Porous carbons-derived from vegetal biomass in the synthesis of quinoxalines. Mechanistic insights

Article abstract of DOI:10.1016/j.cattod.2019.06.043

We report herein for the first-time acid biomass-derived carbons from vegetal biomass, with high developed porosity, prepared through integrating method comprising pyrolysis and surface phosphonation, able to efficiently catalyze the synthesis of quinoxalines from 1,2-diamines and α-hydroxi ketones, under aerobic conditions. The obtained results indicate that the reaction is mainly driven by a combination of acid function strength and textural properties in terms of conversion and selectivity. Furthermore, our experimental and theoretical observations suggest that the preferred reaction pathway for this transformation, in the presence of the investigated acid carbon catalysts, involves cascade reactions including imination reaction between reactants, successive imine-enamine and keto-enol tautomerisms, heterocyclization followed by dehydration, and aromatization. While the acid sites seem to be a relevant role in each reaction step, the system formed by activated carbon and molecular oxygen could be behind the last oxidative reaction to give the corresponding nitrogen heterocycles.

Full text of DOI:10.1016/j.cattod.2019.06.043

Accepted Manuscript
Title: Porous carbons-derived from vegetal biomass in the
synthesis of quinoxalines. Mechanistic insights
´
Authors: M. Godino-Ojer, R. Blazquez-Garcıa, I. Matos, M.
´
Bernardo, I.M. Fonseca, E. Perez Mayoral
PII:
S0920-5861(18)31357-9
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2019.06.043
CATTOD 12290
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
11 January 2019
8 May 2019
13 June 2019
´
Please cite this article as: Godino-Ojer M, Blazquez-Garcıa R, Matos I, Bernardo
´
M, Fonseca IM, Perez Mayoral E, Porous carbons-derived from vegetal biomass
in the synthesis of quinoxalines. Mechanistic insights, Catalysis Today (2019),
https://doi.org/10.1016/j.cattod.2019.06.043
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

POROUS CARBONS-DERIVED FROM VEGETAL BIOMASS IN THE SYNTHESIS OF
QUINOXALINES. MECHANISTIC INSIGHTS
M. Godino-Ojer1,2, R. Blazquez-García1, I. Matos3,*, M. Bernardo3, I. M. Fonseca3, E. Pérez
Mayoral1,*
1 Departamento de Química Inorgánica y Química Técnica, Universidad Nacional de Educación
a Distancia, UNED, Paseo Senda del Rey 9, Facultad de Ciencias, 28040-Madrid (Spain)
2
Facultad de Ciencias Experimentales, Universidad Francisco de Vitoria, UFV, Ctra. Pozuelo-
Majadahonda km 1.800, 28223, Pozuelo de Alarcón (Madrid, Spain)
3
LAQV/REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia,
Universidade Nova de Lisboa, 2829-516 Caparica (Portugal)
* Corresponding authors:
eperez@ccia.uned.es; Phone (+34) 913989047; Fax (+34) 913986697
ines.matos@fct.unl.pt; Phone (+351) 212948 300
Graphical abstract

Highlights




Hedychium gardnerianum biomass was successfully transformed into highly porous
acid ACs
Acid biomass-derived carbons efficiently catalyze the synthesis of nitrogen
heterocycles
Reaction is mainly controlled by a combination of acid functions strength and textural
properties
Experimental and theoretical results suggest a new reactive route for the synthesis of
quinoxalines
Dedication:

This paper is dedicated to Prof. Ziolek on the occasion of her 70th birthday. “Metal-free
catalysts in honor of our own Niobium Queen”
Abstract
We report herein for the first-time acid biomass-derived carbons from vegetal biomass, with
high developed porosity, prepared through integrating method comprising pyrolysis and
surface phosphonation, able to efficiently catalyze the synthesis of quinoxalines from 1,2-
diamines and -hydroxi ketones, under aerobic conditions. The obtained results indicate that
the reaction is mainly driven by a combination of acid function strength and textural properties
in terms of conversion and selectivity. Furthermore, our experimental and theoretical
observations suggest that the preferred reaction pathway for this transformation, in the
presence of the investigated acid carbon catalysts, involves cascade reactions including
imination reaction between reactants, successive imine-enamine and keto-enol tautomerisms,
heterocyclization followed by dehydration, and aromatization. While the acid sites seem to be
a relevant role in each reaction step, the system formed by activated carbon and molecular
oxygen could be behind the last oxidative reaction to give the corresponding nitrogen
heterocycles.
Keywords
Porous carbons, Nanocatalysts, Fine chemicals, Quinoxalines, Computational methods
Highlights
Hedychium gardnerianum biomass was successfully transformed into highly porous acid ACs
Acid biomass-derived carbons efficiently catalyze the synthesis of nitrogen heterocycles

Reaction is mainly controlled by a combination of acid functions strength and textural
properties
Experimental and theoretical results suggest a new reactive route for the synthesis of
quinoxalines
1.
Introduction
The use of biomass plays an important role on the development of renewable and sustainable
systems presently known as circular economy. Lignocellulosic biomass can be easily converted
into low cost and extraordinary versatile carbon materials with interesting physicochemical
properties [1]. In the last decades, biomass has gained attention for energy applications,
conversion into chemicals and char production, highlighting the related porous activated
carbons (ACs) for adsorption processes [2].
Hedychium gardnerianum is an original plant of Asia (India, between the Himalayas and Nepal),
which was introduced in the Portuguese islands of Azores and Madeira for ornamental
reasons. It is an invasive plant that propagates preferentially in wetlands, forest and
agricultural areas. Its rapid growth leads to formation of vast and dense colonies that suffocate
the development of native vegetation, which may lead to the modification of the natural
habitat of native animals and possible threats to the integrity of forest ecosystems, and it is
necessary its periodic removal. This lignocellulosic material can therefore be considered a
suitable raw material to obtain ACs with catalytic properties [3].
ACs have been widely investigated presenting a great variety of applications including catalysis
[4]. In fact, the unique properties of these materials such as chemical and thermal stability,
high surface areas, tunable textural properties and adapted surface chemistry make them
almost ideal supports or catalysts to synthetize valuable compounds [5,6]. The porous
structure can be strictly controlled by varying the synthesis conditions, aspect of supreme

