Molecular sieving carbon catalysts for liquid phase reactions: Study of alkene hydrogenation using platinum embedded nanoporous carbon

Article abstract of DOI:10.1016/j.molcata.2012.10.026

We present a simple method to synthesize shape selective carbon catalysts for large alkene hydrogenation reactions by tailoring porosity of platinum embedded in polyfurfuryl alcohol derived carbons. A small amount of mesoporosity, in addition to the intrinsic microporous nature of the carbon, shortens the diffusion length for the reactant molecule, enabling these materials to be used for catalysis in the liquid phase. A systematic study of hydrogenation reactions of liquid phase alkenes is reported. The molecular sieving effect of the catalyst was examined by varying molecular length, size, double bond position, stereoregularity and the number of double bonds in the alkenes.

Full text of DOI:10.1016/j.molcata.2012.10.026

Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Molecular sieving carbon catalysts for liquid phase reactions: Study of alkene
hydrogenation using platinum embedded nanoporous carbon
Billy Paul M. Holbrooka, Ramakrishnan Rajagopalanb, Krishna Dronvajjalac,
Yogesh Kumar Choudharyc, Henry C. Foleya,b,c,∗
a Department of Chemistry, University Park, PA 16802, United States
b Department of Chemical Engineering, University Park, PA 16802, United States
c Materials Research Institute, University Park, PA 16802, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
We present a simple method to synthesize shape selective carbon catalysts for large alkene hydrogenation
reactions by tailoring porosity of platinum embedded in polyfurfuryl alcohol derived carbons. A small
amount of mesoporosity, in addition to the intrinsic microporous nature of the carbon, shortens the
diffusion length for the reactant molecule, enabling these materials to be used for catalysis in the liquid
phase. A systematic study of hydrogenation reactions of liquid phase alkenes is reported. The molecular
sieving effect of the catalyst was examined by varying molecular length, size, double bond position,
stereoregularity and the number of double bonds in the alkenes.
Received 27 September 2012
Received in revised form 25 October 2012
Accepted 26 October 2012
Available online 5 November 2012
Keywords:
Molecular sieving carbon catalyst
Shape selectivity
© 2012 Elsevier B.V. All rights reserved.
Alkene hydrogenation
Transition shape selectivity
Microporous carbon
1. Introduction
and detailed understanding of the origin of their regularly sized
pores that we enjoy with zeolites, we are not at a total loss if insight.
Shape selective catalysis is used for applications ranging from
acid catalyzed hydrocarbon cracking in petroleum manufacture,
to the production of monomers such as ethylbenzene, xylene, p-
ethyl toluene for polymerization [1]. Currently, zeolites are the only
shape selective catalysts to have found wide usage in practice; in
addition to their high selectivity for certain reactions, they have
adequate to good thermal stability and uniform pore structure.
However, there are numerous reactions where the acidic nature of
zeolites leads to unfavorable side reactions and deactivation. Thus
shape selective catalysis without zeolites is intriguing.
An alternative to shape selective zeolites is molecular sieve car-
bon (MSC), or nanoporous carbon (NPC) [2–4]. NPC is thermally
stable, possesses good mechanical strength and it surface is chemi-
cally fungible. In contrast to the inherent surface acidity of zeolites,
NPC can be acidic, basic or neutral. That said NPC are complex solids;
they are not crystalline, unlike zeolites. In spite of their global dis-
order, the NPC derived directly from polyfurfuryl alcohol (PFA) has
a narrow pore size distribution with an average of about 0.5 nm,
much like the ZSM-5 zeolites. Although we lack the kind of ready
It is considered most likely that the regular porosity is due to the
misalignment of polyaromatic domains in the carbon framework
arising from the inclusion of curvature induced by the presence
of five and seven membered rings being in close proximity to six
membered rings. While this is quite a contrast to the crystalline
structure of zeolites their molecular sieving capabilities are similar.
