Characterization of a carbon-nitrogen network solid with NMR and high field EPR

Article abstract of DOI:10.1021/jp0004157

Considerable attention has been focused on developing a synthetic route to a carbon-nitrogen material with mechanical and thermal properties comparable or superior to those of diamond. To date, no substance with the desired C₃N₄ stoichiometry in a silicon-nitride crystal lattice has been reported. One of the principal difficulties in the pursuit of ultrahard carbon-nitrogen (CN) solids is the characterization of amorphous CN samples. We describe a solid-state NMR study of a paracyanogen-like solid utilizing ¹³C-¹⁵N adiabatic-passage Hartmann-Hahn cross-polarization (APHH-CP) to perform dipolar filtering and show that this method is well-suited for recoupling ¹³C-¹⁵N in network solids. In addition, high-frequency electron paramagnetic resonance (EPR) indicates a density of electron spins of approximately 1 × 10¹⁷ e^-/cm³. We conclude by discussing how NMR and EPR data may be useful for optimizing CN-polymer samples as potential precursors for ultrahard carbon nitrogen solids.

Full text of DOI:10.1021/jp0004157

J. Phys. Chem. B 2000, 104, 9817-9822
9817
Characterization of a Carbon-Nitrogen Network Solid with NMR and High Field EPR
†
†
†
†
‡
David Rovnyak, Marc Baldus, Boris A. Itin, Marina Bennati, Andrew Stevens, and
Robert G. Griffin*,†
MIT/HarVard Center for Magnetic Resonance and Department of Chemistry,
Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, and Department of Chemistry,
HarVard UniVersity, Cambridge, Massachusetts 02139
ReceiVed: February 2, 2000; In Final Form: July 11, 2000
Considerable attention has been focused on developing a synthetic route to a carbon-nitrogen material with
mechanical and thermal properties comparable or superior to those of diamond. To date, no substance with
3 4
the desired C N stoichiometry in a silicon-nitride crystal lattice has been reported. One of the principal
difficulties in the pursuit of ultrahard carbon-nitrogen (CN) solids is the characterization of amorphous CN
13
15
samples. We describe a solid-state NMR study of a paracyanogen-like solid utilizing C- N adiabatic-passage
Hartmann-Hahn cross-polarization (APHH-CP) to perform dipolar filtering and show that this method is
well-suited for recoupling 13C- N in network solids. In addition, high-frequency electron paramagnetic
15
1
7
-
3
resonance (EPR) indicates a density of electron spins of approximately 1 × 10 e /cm . We conclude by
discussing how NMR and EPR data may be useful for optimizing CN-polymer samples as potential precursors
for ultrahard carbon nitrogen solids.
Introduction
in network solids, where sensitivity is at a premium, without
resorting to conventional 2D chemical shift correlation experi-
Modern, high-performance materials are required for applica-
tions in computing (heat conduction and insulation), industrial
processes (hardness and machinability), and medicine (bio-
implants and surgical tools). Examples of synthetic materials
that have emerged to address these markets include phos-
phazenes, synthetic diamonds, boron nitride, and many plastics.
A recent approach for the development of new materials is the
use of theoretical models to predict, from first principles only,
properties of hypothetical solids. It has been proposed that a
crystalline solid with C3N4 stoichiometry in a â-Si3N4 crystal
lattice would have a bulk modulus comparable or greater to
ments.
It is of additional interest that network solids can possess
significant concentrations of paramagnetic impurities.1 Spec-
troscopically, paramagnetic defects may be useful for improving
spectral density, for electron paramagnetic resonance (EPR)
studies, and as paramagnetic centers to enhance NMR spectra
5,16
15,17,18
through dynamic nuclear polarization.
EPR studies at high
frequencies (>100 GHz) are attractive since sample sizes are
very small and isotopic labeling is not required. We have
therefore also acquired 139.5 GHz electron paramagnetic
resonance (EPR) spectra for additional insight into paracyano-
gen-like materials.
that of diamond, have a high thermal conductivity, and be
optically transparent.1-3 Synthetic routes aimed at such carbon-
nitride samples commonly produce amorphous powders or films
Experimental Section
4-10
with few or no crystalline regions.
The analysis of amor-
1
3
15
13 15
phous powders is difficult since many techniques such as
Rutherford backscattering (RBS), X-ray photoelectron spec-
troscopy (XPS), and electron energy loss spectroscopy (EELS)
probe primarily sample surfaces. While quantitating atom ratios
and revealing hybridization, these methods are unable to describe
complete local structural motifs of the bulk material.
