The selective oxidation of glycerol over metal-free photocatalysts: insights into the solvent effect on catalytic efficiency and product distribution

Article abstract of DOI:10.1039/d1cy00360g

Selective oxidation of glycerol to high value-added derivatives is a promising biomass conversion pathway, but the related reaction mechanism, in particular the solvent effect, is rarely studied. In this work, O-doped g-C₃N₄was used as a metal-free catalyst to catalyze the selective oxidation of glycerol in different solvents. It was found that solvents can affect both catalytic efficiency and product distribution. A series of controlled experiments and theoretical calculation were applied to attest that the difference in interaction between glycerol and catalysts in different solvents is the main factor: competitive adsorption and hydrogen bond network from water inhibit the adsorption and activation of glycerol on the catalyst surface and reduce the conversion efficiency, while in acetonitrile, the stronger adsorption makes the oxidation reaction continue to yield esters. Two reaction routes in different solvents over O-doped g-C₃N₄are proposed for the first time, which is helpful for people to better understand the related reaction mechanism.

Full text of DOI:10.1039/d1cy00360g

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
The selective oxidation of glycerol over metal-free
photocatalysts: insights into the solvent effect on
catalytic efficiency and product distribution†
Cite this: DOI: 10.1039/d1cy00360g
a
a
a
Chengguang Yue, Mingming Fan, Agus Haryonob,b
*
*
Pingbo Zhang,
Yan Lenga and Pingping Jiang
a
Selective oxidation of glycerol to high value-added derivatives is a promising biomass conversion pathway,
but the related reaction mechanism, in particular the solvent effect, is rarely studied. In this work, O-doped
g-C3N4 was used as a metal-free catalyst to catalyze the selective oxidation of glycerol in different
solvents. It was found that solvents can affect both catalytic efficiency and product distribution. A series of
controlled experiments and theoretical calculation were applied to attest that the difference in interaction
between glycerol and catalysts in different solvents is the main factor: competitive adsorption and
hydrogen bond network from water inhibit the adsorption and activation of glycerol on the catalyst surface
and reduce the conversion efficiency, while in acetonitrile, the stronger adsorption makes the oxidation
reaction continue to yield esters. Two reaction routes in different solvents over O-doped g-C3N4 are
proposed for the first time, which is helpful for people to better understand the related reaction
mechanism.
Received 28th February 2021,
Accepted 19th March 2021
DOI: 10.1039/d1cy00360g
rsc.li/catalysis
promising high value-added conversion method. Glycerol can
be selectively oxidized to a variety of fine chemicals with high
Introduction
Due to the global energy crisis and environmental pollution,
the development of green renewable energy is very rapid. As a
green and renewable alternative to fossil fuels, biodiesel has
become a popular biomass energy source, and its output has
increased year by year. For every 10 tons of biodiesel
produced, 1 ton of glycerol is produced as a by-product; as a
result, the rapid development of the biodiesel industry has
made the already saturated glycerin market even more
sluggish.1 The price of glycerol is continuously falling,
making glycerol purification unprofitable; many factories
choose to treat crude glycerin directly as waste, which not
only leads to the waste of biomass resource, but also may
cause environmental pollution. Therefore, the development of
high value-added conversion methods for glycerol has become
one of the current research hotspots in the field of biomass
conversion.2
added values such as glyceraldehyde, glyceric acid,
dihydroxyacetone, and hydroxypyruvate.3 Most studies on the
aerobic oxidation of glycerol are focused on the traditional
heterogeneous catalysis with noble metals as the active center
of catalysts such as the Pt–Bi/C and Au/metal oxidation.4–8
Considering that noble metal catalysts have problems such
as easy leaching and deactivation, and noble metals are
scarce and expensive, the development of non-precious metal
or metal-free catalysts with high activity shows better
development prospects. In recent years, researchers have
found that while using semiconductor materials with suitable
valence band positions (such as Bi2WO6, Bi/Bi3.64Mo0.36O6.55
,
and BiVO4),9–11 glycerol can achieve photocatalytic aerobic
oxidation on non-precious metal catalysts. However, there
seems to be no report about the photocatalytic selective
oxidation of glycerol using metal-free catalysts.