Therefore, the NPC materials offer intriguing possibilities for com-
bining catalysis with shape selective effects in the synthesis and
production of fine chemicals. It can be imagined that NPC could be
found useful and better in such reactions than metals on activated
carbon, but only if the metals within them can be made accessible
to large molecules for liquid phase reactions and if these metals can
be stabilized against sintering.
Catalysis with platinum embedded within nanoporous carbon
has been demonstrated in nearly defect-free, platinum-loaded
NPC membranes [5–7]. The platinum-containing catalytic mem-
branes displayed significant size selectivity in the hydrogenation
of different simple gas phase alkenes. The ratios of the intrin-
sic rate constants for differently sixed light alkenes hydrogenated
within these membranes were much more sensitive to the alkene
chain length (propylene/isobutylene and 1-butene/isobutylene)
than were the same ratios derived for more standard platinum
supported on carbon catalysts [7]. Since the analysis of the reac-
tion data had already taken into account the different diffusivities
∗
Corresponding author at: Department of Chemical Engineering, University Park,
PA 16802, United States. Tel.: +1 814 865 6332; fax: +1 814 865 7846.
E-mail addresses: hcf2@psu.edu, hankfoley@engr.psu.edu (H.C. Foley).
1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.10.026

62
B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
as a function of size for the alkenes, it was considered likely that
the magnitudes of the intrinsic rate constant ratios were a result
of the carbon’s micropores (average size ∼0.5 nm) exerting steric
effects on the intermediates and possibly even on transition states
that led from reactants to products. This hypothesis was bolstered
by HRTEM images, which showed that the regular carbon chan-
nels extended from the bulk of the carbon matrix to the surface of
platinum particles. Hence molecular ingress to and egress from the
platinum particles surfaces was only through these carbon chan-
nels.
Platinum embedded nanoporous carbon catalysts were pre-
pared using tetrohydrofuran (THF) as diluent instead of Triton
X-100. We used a procedure nearly identical to that described
above to make these catalysts. This catalyst was labeled as Pt-PFA-
THF. In addition to the synthesized catalysts, a standard 10 wt.%
platinum catalyst dispersed on the surface of carbon black was
purchased from Alfa Aesar Co. and used as a control for the hydro-
genation reactions.
2.3. Transmission electron microscopy (TEM)
More recently a novel synthetic procedure to embed monodis-
perse platinum nanoparticles in the matrix of nanoporous carbon
was reported [8]. Embedding prevents the particles from sintering
even at temperatures as high as 800 ◦C in flowing hydrogen [8].
Our results show that these catalysts also exhibit shape selectiv-
ity effects in the gas phase hydrogenation of light alkenes and that
these catalysts were highly active at room temperature. However,
given that these reactants were such small molecules, to be of more
significant interest in catalytic synthesis of high value fine chemi-
cals, the platinum embedded within NPC catalysts must be active
for reactions of larger molecules in the liquid phase. In this paper,
we examine the catalytic activities of these platinum embedded
within nanoporous carbon catalysts for liquid phase hydrogenation
of alkenes to better understand and demonstrate their molecular
sieving and how that affects their activity and selectivity.
The images of platinum embedded nanoporous carbon were
examined using an Hitachi HF2000 high resolution transmission
electron microscopy. In addition to the high-resolution images, the
size of the platinum nanoparticles was calculated using images
taken by HD2000 scanning transmission electron microscopy.
2.4. Scanning electron microscopy
Surface topography of the samples was studied using ultra high
resolution low KV Hitachi S-5200 microscope. The morphologies
of the samples were studied and the dispersion of platinum was
confirmed by taking the back-scattered images of the samples.
2.5. Pore size distribution
2. Experimental
The total pore volume and the average pore size were cal-
culated from methyl chloride adsorption isotherms using the
Horvath–Kawazoe slit shape pore model [9]. Before the methyl
chloride adsorption, the catalyst was heated to 350 ◦C and held
there for 4 h under vacuum. The catalyst was then allowed to cool
to room temperature and then the adsorption isotherm was mea-
sured based on gravimetric uptake of methyl chloride at different
pressures. Typically, the uptake equilibration time was between 2
and 3 h at each pressure. The adsorption isotherm was measured
at 273 K, 295 and 304 K for the pressure range of 5–700 Torr.
