5-Hydroxymethylfurfural catalytic oxidation under mild conditions by Co (II), Fe (III) and Cu (II) Salen complexes supported on SBA-15: Synthesis, characterization and activity

Article abstract of DOI:10.1016/j.apcata.2017.08.035

In the present work, an alternative system of 5-hydroxymethylfurfural (HMF) oxidation was studied, in an effort to avoid the use of expensive precious metal catalysts, high pressures and systems not environmentally friendly. The catalysts tested were Co (II), Fe (II) and Cu (II) Salen complexes supported on SBA-15 (Co Salen/SBA-15, Fe Salen/SBA-15 and Cu Salen/SBA-15 respectively). The reaction was carried out in aqueous medium at room temperature (25 °C ± 1) and atmospheric pressure, using hydrogen peroxide (H₂O₂) as the only oxidant agent. The effect of various reaction conditions such as H₂O₂ content, oxidant agent amount, different transition metals present in the catalyst and temperature were evaluated. The HMF and the reaction products such as; 2,5-diformylfuran (DFF), 2,5-furandicarboxylic acid (FDCA), 5-hydroxymethyl-2-furancarboxylic acid (HMFCA) and 5-formyl-2-furancarboxylic acid (FFCA) were followed by high performance liquid chromatography (HPLC) measurements. The selectivity and reaction kinetics under different reaction conditions was studied. The different synthesized catalysts were extensively characterized by UV–vis spectroscopy, subsequently by FTIR, DR-UV–vis, AA and XPS spectroscopy and through nitrogen physorption, which was related to its catalytic activity. The successful synthesis of SBA-15 support was confirmed by its textural properties and STEM. From the experimental data obtained a reaction mechanism with parallel reactions was proposed founding that the oxidation route of the alcohol group in HMF is favored over of the oxidation of the aldehyde group on the other side of the molecule.

Full text of DOI:10.1016/j.apcata.2017.08.035

Accepted Manuscript
Title: 5-Hydroxymethylfurfural Catalytic Oxidation Under
Mild Conditions by Co (II), Fe (III) And Cu (II) Salen
Complexes Supported on SBA-15: Synthesis, Characterization
and Activity
Authors: Daniela Xul u´ Mart ´ı nez-Vargas, Javier Rivera De La
Rosa, Ladislao Sandoval-Rangel, Jorge Luis Guzm a´ n-Mar,
Marco A. Garza-Navarro, Carlos J. Lucio-Ortiz, David A. De
Haro-Del R ´ı o
PII:
S0926-860X(17)30416-7
DOI:
Reference:
http://dx.doi.org/10.1016/j.apcata.2017.08.035
APCATA 16390
To appear in:
Applied Catalysis A: General
Received date:
Accepted date:
11-8-2017
21-8-2017
Please cite this article as: Daniela Xul u´ Mart ´ı nez-Vargas, Javier Rivera De La
Rosa, Ladislao Sandoval-Rangel, Jorge Luis Guzm a´ n-Mar, Marco A.Garza-Navarro,
Carlos J.Lucio-Ortiz, David A.De Haro-Del R ´ı o, 5-Hydroxymethylfurfural Catalytic
Oxidation Under Mild Conditions by Co (II), Fe (III) And Cu (II) Salen Complexes
Supported on SBA-15: Synthesis, Characterization and Activity, Applied Catalysis A,
Generalhttp://dx.doi.org/10.1016/j.apcata.2017.08.035
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.