Among the reported high value-added conversion paths of
glycerol, selective oxidation of glycerol is considered to be a
Metal-free catalysts such as carbon quantum dots, carbon
nanotubes, graphenes, and graphitic carbon nitride are
considered to be an important part of sustainable chemistry,
considering they are rich in resources and stable in structure,
and the unique physical and chemical properties of metal-
free catalysts make them comparable to metal-containing
catalysts in the catalytic fields of organic oxidation, hydrogen
evolution, oxygen reduction, etc.12–15 It was reported that
N-doped carbon nanotubes were applied as metal-free
catalysts to catalyze the selective oxidation of glycerol using
a The Key Laboratory of Synthetic and Biological Colloids, Ministry of Education,
School of Chemical and Material Engineering, Jiangnan University, Wuxi 214122,
P. R. China. E-mail: pingbozhang@126.com, fanmm2000@126.com
b Research Center for Chemistry, Indonesian Institute of Sciences (LIPI), Kawasan
Puspiptek, Serpong 15314, Indonesia
† Electronic supplementary information (ESI) available: Standard curves, LC-MS/
MS methods, optimum geometries and catalyst characterization methods and
results. See DOI: 10.1039/d1cy00360g
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
tert-butyl hydroperoxide as an oxidant, which can realize the
Carmine D'Agostino et al. found the competitive adsorption
between the solvent water and the substrates in the catalytic
oxidation of 1,3-propanediol and 1,4-butanediol, using
nuclear magnetic resonance (NMR) relaxation time
measurements.34,35 As a result, the conversion rate of the
alcohols decreased significantly in the presence of water. It
has been noticed that among the research results on the
selective oxidation of glycerol in aqueous solutions that have
been reported so far, the conversion rate of glycerol seems to
be relatively low; hence, we speculate that water which as
solvent in the glycerol selective oxidation system may have a
certain impact on the catalytic efficiency. Moreover,
considering that glycerol contains multiple hydroxyl groups,
it is theoretically feasible for other derivative reactions (such
as oxidative esterification) to occur in addition to oxidation
under anhydrous conditions. However, there is little
systematic study of elucidation of the solvent effect on the
glycerol oxidation system, which is of great significance for
the in-depth understanding of the reaction mechanism and
the design and optimization of subsequent catalysts.
Herein, O-doped g-C3N4 (OCNN) prepared by an in situ
calcination method was used as a metal-free catalyst to
catalyze the selective oxidation of glycerol for the first time.
The influence of solvent effect on the substrate conversion
efficiency and product distribution was systematically
investigated using different solvents and reference reactants.
Moreover, to better understand the difference in the glycerol
adsorption state on the OCNN surface under different solvent
conditions from the molecular level, we conducted density
functional theory (DFT) calculations using the Gaussian 16
software and independent gradient model (IGM) isosurface
analysis with Multiwfn 3.6. A new reaction path: one-step
oxidative esterification of glycerol to yield ester compounds
highly
selective
conversion
of
glycerol
into
dihydroxyacetone.16 Inspired by the above-mentioned work,
we supposed that a metal-free photocatalyst with a suitable
valence conduction band position for activating oxygen and
glycerol can be used to replace the nitrogen-containing
carbon nanotubes to generate reactive oxygen species from
oxygen under illumination, which may avoid the use of
environment-unfriendly tert-butyl hydrogen peroxide to
realize the selective oxidation of glycerol using a metal-free
catalyst and oxygen. This catalytic reaction system should be
more in line with the development concept of green and
sustainable chemistry.
Polymeric graphitic carbon nitride (g-C3N4) as a metal-free
semiconductor material, can achieve visible light response
and participate in a variety of photocatalytic reactions owing
to its suitable conduction and valence band edge positions,
such as water splitting, carbon dioxide reduction, alcohol
oxidation and degradation of organic pollutants.17–20
However, bulk g-C3N4 shows low photocatalytic efficiency due
to its small specific surface area and high recombination rate
of photo-induced electron–hole pairs.21 O-Doping is an
efficient approach to improve the photocatalytic efficiency of
g-C3N4; O-doped g-C3N4 has been widely applied in
photocatalytic hydrogen production, pollutant degradation,
organic matter transformation, hydrogen peroxide detection
and other fields and has shown good catalytic
performance.22–25 The reported oxygen-doping methods
include acid treatment, hydrothermal, hydrogen peroxide
oxidation and in situ calcination.26–29 Among them, the
preparation process of in situ calcination is simple and easy,
which does not need the participation of strong acids or
concentrated hydrogen peroxide. By mixing
a certain
proportion of oxygen-doping sources such as ammonium
acetate or other oxygen-rich organic substances in the
precursor, oxygen atoms can be doped in the CN framework
during the high-temperature calcination.
was
discovered
and
elucidated
combining
the
characterizations by EPR and LC-MS/MS, which is helpful for
people to better understand the reaction path and
mechanism of photocatalytic oxidation of glycerol.
In the heterogeneous catalytic conversion of biomass,
there are many factors that affect the results of the reaction,
among which the solvent effect has been proven to play an
important role on reaction rates and product selectivity.30
Existing studies have shown that in the conversion of glucose
into 5-hydroxymethylfurfural, as for increasing the amount of
catalyst and increasing the reaction temperature to increase
the conversion rate of the substrate, changing the reaction
solvent can often achieve a multiplier effect with half the
effort.31,32 In addition, the solvent effect can influence the
distribution of products by modifying a solvent to selectively
stabilize or destabilize a given product. Deng et al. discovered
that the addition of water to the oxidative system can change
the products of selective oxidative esterification of
Experimental
Preparation of OCNNs
OCNNs were prepared according to the procedure reported in
the literature29 with some modifications. First, 10 g urea was
mixed with different weights of ammonium acetate to get the
precursors; the mixture was ground for 15 min and
transferred into a 30 mL porcelain crucible with a cover and
sealed with a tinfoil. Then, the sealed crucibles were calcined
at 550 °C for 3 h at a heating rate of 10 °C min−1 in a muffle
furnace. The yellow solid in the crucible was taken out and
ground into powder when the muffle furnace was naturally
cooled to room temperature. These yellow powders were
named OCN-1, OCN-2, OCN-3 and OCN-4, whose precursors
contained different amounts of ammonium acetate (0.2, 0.4,
0.6 and 0.8 g, respectively). Finally, the OCNs were moved
into the porcelain crucible without any cover or tinfoil, and
heated to 500 °C for 2 h at a ramping rate of 2 °C min−1 in
5-(hydroxymethyl)furfural
from
dimethyl
furan-2,5-
dicarboxylate to methyl
5-(hydroxymethyl)furan-2-
carboxylate.33 In other polyol catalytic oxidation systems that
own a similar structure to glycerol, researchers have found
that water as
a solvent affects the catalytic efficiency.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
the muffle furnace. The obtained nanosheets were named
OCNN-1, OCNN-2, OCNN-3 and OCNN-4, whose precursors
contained different amounts of ammonium acetate (0.2, 0.4,
0.6 and 0.8 g, respectively). In addition, the bulk g-C3N4 (CN)
and g-C3N4 nanosheets (CNNs) were prepared by the same
method as OCNs or OCNNs without the addition of
ammonium acetate.
