Polyimide aerogels by ring-opening metathesis polymerization (ROMP)

Article abstract of DOI:10.1021/cm200323e

Polyimide aerogel monoliths are prepared by ring-opening metathesis polymerization (ROMP) of a norbornene end-capped diimide, bis-NAD, obtained as the condensation product of nadic anhydride with 4,4′-methylenedianiline. The density of the material was varied in the range of 0.13-0.66 g cm ^-3 by varying the concentration of bis-NAD in the sol. Wet gels experience significant shrinkage, relative to their molds (28%-39% in linear dimensions), but the final aerogels retain high porosities (50%-90% v/v), high surface areas (210-632 m² g^-1, of which up to 25% is traced to micropores), and pore size distributions in the mesoporous range (20-33 nm). The skeletal framework consists of primary particles 16-17 nm in diameter, assembling to form secondary aggregates (by SANS and SEM) 60-85 nm in diameter. At lower densities (e.g., 0.26 g cm^-3), secondary particles are mass fractals (D_m = 2.34 ± 0.03) turning to closed-packed surface fractal objects (D_S = 3.0) as the bulk density increases (≥0.34 g cm^-3), suggesting a change in the network-forming mechanism from diffusion-limited aggregation of primary particles to a space-filling bond percolation model. The new materials combine facile one-step synthesis with heat resistance up to 200 °C, high mechanical compressive strength and specific energy absorption (168 MPa and 50 J g^-1, respectively, at 0.39 g cm^-3 and 88% ultimate strain), low speed of sound (351 m s^-1 at 0.39 g cm^-3) and styrofoam-like thermal conductivity (0.031 W m^-1 K^-1 at 0.34 g cm ^-3 and 25 °C); hence, they are reasonable multifunctional candidate materials for further exploration as thermal/acoustic insulation at elevated temperatures.

Full text of DOI:10.1021/cm200323e

ARTICLE
pubs.acs.org/cm
Polyimide Aerogels by Ring-Opening Metathesis
Polymerization (ROMP)
Nicholas Leventis,*,† Chariklia Sotiriou-Leventis,*,† Dhairyashil P. Mohite,† Zachary J. Larimore,‡
Joseph T. Mang,*,§ Gitogo Churu,^ and Hongbing Lu*,^
†Department of Chemistry, Missouri University of Science and Technology, Rolla, Missouri 65409, United States
‡Department of Mechanical Engineering, Missouri University of Science and Technology, Rolla, Missouri 65409, United States
§Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States
^Department of Mechanical Engineering, The University of Texas at Dallas, Richardson, Texas 75080, United States
ABSTRACT: Polyimide aerogel monoliths are prepared by ring-opening metathesis poly-
merization (ROMP) of a norbornene end-capped diimide, bis-NAD, obtained as the condensa-
tion product of nadic anhydride with 4,40-methylenedianiline. The density of the material was
varied in the range of 0.13ꢀ0.66 g cmꢀ3 by varying the concentration of bis-NAD in the sol.
Wet gels experience significant shrinkage, relative to their molds (28%ꢀ39% in linear
dimensions), but the final aerogels retain high porosities (50%ꢀ90% v/v), high surface areas
(210ꢀ632 m2 gꢀ1, of which up to 25% is traced to micropores), and pore size distributions in
the mesoporous range (20ꢀ33 nm). The skeletal framework consists of primary particles
16ꢀ17 nm in diameter, assembling to form secondary aggregates (by SANS and SEM)
60ꢀ85 nm in diameter. At lower densities (e.g., 0.26 g cmꢀ3), secondary particles are mass
fractals (Dm = 2.34 ( 0.03) turning to closed-packed surface fractal objects (DS = 3.0) as the
bulk density increases (g0.34 g cmꢀ3), suggesting a change in the network-forming mechanism from diffusion-limited aggregation
of primary particles to a space-filling bond percolation model. The new materials combine facile one-step synthesis with heat
resistance up to 200 ꢀC, high mechanical compressive strength and specific energy absorption (168 MPa and 50 J gꢀ1, respectively,
at 0.39 g cmꢀ3 and 88% ultimate strain), low speed of sound (351 m sꢀ1 at 0.39 g cmꢀ3) and styrofoam-like thermal conductivity
(0.031 W mꢀ1 Kꢀ1 at 0.34 g cmꢀ3 and 25 ꢀC); hence, they are reasonable multifunctional candidate materials for further exploration
as thermal/acoustic insulation at elevated temperatures.
KEYWORDS: ROMP, norbornene, end-capped, polyimides, aerogels
polymerization (ROMP) of the monomer,15 and polyimides.16,17
1. INTRODUCTION
Several other types of aerogels based on soluble polymers such as
Aerogels are low-density nanoporous solids with high surface
polystyrene,18 polyacrylonitrile,19 and cellulose20 are prepared by
area, low thermal conductivity, and high acoustic attenuation.1,2
inducing phase separation of preformed polymers. The accelerated
They are prepared by converting and removing the pore-filling
interest in organic aerogels is driven by the facile tailoring
solvents of suitable wet gels as supercritical fluids (SCFs).3,4
of their properties by choosing the polymer, the straightforwardness
Inorganic aerogels are mostly based on silica and have been
of the polymerization process that facilitates synthesis, and the fact
studied more extensively. They are fragile materials, and con-
that inorganic aerogels, whose skeletal framework has been cross-
firmed applications have been only in specialized environments
linked covalently by a conformal polymer coating, demonstrate
(for example, as thermal insulators aboard planetary vehicles and
dramatically increased mechanical strength, rendering this class
as Cherenkov radiation detectors in certain nuclear reactors).
of materials suitable for applications previously inconceivable for
Other oxide aerogels are evaluated as energetic materials, or
aerogels for example, in ballistic protection (armor).21 Since the
starting materials for porous metals and ceramics.5ꢀ7
mechanical properties of the latter materials are dominated by the
polymer, purely polymeric aerogels with the structure and inter-
On the other hand, organic aerogels were first reported
together with the inorganic counterparts;3,4 however, early
particle connectivity of polymer-cross-linked aerogels should have
emphasis on the latter delayed their systematic investigation for
similar mechanical properties.
almost 60 years, until Pekala reported the bottom-up synthesis
In that context, interest in polyimide aerogels stems from the
of phenolic resin-type aerogels via condensation of resorcinol with
high mechanical strength and high thermal stability of the
formaldehyde (RF).8 Subsequently, several other types of bot-
tom-up polymer aerogels were reported, first by variation of the
phenolic resin chemistry (phenolꢀfurfural,9 cresolꢀformalde-
hyde,10 melamineꢀformaldehyde11), and then based on poly-
urethane,12 polyurea,13 and, more recently, polybenzoxazine,14
poly(bicyclopentadiene) synthesized via ring-opening metathesis
polymer,22 which would render this class of aerogels suitable
for high-temperature thermal and acoustic insulation. In general,
Received: February 1, 2011
Revised:
March 12, 2011
Published: March 30, 2011
r
2011 American Chemical Society
2250
dx.doi.org/10.1021/cm200323e Chem. Mater. 2011, 23, 2250–2261
|

Chemistry of Materials
ARTICLE
Table 1. Formulations of bis-NAD-xx Aerogels
sample
bis-NAD (% w/w versus NMP)
GC-II (% w/w versus bis-NAD)
GC-II (% mol versus bis-NAD)
bis-NAD: GC-II (mol:mol)
bis-NAD-2.5
bis-NAD-5
2.5
5.0
4.0
2.0
2.30
1.15
0.58
0.43
0.29
43.3
86.6
bis-NAD-10
bis-NAD-15
bis-NAD-20
10.0
15.0
20.0
1.0
173.2
231.0
346.4
0.75
0.50
there are two classic routes to consider for polyimide aerogel
synthesis: the first (DuPont process) yields linear polyimides and
involves reaction of dianhydrides with diamines,23 whereas the
second one, which is referenced as the polymerization of
monomeric reactants (PMR) route yields thermoset resins and
involves the synthesis and polymerization of norbornene-capped
imide oligomers.24 The DuPont route proceeds through a linear
polyamic acid that is dehydrated to the imide either chemically
(e.g., with sacrificial reagents such as acetic anhydride/pyridine),
or thermally at high temperatures. The PMR route has been
strictly a high-temperature process involving cross-linking of the
norbornene end-caps.
