Reuseable monolithic nanoporous graphite-supported nanocatalysts (Fe, Au, Pt, Pd, Ni, and Rh) from pyrolysis and galvanic transmetalation of ferrocene-based polyamide aerogels

Article abstract of DOI:10.1021/acs.chemmater.6b02364

Polyamide aerogels with ferrocene as a monomer repeat unit were prepared in one step from ferrocene dicarboxylic acid and tris(4-isocyanatophenyl)methane. Pyrolysis at ≥800 °C yielded nanoporous carbons doped throughout with crystallites of α-Fe (about 50 nm in diameter), which in turn were shrouded in graphitic ribbons (<30 graphene layers thick). Transmetalation was carried out with aqueous solutions of Au, Pt, Pd, Rh, and Ni salts, via a path akin to galvanic corrosion, whereas graphitic ribbons separated anodes (α-Fe particles) from cathodes (defects along the ribbons). The new metallic phases formed clusters of smaller crystallites (10-20 nm in diameter) on the graphitic ribbons, leaving behind empty cage-like formations previously occupied by the Fe(0) nanoparticles. All metal-doped carbons were monolithic and over 85% porous. Catalytic activity was demonstrated with the oxidation of benzyl alcohol to benzaldehyde catalyzed with carbon-supported Au or Pt, the reduction of nitrobenzene by hydrazine to aniline catalyzed with carbon-supported Fe, and two Heck coupling reactions of iodobenzene with styrene or butyl acrylate, catalyzed with carbon-supported Pd. The distinguishing feature of those catalysts was that they could be just picked up, for example, with a pair of tweezers, and redeployed in a new reaction mixture immediately, thus bypassing less efficient recovery processes like filtration.

Full text of DOI:10.1021/acs.chemmater.6b02364

Article
pubs.acs.org/cm
Reuseable Monolithic Nanoporous Graphite-Supported
Nanocatalysts (Fe, Au, Pt, Pd, Ni, and Rh) from Pyrolysis and Galvanic
Transmetalation of Ferrocene-Based Polyamide Aerogels
Adnan M. Saeed, Clarissa A. Wisner, Suraj Donthula, Hojat Majedi Far, Chariklia Sotiriou-Leventis,*
and Nicholas Leventis*
Department of Chemistry, Missouri University of Science and Technology, Rolla, Missouri 65409, United States
S
* Supporting Information
ABSTRACT: Polyamide aerogels with ferrocene as a monomer repeat unit were
prepared in one step from ferrocene dicarboxylic acid and tris(4-isocyanatophenyl)-
methane. Pyrolysis at ≥800 °C yielded nanoporous carbons doped throughout with
crystallites of α-Fe (about 50 nm in diameter), which in turn were shrouded in graphitic
ribbons (<30 graphene layers thick). Transmetalation was carried out with aqueous
solutions of Au, Pt, Pd, Rh, and Ni salts, via a path akin to galvanic corrosion, whereas
graphitic ribbons separated anodes (α-Fe particles) from cathodes (defects along the
ribbons). The new metallic phases formed clusters of smaller crystallites (10−20 nm in
diameter) on the graphitic ribbons, leaving behind empty cage-like formations previously
occupied by the Fe(0) nanoparticles. All metal-doped carbons were monolithic and over
85% porous. Catalytic activity was demonstrated with the oxidation of benzyl alcohol to
benzaldehyde catalyzed with carbon-supported Au or Pt, the reduction of nitrobenzene by
hydrazine to aniline catalyzed with carbon-supported Fe, and two Heck coupling
reactions of iodobenzene with styrene or butyl acrylate, catalyzed with carbon-supported
Pd. The distinguishing feature of those catalysts was that they could be just picked up, for example, with a pair of tweezers, and
redeployed in a new reaction mixture immediately, thus bypassing less efficient recovery processes like filtration.
aerogels,12−15 and at that point our attention shifted toward a
1. INTRODUCTION
Heterogeneous catalysis is a surface phenomenon,1,2 thereby
better dispersion of the metallic nanoparticles in the carbon
matrix. For this, a sensible approach is to incorporate a
heterogeneous catalysts consist of high surface-to-volume
precursor of metallic nanoparticles into every monomer repeat
unit of a carbonizable polymer. In that regard, we chose
ferrocene, a known precursor of iron nanoparticles.16
At the implementation level, that system design was carried
out with a hyperbranched polyamide aerogel prepared from
ferrocene dicarboxylic acid (Fc(COOH)2) and tris(4-
isocyanatophenyl)methane (TIPM) as shown in Scheme 1.
TIPM is an inexpensive monomer, and it was chosen because it
is also a known precursor of several carbonizable polymeric
aerogels,17,18 including polyamides.19
Indeed, pyrolysis of ferrocene-based polyamide aerogels gave
monolithic carbon aerogels bearing Fe(0) nanoparticles
dispersed throughout their volume. Apart from their own
catalytic activity, those materials were transmetalated with
selected metal ions, replacing Fe(0) nanoparticles with Au, Pt,
Pd, Ni, and Rh. All materials were characterized along all
processing steps in terms of their chemical composition and
their micro-/nanomorphology. Catalytic activity was demon-
catalytic particles on inert supports. The latter include mostly
oxides, carbides, nitrides, and activated carbon.3,4 Supports,
typically in particulate/coarse-powder form, provide high
surface area for the smaller catalytic particles to latch on and
pack in a way that ensures low mass transfer resistance to and
from the catalytic sites. Recovery of the catalyst at the end of a
catalytic cycle is typically carried out by filtration,5 which is
time-consuming and may lead to catalyst loss. At the other
extreme, industrial, continuously recyclable heterogeneous
catalysts, like those employed in fuel cells6 or in catalytic
converters,7,8 operate with gas-phase reagents and are three-
level structures, in which the two lower levels incorporate the
general principles set forth above, while, out of engineering
necessity, the top level adds their key macroscopic feature:
monolithicity. Transferring that feature to any other system is
conceptually and practically equivalent to having monolithic
catalysts that could be picked up and redeployed immediately
in any suitable situation. Clearly, such catalysts will have to be
based on porous supports, and in that regard oxide or ceramic
aerogels9,10 doped with metallic nanoparticles11 could comprise
a reasonable option. However, in general those types of
aerogels are delicate materials that would not survive harsh
reaction environments. Thus, we opted for sturdier carbon
Received: June 12, 2016
Revised: June 17, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Scheme 1. Reaction of a Triisocyanate (TIPM) with Ferrocene Dicarboxylic Acid Toward a Polyamide (atom labeling is used
for NMR peak assignment)
strated with selected reactions catalyzed by the corresponding
metal nanoparticles.
2. RESULTS AND DISCUSSION
This section moves from the synthesis and characterization of
ferrocene-based polyamide aerogels (Section 2.1) to their
pyrolytic conversion to Fe(0)-doped carbon (Section 2.2), to
transmetalation (Section 2.3), and finally to a demonstration of
catalytic activity as set forth above (Section 2.4). Emphasis is
placed on the metal content, particle size, and pore structure.
2.1. Synthesis and Characterization of Ferrocene-
Figure 1. Photograph of samples after different stages of processing as
shown. For relevant data refer to Table 1.
Based Polyamide Aerogels (FcPA-xx). The reaction of
Scheme 1 was implemented according to Scheme 2. Fc-
Scheme 2. Implementation of the Reaction of Scheme 1 in
the Synthesis of Ferrocene-Based Polyamide Aerogels
In brief, FcPA-xx shrunk 35−41% in linear dimensions
relative to the molds and in reverse order with xx. Skeletal
densities (ρs) did not vary in any systematic way that could
imply closed porosity. Bulk densities (ρb) and porosities (Π =
100 × (ρs − ρb)/ρs) varied from 0.12 g cm−3 and 92% (v/v)
(xx = 05) to 0.49 g cm−3 and 63% (v/v) (xx = 25). The porous
structure was probed with N2-sorption porosimetry (Appendix
III in Supporting Information). Isotherms started off (at xx =
05) with very narrow hysteresis loops, becoming wider at
higher densities with well-defined saturation plateaus at xx ≥
15. Accordingly, lower-density samples were dominated by
larger macropores, shifting to smaller macropores and
mesopores as density increased (volume ratio, V>300_nm
/
V1.7−300_nm = 4.2 for FcPA-05, versus 0.8 for FcPA-25see
Table S.2). BET surface areas, σ, were in the 460−260 m2 g−1
range, in descending order with xx. Microscopically, the skeletal
framework of all FcPA-xx consisted of random assemblies of
nanoparticles (Figure 2A). Particle radii calculated from N2-
sorption data via r = 3/(ρs × σ) were in the 4.5−8.7 nm range
and agreed well with primary particle radii calculated from
small-angle X-ray scattering data (SAXS: 2.9−6.5 nmsee
Appendix IV in Supporting Information). SAXS further showed
that primary particles aggregated into surface fractal secondary
particles with diameters in the range of 35−60 nm (white
circles in Figure 2A).
