A family of oxide-carbide-carbon and oxide-nitride-carbon nanocomposites

Article abstract of DOI:10.1039/c3cc47480a

This paper describes a powerful and versatile approach that combines the benefits of sol-gel processing with controlled phase separation to yield oxide-carbide-carbon or oxide-nitride-carbon nanocomposites. the Partner Organisations 2014.

Full text of DOI:10.1039/c3cc47480a

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
A family of oxide–carbide–carbon and
oxide–nitride–carbon nanocomposites†
Cite this: Chem. Commun., 2014,
0, 5364
5
ab
ac
d
e
e
e
Z. Schnepp,* M. J. Hollamby, M. Tanaka, Y. Matsushita, Y. Xu and Y. Sakka
Received 30th September 2013,
Accepted 25th November 2013
DOI: 10.1039/c3cc47480a
www.rsc.org/chemcomm
This paper describes a powerful and versatile approach that combines
Recently, we have shown that abundant biopolymers such as
14
15
the benefits of sol–gel processing with controlled phase separation to alginate are effective precursors for synthesis of oxide, nitride or
yield oxide–carbide–carbon or oxide–nitride–carbon nanocomposites. carbide nanostructures. Biopolymers are particularly suited to
16
control of particle size in ceramics due to their ability to bind
17
Transition metal carbides and nitrides are valuable materials strongly to metal cations. Remarkably, it is also possible to force
1
for applications such as fuel cell electrocatalysis, ammonia oxide/carbide or oxide/nitride phase separation from these homo-
2
3
synthesis and photocatalytic water splitting. In these systems, the geneous precursors by exploiting differences in stability of inter-
18
properties often depend on control of particle size, morphology mediate oxides. So far, however, full separation was only achieved
and accessible surface area. Sol–gel, solvothermal and tem- in one system (Fe C–MgO). In this paper, we demonstrate the
plating processes have all been used to generate carbides and versatility of the sol–gel route in the synthesis of a family of
4
5
6
19
3
7
nitrides with small particle size or porosity. A major hurdle that nanocomposites. This new general route emphasizes the potential
remains is the controlled and tunable synthesis of composites, of sol–gel methods for synthesis of multifunctional materials.
8
particularly carbides or nitrides combined with metal oxides.
Nanocomposites of TiO
C and MgO–WN were synthesized by mixing
aqueous (or ethanolic) metal salts with hot gelatin solution,
2 3 2 2 3 2
–Fe C, TiO –WN, CeO –Fe C, CeO –
Such architectures are critical for applications such as noble WN, MgO–Fe
3
9
metal-replacement in solar water splitting.
Combining two different ceramics in a nanocomposite pre- stirring vigorously and then drying at 80 1C. The resulting
sents a significant synthetic challenge. Historically, it has been sponge-like precursors were calcined under nitrogen. Samples
achieved simply by mechanical mixing, which can seriously are denoted TF, TW, CF, CW, MF and MW. A range of molar
compromise the properties of one or both components. More ratios (100 : 0, 75 : 25, 50 : 50 and 0 : 100) were prepared for each
1
0,11
effective methods use ammonolysis
or temperature pro- sample and labelled 0, 25, 50 and 100 respectively. For example,
1
2
grammed reduction of oxide–oxide nanocomposites but have sample TF25 was prepared with 75 mol% of Ti and 25 mol% Fe.
the disadvantage of using hazardous synthesis gases, or multi-step Full sample details are listed in Table S5 (ESI†). The extent of
processes. Pinpointing more efficient and sustainable synthesis foaming depended on the nature of the metal precursors and a
routes is critical to future materials design. On this basis, a higher amount of nitrate led to increased foaming, suggesting
generalised one-pot sol–gel method to composites of two or more that foaming resulted from nitrate-oxidation of the polymer.
ceramic materials would have the key advantages of: (i) a single Scanning electron microscopy (SEM) images show that the
heating step, (ii) fine-tuning of particle size while achieving good open, sponge-like structure is maintained even after calcina-
13
1
9
interspersion and (iii) standard, inexpensive and non-hazardous tion, as observed previously (Fig. S1, ESI†).
synthesis atmospheres such as nitrogen.
Synchrotron X-ray diffraction (SXRD) patterns (Fig. 1, Fig. S2–S7,
ESI†) reveal two distinct crystalline phases in each sample: an
oxide (MgO, TiO or CeO ) combined with either Fe C or WN.
2 2 3
a
The broad peaks suggest very small crystallites, which could
indicate that intermixing of two crystalline phases restricts
sintering compared to the sol–gel synthesis of a single phase.