The first polyimide aerogels were prepared via the DuPont
process, using both chemical dehydration and high-temperature
treatment to complete imidization.16,17,25 Those conditions com-
pound the inherent economic disadvantage of supercritical fluids in
the aerogel synthesis. In that regard, recently, we introduced an
alternative route whereas polyimide aerogels can be obtained at
room temperature via reaction of dianhydrides with diisocya-
nates; thus, aerogels prepared from pyromellitic dianhydride and
4,40-methylenedianiline are chemically identical to those pre-
pared from the same anhydride and methylene diphenyl-
p-diisocyanate.26 Here, we introduce a second low-temperature
process to polyimide aerogels via the PMR route, whereas the
norbornene end-caps of a suitable bisnadimide, bis-NAD, are
polymerized via ROMP using the second-generation Grubbs’
catalysts GC-II.27 Evidently, bis-NAD-derived polyimide
aerogels are extremely robust multifunctional materials, com-
bining styrofoam-like thermal conductivity and mechanical
properties comparable to those of polymer-cross-linked silica
aerogels.21
7-methanoisoindole-1,3(2H)-dione)] was prepared from nadic anhy-
dride and MDA by a modification of literature procedures.28 In turn,
nadic anhydride was prepared via a DielsꢀAlder reaction of fresh
cyclopentadiene with maleic anhydride. Cyclopentadiene was prepared
via a reverse DielsꢀAlder reaction by refluxing DCPD at 180 ꢀC.
Cyclopentadiene was collected in an ice-cooled receiver and used for
further reaction with maleic anhydride. The latter (4.00 g, 0.0408 mol)
was first dissolved in ethyl acetate (15.0 mL) at room temperature under
magnetic stirring. The solution was cooled for 15 min in an ice bath, and
freshly prepared cyclopentadiene (4.0 mL) was added in the solution.
The reaction mixture was stirred for another 20 min and 15.0 mL of
hexane was added to complete precipitation of the crude product. The
precipitate was separated by vacuum filtration and purified by recrys-
tallization from ethyl acetate/hexane. Subsequently, nadic anhydride
(1.656 g, 0.0101 mol) was dissolved in anhydrous NMP (15.0 mL) at
room temperature under magnetic stirring. MDA (1.000 g, 0.005 mol)
was added and the reaction mixture was stirred at room temperature for
24 h under N2. At the end of the period, acetic anhydride (6.180 g,
0.0606 mol) and pyridine (1.0 mL) were added and the reaction mixture
was heated at 100 ꢀC for 6 h. The mixture was allowed to cool at room
temperature, and the precipitate was washed with methanol, followed by
drying under vacuum at 70 ꢀC for 24 h. Yield: 2.0 g (75%); mp
1
243ꢀ245 ꢀC (lit.28a mp 244 ꢀC for the endo,endo- isomer). H NMR
(400 MHz, CDCl3) δ 7.22 (d, J = 8.4 Hz, 4H), 7.05 (d, J = 8.4, 4H), 6.24
(t, J = 1.8, 4H), 3.98 (s, 2H), 3.49 (m, 4H), 3.41 (dd, J = 1.5 Hz, J = 2.9
Hz, 4H), 1.75 (dt, J = 1.6 Hz, J = 8.8 Hz, 2H), 1.58 (dt, J = 1.6 Hz, J = 8.8
Hz, 2H); 13C NMR (100 MHz, CDCl3) δ 176.78, 140.76, 134.49,
129.88, 129.56, 126.56, 52.13, 45.66, 45.38, 41.01; IR (KBr) 2990, 1770,
1710, 1510, 1380, 1180, 840, 745, 620 cmꢀ1. Elemental Analysis, (CHN
% w/w). Found: C: 75.47; H: 5.04; N: 5.71. Theoretical: C: 75.92; H:
5.31; N: 5.71.
2.3. Synthesis of Polyimide Aerogels from bis-NAD via
ROMP. Polyimide aerogels were prepared by mixing two solutions, one
containing bis-NAD in NMP and one with moisture-tolerant Grubbs’
catalyst GC-II in toluene. Different sets of samples with different bulk
densities were obtained by varying the concentration of bis-NAD.
Aerogel samples are abbreviated as bis-NAD-xx, where the extension -
xx stands for the weight percentage of bis-NAD in the bis-NAD plus
NMP mixture. All formulations are summarized in Table 1. Because bis-
NAD has limited solubility in NMP at room temperature, heating at
60 ꢀC was required in order to make the 2.5% and 5% w/w bis-NAD
solutions, while the 10%, 15%, and 20% w/w bis-NAD solutions were
heated at 90 ꢀC. GC-II in 50 μL of toluene (see Table 1) was added to
the bis-NAD solution, and the mixture was shaken vigorously and was
poured into molds (Wheaton polypropylene OmniVials, Part No.
225402, 1 cm in diameter, or 15-cm3 Fisherbrand Class B amber glass
threaded vials, 1.8 cm in inner diameter, Part No. 03-339-23D; the latter
molds were used for samples prepared for compression testing). All
solutions gelled within 10ꢀ20 min, except the 20% w/w bis-NAD sol,
which gelled within 1 min. The resulting wet gels were aged in their
molds for 12 h at 90 ꢀC, washed with NMP (4 washes, 8 h per wash), 1,
4-dioxane (4 washes, 8 h per wash), acetone (4 washes, 8 h per wash),
and dried in an autoclave with liquid CO2 to yield bis-NAD-xx
polyimide aerogels.
2. EXPERIMENTAL SECTION
2.1. Materials. All reagents and solvents were used as received,
unless noted otherwise. Dicyclopentadiene (DCPD), a second-genera-
tion Grubbs’ catalyst, GC-II, and anhydrous N-methyl-2-pyrrolidone
(NMP) were purchased from Aldrich Chemical Co. Maleic anhydride,
4,40-methylenedianiline (MDA), and 1,4-dioxane were purchased from
Acros Organics. Methanol and laboratory-grade NMP were purchased
from Fisher.
2.2. Synthesis of bis-NAD. Bis-NAD [IUPAC name: 2,
20-(methylenebis(4,1-phenylene))bis(3a,4,7,7a-tetrahydro-1H-4,
2251
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
2.4. Methods. Drying with liquid CO2, taken out as a supercritical
fluid (SCF), was conducted in an autoclave (SPI-DRY Jumbo Super-
critical Point Dryer, SPI Supplies, Inc. West Chester, PA). Bulk densities
(Fb) were calculated from the weight and the physical dimensions of
the samples. Skeletal densities (Fs) were determined with helium
pycnometry, using a Micromeritics AccuPyc II 1340 instrument. Por-
ositiy (Π) values were determined from Fb and Fs values:
bis-NAD-xx were multiplied by the calibration factor (0.920 ( 0.028)
determined with rutile, KCl, Al, graphite, and corundum just before our
experiments.
3. RESULTS AND DISCUSSION
3.1. Synthesis and Characterization of bis-NAD. The
monomer, bis-NAD, was prepared in high yield (75%) via the
DuPont route from nadic anhydride and 4,40-methylenedianiline
(MDA) via chemical dehydration of the intermediate diamic acid
(see Scheme 1) and was characterized by elemental analysis,
infrared (IR) spectroscopy, 1H nuclear magnetic resonance
(NMR), and 13C NMR. The IR spectrum (Figure 1) is dominated
by the imide CdO symmetric and asymmetric stretches at 1710 and
"
#
ꢁ
ꢀ
ꢁ
ꢀ
1=Fb ꢀ 1=Fs
1=Fb
Π ¼ 100 ꢁ
Surface areas and pore size distributions were measured by N2 sorption
porosimetry, using a Micromeritics ASAP 2020 surface area and porosity
analyzer. Samples for surface area and skeletal density determination
were outgassed for 24 h at 80 ꢀC under vacuum before analysis. Average
pore diameters were determined by the 4 ꢁ VTotal/σ method, where
VTotal is the total pore volume per gram of sample and σ, the surface area
determined by the BrunauerꢀEmmettꢀTeller (BET) method. VTotal
can be calculated either from the single highest volume of N2 adsorbed
along the adsorption isotherm or from the relationship VTotal = (1/Fb) ꢀ
(1/Fs). If the two average pore diameters coincide, it is taken as proof
that the material lacks macroporosity. Liquid 1H and 13C NMR of bis-
NAD were obtained with a 400 MHz Varian Unity Inova NMR
instrument (100 MHz carbon frequency). Elemental analysis was
conducted using a PerkinꢀElmer elemental analyzer (Model 2400
CHN). Infrared (IR) spectra were obtained in KBr pellets, using a
Nicolet-FTIR Model 750 spectrometer. Solid-state 13C NMR spectra
were obtained with samples ground into fine powders on a Bruker
Avance 300 spectrometer with a carbon frequency of 75.475 MHz, using
magic-angle spinning (at 7 kHz) with broadband proton suppression
and the CPMAS TOSS pulse sequence for spin sideband suppression.