importance in liquid phase reactions in which diffusional limitations can exist. Biomass-derived
carbon surface can be also functionalized in only one-step during the preparation process of
the carbon or after that by using post-synthesis strategies, if specific functional groups do not
tolerate the carbonization temperature. Particularly, biomass-derived carbons functionalized
with hydrophilic groups have been reported in several organic transformations such as
esterification/transesterification reactions involved in biodiesel production, etherification,
acetylation, alkylation, dehydration, rearrangements, biomass transformations, among others
[7]. However, to date there are no reports involving these kinds of carbon materials in
catalyzing cascade reactions for the synthesis of biologically relevant heterocyclic systems
[8][9]. In this context, the goal of this paper is focused onto the one-pot synthesis and
characterization of phosphorous functionalized biomass-derived carbons, through the
integrating method of pyrolysis and surface phosphonation, efficiently catalyzing the synthesis
of quinoxalines. Quinoxaline is an important heterocyclic motif rare in nature but showing a
wide variety of biomedical applications [10]. The most investigated synthetic approaches to
prepare quinoxalines consist of the reaction between 1,2-diamines and 1,2-dicarbonyl
components, -halo ketones or -hydroxi ketones. Among the catalysts reported for the
synthesis of quinoxalines from -hydroxi ketones are iron exchanged molybdophosphoric acid
[11], silica-supported sulfuric acid [12] and perchloric acid[13], silica gel [14], zeolites [15],
metal-organic frameworks, particularly CuBDC [16], and more sophisticated carbon catalysts
such as Ru/C [17] and carbon nanotube–gold nanohybrid [18]. In this context, we propose a
new family of eco-friendly porous carbons, including both biomass-derived and synthetic
carbon materials, with different textural properties and compositions, in which the porous
surface is functionalized with different oxygen functional groups emphasizing on acidic
functions such as CO2H, SO3H and PO3H2, tested in the formation of quinoxaline 1 from o-
phenylendiamine 2 and benzoin 3 under aerobic conditions. This work is based on our previous
studies concerning acid metal-free carbons as highly efficient and environmental catalysts

involved in cascade reactions for the synthesis of interesting heterocyclic systems, particularly
quinolines and related compounds [19,20]. We seek to study the effect of the sample porosity
and acidity in the synthesis of quinoxaline 1, also considering the biomass valorization for the
production of porous carbons highly actives. Additionally, mechanistic insights are also
envisioned as function of our experimental and theoretical observations.
2.
Experimental
2.1. Preparation of the catalysts
Three carbon series with different porous structure were prepared: (1) biomass-derived
carbons obtained from the chemical activation of the invasive plant Hedychium gardnerianum
with H3PO4. Two impregnation ratios (g biomass: g H3PO4) were used, 1:1 and 1:3; (2) xerogel
mesoporous carbons prepared by resorcinol-formaldehyde polymerization. Modification of the
pure carbon surface was carried out by using concentrated HNO3 or HNO3 and H2SO4
treatments as reported by Matos et al. [21]; and (3) ordered mesoporous carbon according to
the experimental protocol reported by Zhao group [22–24]. This method consists of the
organic-organic self-assembly approach, using a soft template as a structure-directing agent.
First a low molecular weight polymer, RESOL, from polymerized phenol and formaldehyde is
prepared to be used as carbon precursor. Subsequently, the pure carbon was submitted to
acid treatment with concentrated HNO3.
Carbon materials were labelled as follow: (1) biomass-derived carbons as RCB1 and RCB2 for
impregnation ratios of 1:1 and 1:3, respectively; (2) xerogel as X followed by N (XN) when X
treated with HNO3 and S when XN was treated with H2SO4 (XNS); (3) ordered pure mesoporous
carbon as CZ followed by N (CZN) when CZ treated with HNO3. In this case, treatment of CZN
with H2SO4 was not achieved because the porous carbon is destroyed.
2.1.1. Synthesis of RCB1 and RCB2