(COOH)2 was prepared following literature procedures20−22
and was codissolved at room temperature in a 1.5:1 mol/mol
ratio with TIPM in variable amounts of DMF/ethyl acetate (see
Experimental Section). All formulations are provided in Table
S.1 of Appendix I in Supporting Information. Gelation was
carried out at 90 °C, and gelation times (included in Table S.1)
varied from 95 min to approximately 3 h, depending on the
monomer concentration. Wet-gels were aged in their molds at
90 °C for 24 h, solvent-exchanged with acetone, dried with
liquid CO2, and vented off at the end as a supercritical fluid
(SCF). The resulting aerogels were sturdy monoliths and are
referred to as FcPA-xx, where xx stands for the total weight
percent concentration of the monomers in the sol, and was
varied in the range 05 ≤ xx ≤ 25. Materials characterization
data for all xx are summarized in Table S.2 of Appendix II in
Supporting Information. A typical monolith is shown in Figure
1, along with materials from further processing.
Solid-state CPMAS 13C NMR spectra (Figure 3) showed the
aliphatic carbon from TIPM at 55 ppm, the cyclopentadienyl
carbons from Fc(COOH)2 as a broad resonance centered at 71
ppm, and the aromatic carbons of TIPM in the 110−140 ppm
range. The peak at 167 ppm was assigned to the amide CO,
and the peak at 154 ppm was assigned to the CO resonance
of TIPM-derived polyurea. The latter has been prepared
independently by reacting TIPM either with water18 or with
boric acid23 and its spectrum is included in Figure 3 as
reference.
B
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 3. Solid-state CPMAS 13C NMR of FcPA-xx (top) and of
TIPM-derived polyurea (second from top). Liquid state 13C NMR of
Fc(COOH)2 in DMSO-d6 (third from the top) and of TIPM in
CDCl3 (bottom two spectrafor peak assignment refer to Scheme 1)
The APT (Attached Proton Test) spectrum of TIPM confirms that C-
4 and C-6 overlap. (S: solvent peak).
(w/w)) carbonizable polymer in its own right18 and thereby is
expected to contribute toward the carbon matrix.
Integration of the two CO resonances in the solid-state
13C NMR spectra of FcPA-xx gives a ratio of ferrocene-based
polyamide and TIPM-derived polyurea equal to 1:1 mol/mol.
Based on (a) that mol ratio, (b) application twice (once for the
polyamide and once for the polyurea) of eq 1 that gives the
Figure 2. Scanning electron microscopy (SEM) of representative
samples along processing. White circles in FcPA-15 denote secondary
particles as identified via SAXS (see Appendix IV in Supporting
Information).
Gn = T − 3(1 − 2n)(D + T) + 3 × 2n × (end‐cap)
The reaction of a carboxylic acid with an isocyanate toward
an amide starts with a condensation step to a mixed carbamic−
carboxylic anhydride (−NH−CO−O−CO−), which either
rearranges intramolecularly to the amide (+CO2) or reacts
intermolecularly with another molecule of carbamic−carboxylic
anhydride toward urea and anhydride;24,25 the latter two
products may also react with one another toward the same
amide obtained via the intramolecular route.26 As confirmed
from FTIR (Figure S.6 of Appendix V in Supporting
Information), the FcPA-xx does not include an anhydride.
On the other hand, based on 13C NMR data obtained during
gelation, the solution in the pores of freshly made wet-gels
contains a soluble product, which is also observed by adding
P2O5 (as a dehydrating agent) in a solution of Fc(COOH)2 in
DMSO-d6. The 13C NMR spectrum of that product was
consistent with the strained intramolecular anhydride
(FcCO)2O, which, however, was unstable and could not be
isolated easily for further characterization. Thus, it is speculated
that owing to the favorable topology of the second −COOH
group of Fc(COOH)2 relative to the primary condensation
product (the carbamic−carboxylic anhydride), the two groups
(−COOH and −NH−CO-O−CO−) react intramolecularly
expelling Fc(CO)2O, a free amine, and CO2. In other words,
the isocyanate acts as a dehydrating agent of Fc(COOH)2. The
free amine reacts with yet unreacted −NCO groups toward
urea. At any rate, TIPM-derived polyurea is a high-yield (56%
(1)
molecular mass (Gn) of hyperbranched polymers from di- (D)
and tri- (T) functional building blocks (n ≥ 0: the generation
number of the hyperbranched polymer),27 (c) assuming one
bidentate Fc(CO−)2 end-cap for every two tips of the
hyperbranched polyamide (see also Section 2.2b below), and
(d) considering the limit of eq 1 for n → ∞, the expected iron
content in FcPA-xx was 8.7% (w/w). (It is noted in passing
that if FcPA-xx consisted exclusively of ferrocene-derived
polyamide, the expected iron content would have been 13.0%
(w/w).) The expected iron content in FcPA-xx agrees well with
the experimental value (9.5 1.1% (w/w)) that was calculated
from data obtained with thermogravimetric analysis (TGA) in
O2 (Figure 4). In that regard, the TGA residue of FcPA-xx at
1000 °C was identified by using powder X-ray diffraction
(PXRD) as Fe2O3, and it was found equal to 13.6 1.5 % (w/
w) (average of 8 samples from different batches). Then, the
weight percent of iron in FcPA-xx was calculated via eq 2:
Fe % (w/w) = [2 × MWFe/MWFe2O3] ×
(TGA Residue % (w/w) at 1000 °C)
2.2. Pyrolytic Conversion of FcPA-xx to Fe(0)-Doped
Carbon. 2.2a. Pilot Runs. Our first objective was to find out
(2)
C
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Figure 4. Representative thermogravimetric analysis (TGA) data
under O2 at 5 °C min−1 of samples as shown. (TGA was repeated
several times with samples from different batches, and results were
used to derive the metal content of various samples as shown in Table
1). The dashed lines and arrow point to the first step of mass loss by
FcPA-15 (about 16% (w/w) at around 310 °C). (Via eq 1, bidentate
Fc(CO−)2 end-caps contribute 14.3% (w/w) to the total mass of
FcPA-15.).
whether and under what conditions FcPA-xx could be
converted to carbon-supported Fe(0). For this, mid-density
FcPA-15 samples (ρb = 0.34 g cm−3) were pyrolyzed under
flowing H2 at various temperatures in the range of 400 to 1400
°C. Samples shrunk uniformly, e.g., 68% at 800 °C and up to
86% at 1400 °C, but remained monolithic. Relevant material
properties at various pyrolytic temperatures are given in Table
S.3 of Appendix II in Supporting Information. Screening for
metallic iron was carried out with powder X-ray diffraction
(PXRDFigure 5). Scheme 3 summarizes the results in terms
of chemical composition of the products and their behavior
toward transmetalation (refer to Section 2.3 below).
Figure 5. Powder X-ray diffraction (PXRD) data after pyrolysis of
FcPA-15 at different temperatures. The (002) reflection of graphite at
2θ = 26.44° first shows up after pyrolysis at 800 °C and dominates the
diffraction pattern above 1000 °C. (Line spectra are included at the
bottom for identification.)