Thermogravimetric analysis (TGA) was conducted under N2 or air with
a TA Instruments Model TGA Q50 thremogravimetric analyzer, using a
heating rate of 10 ꢀC minꢀ1. Scanning electron microscopy (SEM) was
conducted with Au-coated samples on a Hitachi Model S-4700 field-
emission microscope. The crystallinity of the polyimide samples was
determined by X-ray diffraction (XRD) using a Scintag Model 2000
diffractometer with Cu KR radiation and a proportional counter
detector equipped with a flat graphite monochromator. The identity
of the fundamental building blocks of the two materials was probed with
small-angle neutron scattering (SANS), using ∼2-mm-thick disks cut
with a diamond saw from cylinders similar to those used for mechanical
testing, on a time-of-flight, low-Q diffractometer (LQD) at the Manuel
Lujan, Jr. Scattering Center of the Los Alamos National Laboratory.29
The scattering data are reported in the absolute units of differential cross
section per unit volume (cmꢀ1), as a function of Q, the momentum
transferred during a scattering event. Quasi-static mechanical testing
under compression was conducted on an Instron Model 4469 universal
testing machine frame, following the testing procedures and specimen
length-to-diameter ratio (2.0 cm/1.0 cm) that was specified in ASTM
D1621-04a (“Standard Test Method for Compressive Properties of
Rigid Cellular Plastics”). The recorded force, as a function of displace-
ment (machine-compliance corrected), was converted to stress as a
function of strain. The thermal diffusivity (R) of the bis-NAD-xx
aerogels was measured at 23 ꢀC with a Netzsch NanoFlash Model
LFA 447 Flash diffusivity instrument, using disk samples ∼1 cm in
diameter, 2.0ꢀ2.2 mm thick (the thickness of each sample was measured
with 0.01 mm resolution and was entered as required by the data analysis
software). Heat capacities at 23 ꢀC of powders of the same samples (4ꢀ
8 mg), needed for the determination of their thermal conductivity, λ, were
measured using a TA Instruments Differential Scanning Calorimeter
Model Q2000 calibrated against a sapphire standard and run from ꢀ10
to 40 ꢀC at 0.5 ꢀC minꢀ1 in the modulated T4P mode, using 60 s
modulation period and 1 ꢀC as modulation amplitude. The raw data with
1770 cmꢀ1, respectively, and by the CꢀN stretch at 1380 cmꢀ1
.
The absorption at 1510 cmꢀ1 is assigned to the CdC stretch,
while the absorption at 1170 cmꢀ1 is attributed to the dCꢀH in-
plane bending from both the nadic and aromatic rings. In 13C
NMR (Figure 2, peak assignment by simulation), all carbons of
bis-NAD are resolved. No impurities are visible, which is
consistent with the elemental analysis data (see the Experimental
Section). The resonance at 177 ppm is assigned to the imide
carbonyl, the one at 134 ppm is assigned to the sp2-carbons of the
norbornene moieties, and the several resonances between 125
and 145 ppm are assigned to the aromatic carbons. The peak at
41 ppm is due the ꢀCH2ꢀ group of MDA, while the peaks
between 43 and 55 ppm are assigned to the aliphatic carbons of
the norbornene end-caps. Using thermogravimetric analysis
(TGA) (Figure 3A), bis-NAD is shown to be thermally stable,
Scheme 1. Synthesis of bis-NAD
Figure 1. Infrared (IR) data for bis-NAD monomer and a representa-
tive ROMP-derived polyimide aerogel.
2252
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
Scheme 2. Primary Thermal Decomposition Mechanism of
bis-NAD
become important in assessing whether all norbornene end-caps
react during the ROMP gelation process, as discussed below.
3.2. Synthesis of bis-NAD-xx Polyimide Aerogels. Bis-
NAD-related molecules (e.g., with 4,40-dioxyaniline bridges)
have been cross-linked before thermally or with microwaves.30
Cross-linking of bis-NAD itself via ROMP is summarized in
Scheme 3. Monoliths with different densities were obtained by
varying the monomer concentration. The amount of the GC-II
catalyst was varied inversely to the monomer concentration, in
order to keep the gelation time under 20 min. Wet gels were aged
in their molds for 12 h at 90 ꢀC to ensure that all monomer is
consumed and incorporated in the gels.31 This was confirmed in
two ways: first, by analyzing the washes for unreacted monomer,
and second by the mass balance between the aerogels and the
amount of bis-NAD used for their preparation. Nevertheless,
aging has not been optimized time-wise. Wet gels were solvent-
exchanged from NMP, through 1,4-dioxane, to acetone before
they were dried with liquid CO2 taken out superscritically at the
end. Right after gelation, wet gels are yellowish-brown; they look
off-white after NMP washes, because of remaining traces of the
catalyst, and completely white after treatment with 1,4-dioxane
and acetone. Dry aerogels were opaque-white (see photograph in
Scheme 3).
Figure 2. (A) Representative CPMAS 13C NMR spectrum of a ROMP-
derived polyimide aerogel (case shown: bis-NAD-10). The resonance at
29.65 ppm is attributed to residual solvent (acetone). (B) Liquid 13C
NMR of the bis-NAD monomer in CDCl3 (marked “S”). For peak
assignments, see the structure described in the text.
3.3. Characterization of bis-NAD-xx Aerogels. 3.3.a. Chem-
ical Characterization. The polymerization of bis-NAD pro-
ceeds according to Scheme 4. ROMP does not alter the identity
of the functional groups, and elemental analysis of bis-NAD-xx
gives results similar to those for the monomer (%w/w: C, 73.99;
H, 5.21; N, 5.61; versus C, 75.47; H, 5.04; N, 5.71 for the
monomer). Similarly, in IR (Figure 1) analysis, the most
prominent differences between bis-NAD and bis-NAD-xx are
associated with the CH2 and CH stretches that move to lower
frequencies after ring opening of the nadimide (from the
2987ꢀ2871 cmꢀ1 range to the 2937ꢀ2855 cmꢀ1 range), and
an increase from 750 cmꢀ1 to 805 cmꢀ1 in the absorption
frequency of the dCꢀH out-of-plane bending. In CPMAS 13C
NMR of bis-NAD-xx, after ring opening, the alkene carbon
resonance moves upfield (from 134 ppm originally), merging
with the aromatic carbons. The resonance of the bridgehead
carbon (labeled as “c”) moves also upfield from 52 to the 30ꢀ42
ppm range, merging with the MDA methylene bridge (Figure 2).
After ring opening, the “b” and “d” aliphatic carbons of the
norbornene ring move slightly downfield, from 45.66 ppm to
48.50 ppm and from 45.38 ppm to 45.66 ppm, respectively.
Those spectroscopic changes, however, do not warrant that all
norbornene moieties have been cross-linked. TGA of bis-NAD-xx
Figure 3. (A) Thermogravimetric analysis (TGA) data for the monomer
(bis-NAD) at 10 ꢀC minꢀ1. (B) TGA data for a representative ROMP-
derived polyimide sample, as shown at the same heating rate.
up to ∼220 ꢀC, undergoing a 16% mass loss between 220 ꢀC and
350 ꢀC, because of a reverse DielsꢀAlder reaction (loss of
cyclopentadiene was confirmed by mass spectrometry). The
observed mass loss corresponds to a loss of one end-cap per
bis-NAD molecule (see Scheme 2), which, in turn, may imply
that the newly created maleimide reacts with the norbornene
end-cap of another molecule to a more stable adduct. This matter
was not investigated further; however, the TGA data of Figure 3A
2253
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
Scheme 3. Flowchart for the Preparation of Polyimide Aerogels from bis-NAD
Scheme 4. Polymerization of bis-NAD via ROMP
blocks of the skeletal framework. This, in turn, implies early
phase separation of a polymer with substantial linearity.
3.3.b. Microstructural Characterization. Porosity and pore
structure are reported as a function of the bulk density (Fb), which, in
turn, is related to the monomer concentration in the sol. Results are
summarized in Table 2. The morphology of the pore walls (i.e.,
the skeletal framework) was inspected via scanning electron
microscopy (SEM) and their composition was investigated at the
fundamental building block level with SANS.