2.1. Materials
Platinum (II) (acetyl acetonate)2 or Pt(acac), furfuryl alcohol
(FA), Triton X-100 and p-toluenesulfonic acid monohydrate (p-
TSA) were purchased from Aldrich Co and used without any further
purification.
2.2. Synthesis of platinum nanoparticles in nanoporous carbon
2.6. Hydrogen chemisorption
To 5 mL of furfuryl alcohol was added, a known amount of plat-
inum (II) acetyl acetonate in a 125 mL round bottom flask and the
mixture was refluxed for 16 h. This resulted in a black colloidal
solution of platinum nanoparticles in furfuryl alcohol. The solu-
tion was allowed to cool to room temperature and an additional
5 mL of furfuryl alcohol along with 5 mL of Triton X-100 surfac-
tant, were stirred and placed in an ice bath. Next, 0.4 g of p-TSA
was dissolved in 5 mL of Triton X-100 and was added gradually to
the colloidal solution using a syringe pump. The polymerization
was allowed to continue for 24 h. The resultant viscous solution
was poured into a glass vial and maintained at 40 ◦C, 60 ◦C and
100 ◦C for three consecutive days, respectively. The solidified resin
was then pyrolyzed under flowing argon in a quartz tube furnace.
The sample was heated to 200 ◦C over the course of 1 h, main-
tained at that temperature for 2 h and then was slowly increased
to 800 ◦C at 2.5 ◦C/min and thermally “soaked” at 800 ◦C for 8 h.
The exact amount of platinum embedded into the carbon was cal-
culated using the final weight of the carbon and initial weight of
platinum (II) acetylacetonate. The catalyst was crushed and sieved
to particles ranging in size from 90 to 180 ␮m. Pretreatment of the
catalyst was done in a quartz tubular flow reactor by heating to
800 ◦C under UHP argon. At 800 ◦C, a mixture of UHP argon and
hydrogen (70:30) was introduced into the reactor and the cata-
lyst was reduced under these conditions for 3 h. The catalyst was
then slowly brought down to room temperature under flowing
argon atmosphere. This catalyst will be referred as Pt-PFA-TX in
the manuscript.
0.3 g of the catalyst was loaded into a Micromeritics Autochem
II chemisorption unit. The catalyst was pretreated at 400 ◦C under
a helium atmosphere. This was followed by reduction under 1 atm
of hydrogen at 400 ◦C for 1 h. The sample was then cooled to 350 ◦C
and the cell was evacuated for half hour. The pretreated sample was
brought to room temperature and the chemisorption was carried
out. The volume of the sample cell was calibrated using helium gas.
2.7. Study of liquid phase olefin hydrogenation
0.5 g of catalyst were reduced under flowing UHP hydrogen at
130 ◦C in a low-pressure stirred reactor (PARR 5100) for 1 h. The
reactor was pressurized to 3 bar with UHP hydrogen and allowed
to cool to room temperature overnight. After depressurizing the
reactor, a 50 mL reactant mixture containing 40 mL of undecane
(solvent) and 10 mL of the reactant was added. The vessel was
then pressurized to 4 bars with hydrogen and the temperature was
maintained at 30 ◦C. The products were analyzed using GowMac
600 series GC using a flame ionization detector and xTi-5 column
from Restek.
Care was taken to ensure that the reaction was not externally
mass transfer controlled by making sure that the rate of mix-
ing, modified by changing the speed of rotation of the impeller
did not have any impact on the observed reaction rate. Turnover
frequency number was reported for each reaction as the ratio of

B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
63
Table 1
0.25
0.20
0.15
0.10
0.05
0.00
Mesopore and micropore size distribution of the synthesized platinum embedded
catalyst as measured using methyl chloride adsorption tests.