Scheme 3. Properties at a Glance: Chemical Composition,
Electrical Conductivity (+), and Ability for Transmetalation
via Reaction with Metal Ions, [M]n+, of the Product from
Pyrolysis of FcPA-15 at Different Temperatures
It is noted that our initial attempts focused on producing
Fe(0)-doped carbons carbothermally,28−31 and accordingly
pyrolysis was carried out under flowing Ar. However, the
only crystalline phase that was observed at ≤1100 °C was Fe3C;
some α-Fe was noted only at 1200 °C. Under H2, on the other
hand, the main crystalline phase was always α-Fe. In fact, at
400/500 °C α-Fe was the only crystalline phase obtained. Some
cementite (Fe3C) started showing up at ≥600 °C, along with
some martensite (Fe1.91C0.09) at 1400 °C (Figure 5). By putting
those pieces of information together, it is reasonable to
speculate that under H2, carbides were produced in a parallel
carbothermal process. As summarized in Scheme 3, at ≥800
°C/H2 samples were electrically conducting, Fe(0) became
chemically accessible (e.g., they would react with ions in
solution filling the poressee Figure S.7 of Appendix VI in
Supporting Information), and PXRD included the (002)
reflection of graphite at 2θ = 26.44° (Figure 5), consistent
with iron being a low-temperature graphitization catalyst.32−34
At ≥1000 °C the (002) reflection of graphite dominated the
PXRD spectra. Samples at the onset of those properties,
namely, those obtained at 800 °C, comprised the basis for
further study and are referred to as Fe@C.
for a 68% of total linear shrinkage relative to the molds.
Evidently, however, mass loss compensated for additional
shrinkage, and Fe@C monoliths were less dense (0.286
0.004 g cm−3 versus 0.340 0.004 g cm−3) and more porous
(88% (v/v) versus 74% (v/v)) than FcPA-15 (Table 1.)
Microscopically, the framework of Fe@C consisted of finer,
fused-together-like particles (Figure 2B). N2-sorption porosim-
etry (Appendix III in Supporting Information) showed that the
balance of pores shifted to larger macropores relative to FcPA-
15 (V>300_nm/V1.7−300_nm = 3.4, vs 1.2 in FcPA-15). In short,
despite that Fe@C were significantly smaller in size than FcPA-
15, they had a higher percentage of internal empty space, and
their pores were larger than those of the latter. The BET
surface area of Fe@C (about 370 m2 g−1) was also somewhat
higher than that of FcPA-15 (about 310 m2 g−1), but the
distinguishing feature here was that a significant portion of the
total BET surface area (175 m2 g−1, 48%) was assigned (via t-
plot analysis, Harkins and Jura Model35) to newly formed
2.2b. Characterization of Fe@C and Chemical Accessi-
bility of Fe(0). The pyrolytic (char) yield of FcPA-15 to Fe@C
was 37.3 1.9 % (w/w). Material properties of Fe@C and its
derivatives from further processing (transmetalation) are
compared with those of FcPA-15 in Table 1. As shown in
Figure 1, Fe@C shrunk (49%) relative to their parent FcPA-15,
D
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
(open) micropores. The apparent particle radius in Fe@C
(based on gas sorption data) was smaller (3.4 nm) than that of
FcPA-xx (7.3 nm). However, owing to the lack of well-defined
nanoparticles (refer to the network morphology in Figure 2B),
this value for the particle radius in Fe@C should be taken only
as an estimator of the overall feature size we are dealing with.
The iron content of Fe@C (11.5 2.1% (w/w), via TGA in
O2Figure 4) was 40−50% (w/w) of the expected value (25.5
3.2%) had all iron in FePA-15 (9.5 1.1 % (w/w)) had been
retained. That loss of iron is consistent with an early loss of the
Fc(CO−)2 end-caps of the ferrocene-based polyamide during
pyrolysis: indeed, based on eq 1, those end-caps contribute
14.3% (w/w) to the mass of FcPA-xx, which in turn is
consistent with the mass loss of about 16% at 310 °C noted in
TGA (Figure 4).
The skeletal density of Fe@C (ρs = 2.4 g cm−3) was the exact
weighted average of iron (7.86 g cm−3 at 11.5% (w/w)the
TGA-derived value from above) and the experimental skeletal
density of the matrix (1.7 g cm−3) after removing Fe with HCl.
(The matrix comprised 88.5% (w/w) of the total mass of Fe@
C.) Despite graphitization, the skeletal density of Fe@C was
lower even than the density of sp2-rich amorphous carbon
(1.8−2.0 g cm−3),36 suggesting that the skeletal framework
included closed pores; this is very likely considering that, along
the newly formed open micropores, some may have been
actually closed. (Assuming that all closed pores were confined
within the matrix (carbon), the volumetric percent of closed
pores on the skeletal backbone varied in the 5−15% (v/v)
rangedepending on the density of the skeletal carbonin
the 1.8−2.0 g cm−3 range.)
A closer topological view of Fe(0) within Fe@C, which turns
out to be important for rationalizing the transmetalation
process (Scheme 3 and Section 2.3), was obtained with TEM
(Figure 6 and S.10 of Appendix VII in Supporting
Information). Fe(0) particles were embedded evenly through-
out the surrounding matrix (Figure 6A). Their size distribution
was broad with a maximum at 52 nm and fwhm = 31 nm. The
average crystallite size (via the Scherrer equation applied to the
(110) peak of Fe(0)see Figure 5 above) was also equal to 52
1 nm. Under higher magnification (Figure 6B), Fe(0)
particles were shrouded in layered ribbons fading away into the
surrounding matrix, presumably amorphous carbon. As ribbons
faded away, they swirled around occasionally forming pockets
similar to those encapsulating Fe(0) (see Figure S.10);
however, the key observation was that Fe(0) particles were
always encased in layered ribbons. The ribbon around the
particle of Figure 6B is 9.48 nm thick and consists of 28
graphene layers, giving an interlayer spacing of 3.39 Å, which is
close to that of graphite (3.35 Å).37 That graphene layer stack
height is consistent with the value (La = 16.9 nm) calculated
from Raman data using Knight’s empirical formula38 and the
ratio of the areas under the characteristic D and G bands of
disordered graphite (at 1326 and 1592 cm−1, respectively; ID/IG
= 2.27see Figure S.11 of Appendix VIII in Supporting
Information). The agreement between the graphene layer stuck
height via TEM and Raman renders graphitic ribbons and cages
like those in Figure 6B a general structural property of the
entire sample.
Graphitic ribbons are a rather common occurrence in low-
temperature Fe(0)-catalyzed graphitization.39−41 In many cases,
shrouding Fe(0) particles in such ribbons (as in Figure 6B)
isolates the metal from its environment.42,43 In Fe@C,
however, Fe(0) particles were chemically accessible and were
E
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
(the V>300_nm/V1.7−300_nm ratio moved from 3.4 in Fe@C up to
6.7 in tm-Pd@C, 12.4 in tm-Au@C, and 15.6 in tm-Pt@C).
The BET surface areas were reduced from about 370 m2 g−1 in
Fe@C (and 282 m2 g−1 in HCl-treated Fe@C) to 100−130 m2
g−1 in tm-M@C, with an even greater reduction in the surface
area assigned to micropores (from about 170 m2 g−1 in Fe@C
and about 110 m2 g−1 in HCl-treated Fe@C, to 5−15 m2 g−1 in
tm-M@C). According to the pore volume data, surface-tension-
driven nanoscopic contraction cannot be ruled out as a
mechanism for the observed changes in the pore structure. On
the other hand, however, surface area data suggest that
transmetalation, as opposed to removing Fe(0) with HCl,
blocked access to micropores. That would increase closed
porosity and would decrease the apparent skeletal densities.
Indeed, the skeletal densities of all three tm-samples were found
lower than that of Fe@C, despite quantitative replacement of
Fe(0) with heavier metals; as summarized in Table 1, the
skeletal densities of the three tm-M@C were in the 1.88−1.98 g
cm−3 range, versus ρs = 2.40 g cm−3 for Fe@C.