Although all bis-NAD is incorporated in the final aerogels, Fb
does not vary linearly with the concentration of the monomer in
the sol: e.g., the Fb value of bis-NAD-20 is 0.660 g cmꢀ3 but that
of bis-NAD-2.5 is 0.134 g cmꢀ3, not 8 times less as expected
from the relative concentrations of the monomer. This is because
all samples shrink in reverse order to the concentration of bis-NAD
in the sol: bis-NAD-2.5 samples shrink 39%, relative to the molds,
while bis-NAD-20 samples shrink less (28%; see Table 2). Minimal
shrinking (1ꢀ3% in linear dimensions) is observed during gela-
tion and aging (syneresis), no further shrinkage takes place
during NMP and 1,4-dioxane washes, while the majority of
shrinkage is observed during the final acetone washes. No
shrinking is observed during SCF drying. Therefore, exhibiting
typical gel-like semipermeable membrane behavior, bis-NAD-xx
wet gels swell until the pressure created by stretching of the
framework -which, therefore, must be quite flexible- is balanced
by the osmotic pressure of the internal “solution.”32 Thus,
changing the polarity of the solvent changes the degree of
swelling. Interparticle covalent bonding is more prevalent in
higher-density samples; hence, they stretch less, swell less, and,
therefore, shrink less.
Figure 4. X-ray diffraction (XRD) spectra of three representative
ROMP-derived polyimide aerogels (as shown).
in N2 (see Figure 3B) shows a small initial mass loss (∼3%) below
100 ꢀC (attributable to residual solvents), and a second small mass
loss (∼3%) in the 190ꢀ240 ꢀC range consistent with a reverse
DielsꢀAlder reaction of a small amount of dangling unreacted
norbornene moieties, according to Scheme 2. These moieties may
become inaccessible to the catalyst in a closely packed polymer. In
that regard, XRD shows that all samples have a high degree of
crystallinity (up to 53%; see Figure 4), suggesting a regular
packing of the polymer chains within the fundamental building
All skeletal density (Fs) values fall in the 1.26ꢀ1.36 g cmꢀ3
range, the variation is significant, but since there is no systematic
trend with the monomer concentration, it is rather attributed to
random error. Porosity (Π) values, which are calculated from the
Fb and Fs values, decrease from 90% to 48% as the bulk density
2254
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
Table 2. Selected Properties of Polyimide Aerogels via ROMP
diameter
(cm)a
shrinkage
(%)a,b
bulk density,
skeletal density,
porosity Π
surface area,
average pore
diameter (nm)e
average pore
diameter (nm)f
a
c
d
sample
Fb (g cmꢀ3
)
Fs (g cmꢀ3
)
(% void space)
(m2 gꢀ1
)
bis-NAD-2.5
bis-NAD-5
0.609 ( 0.005
0.639 ( 0.005
0.685 ( 0.005
0.705 ( 0.007
0.725 ( 0.005
39 ( 1
36 ( 1
32 ( 1
30 ( 1
28 ( 1
0.134 ( 0.002
0.261 ( 0.005
0.341 ( 0.011
0.507 ( 0.014
0.660 ( 0.019
1.360 ( 0.014
1.325 ( 0.008
1.260 ( 0.003
1.292 ( 0.004
1.260 ( 0.002
90.1
80.3
72.9
60.7
47.6
632 [180]
524 [124]
438 [72]
298 [29]
210 [19]
13.8 [42.6]
16.7 [23.5]
16.4 [19.6]
14.8 [16.1]
12.9 [13.7]
33.4 [69.7]
42.8 [67.6]
42.6 [44.2]
30.7 [12.8]
20.4 [5.0]
bis-NAD-10
bis-NAD-15
bis-NAD-20
a Average of three samples. (Mold diameter: 1.0 cm). b Shrinkage =100 ꢁ (sample diameter ꢀ mold diameter)/(mold diameter). Single sample,
average of 50 measurements. d BET surface area of the micropore (by t-plot, using the Harkins and Jura method). e By the 4 ꢁ VTotal/σ method. For the
first number, VTotal was calculated by the single-point adsorption method; for the number in brackets, VTotal was calculated via VTotal = (1/Fb) ꢀ (1/Fs).
f From the BJH desorption plot. First numbers are the peak maxima; numbers in brackets are the widths at half maxima.
c
mesoporous materials and “ink bottle”-type pores.33 Quantita-
tively, average pore diameters calculated by the 4 ꢁ VTotal/σ
method, using VTotal either from the maximum volume adsorbed
from the isotherms (captures mesopores), or from VTotal = (1/
Fb) ꢀ (1/Fs) (captures all pores) diverge significantly at lower
densities (signifying macroporosity), but converge for the denser
samples (signifying mesoporosity; see Table 2). Similarly, BJH-
desorption plots (shown as insets in Figure 5) give broad (with
hints for bimodal) pore-size distributions for the lower density
samples, but they are quite narrow and monomodal at higher
densities. (It is noted that, although BJH maxima are also
summarized in Table 2, they should not be considered quantita-
tively, because all adsorptionꢀdesorption isotherms are consis-
tently open-looped, indicating swelling of nonrigid pores33,34 in
agreement with conclusions reached above from shrinkage data.)
At the low P/P0 end of the isotherms, the significant quick rise of
the volume of N2 adsorbed indicates the presence of a significant
fraction of micropores. Data analysis within the range of 0.05 <
P/P0 < 0.3, according to the BET model (Table 2), shows that, at
low densities, the surface area is quite high (up to 632 m2 gꢀ1 for
bis-NAD-2.5), decreasing (but remaining quite significant) to
210 m2 gꢀ1 at Fb = 0.660 g cmꢀ3 (bis-NAD-20 samples). t-Plot
analysis of the isotherms within the range of 0.05 < P/P0 < 0.5,
using the Harkins and Jura method,35 shows that, at low
densities, up to 35% of the surface area comes from micropores,
decreasing to <10% in the denser samples (bis-NAD-15 and bis-
NAD-20).
SEM analysis (Figure 6) shows that all of the samples consist
of particles that are agglomerating together to form larger
clusters. Lower-density samples (<0.5 g cmꢀ3) show clearly
the presence of macropores, which is consistent with the
N2-sorption analysis above. The minimum particle diameter
observed by SEM (∼20 nm) is rather uniform throughout all
densities. Those smallest particles assemble to larger aggregates
(50ꢀ100 nm in diameter), but they are best defined (discernible) in
the lowest and highest density samples (bis-NAD-5 and bis-NAD-20,
respectively). The medium density samples (bis-NAD-10) arefuzzier
and the smallest particles look as if they are fused together into larger
clusters. The smallest particles in the highest density samples (bis-
NAD-20) are dispersed rather uniformly in space, rendering identi-
fication of the larger aggregates difficult. A quantitative assessment of
the makeup of the skeletal framework was obtained with SANS; the
data are included in Figure 6. In addition, the results obtained by
applying the Beaucage Unified Model analysis,36 which models the
samples as having multiple levels of structure, each with a distinct
characteristic length, corresponding hierarchically (starting from high
Q values) to a particle, aggregate, and agglomerate, are summarized in
Figure 5. N2 sorption isotherms for the bis-NAD-xx aerogels, as a
function of the bulk density (Fb). Insets show BJH desorption plots.
Data are summarized in Table 2.
increases (see Table 2). N2-sorption porosimetry (Figure 5) suggests
that the most dense samples are strictly mesoporous. Specifically, at
lower densities (e0.34 g cmꢀ3), the N2 absorption isotherms rise
above a value of P/P0 = 0.9 and do not reach a well-defined
saturation plateau, indicating that a significant portion of the
porosity is due to macropores (defined as pores with diameters of
>50 nm); on the other hand, the same isotherms do show narrow
hysteresis, which is an indication of some mesoporosity. As the
bulk density increases (g0.5 g cmꢀ3) the onset of the quick rise
in the volume of N2 adsorbed moves to lower P/P0 values (to P/
P0 ≈ 0.8), the isotherms reach saturation and they show large
H2-type hysteresis loops, all of which is consistent with mostly
2255
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
Figure 6. Scanning electron microscopy (SEM) and small-angle neutron scattering (SANS) analyses of ROMP-derived bis-NAD-xx aerogels, as a
function of the bulk density (Fb). Refer to Table 3 for information from the SANS data and for the meaning of Regions IꢀIV .