(a)
Catalyst
Vmicro (cc/g)
Vmeso (cc/g)
VTotal (cc/g)
12 wt.% Pt-PFA-THF
12 wt.% Pt-PFA-TX
Control
0.18
0.15
0.33
0.06
0.01
0.19
0.24
0.16
0.52
observed reaction rate to the number of active sites measured using
chemisorption.
Pt-PFA-TX
Pt-PFA-THF
3. Results
1.00
10.00
100.00
1000.00
Pore Size (Angstroms)
3.1. Synthesis and characterization of platinum embedded
nanoporous carbon catalysts
0.3
0.25
0.2
(b)
In our previous investigation, we reported the synthesis of well-
dispersed platinum nanoparticles sequestered within a matrix of
nanoporous carbon [1]. During the synthesis of the catalyst, we
added Triton X-100 as a surfactant to disperse the nanoparticles.
However, when a large volume of Triton X-100 (1:1 by volume
with furfuryl alcohol) was added, it also began to act as a pore
former. Table 1 shows the cumulative micro- and mesopore vol-
ume of 12 wt.% platinum embedded carbon made by the pyrolysis of
platinum nanoparticles/polyfurfuryl alcohol composites that were
synthesized using THF and Triton X-100 as diluent, respectively.
The porosity of the Pt-PFA-THF catalyst was predominantly micro-
porous (∼0.15 cc/g) with pore sizes less than 2 nm as shown in Fig. 1.
By contrast, Pt-PFA-TX had both the micropore (<2 nm) and meso-
pores (2–50 nm). The micropore volume of Pt-PFA-TX was similar
(∼0.18 cc/g) while the mesopore volume was almost six times more
than the Pt-PFA-THF catalyst (∼0.06 cc/g). Commercial 10 wt.% plat-
inum impregnated catalyst had significant amount of micropore
and mesopore volume.
Pt-PFA-TX
Pt-PFA-THF
0.15
0.1
0.05
0
1.00
10.00
100.00
Pore Size (Angstroms)
1000.00
Fig. 1. (a) Cumulative pore volume of 12 wt.% Pt-PFA-TX and 12 wt.% Pt-PFA-THF
and (b) pore size distribution of 12 wt.% Pt-PFA-TX and Pt-PFA-THF as measured
using methyl chloride adsorption tests.
Fig. 2. Low kV high resolution scanning electron micrograph of (a) nanoporous carbon derived from polyfurfuryl alcohol and (b) 12 wt.% Pt-PFA-TX.

64
B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
Fig. 3. (a) HRTEM image of 12 wt.% Pt-PFA-TX and (b) an enlarged image of platinum nanoparticles submerged in the nanoporous carbon matrix.
The surface topography of platinum embedded carbon was
most position along the alkyl chain, the reactivity decreased. The
order of reactivity was as follows: 1-hexene > 2-hexene > 3-hexene
(Fig. 5).
examined using a low KV high resolution SEM. Nanoporous carbon
made with polyfurfuryl alcohol had relatively smooth morphology
as shown in Fig. 2. In contrast, the Pt-PFA-TX had a porous texture.
The presence of 10 nm pores was evident in the 12 wt.% Pt-PFA-TX.
Transmission electron microscopy and XRD studies showed that
the platinum nanoparticles were well dispersed in all samples and
that the average particle size was about 3–4 nm as shown in Fig. 3.
We probed the effect of steric hindrance on the double bond of
the olefins by performing hydrogenation reactions of 1-hexene, 2-
methyl-1-pentene, 3-methyl-1-pentene and 4-methyl-1-pentene
using both control and our catalyst as shown in Fig. 6. In
the case of commercial catalyst (control), all the reactants, 2-
methyl-1-pentene, 3-methy-l-pentene and 4-methyl-1-pentene,
hydrogenated completely within 60 min. The reactivity for all
the reactants was very similar, regardless of the placement of
the alkyl groups. However, the reactivity of the alkenes with
the embedded catalysts was significantly affected by the prox-
imity of the methyl groups to the unsaturated double bond.