Sample Preparation. Samples of K C N and Na C N
were obtained from Cambridge Isotope Laboratories (Andover,
MA 01810) and used without further purification. Preparation
of Hg(CN) : HCN was generated by addition of NaCN or KCN
in aqueous solution to sulfuric acid; the HCN was transferred
through a condenser tube into a reaction vessel containing an
2
Solid-state nuclear magnetic resonance (SSNMR) is used in
a diverse range of disciplines to characterize local structure in
excess of mercuric oxide, HgO, suspended in water. Hg(CN)2
is very water-soluble and hence the excess HgO is removed by
filtration; subsequent evaporation with heat yields a white
11,12
crystalline and amorphous solids.
Here, we utilize Bloch-
decay and cross-polarization experiments, together with the
powder. Hg(CN) is both heat- and light-sensitive. Preparation
2
adiabatic passage Hartmann-Hahn cross-polarization (APHH-
of paracyanogen-like solid: The Hg(CN) was annealed under
2
CP)1
3,14
method to transfer polarization between C and
13
15
N
19
vacuum and at high temperature; for details see Stevens et al.
nuclei in a paracyanogen-like material. The dipolar filtering of
APHH-CP is able to improve resolution, facilitate assignments,
and provide connectivity information. It is a principle aim of
this study to demonstrate that such information can be obtained
Samples were baked at 100 °C for approximately 12 h before
they were packed into rotors. Caution: Consult your depart-
mental and institutional safety officers before working with
HCN.
Nuclear Magnetic Resonance. NMR experiments were
performed at room temperature on custom-designed spectrom-
*
To whom correspondence should be addressed.
MIT/Harvard Center for Magnetic Resonance and Department of
†
1
eters operating at 4.8 T (200 MHz for H) and 7.4 T (317 MHz
Chemistry.
‡
1
Department of Chemistry, Harvard University.
for H). Three-channel, custom-designed probes incorporated
1
0.1021/jp0004157 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/29/2000

9818 J. Phys. Chem. B, Vol. 104, No. 42, 2000
Rovnyak et al.
Figure 1. Diagram of the adiabatic passage Hartmann-Hahn cross-
polarization pulse sequence, shown here for polarization transfer first
1
13
from H to C via conventional cross-polarization followed by APHH-
CP from 13C to N. An even offset profile of at least 10 kHz is obtained
when ∆ ) 2-3 kHz.
15
Figure 2. (a) Chemical structures proposed for paracyanogen that
5
and 4 mm Chemagnetics spinner assemblies and implemented
possess a high degree of order and extended conjugation; (b) amorphous
transmission line principles to deliver radio frequency power
to the coil. The APHH-CP experiment is diagrammed in Figure
2
23,37
chains of sp carbon and nitrogen with pendant nitriles proposed
to
explain the low order and electrically insulating nature of paracyanogen.
1
3,14
1.
The value of ∆ in Figure 1 was set in the 2-4 kHz range
to provide an even polarization transfer profile of at least (5
kHz.1
3,14
Typical RF fields for cross-polarization and APHH-
CP ranged from 30 to 50 kHz depending upon the matching
conditions. An 15N-acetyl-valine (NAV) sample was used to
set the appropriate RF matching condition for the APHH-CP.
Proton decoupling fields of 100 kHz or greater were employed.
Magic-angle spinning rates were controlled to within a few hertz,
and typical rates used in this study were in the 6-8 kHz range.
1
3
All C chemical shifts are referenced externally to glycine
1
5
such that the carbonyl line is defined to be 177 ppm. All
N
chemical shifts are referenced externally to NAV such that the
1
5
nitrogen shift in NH3 (l) is defined to be 0 ppm.
Electron Paramagnetic Resonance. EPR data were recorded
at room and low temperature (10 K) with a custom-designed
2
0
pulsed EPR spectrometer operating at 139.5 GHz (5 T).
Results
For our study of NMR in network-bonded CN solids, we
selected a material that is prepared by thermal annealing of
mercuric(II) cyanide Hg(CN)2.21 This synthesis was first
performed by Guy-Lussac, and the resulting solid is believed
to be an amorphous carbon-nitrogen polymer, often denoted
paracyanogen (pCN). Paracyanogen has been explored as a
Figure 3. Echo-detected 13C cross-polarization spectra of pCN( C)
13
13
with proton decoupling at (A) 4.8 T (ω
o
/2π[ C] ) 50 MHz) and (B)
/2π[ C] ) 80 MHz). Asterisks denote spinning sidebands.