Table 3. Small-Angle Neutron Scattering (SANS) Data for Polyimide Aerogels via ROMP (Footnoted Terms Refer to Figure 6)
Primary Particles
Secondary Particles
Porod shapea
RG (nm)b []c
RG (nm)d []c
Dm
DS
e
e
sample
bis-NAD-5
bis-NAD-10
bis-NAD-20
4.0 ( 0.1
4.0 ( 0.1
4.0 ( 0.1
6.3 ( 1.1 [16.4]
6.5 ( 1.2 [16.9]
6.1 ( 1.7 [15.8]
f
2.34 ( 0.03
32.8 ( 2.5 [85.2]
23.1 ( 1.1 [60]
2.9 ( 0.1
3.0 ( 0.5
a From Region I. b From Region II. c The value in brackets represents the particle diameter, which is defined as 2R, where R is the particle radius. The
radius of gyration is given as RG = 0.77R. d From Region IV. e From Region III. f Region IV in this sample was beyond the experimentally accessible range
of the scattering vector Q; thus, RG could not be estimated.
Table 3. Furthermore, the Unified Model allows analysis of such
hierarchical structures over the full range of Q, the momentum
transferred in a scattering event, allowing deconvolution of over-
lapping length scales that can lead to subtle changes in the data, such
as a change in slope. Over all densities and the entire range of the
scattering Q, SANS shows two Guinier regions (knees), which are
indicative of characteristic length scales and two power-law regions
(which appear to be linear on the logꢀlog plots in Figure 6). Plots of
the SANS data found in Figure 6 have been divided into four regions
(in Q), for ease of identifying the different structural levels. The
Unified Model provides the radii of gyration (RG) from the Guinier
knees (Regions II and IV in the data of Figure 6) and the fractal
dimensions of the secondary particles from the linear region (Region III).
The slope of the data in Region I provides information about the
surface characteristics of the primary particles themselves. Matching
and comparing the SANS data of Table 3 with the SEM data of
Figure 6 identifies the minimum SEM particles as polydisperse, but
monomodal primary particles, 16ꢀ17 nm in diameter, with smooth
(nonfractal) interfaces (the slopes in Region I are all uniformly equal
to 4.0). The size of the primary particles does not change with
density (i.e., the concentration of bis-NAD in the sol), in analogy
to silica.37 Secondary aggregates are larger for medium-density
samples (85.2 nm in diameter for bis-NAD-10), but their size
decreases as the density increases (60 nm for bis-NAD-20).
Guinier Region IV for bis-NAD-5 was located at the edge of the
accessible Q-range and the secondary particle size could not be
2256
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
measured. Nevertheless, in those lower-bulk-density samples,
primary particles assemble into the secondary aggregates fractally
(mass fractal dimension, Dm = 2.34 ( 0.03), which suggests
diffusion-limited aggregation as the growth mechanism.38 As the
bulk density increases (bis-NAD-10 and bis-NAD-20 samples),
the Region III exponent increases and falls at the limit between
mass and surface fractals. The level of uncertainty associated with
the large size of both the primary and secondary particles causes
significant overlap of Guinier Regions II and IV with the linear
Region III, from which the fractal dimension is estimated, making
assessment by SANS alone inconclusive. However, combining
the SANS and SEM results suggests that we are dealing with
surface fractals of nonfractal objects. With the Region III slope
being attributed to surface fractals with DS = 3.0, secondary
particles of bis-NAD-10 and bis-NAD-20 are then classified as
surface fractal closed-packed objects. Transition from the more-
open fractal structure to the more-dense nonfractal one justifies
the decrease in the percentage of micropore surface area, from
35% in the lower-density samples to <10% in the higher-density
samples.
Considering the XRD and SANS data together—namely, the
closed packing of the polymeric strands implied by XRD and the
invariance of the primary particle size revealed by SANS—
supports fast polymerization to mostly linear oligomers that
always reach their solubility limit at the same point, irrespective
of the concentration of bis-NAD in the sol, and become phase-
separated into uniform-sized primary particles, which are sur-
face-active through dangling norbornene moieties, or catalyst-
terminated polymer strands. At the lowest concentrations (e.g.,
bis-NAD-5 samples), primary particles react with each other via a
diffusion-limited mechanism to form fractal secondary particles that,
in turn, form a gel. Values of DS = 3.0 at higher concentrations (e.g.,
bis-NAD-10 and bis-NAD-20 samples) might be associated with
fast ROMP, which fills the sol with primary particles that react
with their next neighbor through a bond percolation model,
yielding non(mass) fractal secondary objects.39 Cross-linking of
the polymer strands most probably continues throughout those
gel-forming processes. In the case of the bis-NAD-20 samples,
extremely fast ROMP consumes all monomer quickly (recall, for
example, that gelation occurs within <1 min in those samples).
Somewhat slower ROMP in the more-dilute bis-NAD-10 sam-
ples is followed by the accumulation of monomer on the
secondary particles (a monomer-cluster-growth-like process),
which explains the fuzziness in SEM observation.
3.3.c. Application-Related Bulk Properties. Polyimides are
thermally stable polymers; therefore, appropriate applications
for bis-NAD-xx aerogels include high-temperature thermal
and acoustic insulation. Relevant properties to monitor in-
clude thermal stability, mechanical strength, and thermal
conductivity.
Thermal Stability. Despite their resemblance to PMR-type
polyimides (both materials are prepared from the same norbor-
nene end-capped oligomers),24 ROMP-derived bis-NAD-xx
aerogels have unsaturated backbones (see Scheme 4). Therefore,
their use in air might be problematic. Indeed, TGA (see Figure 3B)
of bis-NAD-xx shows a mass increase above ∼200 ꢀC, presumably
by reaction with oxygen. PMR-type polyimides are rated for
operation up to 10 000 h at 290 ꢀC,24 which is obviously not
possible with as-prepared bis-NAD-xx. Increasing the molecular
weight of the monomer, or post-gelation saturation of the double
bonds might be approaches around this issue. Conveniently, that
process could be coupled with increasing the hydrophobicity and
reducing the flammability of the material.
Figure 7. Top: Stressꢀstrain curves under quasi-static compression of the
bis-NAD-xx aerogels, as a function of the bulk density (bis-NAD-5, Fb =0.24
g cmꢀ3 (curve a); bis-NAD-10, Fb = 0.39 g cmꢀ3 (curve b); bis-NAD-15,
Fb =0.53gcmꢀ3 (curvec);andbis-NAD-5,Fb =0.63gcmꢀ3 (curved)).Inset
shows a magnification of the early, almost-elastic region, including loadingꢀ
unloading data for a bis-NAD-15 sample. Bottom of the figure shows
photographs of two samples as indicated, before and after compression,
showing the different modes of failure.
Mechanical Properties. As stated in the Introduction, organic
aerogels are pursued partly as alternatives to polymer cross-
Table 4. Mechanical Characterization Data under Quasi-static Compression of Polyimide Aerogelsa
Specific Energy, abs
bulk density,
strain
rate
Young’s
modulus,
E (MPa)
speed of
sound
yield stress
at 0.2% offset
strain (MPa)
ultimate
strength,
ultimate
strain
(%)
Fb
b
sample
(g cmꢀ3
)
(sꢀ1
)
(m sꢀ1
)
UCS (MPa)
Poisson’s ratio
(J gꢀ1
)
(J cmꢀ3
)
bis-NAD-5
bis-NAD-10
bis-NAD-15
bis-NAD-20
0.240
0.390
0.528
0.625
0.035
0.035
0.035
0.035
NA
NA
0.36 ( 0.02
45.0 ( 21.6 85.1 ( 2.1 0.267 ( 0.037 16.7 ( 3.9
4
20
27
9
48 ( 8
350.8
572.4
678.7
2.25 ( 0.12 168.4 ( 18.6 88.1 ( 1.6 0.269 ( 0.041 50.2 ( 2.4
6.05 ( 0.17 127.4 ( 14.1 79.6 ( 3.1 0.286 ( 0.006 50.1 ( 2.0
173 ( 13
288 ( 0.5
11.2 ( 0.079
27.7 ( 0.8
40.6 ( 6.8 0.299 ( 0.008 14.7 ( 1.6
aAverage of two samples. b Per unit volume energy absorption, calculated from the energy absorption per unit mass multiplied by bulk density.
2257
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
thickness (refer to SEM) prevents pore walls from buckling
during compression of the sample. As a result, hardening at
strains prior to 70% are attributed to nanobending deformations.
