Structural Insights on Nitrogen-Containing Hydrothermal Carbon Using Solid-State Magic Angle Spinning 13 C and 15 N Nuclear Magnetic Resonance

Abstract

In the past 10 years, the hydrothermal treatment of biomass has gained an increasing interest in the field of material sci-ence. Many studies already focus on both applications 1 and fundamental 2 aspects motivated by the interest in producing carbonaceous powders with tunable sizes and surface properties directly from raw, processed, and even waste biomass. 3 Recent review articles can be consulted for an overview on the progress regarding this topic over the past few years. 4 The most interesting point of this approach is undoubtedly the perspective of proces-sing waste biomass based on lignins, cellulose into valuable carbonaceous materials, although the main research so far has been done using pure carbohydrates (glucose, sucrose) normally present in biomass composition. One of the main reasons for that is the fact that the chemical reactions that transform saccharides into hydrothermal carbons are extremely complex, and they had first to be understood starting from simple model systems. 5 It is widely known that saccharides dehydrolize to form furans, 6 and we have already pointed out the importance of such intermediate reactants in the formation of the final material. 5 Different techniques like Fourier transform infrared (FT-IR), 2b X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD) 7 have been used to identify the structure of the amor-phous material, but so far, the best results regarding the final structure of such carbon materials were obtained by using advanced solid state NMR techniques, 2a which attested to the presence of an abundant polyfuran core instead of the expected aromatic network, previously reported by other authors. 2b One important advantage of hydrothermal carbons (HC) is undoubtedly the possibility of modifying the surface 8 or the bulk of the carbonaceous network by introducing heteroatoms such as transition metals 1i,q,9 or nitrogen. 10 Hybrid metalÀcarbon nano-structures have been obtained with potential application in catalysis. The introduction of N-atoms within the bulk or at the surface structure of these materials is extremely interesting given the multitude of applications such as, for example, pH-responsive adsorbents, 11 supercapacitors, 12 fuel cell electrodes, 13 CO 2 adsorbers, 14 and materials with extremely high conductivities. 15 We have recently shown the possibility of making nitrogen-rich carbon nanoparticles using a mixture of glucose and ovalbumin, 16 while glucosamine and chitosan 10 can also be used as simulta-neous nitrogen sources. Preliminary results show a very good and highly selective CO 2 uptake, suggesting that such materials could be successful candidates as CO 2 sequestration agents 17 or for use in supercapacitors. 18 For this reason, we want to address the problem of structural identification no matter which natural