Recycle times were 4 and 8 s (4.8 T, 7.4 T). The spectra are referenced
to an external glycine sample.
22
precursor for metallic CN polymers and ultrahard carbon-
13
7
.4 T (ω
0
19,23
nitride.
A convenient isotopic enrichment method exists for
paracyanogen through the synthesis and subsequent annealing
13
15
of C, N-mercuric cyanide (see Experimental Section for
details). Moreover, the structure of paracyanogen has not been
(0-70 ppm). The lines at 163 and 157 ppm correspond to sp2
carbon, while the broad 120 ppm resonance is assigned to nitrile
groups. The C chemical shift in nitriles is very sensitive to
demonstrated in the literature, although recent studies have ruled
out several proposed models,2
2,23
shown in Figure 2a, and
13
explored the high-temperature and high-pressure phase dia-
local structure, and hence this feature experiences considerable
amorphous broadening. Increasingly longer recycle delays were
required at 7.4 and 9.4 T (15-20 s recycle times; data not
shown), suggesting a sharply declining spectral density. We were
unable to obtain useful spectra at 14.1 T and attribute this to
low spectral density. The spectral resolution is only slightly
improved at 7.4 T, indicating at least partial amorphous character
of this solid. Asymmetric doublets for spin-1/2 nuclei coupled
19
gram of pCN. Structures such as those in Figure 2a are highly
ordered, and their extended conjugation would lead to metallic
2
2
character. More recently, the structure of Figure 2b, which is
2
composed of chains of sp carbon and nitrogen with pendant
nitriles, was suggested to be more representative of the physical
1
9,23
properties of pCN.
Two samples were made for this study,
13
13 15
13
denoted pCN( C) and pCN( C, N). pCN( C) is a fine, black
powder and was generated by high-temperature, vacuum an-
13
to quadrupolar nuclei are most readily observed for
C
1
3
24
nealing of 25%- C, Hg(CN)2. The second sample, pCN-
resonances in nitrile groups; however, the nitrile region is too
1
3
15
15
13
13
(
C, N), was generated by annealing of 100%- N, 25%- C,
Hg(CN)2.
Nuclear Magnetic Resonance. In Figure 3 we show 13C
echo-detected CP-MAS spectra of pCN( C) obtained at 4.8 and
.4 T. Figure 3 exhibits lines at 163 and 157 ppm as well as a
broad to observe this effect in our spectra. The C line widths
1
3
13 15
of pCN( C) and pCN( C, N) are also not measurably different.
This is reasonable considering that asymmetric splitting was
13
29
25
not observed in Si spectra of â-Si3N4.
1
7
It is remarkable that the H cross-polarization experiment can
be applied as this sample is expected to be aprotic. To better
3
broad feature at 120 ppm. There are no lines in the sp region

NMR and EPR of Carbon-Nitrogen Network Solid
J. Phys. Chem. B, Vol. 104, No. 42, 2000 9819
Figure 4. Echo-detected 13C MAS NMR spectra of pCN( C) via
direct-excitation (dashed trace) and cross-polarization (solid trace).
Spectra have been normalized to the broad feature at 120 ppm to
emphasize the relative excitation efficiencies.
13
Figure 6. 13C-detected APHH-CP at 4.8 T, as per the pulse sequence
1
13
of Figure 1 modified for initial CP from H to C followed by APHH-
CP from 13C to N. The mechanism for coherence transfer is determined
by the parameter n in the APHH-CP matching condition, where a value
of 0 permits both proton driven spin diffusion and J-coupled transfer,
and n ) (1 allows only for dipolar coherence transfer.
15
or direct excitation. From 15N NMR spectroscopy in silicon
2
5,28
nitrides and other nitrogen ceramics,
this sensitivity of the
1
5
15
N shifts is expected. From an N CP rise-time experiment, it
1
5
was found that the upfield portion of the N spectrum (130
ppm, 101 ppm) reaches its maximum intensity at about 150 µs,
but the downfield portion (252 ppm, 191 ppm) continues to
increase in signal after 5 ms. Therefore, the upfield portion of
the spectrum corresponds to nitrogen nuclei with directly bonded
protons, whereas the downfield portion is coupled very weakly
to the proton bath. This is illustrated by the shaded regions in
Figure 5. From Figure 5, it is possible to identify regions
corresponding to imine (191 ppm), secondary amine (130 ppm,
101 ppm), and nitrile (252 ppm) groups.
