identical reaction conditions (Scheme 1 and Table 2). Control reactions using as-received SWNTs showed
no reactivity despite EDX and TEM (see supporting information Figure S-1) showing the presence of residual
Ni/Y catalyst from nanotube synthesis implying all of the nickel is completely passivated by layers of graphitic
carbon shells (Table 2).
Scheme 1. Hydrogenation of norbornene and benzonorbornadiene in the presence of RuNPs@SWNT.
Table 2. Hydrogenation reactions of norbornene and benzonorbornadiene in the presence of RuNPs@SWNT,
RuNPs/C60@SWNT, RuNPs@GNF and commercial Ru/C using a high pressure scCO2 batch reactor.
9

Journal of Materials Chemistry A
Page 10 of 22
View Article Online
DOI: 10.1039/C7TA03691D
Yieldc of Products (%) / TONd
Ratio of TONs for
Catalyst
1b:2b
1b
2b
SWNTa
0
N/A
N/A
RuNPs@SWNT
C60@SWNT
RuNPs/C60@SWNT
GNFb
91 / 2959
0
61 / 1983
N/A
1.5:1
N/A
12 / 675
0
7 / 393
N/A
1.7:1
N/A
RuNPs@GNF
Ru/C
46 / 11216
51 / 7428
17 / 4145
21 / 3058
2.7:1
2.4:1
Reaction conditions: alkene (1 mmol), catalyst (equivalent of 0.0017 mmol of ruthenium in the reaction mixture), H2 (10
bar), scCO2 (100 bar), 24 h, 110 °C. a SWNT was annealed at 380 °C for 20 minutes to open their termini prior to use. b GNF
was annealed at 450 °C for 1 hour prior to use. c Yield determined by 1H NMR with an error of ± 2 %. dThe turnover number
(TON) was calculated as the ratio of the number of molecules of substrate consumed in the reaction per number of
available Ru active sites in catalyst used in the reaction (see SI file for details of the calculation of the theoretical number of
active Ru sites in each catalyst).
Crucially, to verify the precise location of the reaction it was important to investigate the effect of the small
portion of larger RuNPs located on the outside of the SWNT which were more readily accessible than the
RuNPs confined within the nanotube channel. A control material was therefore synthesised, in which RuNPs
are located solely on the outside of SWNT, RuNPs/C60@SWNT. This material was prepared by filling of the
internal cavities of the SWNT support material with C60 to block the channels prior to exposure to the RuNP
precursor, Ru3CO12 vapour (see ESI Figure S-4a for HRTEM of the C60@SWNT support material). HRTEM and gas
absorption measurements (Figure S-14-15), which reveal an absence of 1-2 nm pores in the C60@SWNT
support, agree with the fact that the C60 molecules completely block the inner channels of the SWNTs, such
that the Ru3CO12 can only deposit on the outside of the nanotubes and subsequent thermal treatment causes
decomposition of the precursor resulting in RuNPs formation exclusively on the outer surface of the SWNT
(Figures 1d and S-4b). Removal of the templating effect of the inner nanotube cavity during NP formation
resulted in the Ru nanoparticles of diameter 2.56 ± 0.66 nm on the nanotube surface exhibiting a wider size
distribution significantly larger as compared to nanoparticles formed inside, but comparible to the NPs located
on the outside of, nanaotubes in RuNPs@SWNT (c.f. average NP sizes of 0.74 ± 0.18 nm and 2.49 ± 0.85 nm for
NPs inside and outside respectively). This allowed comparison of the catalytic activity of RuNPs located on the
outside of the SWNT (RuNPs/C60@SWNT) with RuNPs located inside the nanotube (RuNPs@SWNT). For
reference C60@SWNT exhibited no catalytic activity when tested in scCO2 hydrogenation reactions, (Table 2).