Here 2-methyl-1-pentene was far less reactive than 3-methyl-1-
pentene, 4-methyl-1-pentene and 1-hexene, demonstrating the
severe steric hindrance effect. Furthermore, 3-methyl 1-pentene
had a higher reactivity than 4-methyl-1-pentene. A significant
3.2. Hydrogen chemisorption
Table 2 summarizes the results of hydrogen uptake as calcu-
lated from the data obtained from the hydrogen chemisorption
experiments. The results show that the apparent number of avail-
able active sites for chemisorption is low. The amount of hydrogen
uptake did increase with increase in platinum loading, but the
apparent dispersion of catalytic active sites was invariant with
loading and was nominally about 4%. The dispersion of platinum
loaded on carbon black (control) was about 13.6%.
3.3. Catalytic activity of platinum embedded carbon
In order to probe the reactant size selectivity for the alkenes,
the hydrogenation reactions were done under similar conditions.
The catalysts showed excellent size selectivity and the reactiv-
ity decreased with increase in the size of the reactant molecules
(1-hexene < 1-octene < 1-nonene < 1-decene) as shown in Fig. 4.
Hydrogenations of 1-hexene, 2-hexene and 3-hexene showed that
the activity of reactant was highly dependent on the position of the
double bond. When the double bond was moved toward the inner
Table 2
Hydrogen chemisorption results of platinum embedded nanoporous carbon
catalysts.
Catalyst
H2 uptake (␮mol/g)
Dispersion
4 wt.% Pt-PFA-TX
6 wt.% Pt-PFA-TX
12 wt.% Pt-PFA-TX
20 wt.% Pt-PFA-TX
Control
4.19
6.65
11.10
12.93
35
4
4
4
3
13.6
Fig. 4. Study of hydrogenation reactions using 12 wt.% Pt-PFA-TX catalyst as a func-
tion of length of the reacting olefins.

B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
65
Fig. 5. Study of hydrogenation reactions as a function of position of double bond in
Fig. 7. Study of hydrogenation reactions using 12 wt.% Pt-PFA-TX as a function of
the reacting olefins using 12 wt.% Pt-PFA-TX.
stereoregularity.
effect of stereoregularity on the extent of hydrogenation of the
reactant molecules was also observed for trans-2-hexene and
cis-2-hexene. A comparison of their reactivities showed that the
cis-2-hexene was far less reactive than the trans-2-hexene (Fig. 7).
Hydrogenation of a diene molecule such as 1,5-hexadiene using
12 wt.% Pt-PFA-TX was compared with a control catalyst (Fig. 8).
Both the catalysts showed two main products namely 1-hexene
(intermediate product) and n-hexane. As expected, the reactivity
of the control catalyst was greater than the embedded platinum
catalyst. The reaction went to 100% conversion within 1 h of reac-
tion time over the control catalyst whereas over Pt-PFA-TX, it took
almost 48 h to undergo complete conversion. Although the reaction
was fast over the control catalyst, the selectivity of 1-hexene/n-
hexane was always less than unity, even at low conversions and
was 0.01 at 95% conversion. However, due to the shape selective
nature of Pt-PFA-TX, we saw that the selectivity of 1-hexene/n-
hexane was greater than unity up to 40% conversion and fell to 0.3
at 95% conversion (Fig. 9).
Table 3
Catalytic turnover frequencies for hydrogenation reactions of different alkenes over
the platinum embedded nanoporous carbon catalysts (Pt-PFA-TX).
Type of olefin
12 wt.% Pt-PFA-TX
10 wt.% Pt control
TOF (s−1
)
TOF (s−1
)
1-Hexene
0.75
0.20
0.004
0.076
0.53
0.43
0.44
0.32
0.29
0.32
0.32
0.32
0.32
0.32
0.32
–
Trans-2-hexene
Trans-3-hexene
Trans 4-methyl-1-pentene
Trans 3-methyl-1-pentene
Trans 2-methyl 1-pentene
1-Octene
1-Nonene
1-Decene
–
–
our studies by looking at the pore size distribution of the catalysts.