Figure 5. Echo-detected 15N cross-polarization MAS spectra with
proton decoupling of pCN( C, N) at 4.8 T; 1440 transients were
collected with a recycle time of 5 s. MAS frequency was 7.9 kHz.
1
3
15
Dipolar recoupled spectra employing the APHH-CP sequence
understand this behavior, we compare the 13C echo-detected
are shown for N to C transfer and C to N transfer in
Figures 6 and 7, respectively. The pulse sequence of Figure 1
was employed, in which, in each case, initial magnetization was
prepared with cross-polarization from protons. The APHH-CP
experiments are shown for n ) 1 and n ) 0 matching conditions
for the middle and top spectra, respectively. The signal-to-noise
15
13
13
15
1
3
spectra of pCN( C) obtained via cross-polarization and one-
pulse excitation, as shown in Figure 4. The spectra are
normalized arbitrarily to the broad 120 ppm feature to facilitate
their comparison. The differing excitation profiles suggest that
the line at 163 ppm is more strongly coupled to the proton bath
than the 157 ppm line due to its enhanced relative intensity in
the CP spectra. We also performed a CP build-up experiment
in which the intensity of the 13C CP-MAS signal is monitored
as a function of the contact time. The build-up of the CP signal
intensity corresponded to a time constant of Tr ) 0.17 ms,
indicating that the observed resonances do not correspond to
2
in the spectra ranges from 10 to 10 , each spectrum requiring
approximately 12 h of acquisition time. The coherence transfer
observed in the n ) 0 cases probably occurs via proton-proton
spin flips and not via the J-coupling. For example, signal is
observed in Figure 7 (n ) 0) for the secondary amines, for which
1
3
15
one bond C- N J-couplings are of the order of 10 Hz, but
1
3
1
26,27
C’s with directly bound H’s.
The proton decoupling
scalar transfer is expected in the nitrile and imino regions (252
1
5
strength was approximately 100 kHz, which is sufficient to
ppm,191 ppm) of the N-detected APHH-CP spectrum since
1
13
13 15
completely remove any H- C dipolar couplings not averaged
these groups will have the largest C- N one-bond J-couplings.
1
3
by MAS in the C spectra.
Mixing times up to 7 ms were employed in an attempt to observe
scalar transfer, but this was unsuccessful. Probably even longer
15
A typical N CP spectrum is shown in Figure 5. Resonances
are observed at 101, 130, and 191 ppm, along with a broad
1
3
contact times are needed as well as a higher enrichment of
C
1
5
15
shoulder at 252 ppm. The N CP-MAS spectrum is clearly
(25% in this study). However the n ) 1 N-detected APHH-
CP spectrum provides crucial information for resolving sites
richer than the 13C spectra obtained either by cross-polarization

9820 J. Phys. Chem. B, Vol. 104, No. 42, 2000
Rovnyak et al.
Figure 8. A 140 GHz EPR spectrum (main) obtained by integration
of the spin-echo intensity over the field sweep at 10 K. A sweep of 0.5
T at room temperature is shown in the inset. The g-anisotropy is
small, and giso ) 2.0032.
-
3
13
13 15
e /cm . EPR spectra for pCN( C) and pCN( C, N) are
identical. Additionally, the observation of spin-echoes at low
temperature indicates that electron-nuclear double-resonance
(ENDOR) experiments, which detect electron-nuclear dipolar
couplings, are possible. It is also noted that dynamic nuclear
polarization (DNP) experiments may be useful to enhance the
NMR spectra and possibly gain insight into which species are
more proximate to paramagnetic impurities; however, this was
beyond the scope of the present study.
Figure 7. 15N-detected APHH-CP at 4.8 T using the pulse sequence
of Figure 1. The matching conditions n are marked on the figure; see
text and Figure 6 caption for more discussion.
and improving assignments. The downfield shoulder of the 15
N
CP spectrum is resolved at 252 ppm in the APHH-CP (n ) 1),
consistent with typical nitrile shifts and with an increased
intensity relative to the 191 ppm resonance that is to be expected
from the larger 13C- N dipolar coupling of the nitrile as
compared to the imine. The relative transfer efficiencies in the
upfield region (130 ppm, 101 ppm) suggest a stronger CN
interaction at 130 ppm than at 101 ppm. Assignments will be
treated in more detail in the discussion.
