Examining the temporal behavior of the hydrocarbonaceous overlayer on an iron based Fischer-Tropsch catalyst

Article abstract of DOI:10.1039/c8ra09731c

In order to examine fundamental processes connected with the use of an unpromoted iron based Fischer-Tropsch synthesis (FTS) catalyst, model studies examining the temporal formation of hydrocarbonaceous species that form over the catalyst are undertaken using a combination of temperature-programmed oxidation, powder X-ray diffraction, Raman scattering, transmission electron microscopy and inelastic neutron scattering (INS). Catalyst samples were exposed to ambient pressure CO hydrogenation at 623 K for defined periods of time-on-stream (3, 6, 12 and 24 h) prior to analysis. INS reveals a progressive retention of hydrogenous species that is associated with the evolution of a hydrocarbonaceous overlayer, as evidenced by the presence of sp² and sp³ hybridized C-H vibrational modes. Correlations between the formation of aliphatic and olefinic/aromatic moieties with post-reaction characterization leads to the proposal of a number of chemical transformations that, collectively, define the conditioning phase of the catalyst under the specified set of reaction conditions. A comparison between the inelastic neutron scattering spectra of the 24 h sample with that of an iron catalyst extracted from a commercial grade Fischer-Tropsch reactor validates the relevance of the experimental approach adopted.

Full text of DOI:10.1039/c8ra09731c

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Examining the temporal behavior of the
hydrocarbonaceous overlayer on an iron based
Cite this: RSC Adv., 2019, 9, 2608
Fischer–Tropsch catalyst†
b
Robbie Warringham,a Alisha L. Davidson,a Paul B. Webb,
Robert P. Tooze,c
*
*
d
a
Russel A. Ewings,d Stewart F. Parker
and David Lennon
In order to examine fundamental processes connected with the use of an unpromoted iron based Fischer–
Tropsch synthesis (FTS) catalyst, model studies examining the temporal formation of hydrocarbonaceous
species that form over the catalyst are undertaken using a combination of temperature-programmed
oxidation, powder X-ray diffraction, Raman scattering, transmission electron microscopy and inelastic
neutron scattering (INS). Catalyst samples were exposed to ambient pressure CO hydrogenation at 623 K
for defined periods of time-on-stream (3, 6, 12 and 24 h) prior to analysis. INS reveals a progressive
retention of hydrogenous species that is associated with the evolution of a hydrocarbonaceous
overlayer, as evidenced by the presence of sp2 and sp3 hybridized C–H vibrational modes. Correlations
between the formation of aliphatic and olefinic/aromatic moieties with post-reaction characterization
leads to the proposal of a number of chemical transformations that, collectively, define the conditioning
phase of the catalyst under the specified set of reaction conditions. A comparison between the inelastic
neutron scattering spectra of the 24 h sample with that of an iron catalyst extracted from a commercial
grade Fischer–Tropsch reactor validates the relevance of the experimental approach adopted.
Received 26th November 2018
Accepted 21st December 2018
DOI: 10.1039/c8ra09731c
rsc.li/rsc-advances
Despite its widespread industrial usage, the mechanism by
which Fe-based catalysts operate is still debated in the litera-
1. Introduction
ture. This is in part due to the complexity of the catalyst
composition under reaction conditions where a mixture of iron
oxides, iron carbides and carbonaceous material can be iden-
tied.2,8–10 Further, the reaction conditions necessary to induce
FTS activity (high pressure/temperature) and the formation of
high molecular weight products means in situ analysis of the
active catalyst system is challenging. Xu and Bartholomew used
Fischer–Tropsch synthesis (FTS) is a non-selective heteroge-
neously catalyzed industrial reaction that can produce a variety
of hydrocarbon products from the reaction of CO and H2
(syngas) obtained from coal, natural gas or biomass sources.1–3
These hydrocarbon products may be further processed to
produce sulfur-free diesel and high value chemicals utilized by
the chemical manufacturing industry.4 If the syngas mixture is
sourced from oil-free resources such as coal, FTS is considered
an alternative route towards the production of fuel, thereby
alleviating the dependency of crude oil derived hydrocarbons.5,6
Several variants of the process operate around the world, with
catalysts predominantly based on Fe and Co containing mate-
rials.3 The high hydrogen content of natural gas is suited to Co
catalysts, whereas hydrogen decient syngas from biomass and
coal is suited to Fe catalysts due to its high activity for water gas
shi (WGS) chemistry.7
¨
high pressure in situ Mossbauer absorption spectroscopy to
study the phase transformation of an Fe-based catalyst under
realistic FTS conditions (10 atm, 538 K).11 Other in situ studies
utilizing X-ray absorption spectroscopy (XAS), X-ray diffraction
(XRD), and scanning transmission X-ray microscopy (STXM)
have also been reported.12–14 An alternative approach to in situ
analysis is to investigate the Fe catalyst undergoing ambient
pressure hydrogenation of CO at elevated temperatures.15
Although no polymerization occurs to form the typical distri-
bution of high molecular weight hydrocarbon products, the
surface chemistry involved in CO/H2 dissociation and C–C/C–H
bond formation is informative and relevant.