International Center for Young Scientists, National Institute for Materials Science
(
NIMS), Tsukuba, Japan
b
c
School of Chemistry, University of Birmingham, UK. E-mail: z.schnepp@bham.ac.uk
University of Keele, Keele, UK
There is no evidence for ternary phases (e.g. FeTiO ) and
d
e
3
Synchrotron X-ray Station at SPring-8, NIMS, Hyogo, Japan
National Institute for Materials Science, Tsukuba, Japan
negligible peak shifting (Fig. S8 and S9, ESI†). This suggests a
high level of phase separation, and phase pure carbide or
nitride phases, which is in stark contrast with our first report
†
Electronic supplementary information (ESI) available: Full experimental details
and extra figures. See DOI: 10.1039/c3cc47480a
5364 | Chem. Commun., 2014, 50, 5364--5366
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
very broad SXRD peaks for WN and the oxides, compared to
the sharper peaks for Fe C. The Fe C particles are also char-
3 3
acterized by ‘onion-like’ layers of graphite (Fig. S12, ESI†). This
is a common observation in the synthesis of Fe C due to
3
2
0
catalytic graphitization. Given that Fe C-catalyzed graphitiza-
3
3
tion is known to occur alongside (and perhaps as a result of) Fe C
2
1
melting, the agglomeration of molten Fe
explain the larger particle size. It should be noted that the Fe
particle size in these systems is still much smaller than in the
3
C droplets could
3
C
16
3
sol–gel synthesis of single-phase Fe C from gelatin. The flexible
nature of this method is demonstrated in the synthesis of TiO2/
Fe N (Fig. S13, ESI†). Forming the nitride phase rather than the
3
carbide Fe C was achieved simply by reducing the organic/metal
3
ratio and lowering the heating temperature to 700 1C. The control
of nitride vs. carbide formation has been observed in single-
2
2
phase systems and is thought to be due to stabilization of a
1
5
metastable nitride phase.
A study of the formation of TiO
2
3
–Fe C has offered some
surprising new insights into the mechanism. Many combina-
tions of elements form stable ternary oxides, some of which are
naturally occurring minerals e.g. ilmenite (FeTiO ) or magne-
3
2 4
sioferrite (MgFe O ). It could therefore be expected that heating
a homogeneous mixture of e.g. Fe:Ti would initially form
2 3
Fig. 1 Synchrotron X-ray diffraction (SXRD) patterns for (a) TiO /Fe C,
2 2 3 2 3
(b) TiO /WN, (c) CeO /Fe C, (d) CeO /WN, (e) MgO/Fe C and (f) MgO/WN.
a ternary Fe–Ti–O phase. Indeed, there are several reports of
The additional sharp peaks in CW50 correspond to minor
WC phases.
W and
2
3,24
sol–gel synthesis of Fe–Ti–O ternary compounds.
However,
SXRD samples quenched at 100 1C intervals (Fig. S14, ESI†) did
not show any ternary phases. Given the broad peaks, it is
of sol–gel synthesis of oxide–nitride mixtures using Ti and W difficult to identify any crystalline phases other than anatase
1
8
25
(Fig. S10, ESI†).
and a small amount of Fe N (a metastable precursor to Fe C).
3 3
Transmission electron microscopy (TEM) images show nano- However, when comparing SXRD patterns for samples TF50
particles interspersed in a carbon matrix (Fig. 2a and Fig. S11, and TF25 at 600 1C (Fig. 3a), very broad peaks of magnetite can
ESI†). High-res TEM images of all samples show individual be observed for TF50. Furthermore, HRTEM images (Fig. 3b)
single crystals that can be assigned to oxide, nitride or carbide clearly show crystallites of magnetite, and STEM (inset) shows
phases (Fig. 2b and Fig. S11, ESI†). In all cases, the phases are distinct Fe and Ti regions. Magnetite has been identified as an
1
5
highly interspersed, supported by elemental mapping and intermediate in the sol–gel synthesis of pure Fe C, but it is
3
energy dispersive X-ray analysis (EDXA) of a sample of TiO2/ surprising that it exists in this mixed system.