Discussion
Materials that exhibit covalent, high-dimensional bonding
may be aprotic and amorphous, and these characteristics can
significantly affect the acquisition of NMR spectra. Recycle
times up to an hour or longer may be required
near absence of local motions in the NMR time scale, or
15
2
5,28,29
due to the
1
6,30,31
amorphous broadening
could limit site resolution to the
Electron Paramagnetic Resonance. Direct-excitation 15N
NMR spectra could not be obtained due to the presence of
intense, short-T2 signals that introduced large distortions in the
spectra. This may be attributed to paramagnetic nitrogens, which
are expected in lattice defects when a nitrogen atom occupies
a carbon position or when a polymer segment terminates with
a divalent nitrogen. We have confirmed the existence of
paramagnetic centers with EPR experiments obtained at 139.5
GHz, shown in Figure 8. A single absorption is observed at
room and low temperature (10 K). The low-temperature
spectrum, obtained by integrating the spin-echo intensity over
the magnetic field sweep, gives an absorptive line shape as
shown in Figure 8. The signal in this sample only permitted
continuous wave EPR (CW-EPR) at room temperature, which
results in a derivative spectrum (inset of Figure 8). A field sweep
range of 0.5 T (about the central field of 5 T) allows detection
of paramagnetic centers with g-values of 2.15-1.85.
point that some materials cannot even be described by repeated
structural units.
Thus, the observation of resolved resonances in the 13C and
1
5
N NMR spectra (Figures 3 and 5) indicates that this pCN-
like material is composed of a repeated structural motif.
However, the degree of order is insuffient to provide NMR line
widths normally observed for crystalline phase samples or to
motionally constrain the solid since NMR recycle times of 3-5
s provide good signal-to-noise at 4.7 T. This disparity is
reconciled first by our knowledge of paramagnetic centers and
NH moieties whose motions may provide frequencies on the
NMR time scale. Second, we note that field-dependent studies
show a rapidly declining spectral density at higher static fields,
indicating some degree of extended, three-dimensional bonding
in this solid. An explanation consistent with these findings is
that a repeated, two-dimensional structural unit exists along with
significant covalent cross-linking, which constrains motions and
introduces structural heterogeneity. Given the low concentration
of paramagnetic defects and the use of echo-detection we believe
that we are observing signals of spins in the bulk material and
not spins that are proximate to a minority phase or defect. A
study of the phase properties of an identically prepared material
also gave no evidence of minority phases in thermally annealed
The line shape is only slightly asymmetric, indicating a small
g-anisotropy with an estimated isotropic g-value of 2.0032.
Magnetic field scans over half a Tesla detected no other
paramagnetic signals, suggesting that other impurities are not
present. The EPR signal intensity increases at lower temperature
in a Boltzmann-like fashion as expected for an isolated
paramagnetic center in an electrically insulating material.
Without knowing the molecular unit of this material, it is
difficult to specify the spin density; however, the spin-echo
intensity was compared to standard solutions of different organic
1
9
pCN.
The current data is insufficient to fully constrain the local
structure of this solid; however, we are able to make assignments
of the spectra to chemical moieties, and we discuss their
1
7
13
radicals, indicating an approximate spin count of 1 × 10
significance here. From the C spectra, we rule out the presence

NMR and EPR of Carbon-Nitrogen Network Solid
J. Phys. Chem. B, Vol. 104, No. 42, 2000 9821
3
2
of sp carbon and identify two sp carbon lines and a broad
nitrile resonance. This observation agrees with previous hy-
this mechanism is not available and, for aprotic network solids,
APHH-CP at n ) 0 has considerable potential.