substrates (glycerol, H2O and CH3CN) were fully optimized,
and optimum geometries are shown in Fig. S4.† The SMD
implicit solvent model41 was used to investigate the influence
of the solvent in this system. One adsorbate molecule was
introduced to the adsorption site of g-C3N4 to determine the
adsorption configuration. The adsorption energy (Ea) of the
adsorbed molecule was calculated using the following
formula:
Photocatalytic oxidation of glycerol and propanediol
Ea = Ecom − (Ead + Esub
)
(3)
Typically, 10 mg catalyst was suspended in 5 mL of glycerol
(or propanediol) acetonitrile solution (50 mmol L−1) in a 25
mL round-bottomed flask. After the solution was bubbled
with O2 for 30 min, the flask was sealed with a balloon that
was prefilled with O2 and placed on the magnetic agitator
with cold trap to maintain the temperature around 25 °C.
The reaction solution was illuminated using a 350 W Xe lamp
(light intensity, 150 mW cm−2) for a certain period. After the
illumination, the mixture was centrifuged at 9500 rpm for 5
min to remove the catalyst particles; the suspension was
collected and filtered through a 0.45 μm nylon syringe filter.
Then, the solution was analyzed using a high-performance
liquid chromatograph (HPLC, Agilent 1100LC) equipped with
a RID, and a Zorbax SAX column (4.6 mm × 250 mm, Agilent)
was used with a mixed solution of H3PO4 (0.5% v/v) in water/
acetonitrile (1/2, v/v), (0.4 mL min−1) as the eluent at 35 °C,
which was performed according to the procedure reported in
the literature with some modifications.6 The amounts of
consumed reactants and produced products were quantified
by an external calibration method. The conversion of glycerol
(or propanediol) and the selectivity of the main products were
calculated using the following equations:
where Ecom, Ead and Esub stand for the total energies of the
complexes, adsorbent and substrate, respectively. In order to
more intuitively analyze the adsorption of glycerol on g-C3N4
in different solvents, the interaction region and strength
between substrate molecules and g-C3N4 were calculated
using an independent gradient model (IGM) isosurface
analysis42 with Multiwfn 3.6.43 The IGM interaction regions
and color-mapped isosurface graphs were obtained using the
VMD 1.9.3 program.44
Radical species trapping experiments
The mechanism of photocatalysis was investigated by a
radical species trapping experiment to determine the main
active species, which involved the photocatalytic oxidation
process of glycerol. Different radical scavengers such as
isopropanol, benzoquinone and furfuryl alcohol, which act as
−
the scavenger of ˙OH, ˙O2 and 1O2 respectively, were added
to the photocatalytic oxidation system to detect the influence
of the active species in the conversion process of glycerol
under identical conditions. The change in glycerol conversion
was taken as the important degree to evaluate the active
species.
Conversion = [(C0 − C)/C0] × 100%
(1)
(2)
Selectivity(main product) = [Cm/(C0 − C)] × 100%
H2O2 determination
We referred to the literature and used potassium titanium(IV)
oxalate as a detection agent to quantitatively analyze the
where C0 is the initial concentration of glycerol (or
propanediol), and C and Cm are the concentrations of
glycerol (or propanediol) and the main products, respectively,
at a certain time after the photocatalytic reaction; the related
standard curves of the detected substances were determined
and shown in Fig. S1.† Besides, the chemical structures of
some unknown products were confirmed by LC-MS/MS (AB
SCIEX, X500R), and the details are shown in section 2 in the
ESI.†
Simulation models and methods
For better studying the difference in the adsorption of
glycerol on the OCNN surface under different solvent
conditions, density functional theory (DFT) calculations were
performed at the level of M062X/6-311 + G(d)36–38 using the
Gaussian 16 software.39 A dispersion correction for density
functional theory (DFT-D3)40 was adopted in the calculations
for accurately describing the interactions. Before the DFT
calculation, the structures of g-C3N4 (Cmc21 space group) and
Fig.