In order to facilitate transport of liquid reactants and products,
it is required to create mesoporosity in addition to the inherent
nanoporous nature of polyfurfuryl alcohol derived carbon.
Previously, the evolution of porosity in polyfurfuryl alcohol was
reported [10]. It was shown that most of the gaseous by-products
were released between 300 ◦C and 500 ◦C. Porosity in these mate-
rials formed as soon as the decomposition of the polymer began.
However, when pyrolysis was done at temperatures greater than
500 ◦C, the mesoporosity so formed collapsed due to the anneal-
ing of the aromatic domains, leaving behind a nanoporous carbon.
Table 3 summarizes the performance of the catalyst for various
alkene hydrogenation reactions.
4. Discussion
Our objective in this investigation was to examine whether or
not we could extend the development of shape selective carbon cat-
alysts previously reported, to liquid phase reactions [8]. We began
Fig. 6. Study of hydrogenation reactions using (a) control and (b) 12 wt.% Pt-PFA-TX as a function of proximity of methyl groups to the double bond in the reacting olefins.

66
B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
Fig. 8. Study of hydrogenation reaction of 1,5–hexadiene using (a) control and (b) 12 wt.% Pt-PFA-TX.
It was concluded that this behavior could be changed if a pore-
forming molecule was added to the mixture as an additive during
the pyrolysis. This objective was easily accomplished by using the
surfactant molecule, Triton X-100, during the synthesis of the plat-
inum embedded in polyfurfuryl alcohol. The presence of Triton
X-100 during pyrolysis created mesoporosity in the carbon that
was retained even after annealing at 800 ◦C for 8 h. This mesoporo-
sity reduced resistance offered by the native PFA-derived carbon to
the transport of the reactant molecules, while preserving the shape
selective nature of the catalyst due to the presence of nanopores in
the immediate vicinity of the highly reactive platinum nanopar-
ticles. As reported previously, we saw that the catalysts prepared
using this method had very high resistance to sintering [8].
Hydrogen chemisorption experiments showed very little hydro-
gen uptake. The apparent dispersion of catalytic sites on the basis
of hydrogen chemisorption was about 4% and was invariant with
loading. The maximum theoretical dispersion for a 3–4 nm par-
ticle, if it were supported on the surface of an activated carbon
and if it were completely accessible to hydrogen would be about
28%. The deviation between expected and measured is a factor of
seven All the platinum particles are completely embedded within
the nanoporous carbon; however, hydrogen can access only those
platinum particles that are accessible in the porous sections of the
carbon. Therefore we cannot measure the dispersion accurately by
hydrogen chemisorption because many of the platinum particles
are inaccessible to hydrogen. However, despite being inaccessible,
we include their mass in the calculation of the measured dispersion
of 4% and this drops that value of the dispersion by a factor of seven
or more below expectation.
We can get at this in another way to confirm the explanation by
considering the actual porosity of the NPC. The available micropore
volume in these materials is about 0.15–0.18 cm3 g−1 and the total
pore volume for Pt-PFA-THF and Pt-PFA-TX was about 0.16 cm3 g−1
and 0.24 cm3 g−1, respectively. The accessible porosity calculated
based on the fraction of void volume to the total volume occupied by
the catalyst amounts to 25% and 33%, respectively. As the platinum
is evenly dispersed within the carbon, almost 75% of the platinum is
in the non-porous regions of the carbon and is rendered inaccessi-
ble. Thus given that the maximum theoretical dispersion for 3–4 nm
platinum particles should have been about 25%, if one third to one
fourth of the platinum is accessible, then the apparent dispersion
due to this effect could be no better than 6.25–8.25%. Moreover,
it has also been shown that the platinum surface can be contam-
inated by carbon and can result in lower dispersion [11–13]. We
measured dispersion of 4%. This result is in good agreement with
the expected value when we factor in the actual porosity of the NPC
and carbon contamination on the surface. It is interesting to note
that the turnover frequencies for these catalysts (see Table 3) are
quite high.