2
3
potheses for the structure of paracyanogen-like solids that
describe chains of sp carbon and nitrogen with cyano groups
hanging off the chains. The absence of sp carbon indicates that
water molecules do not add across double bonds, an effect seen
The problem of observing an NMR signal for a nucleus
dipolar coupled to one or more unpaired electrons has been
2
3
34
discussed by Nayeem et al., who showed that the interaction
is inhomogeneous even in the presence of dipolar coupled,
2
3
in paracyanogen-like solids prepared by reactive sputtering
unpaired electrons, but only if the g-anisotropy is large. Such
and also in protonated silicon nitride powders.29 Other studies
have observed that reactive sputtering leads to dangling bonds
on carbon nuclei, which are reactive centers for water addition
34
an interaction was observed for proton-electron pairs but also
1
3
in C spectra by Pan et al., who observed spinning sidebands
16
over a 2000 ppm range in amorphous carbon. Extended
1
6
13
reactions. In contrast, our sample was prepared by a thermal
annealing process. Although samples were stored in air before
measuring, controlled exposures to moisture were not under-
taken. The addition of water in this fashion would result in
protons dipolar coupled to sp carbon nuclei, which would thus
be easily detected in CPMAS spectra. The sp chemical shifts
agree with imine resonances observed in nitrogenated amor-
phous C:H films.32 It is also significant that the difference in
chemical shift between the two sp carbon species is well-
resolved (5 ppm). A distribution of carbon bound to one, two,
or three nitrogens, or one or more cyano groups, would be
expected to result in a completely amorphous spectrum. This
reinforces our assertion above that resolved lines must indicate
the existence of a regular chemical unit. Finally, a broad signal
is present at about 120 ppm, consistent with literature values
spinning sideband manifolds were not observed in the C or
1
5
N spectra of our compounds, in agreement with our direct
observation of minimal g-anisotropy. The observation of a
single, almost featureless EPR line leaves open the possibility
of free electrons localized on carbons instead of nitrogens, for
which a substantial hyperfine coupling, resulting in typical three
3
2
1
4
15
line ( N) or two line ( N) splittings up to 30 G, is usually
15
expected (see refs 35 and 36 and references therein). The
2
narrow observed EPR line width can be due either to small
hyperfine couplings or to electron exchange (see ref 36, p 199).
1
5
Although the N NMR spectra suggest free electrons localized
on nitrogens, the EPR spectra do not unambiguously support
15
this assignment. N ENDOR experiments are required to resolve
this disparity.
Finally, we evaluate the potential for magnetic resonance data
to provide a set of structural criteria in the bulk material that
may determine the suitability of such samples as possible
3
3
for nitrile groups.
Turning to the nitrogen spectra, a nitrile region at 252 ppm
is resolved by correlating the 15N-CPMAS spectrum with the
APHH-CP (n ) 1) spectrum, as shown in Figure 7. Without
3
precursors for high-dimensional, sp -hybridized, carbon-
nitrogen network solids. The temperature-pressure composition
1
3
15
the C- N correlation to resolve the cyano region, this
assignment would be difficult as the signal at 252 ppm is weak.
diagram of a similarly prepared sample has been previously
1
9
reported in which it was found that kinetic decomposition to
1
5
2
In fact, the weak polarization of the cyano region in the
CPMAS spectrum probably explains the absence of intensity
for the nitrile region in the C-detected APHH-CP (Figure 5,
n ) 1). We assign the 191 ppm 1 N region to imine groups
N
form N2 gas was favored over thermodynamic conversion of sp -
3
to sp -carbon, even at very high temperatures and pressures.
13
Such a material is unsuitable as a precursor for network bonded,
covalent CN solids so long as the decomposition to N2 remains
favored. Both NMR and EPR data suggest means by which a
paracyanogen-like material can be tested for suitability for high-
pressure and high-temperature composition analyses. The line
5
(
CdN-R) based on the large intensity of the peak in the APHH-
CP (n ) 1), which suggests strong CN dipolar coupling, and
the chemical shift that is well outside the nitrile and amino
regions. The imino region may be broadened by heterogeneity
in R, which may be C or N, but is still well-resolved from the
secondary amine groups at 130 and 101 ppm, which are typical
for R*NHR shifts. Here we propose to assign these to C*NHC
and C*NHN moieties respectively because their intensities are
similar in CPMAS spectra but differ by a factor of 2 in the
APHH-CP. This is reasonable considering that the adiabatic-
transfer efficiency should be identical for both sites due to the
even offset profile of APHH-CP and the small difference in
isotropic shifts. A significant observation is that the breadth of
the imine region and the presence of two distinct amine
resonances both suggest N-N bonding, a feature that is absent
from Figure 2b. The presence of N-N bonding has also been
supported by thermal decomposition studies.19
1
5
at 101 ppm in the N spectra indicates N-N bonds that can
condense to N2 during decomposition. Synthetic preparations
can be attempted to minimize this feature. Similarly, elimination
33
1
of cyano groups removes sp carbon from the sample and may
assist in favoring the thermodynamic process over the kinetic
13
15
one. Monitoring the appropriate lines in the C and N spectra
will report on the relative cyano concentrations in the bulk. Of
1
3
3
course, observation of C signals in the sp region would be
valuable. The presence of directly bound protons can be tested
with simple cross-polarization build-up experiments, as in this
study, or with dipolar dephasing or separated local field
spectroscopies (for a recent review, see ref 12, Schmidt-Rohr
and Spiess, pp 205-235).