Previous studies from this group have utilized this test
reaction to study how hydrogen is partitioned within the cata-
lyst matrix of reacted Fe FTS catalysts through the application of
inelastic neutron scattering (INS).16–19 From measuring indus-
trially reacted and laboratory prepared Fe FTS catalysts, the
formation of hydrocarbonaceous deposits during the large scale
aSchool of Chemistry, Joseph Black Building, University of Glasgow, Glasgow, G12
8QQ, UK. E-mail: David.Lennon@glasgow.ac.uk; Tel: +44-141-330-4372
bSchool of Chemistry, University of St Andrews, St Andrews, KY16 9ST, UK
cGroup Technology, Research and Technology, Sasol Ltd, UK
dISIS Facility, STFC Rutherford Appleton Laboratory, Chilton, Didcot, Oxon OX11
0QX, UK
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c8ra09731c
2608 | RSC Adv., 2019, 9, 2608–2617
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
FTS operation and the representative laboratory reaction was All gas ows were monitored using an in-line quadrupole mass
established. A further study found that H2 pre-treatment before spectrometer (Hiden Analytical, HPR-20) attached to the reactor
the CO hydrogenation reaction increased the level of hydro- exit line via a differentially-pumped, heated quartz capillary.
carbonaceous species present on the sample, particularly sp2 Under a continuous feed of reactant gases, the sample was
hybridized carbon moieties.18 These investigations led to the subjected to a temperature ramp of 5 K minꢀ1 to 623 K and held
proposal that the polyaromatic overlayer found on these reacted for a pre-determined length of time (x), aer which the reactant
Fe catalysts could have a role in the mediation and transfer of ows were stopped and the sample was cooled to room
reactive species to catalytic active sites.
temperature under the helium carrier gas. The micro-reactor
For the present study, the temporal dependence of the samples are referred to as MR-x, where x indicates the time
hydrocarbonaceous overlayer formed on a Fe FTS catalyst on stream in hours. For ex situ characterization, reacted
during ambient pressure CO hydrogenation (H2 : CO ¼ 2 : 1, samples were subjected to a passivation procedure involving the
623 K) is analyzed post-reaction using a combination of char- introduction of small amounts of O2 to the reactor, gradually
acterization techniques: temperature-programmed oxidation increasing to atmospheric levels (i.e. 20% O2 in the gas feed).22
(TPO), powder X-ray diffraction (XRD), Raman spectroscopy, As mentioned elsewhere,17 electing to operate syngas reactions
transmission electron microscopy (TEM) and INS. For the rst at ambient pressure permits access to Fe/CO/H2 surface chem-
time, the INS spectra reveal a progressive retention of hydrog- istry whilst ensuring the C–C propagation stage is inaccessible,
enous species as a function of time-on-stream that is associated as the latter would otherwise lead to products that would
with the evolution of the hydrocarbonaceous overlayer. Corre- dominate the INS spectrum. Thus, the approach adopted here
lations between the formation of aliphatic and olfenic/aromatic enables the investigation of surface species to take place
moieties with post-reaction TPO proles, plus other analytical without such species being ‘swamped’ by a high molecular
probes, leads to the proposal of a number of chemical trans- weight product slate.
formations that collectively dene the conditioning phase of the
catalyst under the specied set of reaction conditions. A
comparison between the INS spectra of a 24 h reacted sample
2.3. Inelastic neutron scattering measurements
with an Fe FTS catalyst extracted from a commercial grade FTS For INS measurements, approximately 10 g of catalyst was
reactor indicates the relevance of the experimental approach loaded into an Inconel reactor cell and attached to a custom
adopted. A qualitative model is proposed to account for the built sample preparation rig.23 For ambient pressure CO
experimental observations.
hydrogenation measurements, the iron oxide catalyst was
heated to 623 K at 5 K minꢀ1 under a ow of CO (75 sccm, CK
gas, 99.9%) and H2 (150 sccm, CK gas, 99.9%) in a carrier gas
(He, 600 sccm, CK gas, 99.9%, total WHSV of 1.47 hꢀ1) and held
at temperature for a pre-determined length of time. The gas
products were analyzed by an in-line mass spectrometer (Hiden
Analytical, HPR20 QMS Sampling System). Note that the MS
instrument utilized for these scaled-up reaction measurements
at the Central Facility is uncalibrated at the time of measure-
ment, therefore the gas traces are a qualitative representation of
the reaction proles. Once the specic reaction had nished,
the reactant gases were stopped and the sample allowed to cool
to room temperature under the carrier gas. The samples
prepared in the large scale reactor are referred to as LR-x, where
x indicates the time on stream in hours. The reactor cell was
isolated and placed in the argon-lled glove box (MBraun Uni-
Lab MB-20-G, [H2O] <1 ppm, [O2] <2 ppm) before being loaded
2. Experimental
2.1. Catalyst preparation and characterization
The iron oxide catalyst sample used for this investigation was
prepared using the co-precipitation of iron nitrate (Sigma
Aldrich, 99.99%) and sodium carbonate (Sigma Aldrich,
99.99%). The preparative procedure utilizes a batch reactor
apparatus for reproducible sample synthesis and is described
elsewhere.19 The procedure produces hematite (a-Fe2O3) with
a surface area of 70.8 m2 gꢀ1 and an absence of promoters/
modiers. All samples were ground and sieved to a particle
size range of 250–500 mm.
2.2. Micro-reactor measurements
Initial reaction testing was performed using a catalyst test line into an aluminum sample holder that is then sealed via an
composed of 1/8 in. diameter stainless steel Swagelok tubing; indium wire gasket.24 All INS measurements were performed
a description of which can be found elsewhere.20,21 Approxi- using the MAPS direct geometry spectrometer.25 Spectra were
mately 40 mg of catalyst was loaded into a 1/4 in. quartz tube recorded at 20 K at an incident neutron energy of 600 meV and
reactor with the reactor plugged with quartz wool. The reactor is 250 meV using the A-chopper package. Quantication of the
housed within a tube furnace (Carbolite MTF 10/15/30) equip- n(C–H) feature obtained by INS was achieved following a cali-
ped with PID control. A thermocouple is positioned within the bration protocol described elsewhere.26 Briey, samples con-
catalyst bed to ensure accurate temperature reading during taining known masses of a polystyrene calibrant (Sigma Aldrich,
measurements. For CO hydrogenation reactions ambient pres- >99.0%, typical molecular weight ¼ 350000) were analyzed
sure gas ows of CO (3.35 sccm, CK Gas, 99.8%), H2 (6.75 sccm, using the MAPS spectrometer at an incident energy of 600 meV.