Full text

Page 1

Published:April 21, 2011
r2011 American Chemical Society
8976
dx.doi.org/10.1021/jp2015512|J. Phys. Chem. C 2011, 115, 8976–8982
ARTICLE
pubs.acs.org/JPCC
Structural Insights on Nitrogen-Containing Hydrothermal
Carbon Using Solid-State Magic Angle Spinning13C and15N
Nuclear Magnetic Resonance
Niki Baccile,*,†Guillaume Laurent,†Cristina Coelho,†Florence Babonneau,†Li Zhao,‡and
Maria-Magdalena Titirici‡
†UPMC Univ Paris 06 and CNRS, UMR 7574, Chimie de la Mati? ere Condens? ee de Paris, F-75005, Paris, France
‡Max-Planck-Institut f€ ur Kolloid-und Grenzfl€ achenforschung, Research Campus Golm, D-14424 Potsdam, Germany
b
S Supporting Information
’INTRODUCTION
In the past 10 years, the hydrothermal treatment of biomass
has gained an increasing interest in the field of material sci-
ence. Many studies already focus on both applications1and
fundamental2aspects motivated by the interest in producing
carbonaceous powders with tunable sizes and surface properties
directly from raw, processed, and even waste biomass.3Recent
review articles can be consulted for an overview on the progress
regardingthistopicoverthepastfewyears.4Themostinteresting
point of this approach is undoubtedly the perspective of proces-
sing waste biomass based on lignins, cellulose into valuable
carbonaceous materials, although the main research so far has
been done using pure carbohydrates (glucose,sucrose) normally
presentinbiomasscomposition.Oneofthemainreasonsforthat
is the fact that the chemical reactions that transform saccharides
into hydrothermal carbons are extremely complex, and they had
first to be understood starting from simple model systems.5It is
widely known that saccharides dehydrolize to form furans,6and
wehavealreadypointedouttheimportanceofsuchintermediate
reactants in the formation of the final material.5Different
techniques like Fourier transform infrared (FT-IR),2bX-ray
photoelectron spectroscopy (XPS), and X-ray diffraction
(XRD)7have been used to identify the structure of the amor-
phous material, but so far, the best results regarding the final
structure of such carbon materials were obtained by using
advanced solid state NMR techniques,2awhich attested to the
presence of an abundant polyfuran core instead of the expected
aromatic network, previously reported by other authors.2b
One important advantage of hydrothermal carbons (HC) is
undoubtedly the possibility of modifying the surface8or the bulk
ofthecarbonaceousnetworkbyintroducingheteroatomssuchas
transition metals1i,q,9ornitrogen.10Hybrid metal?carbon nano-
structures have been obtained with potential application in
catalysis. The introduction of N-atoms within the bulk or at the
surface structure of these materials is extremely interesting given
themultitudeofapplicationssuchas,forexample,pH-responsive
adsorbents,11supercapacitors,12fuel cell electrodes,13CO2
adsorbers,14and materials with extremely high conductivities.15
We have recently shown the possibility of making nitrogen-rich
carbonnanoparticlesusingamixtureofglucoseandovalbumin,16
while glucosamine and chitosan10can also be used as simulta-
neousnitrogensources.Preliminaryresultsshowaverygoodand
highlyselectiveCO2uptake,suggestingthatsuchmaterialscould
besuccessfulcandidatesasCO2sequestrationagents17orforuse
in supercapacitors.18For this reason, we want to address the
problem of structural identification no matter which natural
Received:
Revised:
February 16, 2011
March 31, 2011
ABSTRACT: Here,13C and15N solid state NMR is used as the
main and most effective characterization technique on nitrogen-
containing hydrothermal carbons obtained from glucose and
glycine. This study represents a model system for other types of
nitrogen-containing hydrothermal carbons, which were shown to
have interesting energy-storage properties (Zhao et al. Adv. Mater.
2010, 22, 5202). These materials are obtained either from
N-containing carbohydrates or from pure carbohydrates in the
presence of natural amino-containing compounds such as proteins or aminoacids. In contrast to what is generally known for this
model system, high molecular weight heterogeneous polymers (e.g., melanoidins) that are formed when sugars and amino acids
combinethrough theMaillard reaction, wefoundanextended nitrogen-containingaromaticnetwork, whichischemicallyboundto
a polyfuran network known to be one of the main components of the biomass-derived hydrothermal carbons. In contrast to the
hydrothermal carbons obtained from pure carbohydrates, these types of N-containing materials have an increased level of aromatic
character already present at 180 ?C, after the hydrothermal treatment.