The PXRD spectra of tm-M@C were dominated by the new
metallic phases (Figure 7). No residual Fe(0) was detected in
Figure 6. Transmission electron microscopy (TEM). A, B: Fe@C at
two different magnifications. Inset in A: Patricle size distribution. C:
Fe@C after HCl-treatment. D: tm-Pt@C. Arrows point at empty
graphitic cages after removing Fe(0). (Additional TEM data are shown
in Appendix VII in Supporting Information.)
removed quantitatively with concentrated aqueous HCl (see
Experimental Section), as confirmed by Figure S.8 of Appendix
VI that compares PXRD spectra of Fe@C before and after HCl
treatment. In addition, general inspection with TEM of HCl-
treated samples (Figure S.10) shows an absence of Fe(0)
particles, while cage-like structures like those previously
shrouding Fe(0) particles abound (Figure 6C). Macroscopi-
cally, HCl treatment caused no further shrinkage, but BET
surface areas were somewhat compromised (about 280 vs 370
m2 g−1 before HCl-treatment); yet, 40% of that area (about 110
m2 g−1) was still assigned to open micropores. (Material
properties of HCl-treated samples are included in Table 1.)
2.3. Transmetalation of Fe@C to tm-M@C. The chemical
accessibility of Fe(0) in Fe@C, as magnifested by our ability to
dissolve Fe(0) away with aq. HCl, made it possible to replace
Fe(0) with Au, Pt, or Pd (M) via reaction with ions of the
corresponding metals, [M]n+,44,45 according to eq 3. In all three
nFe(0) + 2[M]n+ → nFe2+ + 2M(0)
(3)
Figure 7. Powder X-ray diffraction (PXRD) data of Fe@C samples
transmetalated with Au, Pt, and Pd, as shown. The PXRD spectrum of
Fe@C is included for comparison. All samples retain the small amount
of Fe3C present in the original Fe@C.
cases, the corresponding reactions are highly exothermic with
standard redox potentials of over 1.0 V. Experimentally, Fe@C
monoliths, right after they came out of the pyrolysis furnace,
were infiltrated quickly with aqueous solutions of [M]n+ using
capillary action under reduced pressure. Transmetalated
monoliths (referred to as tm-M@C) were washed with water
and acetone and were dried under ambient pressure without
noticeable changes in their size relative to their Fe@C
precursors (Figure 1). Microscopically, tm-M@C appeared
very similar to Fe@C, although voids seemed to be larger
(Figure 2C). Quantitative materials characterization data of the
three tm-M@C are included in Table 1. Thus, porosities (86−
90% (v/v)) remained in the same range as those of Fe@C
(88% (v/v)), but pore sizes shifted to even larger macropores
any of the transmetalated samples, and minor additional
reflections were assigned to Fe3C. TGA-in-O2 (Figure 4)
supported quantitative replacement of Fe(0) by M; that is,
starting with 11.5 2.1% (w/w) of Fe(0) in the parent Fe@C,
the expected/found values of M were (% (w/w)): Au, 31.4
4.4/33.6 8.8; Pt, 31.2 4.4/34.5 0.8; and Pd, 19.9 2.1/19.8
1.8 (Table 1). The average crystallite sizes (via the Scherer
equation applied to the (111) reflections) were Au, 15.2 0.5
nm; Pt, 20.9
0.7 nm; Pd, 21.4
0.7 nm), namely,
significantly smaller than those of Fe(0) in Fe@C (52 1 nm).
According to TEM (Figure 6D), the graphitic pockets
F
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
shrouding the parent iron particles in Fe@C were now empty
(like after treatment with HClcompare with Figure 6C), and
the new metallic particles clustered on the graphitic ribbons. In
no occasion (see also Figure S.10) new metallic nanoparticles
were formed inside cages previously occupied by Fe(0).
Macroscopically, bulk iron dipped in gold plating solutions is
passivated by a layer of Au(0).46 Microscopically, redox
transmetalation typically yields core−shell45 or hollow
particles.47 Here, complete consumption of Fe(0) nano-
particles, together with (a) the location of the new metallic
particles (on the ribbons rather than in the pockets) and (b)
their smaller size relative to the parent Fe(0) points to a
galvanic corrosion mechanism (Scheme 4),48 in which graphitic
2.4. Evaluation of Fe@C and tm-M@C as Catalysts.
Fe@C and tm-M@C monoliths similar to those shown in
Figure 1 were dipped under vigorous magnetic stirring in
reaction mixtures that could be catalyzed by the respective
metals.52−54 Reactions, run times, and results are summarized
in Table 2. Based on the weight of each monolith and its metal
Table 2. Use of Fe@C and tm-M@C (M: Au, Pt, Pd) in
Catalysis: Yields as a Function of Catalytic Cycle (yield %
a
mol/mol)
catalytic cycle
yield % mol/mol
b
catalyst
RXN (t)
1st
2nd
3rd
4th
5th
Scheme 4. Galvanic Transmetalation Mechanism of Fe@C
to tm-M@C (e.g., M = Au)
Fe@C
1 (24 h)
2 (24 h)
2 (24 h)
3 (24 h)
4 (2 h)
89
98
85
98
91
84
89
84
84
92
75
75
79
74
79
83
71
80
62
73
86
70
71
75
71
tm-Au@C
tm-Pt@C
tm-Pd@C
tm-Pd@C
a
In all cases (catalytic metal):(limiting reagent) = 5% mol/molsee
Experimental Section; for the raw data and data analysis see Appendix
IX in Supporting Information). RXN (t): Reaction (run time). 1.
b
nitrobenzene + hydrazine → aniline. 2. benzyl alcohol + O2
→
benzaldehyde. 3. iodobenzene + styrene → cis- + trans-stilbene. 4.
iodobenzene + butyl acrylate → butyl cinnamate
content (Table 1), the metal-to-limiting reagent ratio was
adjusted at 5% mol/mol. Aliquots were taken in regular
intervals and were analyzed using gas chromatography (GC
see Experimental Section; for the raw data see Appendix IX in
Supporting Information). Reactions were stopped by removing
the catalyst shortly after their limiting reagents had disappeared
from the GC traces (see Figures S.12−S.20; typically 24 h,
except Heck coupling of iodobenzene with butyl acrylate that
proceeded much fasterTable 2). In order to remove them,
catalytic monoliths were picked up with a pair of tweezers, were
rinsed briefly with the reaction solvent, and were transferred
immediately into new reaction mixtures to continue catalysis.
The procedure was repeated five times with each catalyst. All
monoliths were robust and survived the prolonged “beating” by
the magnetic stirrer. Yields remained ≥70% for all catalysts, all
cycles (Table 2). No product was observed in any of the
controls that included swapping tm-M@C with Fe@C or with
Fe@C monoliths after HCl-treatment. In the case of tm-Au@
C, catalysis was also attempted with a 1 in.2 gold foil left in the
reaction mixture for several daysno reaction was observed.
Optimization of catalytic activity, e.g., in terms of turn over
numbers (TON), was beyond the scope of this work.
Nevertheless, TON were found in the range of 70−100 and
compare favorably with C-supported catalysts in powder
form.53 Oxidation of benzyl alcohol with either tm-Au@C or
tm-Pt@C gave practically identical results, pointing to the
effectiveness of galvanic replacement to furnish catalytically
active nanosized Au.
ribbons separated physically and connected electrically anodes
and cathodes. The former were the Fe(0) nanoparticles; the
latter were end-points and defects on ribbons serving as
concentrators of the electric field, accelerating electrode
kinetics. Statistically, there are more such defects than Fe(0)
particles; therefore, we expect clusters of smaller Au, Pt, or Pd
crystallites than those of the parent Fe(0), as observed. This
galvanic mechanism also explains the curious formation of a
yellow gold layer on the outer geometric surface of all tm-Au@
C monoliths (see Figure 1): while reduction of ions infiltrated
quickly within the monolith still goes on, namely, while there is
still unreacted Fe(0), ions diffusing from the surrounding
solution toward the monolith get reduced at the point of their
first encounter with the conducting carbon network. In further
support of the galvanic reduction mechanism, initial trans-
metalation attempts of Fe@C with nickel were unsuccessful,
despite that the standard redox potential of the reaction Ni2+
+
Fe(0) → Ni(0) + Fe2+ is positive (but low: 0.183 V).