BOC Ltd, 99.8%) and He (21.25 sccm, BOC Ltd, 99.9%) were The resulting spectra were baseline corrected and tted with
established over the bypass before introduction to the catalyst Gaussian functions using the Origin graphical soware package
reactor (total weight hourly space velocity (WHSV) of 60.8 hꢀ1). (MicroCal Origin, Version 8.0, Fig. S1†). The integrated
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 2608–2617 | 2609

View Article Online
RSC Advances
Paper
response was related to hydrogen content of the calibrant by CO to produce CO2 (Stage I), the simultaneous production of
material; the calibration plots are presented in the ESI CO2, H2O, and CH4 with a corresponding consumption of CO
(Fig. S1†).
and H2 at 623 K (Stage II), and a decrease in product yield
towards steady-state operation (Stage III). A degree of WGS
activity could additionally be contributing to the observed
reaction prole. CO conversion approximates to <1% and is
minimal during the course of the reaction measurement, cor-
responding to an iron time yield of 1.6 ꢂ 10ꢀ5 mol CO gFeꢀ1 sꢀ1
(Fig. 1a). This is a comparable rate to values reported for similar
reaction conditions.28 The catalyst retains this low activity
throughout the course of the 24 h reaction without deactivation.
In situ XRD identies the morphological transformations
during this period, with the reduction of a-Fe2O3 to Fe3O4
leading to the formation of iron carbides (Fig. 1b). It is noted
that the gas exchange characteristics of the in situ XRD cell are
different to the micro-reactor set-up (not least, a larger sample
mass in the in situ XRD reaction cell). This is thought to lead to
a prolongation of the reduction period, i.e. Stage I (Fig. S2†),
with the consequence that the presence of iron oxide phases
persists for longer in the XRD cell than occurs in the micro-
reactor. From 6.6 h onwards (not shown) there was a substan-
tial decrease in signal-to-noise attributed to carbon build-up.
In situ TPO has been demonstrated to be a useful tool in
discriminating and quantifying transient carbon species
formed under CO hydrogenation conditions at 623 K.19 From
our previous study, three peaks were observed aer 6 h on-
stream and were assigned to reactive adsorbed carbon (a),
amorphous-like carbon (b) and iron carbides (g) (Fig. 1c).19 Here
we observe the formation of these peaks occurring between 0 to
3 h at 623 K (Fig. 1c), concurrent with the reduction of a-Fe2O3
to Fe3O4 and beginning of iron carbide formation. Therefore, we
propose the origins of these peaks are related with carbon
2.4. Post-reaction analysis
TPO of the micro-reactor samples was performed post-reaction in
situ whilst the large scale reactor was performed ex situ. Oxygen
(5% in He, 70 sccm, BOC Ltd, 99.5%) was introduced to the
sample (ca. 40 mg) and the reactor heated to 1173 K at 5 K minꢀ1
using the mass spectrometer to monitor the eluting gases.
Quantication of the CO2 peak area was achieved by measuring
the CO2 response from the in situ TPO of known masses of
graphite (Sigma Aldrich, 99.9%).27 Powder XRD was performed
using a Siemens D5000 diffractometer, with a Cu Ka radiation in
Bragg–Brentano geometry in the 2q range 5–85ꢁ (step size 0.02ꢁ
sꢀ1). For in situ XRD studies ca. 200 mg of ground sample was
placed in an Anton Paar XRK-900 reaction chamber with a K-type
thermocouple was housed in the reaction chamber. Temperature
control was maintained by an Anton Paar TCU 750 temperature
control unit equipped with a PID control (Eurotherm 2604). A
H2 : CO mixture (2 : 1, 10 sccm, CK Gases, 99.5%) in carrier gas
(Ar, 20 sccm, BOC Ltd, 99.9%) were introduced via 1/4 in. Swa-
gelok tube gas lines, with a thermocouple positioned within the
catalyst bed to ensure accurate temperature reading during
measurements. The sample was heated to 623 K at 5 K minꢀ1 and
maintained at 623 K for 24 h. Diffractograms were recorded every
hour. Reections were assigned based on the following reference
diffraction patterns; a-Fe2O3, JCPDS #13-534; Fe3O4, JCPDS #19-
629; Fe5C2, JCPDS #36-1248; Fe3C, JCPDS #32-0772. Ex situ
Raman scattering was performed using a Horiba Jobin Yvon
LabRam HR confocal Raman microscope and a 532 nm laser
source at <20 mW power. Measurements were taken for approx-
imately 5 min. TEM was performed using a Tecnai T20 micro-
scope equipped with a post-column Gatan Imaging Filter and
operated with an accelerating voltage of 200 keV. Elemental maps
were produced using energy ltered TEM (EFTEM), adopting the
three window technique to distinguish the O–K, C–K and Fe-L2.3
EELS edges from a power law background. Samples were ground
and suspended in MeOH (Sigma Aldrich, 99.8%) before deposi-
tion on a holey carbon grid and dried for insertion in to the
microscope chamber.