1 Conversion of glycerol over different catalysts (reaction
conditions: 5 mL 50 mM glycerol acetonitrile solution, 10 mg catalyst,
O2, 20 °C, 350 W Xe lamp, 1 h).
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
H2O2 content in the system.45,46 Briefly, 2 mL of the solution
after 3 hours of reaction was taken out, the catalyst powder
was removed using a 0.45 μm nylon syringe filter, 2 mL of
potassium titanium oxalate solution (50 mmol L−1) was
added, and then 1 mL of ultrapure water was added to dilute
the solution to 5 mL. The absorbance of the solution was
obtained using an ultraviolet-visible spectrophotometer
(Jinghua Instruments 752) at 400 nm. The H2O2
concentration was quantified by an external calibration
method, and the standard curve of absorbance and hydrogen
peroxide concentration is shown in Fig. S5.†
adsorption strength of the substrate molecule on the surface
of catalyst is related to the polarity, we believe that the strong
polar solvent formed competitive adsorption with glycerol
molecule on the surface of OCNN and blocked the reaction
sites.
In order to verify the above-mentioned inference, we
added a small amount of water into the glycerol acetonitrile
solution to observe the effect of addition of different volumes
of water on the glycerol conversion. The results are shown in
Fig. 2b; as the volume of water increased (50, 200 and 500
μL), the conversion rate of glycerol was significantly reduced.
In addition, we conducted the catalyst adsorption test to
compare the adsorption of glycerol in acetonitrile and water
over the same quality of the OCNN-2 catalyst (Fig. 2c). The
results indicated that OCNN-2 adsorbed more glycerol
molecules in the glycerol acetonitrile solution (3.9% in
acetonitrile to 0.5% in water). The above-mentioned
experimental results verify our speculation that strong polar
solvents, such as water, can compete with glycerol molecules
for adsorption on OCNN-2. The competitive adsorption will
affect the adsorption and activation of glycerol molecules,
which, in turn, reduces the catalytic efficiency of the reaction
system. In addition, we compared the adsorption of glycerol
on the CN and OCN-2 catalysts with similar specific surface
areas (as shown in Fig. S6†). The results indicate that
O-doping can increase the adsorption of glycerol on the
surface of the catalyst. This may be due to the fact that O has
a greater electronegativity than N, which facilitates the
adsorption of glycerol molecules.
Considering that glycerol, as a polyol, has a complicated
hydrogen bond network in its aqueous solution, it is
generally believed that the existence of the hydrogen bond
network will also affect the adsorption and activation of the
substrate. Therefore, we speculated that in a glycerol aqueous
solution, the existence of the hydrogen bond network is also
a factor leading to the lower catalytic efficiency of the
reaction system. To verify this conjecture, we used
1,2-propanediol (1,2-PG) and 1,3-propanediol (1,3-PG) as
reference substrates, which possess fewer hydroxyl groups
and thus the strength of hydrogen bond network in their
aqueous solution is believed to be weaker than that of
glycerol. The difference in reactivity between glycerol and
propylene glycol in different solvents (acetonitrile and water)
was investigated, and the results are shown in Fig. 2d. The
results indicated that among the three aqueous alcohol
solutions, the conversion rate of glycerol was the lowest and
the conversion rate of 1,3-PG was the highest, which just
showed a negative correlation with the strength of the
Results and discussion
First, we evaluated the activity of the prepared catalysts in a
glycerol acetonitrile solution, and the results are shown in
Fig. 1. It is clear that O-doping significantly improved the
catalytic performance of g-C3N4 in this reaction system and
the OCNN-2 catalyst showed the best catalytic performance
(43.2% glycerol conversion in
1
h).
A
series of
characterizations were performed to find out the effect of
O-doping on the structure and properties of OCNN, as shown
in section 4 in the ESI.† The results indicated that OCNN-2
has a thinner sheet structure, larger specific surface area,
higher light absorption ability and lower photocarrier
recombination rate, which is consistent with the results of
the reported work,25 and these changes are helpful to
improve the catalytic performance.
The reactivity of glycerol on OCNN-2 in different solvents
(acetonitrile, N,N-dimethylformamide, methanol and water)
was explored, and the results are shown in Fig. 2a. It can be
seen that as the polarity of the reaction solvent increases, the
conversion rate of glycerol decreases significantly (43.2% in
acetonitrile to 4.4% in water). Considering that the
Fig. 2 (a) Glycerol conversion in different solvents. (b) Influence of
adding different volumes of water on the reaction activity. (c)
Adsorption of glycerol over OCNN in different solvents. (d) Alcohol
conversion in acetonitrile or water (reaction conditions: 5 mL 50 mM
alcohol solution, 10 mg OCNN-2, O2, 20 °C, 350 W Xe lamp, 1 h for (a)
and (b) and 3 h for (d)).
Scheme
system.
