Application of heteroatom-containing iron(II) piano-stool complexes for the synthesis of shaped carbon nanomaterials

Article abstract of DOI:10.1016/j.jorganchem.2014.12.038

Based on iron(II) piano-stool complexes, five compounds were synthesized by varying the organo-heteroatom ligands in their structure. Two different heteroatoms were considered, namely, nitrogen and sulfur, as constituents of the ligands on the complexes.

Full text of DOI:10.1016/j.jorganchem.2014.12.038

Journal of Organometallic Chemistry 780 (2015) 13e19
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Application of heteroatom-containing iron(II) piano-stool complexes
for the synthesis of shaped carbon nanomaterials
Vincent O. Nyamori*, Muhammad D. Bala, Dennis S. Mkhize
School of Chemistry and Physics, University of KwaZulu-Natal, Private Bag X54001, Durban 4000, South Africa
a r t i c l e i n f o
a b s t r a c t
Article history:
Based on iron(II) piano-stool complexes, five compounds were synthesized by varying the organo-
heteroatom ligands in their structure. Two different heteroatoms were considered, namely, nitrogen
and sulfur, as constituents of the ligands on the complexes. Compounds 1 and 2 contained the nitrogen
heteroatom in their ligands whilst 3 to 5 contained sulfur. These compounds were used as catalysts in the
synthesis of shaped carbon nanomaterials (SCNMs) by means of the floating catalyst chemical vapour
deposition method (CVD). The nitrogen-containing catalysts produced nitrogen-doped carbon nanotubes
(N-CNTs) with bamboo morphology. On the other hand, the sulfur-containing catalysts produced mainly
carbon spheres (CS) and some amorphous carbon (AC) as the bulk of the resulting SCNM products. In the
case of compounds 1 and 2, higher nitrogen content and reaction temperatures were shown to promote
nitrogen-doping which led to more disordered N-CNTs and an increase in the outer diameter of the N-
CNTs. In the case of compounds 3 to 5, higher reaction temperatures led to more graphitic CS and an
increase in the average diameter of the CS. The various results and analyses show that the synthetic
variations of not only reaction conditions, but also the type of substituent on the organometallic catalyst
ligand and its structure, can lead to a level of control towards the resulting carbon nanostructures.
© 2015 Elsevier B.V. All rights reserved.
Received 15 October 2014
Received in revised form
22 December 2014
Accepted 30 December 2014
Available online 9 January 2015
Keywords:
Nitrogen-doped carbon nanotubes (N-CNTs)
Carbon spheres
Organometallic compounds
Chemical vapour deposition (CVD)
Iron(II) complexes
Introduction
production of a particular SCNM [12]. A greener CVD method that
has been used involves the introduction of a catalyst and any other
Shaped carbon nanomaterials (SCNMs) display interesting
properties that are closely related to their shapes and sizes. Hollow
or filled carbon spheres (CSs), carbon nanotubes (CNTs) and ful-
lerenes, amongst other SCNMs, have been shown to have inter-
esting properties and applications across most scientific and
technological fields [1,2]. Some of the properties exhibited by
SCNMs include high surface areas, which permit applicability in
areas such as catalyst supports [3,4], gas adsorption [4,5] and water
purification systems [6,7]. Whilst the formation mechanisms of
SCNMs is still a fascinating phenomenon [8], the synthesis pro-
cedures require a certain level of control towards the desired
products and their dimensions.
reactant/s into a closed environment (autoclave), which is then
heated under autogeneous conditions. This technique avoids the
use of a catalyst support and eliminates the related difficult and
expensive support removal procedures [13].
The addition of heteroatoms during the synthesis of SCNMs by
the CVD method is one way to improve the selectivity and yield
towards specific SCNMs [14]. Depending on the heteroatom
introduced, the carbons on the structure of the SCNMs can be
replaced with the added heteroatom, as a dopant. This can be
achieved by the addition of a heteroatom-containing carbon
source, heteroatom-containing catalysts or a combination of both
approaches.
SCNMs can be synthesized by a variety of techniques which
include chemical vapour deposition (CVD) [9], laser ablation [10]
and arc discharge [11]. Amongst the three, CVD has proven to be
the most facile method that can be made selective towards the
The use of nitrogen-containing compounds has been shown to
result in the incorporation of N-atoms and the production of
nitrogen-doped CNTs (N-CNTs) [15e19]. N-CNTs display a “bamboo-
like” morphology because of the disorder introduced by the nitrogen
[17]. The extent or level of nitrogen-doping in N-CNTs is related to
the distances (segmental length) of the bamboo compartments [18].
The nitrogen atom in N-CNTs may also serve as a functionalization
point for the introduction of a number of other functional groups.
* Corresponding author. Tel.: þ27 31 260 8256; fax: þ27 31 260 3091.
E-mail address: nyamori@ukzn.ac.za (V.O. Nyamori).
http://dx.doi.org/10.1016/j.jorganchem.2014.12.038
0022-328X/© 2015 Elsevier B.V. All rights reserved.

14
V.O. Nyamori et al. / Journal of Organometallic Chemistry 780 (2015) 13e19
Sulfur is another heteroatom that has been introduced during
report from Mohlala et al. has described the synthesis of pristine
multi-walled CNTs by using iron(II) piano-stools [29]. However, we
are now introducing more variable heteroatoms within the catalyst
and exploring the kind and properties of SCNMs obtained at
different conditions.