deposition via
a
combination of CO dissociation/
disproportionation and carbide formation, indicative of the
'evolutionary' nature of iron oxide FTS catalysis as discussed by
Schulz.29 Table 1 gathers the quantied peak areas for the a,
b and g peaks. Aer 6 h on-stream, the a peak is lost, whilst the
b and g peaks remain. The b peak, attributed to amorphous
carbon, continues to increase in mass during the course of the
reaction. The g peak, ascribed to iron carbide species, levels
aer about 12 h on-stream. The increase in carbon retention is
also observable by Raman spectroscopy (Fig. 1d) and, combined
with the CO conversion, in situ XRD and TPO data, indicates
that the carbon speciation is still in dynamic ux aer 24 h
reaction. Visualization of the 24 h reacted sample by ex situ TEM
identies a core shell structure containing a thick outer layer of
3. Results and discussion
3.1. Micro-reactor characterization
Previous characterization efforts have established the phase of approximately 5 nm (Fig. 2), consistent with other microscopy
the iron oxide catalyst sample, investigated for this study, to be studies of reacted iron FTS catalysts.17,19,30–32 The lattice fringes
a-Fe2O3.19 Those studies also discussed the reaction chemistry observable within the central core (site A) were found to be
¨
during the representative FTS reaction (ambient pressure CO 0.190 nm, consistent with iron carbide (e.g. Hagg carbide,
hydrogenation at 623 K for 6 h).19 The focus for this particular d(600) ¼ 0.191 nm),18 whilst the fringes surrounding the core
investigation is on the temporal behavior of the hydrocarbon suggest a graphitic character to the encapsulating carbon.
retention process during CO hydrogenation. The reaction was Energy ltered micrographs conrm the presence of iron and
repeated for several samples whilst varying the time-on-stream carbon throughout the cores, without signicant oxygen
(0, 3, 6, 12, and 24 h). Fig. S2† presents the reaction prole content. Low levels of oxygen are observed, however, within
obtained aer 24 h. As for the previously reported 6 h reaction,19 a minority of small dense regions that also contain iron but are
three stages are identied: the reduction of the a-Fe2O3 to Fe3O4 smaller than the dominant nanoparticles (Fig. S3†).
2610 | RSC Adv., 2019, 9, 2608–2617
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
Fig. 1 (a) The conversion of CO and selectivity to CO2 and CH4 over the a-Fe2O3 sample during CO hydrogenation at 623 K in the micro-reactor.
(b) In situ XRD with identified reflections attributed to a-Fe2O3 (circle), Fe3O4 (square), and iron carbides (triangle). (c) In situ TPO MS profiles and
(d) selected ex situ Raman spectra of the a-Fe2O3 sample after CO hydrogenation at 623 K in the micro-reactor. TPO profiles are stacked to
facilitate comparison.
These types of structure were also reported in our previous
studies and could either represent unreduced iron oxide or
metallic iron which has re-oxidized under the passivation
conditions exposed to the catalyst before the ex situ analysis.17,19
Table 1 A comparison of the quantified peak area and temperature
max from the TPO studies involving samples from the micro-reactor
and large scale reactor samples
a peak
b peak
g peak
Sample C contentc Tmax
C content Tmax C content Tmax
d
MR-0a
MR-3
MR-6
MR-12
MR-24
LR-3b
LR-6
—
3.32
4.14
—
—
1.04
13.83
16.81
18.09
33.43
16.80
18.42
23.41
26.58
600
600
611
592
618
585
595
597
606
2.07
5.89
8.04
18.51
24.04
2.50
7.30
5.95
11.66
653
628
649
660
674
619
629
630
639
477
494
2.65
465
482
487
500
2.69
10.13
15.98
LR-12
LR-24
a
b
c
Fig. 2 Ex situ HRTEM micrograph of the post-reaction (24 h) a-Fe2O3
sample.
MR – micro-reactor. LR – large scale reactor. Carbon content in
mmolesC. d Tmax in K.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 2608–2617 | 2611

View Article Online
RSC Advances
Paper
longer run times are required to fully reduce the a-Fe2O3 start-
ing material when compared with the micro-reactor arrange-
ment. Ex situ XRD indicates the complete reduction of a-Fe2O3
from 12 h (Fig. 3b). Further, there is an increase in iron carbide
reections between 40 and 50ꢁ up to 24 h. The increase is less
than observed for the in situ measurement where, aer
approximately 5 h on-stream, the diffractogram was dominated
by iron carbide features (Fig. 1b). This observation is consistent
with the inferior solid/gas exchange dynamics of the INS
reactor.
In contrast to the micro-reactor TPO proles (Fig. 1c), the
TPO proles obtained for the large scale reactor samples are
similar in prole shape; the a peak is sharp and dened, whilst
the b peak is broad, containing a high temperature shoulder (g
peak, Fig. 3c). The quantied peak areas are collected in Table 1.
As with the micro-reactor studies, carbon retention increases
during the course of the reaction, particularly the reactive
adsorbed carbon associated with the a peak.