1
Presumed reaction path of the esters in the reaction
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
Table 1 Results of photocatalytic oxidation of glycerol and propylene glycol in the aqueous solution
Reactant
Conversion (%)
Products concentration (mM L−1) and selectivity (%)
Glycerol
1,2-Propanediol
1,3-Propanediol
6.0
14.5
23.7
DHA 1.8 (62.6)
LA 5.6 (76.8)
3-HPA 16.0 (85.5)
GLAD 0.5 (15.7)
Others 21.7
Others 23.2
Others 14.5
Reaction conditions: 5 mL 50 mM alcohols aqueous solution, 10 mg cat., O2, 20 °C, 350 W Xe lamp, 3 h.
hydrogen bond network in the aqueous solutions of the three
alcohols. This negative correlation indicated that the
hydrogen bond network did affect the reactivity of polyols in
the system. Another interesting phenomenon was that the
glycerol acetonitrile solution showed the highest conversion
rate, while the conversion rate of the propylene glycols was
basically the same. This phenomenon indicated that the
reactivity of alcohols on the OCNN surface was related to the
polarity of its molecular structure (glycerol molecules have
more hydroxyl groups than propylene glycol) without the
existence of solvent competitive adsorption or hydrogen bond
network in polyol acetonitrile solution, which proved that the
adsorption activation of hydroxyls on the catalyst surface is a
key step in the photocatalytic oxidation of glycerol and
propylene glycol.
Except the solvent effect on the glycerol conversion rate, it
was found that different solvents also affected the
distribution of oxidation products in the polyol reaction
system, particularly for glycerol. We detected several ester
compounds in the reaction system of the glycerol acetonitrile
solution and completed the preliminary characterization of
the new compounds using LC-MS/MS for the first time. The
measured molecular weight was basically consistent with the
calculated molecular weight (Table S1†). Moreover, the
compound structural formulas of these esters were confirmed
using the MS/MS fragments of the parent ions (Fig. S2†), and
we presumed that these esters can achieve progressive
transformation through the cleavage of carbon–carbon
bonds, as shown in Scheme 1.
Table 1 shows the results of photocatalytic oxidation of
glycerol and propylene glycol aqueous solution. The main
products of glycerol were dihydroxyacetone (DHA, 62.6% in
selectivity), glyceraldehyde (GLAD, 15.7% in selectivity), and
hydroxypyruvaldehyde (HPAD, the presence of this substance
was confirmed by LC-MS as shown in Fig. S3†). The oxidation
products of propylene glycol were mainly hydroxypropionic
acid compounds (lactic acid (LA, 76.8% in selectivity) for
1,2-PG and 3-hydroxypropionic acid (3-HPA, 85.5% in
selectivity) for 1,3-PG). It should be noticed that the oxidation
products of glycerol were aldehydes and ketones, and no
acidic compound was found. We speculated that the
existence of the strong hydrogen bond network in the
glycerol aqueous solution promoted desorption of the
products and prevented the further oxidation of glycerol-
oxidized derivatives to acids in this reaction system.
Different from the high selectivity of C3 products in the
aqueous solution, the oxidation products of the alcohol
acetonitrile solution showed more C2 and C1 products
generated by the cleavage of the carbon–carbon bond. We
recorded the substrate conversion rate and product
distribution as the reaction time increased in the polyol
acetonitrile solution, as shown in Fig. 3. The composition of
the oxidation products in the glycerol acetonitrile solution
was relatively complicated (Fig. 3a). In addition to the above-
mentioned ester products, we discovered other glycerol
oxidation derivatives such as formic acid (FA), glycolic acid
(GLCA), dihydroxyacetone (DHA), glyceraldehyde (GLAD),
glyceric acid (GLA) and hydroxypyruvaldehyde (HPAD). It can
be seen that as the reaction time prolonged, the
concentration of FA in the system gradually increased. The
concentration of DHA and GLCA increased within 2 h and
then decreased, indicating that DHA and GLCA were further
oxidized to other products. Although the presence of GLA
and GLAD were detected in the system, the content was very
low and the concentration change was not obvious,
Table 2 Adsorption energy of the substrates on g-C3N4 in different
solvents
Solvent
Water
Substrate
Adsorption energy (kJ mol−1
)
Fig. 3 Results of photocatalytic oxidation of glycerol (a), 1,2-PG (b),
and 1,3-PG (c) in the acetonitrile solution and stability test of OCNN-2
in the glycerol acetonitrile solution (d) (reaction conditions: 5 mL 50
mM alcohol acetonitrile solution, 10 mg cat., O2, 20 °C, 350 W Xe
lamp).
H2O
−24.6
−49.8
−7.7
Glycerol
CH3CN
Glycerol
Acetonitrile
−45.2
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
indicating that GLA and GLAD were intermediate products in
the glycerol oxidation process and will be quickly converted
into other compounds. It is worth noting that
hydroxypyruvate was not detected in the system, and hence,
we speculated that the ester 2-hydroxy-3-((3-hydroxy-2-
oxopropanoyl)oxy)propanoic acid was obtained from HPAD,
which acted as the precursor. The main oxidation products of
propylene glycols in the acetonitrile solution were
hydroxypropionic acids (LA for 1,2-PG and 3-HPA for 1,3-PG)
in the beginning. Then, C2 products (acetic acid (AA) and
ethanol (ET)) and C1 products (formic acid) were formed
owing to the peroxidation of the main products as the
reaction preceded; the details are shown in Fig. 3b and c.