In this study we report the synthesis of five iron(II) piano-stool
complexes and their use as catalysts for the synthesis of SCNMs via
the CVD method. The first two compounds possess ligands func-
tionalized by nitrogen and oxygen heteroatoms with different
structures and amounts of nitrogen and oxygen. Hence, it would be
interesting to investigate how these two iron(II) piano-stool com-
plexes will affect the level of nitrogen-doping in the resulting
SCNMs. The other three compounds contain the sulfur-heteroatom
incorporated in different chain lengths as ligands. These are also
explored to establish how the presence of sulfur and varying the
ligand chain length affects the yields, morphology and selectivity
towards the SCNMs produced.
SCNM synthesis by means of the CVD method [20]. Unlike nitrogen,
a dopant, sulfur has been reported to act as a promoter when
introduced during the synthesis of SCNMs [21]. Hence, the sulfur
atom is not incorporated into the structures of the SCNMs but has
been shown to increase yields [14] and improve selectivity towards
certain types of SCNMs [8,9]. For example, sulfur as an additive, has
been known to introduce Y-shaped CNTs and aid in the formation of
helical carbon nanotubes (HCNTs) [22]. In addition, the presence of
sulfur, during the synthesis of SCNMs via the CVD method, has been
shown to enhance the encapsulation of ferromagnetic moieties,
thus improving the soft magnetic properties of the CNTs, making
them suitable for applications in electromagnetic devices.
Owing to their volatile nature, organometallic compounds are
commonly used as catalysts in the synthesis of SCNMs via the CVD
method [23]. A good example of such organometallic compounds is
iron pentacarbonyl, Fe(CO)5, which has been successfully used to
synthesize SCNMs [23]. The use of polymers containing organo-
metallic complexes to provide a metal source for CNT synthesis has
also been reported to yield a controlled production of CNTs [24,25].
However, there is need to explore a wider scope of catalysts to
deepen the understanding of catalyst structure-product property
relationships. In the past, we have synthetically modified ferrocene
to incorporate a variety of ligands that alter the catalyst properties
and thus alter the properties of resulting SCNMs [19,26,27]. Incor-
poration of heteroatoms directly into the ligands of the catalyst has
also been shown to be a more effective way of modifying the SCNMs
produced compared to their introduction as external dopants [28].
Our current focus is on the use of iron(II) piano-stool complexes
as catalysts for the synthesis of CNTs and other SCNMs. An initial
Results and discussion
Synthesis of iron(II) piano-stool complexes
The synthesis of iron(II) piano-stool complexes proceeded via a
relatively simple procedure (Scheme 1). This involved reduction of
the dimer (compound A) with a dilute amalgam (sodium and
mercury in THF) at room temperature to form the cyclopentadienyl
dicarbonyl iron anion or cyclopentadienyl dicarbonyl metalate
(Compound B). The nucleophilic anion was reacted in situ with the
desired electrophile in which the various reactions proceeded via a
nucleophilic substitution reaction. The reactions involved initial
Scheme 1. Synthesis of iron(II) piano-stool complexes used as catalysts for the synthesis of SCNMs.

V.O. Nyamori et al. / Journal of Organometallic Chemistry 780 (2015) 13e19
15
Table 2
attack of the nucleophilic anion (metalate) on the electron deficient
carbon centre of the electrophile. The carbon centre of the elec-
trophile was directly bonded to an electronegative halogen which
serves as the driving force for the reaction to proceed.
Yield, distribution, size and thermal stability of SCNM products synthesized with
catalysts 3, 4 and 5 at 800, 850 and 900 ꢀC.
Catalyst Tmax Yield SCNMs
(ꢀC) (mg) (distribution %)
Mean diameter Decomposition
(nm)
temperature (ꢀC)
3
4
5
800
850 183
900 300
35
CSs (89), AC (11)
CSs (95), AC (5)
CSs (96), AC (4)
CSs (93), AC (7)
CSs (95), AC (5)
CSs (98), AC (2)
CSs (94), AC (6)
CSs (98) AC (2)
CSs (98) AC (2)
827
1003
1433
875
1242
1574
988
592
635
611
648
671
685
431
458
465
SCNMs synthesized by using iron(II) piano-stool complexes as
catalysts
800
72
850 244
900 290
800 210
850 234
900 254
Yield of SCNMs
The yields of SCNMs obtained from the nitrogen-containing
catalysts (1 and 2) are summarized in Table 1. For both catalysts,
the yield increased with increase in temperature from 800 to
900 ꢀC. This could imply that the relatively lower temperatures
were not conducive for the carbon precursor dissociation and
dissolution into the iron catalyst or the iron particles were not of
appropriate nanosize. Thus, increasing the reaction temperature
favoured a faster supply of carbon and its dissolution into suitable
catalyst nanoparticles and, subsequent precipitation, resulting in
the formation of more SCNMs (Table 1). The results show an overall
increase in yield with increase in temperature.
1239
1448
CSs: Carbon Spheres; AC: Amorphous Carbons; Tmax
temperature.
:
Maximum set reaction
the sulfur could effectively react with carbon in the gas phase thus
retarding the rate of carbon supply or with the iron nanoparticles
leading to catalyst poisoning.
Also, in general, catalyst 2 (containing one oxygen and two ni-
trogen atoms) showed a higher yield than catalyst 1 (containing
one nitrogen and two oxygen atoms). The higher yield observed
with catalyst 2 is attributed to the relatively lower oxygen content.