The next objective was to clearly demonstrate the molecular
sieving nature of the platinum embedded in carbon catalyst by
doing liquid phase hydrogenation reactions. The reactions were
varied on the basis of molecular size and shape, on the position
of double bond, on stereoregularity and number of double bonds
in the reactant molecule. There have been extensive investigations
using single crystal studies on platinum aimed at elucidating the
actual mechanism of hydrogenation of alkenes [14–17]. It has been
shown that alkenes may be adsorbed either by forming a strong
␴-bond or a weak ␲-bond with the single crystal surface of plat-
inum. Spectroscopic studies showed that the ␴-bonded alkenes do
not participate in the hydrogenation and it is believed that the
weakly ␲-bonded alkenes on the Pt (1 1 1) crystal are the ones that
undergo hydrogenation. If in fact the ␲-bond formation is criti-
cal, then the reactant molecule should sit parallel to the platinum
surface. This results in an interesting conformational problem in
these embedded platinum catalysts as the pores of the carbon are
on average only 0.5 nm and these surround the catalyst particles
and extend to their surfaces. Our hydrogenation results confirm
this effect.
In the hydrogenation of alkenes that have kinetic diame-
ters ranging from 0.5 nm (n-hexene) to 0.8 nm (n-octene), the
reactivity decreased with the increase in the chain length of
the reacting alkene. We also explored the effect of more sub-
tle variations such as the presence of steric hindrance due to
Fig. 9. Selectivity of 1-hexene/n-hexane formed during hydrogenation of 1,5-
hexadiene using (a) 12 wt.% Pt-PFA-TX and (b) control.

B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
67
Fig. 11. Structural models of (a) trans-2-hexene and (b) cis-2-hexene.
selectivities of 1-hexene/1-hexane were different for the to cat-
alysts. In the case of the standard catalyst, the selectivity of
1-hexene/hexane was less than one even at very low conversions
(<10%), while for the embedded platinum catalyst this ratio was
greater than one for conversions as high as 40% conversion. This
may be evidence of a transition state shape selective effect. Fur-
thermore, at very high conversions, we see that the selectivity ratio
drops precipitously for standard catalyst (0.01), while it remains at
about 0.3 even at 95% conversion with the embedded catalyst.
Fig. 10. Structural models of (a) 2-methyl-1-pentene, (b) 3-methyl-1-pentene and
(c) 4-methyl-1-pentene.
methyl groups near the double bond of the alkenes. We com-
pared our catalyst with a standard catalyst that supports platinum
on the surface of carbon. The reactivities of 2-methyl-1-pentene,
3-methyl-1-pentene and 4-methyl-1-pentene were very similar
over the control catalyst. However, comparison of these hydro-
genation reactions on the embedded platinum catalyst showed
that reactivity of 1-hexene > 3-methyl-1-pentene > 4-methyl-1-
pentene > 2-methyl-1-pentene. The transport limitation due to the
presence of molecular sieving in addition to the conformational
restraint on the alkenes, gives rise to differences in the reactivity
for hydrogenation of the alkenes as shown in Fig. 10. In the case of
1-hexene, three hydrogens and alkyl groups surround the double
bond. This facilitates the double bond making it to the surface as
the alkyl groups move either up or, out, away from the surface. By
contrast 2-methyl-1-pentene has a methyl and propyl group totally
out of plane with the double bond making it more difficult for the
double bond to align parallel to the catalyst surface. For 3-methyl-
1-pentene and 4-methyl-1-pentene, although the branched methyl
group is closer to the double bond, in 3-methyl-1-pentene versus
4-methyl-1-pentene, the double bond reacts faster. Considering
structures, the double bond in 3-methyl-1-pentene is out of plane
with respect to all other carbon atoms allowing the double bond
to come into close proximity with the platinum particle surface as
compared to the double bond in the 4-methyl-1-pentene.