Page 2

8977
dx.doi.org/10.1021/jp2015512 |J. Phys. Chem. C 2011, 115, 8976–8982
The Journal of Physical Chemistry C
ARTICLE
nitrogen sources, like aminosugars, aminoacids, or proteins, are
used.Wewilluseboth13Cand15NsolidstateNMRexperiments
toidentifycarbonandnitrogensitesandcomparethestructureof
a nitrogen-containing material to a nitrogen-free one, pointing
out the most important structural differences. In particular, a
doubleisotope-enrichedglucose?glycinemixturewillbeusedas
modelsystemandcomparedtonon-enrichableones.Inaddition,
we will show that the hydrothermal treatment of glucose and
glycine,whichistheclassical reaction tomakemelanoidin resins,
actually leads to a highly complex material with an unexpected
structure of very different nature than typical melanoidin resins
themselves,whichwererecentlyinvestigatedbyacombinationof
advanced13C and15N solid state NMR experiments.19
Here, we will use a multinuclear NMR approach to show that
the combination of longer reaction times and higher tempera-
turesunderhydrothermalconditionsdrasticallybringsthewidely
known Maillard reaction between glucose and glycine to its
extreme condensation products characterizedwith a high load of
aromatic nitrogen-containing groups, which were not observed
before in the final raw material. Interestingly, no matter the
complexity of the nitrogen source, the final material seems to
have very close structural features. The better knowledge of the
nitrogen-containing carbonaceous structure should help under-
standing the energy-storage properties of these materials and,
eventually, improve and control them through synthesis.
’EXPERIMENTAL SECTION
Sample HC glu-C13-N15-gly Preparation. Ten milliliters of
a deionized water solution containing 1 g of fully labeled
D-glucose-13C6(Aldrich; CAS, 110187-42-3) and 0.4 g of glyci-
ne-15N(Aldrich;CAS,7299-33-4)wasinsertedintoaglassvialinside
atypicalpoly(tetrafluoroethylene)-linedautoclavesystemforhydro-
thermal reaction at 180 ?C for 12 h. After reaction, the autoclave is
cooled down in a water bath at room temperature. The obtained
black solid powder was then separated from the remaining aqueous
solutionbycentrifugation(7000rpmfor20min)anddriedat80?C
in an oven under vacuum for 12 h.
Scanning electron microscopy (SEM) images were acquired
on a LEO 1550/LEO GmbH Oberkochen provided with an
Everhard Thornley secondary electron and in-lens detectors. Che-
mical analysis was performed on a (C, N, O, S, and H) Elementar
Vario Micro Cube.
1H and13C solid-state magic angle spinning (MAS) NMR
experiments have been acquired on a Bruker Avance 300 MHz
(7.0 T) spectrometer using 4 mm zirconia rotors spinning at a
MAS frequency of νMAS= 14 kHz.1H and13C chemical shifts
were referenced relative totetramethylsilane (TMS;δ=0ppm).
Details on one pulse MAS, cross-polarization (CP) MAS, inver-
sion recovery cross-polarization (IRCP), and CP MAS homo-
nuclearsingle-quantumdouble-quantum(SQ-DQ)13Ccorrelation
experiments have been published elsewhere.2a
15N one pulse experiments were performed on a Bruker
Avance 500 MHz (11.7 T) spectrometer using 7 mm zirconia
rotors at νMAS= 5 kHz. An antiring pulse program (ARING)
witha90?pulseangleof7.70 μswasused.Arecycledelay of30s
and number of transients of 2000 were found to be optimal
conditions.15N{1H} CP MAS experiments were done using a
recycle delay of 3 s and a contact time of 2 ms, TPPM decoupling
was applied during signal acquisition, and the number of transients
was 7200.15N chemical shift were referenced to labelled glycine
at ?348 ppm on the nitromethane scale (CH3NO2; δ = 0 ppm).
Triple resonance CP MAS13C{1H,15N} heteronuclear corre-
lation was recorded on a 700 MHz (16.4 T) spectrometer with a
3.2 mm zirconia rotor (νMAS= 22 kHz) using a triple1H-X-Y
probe. A standard double CP pulse scheme (Figure 1 in the
Supporting Information) was used. Typical acquisition values
were as follows: proton 90? flip angle; t90?= 2.75 μs;15N{1H}
and13C{15N} contact times, tc1= 5 ms and tc2= 1 ms; 64
increments in the indirect dimension; and 1200 transients.
Quadrature detection was achieved using the States method.20
’RESULTS
As pointed out already in the Introduction, given the amor-
phous nature of hydrothermal carbons (HC) and their chemical
heterogeneity, it is impossible to perform a clear-cut structural
study using characterization techniques like XRD, infrared
spectroscopy, or even XPS, whose poor resolution does not
provide full pieces of information. Recently, we have used solid
state NMR to solve the structure of hydrothermal carbon from
glucose2aand identify nitrogen sites in glucosamine-derived HC
samples.10Similar experiments are very helpful whenever che-
mical enrichmentis possible, which is seldom the case, especially
if raw biomass is employed in the synthesis of HC. For this
reason,weconcentratedourattentionontotheglycine?glucose-
derived HC as a model system. We employed15N-glycine and
13C-glucose both as spin-1/2 nitrogen and carbon sources.
Some preliminary morphological studies of nitrogen-doped
hydrothermalcarbonpowdersobtainedfromglucoseandglycine
are very similar to the previously synthesized carbon materials.5,2a
Figure2intheSupportingInformationshowsSEMmicrographs
of typical spherical microparticles as usually obtained from
the hydrothermal carbonization process. The elemental
composition shows that about 8.4 ( 0.5 w% of nitrogen is
incorporated inside the material (C, 63.3 ( 1.0 w%; H, 5.5 (
1.0 w%); this value is comparable with the ones obtained in the
presence of the ovalbumin protein16
glucosamine.10Even if elemental analysis shows that nitrogen
is successfully incorporated in the sample, nothing can be told
about the nature of the N sites and whether the glycine motif is
keptintactasinthecaseofmelanoidins19orunderwentchemical
modification.13C and15N solid state NMR experiments dis-
cussed below will bring a detailed answer to these questions.
The first point to be addressed is the validity of using the
glucose?glycine mixture as model compound. Figure 1 shows
the
samples obtained from the hydrothermal treatment of glucose/
albumin (Figure 1b),16 15N-glucosamine10(Figure 1c), the mixtures
13C-glucose/15N-glycine (Figure 1d) compared with pure glu-
cose (Figure 1a). We have previously observed2athat a contact
time of 3 ms is enough to homogeneously excite the whole13C
population,allowingaqualitativebutreliablecomparisonamong
the relative intensities of the
comparison among the spectra shows very similar features:
The amount of CdO groups (200?210 and 175 ppm) and
aliphatic carbons (10?60ppm)issignificant unlessglucosamine
aloneisused;thearomaticregionisdominatedbyalargeintense
peak between 123 and 130 ppm (shaded region in Figure 1),
which is generally absent when glucose (Figure 1a) alone is
processed under the same conditions, and two peaks at 150 and
110 ppm, assigned to, respectively, CdC?O and HCdC?O
groups.Theresonance inthearomaticregionaround125ppmis
very intense for the HC glu-C13-N15-gly system, indicating a
as well as from
13C CP MAS response of a set of nitrogen-containing
13C resonance peaks. A close