In comparison with RuNPs@SWNT, RuNPs/C60@SWNT showed very low catalytic performance in
scCO2 reactions implying that the larger NPs on the outer walls of the nanoreactor are less active than the
10

Page 11 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
smaller RuNPs confined within the nanoreactor.37-38 Interestingly, the RuNPs on the outside of the C60@SWNT
support (dNP = 2.56 nm) are similar in size to the RuNPs inside the GNF in the RuNPs@GNFs catalyst (dNP = 3.58
nm), see below for details, but give siginificantly lower reaction yields, (Table 2). It is hypothesised, therefore,
that the lack of activity of the RuNPs/C60@SWNT catalyst is most likely a result of the lack of confinement for
the reactions, which cannot create a similarly high local concentration of the reactant molecules around the
catalyst nanoparticles in RuNPs/C60@SWNT as in RuNPs@SWNT (and RuNPs@GNF) leading to a lower yield of
the products in the case of the former material. This is consistent with the previously reported examples of
enhanced local concentration of reactants inside nanotubes resulting in higher yields and is consistent with our
observation of low norbornene conversion (~10 %) in the reaction catalyzed by RuNPs@SWNT performed in
cyclooctane solvent in which the reaction only takes place on the accessible Ru nanoparticles located on the
outside surface of SWNT in the absence of high pressure scCO2 (c.f. RuNPs/C60@SWNT showed 6% norbornene
conversion to norbornane using the same conditions).18-19, 26, 31-32 It is therefore concluded that the RuNPs
located inside the carbon nanoreactor are significantly more active and are responsible for the majority of
product formation, and thus, the contribution of the minority of RuNPs on the outside of the SWNT (~10% by
HRTEM) and GNF (~7% by STEM, Figure 3c) for RuNPs@SWNT and RuNPS@GNF catalysts respectively, is
neglibible (see ESI for details).
The general observed trend of higher TONs for norbornene, compared to benzonorbornadiene, is
most likely due to the steric bulk of the additional benzyl group hindering adsorption of benzonorbornadiene
on the Ru NP surface, as well as potentially an electron withdrawing effect of the additional benzene ring
compared to norbornene.39 Interestingly, however, there is a significant difference in the relative TONs of
reactions of 1b and 2b for each of the SWNT catalysts, highlighted by considering the ratio of TONs, Table 2,
implying the shape of the catalyst affects the reactions in an appreciable way. The extremely confined
RuNPs@SWNT system, where the ratio of TONs of 1b:2b is dramatically lower than for the unconfined
RuNPs/C60@SWNT system, appears to have a higher affinity for 2b, potentially due to favourable π-π
interactions between the aromatic group and the nanotube sidewall. Cleary understanding the nature and
extent of the confinement in these systems is crucial to selecting the correct environment for a specific
reaction.
To fully understand the impact of the restricted reaction space within the SWNT,27 hydrogenation of
norbornene and benzonorbornadiene was performed with catalysts which provide lower (RuNPs@GNF) and
no confinement (RuNPs/carbon black) of reactants around the catalytic nanoparticles. To achieve this, GNFs
were used, which have an inner diameter of 52 ± 13 nm and an outer diameter of 99 ± 25 nm, and are, like
MWNTs, significantly wider than SWNTs, allowing reactants in (and products out) more readily and thus can be
considered as providing a lower level of confinement. In addition, unlike the entirely smooth MWNTs, GNFs
have a unique step-edge internal structure, which stabilses the RuNPs and has the capacity to impart
additional confinement effects on reactions which will be probed in this study. RuNPs within the GNF
(RuNPs@GNF) were formed using similar conditions to RuNPs@SWNT, however due to the significantly lower
surface area of the GNFs only 1% Ru by weight was used, to ensure that RuNPs were only formed in the cavity
11

Journal of Materials Chemistry A
Page 12 of 22
View Article Online
DOI: 10.1039/C7TA03691D
of the GNFs.18 In identical fashion to SWNT, it is the internal structure of the GNF which templates the
formation of the RuNPs, stabilising the NPs and controlling their size and shape. HRTEM imaging confirms that
RuNPs are located solely inside the GNF, at the step-edges formed by rolled-up graphitic sheets, with a mean
nanoparticle diameter of 3.58 ± 1.14 nm (Figure 3a, Table 1). HRTEM and XRD analyses (see ESI for full details)
confirm the metallic nature of the RuNPs and STEM, (Figures 3c and S-3), approximates the amount of RuNP
material on the outer surface to be ~7 %. XPS reveals a characteristic Ru 3d 5/2 peak at 280.2 eV for all three
carbon nanostructure supported RuNPs materials which is consistent with the previously reported literature
for metallic ruthenium.40
In addition, commercially available Ru/C, containing 5% Ru by weight, was used as a control material
in which the RuNPs are located on the surface of carbon black which imparts no confinement effects. HRTEM
confirmed that the metal nanoparticles are distributed throughout the carbon support in Ru/C and showed a
wide distribution of NP sizes, with a mean diameter of 6.63 ± 2.45 nm (Figure 3f, Table 1). It is important to
note that as the RuNPs are not identical sizes across the catalysts in this study, it is not possible to
unambiguously assign all differences in reactivity soley to the effect of confinement. However, as the shape
and size of the NPs in each material is a direct result of the confinement experienced during formation, and
the differences in reactivity cannot be explained solely by considering the size of the NPs, vital information can
be obtained by comparing the reactivity of the catalysts.