Fe
oxide or WN particles (B2–5 nm). This is consistent with the in catalytic methanol decomposition (CH
Methanol is of interest as a ‘hydrogen carrier’: storing H
3
C (Fig. 2c–e). The Fe
3
C particles are larger (B20 nm) than the
As a proof of concept, the MgO/Fe C/C sponges were tested
3
3
OH - 2H
2
+ CO).
in a
C system was
chosen because of the previous high activity of this system in
2
2
6
safer and more manageable form. The MgO–Fe
3
1
9
the oxygen reduction reaction (ORR). Also, iron carbides are
Fig. 3 (a) PXRD patterns for samples with Ti : Fe molar ratios of 50 : 50 and
75 : 25 (denoted TF50 and TF25 respectively), quenched at 600 1C during
heating under N (NB: dashed line added to highlight broad magnetite
2
Fig. 2 (a) TEM (b) high resolution TEM, (c) EDXA line scan, (d) dark field
peaks). (b) High-res TEM and (inset) STEM-EDXA line analysis of sample
STEM and (e) corresponding elemental map for sample TiO
2
/Fe
3
C.
TF50 quenched at 600 1C.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 5364--5366 | 5365

View Article Online
ChemComm
Communication
and simultaneous evolution of two distinct crystalline phases
minimizes the particle size (o20 nm). Furthermore the two
phases are highly interspersed, which offers high active surface
areas, as demonstrated by enhanced catalytic activity per gram.
This simple and robust and flexible method therefore offers
considerable advantages over previous multi-step methods for
generating oxide–carbide and oxide–nitride composites.
The authors thank NIMS-ICYS, JSPS, SPring-8 (2012A4511)
Fig. 4 (a) Percentage methanol conversion with increasing temperature and University of Birmingham.
for samples prepared using Mg : Fe molar ratios of 75 : 25, 50 : 50 and
0
: 100, denoted MF25, MF50 and MF100 respectively (inset BET surface
areas). (b) H production rate per gram of iron for MF25 and MF50
calculated from elemental analysis).
2
Notes and references
(
1
2
3
4
5
Z. Wen, S. Ci, F. Zhang, X. Feng, S. Cui, S. Mao, S. Luo, Z. He and
J. Chen, Adv. Mater., 2012, 24, 1399–1404.
A.-M. Alexander and J. S. J. Hargreaves, Chem. Soc. Rev., 2010, 39,
promising materials for many other catalytic processes such as
4388–4401.
2
7
28
Fischer–Tropsch and the ORR. Fig. 4 shows the temperature
dependent activity for samples MF50 and MF25 (50 and 25 mol%
Fe respectively) compared to a control (100 mol% Fe) sample
labelled MF100. These data show an onset of decomposition of
B300 1C followed by a sharp rise up to 520 1C. After this point,
the activity drops due to carbon build-up poisoning the catalyst.
This deposition of carbon onto iron-containing catalysts (and
some other transition metal compounds) is common in catalytic
A. T. Garcia-Esparza, D. Cha, Y. Ou, J. Kubota, K. Domen and
K. Takanabe, ChemSusChem, 2013, 6, 168–181.
L. Borchardt, C. Hoffmann, M. Oschatz, L. Mammitzsch, U. Petasch,
M. Hermann and S. Kaskel, Chem. Soc. Rev., 2012, 41, 5053–5067.
Y. Kido, K. Nakanishi, A. Miyasaka and K. Kanamori, Chem. Mater.,
2012, 24, 2071–2077.
B. Mazumder and A. L. Hector, J. Mater. Chem., 2008, 18, 1392–1398.
Y. Shi, Y. Wan and D. Zhao, Chem. Soc. Rev., 2011, 40, 3854–3878.
M. Feyen, C. Weidenthaler, R. G u¨ ttel, K. Schlichte, U. Holle, A.-H. Lu
and F. Sch u¨ th, Chem.–Eur. J., 2011, 17, 598–605.
6
7
8
2
9
9 Y. Kim, H. Irie and K. Hashimoto, Appl. Phys. Lett., 2008, 92, 182107.
0 N. Liu, L. Nie, N. Xue, H. Dong, L. Peng, X. Guo and W. Ding,
ChemCatChem, 2010, 2, 167–174.
reactions involving carbon-rich gases. Significantly, before the
catalysts are deactivated, the activity of all three is very similar.
1
This is surprising given that the samples contain very different 11 P. Krawiec and S. Kaskel, Top. Catal., 2009, 52, 1549–1558.