Page 3

8978
dx.doi.org/10.1021/jp2015512 |J. Phys. Chem. C 2011, 115, 8976–8982
The Journal of Physical Chemistry C
ARTICLE
larger aromatic network, whose existence is hard to prove in the
HC glu sample. In fact, we have previously2ashown that
hydrothermal treatment of glucose provides a carbonaceous
scaffoldwhoseamorphousstructureisdominatedbyanextended
furan type network. For this reason and given the possibility to
use13C-enriched glucose and15N-enriched glycine, we show
hereafter a full structural study of their hydrothermal derivative
with our discussion based on the C?H, C?N, and C?C local
environments.
The structural investigation of the nitrogen-containing carbo-
naceous material is done according to a previously published
experimental line.2aIn that occasion, different NMR techniques
were combined together to identify the main constituents of the
carbonaceous scaffold obtained from hydrothermal treatment of
glucose. Here, a similar approach is used; in particular, one-pulse
13C and15N MAS NMR experiment allows the quantitative
identificationofallcarbonandnitrogensitesandarepresentedin
Figures 2a and 3a; CP and IRCP experiments give an unambig-
uous insight on protonated carbons and nitrogen sites; in parti-
cular, IRCP helps discriminating between CH, CH2, and CH3
groups.21Figures 2b and 3b present13C and15N CP MAS
experiments performed at a given contact time, but the full set of
data on CP and IRCP experiments can be found in Figures 3 and 4
in the Supporting Information. Finally, two-dimensional
SQ-DQ13C and HETCOR15N{1H}?13C experiments allow
identification of, respectively, carbon environments within one
C?C atomic bond2a,22and NH?C proximities.
Carbon?ProtonProximities.Chemicalshiftanalysiscombined
with CP experiments was used to discriminate between carbonyl,
aromatic, and aliphatic groups (Figure 2b). In particular, variation of
contact time in CP experiments (Figure 3c in the Supporting
Information) nicely allows to discriminate between protonated
andnon-protonatedsp2carbons,whileIRCPexperimentsgivean
insight on the degree of protonation of aliphatic carbon atoms
(Figure 3a,b in the Supporting Information).
Carbon?Nitrogen Proximities. The structural analysis was
completed by one-pulse nitrogen-15 and15N{1H} CP NMR
experiments (Figure 3). First of all, in good agreement with
elemental analysis (8.4 ( 0.5 w%), these experiments clearly
show that nitrogen is well introduced inside the carbonaceous
network, and it is surrounded by different carbon environments.
Chemical shifts at ?75 and ?250 ppm indicate, respectively,
sp2and sp3nitrogen?carbon bonds.15N{1H} CP (Figure 3b)
onlyshowsprotonatednitrogensites,whosechemicalshiftrange
(?200/?300 ppm) indicates the presence of amide or pyrrole
groups;23interestingly,aminegroupsthatareinitiallyintroduced
via the glycine molecule have massively reacted since the NH2
resonance at ?330 ppm is very low; additionally, the15N one
pulse experiment (Figure 3a) shows nonprotonated aromatic,
pyridinic, and pyrazine-like groups (0?150 ppm).
This preliminary analysis shows very interesting features, which
make this material very different from the hydrothermal carbon
obtained from pure glucose;2ain particular, the introduction of
nitrogen induces a systematic aromatization of the carbonaceous
network.
Figure1.
C13-glucose-derived carbons (HC glu) and several nitrogen-containing
hydrothermal carbon materials: (a) sample HC glu, (b) HC glu-
albumin, (c) HC glu-N15-glucosamine, and (d) HC glu-C13-N15-gly.
In a, glucose is used alone, but it can co-react with (b) ovalbumin16and
(c)
15N-glycine.Notethatthebettersignal-to-noiseratioofHCgluandHC
glu-C13-N15-gly derives from13C enrichment.
13CCPMASNMRexperiments(tc=3ms)recordedonpure
15N-glucosamine.10In d,
13C-glucose reacts in the presence of
Figure 2. Comparison between13C MAS NMR experiments acquired
on the HC glu-C13-N15-gly system under the following conditions: (a)
one-pulse, (b) CP performed at tc= 50 μs, (c) projection from the SQ-
DQ13C experiment attc=3ms, and (d) projection fromthe HETCOR
15N?13C experiment.
Figure 3.
spectra ofHCglu-C13-N15-glyand15Nprojectionofthe13C{1H,15N}
HETCOR experiment.
15N one-pulse and CP MAS (with 2 ms of contact time)