In similar fashion to the SWNT nanoreactors, the catalytic activities of RuNPs@GNF and Ru/C, along
with the as-supplied GNFs as
a control, were investigated in hydrogenation of norbornene and
benznorbornodiene (Table 2). The empty GNFs displayed no catalytic activity at all. Interestingly, the
RuNPs@GNF catalyst exhibited turnover numbers for norbornene (11216) and benzonorbornadiene (4145)
conversion, which are significantly higher when compared to RuNPs@SWNT (c.f. (2959) and (1983)
respectively), and the unconfined Ru nanoparticles in commercial Ru/C or in RuNPs/C60@SWNT. This is
possibly due to the RuNPs inside of the much wider GNF cavity being more accessible for the reactant
molecules coupled with the formation of favourable π-π interactions with both 1a and 2a at the step-edge
enhancing the local concentration of the reactants near the RuNPs, and therefore the rate of reaction.41
12

Page 13 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
Figure 3. a-b) HRTEM images of RuNPs@GNF loaded with RuNPs (1 % by wt), with enlarged region (b) showing
the Ru [002] lattice spacing (d = 0.21 nm) of an individual, hcp structured, metallic NP which corresponds to
the 2θ peak centred at 42.2° in the XRD (see ESI). Black arrows show RuNPs on the step edges inside the GNF.
c) Dark-field scanning TEM image of RuNPs@GNF loaded with RuNPs. Only a very small amount of Ru material
(~7 % approximated by TEM analysis), is located on the outer surface of the GNF so all conversion is assumed
to be a result of reactions catalysed by RuNPs on the inside of the GNF. d) HRTEM image of commercial Ru/C
loaded with RuNPs (5 % by wt.). White arrows show RuNPs located on the carbon support. e) The size
distribution of RuNPs on the outside of the GNFs (green), and d), the size distribution of RuNPs on the carbon
(purple) shows the wide distribution of diameters of the RuNPs in the material. Scale bars: 10 nm in all images.
Previous studies have revealed that both the size and crystal structure of RuNPs have important effects on the
catalytic activity of RuNPs in hydrogentation.42-45 Li et al. reported that the size and loading of hcp RuNPs
supported on mulit-walled carbon nanotubes for the hydrogentation of long-chain alkenes was optimum for
1.3 nm NPs and and metal loadings of 1% by wt., while in solution 3.1 nm RuNPs were observed to be the most
active.42,43 Dupont et al. reported that 2.6 ± 0.4 nm RuNPs were the most active for the partial hydrogenation
of benzene, with TON of up to 165 for supported RuNPs catalysts.44 It is also observed that distortions to the
13

Journal of Materials Chemistry A
Page 14 of 22
View Article Online
DOI: 10.1039/C7TA03691D
lattice planes of hcp RuNPs45 or changes in crystal packing to face-centred cubic Ru36 results in significant
effects to both the activity and selectivity of RuNP catalysed reactions.
Our hydronegation results, in which all RuNPs catalysts have the same hcp structure (confirmed by
HRTEM, see ESI Figure S-5), show an interesting trend in activity, consistent for indivudal reactions with both
alkene starting materials with TONs in the order of RuNPs@GNF>Ru/C>RuNPs@SWNT>RuNPs/C60@SWNT. It is
important to highlight that the mass of catalyst in each reaction was scaled so the same molar percent of of Ru
was present in all our reactions (0.0017 mmol% of Ru), and the trend in activity cannot be rationalised by
simply considering the average NP size or total NP surface area for each sample, (for our series, the total
surface area of RuNPs@SWNT>RuNPs/C60@SWNT>RuNPs@GNF>Ru/C), nor the extent of confinement, where
the confinement in RuNPs@SWNT> RuNPs@GNF >RuNPs/C60@SWNT >Ru/C. Therefore the enhancement in
activity observed upon confinement inside GNF must be a result of a balance of surface area of RuNP,
confinemement of the reaction and accessibility of the catalytic centres. Indeed, GNF provides an optimum
balance of RuNP confinement, leading to stabilisation and enhancement of local concentrations of rectants
around the nanoparticles, and the ease of accessibility of the reactants to catalytic centres (hindered in the
case of RuNP@SWNT). Recently, similar effects were observed in the oxygen reduction reaction catalysed by
PtNP@GNF.46
In addition the RuNPs@GNFs catalyst shows a marked enhancement in reactivity for the production
of 1b (1b:2b ratio = 2.7:1) compared to the unconfined Ru/C catalyst (ratio = 2.4:1) (Table 2) implying
remarkable TON for 1b, rationalised as a result of faster diffusion of the non-aromatic molecules to and from
the step-edge compared to the aromatic benzonorbornadiene which may interact more strongly with the sp2
carbon of the step-edge and therefore diffuse to and from the RuNPs less rapidly. In contrast RuNPs@SWNT
shows the opposite effect exhibiting and enhanced reactivity for the production of 2b (1b:2b ratio = 1.5:1),
potentially due to strong interactions between the aromatic core of 2a and the narrow inner channel of the
SWNT.