Macroscopically, in no case do the samples buckle during
compression, and all Poisson’s ratios are in the range of
0.27ꢀ0.30, reflecting little lateral expansion, unless during the
late stages of the test, when pores have been substantially closed
and samples start to expand radially. Ultimately, lower-density
samples (e0.5 g cmꢀ3) undergo compressive failure at >80%
strain, but the most dense samples (bis-NAD-20, Fb = 0.6
g cmꢀ3) fail catastrophically by fragmentation at much lower
strains (∼40%, see Figure 7). The ultimate compressive strains
roughly follow the corresponding porosities of the samples
(compare Tables 2 and 4). The Young’s modulus (E, calculated
from the slope of the early linearly elastic range), the speed of
sound (calculated from the Young’s modulus and the bulk
density via (E/Fb)0.5), and the yield stress at 0.2% offset strain
all increase as the bulk density increases. Specifically, the Young’s
modulus follows a power law relationship with bulk density
(Figure 8A) of the type E ≈ (Fb)3.35. The sensitivity (exponent)
is higher than that observed with native silica aerogels (∼3.0),41
cross-linked silica aerogels (3.10),42 cross-linked vanadia aero-
gels (1.87),41c and polyurea organic aerogels,43 signifying the
vastly different nature of the interparticle bridging: in bis-NAD-
xx, the neck zones are purely polymeric, whereas in polymer-
cross-linked aerogels, they are mixed organicꢀinorganic. On the
other hand, the ultimate strength, as well as the ability for the
material to store energy (referenced as toughness and quantified by
the integral of the stressꢀstrain curve) vary nonmonotonically with
density: as shown in Figures 8B and 8C, they both increase with
density in the beginning, then they reach a maximum, and,
afterward, they decline. That decline in strength and toughness
coincides with both the change in the failure mode (see photo-
graphs in Figure 7) and the decline in the ultimate strain at failure
(see data in Table 4).
A monotonic variation of the Young’s modulus with density
and a simultaneous nonmonotonic variation of the ultimate
strength and toughness (see Figure 8) also have been observed
with polymer cross-linked silica aerogels, and that behavior is
independent of the cross-linking polymer.44,45 In those materials,
the reinforcing polymeric tethers are placed on a preformed
inorganic framework, and while all accumulated polymer con-
tributes to stiffness,46 only the bridging tethers between the
nanoparticles contribute to strength and toughness.44,45 In
agreement with conclusions reached with silica aerogels,47
changes in the fractal dimension (and, therefore, the connectivity
within secondary particles) should not be relevant to the decline
of the strength and toughness as the density increases. Indeed,
the higher connectivity within the secondary particles of bis-
NAD-10 and bis-NAD-20, as indicated by their fractal dimen-
sion (Table 3), is not associated with an identifiable trend in their
mechanical properties (see Table 4). Therefore, the trends in
Figure 8 should be traceable to the intersecondary particle
connectivity. Based on the microscopic characterization data, it
was concluded that the growth mechanism of medium-density
bis-NAD-xx samples is reminiscent of the cross-linking process
of silica aerogels, in that gelation is followed by a monomer-
cluster growth process, where particles continue to grow in size
via continual reaction with the remaining monomer. That
reinforces the intersecondary particle necks (in a cross-linked
aerogel fashion), yielding stronger materials. At even higher
monomer concentrations, reactions proceed fast, consuming all
Figure 8. (A) Logꢀlog plot of the Young’s modulus versus bulk density
of various bis-NAD-xx aerogels. (B, C) Variation of the ultimate
compressive strength (panel B) and energy absorption (panel C) of
the same bis-NAD-xx aerogels, each as a function of their bulk density.
(Lines have been added to guide the eye.)
linked silica aerogels for their facile one-step synthesis and for
their mechanical properties. In that regard, ROMP-derived
polyimides were investigated under quasi-static compression.
The stressꢀstrain curves (Figure 7) show very short, almost-
elastic ranges up to ∼3% strain, followed by plastic deformation
and hardening up to 70% strain, because of pore collapse. The fact
that the early part of the stressꢀstrain curves is almost elastic was
confirmed by conducting loading and unloading tests (Figure 7,
inset); it was found that the unloading curve after loading to 3%
strain almost follows the loading curve with 0.2% remaining
strain. (By comparison, loading up to 5.5% strain results in 2%
unrecovered strain, and loading up to 8% strain gives 4%
unrecovered strain.) Interestingly, after reaching the 0.2% offset
yield stress, which is a conventional measure of the incipient of
plastic deformation, the stress continues to increase with strain.
This phenomenon is different from plastic foams40 in which, after
reaching the yield strength, the stressꢀstrain curve shows a
plateau associated with the collapse of pores, because of cell-wall
buckling. It is likely that the small ratio of pore size to wall
2258
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
Table 5. Thermal Conductivity Data for Selected bis-NAD-xx Samples a
b
b
b
sample
bulk density, Fb (g cmꢀ3
)
heat capacity, cp (J gꢀ1 Kꢀ1
)
thermal diffusivity, R (mm2 sꢀ1
)
thermal conductivity, λ (W mꢀ1 Kꢀ1
)
bis-NAD-10
bis-NAD-15
bis-NAD-20
a Average of three samples. b At 23 ꢀC.
0.338 ( 0.003
0.568 ( 0.003
0.622 ( 0.002
0.995 ( 0.030
1.088 ( 0.033
1.062 ( 0.032
0.091 ( 0.005
0.085 ( 0.001
0.096 ( 0.004
0.031 ( 0.001
0.053 ( 0.002
0.063 ( 0.003
of the monomer quickly, and yield smaller particles with
weaker interparticle necks, leading to a decline in ultimate strain
and a concomitant decrease in ultimate strength and energy
absorption.
Overall, at their best (i.e., at the medium density range), bis-
NAD-10 and bis-NAD-15 aerogels compete favorably with (and,
in many aspects, they are better than) polymer cross-linked silica
aerogels. For example, polyurea cross-linked silica at Fb = 0.304
g cmꢀ3, process-optimized by statistical design of experiments
(DoE) methods, are 77% porous with σ = 147 m2 gꢀ1, a Young’s
modulus of 32 MPa, a yield stress at a 0.2% offset strain of 1.12
MPa, and an ultimate strength of 237 MPa.45 By comparison, bis-
NAD-10 samples (Fb = 0.341 g cmꢀ3) are 73% porous, with σ =
438 m2 gꢀ1 (Table 2), a Young’s modulus of 48 MPa, a yield
stress at a 0.2% offset strain of 2.25 MPa, and an ultimate strength of
168 MPa, and they can absorb up to 50 J gꢀ1 of energy (see Table 4).
The latter figure renders them better than strong materials typically
used for ballistic protection, such as 4130 steel (15 J gꢀ1 at 7.84
g cmꢀ3), Kevlar-49 epoxy composites (11 J gꢀ1 at 1.04 g cmꢀ3),
and SiC ceramics (20 J gꢀ1 at 3.02 g cmꢀ3).48 Now, from an
engineering design perspective, a fair comparison with standard
materials should also extend from energy absorption per unit
mass (J gꢀ1) to energy absorption per unit volume (J cmꢀ3).
Using the latter metric, steel and SiC (117.6 J cmꢀ3 and 60.4
J cmꢀ3, respectively) remain superior to bis-NAD-xx aerogels
(27 J cmꢀ3 at their best; see Table 4), but the latter still surpass
Kevlar-49 fiber-epoxy composites (11.4 J cmꢀ3). However, since
fiber-epoxy composites are rapidly replacing steel and ceramics in
armor,49 it is concluded that the additional volume requirement
for absorbing a fixed amount of energy by Kevlar composites is
easily taken into account in practice. Therefore, by and large, bis-
NAD-xx aerogels are reasonable and, in fact, better alternatives.
Thermal Conductivity (λ). This was calculated from the
thermal diffusivity (R) and the heat capacity (cp) of bis-NAD-
xx disks ∼2.0 mm thick, using eq 1. The thermal diffusivity,
Figure 9. Temperature rise curve of the back face of a bis-NAD-15
aerogel disk (9.32 mm in diameter, 2.17 mm thick, Fb = 0.568 g cmꢀ3
coated with gold and carbon on both faces, following a heat pulse
incident to the front face. Dashed reference lines indicate t50, which is the
time required for the detector voltage (proportional to temperature) to
reach half its maximum value. Data have been fitted to the pulse-
corrected Cowan model (see text).