Oxygen is known to have a ‘cleaning effect’ since it can react with
carbon (dangling carbon structures on the walls of SCNMs and
amorphous carbon) to form COx which exit the reaction chamber as
exhaust gases [26]. However, an excessive increase in oxygen
abundance reduces the supply of carbon required to form SCNMs
which results in lower yields. In the case of catalyst 1, at 850 ꢀC, it
showed a higher yield than catalyst 2. This is difficult to conclude
on, but we can only stipulate that at this particular temperature
catalyst 1 gives more products, mainly carbon spheres and amor-
phous carbon, but the conditions are not conducive enough for the
formation of CNTs. A similar observation was made by Bepete et al.
who reported a decrease in yield with increase in temperature in
the presence of oxygen [30]. Hence, the reaction mechanism at
different temperatures and the intermediate products formed need
to be probed further.
The yields of SCNMs obtained from the sulfur-containing cata-
lysts (3, 4 and 5) are presented in Table 2. For these catalysts, the
yield increased with increase in temperature in a similar manner as
that observed for the nitrogen-containing catalysts. This further
suggests that higher temperatures are more favourable for the
synthesis of SCNMs when using heteroatom (nitrogen or sulfur)
functionalized iron(II) piano-stool complexes as catalysts. At a
synthesis temperature of 800 ꢀC the yield increased with increase
in the catalyst carbon content (i.e. 3 < 4 < 5). At synthesis tem-
peratures of 850 and 900 ꢀC the yields tended to decrease with
increase in ligand chain length at similar temperature conditions,
except for catalyst 4 at 850 ꢀC. This phenomenon was contrary to
expectations; a possible explanation is that at higher temperatures
Effect of nitrogen and sulfur on the morphology of SCNMs
TEM and SEM analysis revealed that the nitrogen-containing
catalysts 1 and 2 formed N-CNTs and CSs as the only SCNMs
(Table 1). At a temperature of 800 ꢀC catalyst 1 yielded CSs as the
only SCNMs. An increase in temperature to 850 and then 900 ꢀC led
to the formation of both N-CNTs and CSs with the higher temper-
ature favouring the formation of more N-CNTs and a decrease in
formation of CSs. In the case of catalyst 2, a mixture of N-CNTs and
CSs formed at all synthesis temperatures (Table 1). As the tem-
perature was increased from 800 to 900 ꢀC the abundance of N-
CNTs increased while that of CSs decreased. This implied that for
both catalysts 1 and 2 higher synthesis temperatures are favourable
for N-CNT production while the relatively lower temperatures
favour CS production. It is important to note that catalyst 2 yielded
N-CNTs at all temperatures indicating that it is more effective in the
synthesis of N-CNTs compared with catalyst 1. Catalyst 2 had a
higher nitrogen content and lower oxygen content than catalyst 1. It
is therefore plausible that the higher nitrogen content facilitated N-
CNT growth while the higher oxygen content retarded N-CNT
growth [31]. Similarly, Mohlala et al. [32] have observed that the
presence of oxygen during synthesis can poison the catalyst thus
reducing CNT formation.
TEM images of N-CNTs obtained by using catalysts 1 and 2
exhibited the presence of bamboo compartments (Fig. 1). Such
bamboo compartments are associated with nitrogen-doping of
CNTs [17]. The presence of nitrogen was also evidenced from
qualitative analysis of the N-CNTs by use of energy dispersive X-ray
spectroscopy (EDS) (Supplementary Material S1). In some instances
the N-CNTs exhibited
a coiled morphology (Supplementary
Material S1) similar to that observed in a previous study in which
nitrogen was used during synthesis [33]. The N-CNTs synthesized
from catalyst 2 contained more bamboo compartments of shorter
Table 1
Yield, distribution, size and thermal stability of SCNM products synthesized with catalysts 1 and 2 at 800, 850 and 900 ꢀC.
Catalyst
Tmax (ꢀC)
Yield (mg)
SCNMs (distribution %)
Bamboo compartment length (nm)
aMean diameter (nm)
Decomposition temperature (ꢀC)
1
800
850
900
800
850
900
63
182
254
88
165
289
CSs (80), AC (20)
e
e
T (25), CSs (70), AC (5)
T (35), CSs (60), AC (5)
T (20), CSs (70), AC (10)
T (40), CSs (50), AC (10)
T (50), CSs (45), AC (5)
65.4
40.1
32.3
30.1
28.8
44
41
22
23
25
509
477
525
517
498
2
T: tubes; CSs: Carbon Spheres; AC: Amorphous Carbons; Tmax: Maximum set reaction temperature.
a
Tube mean outer diameter (OD).

16
V.O. Nyamori et al. / Journal of Organometallic Chemistry 780 (2015) 13e19
Fig. 1. TEM images of N-CNTs obtained at 900 ꢀC from (a) catalyst 1 showing long bamboo compartments and (b) catalyst 2 showing short bamboo compartments.
lengths compared with those synthesized from catalyst 1 (Fig.1 and
Table 1). Since the bamboo compartments are brought about by
nitrogen-doping, the higher quantities and reduced lengths of the
bamboo compartments implied a higher N-content in N-CNTs from
catalyst 2 [18]. This is attributed to the presence of more nitrogen
atoms in catalyst 2 than in catalyst 1, which resulted in increased
nitrogen-doping levels of N-CNTs.
It is also important to note that the length of the bamboo
compartments of N-CNTs synthesized from catalyst 1 and 2
decreased with increase in synthesis temperature (Table 1). This is
an indication that as the temperature was increased the level of
nitrogen-doping also increased which suggests that a higher tem-
perature is a more favourable condition for N-CNT formation.