3.2. Large scale reactor characterization
Before INS analysis, the a-Fe2O3 sample was exposed to scaled-
up CO hydrogenation conditions using a custom built appa-
ratus for the preparation of large catalyst samples (ca. 10 g)
suitable for INS analysis.23 The MS reaction prole for the 24 h
reacted sample is presented in Fig. S4.† As with the micro-
reacted sample (Fig. S2†), the large scale reaction prole can
be dened by several stages: the simultaneous production of
CO2, H2O and CH4 once at the reaction temperature, followed
by pseudo-steady state operation aer approximately 12 h on-
stream. It should be noted no wax products were formed
during the course of this measurement, conrming that the
selected operational regime does not induce the formation of
long chain hydrocarbons usually associated with the FTS
process that is favored at elevated pressures. The primary route
of formation of CO2 in the micro-reactor set-up was the reduc-
tion of a-Fe2O3 to Fe3O4 by CO. From the selectivity: conversion
prole in Fig. 3a the sharp decrease in CO2 signal coincides with
the CO signal approaching steady-state. This would indicate the
a-Fe2O3 has been reduced to Fe3O4, decreasing the routes of
CO2 production. Reecting the larger sample mass utilized for
the large scale reaction (10 g versus 40 mg in the micro-reactor),
This species is only present for the rst 6 h of reaction in the
micro-reactor (Fig. 1c), suggesting the large scale operation
enables this carbon species to persist for longer periods of time-
on-stream consistent with the inferior gas/solid exchange
Fig. 3 (a) The conversion of CO and selectivity to CO2 and CH4 over the a-Fe2O3 sample during CO hydrogenation at 623 K in the large scale
reactor. (b) Ex situ XRD with identified reflections attributed to a-Fe2O3 (circle), and iron carbides (triangle). (c) Ex situ TPO MS profiles and (d)
selected ex situ Raman spectra of the a-Fe2O3 sample after CO hydrogenation at 623 K in the large scale reactor. TPO profiles are stacked to
facilitate comparison.
2612 | RSC Adv., 2019, 9, 2608–2617
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
dynamics. Ex situ Raman spectra are dened by the ‘D’ and ‘G’ feature consisting of a main peak at 3052 cmꢀ1 with a notice-
bands indicative of disordered and ordered carbonaceous able shoulder at ca. 2934 cmꢀ1 (Fig. 4a). These bands are
deposits (Fig. 3d).33–35 The small increase in intensity from 3 to assigned to the n(C–H) modes of sp2 and sp3 hybridized
24 h is consistent with the increase in carbon retention carbon.18,19 There is a small but noticeable peak at 3674 cmꢀ1
observed by TPO (Fig. 3c, Table 1).
for the 3 h reacted sample attributed to the n(O–H) mode of
terminal hydroxyl groups. With time-on-stream the intensity of
the n(C–H) features increase, whilst the n(O–H) mode is not
present from 6 h onwards indicating a progressive deposition of
hydrocarbonaceous material and the removal of the small
population of surface hydroxyls.
3.3. Inelastic neutron scattering studies
Fig. 4 and 5 presents the INS spectra of the reacted samples
recorded at an incident energy of 600 and 250 meV respectively.
At 600 meV, the spectra are dened by the presence of a n(C–H)
An advantage of INS is the spectral intensity for a specic
mode is directly proportional to the hydrogen concentration.36
Integration and quantication of the n(C–H) features via
described calibration procedures then enables mode specic
hydrogen retention as a function of time-on-stream to be
established (Fig. S4b and S5†).26 The aliphatic sp3 hybridized
C–H mode increases during the initial stages of reaction before
saturating aer ca. 8 h. In contrast, olenic/aromatic sp2
hybridized C–H species continually increase during the reac-
tion. Correlating the quantied INS trends with the TPO peak
proles (Fig. 3c, Table 1) suggest the sp2 hybridized C–H entities
are related to the steadily increasing b peak, i.e. amorphous
carbon. However, the sp3 hybridized n(C–H) signal does not
correlate with any of the a, b or g peaks present in Fig. 3c. It is
possible that desorption or partial oxidation events could be
affecting the TPO proles.
Spectra recorded at an incident energy of 250 meV are pre-
sented in Fig. 5. Several weak features are present aer 3 h on-
stream (1451, 1389 and 1160 cmꢀ1) that are assigned to
aromatic d(C–H) modes.16–19 The 1451 cmꢀ1 feature is assigned
to semi-circle ring deformation modes that are possibly linked
with a d(C–H) mode associated with perimeter carbons of an
Fig. 4 (a) INS spectra (recorded at 600 meV) of the a-Fe2O3 after CO
hydrogenation at 623 K in the large scale reactor. Spectra are stacked Fig. 5 INS spectra (recorded at 250 meV) of the a-Fe2O3 sample after
to facilitate comparison. (b) The quantified hydrogen content of the CO hydrogenation at 623 K in the large scale reactor. Spectra are
2934 cmꢀ1 (hollow) and 3052 cmꢀ1 (solid) features identified in part (a). stacked to facilitate comparison.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 2608–2617 | 2613

View Article Online
RSC Advances
Paper
extended polycyclic aromatic network.37 The 1389 cmꢀ1 band is on each set of samples are composed of similar material,
assigned to an in-plane ring deformation of a naphthalene-type namely a polyaromatic over-layer possessing some aliphatic
molecule and the 1160 cmꢀ1 band is assigned to a CC–H in- functionality. Moreover, Fig. 7 shows how the density of the
plane deformation mode of a polyaromatic hydrocarbon. The catalyst extracted from the large scale INS reactor is modied as
relatively weak intensity of these features suggests the presence a function of time-on-stream. Catalyst density is seen to
of aromatic species at this stage of the reaction is relatively low. decrease sharply over the period 0–6 h before levelling off at
The remainder of peaks at 3 h can be assigned to the A1g Fe–O longer reaction times. van der Loosdrecht and co-workers report
phonon mode of Fe3O4 (591 cmꢀ1) and a combination of alkenic iron based FTS catalysts to exhibit a decrease in density during
d(C–H) and a possible magnetic interaction associated with reaction; a matter that causes problems in the process envi-
Fe3O4 (941 cmꢀ1).19,38 With time-on-stream, the intensity of the ronment.3 To a rst approximation, the trend evident in Fig. 7 is
sp2 hybridized carbon features between 1000 to 1500 cmꢀ1 a further indication on the relevance of the selected test regime
increase. The signals observed at 871 and 801 cmꢀ1 for the 6, 12 connects with changes in catalyst structure and composition as
and 24 h reacted sample spectra can be assigned to the out-of- encountered in the industrial scenario.
plane C–H deformations of either an olenic or aromatic
Niemantsverdriet's respected competition model states
group.39 The band at 506 cmꢀ1 is assigned to a C–C torsion considers the rates of FTS activity with respect to that of carbide
mode of edge carbon atoms contained within a polycyclic formation.8,40 Here, we adopt those concepts but shi the
aromatic network.37
emphasis to the not unimportant matter of hydrogen supply.