In summary, these results reveal that the confinement imposed by the carbon nanoreactor can not
only increase the activity of the Ru catalyst by both templating the formation of small, very active and stable
RuNPs and potentially by increasing the local concentration of reactant molecules in the vicinity of the Ru
catalyst but can also dramatically affect the efficiency of individual reactions depending on the size and shape
of the reactant molecules. This is highlighted by the RuNPs@SWNT catalyst which shows that RuNPs confined
in narrow SWNT are particularly efficient at converting aromatic 2a whilst the less constrained GNF catalyst are
highly active and show enhanced relative conversion of 1a compared to the unconstrained (carbon black)
reaction environment. Understanding the accessibility of the different catalysts is essential in order to fully
appreciate the role that confinement plays in the performance of these materials.
These results are in agreement with previous observations that performing reactions in confinement
has a number of important effects on catalysis and alters the outcome of reactions in a complex fashion;15 by
enhancing the activity of nanaoparticle cataltysts,47 increasing the local concentration of reagents and thus
14

Page 15 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
increasing the rate of reactions,48 and by imposing restritions on both the transition states of intermediates41
and the flow of reactants in and products out of the nanoreactor.27
To investigate the 3D structure of the materials the experimentally measured active surface area of
RuNPs@SWNT, RuNPs@GNF and commercial Ru/C catalysts were compared with theoretically calculated
surface areas based on ideal models of the materials, i.e. assuming the entire catalyst is perfectly formed and
accessible to reagents (see ESI for full details). Brunauer-Emmett-Teller (BET) measurements based on
isothermal N2 gas adsorption at -196 °C (77 K) were performed for RuNPs@SWNT, RuNPs@GNF and
commercial Ru/C catalysts along with control measurements of the support materials, i.e. SWNT, GNF and
amorphous carbon materials to quantify both surface area (Table 3) and pore volume and size distribution (see
supporting information Figure S-14-17 and Table 3-4). This provided a reasonable gauge of how much of the
internal volume of the catalysts were accessible to gaseous N2, and comparison to the theoretical maximum
values, revealed the portion of the internal volume which is either blocked by Ru nanoparticles large enough
to completely fill the channels or other inherent material such as amorphous carbon and residual Ni/Y catalyst
present following the initial nanotube synthesis. BET measurements showed that the commercial Ru/C (929
m2/g) has the largest surface compared to that of RuNPs@SWNT (352 m2/g) and RuNPs@GNF (29 m2/g) (see
Table 3).
Table 3. Theoretical and BET surface area calculations for SWNT, GNF, RuNPs@SWNT, RuNPs@GNF and
commercial Ru/C.
Theoretical Surface Area
Catalyst
BET Surface Area (m2/g)
(m2/g)
SWNTa
2630.00
507.90
351.94
120.99
164.82
12.08
RuNPs@SWNT
C60@SWNT
RuNPs/C60@SWNT
GNFb
2637.46
1315.00
1319.31
35.00
RuNPs@GNF
Carbon Black
Ru/Cc
35.99
28.53
N/A
900.00
N/A
929.41
a SWNT was annealed at 380 °C for 20 min prior to use. b GNF was annealed at 450 °C for 1 hour prior to use. c
BET surface area value for the carbon black used to synthesize the Ru/C was obtained from the manufacturer
(Sigma-Aldrich).
15

Journal of Materials Chemistry A
Page 16 of 22
View Article Online
DOI: 10.1039/C7TA03691D
The increased BET surface area of Ru/C and RuNPs@GNF compared to the background carbon and empty GNF
supports respectively is attributed to the presence of the RuNPs. In contrast RuNPs@SWNT (352 m2/g)
exhibited a lower BET surface area than that of the background SWNT (508 m2/g). This is rationalised as a
consequence of a small number of large (>1.2 nm) RuNPs blocking a significant percentage (c.a. 30-40 %) of the
entrances and/or channels/pores of the narrow nanotubes and not allowing N2 to access the full internal
volume of the material.