The biomass-derived ACs were prepared by using the stalks from the Hedychium
gardnerianum plant collected at São Miguel Island, Azores. The stalks were washed thoroughly
with water and then oven dried at 100 °C for several hours. Afterwards, the dried stalks were
mill grounded to obtain particles with around 10 mm long x 2 mm wide. The biomass particles
were then impregnated with H3PO4 (85 % w/w) at two different impregnation ratios, defined
as the mass ratio of biomass to activating agent, 1:1 (RCB1) and 1:3 (RCB2). Basically, the
biomass particles were immersed into acidic aqueous solutions with the known mass of
activating agent and kept stirring at 50 °C for 5 h. After impregnation, the mixtures were oven
dried for 1-2 days at 70 °C to remove water. The impregnated samples were subsequently
carbonized in a quartz reactor placed in an electric tube furnace at vertical configuration. The
samples were heated to 500 °C, at 5 °C min-1, and kept for 2 h under a N2 flow of 150 cm3 min-1.
After cooling under N2 flow, the samples were Soxhlet washed with deionized water up to
constant pH value (pH of deionized water). The washed samples were oven dried at 105 °C
overnight.
2.1.2. Synthesis of xerogel carbons, X
The xerogel carbon samples were prepared according to Lin and Ritter [25] by sol–gel
technique. The method includes the preparation of a solution containing 5% (w/v) solids with a
resorcinol/formaldehyde molar ratio of 1:2 and a resorcinol/sodium carbonate molar ratio of
50:1. The solution pH was adjusted with diluted HNO3 until 6.2. Then, the solution was placed
in an oven for 7 days at 85 °C, the resulting gel was washed with acetone and dried at 65 °C.
Subsequently, the sample was subjected to a thermal treatment under a 100 cm3 min-1 N2 flow
using a heating rate of 0.5 °C min-1, first for 5 h at 110 °C and then at 800 °C for 3 h. The
obtained carbon was labelled carbon X.
The xerogel carbon (X) was oxidized by treatment with a nitric acid solution (13 M) for 6 h (1
g/20 cm3) at 90 °C, then washed with deionized water in Soxhlet until constant pH and dried in
oven at 110 °C resulting in carbon XN.

The oxidized carbon (XN) was heated at 150 °C with concentrated sulfuric acid solution (1 g
carbon/20 cm3 sulfuric acid solution) for 13 h under N2 atmosphere, washed with deionized
water in Soxhlet until pH 7 and then dried in oven at 110 °C (carbon XNS).
2.1.3. Synthesis of ordered mesoporous carbons, CZ
Phenol (8 g) was dissolved in a 20% NaOH solution (0.371 g) and after formaldehyde (37 wt%,
14.5 g) was added dropwise. The mixture was kept at 75 °C for 1.5 h with stirring. The solution
was neutralized with 1 M HCl and the water present was removed by vacuum distillation at
temperature below 50 °C. The product was then dissolved in ethanol in order to obtain 20 wt%
solution of RESOL. This solution was filtered to remove some solid NaCl, thus obtaining the
final RESOL solution. After the RESOL was prepared, the mesoporous polymer was synthesized.
Pluronic F-127 (4.002 g) was dissolved in ethanol (40 g), and the ethanol RESOL solution (4.030
g) was added dropwise to the F-127 solution. The resulting mixture was transferred to dishes
and allowed to dry at room temperature overnight. The resulting membrane was heated at
100 °C for 24 h in an oven. The carbonization of the mesoporous polymer was carried out at a
heating rate of 1 °C min-1 and held for 2 h at 350 °C and 500 °C and finally for 5 h at 750 °C,
under N2 flow of 100 cm3 min-1, resulting in the ordered mesoporous carbon CZ. CZ surface was
oxidized with HNO3 with a 13 M solution, at 90 °C for 6 h (1 g/20 cm3). The materials were then
washed with deionized water in Soxhlet apparatus until pH 7 and oven dried at 110 °C (carbon
CZN).
2.2. Characterization of the catalysts
Textural characterization was performed by N2 adsorption at -196 °C on an ASAP 2010 V1.01 B
Micromeritics equipment. Elemental analysis of the carbon samples was carried out in a CHNS
Analyser (Thermo Finnigan—CE Instruments, model Flash EA 1112 CHNS series). The pH at the
point of zero charge (pHPZC) was determined by reverse mass titration following the method
proposed by Noh and Schwarz [26]. X-ray photoelectron spectroscopy measurements were
done with an Axis Supra by Kratos Analytical using monochromated Al Kα radiation. With an X-

ray power of 225 W, the survey was recorded at a pass energy of 160 eV and served for
elemental quantification. S 2p and P 2p detail spectra were recorded at 20 eV, whereas C 1s at
5 eV. Data analysis was done with CasaXPS. Fourier transform infrared (FTIR) spectra was
obtained over the range 400−4000 cm−1 using a Cary 630 FTIR spectrometer equipped with a
diamond attenuated total reflectance (ATR) of Agilent Technologies, with a thermoelectrically
cooled dTGS detector and KBr standard beam splitter. All the spectra were recorded via ATR
method with a resolution of 1 cm−1.
2.3. Catalytic performance
Briefly, a mixture of o-phenylendiamine 2 (1 mmol) and benzoin 3 (1 mmol), in toluene (5 mL),
in a three-necked vessel, equipped with thermometer, was placed on a multiexperiment work
station StarFish (Radley´s Discovery Technologies IUK). When the temperature reaches 100 °C,
the catalyst was added (25 mg) and the reaction mixture was stirred during 240 min. The
samples were periodically taken at 15, 30, 60, 120, 180 and 240 min and the catalyst filtered
off and the solvent evaporated in vacuo. The catalysts were milled and sieved to dp < 0.25 mm,
in order to avoid mass transfer limitations. Before use, carbon catalysts were activated at 60 °C
overnight.
The reactions were followed by TLC chromatography performed on DC-Aulofolien/Kieselgel 60
F245 (Merk) using mixtures of CH2Cl2/EtOH 98:2 as eluent. The reaction products were
characterized by 1H NMR spectroscopy. NMR spectra were recorded by using a Bruker AVANCE
1
DPX-300 spectrometer (300 MHz for 1H). H chemical shifts in [D6]DMSO or CDCl3 are
referenced to internal tetramethylsilane. Experiments of Ultra Performance Liquid
Chromatography-Mass Spectrometry (UPLC-MS) was achieved using the chromatograph UPLC
Waters Acquity I Class coupled to mass spectrometer maXis II, Bruker (MS-ESI electrospray
ionization, positive mode)