Finally, we tested the stability of the OCNN-2 catalyst, as
shown in Fig. 3d. The result indicated that OCNN-2 still
maintained more than 80% catalytic activity after repeated
use for five times.
In brief, in this glycerol photocatalytic reaction system,
when water was used as a solvent, the adsorption and
activation of glycerol substrate molecules on the catalyst
surface was inhibited due to the existence of competitive
adsorption and hydrogen bond network, which made the
catalytic efficiency low, while the timely desorption of the
oxidation product in the aqueous solution avoided the
peroxidation of the oxidation product, which made the
selectivity of C3 product (DHA) relatively high. As for the
glycerol acetonitrile system, due to the efficient adsorption of
substrate molecules on the surface of OCNN-2, the catalytic
efficiency became higher, but the desorption of oxidation
products in the weakly polar solvent acetonitrile solution was
relatively slow, which resulted in the C3 oxidation products
participating in the next esterification or peroxidation
reaction to generate ester compounds or C2 and C1 products.
To better understand the difference in the glycerol
adsorption state on the OCNN surface under different solvent
conditions from the molecular level, we conducted density
functional theory (DFT) calculations using the Gaussian 16
software32 to figure out the adsorption energy of the
substrates on g-C3N4 in different solvents. The related results
based on the interaction models (as shown in Fig. S4†) are
given in Table 2. As we can see, the adsorption energy
Fig. 5 (a) Results of controlled experiments using different active
oxygen radical scavengers. (b) EPR spectra of the singlet oxygen and
superoxide radicals over OCNN-2 (isopropanol (IPA) as a scavenger for
hydroxyl radicals; furfural alcohol (FFA) as a scavenger for singlet
oxygen; p-benzoquinone (BQ) as a scavenger for superoxide radicals).
between the CH3CN molecule and g-C3N4 was significantly
smaller than that of glycerol in the acetonitrile solvent, which
indicated that CH3CN molecules could not form competitive
adsorption with glycerol molecules on the catalyst surface;
while in water, due to the presence of the hydroxyl group
with strong polarity in the structure of H2O molecules, the
interaction between H2O and g-C3N4 is significantly greater
than that of CH3CN but less than that of glycerol, but it
should be noticed that the volume of glycerol molecules is
significantly larger than that of H2O; in other words, the
adsorption energies of four or five H2O molecules
(approximately equal to the size of one glycerol molecule)
must be stronger than that of one glycerol molecule. This
result indicated that it is possible to form competitive
adsorption between solvent water and glycerol molecule on
the catalyst surface.
Additionally, the weak interactions between the substrates
and g-C3N4 were calculated using the independent gradient
model (IGM) approach,35 and the results are shown in Fig. 4.
As observed, when water was used as the solvent, the
interactions between H2O molecules and g-C3N4 (Fig. 4a)
involved hydrogen bonding (shown in blue) and van der
Waals forces (shown in green), while the interaction between
the glycerol molecule and g-C3N4 (Fig. 4b) only involved van
der Waals forces (shown in green). This result indicated that
the interaction of H2O and g-C3N4 is stronger than that of
the glycerol molecule in water. When the solvent water was
replaced with acetonitrile, the interaction between the CH3-
CN molecule and g-C3N4 (Fig. 4d) was much weaker, which
was consistent with the calculated result of adsorption
energy. Without the interference of strong polar solvent
water, the interaction between the glycerol molecule and g-
C3N4 (Fig. 4e) showed a stronger interaction (hydrogen
bonding interaction, shown in blue) besides van der Waals
interaction (shown in green). The specific distinction of the
Table 3 Concentration of H2O2 during the photocatalytic reaction in
acetonitrile
Fig.
4 IGM isosurface color maps of different adsorption
Substrate
Glycerol
1,2-PG
1,3-PG
configurations: (a) H2O and (b) glycerol on g-C3N4 in water and (c) the
enlargement of the circle in (b); (d) CH3CN and (e) glycerol on g-C3N4
in acetonitrile and (f) the enlargement of the circle in (e).
Conversion (%)
90.7
1117.1
70.4
671.1
69.6
671.1
CH O (μmol L−1
)
2
2
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
While in the acetonitrile solution, glycerol molecules strongly
adsorbed on the catalyst surface, which were oxidized to form
HPAD and GLAD, which could not desorb immediately and
further the esterification reaction occurred to generate
2-hydroxy-3-((3-hydroxy-2-oxopropanoyl)oxy)propanoic acid or
other compounds.
Conclusions
Scheme
2 Possible photocatalytic reaction routes for glycerol
A green and sustainable conversion route for the selective
oxidation of glycerol and its polyol derivatives at ambient
temperature and atmosphere with molecular oxygen as the
oxidant on a metal-free catalyst (O-doped g-C3N4) under
illumination was provided in this work. Through a series of
controlled experiments and theoretical calculation, we
systematically studied the solvent effect on this catalytic
reaction system at the molecular level. It was found that the
interaction between glycerol and the catalyst surface was
significantly different in different solvent environments,
leading to different conversion efficiencies and product
distributions. Through the active radical capture experiments
and EPR characterization, we investigated the reaction
mechanism of this system, and singlet oxygen and
superoxide radical were proved as the main reactive oxygen
species. The oxidation reaction path of glycerol in different
solvents was given. In particular, it has been found for the
first time that the oxidative esterification reaction can occur
to generate new ester compounds in the glycerol acetonitrile
solution. These findings should have certain reference values
for the subsequent design and development of related
catalysts.
oxidation over OCNN-2 in water and acetonitrile.
interactions between the glycerol molecule and g-C3N4 in
water or acetonitrile can be clearly seen in Fig. 4c and f; this
result indicated that glycerol showed a stronger interaction
on the catalyst surface in the acetonitrile solvent, which was
beneficial for the adsorption and activation of glycerol
molecules and improve the conversion of glycerol.