Typically, carbon nanomaterials have very large surface areas due to their nanometre scale features;
see theoretical specific surface calculated for each sample (Table S-3 in ESI). Peigney et. al. investigated the
theoretical surface area of CNT and reported that the specific surface area of SWNT, independent of its
diameter and length, is the same as that of both sides of a graphene sheet, i.e. 2630 m2/g.49 As we opened the
ends of the nanotubes prior to use, i.e. both internal and external walls of the nanotubes can be assumed to
be accessible, the theoretical surface area for the SWNT sample is 2630 m2/g. However, in reality the majority
of SWNT form bundles and therefore it is difficult to approximate an accurate number for the theoretical
surface area for nanotubes as the outer surface of most single nanotubes are not available as they are
contained within the bundles. Clearly the experimental value for SWNT, 508 m2/g, is dramatically lower and
can be considered as the effect of bundling.
When considering the accessibility of GNFs it is important to understand that their structure is
different to the concentric tubes of traditionally MWNTs and consists of stacked truncated cones of graphite
layers arranged at an angle along the main axis.50 The internal surface has a succession of step edges which
can act as anchoring points for guest molecules while the exterior surfaces are atomically flat. Therefore there
is no reported framework for calculating a theoretical value for the surface area of GNFs.
In our approach we approximated the internal surfaces of a GNF to be flat and the sidewalls to consist
of ~75 graphene layers (using an average thickness value of a GNF as 25 nm). Therefore utilizing the same
mathematical approach used by Peigney et al. for MWNTs, we calculated as theoretical surface area of 35
m2/g, which is significantly higher that of the BET measurement (12 m2/g). This is rationalized as the effect of
the overlapping of the truncated cones in the internal channels decreasing the internal surface area compared
to our theoretical model, but could also be a result of the presence of a number of GNFs in the sample with
thicker sidewalls. Therefore, though it is difficult to comment on the RuNPs@SWNT sample, as a significant
portion of the material appears to be blocked and inaccessible, c.f. an observed surface area of >4 times lower
than predicted. This comparison does reveal that even though the Ru/C catalyst has a higher surface area than
RuNPs@GNF, both in terms of total catalyst and the RuNPs, it is significantly less active. This must be due to
the size and shape of the pores, i.e. the nature of the confinement imposed.
16

Page 17 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
Scheme 2. Competitive reactions of norbornene and benzonorbornadiene in the presence RuNPs@SWNT
using a scCO2 high pressure rig.
Figure 4. Schematic representation of the competitive hydrogenation of norbornene (red) and
benzonorbornadiene (green) over RuNPs (blue) within SWNT. This shows how the extremely confined space
within the nanotube channel traps the reactants close to the catalytic nanoparticles, effectively increasing the
local concentration of reagents and increasing the rate of reaction.
Table 4. Competitive hydrogenation reactions of norbornene and benzonorbornadiene in the presence of
RuNPs@SWNT, RuNPs/C60@SWNT, RuNPs@GNF and commercial Ru/C using a high pressure scCO2 batch
reactor.
Catalyst
Yielda of Products (%) / TONb
Ratio of TONs for
1b:2b
1b
2b
RuNPs@SWNT
RuNPs/C60@SWNT
RuNPs@GNF
Ru/C
70 / 1138
89 / 1447
0.79:1
0.54:1
0.5:1
7 / 197
21 / 2560
52 / 3787
13 / 366
42 / 5120
53 / 3860
0.98:1
Reaction conditions; substrates (0.5 mmol of each alkene), catalyst (0.0017 m mol% of Ru in each reaction system), H2 (10
bar), scCO2 (100 bar), 24 h, 110 °C. a Yield determined by 1H NMR with an error ± 2 % (See Figure S-9 in Electronic
Supplementary Material). bThe turnover number (TON) was calculated as the ratio of the number of molecules of substrate
consumed in the reaction per theoretical number of true active Ru sites in catalyst used in the reaction.
To probe these effects of confinement further competitive reactions were investigated in which equimolar
amounts of norbornene and benzonorbornadiene were simultaneously reduced within the nanoreactor
catalyst (Scheme 2 and Figure 4).
Competitive reactions reveal that the confined space of carbon nanoreactors considerably affects the
selectivity of the reactions compared to reactions performed in the absence of confinement (Ru/C) (Table 4).