Comparable observations were made by Keru et al., who also re-
ported an increase in nitrogen-doping with increase in temperature
[19]. Other reports, however, indicate that an increase in temper-
ature results in a decrease in the amount of nitrogen incorporated
into the CNTs [34]. The discrepancy between the results we ob-
tained and those reported by other authors is attributed not only to
the different methods used during synthesis but also the kind of
catalyst and nitrogen source used.
The average outer diameters (OD) of N-CNTs obtained with
catalysts 1 and 2 are summarized in Table 1. Representative OD
distribution histograms are available in the Supplementary
Material (S2). The mean ODs of N-CNTs obtained with catalyst 2
are smaller than those from catalyst 1. This can be due to more
nitrogen-doping in N-CNTs obtained from catalyst 2 or smaller iron
catalyst nanoparticles formed during synthesis [35e37]. It was also
observed that Tmax influenced the OD of the N-CNTs. The OD of the
N-CNTs obtained from catalyst 2 increased with increase in tem-
perature. The increase in OD with increasing temperature is in
accordance with other reports [19]. A possible explanation is that as
the temperature increases, collisions between iron catalyst parti-
cles increased resulting in more frequent collisions of iron nano-
particles (NPs) and more coalescence, which resulted in larger iron
NPs and thus the production of larger ODs for the N-CNTs. However,
very large particle sizes are known to be unsuitable for N-CNT
formation, instead they are known to favour the formation of CSs,
and if the size becomes too large, then AC is realized [26].
The sulfur containing catalysts 3, 4 and 5 yielded CSs as the only
SCNMs (Table 2 and Fig. 2). Thus, the presence of sulfur in the
catalysts seemed to promote the formation of CSs and not CNTs or
any other SCNMs. Other literature reported the effects of sulfur to
be concentration dependant, with high concentrations of sulfur
resulting in catalyst poisoning [26]. Although the amount of sulfur
atoms introduced in the reactor via the catalyst is low, the inter-
atomic distance of the heteroatom to the iron centre in organo-
metallic complexes has been shown to lead to the same effects as
high concentrations of the heteroatom. It is possible that poisoning
of the catalyst occurred and that spheres were formed without the
aid of the metal catalyst. In a related study, Mohlala et al. have
studied the effects of sulfur on the synthesis of SCNMs by using
ferrocenyl sulfide or ferrocene mixed with either thiophene or S8
[38]. Their findings show that the presence of large amounts of
sulfur in the reactant mixture generated only amorphous carbon,
while lower amounts of sulfur led to mixtures of MWCNTs and
carbon fibres. Also, the product distribution, yield and tube di-
ameters varied with the sulfur content.
The diameters of the CSs obtained with catalysts 3, 4 and 5 were
measured from the TEM images. For each reaction temperature, at
least 100 CSs were measured and the mean diameters are pre-
sented in Table 2. Representative histograms of the diameter dis-
tribution of CSs are available in the Supplementary Material (S3).
The average diameters of the CSs increased with an increase in the
Tmax for all the sulfur-containing catalysts. As the temperature
increased, more carbon deposition may have occurred on the CSs
which lead to an increase of the mean diameter.
As the ligand chain length increased, the mean diameters of CSs
increased for samples obtained at a temperature of 800 ꢀC (i.e.
827 < 875 < 988 nm). However, at 850 and 900 ꢀC the mean di-
ameters of CSs increased with increase of chain length from catalyst
3 to 4 then reduced with catalyst 5. This could imply that at 800 ꢀC,
as the ligand chain length increases, more carbon is present in the
catalysts leading to more carbon deposition on the CSs, which re-
sults in an increase of the diameters. In the case of catalyst 5, there
was a general increase of diameter (988 < 1239 < 1448 nm) with
increase in temperature, but compared with catalyst 4, specifically
at 850 and 900 ꢀC, we would have expected larger ODs, i.e. 1242 vs.
1239 and 1574 vs. 1448 nm at the respective temperatures. We can
only stipulate that the OD size indifference, which in this case is not
very significant, could have been as a result of other factors such as
the difference in geometry (phenyl vs. alkane structure) and other
physical properties of the three catalysts.
The TEM and SEM images of the CSs showed the presence of
fused carbon spheres (Fig. 2). It is not unusual for spheres to
aggregate and form necklace-like structures held together by van
der Waals forces when they have diameters of 1000 nm or less [1].
These are normally formed during the synthesis and are not a result
of post-synthetic treatments [1].
Crystallinity of SCNMs
The crystallinity of the SCNMs synthesized with catalysts 1 to 5
was determined by means of Raman spectroscopy. Two prominent
peaks were observed at 1553 and 1350 cm-1. The peak at 1553 cm-1
represents the graphitic band (G-band) associated with graphitic
carbons [39], while the peak at 1350 cm-1 represents the disorder

V.O. Nyamori et al. / Journal of Organometallic Chemistry 780 (2015) 13e19
17
Fig. 2. (a) TEM image and (b) SEM image of CSs synthesized by using catalyst 5 at 850 ꢀC.
band (D-band) associated with vibrations between carbon atoms of
disordered graphene sheets [40]. The ratio of the integrated areas of
the D- and G-bands (ID/IG) was used to provide an indication of the
crystallinity of the SCNMs. A relatively high ID/IG ratio is indicative
of a less crystalline material, whilst a lower ratio indicates a more
crystalline material. Fig. 3 displays the ID/IG ratios for SCNMs ob-
tained from all catalysts at different temperatures.