14
Additionally, preliminary 13C electron-nuclear double-
resonance (ENDOR) experiments have been performed at low
temperature which reveals a complex distribution of dipolar
couplings. Detailed modeling of this distribution may lead to a
The APHH-CP spectra provide dipolar filtering information
that improves resolution and facilitates assignments, as in the
case of the nitrile region which is resolved at 252 ppm in Figure
(n ) 1). However, in the n ) 0 case, the 15N-detected APHH-
1
3
measure of the proximity of C to paramagnetic centers.
7
CP experiment suggests scalar carbon-to-nitrogen transfer for
the secondary amines which is clearly not physical based on
typical JCN values for a sigma bond. Mixing times in excess of
Using NMR and EPR to provide feedback data versus
variations in the synthetic parameters leading to CN-polymers,
it may be possible to develop a paracyanogen-like material with
minimal N-N bonding, reduced hydrogenation, and greater sp3
character, and thus whose temperature/pressure composition
diagram may favor thermodynamic rearrangement over kinetic
decomposition.
7
ms and an increased 13C enrichment level (>25%) are
proposed to overcome the difficulty of observing the scalar
transfer, but the presence of proton-assisted polarization transfer
is a complicating factor in this sample. In the absence of protons,

9822 J. Phys. Chem. B, Vol. 104, No. 42, 2000
Rovnyak et al.
Conclusion
(8) Fernandez, A.; Prieto, P.; Quiros, C.; Sanz, J. M.; Martin, J.-M.;
Vacher, B. Appl. Phys. Lett. 1996, 69(6), 764.
(9) Husein, I. F.; Zhou, Y.; Li, F.; Allen, R. C.; Chan, C.; Kleiman, J.
I. Mater. Sci. Eng. 1996, A209, 10.
We have performed a study of an isotopically enriched,
paracyanogen-like solid, which possesses relatively short re-
laxation times and resolved spectral regions. These spectral
characteristics permitted the use of adiabatic-passage Hart-
mann-Hahn CP between 13C and N to identify and assign
three carbon sites and four nitrogen sites. We find that APHH-
CP is a robust recoupling method for network CN solids,
providing information not readily available in conventional
direct-excitation spectra while avoiding costly 2D experiments.
We also showed the existence of two NH moieties by CP build-
up experiments and note that preparations of pCN-like solids,
whether by RF sputtering or thermal annealing, seem to be
susceptible to hydrogenation mechanisms. Although we are not
able to unambiguously determine the local structure of this pCN-
like solid, we have demonstrated that water molecules do not
add accross CdN double bonds, confirmed the presence of
nitrile groups, provided evidence for N-N bonding, and shown
the existence of paramagnetic centers. In particular, we note
that the good resolution of many spectral regions is difficult to
reconcile for an amorphous solid. Thus, in consideration of a
rapidly declining spectral density, we propose that a repeated,
two-dimensional structural motif exists with three-dimensional
cross-linking. Finally, we have described methods that may
permit the optimization of amorphous CN materials for high-
temperature, high-pressure composition analyses.
(
10) Tabbal, M.; Merel, P.; Moisa, S.; Chaker, M.; Ricard, A.; Moisan,
M. Appl. Phys. Lett. 1996, 69(12), 1698.
11) Mehring, M. Principles of High-Resolution NMR in Solids. 2nd
(
15
ed.; Springer-Verlag: New York, 1983.
(12) Schmidt-Rohr, K.; Spiess, H. W. Multidimensional Solid-State NMR
and Polymers; Academic Press: New York, 1994.
(
13) Baldus, M. A. Ph.D. Dissertation, Eidgenossische Technische
Hochschule, Zurich, 1996.
14) Baldus, M.; Geurts, D. G.; Hediger, S.; Meier, B. H. J. Magn. Reson.