Meanwhile, Raman (Figure S.11) on one hand shows that at
higher processing temperatures the grapheme layer stack high
does not change much (Lα = 21 nm at 1400 °C, versus Lα = 17
nm in Fe@C), but PXRD on the other hand (Figure 5) shows
that higher-temperature samples contain more graphite.
Reasoning that the internal ohmic drop along the skeletal
framework in samples with higher graphite content would be
lower, FcPA-15 processed at 1200 °C were transmetalated
successfully with Ni2+ and also with another precious metal, Rh,
Finally, because of the way the catalytic experiments were
designed and executed, namely, (a) in all catalytic runs the
(metal inside the catalyst):(limiting regent) ratio was fixed at
5% mol/mol and (b) all reactions were monitored and stopped
shortly after the limiting reagent had been consumed, it was
deemed reasonable to average the yields of the five reactions
after each catalytic cycle and plot them against the latter
(Figure 8). According to Figure 8, all catalysts in the family of
catalysts of this report behave similarly, namely, independent of
in which case the crystallite size was found even smaller (9
1
nm) than that of all other metals in tm-M@C. (For PXRD data
of tm-Rh@C and tm-Ni@C see Figure S.9 of Appendix VI in
Supporting Information.) The ability of galvanic trans-
metalation to step particle size down is particularly relevant
to catalysis as, for example, only nanosized Au is catalytic.49−51
G
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Covestro LLC as a 27% (w/w) solution in dry ethyl acetate under the
trade name of Desmodur RE. Desmodur RE was also purchased
independently from M.F. Cachat (Lakewood, OH, www.mfcachat.
com). Deuterated DMSO (DMSO-d6) and CDCl3 were obtained from
Cambridge Isotope Laboratories Inc. Argon (99.99999%) and H2
(99.999%) were purchased from Ozark Gas (Rolla, MO).
1,1′-Ferrocene Dicarboxylic Acid (Fc(COOH)2). (Fc(COOH)2) was
prepared in two steps from ferrocene according to literature
procedures (Scheme 5).20−22 Yield: 63%; mp >250 °C. 1H NMR
Scheme 5. Synthesis of 1,1′-Ferrocene Dicarboxylic Acid
(Fc(COOH)2)
Figure 8. Cumulative graph of % Yield versus Catalytic cycle for all
catalysts used in this study (data from Table 2). The solid black line
connects the average % yields after each cycle. Error bars are one
standard deviation from the average % yields.
(400 MHz, DMSO-d6) δ 4.47 (d, 4H), 4.72 (d, 4H), 12.34 (s, 2H).
13C NMR (400 MHz, DMSO-d6) δ 171, 73, 72, 71. IR (KBr) 3429,
1687, 1495, 1301, 514 cm−1. Elemental Analysis, (CHN): Theoretical
% (w/w) for C12H10O4Fe: C, 52.55; H, 4.38. Found: C, 51.83; H, 4.13.
Synthesis of Ferrocene Polyamide Aerogels (FcPA-xx). In a typical
procedure, Fc(COOH)2 (4.11 g, 0.015 mol) was dissolved in
anhydrous DMF, and the solution was added to 13.6 g of Desmodur
RE (containing 3.67 g, 0.01 mol of TIPM). The resulting sol was
stirred at room temperature under N2 for 20 min, and was poured in
molds (Wheaton 4 mL Polypropylene Omni-Vials 1.04 cm in inner
diameter, Fisher part No. 225402), which were then sealed and left for
gelation at 90 °C. The total weight percent of monomers (TIPM +
Fc(COOH)2) in the sol was varied by varying the amount of solvent
(DMF) and is denoted by extension -xx in the sample names. All
formulations and gelation times are summarized in Table S.1 of
Appendix I in Supporting Information. Gels were aged for 24 h at 90
°C in their molds. Subsequently, wet gels were removed from their
molds, washed with DMF (3×, 8 h each time), acetone (4×, 8 h each
time, using 4× the volume of the gel for each wash) and were dried in
an autoclave with liquid CO2, which was taken out at the end as a
supercritical fluid (SCF). Elemental Analysis, (CHN): Found % (w/
w): C, 64.34; H, 4.86; N, 8.57.
Preparation of Fe(0)-Doped Nanoporous Carbons. FcPA-15
aerogel monoliths were transferred into a MTI GSL1600X-80 tube
furnace (alumina 99.8% pure, 72/80 mm inner/outer diameters, 457
mm heating zone). The temperature of the furnace was raised to the
desired temperature (400−1400 °C) at 5 °C min−1 under flowing H2
(150 mL min−1). Samples were heated at the prescribed temperature
for 5 h. At the end of the heating period the temperature returned to
room temperature at 5 °C min−1 under constant flow of H2. FcPA-15
samples processed at 800 °C/H2 comprised the basis for further study
and are referred to as Fe@C. For control purposes, FcPA-15 samples
were also treated at 800 and 1200 °C under flowing Ar (150 mL
min−1) for 5 h. Fe(0) was leached out of Fe@C by dipping monoliths
in concentrated HCl under vacuum for 24 h. The HCl solution was
changed every 6 h, and after the second wash it was noted that it
remained colorless. Subsequently, samples were washed with water
and acetone (2× with each solvent, 6 h each time) and were air-dried.
Transmetalation of Fe@C to Noble-Metal (M)-Doped Nano-
porous Carbons (tm-M@C). Fe@C monoliths were transmetalated
with noble metals (tm-M@C, M: Au, Pt, Pd) by dipping in the
corresponding metal ion solutions ([Au3+] = 0.018 M; [H2PtCl6] =
0.035 M; [PdCl2] = 0.035 M) under reduced pressure, right after they
came out of the furnace. The volume of each precious metal solution
was adjusted based on the expected amount of Fe(0) in each Fe@C
monolith (11.5 2.1% (w/w)); in general, in all cases the (volume of
metal ion solution):(volume of the Fe@C monolith) ratio was equal
to about 80. After 5 h in the respective transmetalation bath, monoliths
were placed in a water bath and were heated at around 50 °C for 5 h
still under reduced pressure. Subsequently, they were washed with
reaction and catalytic metal. Therefore, the common catalyst
“fatigue” observed in Figure 8 is attributed to the matrix, which,
owing to its open porous structure, may have allowed for
catalyst loss. This subject goes beyond the immediate scope of
this report but certainly warrants further investigation.
3. CONCLUSION
Although it is known that pyrolysis of mixtures of carbon or
carbon precursors with iron salts yields similar nanostructures
to those described here,39−43 the key feature of this work is that
macroscopically the resulting porous materials are robust
monoliths, a key requirement for easily redeployable catalysts.
The conceptual point of departure for this configuration was
the structure of catalytic converters. In that regard, it is
conceivably possible to transfer back and apply our findings to
our point of departure, but that requires supports inert toward
O2 at high temperatures (e.g., SiC aerogels55). With regards to
the C-supported catalysts at hand, it is rather straightforward to
expand the present approach to other metallocenes. However,
loss of ferrocene at two stages during processing (i.e., during
gelation and during the early stages of pyrolysis) renders that
route less attractive for expensive metals. On the other hand,
galvanic replacement is atom-efficient, steps-down the particle
size, and is a room-temperature process, which removes any
possibility for sintering. Our current attention is focusing on
carbide-free materials, like those obtained at 400/500 °C, and
in order to render Fe(0) nanoparticles accessible, we are
exploring how to deconvolute reduction and carbonization via,
for example, stepwise pyrolysis.