Page 4

8979
dx.doi.org/10.1021/jp2015512 |J. Phys. Chem. C 2011, 115, 8976–8982
The Journal of Physical Chemistry C
ARTICLE
Deeper insights on the type of C?C and N?C bonds can be
obtained from the13C{1H,15N} triple resonance CP NMR
experimentsshowninFigure4.Here,aninitial15N{1H}CPstep
at an optimum contact time (tc1= 5 ms) excites all protonated
nitrogen atoms (Figure 3); then, polarization is transferred to
13C nuclei via a second contact time (tc2= 1 ms), which was
chosenlongenoughtodetectasignalwithareasonablesignal-to-
noiseratiobutshortenoughtoprobetheshortestpossibleC?N
distances. The use of the proton bath to excite the15N spins
limits the accessibility to solely N?H groups (?150/?300
ppm) with respect to the entire nitrogen population, which also
includes aromatic sites (0/?150 ppm) (Figure 3a). Any attempt
torunacomplementary15N{1H,13C}experimentfailed.Forthis
reason, we stress the fact that the results obtained from this
experiment only concern protonated nitrogen environments.
The 2D13C{1H,15N} map in Figure 4 provides a direct
fingerprint of the nitrogen-to-carbon proximities. The
15N-filtered13C projection shows two main regions, one with
twopeaksat130.0and136.7ppmandtheotherat70ppm,while
the15Nprojectionshowsthreeresonancesatabout?195.0,?230.0,
and ?250.0 ppm. As expected, these values are comparable with
simple15N{1H} CP experiments in Figure 3. A deep look at the
2D map shows that CdCH?NH groups are the most abundant
species in the material, followed by CHx?NH species. On the
contrary, no specific correlation is found between (protonated)
nitrogen and carbonyl groups, suggesting that amides are not
relevant. This result is quite interesting, as it suggests that the
signal between 130.0 and 136.0 ppm mainly relates to pyrolic
species.
Carbon?Carbon Connectivities. Two SQ-DQ correlation
experiments were recorded to select different types of carbon
environments. In the first one (Figure 5a), a13C{1H} contact
time (tc) of 35 μs was employed before the excitation of DQ13C
coherences, while inthe second one (Figure 5b) the tcwas much
longer (3 ms). The shorter the tc, the higher the selectivity to
detect protonated carbons. In both cases, the SQ-DQ
excitation and reconversion time was the same (285.7 μs) and
chosen to observe only the direct carbon?carbon bonds.2a,22
13C
Atshortcontacttime(35μs,Figure5a),mostoftheconstitutive
chemical units can be identified. In particular, 1 and 1a refer to the
classicalfuraneunitconnectedtoaliphaticgroupsviatheR-carbon
atomwhile4showsketo-aliphaticandintra-aliphaticbonds.These
specieswerealreadyobservedontheglucose-derivedhydrothermal
carbon and are quite typical for these materials.2aAdditionally,
more specific bonds can be identified in this nitrogen-containing
sample. First of all, the on-diagonal cross-peak at 127.5 ppm (2)
now dominates the 2D map (Figure 5a,b); in particular, its on-
diagonalpositionsuggeststheexistenceofabundantaromaticunits,
whiletheoff-diagonal(2a)peaksatδ=151and127.5ppmsuggest
thepresenceofpossiblephenols.Thecross-peakat127.5ppmalso
indicates the existence of possible pyrazine or pyridine units that
canberelatedtotheexistenceofthebroadpeakat?75ppminthe
one-pulse
aromatic bonds could also be proposed (2a). Pyrrole type com-
pounds, proposed after the13C{1H,15N} HETCOR experiment,
could be justified by cross-peaks (3), while NH?CH groups
should be observed in 5. At longer contact times, the SQ-DQ
experimentdoesnotbringfurtherinformationonthematerialcore
network, but it shows the existence of newly appearing and self-
correlating cross-peaks, as outlined by the connecting black arrow.
Inparticular,theCOOHgroupat176ppmisreadilyconnectedto
a cheto-bond (202 ppm) through a chemical group resonating at
106 ppm, whose exact attribution constitutes a harsh task at the
moment. These cross-peaks probably identify isolated or grafted
molecules whose lack of signal at short contact times rather
depends on their mobility (less efficient CP transfer) than on the
lackofprotonatedenvironment;infact,CPandIRCPexperiments
show that carbon peaks at 70.6 and 106.7 ppm are indeed
protonated. In Figure 5b, we propose the possible molecular
structure, which, according to the exact nature of R and R0groups,
may be related to organic acids derivatives (e.g., dihydroxyfumaric
acid,tartaricacid,ordehydroacorbicacid).ThepresenceofN-atom
inthesecompoundscannotbeexcluded,assuggestedbythecross-
peak in Figure 4 resonating at δ(13C) = 70 ppm.
After combining the pieces of information obtained from one-
pulse, CP, IRCP, HETCOR15N?13C, and SQ-DQ13C experi-
ments presented above, the deconvolution of the one-pulse13C
spectrum (Figure 4 in the Supporting Information) of the HC
glu-13C-15N-gly sample (Figure 2a) shows not less than 25
different carbon species, whose list of resonances, chemical shift
values, and tentative attribution is listed in Table 1 in the
Supporting Information.
Ifcomparedwithliterature,theglucose/glycinesystemiswell-
known for its reactivity in hydrolytic medium undergoing
Maillard reaction, which is at the origin of the nonenzymatic
browning in food chemistry and in the synthesis of melanoidin
resins.19,24Nevertheless, the conditions employed in the HC
process at higher temperatures (T > 160 ?C, self-generated
pressures—10/30 bar—and residence times between 12 and
24 h) produce a different material: Typical melanoidins keep up
to 60% of glycine molecules intact, and the formation of
pyrazines and pyridines is generally not observed, while13C
and15N NMR experiments on glycine-derived HCs show no
trace of leftover glycine, and the overall amount of NH2groups
has almost quantitatively reacted further. Maillard reaction
cascades are the most probable pathways that transform sugar,
sugar derivatives, and amines into cyclic compounds, but the
formation of modified pyrazines,25furanones, pyranones,26and
pyrroles24bis largely documented more as degradation com-
pounds rather than as network-forming bricks.
15N spectrum in Figure 3. Finally, specific furanic-
Figure4. 2DmapshowingthedoubleCP13C{1H,15N}experimenton
the HC glu-C13-N15-gly system and obtained with pulse sequence
showninFigure1intheSupportingInformation.Contacttimesforeach
CPsteparegiveninthefigure.Insightonthepossiblechemicalgroupsis
reportedinFigure5.Notethathere“R”isjustageneralnotationtorefer
to C or H elements.