5. Conclusions
In this investigation, we show that molecular sieving car-
bon catalysts can be synthesized for liquid phase hydrogenation
reactions by pyrolyzing platinum nanoparticles in PFA result-
ing in platinum embedded in PFA-carbon nanocomposites. The
porosity of the carbon can be improved significantly by using
a mesopore former such as Triton X-100 during the synthesis
of the polymer precursor and during its subsequent pyroly-
sis. Steric and conformational restriction effects on size and
shape selectivity of the catalysts were demonstrated using
several different liquid phase alkene hydrogenation reactions.
Based on these findings molecular sieving carbon catalysts
can be developed further for reactions that are important
to the biological, pharmaceutical and fine chemical indus-
tries.
Acknowledgments
We would like to acknowledge Dr. Larry Allard at Oakridge
National Laboratory for assisting us with HRTEM images. We would
also like to thank Prof. Abhaya Datye, University of New Mexico for
helping us characterize the catalysts using low KV high resolution
Scanning Electron Microscopy.
We also looked at the effect of stereoregularity by comparing
trans-2-hexene and cis-2-hexene. For trans-2-hexene, there is less
resistance as the molecule moves through the pores of the carbon
to the catalytic site. By comparison cis-2-hexene has a less stream-
lined structure that increases resistance to transport through the
porous carbon as shown in Fig. 11.
References
The effect of having multiple double bonds in the molecule
was probed using 1,5-hexadiene. Once again we compared the
performance of the embedded catalysts with the control catalyst
in which platinum is dispersed on the carbon surface. For both
catalysts, the same two products, 1-hexene and 1-hexane were
observed. The transport resistance in the embedded platinum cat-
alyst was evidenced by its slower rate of hydrogenation when
compared to the control. But, it was also interesting to note that the
[1] S.M. Csicsery, Pure Appl. Chem. 58 (1986) 841.
[2] J.L. Schmitt, P.L. Walker Jr., Carbon 9 (1971) 791.
[3] J.L. Schmitt, P.L. Walker Jr., Carbon 10 (1972) 87.
[4] D.L. Trimm, B.J. Copper, J. Catal. 31 (1973) 287.
[5] D.S. Lafyatis, R.K. Mariwala, E.E. Lowenthal, H.C. Foley, in: M.L. Occelli, H. Robson
(Eds.), Design and Synthesis of Carbon Molecular Sieves for Separation and
Catalysis, Van Nostrand Reinhold, New York, 1992, pp. 318–332.
[6] K. Miura, J. Hayashi, T. Kawaguchi, K. Hashimoto, Carbon 31 (1993) 667.
[7] M.S. Strano, H.C. Foley, AICHE J. 47 (2001) 66.

68
B.P.M. Holbrook et al. / Journal of Molecular Catalysis A: Chemical 367 (2013) 61–68
[8] R. Rajagopalan, A. Ponnaiyan, P.J. Mankidy, A.W. Brooks, B. Yi, H.C. Foley, Chem.
Commun. 21 (2004) 2498.
[13] H. Chen, H. Yang, Y. Briker, C. Fairbridge, O. Omotoso, L. Ding, Y. Zheng, Z. Ring,
Catal. Today 125 (2007) 256.
[9] R.K. Mariwala, H.C. Foley, Ind. Eng. Chem. Res. 33 (1994) 2314.
[10] C.L. Burket, R. Rajagopalan, A.P. Marencic, K. Dronavajjala, H.C. Foley, Carbon
44 (2006) 2957.
[11] N. Krishnankutty, M.A. Vannice, J. Catal. 155 (1995) 312.
[12] N. Krishnankutty, M.A. Vannice, J. Catal. 155 (1995) 327.
[14] P.S. Cremer, X. Su, Y.R. Shen, G.A. Somarjai, J. Am. Chem. Soc. 118 (1996)
2942.
[15] A.L. Marsh, G.A. Somarjai, Top. Catal. 34 (2005) 121.
[16] D. Chrysostomou, C. French, F. Zaera, Catal. Lett. 69 (2000) 117.
[17] I. Lee, F. Zaera, J. Phys. Chem. B 109 (2005) 2745.