(
A. 1996, 118, 140-144.
(15) Duijvestijn, M. J.; Van Der Lugt, C.; Smidt, J.; Wind, R. A.; Zilm,
K. W.; Staplin, D. C. Chem. Phys. Lett. 1983, 102(1), 25.
(
16) Pan, H.; Pruski, M.; Gerstein, B. C.; Li, F.; Lannin, J. S. Phys ReV
B. 1991, 44(15), 6741-6745.
17) Hall, D.; Maus, D. A.; Gerfen, G. J.; Inati, S. J.; Becerra, L. R.;
(
Dahlquist, F. W.; Griffin, R. G. Science 1997, 276, 930-932.
(18) Abragam, A. The Principles of Nuclear Magnetism; University
Press: London, 1961; Ch. 9.
(
19) Stevens, A. J.; Koga, T.; Agee, C. B.; Aziz, M. J.; Lieber, C. M.
J. Am. Chem. Soc. 1996, 118, 10900.
20) Bennati, M.; et al. J. Magn. Reson. 1999, 138, 232-243.
(
(21) Gay-Lussac, L. J. Ann. Chim. (Paris) 1815, 95, 175.
(22) Jenneskens, L. W.; Mahy, J. W. G.; Vlietstra, E. J.; Goede, S. J.;
Bickelhaupt, F. J. Chem. Soc., Faraday Trans. 1994, 90, 327.
(23) Maya, L. J. Polym. Sci. A: Polym. Chem. 1993, 31, 2595.
(
24) Stoll, E.; Vaughan, R. W.; Saillant, R. B.; Cole, T. J. Chem. Phys.
1
974, 61(7), 2896-2899.
(
(
25) Harris, R. K.; Leach, M. J. Chem. Mater. 1990, 2, 320-323.
26) Wu, X.; Zilm, K. W. J. Magn. Reson. A. 1993, 102, 205.
Acknowledgment. We are grateful to Dr. Gang Wu and Dr.
Phil Costa for advice and discussions during the project and to
Chad Rienstra for guidance in the construction of the transmis-
sion line probes. Additional support from Prof. Charles Lieber
is gratefully acknowledged.
(27) Wu, X.; Burns, S. T.; Zilm, K. W. J. Magn. Reson. A. 1994, 111,
9.
2
(28) Harris, R. K.; Leach, M. J. Chem. Mater. 1992, 4, 260-267.
(29) Leone, E. A.; Curran, S.; Kotun, M. E. J. Am. Ceram. Soc. 1996,
7
9(2), 513.
(30) Kaplan, S.; Jansen, F.; Machonkin, M. Appl. Phys. Lett. 1985, 47,
750.
References and Notes
(31) Beamson, G.; Brennan, W. J.; Clayden, N. J.; Jennings, R. C. K.
J. Polym. Sci.: Part B: Polym. Phys. 1993, 31, 1205.
(
(
(
(
(
1) Cohen, M. L. Mater. Sci. Eng. 1996, A209, 1.
2) Liu, A. Y.; Cohen, M. L. Science 1989, 245, 841.
3) Liu, A. Y.; Cohen, M. L. Phys. ReV. B 1990, 41, 10727.
4) Niu, C.; Lu, Y. Z.; Lieber, C. M. Science 1993, 261, 334-337.
5) Kouvetakis, J.; Bandari, A.; Todd, M.; Wilkens, B. Chem. Mater.
(32) Ricci, M.; et al. J. Mater. Res. 1993, 8(3), 480-488.
(33) Duncan, T. M. A Compilation of Chemical Shift Anisotropies;
Farragut Press: Chicago, 1990.
(
(
34) Nayeem, A.; Yesinowski, J. J. Chem. Phys. 1988, 89, 4600.
35) Poole, C. P. Electron Spin Resonance, 2nd ed.; Dover Publications:
1
996, 6, 811-814.
6) Merchant, A. R.; McColluch, D. G.; McKenxie, D. R.; Yin, Y.;
Hall, L. J. Appl. Phys. 1996, 79(9), 6914.
7) Xin, H.; Lin, C.; Xu, W.; Wang, L.; Zou, S.; Wu, X.; Shi, X.; Shu,
H. J. Appl. Phys. 1996, 79(5), 2364.
Mineola, New York, 1983.
(36) Carrington, A.; McLachlan, A. D. Introduction to Magnetic
Resonance; Harper & Row: New York, 1967.
(
(
(37) Brotherton, T. K.; Lynn, J. W. Chem. ReV. 1959, 59, 841.