Eqn (1) describes the reaction between chemisorbed carbon and
hydrogen atoms to form C–H products. Under the reaction
conditions utilized for this study, it is suggested that the
formation of amorphous carbon (eqn (2)) is a result of
a hydrogen deciency within the reaction system, where
adsorbed carbon is produced via the dissociative adsorption of
CO. It is assumed that the carbon resides at an iron carbide site
that supports dissociative adsorption of CO.10 Adsorbed
hydrogen atoms diffuse to this site and form C–H bonds. Whilst
a rich hydrogen supply will convey saturated C–H bonds,
a constrained hydrogen supply will lead to unsaturated C–H
bonds. Similarly, the formation of amorphous/graphitic carbon
(C(poly)) or hydrocarbon depends on the hydrogen supply.8,40
3.4. Proposed scheme
Previous studies from this group have investigated an indus-
trially reacted iron based FTS catalyst by INS using the TOSCA
instrument.16 The spectrum is indicative of hydrocarbonaceous
deposits consisting of a polyaromatic overlayer that retains
some aliphatic character.16 As the MAPS spectrometer is char-
acteristically different from the TOSCA spectrometer (direct
versus indirect geometry instruments),25 the 24 h reacted sample
was also analyzed using the TOSCA spectrometer for direct
comparison with the industrially reacted iron sample (Fig. 6).
The spectra exhibit comparable characteristics, particularly in
the deformation region between 800 to 1500 cmꢀ1. These
similarities indicate hydrocarbonaceous components present
xC(ad) + yH(ad) / CxHy
nC(ad) / C(poly)n
(1)
(2)
Fig. 8 is schematic representation that elaborates on
a previously suggested model.16 Roles have now been included
Fig. 6 INS spectral comparison of (a) a reacted commercial grade Fe
FTS catalyst after solvent extraction16 and (b) the a-Fe2O3 sample after Fig. 7 A plot of density of the catalyst extracted from the large scale
CO hydrogenation at 623 K for 24 h using the large scale reactor set- INS reactor as a function of time-on-stream (CO hydrogenation at
up. Both spectra were obtained with the TOSCA spectrometer.
623 K).
2614 | RSC Adv., 2019, 9, 2608–2617
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
for (i) a hydrocarbonaceous overlayer and (ii) the effect of lead to CH4 formation provided the hydrogen supply is suffi-
hydrogen supply. Sites for the dissociation of CO and H2 are cient (eqn (1), site B1). A constrained hydrogen supply will favor
identied with the hydrocarbonaceous overlayer acting as the formation of amorphous/graphitic carbon (eqn (2), site B2).
a mediator for the supply of carbon and hydrogen atoms from Under classical FTS conditions (elevated temperature and high
these sites. The following statements represent a hypothesis, pressure), eqn (1) is supplemented by a C–C propagation reac-
illustrated in Fig. 8, which seeks to describe how the catalyst tion; a rich and dynamic hydrogen supply to this site will then
conditioning regime results in an active catalytic material support formation of high molecular weight saturated products.
capable of sustaining reagent turnover. Firstly, in the reducing The hydrocarbonaceous overlayer may therefore represent the
environment of a syngas feed at elevated temperatures, the a- precursor to both desired FT product and carbonaceous, or
Fe2O3 pre-catalyst is reduced to Fe3O4 which then induces hard, carbon. The fate of the hydrocarbonaceous overlayer will
carbide formation. At this stage, as indicated in Fig. 4 and 5, be dependent on hydrogen supply with hard carbon formation
a hydrocarbonaceous overlayer forms. It is assumed that this occurring under hydrogen lean conditions.
overlayer forms at the surface of the iron carbide.17 Adopting
Clearly, further work is required to evaluate the feasibility of
a similar line-of-thought to that considered for certain hydro- the proposed hypothesis but, nevertheless, it can be used to
genation reactions,41 the coverage of the hydrocarbonaceous rationalize established trends of Fe based catalysts applied to
overlayer is partial, leaving two distinct ‘open’ sites: A and B. CO hydrogenation and FTS reactions. The novelty of this
Site A is associated with the dissociative adsorption of hydrogen particular work is that the INS spectra show, for the rst time,
that provides a continuous source of hydrogen atoms. Site B is that the formation of a hydrocarbonaceous overlayer is an
connected with the dissociative adsorption of carbon monoxide integral part of the catalyst conditioning phase of this intricate
to produce chemisorbed carbon and oxygen atoms. Reaction of catalytic system. Moreover, the INS spectrum of the 24 h sample
adsorbed oxygen atoms at this site with either carbon or is comparable to a commercial grade FTS catalyst extracted
hydrogen will result in the formation of, respectively, CO2 and from
a large-scale coal-to-liquids reactor indicating that
H2O. Alternatively, site B could support reactions described by a hydrocarbonaceous overlayer is present in both samples
eqn (1) (site B1) and 2 (site B2). Under CO hydrogenation (Fig. 6). This outcome validates the use of CO hydrogenation
conditions (elevated temperature but low pressure), this will over a representative iron based FTS catalyst as a test reaction
compatible with the acquisition of informative INS spectra.