1
1
2 D.-V. N. Vo and A. A. Adesina, Appl. Catal., A, 2011, 399, 221–232.
3 A Sustainable Global Society (White paper), Royal Society of Chem-
istry, 2011.
amounts of Fe. The total surface area (Fig. 4a) is similar for all
samples, implying that active surface area is the key factor. This
2
can be clearly seen in a plot of H production rate per gram of 14 Z. Schnepp, S. R. Hall, M. J. Hollamby and S. Mann, Green Chem.,
2
011, 13, 272–275.
iron (calculated from elemental analysis). In this case, the
activity of sample MF25 is much higher than that of MF50,
1
5 Z. Schnepp, M. Thomas, S. Glatzel, K. Schlichte, R. Palkovits and
C. Giordano, J. Mater. Chem., 2011, 21, 17760–17764.
indicating that the active surface area in MF25 is higher. This is 16 Z. Schnepp, S. C. Wimbush, C. Giordano and M. Antonietti, Chem.
Mater., 2010, 22, 5340–5344.
3
consistent with the much smaller Fe C particle size in MF25, as
observed by TEM and SAXS, which is a result of the higher
1
1
7 Z. Schnepp, Angew. Chem., Int. Ed., 2013, 52, 1096–1108.
8 Z. Schnepp, M. J. Hollamby, M. Tanaka, Y. Matsushita, Y. Katsuya
and Y. Sakka, Sci. Tech. Adv. Mater., 2012, 13, 035001.
1
9
amount of MgO reducing sintering of Fe C. Selectivity is
3
1
2
2
2
9 Z. Schnepp, Y. Zhang, M. J. Hollamby, B. R. Pauw, M. Tanaka,
Y. Matsushita and Y. Sakka, J. Mater. Chem. A, 2013, 1, 13576–13581.
0 Y. Homma, Y. Kobayashi, T. Ogino, D. Takagi, R. Ito, Y. J. Jung and
P. M. Ajayan, J. Phys. Chem. A, 2003, 107, 12161–12164.
high (Fig. S15–S17, ESI†) and the H /CO ratio stabilizes at 2
2
(
Fig. S18, ESI†). The high initial H
reactions such as 2CO(g) " C(s) + CO2(g). Carbon balances
Tables S1–S3, ESI†) give very small variation (o5%).
The activity of this Fe C/MgO catalyst for methanol decom-
position is lower than for optimized catalysts.
2
/CO ratio could indicate side
1 S. Glatzel, Z. Schnepp and C. Giordano, Angew. Chem., Int. Ed., 2013,
(
52, 2355–2358.
3
2 C. Giordano, C. Erpen, W. Yao and M. Antonietti, Nano Lett., 2008, 8,
4659–4663.
3 X. Tang and K. Hu, J. Mater. Sci., 2006, 41, 8025–8028.
4 Y. G. Sharma, M. Kharkwal, S. Uma and R. Nagarajan, Polyhedron,
3
0,31
However, this
2
2
is a compelling demonstration of improved mass-activity through
optimizing particle size and dispersion. The data offer a good
2009, 28, 579–585.
starting point for investigation of oxide/carbide and oxide/nitride 25 K. H. Jack, Proc. R. Soc. London, Ser. A, 1948, 195, 34–40.
2
2
2
2
6 S. S ´a , H. Silva, L. Brand ˜a o, J. M. Sousa and A. Mendes, Appl. Catal.,
B, 2010, 99, 43–57.
7 E. De Smit and B. M. Weckhuysen, Chem. Soc. Rev., 2008, 37,
2758–2781.
8 Z. Wen, S. Ci, F. Zhang, X. Feng, S. Cui, S. Mao, S. Luo, Z. He and
J. Chen, Adv. Mater., 2012, 24, 1399–1404.
9 C. H. Bartholomew, Catal. Rev. Sci. Eng., 1982, 24, 67–112.
materials as catalysts. This field has recently attracted a lot of
32
attention and the simple, sol–gel preparation of carbide or
nitride composites with metal oxides offers a cost-effective and
sustainable alternative to existing multi-step routes. Furthermore,
there are many simple ways to optimize the sol–gel synthesis
(
e.g. biopolymer type, or drying rate).
In summary, we have demonstrated a flexible and general
30 G. Marb ´a n, A. L o´ pez, I. L o´ pez and T. Vald ´e s-Sol ´ı s, Appl. Catal., B,
010, 99, 257–264.
2
3
1 J. H. Jang, Y. Xu, M. Demura, D. M. Wee and T. Hirano, Appl. Catal.,
A, 2011, 398, 161–167.
method for producing composites of transition metal nitrides or
carbides with metal oxides. The homogeneous sol–gel precursor 32 J. S. J. Hargreaves, Coord. Chem. Rev., 2013, 257, 2015–2031.
5366 | Chem. Commun., 2014, 50, 5364--5366
This journal is ©The Royal Society of Chemistry 2014