For all reactions in carbon nanoreactors benzonorbornadiene was preferentially reduced over norbornene in
the presence of carbon nanoreactors due to the strong aromatic character of the nanotube. In fact the TONs of
2b in the competitive reactions are larger than the individual reactions for RuNPs@GNF and Ru/C despite a
reduction in the starting concentration (Table 2), indicating that the presence of equimolar quatities of 1a may
17

Journal of Materials Chemistry A
Page 18 of 22
View Article Online
DOI: 10.1039/C7TA03691D
have an effect on π-π interactions of the aromatic ring of 2a and the sp2 hybridized carbon network of the
interior of the nanotubes being amplified in the rpesense of the non-aromatic 1a leading to an increased local
concentration of the aromatic reactant in proximity to RuNPs catalyst thus promoting reduction of 2a.41
RuNPs@GNF exhibited both the highest activity (total TON = 5680) and also the greatest selectivity
towards 2b formation (ratio 1b:2b = 0.5:1) compared to RuNPs@SWNT (ratio of 0.79:1) (Table 4). This can be
explained by considering the nature of the confinement imposed by the step-edges of the internal channel of
the GNF. Any increase in the local concentration of the reactants at the step-edges compared to the local
concentration inside the nanotube cavity and bulk would result in a much higher rate of reaction.48 As shown
in Figure 3a, the step-edges enable a very well-ordered distribution of RuNPs. The height of the step-edges, ~3
nm, provides a controlled space, which forms favourable π-π interactions between the aromatic 2a and the sp2
hybridized carbon step-edges and thus increases the concentration of 2a in the vicinity of the RuNPs and
consequently increases the observed conversion of the aromatic compound. The same effect is observed for
SWNT, with the formation of 2b observed to be enhanced compared to 1b in the individual reactions (see ratio
of TONs in Table 2), however, the physical size of the aromatic compound means that there will be a significant
steric barrier to the movement of the molecules into the narrow SWNT channels compared to the relatively
larger space at the step edges in GNF which explains the lower observed TON for the SWNT catalyst.
4. Conclusions
In conclusion, a series of RuNPs@carbon nanoreactor catalysts have been synthesized and the catalytic
properties of these materials have been assessed and the role of confinement explored. Overall, the
confinement imposed by RuNPs@SWNT and RuNPs@GNF results in dramatic changes to reactions compared
to commercially used Ru/C which exhibits no confinement. The affinity of aromatic groups for the interior
channels of carbon nanoreactors results in enhanced conversion of aromatic reagents in competitive
hydrogenation reactions providing the ability to alter the selectivity of chemical reactions using
support/reactant interactions.
Interestlingly the extreme confinement imposed by the shape of the SWNT nanoreactors is found to be a
double edged sword, as though they exhibit enhanced selectivity towards aromatic substrates, as a result of
strong interactions between the aromatic species and nanotube sidewall, the constricted space inside the
SWNT lowers TONs in general. Remarkably it is the wider, structured interiors of GNF which have a greater
effect on reactions, enhancing the activity and resulting in much higher TONs than the extremely narrow
SWNT and thus offer the best of both worlds, in that they are wide enough not to limit diffusion of reactants
but structured enough to impart the positive effects of reaction confinement.
5. Acknowledgements
This work was supported by Turkish Ministry of Education, European Research Council (ERC) and the EPSRC
(EP/K005138/1 for the Kratos LiPPS XPS instrument). We would like to thank the Nanoscale and Microscale
Research Centre (nmRC) for access to XPS, HRTEM and STEM instruments, Dr. A. La Torre for TEM support, Ben
18

Page 19 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
Pointer-Gleadhill for help with ICP-OES analysis, Dr. David Morgan for helpful discussions regarding the XPS
peak assignments and Helena Matabosch Coromina for help to use the BET equipment.
Electronic Supplementary Material: Supplementary material (containing additonal figures showing HRTEM
images, size distribution and EDX data for RuNPs@SWNT, XPS data, XRD data for RuNPs@SWNT,
RuNPs/C60@SWNT and RuNPs@GNF with the reference data obtained from the ICSD database of Royal Society
1
of Chemistry, H NMR spectra for competitive hydrogenation of norbornene and benzobnorbornadiene, TGA
and ICP-OES data for RuNPs@SWNT, TGA data for RuNPs@GNF, details of the TON calculations used for single
and competitive hydrogenation reactions, BET surface area, theoretical surface area calculations and pore size
distributions for RuNPs@SWNT, RuNPs/C60@SWNT, RuNPs@GNF and Ru/C) is available in the online version
of this article at ..