4. Conclusions
CO hydrogenation has been investigated as a function of time
over an un-promoted Fe based FTS catalyst and characterized
using TPO, XRD, Raman, TEM and INS. The following conclu-
sions can be drawn.
ꢃ A temporal prole is constructed from the TPO and INS
studies for both the carbonaceous and hydrocarbonaceous
components of the reacted catalyst samples.
ꢃ TPO proles are assigned to reactive carbon species,
amorphous carbon and iron carbide species; the development
of which are dependent on the gas/solid exchange dynamics
afforded by either a micro-reactor or a large scale INS reactor.
ꢃ INS analysis of the reacted catalyst samples identify the
presence of sp2 and sp3 hybridized C–H species, with quanti-
cation of the n(C–H) modes indicating the concentration of
aliphatic C–H species to saturate aer approximately 6 h time-
on-stream. In contrast, the concentration of olenic/aromatic
C–H entities progressively increase as a function of time-on-
stream.
ꢃ The INS spectrum for a catalyst that has experienced a 24 h
reaction period yields a comparable INS spectrum to that of
a commercial grade Fe based FTS catalyst that has been
extracted from a large scale coal-to-liquids unit operation. This
establishes the relevance of the hydrocarbonaceous overlayer
concept to FTS and, moreover, indicates that the test reaction
explores aspects of iron oxide/CO/H2 surface chemistry that has
tangible connectivity to the economically relevant process
chemistry.
Fig. 8 A schematic indicating the composition of the Fe FTS catalyst
after ca. 6 h ambient pressure CO hydrogenation at 623 K. The
hydrocarbonaceous overlayer (highlighted area) is suggested to
mediate the transfer of hydrogen, where under H2-rich conditions,
CxHy prevails (site B1) whilst under H2-lean conditions, carbon poly-
merization occurs (site B2).
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 2608–2617 | 2615

View Article Online
RSC Advances
Paper
ꢃ A schematic illustration (Fig. 8) is presented to dene the 14 E. de Smit, J. F. Creemer, H. W. Zandbergen,
range of components and probable distribution of sites con-
nected with an active Fe based FTS catalyst.
B. M. Weckhuysen and F. M. F. de Groot, J. Phys.: Conf.
Ser., 2009, 190, 012161.
15 T. Herranz, S. Rojas, F. Perezalonso, M. Ojeda, P. Terreros
and J. Fierro, J. Catal., 2006, 243, 199–211.
Conflicts of interest
´
16 N. G. Hamilton, I. P. Silverwood, R. Warringham, J. Kapitan,
L. Hecht, P. B. Webb, R. P. Tooze, S. F. Parker and D. Lennon,
Angew. Chem., Int. Ed., 2013, 52, 5608–5611.
There are no conicts to declare.
17 N. G. Hamilton, R. Warringham, I. P. Silverwood, J. Kapitan,
L. Hecht, P. B. Webb, R. P. Tooze, W. Zhou, C. D. Frost,
S. F. Parker and D. Lennon, J. Catal., 2014, 312, 221–231.
18 R. Warringham, N. G. Hamilton, I. P. Silverwood, C. How,
P. B. Webb, R. P. Tooze, W. Zhou, C. D. Frost, S. F. Parker
and D. Lennon, Appl. Catal., A, 2015, 489, 209–217.
19 R. Warringham, A. R. McFarlane, D. A. MacLaren,
P. B. Webb, R. P. Tooze, J. Taylor, R. Ewings, S. F. Parker
and D. Lennon, J. Chem. Phys., 2015, 143, 174703.
Acknowledgements
Sasol Ltd., the University of Glasgow and the EPSRC (grant
numbers EP/P504937/1 and EP/P505534/1) are thanked for the
provision of postgraduate studentships (RW and ALD). The
STFC Rutherford Appleton Laboratory is thanked for access to
neutron beam facilities. Technical assistance in electron
microscopy was provided by Dr Donald MacLaren and Mr Colin
How (School of Physics and Astronomy, University of Glasgow).
The Royal Society is thanked for the provision of an Industrial
Fellowship (PBW).
20 A. R. McFarlane, I. P. Silverwood, E. L. Norris,
R. M. Ormerod, C. D. Frost, S. F. Parker and D. Lennon,
Chem. Phys., 2013, 427, 54–60.
21 A. R. McFarlane, I. P. Silverwood, R. Warringham,
E. L. Norris, R. M. Ormerod, C. D. Frost, S. F. Parker and
D. Lennon, RSC Adv., 2013, 3, 16577–16589.
References
1 A. P. Steynberg, in Studies in Surface Science and Catalysis, ed. 22 M. D. Shroff and A. K. Dayte, Catal. Lett., 1996, 37, 101–106.
A. Steynberg and M. Dry, Elsevier, 2004, ch. 1, vol. 152, pp. 1– 23 R. Warringham, D. Bellaire, S. F. Parker, J. Taylor,
63.
R. A. Ewings, C. M. Goodway, M. Kibble, S. R. Wakeeld,
M. Jura, M. P. Dudman, R. P. Tooze, P. B. Webb and
D. Lennon, J. Phys.: Conf. Ser., 2014, 554, 012005.
2 E. de Smit and B. M. Weckhuysen, Chem. Soc. Rev., 2008, 37,
2758–2781.