Characterization data of quinoxalines 1, 7 and 8 are in good agreement with those previously
reported using other catalytic systems [27].
2.4. Computational methods
All calculations were performed with the Gaussian 09 software package [28]. All geometries
were optimized using B3LYP/6-31+G(d,p). The stationary points were characterized by means
of harmonic vibrational frequency analysis. Thus, the transition structures were confirmed to
be first-order saddle points. The imaginary frequency was inspected in each to ensure it
represented the desired reaction coordinate. For key transition states the intrinsic reaction
coordinate (IRC) was followed to ensure it connects the reactants and products.
Based on previous studies [20], we select the most reduced models simulating the active
centers of the catalysts under study (Chart 1).
Chart 1. Selected reduced models simulating the acid active centers in the catalysts under
study. (a) Carbons treated with HNO3 (XN and CZN), (b) xerogel treated with HNO3 but also
with H2SO4 (XNS) and (c) biomass-derived carbons activated with H3PO4 (RCB1 and RCB2).
3.
Results and discussion
3.1.
Synthesis and characterization of the catalyst
We prepared three different series of acid metal-free porous carbons. Firstly, we synthetized
two biomass-derived carbons by activation of invasive vegetal specie with different amounts of
H3PO4. Furthermore, for comparison purposes we prepared two different pure carbonaceous
materials, a mesoporous xerogel by resorcinol-formaldehyde polymerization and an ordered
mesoporous carbon, which were subsequently modified by acid treatments to obtain carbons

with different textures and acid character. The carbons were characterized to determine the
features of the catalysts.
The textural characteristics of the biomass derived carbons presented in Table 1 allow
concluding that the chemical activation of the Hedychium gardnerianum biomass with H3PO4 is
a suitable methodology to prepare ACs with highly developed porosity. It was possible to
obtain a carbon with a BET area of 2197 m2 g-1 and a pore volume of 1.39 cm3 g-1 (RCB2) which
is comparable or even higher than many commercial carbons. RCB1 and RCB2 carbons present
isotherms of type I (b) which means that these materials have pore size distributions over a
wide range including wider micropores. Also, their isotherms present large H4 hysteresis,
indicating that a large mesopore volume is also present [29]. It is also possible to observe that
the increase in H3PO4 amount in the preparation procedure favored the development of
microporosity increasing around 2.5 times the microporous volume (from 0.23 in RCB1 to 0.59
cm3 g-1 in RCB2) while the mesoporosity only increase 1.3 times (Table 1).
Pore size distribution of the biomass derived carbons was obtained from the Non-Local Density
Functional Theory (NLDFT) equilibrium model for carbon slit-shaped pores (Figure 1). Based on
these calculations, these carbon samples have primarily micropores with a pore width around
1-1.8 nm. It can also be observed that RCB2 sample presents a higher content of mesopores
with 2.3 - 4 nm width, but the content in mesopores with widths above 5 nm is very similar for
both carbon samples.

1.20
1.10
1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
RCB1
RCB2
1.0
1.8
2.6
3.4
4.2
5.0
5.8
6.6
7.4
8.2
9.0
9.8
Pore width (nm)
Figure 1. NLDFT pore size distribution of the biomass derived carbons.
The synthesized carbon CZ presents a BET area of 749 m2 g-1, value comparable to those of the
literature [30], and a typical isotherm (type IV) and hysteresis (H2) for ordered mesoporous
material. However, these carbons also present microporosity, probably developed in the walls
of larger pores. The XN sample predominantly shows meso- and macroporosity presenting
smaller SBET
.
The surface characteristics of the samples were also analysed by elemental analysis, XPS and
pHPZC (Table 2 and 3). The elemental analysis of the biomass derived carbons revealed the high
oxygen content of the samples, these results combined with the low values of pHpzc of the
RCB1 and RCB2 suggests that the activation with phosphoric acid alone resulted in not only
high porosity, but also functionalization of the carbon surface. This effect is also probably due
to the presence of phosphorous groups at the surface of the catalysts, which was confirmed by
XPS studies.
Table 1. Textural parameters obtained from N2 adsorption-desorption isotherms of carbon
samples.
Catalyst SBET (m2 g-1) Vmicro (cm3 g-1) Vmeso (cm3 g-1)
VT (cm3 g-1)
DpBJH (nm)

B.E.T.
552
t-plot method BJH method
Gurvitsh
0.28
BJH method
1.7
X
0.16
0.11
0.15
0.15
0.10
0.23
0.59
0.11
0.45
0.57
0.29
0.22
0.58
0.77
XN
347
1.14
1.53
XNS
CZ
448
1.27
1.71
749
0.41
3.19
CZN
RCB1
RCB2
576
0.37
3.40
1076
2197
0.81
1.94
1.39
1.94
SBET – apparent surface area; B.E.T. - Brunauer, Emmet, and Teller method; Vmicro – micropore volume;
Vmeso – mesopore volume; VT – total pore volume; Dp – pore diameter; BJH - Barrett, Joyner, and
Halenda method.
The oxidation of the carbons surface resulted in the increase of oxygen content and
consequent decrease in pHpzc attributed to the introduction of acid functional groups. The
treatment of XN with sulfuric acid resulted in a decrease of oxygen content but the acidity of
the sample XNS was increased, revealing the successful introduction of strong acid sulfonic
groups in the sample.
Table 2. Surface chemistry characterization. Elemental composition (expressed on wt%)
determined by elemental analysis (E.A).and pHPZC of the carbon materials.
Sample
X
C
H
N
S
O
pHPZC
9.7
89.9
73.9
0.9
1.3
0.0
0.4
0.0
0.0
9.2
XN
27.8 3.3
17.5 2.8
7.4 9.0
24.8 3.2
35.0 2.4
8.7 3.0
XNS
CZ
84.2
91.9
73.3
62.3
89.8
1.3
0.7
1.2
2.3
1.0
0.03
0.0
0.7
0.5
0.5
0.4
0.0
0.0
0.0
0.0
CZN
RCB1
RCB2