For the sulfur-containing catalysts, a reaction temperature of
800 ꢀC generally yielded the most disordered CSs as they had a
higher ID/IG ratio compared with CSs obtained at 850 and 900 ꢀC.
Crystallinity was observed to increase as the temperature was
increased (decreasing ID/IG). This is in agreement with literature,
where higher temperatures have been shown to be conducive for the
formation of more graphitic CSs [1]. Under similar Tmax, the crys-
tallinity of the CSs was observed to increase with increase in ligand
chain length, with catalyst 5 producing the most crystalline CSs.
For the nitrogen-containing catalysts 1 and 2, the ID/IG ratios
generally increased with rise in synthesis temperature. This indi-
cated that the N-CNTs became less crystalline as the temperature
was increased. A decrease in crystallinity can be caused by
increased nitrogen-doping levels in N-CNTs or an increase in the
amount of AC [19]. Since the amount of AC did not increase with
increase in temperature (Table 1), the decreased crystallinity was
mainly attributed to the increase of nitrogen content in the N-CNTs
with increasing temperatures. This observation further supports
the increase in nitrogen-doping with increase in temperature
observed from TEM analysis (see Effect of nitrogen and sulfur on
the morphology of SCNMs section). Additionally, a higher ID/IG ra-
tio was observed for the samples obtained from catalyst 2
compared with those of catalyst 1. This was attributed to higher
incorporation of the nitrogen atoms into the structures of the N-
CNTs (nitrogen-doping) obtained with catalyst 2 compared with
those produced from catalyst 1. This is also consistent with data
presented in Effect of nitrogen and sulfur on the morphology of
SCNMs section under the discussion on the bamboo compart-
ments, sizes and implications.
Dependence of thermostability of SCNMs on nitrogen and sulfur
Thermogravimetric analysis (TGA) measurements were con-
ducted to obtain information on the thermal stability of the SCNMs
obtained from the various catalysts. The TGA runs were conducted
in triplicates to verify reproducibility. The TGA decomposition
temperatures of the synthesized materials are listed in Tables 1 and
2 The TGA curves of SCNMs obtained from all catalysts are provided
in the Supplementary Material (S4). For the SCNMs synthesized
with catalysts 1 and 2, there was a decrease in thermal stability as
the temperatures increased. This correlated with the increased
level of nitrogen-doping in the N-CNTs as the temperature was
increased (see Effect of nitrogen and sulfur on the morphology of
SCNMs section). The increase in nitrogen-doping levels increased
the density of structural defects in N-CNTs thus reducing their
thermal stability [30].
In the case of the sulfur-containing catalysts, it was generally
observed that the decomposition temperatures increased as the
Tmax was increased. This trend implied that higher temperatures
enhanced the production of more graphitic spheres. This observa-
tion also supports the observed decreasing trend in the ID/IG ratios
as the temperature increased. However, catalyst 3 did not follow
this trend, with a temperature of 850 ꢀC, producing the most
thermally stable spheres for this catalyst. A possible explanation
could be that the CSs obtained with catalyst 3 at a reaction tem-
perature of 850 ꢀC had a narrower diameter distribution compared
with the other two reaction temperatures for this catalyst. A nar-
row diameter distribution of spheres can exhibit narrow decom-
position temperature ranges. The CSs obtained at 800 and 900 ꢀC
could possibly have a wider diameter distribution leading to a
combination of low and high decomposition temperatures within
the same sample thus resulting in an overall lowering of the
decomposition temperature of the whole sample.
The most thermally stable CSs were obtained with catalyst 4. As
the ligand chain length increased from catalyst 3 to 4, thermal
stability of CSs synthesized under similar Tmax conditions was
observed to also increase. A further increase in the ligand chain
length (catalyst 5) resulted in a decrease of the thermal stability of
the CSs. This trend cannot be easily and conclusively explained
Fig. 3. ID/IG ratios of SCNMs synthesized with catalysts 1 to 5 at 800, 850 and 900 ꢀC.

18
V.O. Nyamori et al. / Journal of Organometallic Chemistry 780 (2015) 13e19
especially when, in general, products from catalyst 5 were deemed
to be more crystalline by Raman analysis (see Crystallinity of
SCNMs section) and the only differences, which were also not
very consistent, were on the effect of diameter size and the dis-
tribution of the SCNMs on the thermal stability. Hence, this needs
further probing.
All solvents were dried and freshly distilled before use. Reactions
were carried out under ultra-high-purity N2 or Ar atmospheres
(Afrox, South Africa). Analytical and reagent grade solvents were
used for reactions. Dicyclopentadiene (diCp) (97%), pentane, tetra-
hydrofuran (THF), toluene, dichloromethane (DCM) and diethyl
ether were products of Merck (Schuchardt, Germany). Deuterated
chloroform (CDCl3) was supplied by SigmaeAldrich Chemical Co. (St.
Louis, USA). Argon gas of ultra-high-purity (99.999e100%,
O2 ꢁ 3 ppm, moisture ꢁ 2 ppm) was purchased from Afrox Limited
Gas Co. (Durban, South Africa). Silica gel (0.040e0.063 mm), neutral
aluminium oxide 90 (0.063e0.200 mm), and thin-layer chroma-
tography plates were all supplied by Merck (Schuchardt, Germany).
Conclusions
Nitrogen- and sulfur-containing iron(II) piano-stool complexes
were successfully synthesized and used as novel catalysts for the
synthesis of N-CNTs and CSs. Nitrogen-containing iron(II) piano-
stool complexes were selective towards N-CNT formation while,
sulfur-containing complexes were selective towards formation of
CSs. Nitrogen and sulfur heteroatoms were also used to modulate
size, crystallinity and thermostability of the formed SCNMs.