6. References
[1] D. K. Mishra, A. A. Dabbawala and J. S. Hwang, J. Mol. Catal. A: Chemical 2013, 376, 63–70.
[2] Q. Lin, G. Yang, Q. Chen, R. Fan, Y. Yoneyama, H. Wan and N. Tsubaki, Chem. Cat. Chem. 2015, 7, 682-
689.
[3] I. Lee, M. A. Albiter, Q. Zhang, J. P. Ge, Y. D. Yin and F. Zaera, Phys. Chem. Chem. Phys. 2011, 13,
2449-2456.
[4] K. Okitsu, A. Yue, S. Tanabe and H. Matsumoto, Chem. Mater. 2000, 12, 3006-3011.
[5] M. Yoshizawa, J. K. Klosterman, and M. Fujita, Angew. Chem., Int. Ed. 2009, 48, 3418.
[6] N. A. Dhas, H. Cohen, and A. Gedanken, J. Phys. Chem. B 1997, 101, 6834-6838.
[7] B. S. Kong, J. X. Geng and H. T. Jung, Chem. Commun. 2009, 274-2176.
[8] G. M. Scheuermann, L. Rumi, P. Steurer, W. Bannwarth, and R. Mulhaupt, J. Am. Chem. Soc. 2009, 131,
8262-8270.
[9] X. Pan, and X. Bao, Acc. Chem. Res. 2011, 44, 553-562.
[10] D. A. Britz, A. N. Khlobystov, K. Porfyrakis, A. Ardavan and G. A. D. Briggs, Chem. Commun. 2005, 37-39.
[11] S. Iijima, Nature 1991, 354.
[12] S. M. Dresselhaus, G. Dresselhaus and A. Jorio, Annu. Rev. Mater. Res. 2004, 34, 247-278.
[13] J. L. Figueiredo and M. F. R. Pereira, Carbon materials for catalysis, Carbon as Catalyst, in Carbon
Materials for Catalysis (eds P. Serp and J. L. Figueiredo), John Wiley & Sons, Inc., Hoboken, NJ, USA, 2008.
[14] J. Zhu, A. Holmen and D. Chen, Chem. Cat. Chem. 2013, 5, 378-401.
[15] A. N. Khlobystov, ACS Nano 2011, 5, 9306-9312.
[16] X. Pan, and X. Bao, Chem. Commun. 2008, 6271-6281.
[17] E. Castillejos, P. J. Debouttiere, L. Roiban, A. Solhy, V. Martinez, Y. Kihn, O. Ersen, K.
Philippot, B. Chaudret and P. Serp, Angew. Chem., Int. Ed. 2009, 48, 2529-2533.
19

Journal of Materials Chemistry A
Page 20 of 22
View Article Online
DOI: 10.1039/C7TA03691D
[18] W. A. Solomonsz, G. A. Rance, M. Suyetin, A. La Torre, E. Bichoutskaia and A. N. Khlobystov, Chem. Eur. J.
2012, 18, 13180-13187.
[19] B. Cornelio, A. R. Saunders, W. A. Solomonsz, M. Laronze-Cochard, A. Fontana, J. Sapi, A. N. Khlobystov
and G. A. Rance, J. Mater. Chem. A. 2015, 3, 3918-3927.
[20] D. Wang, G. Yang, O. Ma, M. Wu, Y. Tan, Y. Yoneyama and N. Tsubaki, ACS Catal. 2012, 2, 1958-1966.
[21] A. Benyounes, M. Kacimi, M. Ziyad and P. Serp, Chin. J. Catal. 2014, 35, 970-978.
[22] X. Pan, Z. Fan, W. Chen, Y. Ding, H. Luo and X. Bao, Nature Mater. 2007, 7, 507-511.
[23] J. Deng, P. Ren, D. Deng, L. Yu, F. Yang and X. Bao, Energy Environ. Sci. 2014, 7, 1919.
[24] T. W. Chamberlain, T. Zoberbier, J. Biskupek, A. Botos, U. Kaiser and A. N. Khlobystov, Chem. Sci. 2012, 3,
1919-1924.
[25] X. Li, T. Hungria C. G. Marcelot, M. R. Axet, P. -F. Fazzini, R. Tan, P. Serp, and K. Soulantica, Chem.
Commun. 2016, 52, 2362-2365.
[26] I. Such-Basánez, M. C. Roman-Martínez and P. Serp, Microporous and Mesoporous Materials 2016, 225,
378-384.
[27] T. W. Chamberlain, J. H. Earley, D. P. Anderson, A. N. Khlobystov and R. A. Bourne, Chem. Commun. 2014,
50, 5200-5202.