Page 5

8980
dx.doi.org/10.1021/jp2015512 |J. Phys. Chem. C 2011, 115, 8976–8982
The Journal of Physical Chemistry C
ARTICLE
Figure5. HCglu-C13-N15-gly.SQ-DQ13Cmapsobtainedforequaldoublequantumexcitationandreconversiontimes(285.7μs).Thecontacttimes
before double quantum coherences excitation are tc= 35 μs (a) and tc= 3 ms (b).

Page 6

8981
dx.doi.org/10.1021/jp2015512 |J. Phys. Chem. C 2011, 115, 8976–8982
The Journal of Physical Chemistry C
ARTICLE
Togofurther,oneshouldnotforgetthatthemaindegradation
productofglucoseishydroxymethylfurfural(HMF),whichisthe
main reactive compound in the formation of hydrothermal
carbon2a,5and that spectroscopic signature of furane rings is still
abundant in the final nitrogen-containing HC material. The C
site in R-position in furan rings and their derivatives can easily
polymerize via radical, cationic, or anionic reactions;27nitrogen-
containing derivatives of HMF also can be synthesized as
precursors for linear and branched polymers,28but the chemical
processes behind are still quite unclear because the reaction
between glucose, furans, and glycine does not stop at the simple
glycineincorporation,andmethodstofollowthereactionsinsitu
are at the moment not yet developed.
’CONCLUSIONS
Nitrogen-containing carbons constitute a new class of inter-
esting energy-storage materials. The use of renewable resources
in combination of hydrothermal treatment also constitutes a
cheap way to introduce nitrogen within a carbonaceous scaffold.
Nevetheless, the complexity of the reaction mechanism and of
thefinalstructureitselfconstitutesabarriertotheunderstanding,
though the improving, of these materials.
Treatingnaturalnitrogen-containingmoleculessuchasamino
acids or proteins with glucose or aminated saccharides such as
chitosan or glucosamine under hydrothermal conditions provide
the formation of nitrogenated materials up to 8?10 w%. Here,
we have carried out a specific structural study on the glycine/
glucose system using solid state NMR and isotopically enriched
(13C and15N) sources. One-dimensional13C experiments (one-
pulse and CP MAS) showed that all hydrothermal carbons have
very similar structures no matter the complexity of the carbon
source. Then, the presence of extended aromatic domains in
coexistance with the expected furan network is documented and
constituted the main feature of these materials.
Finally, the combination of one-dimensional (CP, IRCP) and
two-dimensional (SQ-DQ13C, double CP13C{1H,15N}) ex-
periments help drawing the carbon-to-carbon and carbon-to-
nitrogen chemical arrangements. We were able to propose
several possible structures that include the presence of pyrazine
motifs connected to furans, and we showed the quasi-absence of
NH2groups.Interestingly,thestructureofsuchamaterialisvery
atypical for the Maillard reaction between glucose and glycine,
whichnormallyreacttogivemelanoidinresinsforwhichonlythe
byproducts were shown to contain aromatic nitrogens but never
the final solid. There is no trace left of the initial glycine and
glucose. Instead, they completely rearranged into an extended
pyrazo-furanic matrix. The exact mechanism is still ambiguous,
but possible reactions between the R-carbons of furans and the
pyrazines are a possible explanation.
’ASSOCIATED CONTENT
b
S
SupportingInformation.
{1H}and13C{15N}pulsesequence,SEMimagesofthenitrogen-
containing hydrothermal carbon particles, evolution of the13C
IRCP and CP and integrated intensities of selected peaks, CP
spectra, and chemical shift listing and attribution of the peaks.
Thismaterialis available free ofchargevia theInternetathttp://