Figure 2. FTIR spectra of the carbon catalysts CZ, CZN, XN and RCB2.
The FTIR spectra of the different samples show some differences in the carbon materials
(Figure 2). The CZ and CZN carbons present spectra with better defined bands probably due to
the ordered nature of the material. All the carbons have bands in the same characteristic
regions for carbon type materials [31]. The absorption bands around 2900 cm-1 are assigned to
C–H stretching in aliphatic series, while the band at approximately 1730 cm-1 is usually
assigned to C=O stretching vibrations of ketones, aldehydes, lactones or carboxyl groups.
Around 1580 cm-1 is the band associated with the C–C vibrations in aromatic rings and the
broad band between 1100-1200 cm-1 is common in oxidized carbons and normally assigned to
the C–O stretching in acids, alcohols, phenols, ethers and/or esters groups. However, the CZN
sample presents two distinct bands at 1370 and 1220 cm-1 also assigned to oxidized forms. In
the same context, the biomass derived carbon presents an even broader band with a shoulder,
which has also been reported as characteristic of carbons obtained using phosphoric acid and
can be assigned to the stretching mode of hydrogen-bonded P=O, to O–C stretching vibrations
in P–O–C (aromatic) linkage and to P=OOH [32].
Table 3. Surface chemistry characterization. Elemental composition (expressed on atomic %) of
the carbon materials determined by XPS.

Sample
C 1s[a]
O 1s[a]
S 2p[a]
P 2p[a]
O/C[b] O/C[c]
XN
XNS
CZ
CZN
RCB2
87.7
93.5
84.0
78.4
88.2
12.2
6.4
16.0
23.6
9.14
0.0
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.3
0.14
0.07
0.14
0.23
0.10
0.28
0.16
0.06
0.25
0.07
[a] Determined by XPS; [b] atomic ratio determined by XPS; [c] atomic ratio determined by E. A.
The XPS study of the biomass derived sample RCB2 was particularly important to determine
the presence of phosphorous in the resulting carbon matrix. The XPS P2p peak was located at
binding energy (BE) around 133 eV and could be fitted to assign the contribution of two peaks
one at 132.8 eV and another at 133.6 eV. These contributions to the P2p zone can be
described as P groups directly bonded to a carbon site (CPO3 or C2PO2 group), or with a bridge
O atom (COPO3 or (CO)3PO groups)[33,34].
Sample XNS is the only one where the sulphur element is observed, the BE values suggesting
an oxidized form of sulphur probably in the form of sulphate or sulfonic groups (OSO3H or
SO3H)[35].
The BE of the C1s spectrum characteristic of ACs are observed in all samples. The fitting of the
C 1s peak of all samples showed a predominant C sp2 component and then a smaller
contribution of oxidized forms of carbon, corresponding to alcohol/ ether (–COH, –COC–, O
C(O–C*)), carbonyl (–C=O), and carboxyl (COOH) functional groups at BEs of 286.3 eV, 287.8 eV
and 288.7 eV respectively [36,37].
For XN sample it is observed the predominance of alcohol or esters functional groups, while
the biomass derived carbon RCB2 presented a more even contribution of the different
functional groups. After oxidation, the sample CZN revealed a sharp increase in surface oxygen
content and a stronger contribution of CO.
Sample CZN seems to present a homogeneous distribution of oxygen functional groups since
the atomic ratio O/C determined by XPS is similar to the one determined by elemental

analysis. The contrary trend is observed for the xerogel samples XN and XNS revealing a
preferential oxidation and oxygen deposition inside the pores, thus not detectable by XPS.
On the other hand, for the RCB2 sample the atomic ratio O/C determined by XPS is similar to
the value obtained for the O/C ratio determined by elemental analysis. Since the oxidation of
the carbons resulted from the activation procedure, it was possible to obtain a material with a
homogeneous distribution of the oxygen surface groups throughout the carbon.
Raman spectroscopy was also used to characterize the porous carbons. Figure 3 shows Raman
spectra obtained for the prepared carbon materials. All the samples show two strong bands at
approximately 1345 and 1588 cm−1, assigned to the D and G bands typically present at
amorphous carbon samples. In particular, the presence of D band in Raman scattering is
associated with a disordered carbon structure and is assigned to phonons of A1g symmetry
originated at defects [38][39] which break the basic symmetry of the graphene sheet of carbon
matrices [40]. The graphitic G band is assigned to phonons of E2g symmetry.
The graphitization degree of a carbon material is generally characterized by the ID/IG intensity
ratio value in the Raman spectra and it may be a tool to identify structural changes in carbon
materials due to their treatment. For the materials prepared in this work, the ratio is in the
range of 0.83-0.90 suggesting a rather low ordered structure of the carbon materials.
Interestingly, no significant changes can be observed in the ID/IG intensity ratio after the acid
treatment. The biomass derived carbon although presenting a very high surface area,
presented the lower ID/IG intensity ratio indicating a higher degree of graphitization.