[28] H. Zhang, X. Pan and X. Bao, Journal of Energy Chemistry 2013, 22, 251-256.
[29] H. Zhang, X. Pan, J. Liu, W. Qian, F. Wei, Y. Huang and X. Bao, ChemSusChem 2011, 4, 975-980.
[30] H. Zhang, X. Pan, X. Han. X. Liu, X. Wang, W. Shenand X. Bao, Chemical Science 2013, 4, 1075-1078.
[31] H. Shiozawa, T. Pichler, A. Grüneis, R. Pfeiffer, H. Kuzmany, Z. Liu, K. Suenaga and H. Kataura, Adv. Mater.
2008, 20, 1443-449.
[32] S. A. Miners, G. A. Rance and A. N. Khlobystov, Chem. Commun. 2013, 49, 5586-5588.
[33] S. A. Miners, G. A. Rance, A. La Torre, S. M. Kenny, A. N. Khlobystov, J. Mater. Chem. C, 2014, 2, 8357-
8363.
[34] J. W. Coe, M. C. Wirtz, C. G. Bashore and J. Candler, Org. Lett. 2004, 6(10), 1589-92.
[35] E. Castillejos, P. J. Debouttiere, L. Roiban, A. Solhy, V. Martinez, Y. Kihn, O. Ersen, K. Philippot, B. Chaudret,
P. Serp, Angew. Chem., Int. Ed. 2009, 48, 2529–2533.
[36] K. Kusada, H. Kobayashi, T. Yamamoto, S. Matsumura, N. Sumi, K. Sato, K. Nagaoka, Y. Kubota, H.
Kitagawa, J. Am. Chem. Soc. 2013, 135, 5493−5496.
[37] W. Chen and S. W. Chen, Angew. Chem. Int., Ed. 2009, 48, 4386–4389.
[38] R. S. Oosthuizen and V. O. Nyamori, Platinum Metals Rev. 2011, 55, 154-169.
[39] Z. -Q. Feng, X. -L. Yang, Y. -F. Ye and L. -Y. Hao, Bull. Korean Chem. Soc. 2014, 35, No. 4.
20

Page 21 of 22
Journal of Materials Chemistry A
View Article Online
DOI: 10.1039/C7TA03691D
[40] J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, 1992 Handbook of X-ray Photoelectron
Spectroscopy (Perkin-Elmer, Minnesota) pp. 44 and 114.
[41] M. A. Lebedeva, T. W. Chamberlain. A. Thomas B. E. Thomas, C. T. Stoppiello, E. Volkova, M. Suyetin, A. N.
Khlobystov, Nanoscale 2016, 8, 11727-11737.
[42] S. Guo, K. Y. Liew, J. Li, J Am Oil Chem Soc 2009, 86, 1141–1147.
[43] B. Xu, K. Y. Liew, J. Li, J Am Oil Chem Soc 2007, 84, 117–122.
[44] E. T.Silveira, A. P.Umpierre, L. M. Rossi, G. Machado, J. Morais, G. V. Soares, I. J. R. Baumvol, S. R. Teixeira,
P. F.P. Fichtner, J. Dupont, Chem. Eur. J. 2004, 10, 3734-3740.
[45] J. Mao, W. Chen, W. Sun, Z. Chen, J. Pei, D. He, C. Lv, D. Wang, Y. Li Angew. Chem., Int. Ed., 2017, 56, 1-6.
[46] M. C. Gimenez-Lopez, A. Kurtoglu, D. A. Walsh, A. N. Khlobystov Adv. Mat. 2016, 28, 9103-9108.
[47] S. Guo, X. Pan, H. Gao, Z. Yang, J. Zhao, X. Bao, Chem Eur J 2010, 6, 5379–5384.
[48] M. J. Ledoux, R. Vieira, C. Pham-Huu, N. Keller, J. Catal. 2003, 216, 333–342.
[49] A. Peigney, C. Laurent, E. Flahaut, R. Bacsa and A. Rousset, Carbon, Elsevier 2001, 39, 507-514.
[50] A. La Torre, M. C. Gimenez-Lopez, M. W. Fay, G. A. Rance, W. A. Solomonsz, T. W. Chamberlain, P. D.
Brown and A. N. Khlobystov, ACS Nano 2012, 6, 2000-2007.
21

Journal of Materials Chemistry A
Page 22 of 22
View Article Online
DOI: 10.1039/C7TA03691D
TOC graphic
Exploratory, competitive hydrogenation reactions reveal the optimum level of confinement to
control chemical reactions.