Increasing the nitrogen content in the catalyst, increased the level
of nitrogen-doping in N-CNTs, which in turn had an effect on the
physical properties of the N-CNTs. Overall, catalyst 2 was better
than catalyst 1 for the synthesis of higher nitrogen-containing N-
CNTs. Varying the chain length of the sulfur containing compounds
altered the physical properties of the formed CSs.
Synthesis and characterization of iron(II) piano-stool complexes
General procedure
Synthesis of dicyclopentadienyl iron carbonyl dimer (compound
A, Scheme 1) was carried out as previously reported [41]. In brief,
dicyclopentadiene and iron pentacarbonyl were reacted under an
atmosphere of argon for 8 h at 150 ꢀC. After cooling and filtering,
the crude product was recrystallized by using a mixture of pentane
and dichloromethane (3:1 v/v) to obtain compound A.
A study on synthesis temperature proved to be a useful approach
to tuning the physical properties of the synthesized SCNMs. For
example, in the case of nitrogen-containing catalysts, higher tem-
peratures favoured higher nitrogen-doping levels in the N-CNTs,
which in turn influenced the crystallinity and thermostability of
these samples. While in the case of the sulfur-containing catalysts,
the crystallinity of CSs increased with increase in temperature for
catalysts 3 and 5 and also, in general, larger diameters were realized.
Thus, synthesis conditions, the heteroatoms present in organome-
tallic catalysts, and the catalyst ligand chain length, play vital roles in
controlling the obtained SCNMs, which in turn can be used to tune
the physical properties of the products obtained.
The procedure for the synthesis of the compounds 1, 2, 3, 4 and 5
(Scheme 1) was similar, and related to a previously reported pro-
cedure [42]. Briefly, an amalgam was formed by reacting mercury
with sodium. Compound A was added to the amalgam to form the
nucleophile/salt, compound B. Compound B was subsequently
introduced into a beaker and reacted with one of the following
electrophiles: N-(bromomethyl)phthalimide, 3-bromomethyl-
2(1H)-quinoxalinone, 2-chloroethyl methyl sulphide, 2-chloroethyl
ethyl sulphide or 2-chloroethyl phenyl sulphide to form com-
pounds 1, 2, 3, 4 or 5 respectively. The experimental procedures for
the synthesis and characterization of compounds 1 to 5 are detailed
in the Supplementary Material (S5). Herein, a brief description of
the experimental procedures for the synthesis and characterization
of two new compounds (1 and 2) are detailed. All the compounds
were characterized by means of standard methods including
melting point determination, 1H NMR, 13C NMR, IR studies and
mass spectrometry as detailed below.
Experimental
Instrumentation
A Bruker 400 MHz Advance or 600 MHz Ultrashield NMR
spectrometer was used to obtain the 1H and 13C NMR spectra at
room temperature in CDCl3 as the solvent. Melting points were
recorded on a Stuart Scientific, model SMP3, melting point appa-
ratus. For each sample three reproducible melting point readings
were taken and averaged. Elemental analyses were performed on a
LECO CHNS-932 elemental analyser which was standardized with
acetanilide. Products of the SCNM synthesis were characterized by
transmission electron microscopy (TEM) (JEOL JEM 1010), scanning
electron microscopy (SEM), (JEOL 2100), thermogravimetric anal-
ysis (TGA) (TA Instruments Q Series™ thermal analyzer DSC/TGA
(Q600)) and Raman spectroscopy (Delta Nu Advantage 532™).
Synthesis of dicarbonyl(ƞ5-cyclopentadienyl) (N-
phthaloylaminomethyl)iron(II) (1)
The general procedure outlined in General procedure section
was followed to obtain compound 1 as a yellow powder. Yield:
(1.074 g, 80%): m.p. 181 ꢀC, 1H NMR (400 MHz, CDCl3, ppm)
dH ¼ 4.05 (2H, s, FeeCH2), 4.91 (5H, s, C5H5), 7.70 (4H, m, C6H4). 13
C
NMR (100 MHz, CDCl3, ppm) dC ¼ 11.14 (FeeCH2), 85.36 (Cp moi-
ety), 122.5 (2C, Ar-CH), 132.8 (2C, AreC), 133.4 (2C, Ar-CH), 168.7
(NCO), 215.5 (terminal CO). IR (ATR, cmꢂ1) 1997 (terminal CO), 1951
(terminal CO), 1751, 1701. CHN analysis for C16H11FeNO4 Calculated:
C; 57.01, H; 3.29, N; 4.15. Found: C; 57.16, H; 3.34, N; 4.18.