G-band 1588
D-band 1345
CZN, ID/IG= 0,90
CZ, ID/IG= 0,88
XS, ID/IG= 0,87
XN, ID/IG= 0,87
X, ID/IG= 0,85
RCB2, ID/IG= 0,83
1900
1700
1500
1300
1100
900
700
Raman Shift (cm-1)
Figure 3. Raman spectra for the carbon materials under study.
3.2. Catalytic performance
The catalysts were tested in the synthesis of quinoxaline 1 from o-phenylendiamine 2 and
benzoin 3, under aerobic conditions, in toluene, at 100 °C (Scheme 1).
Scheme 1. Synthesis of quinoxaline 1 from o-phenylendiamine 2 and benzoin 3, under aerobic
conditions, in toluene, at 100 °C, catalyzed by porous carbon materials.
Firstly, we check the catalytic behavior of pure carbons CZ and X for comparison with the
modified samples CZN and XN. Figure 3 depicts the conversion of diamine 2 vs time for both
types of carbons. As it can be observed, both pure carbons, CZ and X, are active in the selected
reaction with high conversion values but diminished selectivity to quinoxaline 1 (Figure 4).

CZ
X
CZN XN
100
80
60
40
20
0
0
60
120
Time (min)
CZ
X
CZN XN
100
80
60
40
20
0
60 min
120 min
Figure 4. Synthesis of quinoxaline 1 from o-phenylendiamine 2 and benzoin 3, under aerobic
conditions, in toluene, at 100 °C, catalyzed by CZ and CZN catalysts.
1
The analysis of the H NMR spectra of the samples, at different reaction times, when using CZ
material reveals the formation of the corresponding quinoxaline 1 (27%, 2h) together with the
intermediate 4 (49%, 2h) as major reaction product, these results are in good agreement with
previous studies using graphite as catalyst [41]. It was also detected the presence of additional
signals in the aromatic region at  = 8.2, almost overlapping the corresponding signals of
aromatic protons in quinoxaline skeleton. The absence of any aliphatic protons strongly
suggests the formation of additional intermediate specie whose structure could be assigned to
compound 5 (Scheme 2) as also suggested by MS experiments. Similar results were obtained
for the X carbon. These results indicate that the carbon matrix predominantly catalyzes the
conversion of reagents into the intermediate compound 4.

Oxidized carbons, CZN and XN, present lower surface area and diminished microporous
volume compared to the parent carbons, CZ and X, this is probably due to the incorporation of
acid oxygen groups in the mouth of the pores. These differences in textural properties may
account for the decreased in conversion. However, the observed increase in selectivity may be
attributed to the presence of acid functions in CZN and XN samples, as indicated by the pHPZC
values (pHPZC 3.2 and 3.3 for oxidized samples vs 9 and 9.7 for the parent carbons, Table 2).
Given the almost equal acidity of both samples, the higher selectivity for CZN could be due to
not only the higher amount of surface oxygen functions (Tables 2 and 3) but also to the
ordered channel structure.
Scheme 2. Reaction mechanisms accepted for the synthesis of quinoxaline 1 (Path a and b).
Alternative pathway for the synthesis of quinoxaline 1 in the presence of acid carbon catalysts
(Path c).
Scheme 2 also summarizes the accepted mechanisms reported for the investigated reaction
consisting of the cascade reactions condensation-dehydration-cyclization and aromatization
(Scheme 2, Path a) [42] or initial benzoin oxidation followed by double condensation and
dehydration to yield 2,3-diphenylquinoxaline 1 (Scheme 2, Path b) [11]. It is noticeable that in
our case benzyl 6, as oxidation product from benzoin, or intermediate compound 7 were not

detected. Therefore, these results open the doors to consider other alternative reaction
pathway for the formation of quinoxaline 1 in the presence of acid carbon catalysts probably
through Path c.
Following our ongoing investigations, we explore the reaction under the same experimental
conditions in the presence of XNS presenting stronger acid functions. Figure 5 depicts the
catalytic behavior observed for the investigated acid xerogels for comparison. XNS catalyst
results in the highest conversion values but also increased selectivity to product 1. In this case,
it seems that the important feature to selectivity is type and strength of acid sites. These
results suggest that the high conversion values are related to the combination of increased
surface area and the presence of stronger acid sites. Additionally, the reaction also occurs at
lower temperature (80 °C) giving lower conversion values (XN = 61%; 3 h) with maintained
selectivity, as expected. The non-activated XN catalyst leads to slight increase of conversion
values but surprisingly superior selectivity at all reaction times, this effect attributed to the
presence of water on the hydrophilic surface, as briefly commented in mechanistic
considerations section.
XN_no activated
XN
XNS
RCB1
RCB2
100
80
60
40
20
0
0
60
120
Time (min)
180
240

XN no activated
XN
XNS
RCB1
RCB2
100
80
60
40
20
0
60 min
120 min
180 min
240 min
Figure 5. Synthesis of quinoxaline 1 from o-phenylendiamine 2 and benzoin 3, under aerobic
conditions, in toluene, at 100 °C, catalyzed by xerogels XN and XNS.
Finally, we check the biomass-derived carbons in the reaction observing almost no differences
on their catalytic behavior in terms of conversion and selectivity (Figure 5). Given the high
microporosity of RCB1 and RCB2 samples complemented by the presence of mesoporous, high
conversion was expected. Selectivity differences observed at lower reaction times could be
attributed to the notably increase in surface area and porous volume of RCB2. Similarly, as for
XN and XNS, the enhanced selectivity seems to be related with the type and strength of acid
sites. Although XN presents higher amount of oxygen acid sites (detected by XPS), the
presence of sulfur or phosphor acid groups selectively drives the reaction for the desired
product. The reaction in terms of conversion and selectivity follows the order RCB2 > XNS > XN
at final reaction times. Therefore, the observed reactivity could be consequence of a
compromise between the acidity and the porosity of the catalyst. Thus, the biomass-derived
carbons were successfully proven to be an effective and sustainable alternative carbon
catalysts.
Figure 6 shows the evolution of the reaction catalyzed by the most effective, RCB2 sample, in
which it can be observed the formation of quinoxaline 1 while both intermediate compounds 4
and 5 disappear.