5-Hydroxymethylfurfural Catalytic Oxidation Under Mild Conditions by Co (II), Fe (III) And Cu
(II) Salen Complexes Supported on SBA-15: Synthesis, Characterization and Activity
1
1,2
1
Daniela Xulú Martínez-Vargas , Javier Rivera De La Rosa* , Ladislao Sandoval-Rangel , Jorge
1
2,3
1
Luis Guzmán-Mar , Marco A. Garza-Navarro , Carlos J. Lucio-Ortiz and David A. De Haro-Del
Río1
1Universidad Autónoma de Nuevo León, UANL, Facultad de Ciencias Químicas, Ave. Universidad
S/N, Cd. Universitaria, San Nicolás de los Garza, N. L., C.P. 64451, México. Tel.: +528180294000
ext 6357; fax: +52818029400 ext 6282.
2
Universidad Autónoma de Nuevo León, UANL, Centro de Innovación, Investigación y Desarrollo
en Ingeniería y Tecnología (CIIDIT), Km 10 de la nueva carretera al Aeropuerto Internacional de
Monterrey, PIIT Monterrey, 66600, Apodaca, Nuevo León, México.
3
Universidad Autónoma de Nuevo León (UANL), Facultad de Ingeniería Mecánica y Eléctrica
(FIME), Av. Universidad S/N, Cd. Universitaria, San Nicolás de los Garza, Nuevo León C.P. 66451,
México
Corresponding author: *javier.riverad@uanl.edu.mx
Graphical abstract
Highlights

Synergistic effect of Cu (II) and Fe(III) Salen catalysts supported on SBA-15 and H
HMF oxidation
2
2
O in



Cu Salen/SBA-15 catalyst active in HMF oxidation under mild conditions
Mechanism of HMF catalytic oxidation step by step
Higher selectivity to oxidation of alcohol group compared with the aldehyde group of HMF
molecule
ABSTRACT
In the present work, an alternative system of 5-hydroxymethylfurfural (HMF) oxidation was studied,
in an effort to avoid the use of expensive precious metal catalysts, high pressures and systems not
1

environmentally friendly. The catalysts tested were Co (II), Fe (II) and Cu (II) Salen complexes
supported on SBA-15 (Co Salen/SBA-15, Fe Salen/SBA-15 and Cu Salen/SBA-15 respectively).
The reaction was carried out in aqueous medium at room temperature (25 °C +1) and atmospheric
pressure, using hydrogen peroxide (H
2
O ) as the only oxidant agent. The effect of various reaction
2
conditions such as H content, oxidant agent amount, different transition metals present in the
2
O
2
catalyst and temperature were evaluated. The HMF and the reaction products such as; 2,5-
diformylfuran (DFF), 2,5-furandicarboxylic acid (FDCA), 5-hydroxymethyl-2-furancarboxylic acid
(HMFCA) and 5-formyl-2-furancarboxylic acid (FFCA) were followed by high performance liquid
chromatography (HPLC) measurements. The selectivity and reaction kinetics under different
reaction conditions was studied. The different synthesized catalysts were extensively characterized
by UV-vis spectroscopy, subsequently by FTIR, DR-UV-vis, AA and XPS spectroscopy and through
nitrogen physorption, which was related to its catalytic activity. The successful synthesis of SBA-15
support was confirmed by its textural properties and STEM. From the experimental data obtained a
reaction mechanism with parallel reactions was proposed founding that the oxidation route of the
alcohol group in HMF is favored over of the oxidation of the aldehyde group on the other side of the
molecule.
KEYWORDS
5
1
-hydroxymethylfurfural oxidation, Co(II) Salen SBA-15, Cu (II) Salen SBA-15, Fe (III) Salen SBA-
5.
1. INTRODUCTION
During the past decade, one of the global concerns has been the use of fossil fuels for the
production of energy and chemicals, due to their limited availability and the pollution caused for
their use [1, 2]. Nowadays, the production of both energy and chemicals requires alternative
resources and a more severe environmental control [2, 3]. Therefore, biomass sources are the most
promising alternative for the replacement of fossil resources, due to their abundance in nature and
low cost. About 75% of naturally available biomass is composed of carbohydrates, then, much
interest has been given to the development of efficient approaches to transform carbohydrates,
2