4. EXPERIMENTAL SECTION
4.1. Materials. All reagents and solvents were used as received,
unless noted otherwise. Ferrocene, aluminum chloride, acetyl chloride,
lithium aluminum hydride, sodium hydroxide, dichloromethane,
hexane, anhydrous N,N-dimethylformamide (DMF), butyl acrylate,
styrene (inhibitor was removed by extraction with 5 M solution of
NaOH followed by drying with anhydrous sodium sulfate),
acetophenone, hexadecane, benzaldehyde, iodobenzene, benzyl
alcohol, butyl cinnamate, triethyl amine, cis- and trans-stilbene,
chloroplatinic acid hydrate, palladium chloride, nickel(II) chloride
hexahydrate, and concentrated HCl (12.1 N) were purchased from
Sigma-Aldrich Chemical Co. A gold plating solution (catalog number:
42307) and rhodium(III) chloride trihydrate were purchased from Alfa
Aesar. Tris(4-isocyanatophenyl)methane (TIPM) was donated from
H
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
water (2×, 8 h each time) followed by acetone (2×, 8 h each time) and
were vacuum-dried overnight at 80 °C.
in 5 mL glass vials, which were ultrasonicated for 20 min to disperse
the smallest particles in the solvent. Immediately afterward, and just
before particle settling was complete, a single drop was taken and
placed on a 200 mesh copper grid bearing a lacey Formvar/carbon
film. The grid was allowed to air-dry overnight before microscopy. At
least six different areas/particles were examined to ensure that the
results were uniform over the entire sample.
The fundamental building blocks of the FcPA-xx aerogels were also
probed with small-angle X-ray scattering (SAXS), using ∼2 mm thick
disks cut dry with a diamond saw. SAXS analysis was carried out with
the same PANalytical X’Pert Pro multipurpose diffractometer (MPD)
described above configured for scattering, using a 1/32° SAXS slit, a
1/16° antiscatter slit on the incident beam side, and a 0.1 mm
antiscatter slit together with a Ni 0.125 mm automatic beam attenuator
on the scattered beam side. Samples were placed in circular holders
between thin Mylar sheets, and scattering intensities were measured by
running 2θ scans from −0.1° to 5° with a point detector in the
transmission geometry. All scattering data were reported in arbitrary
units as a function of Q, the momentum transferred during a scattering
event. Data analysis was conducted using the Beaucage Unified
Model56,57 applied with the Irena SAS tool for modeling and analysis
of small angle scattering within the Igor Pro application (a commercial
scientific graphing, image processing, and data analysis software from
Wave Metrics, Portland, OR).
4.2. Methods. Drying Procedure. Drying of wet-gels with
supercritical fluid (SCF) CO2 was carried out in an autoclave
(SPIDRY Jumbo Supercritical Point Dryer, SPI Supplies, Inc. West
Chester, PA, or in a Spe-edSFE system, Applied Separations,
Allentown, PA). Samples were loaded into the autoclave and acetone
was added until all samples were submerged. The pressure vessel was
closed and liquid CO2 was allowed in at room temperature. Acetone
was drained out from the pressure vessel as it was being displaced by
liquid CO2. Subsequently, more liquid CO2 was allowed into the vessel
and was drained out several more times until all acetone was extracted
out of the pores of the samples. The criterion for that was that vented-
out CO2 started forming dry ice. Subsequently, the temperature of the
autoclave was raised to 40 °C, and SCF CO2 was vented off as a gas.
Physical Characterization. Bulk densities (ρb) were calculated from
the weight and the physical dimensions of the samples. Skeletal
densities (ρs) were determined with helium pycnometry using a
Micromeritics AccuPyc II 1340 instrument. Porosities, Π, as percent of
open empty space were calculated from ρb and ρs via Π = 100 × [ρs −
ρb]/ρs.
Chemical Characterization. Elemental analysis (CHN) was
conducted with a PerkinElmer elemental analyzer (Model 2400
CHN).
Infrared (IR) spectra were obtained in KBr pellets, using a Nicolet-
FTIR Model 750 spectrometer. Raman spectroscopy of carbon
samples was conducted with a Jobin-Yvon micro-Raman spectrometer
with a 632.8 nm He−Ne laser as the excitation source.
Evaluation of Fe@C and tm-M@C as Catalysts. All four catalysts
(Fe@C, tm-Au@C, tm-Pt@C, tm-Pd@C) were used at a 5% mol/mol
ratio relative to the limiting reagent in the corresponding reaction
mixture. Starting materials and products were quantified using gas
chromatography and internal standards. Response factors were equal
to the slopes of calibration curves that were constructed with a series
of samples containing known concentrations of each reactant, product,
and internal standard.
Liquid 1H and 13C NMR spectra were recorded in deuterated
solvents using a 400 MHz Varian Unity Inova NMR instrument (100
MHz carbon frequency). Solid-state 13C NMR spectra were obtained
with samples ground into fine powders on a Bruker Avance III 400
MHz spectrometer with a carbon frequency of 100 MHz, using magic-
angle spinning (at 5 kHz) with broadband proton suppression and the
CPMAS TOSS pulse sequence for spin sideband suppression. Solid-
state 13C NMR spectra were referenced externally to glycine (CO:
176.03 ppm).
Reduction of nitrobenzene52 was catalyzed with Fe@C and was
carried out in a thick-jacketed round-bottom pressure flask with a
Teflon screw-cap. Nitrobenzene (0.984 g, 8 mmol) was dissolved in
THF (50 mL). Hydrazine hydrate (0.641 g, 20 mmol) was added as a
reducing agent and hexadecane (1000 μL, 3 mmol) as an internal
standard. The flask was sealed and the reaction mixture was stirred for
24 h at 100 °C. Aliquots (100 μL) were taken every 2 h by cooling the
flask temporarily to 40 °C and were analyzed immediately with GC.
After 24 h the reaction mixture was cooled down to room temperature,
and the Fe@C monolith was picked up with a pair of tweezers, rinsed
briefly with THF, and transferred immediately to a new reaction
mixture for the next cycle. The whole process was repeated five times.
Oxidation of benzyl alcohol53 was catalyzed either with tm-Au@C
or tm-Pt@C. Benzyl alcohol (0.864 g, 8 mmol) was dissolved in
distilled water (50 mL), and acetophenone (1000 μL, 8 mmol) was
added as an internal standard. The reaction mixture was heated with
an oil bath to 60 °C under constant bubbling of oxygen and vigorous
magnetic stirring. Aliquots (100 μL) were taken every 2 h, extracted
with diethyl ether (2 mL), and analyzed immediately with GC. After
24 h the reaction mixture was cooled down to room temperature, and
the tm-Au@C and tm-Pt@C monoliths were harvested out as above,
rinsed briefly with water, and transferred to new reaction mixture for
the next cycle. The whole process was repeated five times.
Powder X-ray diffraction (PXRD) was conducted with a
PANalytical X’Pert Pro multipurpose diffractometer (MPD) with Cu
Kα radiation (λ = 1.54 Å) and a proportional counter detector
equipped with a flat graphite monochromator. Phase composition was
estimated via Rietveld refinement of the X-ray diffraction patterns
utilizing RIQAS software (Materials Data, Inc., version 4.0.0.26).
Structural information for crystalline phases was obtained from the
ICSD database version 2.01. Crystallite sizes were calculated using the
Scherrer equation and the fwhm of the lowest-angle diffractions. A
Gaussian correction was applied utilizing NIST SRM 660a LaB6 to
determine the instrumental broadening.
Thermogravimetric analysis (TGA) was conducted under O2 at 5
°C min−1 using a TA Instruments Model TGA Q50 thermogravimetric
analyzer. The residue from Fe@C after TGA analysis was collected
and was analyzed with XRD. The only crystalline phase found was
Fe2O3.
Gas chromatography (GC) was carried out with a Hewlett-Packard
5890 Series II gas chromatograph equipped with a DB-5 capillary
column (30 m/0.25 mm) and a flame ionization detector (FID).
Characterization of the Porous Structure. Brunauer−Emmett−
Teller (BET) surface areas, pore volumes, and pore size distributions
for pores in the 1.7−300 nm range were determined with N2-sorption
porosimetry at 77 K using a Micromeritics ASAP 2020 surface area
and porosity analyzer. Pore size distributions were calculated by
applying the Barrett−Joyner−Halenda (BJH) equation to the
desorption branch of the isotherms.
Heck coupling reactions54 were catalyzed with tm-Pd@C at 80 °C,
with constant magnetic stirring under N2 in DMF (5 mL) using a
mixture of iodobenzene (1.632 g, 8 mmol), triethyl amine (0.809 g, 8
mmol), and butyl acrylate (1.28 g, 10 mmol) or styrene (1.04 g, 10
mmol). In both cases, hexadecane was added as an internal standard
(1000 μL, 3 mmol). The same protocol as above was observed for GC
analysis and recycling of the catalyst.