In order to ascertain the active oxygen species playing
major roles in the photocatalytic oxidation system of the
glycerol–acetonitrile solution, we carried out active radical
capture experiments and EPR characterization. In Fig. 5a, the
addition of furfural alcohol (singlet oxygen scavenger) and
p-benzoquinone (superoxide radical scavenger) resulted in a
decrease in the conversion rate of glycerol, while the addition
of isopropanol as a hydroxyl radical scavenger showed almost
no change in the conversion rate of glycerol. At the same
time, in the EPR spectra of the OCNN-2 catalyst (Fig. 5b),
obvious signals of singlet oxygen and superoxide radicals
were observed, indicating that these two active oxygen species
were produced in the system. The above results showed that
singlet oxygen and superoxide radicals participated in the
photocatalytic oxidation process of glycerol and acted as
main reactive oxygen species in this system.
Conflicts of interest
There are no conflicts to declare.
To figure out the product of molecular oxygen in this
catalytic system, we carried out the H2O2 determination in
the solution after the reaction, and the results are given in
Table 3. It can be seen that there is a certain amount of H2O2
in the polyol acetonitrile solution after the reaction, and the
content of H2O2 is proportional to its conversion rate; hence,
we speculate that oxygen will become H2O2 after participating
in the reaction. With reference to the existing reports in the
literature,47,48 we believe that H2O2 is produced via two-
electron reduction of O2. In addition, the concentration of
H2O2 in the system is relatively low, which may be due to the
faster decomposition of H2O2 in an anhydrous
environment.47
Combining the above-mentioned results, the possible
photocatalytic routes for glycerol oxidation over OCNN-2 are
proposed, as shown in Scheme 2. Glycerol molecules
adsorbed on the surface of the catalyst can be activated and
oxidized to oxidation products under the interaction of
photogenic holes and reactive oxygen species. In aqueous
solutions, the oxidation products (mainly DHA and GLAD)
quickly desorbed from the surface of the catalyst, thereby
avoiding the occurrence of peroxides or other side reactions.
Acknowledgements
The financial supports from the Natural Science Foundation
of China (NSFC) (No. 21978112) and MOE & SAFEA for the
111 Project (B13025) are gratefully acknowledged.
References
1 C. H. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. Lu, Chem.
Soc. Rev., 2008, 37, 527–549.
2 B. Katryniok, H. Kimura, E. Skrzyńska, J.-S. Girardon, P.
Fongarland, M. Capron, R. Ducoulombier, N. Mimura, S.
Paul and F. Dumeignil, Green Chem., 2011, 13, 1960–1979.
3 G. Dodekatos, S. Schünemann and H. Tüysüz, ACS Catal.,
2018, 8, 6301–6333.
4 Y. Wang, D. Yuan, J. Luo, Y. Pu, F. Li, F. Xiao and N. Zhao,
J. Colloid Interface Sci., 2020, 560, 130–137.
5 F.-F. Wang, S. Shao, C.-L. Liu, C.-L. Xu, R.-Z. Yang and W.-S.
Dong, Chem. Eng. J., 2015, 264, 336–343.
6 D. Liang, J. Gao, H. Sun, P. Chen, Z. Hou and X. Zheng,
Appl. Catal., B, 2011, 106, 423–432.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
7 H. Kimura, Appl. Catal., A, 1993, 105, 147–158.
8 P. Yang, J. Pan, Y. Liu, X. Zhang, J. Feng, S. Hong and D. Li,
ACS Catal., 2018, 9, 188–199.
9 D. Liu, J. C. Liu, W. Cai, J. Ma, H. B. Yang, H. Xiao, J. Li,
Y. Xiong, Y. Huang and B. Liu, Nat. Commun., 2019, 10,
1779.
31 H. Zhao, J. E. Holladay, H. Brown and Z. C. Zhang, Science,
2007, 316, 1597–1600.
32 Y. J. Pagán-Torres, T. Wang, J. M. R. Gallo, B. H. Shanks and
J. A. Dumesic, ACS Catal., 2012, 2, 930–934.
33 K. S. Kozlov, L. V. Romashov and V. P. Ananikov, Green
Chem., 2019, 21, 3464–3468.
10 Y. Zhang, N. Zhang, Z.-R. Tang and Y.-J. Xu, Chem. Sci.,
2013, 4, 1820–1824.
11 S. Zhao, Z. Dai, W. Guo, F. Chen, Y. Liu and R. Chen, Appl.
Catal., B, 2019, 244, 206–214.