derived from biomass, into value-added chemicals [4, 5]. However, the use of biomass has caused
considerable controversy due to ethical issues, employing the use of edible products for biomass
such as corn and sugar cane, which are used as raw materials in bioethanol production [1, 6]. Due
to these precedents, research is being directed to the use of biomasses derived from agroindustrial
waste, uncultivable or inedible raw materials.
Catalytic oxidation, either homogeneous or heterogeneous is a very powerful tool in the synthesis of
fine chemicals [7, 8]. In such synthesis, compounds derived from furfural which have functional
groups in position five have played an important role [1]. Therefore, the 5-hydroxymethylfurfural
(HMF) is considered an important and versatile platform precursor for the synthesis of this kind of
compounds with applications in pharmaceutical and polymers industry, whose syntheses are
attempted from carbohydrates such as fructose, glucose, cellulose, etc. [2, 9-14]. HMF oxidation
leads to the formation of different products such as 5-hydroxymethyl-2-furancarboxylicacid
(HMFCA), 2,5-diformylfuran (DFF), 5-formyl-2-furancarboxylicacid (FFCA) and 2,5-
furandicarboxylicacid (FDCA) [7, 15]. FDCA is a promising biomass-derived chemical building block
with a great market potential, due to being considered as a monomer for synthesis of a new
polymer class, and also as a potential replacement of terephthalic acid, mainly used as precursor in
the production of polyethyleneterephtalate (PET) [1, 2, 16, 17]. DFF is considered as a highly useful
chemical with applications in pharmaceutics, fungicides, organic conductors and macrocyclic
ligands [7, 18-21]. It has attracted great attention as a monomer in the production of multifunctional
polymeric materials with special applications [22-26]. HMFCA and FFCA are intermediary products.
At the same time, the oxidative cleavage of the furan ring of HMF could take place; resulting in the
formation of succinic acid, 2-oxoglutaric acid, maleic acid and fumaric acid [27, 28]. Leuvinic acid
and formic acid could also be present due to HMF hydration in the presence of an acid catalyst [28,
2
9]. The classical synthesis of FDCA and DFF from HMF usually involves high pressure and
temperature, metal salts (generally based on noble metals) and organic solvents [7, 30], which
makes the process expensive and polluting [2, 31-33].
The use of expensive, noble metal-based catalysts such as Ru, Pt [34-38], Pd [35, 39-41], Rh and
Au [35, 36] in the HMF oxidation has been recently reported; Nie et al.[30] studied Ru, Pt, Pd, Rh
3

and Au [36] catalysts using activated carbon (C), Mg
CeO as supports; they found that the Ru/C catalyst showed the best performance with a DFF yield
of 96% at 110°C, high pressure of pure O (2 MPa) and toluene as solvent, known by its toxicity.
Antonyraj et al. [42] also found a high DFF yield of 97% with a 99% conversion of HMF employing
2 x
AlO , MgO, Al
2
O
3 2
, ZSM-5, TiO , ZrO
2
and
2
2
2
Ru/-Al O
3
and toluene as a solvent with 0.28 MPa of pure oxygen. Although other catalysts such
2+
2+
as VO /Cu on sulfonated carbon [7], copper complexes [43, 44], Mn (III) Salen complex [45] and
gold nanoparticles on hydrotalcite have also been studied and shown high selectivity to DFF or
FDCA, most of them have serious disadvantages such as being expensive and/or homogeneous, in
the presence of unfavorable conditions like organic solvents, high pressures, promoters and
different oxidizing agents (sometimes using more than one at a time). Vanadium species are known
for their toxicity and also oxidizing agents such as chromate (CrO 2
−), dichromate (Cr
O
2−) and
4
2
7
permanganate (MnO
4
−), that have been often employed to oxidize HMF [44]. Therefore, it is still
necessary to improve processes through non-expensive and readily available catalysts, which can
operate at more favorable conditions such as low pressure and temperature (which will decrease
costs and will make the processes safer and simpler) in addition to a decreased use of solvents and
harmful materials to health and the environment in the reaction system.
In the present study, HMF oxidation was investigated under mild conditions such as aqueous media
with H
2
O , which is considered as environmentally friendly and green oxidation reagent, with metal
2
(Co (II), Cu (II) and Fe (III)) Salen complex catalysts supported on SBA-15 (Co Salen/SBA-15, Cu
Salen/SBA-15 and Fe Salen/SBA-15) as a first study of this system as an alternative for the
synthesis of compounds of high added value, a system which has not been previously studied.
Salen metal complexes can be used to simulate the catalytic behavior of an enzyme [44], which
makes them attractive for their use in catalysis, although they have been used mainly in
homogeneous reaction systems. An extensive characterization of the heterogeneous catalysts
synthesized is reported in this work, also the effect of different conditions in the HMF reaction
system and a reaction mechanism is proposed.
2. EXPERIMENTAL
2.1 Materials and methods.
4