)
Because of the small variation of the thermal diffusivity with
density and the constant value of the heat capacity, thermal
conductivity scales linearly with the bulk density within the limits
investigated and expresses the contribution of the through-lattice
heat transfer. The thermal conductivity of the bis-NAD-10
samples is determined to be equal to 0.031 W mꢀ1 Kꢀ1, which
compares favorably with that of polyurea cross-linked silica
aerogels (0.041 W mꢀ1 Kꢀ1 at 0.451 g cmꢀ3),42 glass wool
(0.040 W mꢀ1
K K
ꢀ1), styrofoam (0.030 W mꢀ1 ꢀ1), and
polyurethane foam (0.026 W mꢀ1 Kꢀ1).53
4. CONCLUSIONS
Bis-NAD-xx aerogels, considered together with other organic
aerogels from the recent literature,8ꢀ17 exemplify the design
parameters for the bottom-up synthesis of polymeric gels that
can be dried to aerogels. The key requirement seems to be phase
separation of surface-reactive nanoparticles that can cross-link
with each other into a three-dimensional network. Phase separa-
tion is induced by reduced solubility of the growing polymer,
which, inturn, is introducedby cross-linkingatthe molecular level.
Significant shrinking (in the present case, up to 40%) seems to
be encountered more frequently with organic aerogels rather
than their inorganic counterparts. Although that is typically a
problem with the latter, leading to cracking, the more-flexible
organic framework seems to accommodate stresses better, and
the materials come out as perfect monoliths. More importantly
though, desirable properties such as high surface area, porosity,
and pore structure do not seem to be affected detrimentally; in
that regard, shrinkage may be difficult to predict, but most
certainly is reproducible and therefore can be engineered into
the final object. As demonstrated herewith, ROMP-derived
λ ¼ Fb ꢁ cp ꢁ R
ð1Þ
was measured using a flash method (see the Experimental
Section),50 where the sample is heated from one side and the
temperature rise is observed as a function of time on the other
side. Coating the samples on both sides with gold and carbon
ensures absorption of the heat pulse and minimizes radiative
pathways and pulse “bleed through.”51 Typical data are shown in
Figure 9. The data analysis software employs the pulse-corrected
Cowan model52 to approximate the heat-transfer equation, using
an initial value for the thermal diffusivity estimated by the time it
takes for the detector voltage to reach its half-maximum value
(marked as t50 in Figure 9). Subsequently, a least-squares fit is
iteratively performed in a defined time range (10 ꢁ t50), and the
value for thermal diffusivity, R, is obtained. (A value of 10 times
t50 has been determined to be a suitable measure of the initial
cooling event after the heat pulse.) Table 5 summarizes the data.
2259
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
polyimide aerogels can be prepared in one step as mesoporous
materials over a wide density range with high porosities, high
surface areas, high modulus, high strength, and high toughness.
Combining the one-step synthesis with mechanical strength,
manageable thermal stability, relatively low thermal conductivity,
and low speed of sound wave propagation render bis-NAD-xx to
be reasonable multifunctional candidates for further investigation
into thermal and acoustic insulation at elevated temperatures.
From a theoretical perspective, bis-NAD-xx emphasize the fact
that nucleation and network growth in organic aerogels is a
complicated process that may not be known a priori, but it has
definite effects on the performance of the materials. It can be
influenced by typical reaction conditions, such as solvent,
temperature, monomer, and catalyst concentration; moreover,
predictability most certainly can be gained through multivariable
optimization studies.
(13) Lee, J. K.; Gould, G. L.; Rhine, W. L. J. SolꢀGel Sci. Technol. 2009,
49, 209–220.
(14) Lorjai, P.; Chaisuwan, T.; Wongkasemjit, S. J. SolꢀGel Sci. Technol.
2009, 52, 56–64.
(15) Lee, J. K.; Gould, G. L. J. SolꢀGel Sci. Technol. 2007, 44,
29–40.
(16) Rhine, W.; Wang, J.; Begag, R. Polyimide Aerogels, Carbon Aerogels,
and Metal Carbide Aerogels and Methods of Making Same, U.S. Patent No.
7,074,880, 2006.
(17) (a) Kawagishi, K.; Saito, H.; Furukawa, H.; Horie, K. Macromol.
Rapid Commun. 2007, 28, 96–100. (b) Meador, M. A. B.; Malow, E. J.; He,
Z. J.; McCorkle, L.; Guo, H.; Nauyen, B. N. Polym. Prepr. (Am. Chem. Soc.,
Div. Polym. Chem.) 2010, 51, 265–266.
(18) Daniel, C.; Giudiee, S.; Guerra, G. Chem. Mater. 2009, 21,
1028–1034.
(19) Goueree, P.; Talbi, H.; Miousse, D.; Tran-van, F.; Dao, L. H.; Lee,
K. H. J. Electrochem. Soc. 2001, 148, A94–A101.
(20) (a) Fischer, F.; Rigacci, A.; Pirard, R.; Berthon-Fabry, S.;
Achard, P. Polymer 2006, 47, 7636–7645. (b) Gavillon, R.; Budtova,
T. Biomacromolecules 2008, 9, 269–277. (c) Surapolchai, W.; Schiraldi,
D. A. Polym. Bull. 2010, 65, 951–960.
’ AUTHOR INFORMATION
Corresponding Authors:
(21) (a) Leventis, N.; Sotiriou-Leventis, C.; Zhang, G.; Rawashdeh,
A.-M. M. Nano Lett. 2002, 2, 957–960. (b) Zhang, G.; Dass, A.; Rawashdeh,
A.-M. M.; Thomas, J.; Counsil, J. A.; Sotiriou-Leventis, C.; Fabrizio, E. F.;
Ilhan, F.; Vassilaras, P.; Scheiman, D. A.; McCorkle, L.; Palczer, A.; Johnston,
J. C.; Meador, M. A. B.; Leventis, N. J. Non-Cryst. Solids 2004, 350, 152–164.
(c) Leventis, N.; Palczer, A.; McCorkle, L.; Zhang, G.; Sotiriou-Leventis, C.
J. SolꢀGel Sci. Technol. 2005, 35, 99–105. (d) Leventis, N. Acc. Chem. Res.
2007, 40, 874–884. (e) Leventis, N.; Vassilaras, P.; Fabrizio, E. F.; Dass, A.
J. Mater. Chem. 2007, 17, 1502–1508. (f) Leventis, N.; Mulik, S.; Sotiriou-
Leventis Chem. Mater. 2008, 20, 6985–6997.
(22) Sroog, C. E.; Endrey, A. L.; Abrmo, S. V.; Berr, C. E.; Edward,
W. M.; Oliver, K. L. J. Polym. Sci., Part A 1965, 3, 1373–1390.
(23) (a) Edward, W. M.; Robinson, I. M. Polyimides of Pyromellitic Acid,
U.S. Patent No. 2,710,853, 1955. (b) Edwards, W. M.; Robinson, I. M.
Preparation of Pyromellitimides, U.S. Patent No. 2,867,609, 1959.
(24) (a) Woodfine, B.; Soutar, I.; Preston, P. N.; Jigajinni, V. B.; Stewart,
N. J.; Hay, J. N. Macromolecules 1993, 26, 6330–6334. (b) Baugher, A. H.;
Espe, M. P.; Goetz, J. M.; Schaefer, J.; Pater, R. H. Macromolecules 1997,
30, 6295–6301. (c) Hu, A. J.; Hao, J. Y.; He, T.; Yang, S. Y. Macromolecules
1999, 32, 8046–8051. (d) Xie, W.; Pan, W.-P.; Chuang, K. C. Thermochim.
Acta 2001, 367ꢀ368, 143–153.
*E-mail address: (N.L.) leventis@mst.edu; (C.S.-L.) cslevent@
mst.edu; (J.T.M.) jtmang@lanl.gov; (H.L.) hongbing.lu@
utdallas.edu.
’ ACKNOWLEDGMENT
We thank the following for their financial support: the Army
Research Office, under Award No. W911NF-10-1-0476 (N.L., C.
S.-L.); and the National Science Foundation, under Agreement
Nos. CHE-0809562 (N.L., C.S.-L.), DMR-0907291 (N.L., H.L.),
CMMI-0653970 (N.L., C.S.-L.), and CMMI-0653919 (H.L.).
We also acknowledge the Materials Research Center of Missouri
S&T for its support in sample characterization (SEM, XRD).
Solids NMR work was conducted at the University of Missouri
Columbia by Dr. Wei Wycoff. This work also benefited from the
use of the SANS instrument, LQD at the Manuel Lujan, Jr.
Neutron Scattering Center of the Los Alamos National Labora-
tory, supported by the DOE Office of Basic Energy Sciences; this
work also utilized facilities supported in part by the National
Science Foundation, under Agreement No. DMR-0454672.