Characterization of the Skeletal Framework. Scanning electron
microscopy (SEM) was conducted with Au-coated samples on a
Hitachi Model S-4700 field-emission microscope. Transmission
electron microscopy (TEM) was conducted with an FEI Tecnai F20
instrument employing a Schottky field emission filament operating at a
200 kV accelerating voltage. The samples were finely ground by hand
in a mortar with a pestle, and the powder was mixed with isopropanol
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.chemma-
ter.6b02364.
I
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
Cross-Linked Resorcinol-Formaldehyde Aerogels. Chem. Mater. 2008,
20, 6985−6997.
(16) Conroy, D.; Moisala, A.; Cardoso, S.; Windle, A.; Davidson, J.
Carbon Nanotube Reactor: Ferrocene Decomposition, Iron Particle
Growth, Nanotube Aggregation and Scale-up. Chem. Eng. Sci. 2010,
65, 2965−2977.
(17) Chidambareswarapattar, C.; Xu, L.; Sotiriou-Leventis, C.;
Leventis, N. Robust Monolithic Multiscale Nanoporous Polyimides
and Conversion to Isomorphic Carbons. RSC Adv. 2013, 3, 26459−
26469.
(18) Leventis, N.; Sotiriou-Leventis, C.; Chandrasekaran, N.; Mulik,
S.; Larimore, Z. J.; Lu, H.; Churu, G.; Mang, J. T. Multifunctional
Polyurea Aerogels from Isocyanates and Water. A Structure−Property
Case Study. Chem. Mater. 2010, 22, 6692−6710.
(19) Leventis, N.; Chidambareswarapattar, C.; Mohite, D. P.;
Larimore, Z. J.; Lu, H.; Sotiriou-Leventis, C. Multifunctional Porous
Aramids (Aerogels) by Efficient Reaction of Carboxylic Acids and
Isocyanates. J. Mater. Chem. 2011, 21, 11981−11986.
(20) Iqbal, N.; Akhter, Z.; Saeed, M. A.; Khan, M. S. Synthesis and
Characterization of Some Novel Organometallic Aromatic Polyamides.
Appl. Organomet. Chem. 2006, 20, 344−350.
(21) Iqbal, N.; Saeed, M. A.; Khan, M. S.; Akhter, Z. Synthesis and
physicochemical studies of ferrocene-containing materials. Appl.
Organomet. Chem. 2007, 21, 862−869.
(22) Rosenblum, M.; Woodward, R. B. The Structure and Chemistry
of Ferrocene. J. Am. Chem. Soc. 1958, 80, 5443−5449.
Appendix I: Formulations of FcPA-xx aerogels. Appendix
II: Materials characterization data for FcPA-xx, Fe@C,
and tm-M@C. Appendix III: N2-sorption data. Appendix
IV: Small-angle X-ray scattering (SAXS) data for FcPA-
xx. Appendix V: FTIR of FcPA-xx. Appendix VI:
Additional PXRD data related to transmetalation and
removal of Fe(0) with HCl. Appendix VII: Additional
TEM of Fe@C and tm-M@C. Appendix VIII: Raman
data for FcPA-15 after pyrolysis at different temper-
atures. Appendix IX: Data from catalysis using Fe@C
and tm-M@C as catalysts (PDF)
AUTHOR INFORMATION
■
Corresponding Authors
*(N.L.) E-mail: leventis@mst.edu.
*(C.S.-L.) E-mail: cslevent@mst.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This project was funded by the Army Research Office under
Award Number W911NF-14-1-0369. We also thank Covestro
LLC (formerly Bayer Corporation U.S.A.) for the generous
supply of Desmodur RE (TIPM) and the Materials Research
Center of the Missouri University of Science and Technology
for support with materials characterization.
(23) Leventis, N.; Sotiriou-Leventis, C.; Saeed, A. M.; Donthula, S.;
Majedi Far, H.; Rewatkar, P. M.; Kaiser, H.; Robertson, J. D.; Lu, H.;
Churu, G. Nanoporous Polyurea from a Triisocyanate and Boric Acid:
A Paradigm of a General Reaction Pathway for Isocyanates and
Mineral Acids. Chem. Mater. 2016, 28, 67−78.
(24) Blagbrough, I. S.; Mackenzie, N. E.; Ortiz, C.; Scott, A. I. The
Condensation Reaction between Isocyanates and Carboxylic Acids. A
Practical Synthesis of Substituted Amides and Anilides. Tetrahedron
Lett. 1986, 27, 1251−1254.
(25) Sasaki, K.; Crich, D. Facile Amide Bond Formation from
Carboxylic Acids and Isocyanates. Org. Lett. 2011, 13, 2256−2259.
(26) Sorenson, W. R. Reaction of an isocyanate and a carboxylic acid
in dimethyl sulfoxide. J. Org. Chem. 1959, 24, 978−980.
(27) For the derivation of eq 1 see Appendix II, Section 3 in the
Supporting Information of ref 23.
(28) Leventis, N.; Chandrasekaran, N.; Sotiriou-Leventis, C.;
Mumtaz, A. Smelting in the Age of Nano: Iron Aerogels. J. Mater.
Chem. 2009, 19, 63−65.
(29) Leventis, N.; Chandrasekaran, N.; Sadekar, A. G.; Mulik, S.;
Sotiriou-Leventis, C. The Effect of Compactness on the Carbothermal
Conversion of Interpenetrating Metal Oxide/Resorcinol-Formalde-
hyde Nanoparticle Networks to Porous Metals and Carbides. J. Mater.
Chem. 2010, 20, 7456−7471.
REFERENCES
■
(1) Ertl, G. Reactions at Surfaces: From Atoms to Complexity
(Nobel Lecture). Angew. Chem., Int. Ed. 2008, 47, 3524−3535.
(2) Mizuno, N.; Misono, M. Heterogeneous Catalysis. Chem. Rev.
1998, 98, 199−217.
(3) Wang, Y.-J.; Wilkinson, D. P.; Zhang, J. Noncarbon Support
Materials for Polymer Electrolyte Membrane Fuel Cell Electro-
catalysts. Chem. Rev. 2011, 111, 7625−7651.
(4) Zaera, F. Nanostructured Materials for Applications in
Heterogeneous Catalysis. Chem. Soc. Rev. 2013, 42, 2746−2762.
(5) Yin, L.; Liebscher, J. Carbon−Carbon Coupling Reactions
Catalyzed by Heterogeneous Palladium Catalysts. Chem. Rev. 2007,
107, 133−173.
(6) Debe, M. K. Electrocatalyst Approaches and Challenges for
Automotive Fuel Cells. Nature 2012, 486, 43−51.
(7) Tomasic, V.; Jovic, F. State-of-the-art in the Monolithic
Catalysts/Reactors. Appl. Catal., A 2006, 311, 112−121.
(8) Twigg, M. V. Catalytic Control of Emissions from Cars. Catal.
Today 2011, 163, 33−41.
(30) Mahadik-Khanolkar, S.; Donthula, S.; Bang, A.; Wisner, C.;
Sotiriou-Leventis, C.; Leventis, N. Polybenzoxazine Aerogels. 2.
Interpenetrating Networks with Iron Oxide and the Carbothermal
Synthesis of Highly Porous Monolithic Pure Iron(0) Aerogels as
Energetic Materials. Chem. Mater. 2014, 26, 1318−1331.
(31) Leventis, N.; Donthula, S.; Mandal, C.; Ding, M. S.; Sotiriou-
Leventis, C. Explosive versus Thermite Behavior in Iron(0) Aerogels
Infiltrated with Perchlorates. Chem. Mater. 2015, 27, 8126−8137.
(9) Husing, N.; Schubert, U. Aerogels-airy Materials: Chemistry,
̈
Structure and Properties. Angew. Chem., Int. Ed. 1998, 37, 22−45.
(10) Pierre, A. C.; Pajonk, G. M. Chemistry of Aerogels and their
Applications. Chem. Rev. 2002, 102, 4243−4265.