Chemical reagents, solvents and gases
Synthesis of dicarbonyl(ƞ5-cyclopentadienyl) (quinoxalin-2(1H)-
onemethyl)iron(II) (2)
All reagents were synthetically pure and used as supplied unless
otherwise stated. Iron pentacarbonyl (97%), 2-chloroethyl methyl
sulfide (97%), 2-chloroethyl ethyl sulfide (98%), 2-chloroethyl
phenyl sulfide (98%), 2-chloro-N,N-dimethylpropylamine hydro-
chloride (98%), N-(bromomethyl)phthalimide (97%) and 2-(4-
chlorophenyl)ethylamine (98%) were products of SigmaeAldrich
Chemical Co. (St. Louis, USA). Sodium (99.5%) and mercury (99.5%)
The general procedure outlined in General procedure section
was followed to obtain compound 2 as a yellow powder. Yield:
(0.952 g, 69%): m.p. 149 ꢀC. 1H NMR (600 MHz, CDCl3, ppm)
dH ¼ 2.66 (2H, s, FeeCH2), 4.87 (5H, s, C5H5), 5.81 (1H, s, NH), 7.35
3
3
(2H, dd, m, C6H2), 7.51 (1H, d, JHH 8.0, C6H), 7.83 (1H, d, JHH 8.0,
C6H). 13C NMR (150 MHz, CDCl3, ppm) dC ¼ 10.03 (FeeCH2), 85.04
(Cp moiety), 114.4 (Ar-CH), 124.1 (Ar-CH), 125.5 (Ar-CH), 130.9 (Ar-
CH), 133.5 (AreC), 156.3 (AreC), 163.8 (NCO), 217.4 (terminal CO). IR
(ATR, cmꢂ1): 3406 (NeH stretch), 1998 (terminal CO), 1942 (ter-
minal CO),1658. CHN analysis for C16H12FeN2O3 Calculated: C; 57.17,
H; 3.60, N; 8.33. Found: C; 57.49, H; 3.65, N; 8.41.
€
€
were supplied by Riedel de Haen (Seelze, Germany) and Rochelle
Chemicals (Johannesburg, South Africa), respectively. The 3-
bromomethyl-2(1H)-quinoxalinone (90%) was supplied by Alfa
Aesar (Karlsruhe, Germany).

V.O. Nyamori et al. / Journal of Organometallic Chemistry 780 (2015) 13e19
19
Synthesis and characterisation of SCNMs
References
[1] A.A. Deshmukh, S.D. Mhlanga, N.J. Coville, Mater. Sci. Eng. R. 70 (2010) 1.
[2] F.L.M. De Volder, H.T. Sameh, R.H. Baughman, A.J. Hart, Science 339 (2013)
A CVD floating catalyst method was used for the synthesis of the
CNTs, N-CNTs and other SCNMs by using compounds 1, 2, 3, 4 and 5
as catalysts. The setup of the reactor used was based on similar de-
535.
[3] L.M. Ombaka, P. Ndungu, V.O. Nyamori, Catal. Today 217 (2013) 65.
[4] Y.C. Liu, X.P. Qiu, Y.Q. Huang, W.T. Zhu, Carbon 40 (2002) 2375.
[5] Y. Feng, W. Yang, N. Wang, W. Chu, D. Liu, J. Anal. Appl. Pyrolysis 107 (2014)
signs reported in the literature [27]. Briefly,
a quartz tube
(850 ꢃ 27 mm i.d.), used as the reactor vessel, was placed inside a
muffle/tube furnace (model no. TSH12/50/610, Elite Thermal Sys-
tems Ltd) fitted with a main zone furnace controller (Eurotherm
2416). At one end of the quartz reactor tube a quartz water-cooled
injection port was fitted, and at the other end a glass cold finger
was fitted. The carrier gas and reducing agent,10% hydrogen in argon
(v/v), was supplied to the system at a rate ofꢂ1100 mL minꢂ1 at 80 kPa.
The furnace was set to heat at 10 ꢀC min to reach the desired
maximum temperature (Tmax ¼ 800, 850 and 900 ꢀC) for the various
reactions. The system was purged for 10 min with the carrier gas, a
mixture of hydrogen/argon gas (10:90H2:Ar (v/v)), before each re-
action. A solution of 2.5 wt.% catalyst/toluene was prepared for all the
reactions carried out, by weighing out 0.275 g of catalyst in 11 mL
toluene. Once the set Tmax was reached and purging completed,
7.5 mL of catalyst solution was injected with a syringe (SS 150 mm L
non-boring 24 gauge bevelled-tip needle) into the quartz reactor.
The furnace was maintained at the set Tmax for 30 min; afterwards
the reactor was allowed to cool to room temperature.
Only carbonaceous material deposited in the hot zone of the
reactor was collected for analysis. The SCNMs produced by these
iron(II) piano-stool catalysts were characterized by TEM and SEM to
determine their type, distribution and morphology. Energy
dispersive X-ray spectroscopy (EDS) was used to provide qualitative
evidence of heteroatom-doping, in this case nitrogen-doping and is
not presented herein. The distribution and average size of the
SCNMs was determined by counting procedures and the results
represent average values obtained from electron micrographs. The
nanotube yield was calculated from the wt.% of the products ob-
tained relative to the mass of the solution injected into the system.
Raman spectroscopy and TGA were used to characterize the crys-
tallinity and thermal stability, respectively. The effects of reaction
temperature (Tmax) on the SCNMs distribution and yield were also
investigated.
204.
[6] X. Wang, J. Liu, W. Xu, Colloids Surf., A: Physicochem. Eng. Asp. 415 (2012)
288.
[7] R. Das, M.E. Ali, S.B.A. Hamid, S. Ramakrishna, Z.Z. Chowdhury, Desalination
336 (2014) 97.
[8] N.J. Coville, S.D. Mhlanga, E.N. Nxumalo, A. Shaikjee, S. Afr. J. Sci. 107 (2011)
01.
[9] I. Martin-Gullon, J. Vera, J.A. Conesa, J.L. Gonzalez, C. Merino, Carbon 44 (2006)
ꢀ
1572.