Pluronic P123 and tetraethyl orthosilicate (TEOS) were used to synthesize SBA-15. Salicylaldehyde
99%) and 3-aminopropyltriethoxysilane (99%) from Aldrich (St. Louis, MO, USA) was used for the
(
ligand synthesis and the cobalt (II) acetate tetrahydrate (>98%) and copper (II) acetate (98%) from
Aldrich (St. Louis, MO, USA) and iron (III) chloride hexahydrate (>99%) from Desarrollo de
Especialidades Químicas (San Nicolás de los Garza, N.L., México) as metallic precursors for the
formation of the respective metal Salen complex. Methanol (ACS Reagent 99.8%) from JT Baker
(Central Valley, PA, USA) was used as a solvent. 5-hydroxymethylfurfural ≥ 99%, 2,5 diformylfuran
(97%), 2,5-furandicarboxaldehyde (97%), 5-formyl-2-furoic acid and 5-hydroxymethyl-2-
furancarboxylic acid (Sigma-Aldrich, St. Louis, MO, USA), were used as standards for analytical
measures. Nanopure water was used as reaction media (50 mL). Hydrogen peroxide 30 p/V% (CTR
Scientific, Monterrey, N.L., México) also was added (50 μL) before the catalyst was incorporated.
In order to follow the oxidation process, the substrate (HMF) and major products of the expected
reaction (2,5 dyformylfuran (DFF), 2,5-furandicarboxaldehyde (FDCA), 5-formyl-2-furoic acid
(FFCA) and 5-hydroxymethyl-2-furancarboxylic acid (HMFCA)) were separated by High-
performance liquid chromatography (HPLC) using the HPLC Perkin Elmer Equipment with UV
detector and an Aminex HPX-87H 300 X 7.8 mm column (cross-linked resin hydrogen ionic form,
sulfonated divinyl benzene-styrene copolymer as support, 9 μm particle size) from Bio-Rad
(Hercules, California, USA). The mobile phase used was H
2
SO 5mM in water. All aqueous
4
solutions were prepared with nanopure water.
The column was conditioned every day before the analyses by flowing the mobile phase through
the column (1.2 mL/min) until no peaks or baseline noise were showed in the chromatogram.
Following this the sample was manually injected when the display marked 0.2 mL/min, in order to
control the migration time (t ) in each injection. 25 μL of the sample were consecutively injected
M
three times, in order to have an excess amount and to ensure repetitiveness of the injections. The
method used consisted of a 30 min run, a 1.2 mL/min flow and room temperature (25 °C), using the
previously described mobile phase. An UV detector was used with  = 250 nm. At the end of each
5