pubs.acs.org.
Schemeofthe double CP15N-
’AUTHOR INFORMATION
Corresponding Author
*Fax: þ33 1 44 27 15 04. E-mail: niki.baccile@upmc.fr.
’REFERENCES
(1) (a) White, R. J.; Budarin, V. L.; Clark, J. H. ChemSusChem 2008,
1,408–411.(b)Yao,C.;Shin,Y.;Wang,L.Q.;Windisch,C.F.;Samuels,
W.D.; Arey,B.W.;Wang,C.;Risen, W.M.;Exarhos, G.J.J.Phys. Chem.
C 2007, 111, 15141–15145. (c) Budarin, V. L.; Clark, J. H.; Luque, R.;
Macquarrie,D.J.;Koutinas,A.;Webb,C.GreenChem.2007,9,992–995.
(d)Budarin,V.;Luque,R.;Macquarrie,D.J.;Clark,J.H.Chem.—Eur.J.
2007, 13, 6914–6919. (e) Yu, S. H.; Cui, X. J.; Li, L. L.; Li, K.; Yu, B.;
Antonietti, M.; Colfen, H. Adv. Mater. 2004, 16, 1636. (f) Tusi, M. M.;
Brandalise, M.; Correa, O. V.; Oliveira Neto, A.; Linardi, M.; Spinac? e,
E. V. Mater. Res. 2007, 10, 171–175. (g) Sevilla, M.; Lota, G.; Fuertes,
A. B. J. Power Sources 2007, 171, 546–551. (h) Qian, H. S.; Lin, G. F.;
Zhang,Y.X.;Gunawan,P.;Xu,R.Nanotechnology2007,18.(i)Cui,X.J.;
Antonietti, M.; Yu, S. H. Small 2006, 2, 756–759. (l) Sun, X. M.; Liu,
J. F.; Li, Y. D. Chem.—Eur. J. 2006, 12, 2039–2047. (m) Sun, X. M.; Li,
Y. D. Angew. Chem., Int. Ed. 2004, 43, 597–601. (n) Yang, H. X.; Qian,
J. F.; Chen, Z. X.; Ai, X. P.; Cao, Y. L. J. Phys. Chem. C 2007,
111, 14067–14071. (o) Zhu, K.; Egeblad, K.; Christensen, C. H. Eur.
J. Inorg. Chem. 2007, 3955–3960. (p) Hu, J.; Li, H.; Huang, X. J. Solid
State Ion. 2007, 178, 265–271. (q) Titirici, M.-M.; Antonietti, M.;
Thomas, A. Chem. Mater. 2006, 18, 3808–3812. (r) Sun, X.; Li, Y.
Angew. Chem., Int. Ed. 2004, 116, 3915–3919.
(2) (a) Baccile, N.; Laurent, G.; Babonneau, F.; Fayon, F.; Titirici,
M.-M.; Antonietti, M. J. Phys. Chem. C 2009, 113, 9644–9654. (b)
Sevilla, M.; Fuertes, A. B. Chem.—Eur. J. 2009, 15, 4195–4203.
(3) White, R.J.; Antonietti, M.;Titirici, M.-M. J.Mater. Chem. 2009,
19, 8645–8650.
(4) (a) Hu, B.; Yu, S.-H.; Wang, K.; Liu, L.; Xu, X.-W. Dalton Trans.
2008, 5414–5423. (b) Titirici, M.-M.; Antonietti, M. Chem. Soc. Rev.
2010, 39, 103–116. (c) Hu, B.; Wang, K.; Wu, L.; Yu, S.-H.; Antonietti,
M.; Titirici, M.-M. Adv. Mater. 2010, 22, 1–16.
(5) Titirici, M.-M.; Antonietti, M.; Baccile, N. Green Chem. 2008,
10, 1204–1212.
(6) (a)Lourvanij,K.;Rorrer,G.L.Appl.Catal.,A1994,109,147.(b)
Lourvanij, K.; Rorrer, G. L. Ind. Eng. Chem. Res. 1993, 32, 11. (c) Antal,
M. J.; Mok, W. S. L.; Richards, G. N. Carbohydr. Res. 1990, 199, 91.
(7) Titirici, M.-M.; Thomas, A.; Antonietti, M. Adv. Funct. Mater.
2007, 17, 1010–1018.
(8) Demir-Cakan, R.; Baccile, N.; Antonietti, M.; Titirici, M.-M.
Chem. Mater. 2009, 21, 484–490.
(9) Sun, X.; Li, Y. Angew. Chem. 2004, 116, 607–611.
(10) Zhao, L.; Baccile, N.; Gross, S.; Zhang, Y.; Wei, W.; Sun, Y.;
Antonietti, M.; Titirici, M.-M. Carbon 2010, 48, 3778–3787.
(11) Jia,Y.F.;Xiao,B.;Thomas,K.M.Langmuir2002,18,470–478.
(12) Hulicova, D.; Kodama, M.; Hatori, H. Chem. Mater. 2006,
18, 2318–2326.
(13) Shao,Y.Y.;Sui,J.H.;Yin,G.P.;Gao,Y.Z.Appl.Catal.,B2008,
79, 89–99.
(14) Plaza, M. G.; Pevida, C.; Arenillas, A.; Rubiera, F.; Pis, J. J. Fuel
2007, 86, 2204–12.
(15) Paraknowitsch, J. P.; Zhang, J.; Su, D.; Thomas, A.; Antonietti,
M. Adv. Mater. 2010, 22, 87–92.
(16) Baccile,N.;Antonietti,M.;Titirici,M.-M.ChemSusChem2010,
3, 246–253.
(17) Zhao,L.;Bacsik,Z.;Hedin,N.;Wei,W.;Sun,Y.;Antonietti,M.;
Titirici, M.-M. ChemSusChem 2010, 3, 840–845.
(18) Zhao, L.; Fan, L.-Z.; Antonietti, M.; Titirici, M.-M. Adv. Mater.
2010, 22, 5202–5206.
(19) Fang, X.; Schmidt-Rohr, K. J. Agric. Food Chem. 2009,
57, 10701–10711.
(20) States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson.
1982, 48, 286.