~
[10] E. Munoz, W.K. Maser, A.M. Benito, M.T. Martínez, G.F. de la Fuente,
Y. Maniette, A. Righi, E. Anglaret, J.L. Sauvajol, Carbon 38 (2000) 1445.
[11] T.W. Ebbesen, P.M. Ajayan, Nature 358 (1992) 220.
[12] K. Mukul, A. Yoshinori, J. Nanosci. Nanotechnol. 10 (2010) 3739.
[13] V.O. Nyamori, N.J. Coville, Organometallics 26 (2007) 4083.
[14] L. Ci, Z. Rao, Z. Zhou, D. Tang, X. Yan, Y. Liang, D. Liu, H. Yuan, W. Zhou,
G. Wang, W. Liu, S. Xie, Chem. Phys. Lett. 359 (2002) 63.
[15] E.N. Nxumalo, V.P. Chabalala, V.O. Nyamori, M.J. Witcomb, N.J. Coville,
J. Organomet. Chem. 695 (2010) 1451.
[16] E.N. Nxumalo, N.J. Coville, Materials 3 (2010) 2141.
[17] X. Li, G. Zhu, Z. Xu, Thin Solid Films 520 (2012) 1959.
[18] S.D. Mhlanga, E.N. Nxumalo, N.J. Coville, V.V. Srinivasu, Mater. Chem. Phys.
130 (2011) 1182.
[19] G. Keru, P.G. Ndungu, V.O. Nyamori, J. Nanomater 2013 (2013) 7.
[20] L. Tapaszto, K. Kertesz, Z. Vertesy, Z.E. Horvath, A.A. Koos, Z. Osvath, Z. Sarkozi,
A. Darabont, L.P. Biro, Carbon 43 (2005) 970.
[21] C. Tongxiang, L. Ruitao, H. Zheng-hong, K. Feiyu, W. Kunlin, W. Dehai,
Nanoscale Res. Lett. 6 (2011) 77.
[22] C. Valles, M. Perez-Mendoza, P. Castell, M.T. Martínez, W.K. Maser,
A.M. Benito, Nanotechnology 17 (2006) 4292.
[23] V.O. Nyamori, S.D. Mhlanga, N.J. Coville, J. Organomet. Chem. 693 (2008) 2205.
[24] C. Hinderling, Y. Keles, T. Stockli, H.F. Knapp, T. de los Arcos, P. Oelhafen,
I. Korczagin, M.A. Hempenius, G.J. Vansco, R. Pugin, H. Heinzelmann, Adv.
Mater. 16 (2004) 876.
[25] E. Lahiff, C.Y. Ryu, S. Curran, A.I. Minett, W.J. Blau, P.M. Ajayan, Nano Lett. 3
(2003) 1333.
[26] R.S. Oosthuizen, V.O. Nyamori, Appl. Organomet. Chem. 26 (2012) 536.
[27] V.O. Nyamori, E.N. Nxumalo, N.J. Coville, J. Organomet. Chem. 694 (2009)
2222.
[28] E.N. Nxumalo, V.O. Nyamori, N.J. Coville, J. Organomet. Chem. 693 (2008)
2942.
[29] M.S. Mohlala, N.J. Coville, J. Organomet. Chem. 692 (2007) 2965.
[30] G. Bepete, Z.N. Tetana, S. Lindner, M.H. Rümmeli, Z. Chiguvare, N.J. Coville,
Carbon 52 (2013) 316.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ €
ꢀ
ꢀ
ꢀ
[31] S. Liu, Y. Zhang, Y. Lin, Z. Zhao, Q. Li, Carbon 69 (2014) 247.
[32] M.S. Mohlala, X.-Y. Liu, N.J. Coville, J. Organomet. Chem. 691 (2006) 4768.
[33] M. Yudasaka, R. Kikuchi, Y. Ohki, S. Yoshimura, Carbon 35 (1997) 195.
[34] P.S.D. Ram Manohar Yadav, T. Shripathi, R.S. Katiyar, O.N. Srivastava, Nano-
scale Res. Lett. 4 (2009) 197.
[35] C.L. Cheung, A. Kurtz, H. Park, C.M. Lieber, J. Phys. Chem. B 106 (2002) 2429.
[36] V. Jourdain, C. Bichara, Carbon 58 (2013) 2.
[37] C. Lu, J. Liu, J. Phys. Chem. B 110 (2006) 20254.
[38] M.S. Mohlala, X.-Y. Liu, M.J. Witcomb, N.J. Coville, Appl. Organometal. Chem.
21 (2007) 275.
[39] C. Thomsen, S. Reich, Phys. Rev. Lett. 85 (2000) 5214.
[40] M.A. Pimenta, G. Dresselhaus, M.S. Dresselhaus, L.G. Cancado, A. Jorio, R. Saito,
Phys. Chem. Chem. Phys. 9 (2007) 1276.
Acknowledgements
This research is financially supported by the National Research
Foundation (NRF)(Grant number 91529 and 76635) and the Uni-
versity of KwaZulu-Natal (UKZN), for which we are grateful. Dr.
Vincent O. Nyamori also thanks Mr Atang Sauli and Miss Lucy M.
Ombaka their contribution in drafting the manuscript. We are also
grateful to Prof. Bice S. Martincigh and Dr. Nolwazi Nombona for
their critical comments of the manuscript.
[41] G. Girolami, T. Rauchfuss, R. Angelici, Synthesis and Technique in Inorganic
Chemistry, third ed., University Science Books, Sausalito, 1999, p. 177.
[42] T.S. Piper, G. Wilkinson, J. Inorg. Nucl. Chem. 3 (1956) 104.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jorganchem.2014.12.038.