12 D. S. Su, J. Zhang, B. Frank, A. Thomas, X. Wang, J.
Paraknowitsch and R. Schloegl, ChemSusChem, 2010, 3,
169–180.
13 J. Long, X. Xie, J. Xu, Q. Gu, L. Chen and X. Wang, ACS
Catal., 2012, 2, 622–631.
34 C. D'Agostino, M. R. Feaviour, G. L. Brett, J. Mitchell, A. P. E.
York, G. J. Hutchings, M. D. Mantle and L. F. Gladden,
Catal. Sci. Technol., 2016, 6, 7896–7901.
35 C. D'Agostino, T. Kotionova, J. Mitchell, P. J. Miedziak, D. W.
Knight, S. H. Taylor, G. J. Hutchings, L. F. Gladden and
M. D. Mantle, Chem. – Eur. J., 2013, 19, 11725–11732.
36 R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem.
Phys., 1980, 72, 650–654.
37 T. Clark, J. Chandrasekhar, G. W. Spitznagel and P. V. R.
Schleyer, J. Comput. Chem., 1983, 4, 294–301.
14 L. Dai, Y. Xue, L. Qu, H.-J. Choi and J.-B. Baek, Chem. Rev.,
2015, 115, 4823–4892.
38 Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
39 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A.
Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J.
Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P.
Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D.
Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B.
Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G.
Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K.
Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. J.
Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N.
Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K.
Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J.
Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R.
Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O.
Farkas, J. B. Foresman and D. J. Fox, Gaussian 16, Revision
C.01, Gaussian, Inc., Wallingford CT, 2019.
15 M. Z. Rahman, M. G. Kibria and C. B. Mullins, Chem. Soc.
Rev., 2020, 49, 1887–1931.
16 N. Gupta, O. Khavryuchenko, A. Villa and D. Su,
ChemSusChem, 2017, 10, 3030–3034.
17 X. Wang, S. Blechert and M. Antonietti, ACS Catal., 2012, 2,
1596–1606.
18 S. Cao, J. Low, J. Yu and M. Jaroniec, Adv. Mater., 2015, 27,
2150–2176.
19 Y. Zheng, L. Lin, B. Wang and X. Wang, Angew. Chem., Int.
Ed., 2015, 54, 12868–12884.
20 Y. Xu, Z. Zhang, C. Qiu, S. Chen, X. Ling and C. Su,
ChemSusChem, 2021, 14, 582–589.
21 W. J. Ong, L. L. Tan, Y. H. Ng, S. T. Yong and S. P. Chai,
Chem. Rev., 2016, 116, 7159–7329.
22 W. Zhu, N. Hao, J. Lu, Z. Dai, J. Qian, X. Yang and K. Wang,
Chem. Commun., 2020, 56, 1409–1412.
23 Y. Huang, L. Ning, Z. Feng, G. Ma, S. Yang, Y. Su, Y. Hong,
H. Wang, L. Peng and J. Li, Environ. Sci.: Nano, 2021, 8,
460–469.
24 J. Ma, D. Jin, Y. Li, D. Xiao, G. Jiao, Q. Liu, Y. Guo, L. Xiao,
X. Chen, X. Li, J. Zhou and R. Sun, Appl. Catal., B, 2021, 283,
119520.
40 S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys.,
2010, 132, 154104.
41 A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys.
Chem. B, 2009, 113, 6378–6396.
25 G. Q. Zhang, Y. S. Xu, C. X. He, P. X. Zhang and H. W. Mi,
Appl. Catal., B, 2021, 283, 119636.
26 H. Wang, S. Jiang, S. Chen, D. Li, X. Zhang, W. Shao, X. Sun,
J. Xie, Z. Zhao, Q. Zhang, Y. Tian and Y. Xie, Adv. Mater.,
2016, 28, 6940–6945.
42 C. Lefebvre, G. Rubez, H. Khartabil, J. C. Boisson, J.
Contreras-Garcia and E. Henon, Phys. Chem. Chem. Phys.,
2017, 19, 17928–17936.
43 T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580–592.
44 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics
Modell., 1996, 14, 33–38.
27 Q. Gao, J. Xu, Z. Wang and Y. Zhu, Appl. Catal., B, 2020, 271,
118933.
45 R. M. Sellers, Analyst, 1980, 105, 950–954.
28 W. Jiang, Q. Ruan, J. Xie, X. Chen, Y. Zhu and J. Tang, Appl.
Catal., B, 2018, 236, 428–435.
46 C. S. Pan and Y. F. Zhu, Environ. Sci. Technol., 2010, 44,
5570–5574.
29 F. Li, M. Han, Y. Jin, L. Zhang, T. Li, Y. Gao and C. Hu, Appl.
Catal., B, 2019, 256, 117705.
30 L. Shuai and J. Luterbacher, ChemSusChem, 2016, 9,
133–155.
47 Y. Shiraishi, S. Kanazawa, Y. Sugano, D. Tsukamoto, H.
Sakamoto, S. Ichikawa and T. Hirai, ACS Catal., 2014, 4, 774–780.
48 L. Shi, L. Yang, W. Zhou, Y. Liu, L. Yin, X. Hai, H. Song and
J. Ye, Small, 2018, 14, 1703142.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021