Page 7

8982
dx.doi.org/10.1021/jp2015512 |J. Phys. Chem. C 2011, 115, 8976–8982
The Journal of Physical Chemistry C
ARTICLE
(21) (a)Cory,D.G.;Ritchey,W.M.Macromolecules1989,22,1611.
(b)Alemany,L.B.;Grant,D.M.;Pugmire,R.J.;Alger,T.D.;Zilm,K.W.
J. Am. Chem. Soc. 1983, 105, 2142.
(22) Brinkmann, A.; Ed? en, M.; Levitt, M. H. J. J. Chem. Phys. 2000,
112, 8539.
(23) (a) Levy, G. C.; Lichter, R. L. Nitrogen-15 Nuclear Magnetic
Resonance Spectroscopy; Wiley: New York, 1979. (b) Gammon, W. J.;
Hoatson, G. L.; Holloway, B. C.; Vold, R. L.; Reilly, A. C. Phys. Rev. B
2003, 68, 195401.
(24) (a) Tressl, R.; Wondrak, G. T.; Garbe, L.-A.; Krueger, R.-P.;
Rewicki, D. J. Agric. Food Chem. 1998, 46, 1765–1776. (b) Adams, A.;
Tehrani, K. A.; Kers::yiene, M.; Venskutonis, R.; De Kimpe, N. J. Agric.
Food Chem. 2003, 51, 4338–4343.
(25) (a) Yaylayan, V. A.; Keyhani, A. J. Agric. Food Chem. 2001,
49,800–803.(b)Haffenden,L.J.W.;Yaylayan,V.A.J.Agric.FoodChem.
2005, 53, 9742–9746.
(26) Ames, J. M.; Bailey, R. G.; Mann, J. J. Agric. Food Chem. 1999,
47, 438–443.
(27) Gandini, A.; Belgacem, M. N. Prog. Polym. Sci. 1997, 22, 1203.
(28) Gandini, A.; Coelho, D.; Gome, M. J. Mater. Chem. 2009,
19, 8656–8664.