(25) Sroog, C. E. Prog. Polym. Sci. 1991, 16, 561–694.
(26) Chidambareswarapattar, C.; Larimore, Z.; Sotiriou-Leventis, C.;
Mang, J. T.; Leventis, N. J. Mater. Chem. 2010, 20, 9666–9678.
(27) Grubbs, R. H. Angew. Chem., Int. Ed. 2006, 45, 3760–3765.
(28) (a) Laguitton, B.; Mison, P.; Sillion, B.; Brisson, J. Macromolecules
1998, 31, 7203–7207. (b) Chen, C.-F.; Qin, W.-M.; Huang, X.-A. Polym. Eng.
Sci. 2008, 48, 1151–1156.
(29) Seeger, P. A.;Hjelm, R. P., Jr.J. Appl. Crystallogr. 1991, 24, 467–478.
(30) Liu, Y.; Sun, X. D.; Xie, X.-Q.; Scola, D. A. J. Polym. Sci., Part A:
Polym. Chem. 1998, 36, 2653–2665.
(31) The polymerization of bis-NAD via ROMP is expected to be a
living process that proceeds until all of the monomer is consumed. For
example, see: Bielawski, C. W.; Grubbs, R. H. Prog. Polym. Sci. 2007,
32, 1–29.
(32) (a) Everett, D. H. Basic Principles of Colloid Science; The Royal
Society of Chemistry: London, U.K., 1988; p 189. (b) Ilmain, F.; Tanaka,
T.; Kokufuta, E. Nature 1991, 349, 400–401.
(33) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti,
R. A.; Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603–619.
(34) Lowell, S.; Shields, J. E.; Thomas, M. A.; Thommes, M. Character-
ization of Porous Solids and Powders: Surface Area, Pore Size and Density;
Kluwer Academic Publishers: Norwell, MA, 2004; pp 44ꢀ45.
(35) Webb, P. A.; Orr, C. Analytical Methods in Fine Particle Technology;
Micromeritics Instrument Corporation: Norcross, GA, 1997; pp 67ꢀ68.
(36) (a) Beaucage, G. J. Appl. Crystallogr. 1995, 28, 717–728.
(b) Beaucage, G. J. Appl. Crystallogr. 1996, 29, 134–146. (c) Mang,
’ REFERENCES
(1) Pierre, A. C.; Pajonk, G. M. Chem. Rev. 2002, 102, 4243–4265.
(2) Morris, C. A.; Anderson, M. L.; Stroud, R. M.; Merzbacher, C. I.;
Rolison, D. R. Science 1999, 284, 622–624.
(3) Kistler, S. S. Nature 1931, 127, 741.
(4) Kistler, S. S. J. Phys. Chem. 1932, 36, 52–63.
(5) Gash, A. E.; Pantoya, M.; Satcher, J. H.; Simpson, R. L. Polym.
Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 2008, 49, 558–559.
(6) Leventis, N.; Chandrasekaran, N.; Sadekar, A. G.; Mulik, S.;
Sotiriou-Leventis, C. J. Mater. Chem. 2010, 20, 7456–7471.
(7) Baumann, T. F.; Gash, A. E.; Chinn, S. C.; Sawvel, A. M.; Maxwell,
R. S.; Satcher, J. H., Jr. Chem. Mater. 2005, 17, 395–401.
(8) Pekala, R. W. Low Density ResorsinolꢀFormaldehyde Aerogels, U.S.
Patent No. 4,873,218, 1989.
(9) Pekala, R. W.; Alviso, C. T.; Lu, X.; Gross, J.; Frickle, J. J. Non-Cryst.
Solids 1995, 188, 34–40.
(10) Li, W.-C.; Lu, A.-H.; Guo, S.-C. J. Colloidal Interface Sci. 2002,
254, 153–157.
(11) Pekala, R. W. MelamineꢀFormaldehyde Aerogels, U.S. Patent No.
5,086,085, 1992.
(12) Biesmans, G.; Martens, A.; Duffours, L.; Woignier, T.; Phalippou, J.
J. Non-Cryst. Solids 1998, 225, 64–68.
2260
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261

Chemistry of Materials
ARTICLE
J. T.; Son, S. F.; Hjelm, R. P.; Peterson, P. D.; Jorgensen, B. S. J. Mater.
Res. 2007, 22, 1907–1920. (d) Tappan, B. C.; Huynh, M. H.; Hiskey,
M. A.; Chavez, D. E.; Luther, E. P.; Mang, J. T.; Son, S. F. J. Am. Chem.
Soc. 2006, 128, 6589–6594.
(37) Orcel, G.; Gould, R. W.; Hench, L. L. In Better Ceramics Through
Chemistry II; Brinker, C. J., Clark, D. E., Ulrich, D. R., Eds.; Materials
Research Society: Pittsburgh, PA, 1986; Vol. 73, p 289.
(38) Hench, L. L.; West, J. K. Chem. Rev. 1990, 90, 33–72.
(39) (a) Zallen, R. The Physics of Amorphous Solids; Wiley: New York,
1983; Chapter 4. (b) Stauffer, D.; Conialio, A.; Adam, M. Adu. Polym. Sci.
1982, 44, 103.
(40) Daphalapurkar, N. P.; Hanan, J. C.; Phelps, N. B.; Bale, H.; Lu,
H. Mech. Adv. Mater. Struct. 2008, 15, 594–611.
(41) (a) Fricke, J. J. Non-Cryst. Solids 1988, 100, 169–173. (b) Gross, J.;
Fricke, J. Nanostructured Mater. 1995, 6, 905–908. (c) Luo, H.; Churu, G.;
Schnobrich, J.; Hobbs, A.; Fabrizio, E. F.; Dass, A.; Mulik, S.; Sotiriou-
Leventis, C.; Lu, H.; Leventis, N. J. SolꢀGel Sci. Technol. 2008, 48, 113–134.
(42) Katti, A.; Shimpi, N.; Roy, S.; Lu, H.; Fabrizio, E. F.; Dass, A.;
Capadona, L. A.; Leventis, N. Chem. Mater. 2006, 18, 285–296.
(43) Leventis, N.; Sotiriou-Leventis, C.; Chandrasekaran, N.; Mulik, S.;
Larimore, Z. J.; Lu, H.; Churu, G.; Mang, J. T. Chem. Mater. 2010, 22, 6692–
6710.
(44) Meador, M. A. B.; Fabrizio, E. F.; Ilhan, F.; Dass, A.; Zhang, G.;
Vassilaras, P.; Johnston, J. C.; Leventis, N. Chem. Mater. 2005, 17, 1085–
1098.
(45) Meador, M. A. B.; Capadona, L. A.; MacCorkle, L.; Papadopoulos,
D. S.; Leventis, N. Chem. Mater. 2007, 19, 2247–2260.
(46) Mulik, S.; Sotiriou-Leventis, C.; Churu, G.; Lu, H.; Leventis,
N. Chem. Mater. 2008, 20, 5035–5046.
(47) Woignier, T.; Reynes, J.; Hafili-Alaoui, A.; Beurroies, I.; Phalippou,
J. J. Non-Cryst. Solids 1998, 241, 45–52.
(48) (a) American Society for Metals. ASM Engineering Materials
Handbook, Composites, Vol. 1; ASM International: Materials Park, OH,
1998; p 178, Table 2. (b) Luo, H.; Chen, W. Intern. J. Appl. Ceram. Technol.
2004, 1, 254–260. (c) Luo, H.; Chen, W.; Rajendran, A. M. J. Am. Ceram. Soc.
2006, 89, 266–273.
(49) Hogg, P. J. Science 2006, 314, 1100–1101.
(50) Parker, W. J.; Jenkins, J. J.; Abbott, G. L.; Butler, C. P. J. Appl. Phys.
1961, 32, 1679–1684.
(51) Lee, D.; Stevens, P. C.; Zeng, S. Q.; Hunt, A. J. J. Non-Cryst. Solids
1995, 186, 285–290.
(52) (a) Cowan, R. J. Appl. Phys. 1961, 32, 1363–1369. (b) Cowan,
R. J. Appl. Phys. 1963, 34, 926–927.
(53) Lide, D. R. CRC Handbook of Chemistry and Physics, 84th ed.;
CRC Press: Boca Raton, FL, 2003; pp 12/22612/227.
2261
dx.doi.org/10.1021/cm200323e |Chem. Mater. 2011, 23, 2250–2261