3 J. Van de Loosdrecht, F. G. Botes, I. M. Ciobica, A. Ferreira, 24 P. C. H. Mitchell, S. F. Parker, A. J. Ramirez-Cuesta and
P. Gibson, D. J. Moodley, A. M. Saib, J. L. Visage,
C. J. Weststrate and J. W. Niemantsverdriet, in Fischer-
J. Tomkinson, in Vibrational Spectroscopy Using Neutrons
with Applications in Chemistry, Biology, Materials Science
and Catalysis. Series on Neutron Techniques and
Applications, World Scientic, Singapore, 2005, vol. 3.
Tropsch
Synthesis:
Catalysts
and
Chemistry.
In
Comprehensive Inorganic Chemistry II, ed. J. Reedijk and K.
Poeppelmeier, Elsevier, Oxford, 2013, ch. 7.20, vol. 7, pp. 25 S. F. Parker, D. Lennon and P. W. Albers, Appl. Spectrosc.,
525–557.
2011, 65, 1325–1341.
4 L. P. Dancuart, R. de Haan, and A. de Klerk, in Studies of 26 I. P. Silverwood, N. G. Hamilton, C. J. Laycock,
Surface Science and Catalysis, A. Steynberg and M. Dry,
Elsevier, 2004, ch. 6, vol. 152, pp. 472–522.
5 H. Schulz, Appl. Catal., A, 1999, 186, 3–12.
J. Z. Staniforth, R. M. Ormerod, C. D. Frost, S. F. Parker
and D. Lennon, Phys. Chem. Chem. Phys., 2010, 12, 3102–
3107.
´
6 Y. Traa, Chem. Commun., 2010, 46, 2175–2187.
7 H. Janhangiri, J. Bennett, P. Mahjoubi, K. Wilson and S. Gu,
Catal. Sci. Technol., 2014, 4, 2210–2229.
27 I. P. Silverwood, N. G. Hamilton, A. R. McFarlane, J. Kapitan,
L. Hecht, E. L. Norris, R. M. Ormerod, C. D. Frost, S. F. Parker
and D. Lennon, Phys. Chem. Chem. Phys., 2012, 14, 15214–
15225.
8 J. W. Niemantsverdriet, A. M. Van der Kraan, W. L. Van Dijk
and H. S. J. Van der Baan, J. Phys. Chem., 1980, 84, 3363– 28 H. M. T. Galvis, J. H. Bitter, T. Davidian, M. Ruitenbeek,
3370.
A. I. Dugulan and K. P. de Jong, J. Am. Chem. Soc., 2012,
134, 16207–16215.
9 M. E. Dry, The Fischer-Tropsch Synthesis, Catalysis: Science and
Technology, ed. J. R. Anderson and M. Boudart, Springer- 29 H. Schulz, Top. Catal., 2003, 26, 73–84.
¨
Verlag, New York, 1981, pp. 160–255.
10 E. de Smit, F. Cinquini, A. M. Beale, O. V. Safonova, W. van
Beek, P. Sautet and B. M. Weckhuysen, J. Am. Chem. Soc.,
2010, 132, 14928–14941.
11 J. Xu and C. H. Bartholomew, J. Phys. Chem. B, 2005, 109,
2392–2403.
30 P. C. Thune, P. Moodley, F. Scheijen, H. Fredricksson,
R. Lancee, J. Kropf, J. T. Miller and J. W. Niemantsverdriet,
J. Phys. Chem. C, 2012, 116, 7367–7373.
31 M. D. Shroff, D. S. Kalakkad, K. E. Coulter, S. D. Kohler,
M. S. Harrington, N. B. Jackson, A. G. Sault and
A. K. Datye, J. Catal., 1995, 156, 185–207.
¨
12 S. Li, S. Krishnamoorthy, A. Li, G. D. Meitzner and E. Iglesia, 32 P. C. Thune, C. J. Weststrate, P. Moodley, A. M. Saib, J. van de
J. Catal., 2002, 206, 202–217.
13 L. Niu, X. Liu, X. Liu, Z. Lv, C. Zhang, X. Wen, Y. Yang, Y. Li
and J. Xu, ChemCatChem, 2017, 9, 1691–1700.
Loosdrecht, J. T. Miller and J. W. Niemantsverdriet, Catal.
Sci. Technol., 2011, 1, 689–697.
2616 | RSC Adv., 2019, 9, 2608–2617
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
33 A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner and 38 I. Chamritski and G. J. Burns, J. Phys. Chem. B, 2005, 109,
¨
U. Poschl, Carbon, 2005, 43, 1731–1742.
4965–4968.
34 M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus, 39 D. Lin-Vien, N. B. Colthup, W. G. Fateley and J. G. Graselli,
L. G. Cancado, A. Jorio and R. Saito, Phys. Chem. Chem.
Phys., 2007, 9, 1276–1291.
35 J. Nemanich and S. A. Solin, Phys. Rev. B, 1979, 20, 392–401.
The handbook of infrared and Raman characteristic
frequencies of organic molecules, Academic Press, Boston,
1991.
36 P. W. Albers, D. Lennon and S. F. Parker, Catalysis, in 40 J. W. Niemantsverdriet and A. M. Van der Kraan, J. Catal.,
Neutron Scattering: Applications in Biology, Chemistry, and 1981, 72, 385–388.
Materials Science, ed. F. Fernandez-Alonzo and D. L. Price, 41 A. Borodzinski and A. Cybulski, Appl. Catal., A, 2000, 198, 51–
Elsevier, 2017, vol. 49, pp. 279–348.
66.
37 P. W. Albers, J. Pietsch, J. Krauter and S. F. Parker, Phys.
Chem. Chem. Phys., 2003, 5, 1941–1949.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 2608–2617 | 2617