working day, the column was cleaned by flowing the mobile phase. Although this procedure
increased the reproducibility, slight shifts in t were observed and considered normal.
M
Sampling was conducted by taking 500 μL of the reaction solution (carried out in deionized water)
and centrifuged to remove the catalyst (6000 rpm for 1 min). Data acquisition was performed using
the TotalChrom software (Perkin Elmer, Shellton, CT, USA). Each mobile phase solution was
prepared daily and degassed by sonication for 15 min before use. The capillary was conditioned
with 5 mM of H
2
SO at least during 30 min.
4
2.2 Characterization.
The amount of metal loading (Co, Cu or Fe active site according to the respective metal Salen
complex on SBA-15) was quantified by Atomic absorption spectroscopy (AAS) using an AA 1
Thermo Jarrell Ash Scan (Franklin, MA, USA) with a hollow cathode lamp as the light source. For
Co analysis the lamp used was No: 62928162928-02, lamp current: 3 mA operating in background
Hiefje Smith (the signal is matched to the intensity of the background), : 240.7 nm pass band: 0.3
nm, Burnt head: long history, air-acetylene flame. Sensitivity of 0.044 absorbance = 1% absorption
about 0.05 μg/mL for the instrument with the parameters described. The standard of 1.2 μg/mL Co
gives an approximate reading of 0.1 A. Linear range of Co: linear concentration above 5 μg/mL
(aqueous solution using the instrumental parameters described above). For the measurment of Cu
the lamp used was No: 63041/63041-02, lamp current: 2.5 mA operating in background Hiefje
Smith (the signal is matched to the intensity of the background), : 324.7 nm pass band: 1.0 nm,
Burnt head: long history, air-acetylene flame. Sensitivity of 0.0044 absorbance = 1% absorption
about 0.03 μg/mL. The standard of 0.7 μg/mL Co gives an approximate reading of 0.1 A. Linear
range of Cu: linear concentration above 4 μg/mL (aqueous solution using the instrumental
parameters described above). Finally, for Fe the lamp used was No: 62810/62810-02, lamp current:
3
mA operating in background Hiefje Smith (the signal is matched to the intensity of the
background), : 248.3 nm pass band: 0.3 nm, Burnt head: long history, air-acetylene flame.
Sensitivity of 0.044 absorbance = 1% absorption about 0.04 μg/mL for the instrument with the
parameters described. The standard of 1.0 μg/mL Fe gives an approximate reading of 0.1 A. Linear
6

range of Fe: linear concentration above 5 μg / mL (aqueous solution using the instrumental
parameters described above). Fourier-transform infrared (FTIR) spectroscopy was performed using
a Nicolet 6700 FTIR spectrometer (Thermo Scientific, Waltham, MA, USA) with KBr disks and also
the Sprectrum One FTIR-ATR spectrometer (Perkin-Elmer, Waltham, MA, USA). UV–Vis spectra
were measured with a Genesys 10UV scanning (Thermo Electron Corporation, Madison, WI, USA)
spectrophotometer, which was operated in the region of 200–800 nm and used quartz cuvettes
(Fisher Scientific, Pittsburgh,PA,USA). UV–Vis diffuse reflectance spectra were recorded on a
Nicolet Evolution 300 PC UV–Vis spectrophotometer (Thermo Scientific, Waltham, MA, USA). N
2
sorption isotherms of silica SBA-15 were recorded in a BEL SORP mini II (BEL JAPAN INC.,
Toyonaka-shi, Osaka 561-0807 JAPAN), plotting relative pressure points of 0.05 – 0.99 and 0.99 –
.030 for adsorption and desorption, respectively. For these analyses, all samples were degassed
at 150 °C for 24 hrs. The measurements were performed using ultra high purity (UHP) grade N as
the adsorbent gas and liquid N at -196°C as the medium. The Barrett-Joyner-Halenda (BJH) and
0
2
2
Brunauer, Emmet and Teller (BET) methods were used to calculate the pore diameter and surface
area, respectively. Surface analyses were performed using an X-ray photoelectron spectrometer K-
Alpha Thermo Scientific, with X-ray spot of 400 nm and flood gun for charge compensation.
Samples were prepared in discs with KBr. Measurements were performed after ion beam etch with
argon ion gun (2 keV during 20 s with 20–40nm estimated thickness); with 400 μm spot size, 50.0
eV pass energy and energy step size of 0.100 eV. The Shirley method was used to estimation of
the background signals in XPS spectra. For all samples the C1s was taken as reference (284.83
eV).
2.3 Preparation of metal Salen catalysts.
All catalysts were prepared using a procedure previously reported by Rajabi et al. [46, 47] and
modified according to previous reports [8] an schematic representation can be seen in Fig 1. First,
0
.217 g (2 mmol) of salicylaldehyde was dissolved in 200 mL of absolute methanol and mixed with
.473 g (2 mmol) of 3-aminopropiltriethoxysilane (APTES), upon which the solution immediately
0
turned yellow. After stirring for 3 h, the metal precursor was added; 0.254 g (1 mmol) of Co
7