Additionally, we carried out the reaction in the presence of RCB2 sample in which the catalyst
was removed from the reaction mixture after 15 min of reaction time (Table 4). When the
catalyst is filtered off, the reaction mainly evolves toward the formation of intermediate 4
resulting on a diminished selectivity to the desired product 1. These results suggest that no
leaching of acidic species is produced.
Table 4. Synthesis of quinoxaline 1 from o-phenylendiamine 2 and benzoin 3, under aerobic
conditions, in toluene, at 100 °C.
Conversion (%) / Selectivity (%)
Catalyst
15 min
71 / 31
67 / 34
--
240 min
94 / 83
93 / 41
50 / 36
RCB2
RCB2*
Blank
* Catalyst was removed from the reaction mixture after 15 min of reaction time
Additionally, recycles experiments were carried out by using RCB2 sample observing
diminished conversion values of around 10% for the second and third cycles with maintained
selectivity to 1 (1st run: 82%; 2nd run: 72%; 3rd run: 74%).
Total
4
5
1
100
80
60
40
20
0
0
60
120
Time (min)
180
240
Figure 6. Synthesis of quinoxaline 1 from o-phenylendiamine 2 and benzoin 3, under aerobic
conditions, in toluene, at 100 °C, catalyzed by biomass-derived porous carbon RCB2. Evolution
of the reaction.

Finally, we study the scope of the methodology by using different -hydroxi ketones and
another diamine, under the same experimental conditions, in the presence of the best
catalyst, RCB2 sample (Scheme 3).
Scheme 3. Synthesis of nitrogen heterocycles catalyzed by biomass-derived carbon RCB2, in
toluene at 100 °C.
Remarkably, quinoxaline 7a was obtained in 71% with total selectivity in only 30 min of
reaction time, while when using acetoin (R1 = R2 =Me) it was observed the formation of 7b in
60 % with significant decreased selectivity (60 %) after 4h. The formation of intermediate
compounds I2 was also detected which was progressively transformed into quinoxaline 7b. It is
noteworthy that in this case total conversion was observed after only 30 min of reaction time
providing 49% of quinoxaline 7b. Additionally, we also prepared the dihydropyrazine 8 starting
from ethylene diamine and benzoin 3 in 96% of yield with almost total selectivity (96%), after
only 30 min of reaction time, which by oxidative aromatization gradually lead to pyrazine 9
(60%, 4h).
In summary, the obtained results are especially relevant because it supposes the easy
technological valorization of the biomass into metal-free acidic porous carbons able to
efficiently catalyze the synthesis of quinoxalines and related compounds from diamines and
different -hydroxi ketones, under aerobic conditions.

3.3.
Mechanistic considerations
Based on our experimental results and considering other alternative reaction mechanism for
the transformation of o-phenylendiamine 2 and benzoin 3 into quinoxaline 1 catalyzed by acid
carbon catalysts, we carried out a succinct theoretical study by analyzing each step as an
elementary reaction in gas phase. We select the most reduced models containing exclusively
acid functions simulating each series of catalysts (Chart 1) but also the simplest -hydroxi
ketone for decreasing the computational cost. Obviously, these models do not represent a
realistic situation only allowing to analyze the effect of different acid centers in the reaction
without any confinement restrictions. It seems reasonable to think that the reaction could
follow alternative routes (Paths b or c) as shown in scheme 4. Additionally, we firstly
investigate the formation of the intermediate compound I6 from I2 in absence or in the
presence of the catalysts, particularly using the N-model, following the path a, as reported
(Scheme 4, Path a). Thus, the computed transition structures for this transformation in
absence (TSI2-I6) or using N-model (N-TSI2-I6) reveal that the catalyst would not participate in the
cyclization reaction giving high and similar relative free energy values,G#, 73.5 and 73.8 Kcal
mol-1 respectively. However, based on our experimental results the reaction is accelerated in
the presence of the investigated catalysts. At this regard, it is important to note that the
reaction in absence of any catalyst gives 50% of conversion with 36% of selectivity to 1 after 4
h of reaction time. Considering the formation of intermediate compounds I2 and I3 in the
presence of the investigated carbon catalysts, as detected by NMR, we could assume that the
preferred operative pathways for the formation of quinoxaline 1 could be the path b or c
instead of the most common reported path a (Scheme 4).

Scheme 4. Alternative reaction route for the synthesis of quinoxaline 1 from o-
phenylendiamine 2 and -hydroxi ketones.
Figure 7 depicts both pathways (Path b or c) for the four systems, uncatalyzed and the models
selected for simulating the active centers of the investigated carbon catalysts.
uncatalyzed
N_model
S_model
P_model
(a)
100.0
80.0
60.0
40.0
20.0
0.0
-20.0
-40.0
Reactants I1
I2
I3
I4
I5
I6
Reaction coordinate