(11) Morris, C. A.; Anderson, M. L.; Stroud, R. M.; Merzbacher, C.
I.; Rolison, D. R. Silica Sol as a Nanoglue: Flexible Synthesis of
Composite Aerogels. Science 1999, 284, 622−624.
́
(32) Maldonado-Hodar, F. J.; Moreno-Castilla, C.; Rivera-Utrilla, J.;
Hanzawa, Y.; Yamada, Y. Catalytic Graphitization of Carbon Aerogels
(12) Pekala, R. W.; Alviso, C. T.; Kong, F. M.; Hulsey, S. S. Aerogels
Derived from Multifunctional Organic Monomers. J. Non-Cryst. Solids
1992, 145, 90−98.
by Transition Metals. Langmuir 2000, 16, 4367−4373.
(33) Lu, A.-H.; Li, W.-C.; Salabas, E.-L.; Spliethoff, B.; Schuth, F.
̈
Low Temperature Catalytic Pyrolysis for the Synthesis of High Surface
Area, Nanostructured Graphitic Carbon. Chem. Mater. 2006, 18,
2086−2094.
(13) Al-Muhtaseb, S. A.; Ritter, J. A. Preparation and Properties of
Resorcinol−Formaldehyde Organic and Carbon Gels. Adv. Mater.
2003, 15, 101−114.
(34) Sun, Z.; Yan, Z.; Yao, J.; Beitler, E.; Zhu, Y.; Tour, J. M. Growth
of Graphene from Solid Carbon Sources. Nature 2010, 468, 549−552.
(35) Webb, P. A.; Orr, C. Analytical Methods in Fine Particle
Technology; Micromeritics Instrument Corporation: Norcross, GA,
U.S.A., 1997; pp 67−68.
(14) Li, W.-C.; Lu, A.-H.; Schuth, F. Preparation of Monolithic
̈
Carbon Aerogels and Investigation of Their Pore Interconnectivity by
a Nanocasting Pathway. Chem. Mater. 2005, 17, 3620−3626.
(15) Mulik, S.; Leventis, N.; Sotiriou-Leventis, C. Macroporous
Electrically Conducting Carbon Networks by Pyrolysis of Isocyanate
J
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX

Chemistry of Materials
Article
(36) McKenzie, D. R.; Muller, D.; Pailthorpe, B. A. Compressive-
stress-induced Formation of Thin-film Tetrahedral Amorphous
Carbon. Phys. Rev. Lett. 1991, 67, 773−776.
(56) Beaucage, G. Approximations Leading to a Unified Exponential/
Power-law Approach to Small-angle Scattering. J. Appl. Crystallogr.
1995, 28, 717−728.
(57) Beaucage, G. Small-angle Scattering from Polymeric Mass
Fractals of Arbitrary Mass-fractal Dimension. J. Appl. Crystallogr. 1996,
29, 134−146.
(37) Pierson, H. O. Handbook of Carbon, Graphite, Diamond and
Fullerenes; Noyes Publications: 1993.
(38) Canca
̧
do, L. G.; Takai, K.; Enoki, T.; Endo, M.; Kim, Y. A.;
Mizusaki, H.; Jorio, A.; Coelho, L. N.; Magalhaes Paniago, R.; Pimenta,
̃
M. A. General Equation for the Determination of the Crystallite Size
La of Nanographite by Raman Spectroscopy. Appl. Phys. Lett. 2006, 88,
163106.
(39) Fu, R.; Baumann, T. F.; Cronin, S.; Dresselhaus, G.;
Dresselhaus, M. S.; Satcher, J. H., Jr. Formation of Graphitic
Structures in Cobalt- and Nickel-Doped Carbon Aerogels. Langmuir
2005, 21, 2647−2651.
(40) Wang, X. X.; Tan, Z. H.; Zeng, M.; Wang, J. N. Carbon
Nanocages: A New Support Material for Pt Catalyst with Remarkably
High Durability. Sci. Rep. 2014, 4, 4437.
(41) Zhao, M.; Song, H. Catalytic Graphitization of Phenolic Resin. J.
Mater. Sci. Technol. 2011, 27, 266−270.
(42) Gom
́
ez de la Fuente, J. L.; Martínez-Huerta, M. V.; Rojas, S.;
Terreros, P.; Fierro, J. L. G.; Pena, M. A. Enhanced Methanol
̃
Electrooxidation Activity of PtRu Nanoparticles Supported on H2O2-
Functionalized Carbon Black. Carbon 2005, 43, 3002−3039.
(43) Yu, C.; Sun, Y.; Fan, X.; Zhao, Z.; Qiu, J. Hierarchical Carbon-
Encapsulated Iron Nanoparticles as a Magnetically Separable
Adsorbent for Removing Thiophene in Liquid Fuel. Part. Part. Syst.
Charact. 2013, 30, 637−644.
(44) Brankovic, S. R.; Wang, J. X.; Adzic, R. R. Metal Monolayer
Deposition by Replacement of Metal Adlayers on Electrode Surfaces.
Surf. Sci. 2001, 474, L173−L179.
(45) Papadimitriou, S.; Tegou, A.; Pavlidou, E.; Armyanov, S.;
Valova, E.; Kokkinidis, G.; Sotiropoulos, S. Preparation and Character-
isation of Platinum- and Gold-coated Copper, Iron, Cobalt and Nickel
Deposits on Glassy Carbon Substrates. Electrochim. Acta 2008, 53,
6559−6567.
(46) Dass, A.; Counsil, J. A.; Gao, X.; Leventis, N. Magnetic Field
Effects on the Open Circuit Potential of Ferromagnetic Electrodes in
Corroding Solutions. J. Phys. Chem. B 2005, 109, 11065−11073.
(47) Selvakannan, P. R.; Sastry, M. Hollow Gold and Platinum
Nanoparticles by a Transmetalation Reaction in an Organic Solution.
Chem. Commun. 2005, 1684−1686.
(48) Callister, W. D.; Rethwisch, D. G. Materials Science and
Engineering: An Introduction, 9th ed.; Wiley: New York, NY, 2014;
Chapter 17, p 681.
(49) Wu, H.; Pantaleo, G.; Venezia, A. M.; Liotta, L. F. Mesoporous
Silica Based Gold Catalysts: Novel Synthesis and Application in
Catalytic Oxidation of CO and Volatile Organic Compounds (VOCs).
Catalysts 2013, 3, 774−793.
(50) Corma, A.; Garcia, H. Supported Gold Nanoparticles as
Catalysts for Organic Reactions. Chem. Soc. Rev. 2008, 37, 2096−2126.
(51) Zhang, Y.; Cui, X.; Shi, F.; Deng, Y. Nano-Gold Catalysis in
Fine Chemical Synthesis. Chem. Rev. 2012, 112, 2467−2505.
(52) Jagadeesh, R. V.; Wienhofer, G.; Westerhaus, F. A.; Surkus, A.-
̈
E.; Pohl, M.-M.; Junge, H.; Junge, K.; Beller, M. Efficient and Highly
Selective Iron-catalyzed Reduction of Nitroarenes. Chem. Commun.
2011, 47, 10972−10974.
(53) Ng, Y. H.; Ikeda, S.; Harada, T.; Morita, Y.; Matsumura, M. An
Efficient and Reusable Carbon-supported Platinum Catalyst for
Aerobic Oxidation of Alcohols in Water. Chem. Commun. 2008,
3181−3183.
́
(54) Anderson, K.; Fernandez, S. C.; Hardacre, C.; Marr, P. C.
Preparation of Nanoparticulate Metal Catalysts in Porous Supports
Using an Ionic Liquid Route; Hydrogenation and C−C Coupling.
Inorg. Chem. Commun. 2004, 7, 73−76.
(55) Leventis, N.; Sadekar, A.; Chandrasekaran, N.; Sotiriou-
Leventis, C. Click Synthesis of Monolithic Silicon Carbide Aerogels
from Polyacrylonitrile-Crosslinked 3D Silica Networks. Chem. Mater.
2010, 22, 2790−2803.
K
DOI: 10.1021/acs.chemmater.6b02364
Chem. Mater. XXXX, XXX, XXX−XXX