(
CH
3
COO)
2 2
·4H O, 0.185 g (1 mmol) of Cu(CH
3
COO)
2
and 0.271 g (1 mmol) FeCl
3
·6H O per each
2
catalyst. Further, the solution was stirred for 3 additional hours until its color changed to green for
Co and Cu Salen complexes or purple for Fe Salen complex. Next, 3 g of the solid support (SBA-
15) were added and kept under stirring overnight (~15 h). Finally, the obtained solid was washed
with methanol and water to remove all physisorbed species (until the washings were colorless) and
dried overnight at 80 °C. The SBA-15 support was prepared according to the procedure reported in
a previous work [8].
(Insert Fig. 1 here)
2.4 HMF oxidation.
The catalytic activity performance of the metal Salen complexes supported on SBA-15 (Co/SBA-15,
Fe/SBA-15 and Cu Salen/SBA-15) in the oxidation of HMF were evaluated. The HMF oxidation
reaction was carried out in an aqueous system at neutral pH (the pH was not adjusted) using H
2
O
2
as oxidant agent. The reaction was performed under mild conditions (aqueous media, neutral pH,
atmospheric temperature and pressure). The system consisted of a 125 mL round-bottom flask with
a refrigerant column to avoid the HMF volatilization. All tests were performed with an initial
substrate HMF 0.4 mM [8], 50 mL reaction volume, H
2
O 30 w/V % (100 μL) as oxidant agent and
2
using 0.05 g of catalyst. Aliquots of 500 μL were taken during 24 h, from which 75 μL were injected
in the chromatograph for their analysis. The samples were taken in short periods of time at the early
minutes of the reaction, and in a longer period as the reaction advanced in order to have enough
information for the kinetic study. The reaction mixture was stirred at a constant 500 rpm. Tests were
done at low temperatures 25, 30 and 40 °C to know how the temperature affects the reaction.
Although the catalyst can be used at higher moderate temperatures, 40 °C level was selected as
the maximum temperature to avoid H
2
O degradation. Temperature levels were recorded with
2
thermocouples previously connected to a temperature monitoring program using the Labview
System Design Software.
8

3
. RESULTS AND DISCUSSION
.1 Catalyst characterization
3
The different metallic Salen complexes Co (II), Cu (II) and Fe (III) were synthesized and
immobilized by covalent bonds onto synthetic silica SBA-15. To follow synthesis and
characterization of the catalysts, Fourier transform infrared spectroscopy (FTIR), UV-Visible
spectroscopy (UV-Vis), Diffuse reflectance UV-Visible spectroscopy (DR-UV-Vis) and X ray
photoelectron spectroscopy (XPS) were performed. UV-Visible absorption spectra confirmed the
formation of the Salen ligand and its coordination with the corresponding metallic atom. FTIR
spectra showed incorporation of metallic Salen complexes and its anchoring to the support was
confirmed by DR-UV-Vis. Metallic content (from the complex) over the catalyst was obtained with
atomic absorption spectroscopy (AAS); scanning electron microscopy (SEM) confirmed the
synthesis of mesoporous hexagonal SBA-15, along with the textural properties results obtained
from nitrogen physisorption characterization.
3.1.1 Fourier transform infrared spectroscopy (FT-IR)
Fig. 2 shows the spectra of Co Salen/SBA-15, Fe Salen/SBA-15 and Cu Salen/SBA-15 catalysts,
along with the support without the complex (SBA-15), for comparative reasons. The catalyst spectra
-
1
show a particular band at 1635 cm attributed to ʋ(C = N) stretching vibration , with a shift to lower
frequency for the complexes, indicating coordination of azomethine nitrogen [48-50]. In this case,
the band is showed at for Co Salen/SBA-15 and Cu Salen/SBA-15 but not for Fe Salen/SBA-15
probably due to its lower concentration. SBA-15 support showed intense and broad signals at 1000-
-
1
1
300, and 800 cm , assigned to SiO
2
structure, which remains present after complex
immobilization.
-
1
The signal at 950 cm , associated with Si-OH bonds, decreases after complex immobilization,
which implies that -OH groups have been consumed due to the anchoring of the complex over silica
via chemical bond [51].
9