uncatalyzed
N_model
S_model
P_model
(b)
100.0
80.0
60.0
40.0
20.0
0.0
-20.0
-40.0
Reactants I1
I2
I3
I4
I5
I7
Reaction coordinate
Figure 7. Free-energy profiles computed for the uncatalyzed, N, S and P-models catalytic
systems following (a) path b and (b) path c in the formation of quinoxaline 1.
Both pathways for the uncatalyzed reaction show that the relative free-energy values for all
the computed transition structures are the highest for each elementary step (close to 40 kcal
mol-1) more than those found for the catalyzed processes. Two interesting findings were
observed for the uncatalyzed reaction: (1) initial reactant complex is formed, in which both
reagents interact by intra- and intermolecular H-bonds, favoring the approaching of the
reagents, and (2) the involvement of one water molecule in TSI1-I2, TSI2-I3 and TSI3-I4 notably
decreasing the activation barrier (from 46.1 to 13.0 kcal mol-1 for TSI1-I2 and TSI1-I2w
,
respectively). This fact has been previously observed by the authors in dehydration and
heterocyclization reactions since water molecules can act as bifunctional acid-base catalyst
[43,44]. In the same context, the reactant complex formation is also observed for the catalyzed
processes in which the acidic proton of OH group in catalysts is H-bounded to the carbonyl
acceptor and the oxygen atom to the adjacent amine function.
In Figure 8 is shown the optimized transition states TSR-I1 for the first step of the reaction,
comprising the nucleophilic attack of an amine group from 2 to acetoin for the four
investigated systems. In all the cases it was observed the formation of CN bond
(1.722871.57017Å), simultaneously initiating the hydrogen migration from N2 to N4 in which

the O-H groups from the catalytic model is also involved, thus assisting the almost formation
of alcohol in I1. It is clear that the acidity plays an important role in the first step of the
reaction since the most acid model shows the most advanced TS (Figure 8c), supported by
experimental observations.
Figure 8. Optimized transition structures for the nucleophilic attack of 2 to acetoin. (a) TSR-I1 in
the absence of any catalyst (b) N-model; (c) S-models and (d) P-model. Relevant distances are
expressed in Å.
Comparing the energy profiles shown in the Figure 7, it is observed reduced free-energy values
(6.6-9.9 kcal mol-1 for TSR-I1 and 6.3-7.3 kcal mol-1 for TSI1-I2w) for the initial nucleophilic
additions and dehydrations in the presence of N and P-models, this decrease being more
pronounced in the case of S-model probably due to the higher acidity of the sulfonic acid
functions as expected. It is important to note that the intermediate I3 was not experimentally
observed when the reaction was carried out when using acetoin instead benzoin probably due
to the additional stabilization by the presence of phenyl groups. The main energy differences
were observed for the following elementary steps comprising successive imine-enamine and
enol-keto tautomerisms followed by heterocyclization to give I5, in which the energy barrier is

notable decreased in the presence of acid catalytic models (Figure 7). Barely energetic
differences exist for the dehydration of I5 to I6 or I7 for the uncatalyzed process. However, the
obtained results suggest that this feature is changed for the catalytic processes; while the
reaction in the presence of N-model seems to occur following path c, S- and P-models would
catalyze the reaction through path b. At this respect, traces of intermediate compound I7 were
exclusively detected in the 1H NMR spectra for the reaction catalyzed by XN (N-model), by the
presence of signal centered at  6.5 which could be assigned to the aromatic protons of the
dyhidroquinoxaline skeleton.
Finally, both intermediate compounds I6 or I7 could evolve to quinoxaline 1 by oxidative
aromatization in the presence of acid carbon catalysts under study. This fact is supported by
studies concerning the oxidative conversion of a great variety of compounds such as benzylic
and allylic alcohols and different nitrogen heterocycles including 2-arylimidazolines, indolines,
pyrazolines, 3,4-dihydropyrimidin-2(1H)-ones, among others, by using activated carbon–
molecular oxygen systems [45]. Hayashi suggests that surface area, micro- and mesoporosity
along with the amount of the oxygen functional groups play an important role in the oxidative
reaction. In this sense, the biomass-derived carbons herein reported with high porosity
development were found to be the most selective catalysts.
4.
Conclusions
We report herein for the first-time acid biomass-derived carbons able to efficiently catalyze
the synthesis of quinoxalines and related compounds. Using the biomass Hedychium
gardnerianum, an undesired invasive plant, as precursor it was easily to obtain ACs with highly
developed porosity by chemical activation with H3PO4. The obtained carbon material (RCB2)
presented very promising textural properties with a BET area of 2197 m2 g-1 and a pore volume
of 1.39 cm3 g-1, comprising both micro- and mesopores. The chemical activation step allowed
the simultaneous activation and functionalization of the carbon surface, with the introduction

of phosphorus base functional groups. These biomass derived carbons revealed high activity
and selectivity in the synthesis of quinoxaline 1 from o-phenylendiamine 2 and benzoin 3, 94%
conversion and 83% selectivity. Our experimental and theoretical results demonstrate that the
reaction is mainly controlled by a combination of acid strength of oxygen functions and
porosity. The reaction could follow an alternative pathway consisting of: (1) nucleophilic
addition between reactants and subsequent dehydration, (2) successive imine-enamine and
keto-enol tautomerisms, (3) heterocyclization followed by dehydration, and finally (4)
aromatization to give nitrogen heterocycles.
5.
Acknowledgements
This work has been supported by Spanish Ministry (CTM 2014-56668-R project) and Associate
Laboratory for Green Chemistry-LAQV which is financed by national funds from FCT/MCTES
(UID/QUI/50006/2019). Ines Matos thanks FCT for the Investigador FCT contract
IF/01242/2014/CP1224/CT0008
and
M.
Bernardo
for
post-doc
fellowship
(SFRH/BPD/93407/2013). Authors also thank to Dr. Helena Cristina Meneses e Vasconcelos
from Universidade dos Açores, for the vegetal precursor. I. Matos also thanks Fundacion
Carolina.