-
1
Once the complex has been anchored, the bands at 1300-1600 cm can be associated with the
-
1
existence of organic ligands. The band at 1544 cm is attributed to azomethin group vibrations (H-
C=N), which is a characteristic signal of Salen ligands [8]. IR spectra showed some evidence of
complex anchoring, which was confirmed by DR-UV-Vis analysis.
(Insert Fig. 2 here)
The Fe complex catalyst spectra showed more differences, since the bands associated with the
organic ligand can be difficult to observe. This may be explained due to a lower metallic loading for
this catalyst (according to elemental analysis described below).
3.1.2 UV-visible spectroscopy (UV-vis)
Fig. 3 shows the formation of imine by comparing UV-Vis spectra obtained from the collected
samples at different times over the synthesis, to subsequently give rise to the metal Salen complex
formation according with the metal precursor used. Bands observed in the 200-400 nm range are
attributed to π -π* and n-π* ligand charge transfer behavior [48, 52-54]. APTES, salicylaldehyde
and salicylaldehyde-APTES mixture were dissolved in methanol and analyzed. No signals were
observed for APTES (Fig. 3a). Salicylaldehyde (Fig. 3b) showed three bands at 212, 250 and 325
nm, similar to salicylaldehyde-APTES (Fig. 3c), but with a small shift to the red (2 nm) for the first
and second signals. Additionally, a decrease in band intensity due to presence of APTES was
observed. As soon as salicylaldehyde is in contact with APTES, a free imine is formed; a change in
color from transparent to yellow evidenced this event. Three bands at 220, 254 and 315 nm were
observed in the sample taken after stirring the salicylaldehyde-APTES and methanol mixture for 3
h, while a new band at 405 nm appears due to the formation of the free imine.
1
0

(Insert Fig. 3 here)
In Fig. 4, the UV-Vis spectra of different metallic Salen complexes Fe(III), Cu(II) and Co(II)]
unsupported) are compared. The dotted lines represent the spectrum after 3 h stirring once the
(
metal precursor was added and continuous lines signify the spectra after an additional 15 h stirring.
In the Co Salen complex spectrum (Fig. 4c) three bands at 230, 265 and 365 nm were observed.
The Cu Salen complex spectrum (Fig. 4b) showed four bands at 230, 273, 305 and 363 nm, while
for the Fe Salen complex spectrum (Fig. 4a) bands at 243, 263, 330 and 410 nm are observed. All
complexes showed bands between 230 and 273 nm, which are attributed to π-π* charge transfers.
The signals between 305 and 365 nm are assigned to n-π* charge and the small band from Fe
Salen at 410 nm can be attributed to ligand to metal charge transfers. The band around 363-365
nm was assigned to charge transfer transition from the Salen ligand [55]. This indicates the
formation of the Salen ligand and the decrease of the free imine band once the Salen ligand signal
appears, confirming imine coordination forming the Salen metallic complex [56]. A characteristic
absorption band at approximately 450 and 550 nm, attributed to d-d charge transfer transitions from
the Co atom, was not observed due to the low concentration of the Salen metallic complex, since
the synthesis was carried out in an excess of methanol [55]. For this work, all catalysts were
prepared using the same methodology. The reacting solution was stirred for 15 h, and finally the
solvent was separated using a rotary evaporator. Since spectra after 3 h and an additional 15 h
stirring are very similar, it is possible that the stirring time can be reduced in order to evaluates the
difference in catalytic activity.
(Insert Fig. 4 here)
3.1.3 Diffuse reflectance UV-visible spectroscopy (DR-UV-Vis)
DR-UV-Vis spectra for the different metal Salen complex catalysts are observed in Fig. 5. The
spectrum for pure SBA-15 support (Fig. 5d) is also shown for comparison. Characteristic signals
can be observed at 280 nm for Co Salen/SBA-15 (Fig. 5c), and 270 nm for Cu Salen/SBA-15
1
1

(Fig.5b) and Fe Salen/SBA-15 (Fig. 5a); both bands are attributed to π-π* from the phenyl ring of
the Salen ligand. Signals at 370 nm for Cu and at 390 nm for Co and Fe, were related with n-π*
transitions from azomethine (C=N). A shoulder was observed at 460 nm for Co and Cu due to
charge transference between metal and nitrogen; additionally, bands at 620 and 650 nm for Co and
Cu respectively were related with d-d transitions. For the Fe Salen/SBA-15 catalyst, a shoulder was
observed at 500 nm, and the signal at 600 nm was not observed, possibly due to an overlapping or
a smaller complex content over SBA-15, which was found in elemental analysis. The mentioned
signals are a clear indicative of the metallic complexes being supported over SBA-15, according
with the previously reported for metal Salen complexes supported [7, 57-59]. The different  values
found for different catalyst may be explained due to their varying nature and structure; Co (II) and
Cu (II) Salen complexes usually present a flat square structure, while Fe (III) Salen complex is
associated to a square pyramidal structure [60-62].
(
Insert Fig. 5 here)
3.1.4 Elemental analysis by Atomic absorption spectroscopy (AAS)
All different solid catalysts Co, Cu and Fe Salen complexes supported over SBA-15 were submitted
to acid digestion and AAS evaluation. The metallic loading (mmol metal/g catalyst) over the support
was Cu (0.22 mmol/g)  Co (0.20 mmol/g)  Fe (0.18 mmol/g). This may be related with atomic
radiuses of the metals: Fe (1.72 Å)  Co (1.67 Å)  Cu (1.57 Å), and atomic weight of Cu (63.55
g/mol) > Co (58.93 g/mol) > Fe (55.85 g/mol). The Cu Salen complex seems to have a smaller size
(being the element with the lowest atomic radius), which could promote dispersion and further
facilitate anchoring over the hexagonal channels of SBA-15 in comparison with the Fe Salen
complex, being the Fe which has the largest atomic radius of the three metals. Therefore a better
catalytic performance with Cu Salen catalyst is expected.
1
2

3
.1.5 N adsorption/desorption isotherms and textural analysis.
2
Fig. 6 shows adsorption and desorption isotherms for the SBA-15 support and Fe Salen/SBA-15,
Cu Salen/SBA-15 and Co Salen/SBA-15 catalysts at 77 K. All the plots showed are isotherms Type
IV, which are characteristic of mesoporous materials. A clear step was observed for pore filling,
which indicates the presence of a narrow pore size distribution [63, 64]. The hysteresis cycle
showed a behavior similar to H1 and H2 types; more specifically, H1 type for SBA-15 and H2 type
for the rest of the catalysts. However, for the Fe Salen/SBA-15 catalysts hysteresis was more
similar to that of the pure SBA-15 (H1), which may be related with a lower complex charge over the
support (which was confirmed with elemental analysis). H1 type hysteresis is frequently associated
with porous materials consisting on agglomerates with nearly uniform arrangements, which tend to
show a narrow pore size distribution, while H2 type is more associated with unordered pores, where
pore size distribution and shape are not very well defined [65]. It can be caused due to the complex
incorporation; however, for this case all samples showed a narrow distribution. It can also be
0
0
observed that the hysteresis loops of all samples begin at p/p = 0.4 and end at p/p = 0.7,
suggesting limited adsorption. This phenomenon is related to the presence of particles that are not
rigidly joined together [66]. The SBA-15 (Fig. 6) reached the highest value of adsorption at the
saturation point (p/p = 1) followed by Fe Salen/SBA-15, Co Salen/SBA-15 and Cu Salen/SBA-15;
0
which is related to the greater surface area, as shown in Table 1, and the better interconnection of
pores within the particles, which can favor percolation phenomena [66]. Nevertheless, the metal
Salen complex is responsible for the activity and the aforementioned confirm the anchoring of the
metal Salen complexes in the SBA-15 channels.
(Insert Fig. 6 here)
In Fig. 7, pore distribution for support and different catalysts are shown. A small shift can be
observed at the right side of the peak, which is more prominent for Cu and Co catalysts. This may
1
3

be related with complex incorporation over the support; the Salen complex enters inside the biggest
pores, decreasing the available spaces. This may also be related with a higher peak maximum. It is
widely known that SBA-15 hexagonal mesoporous silica is interconnected with nonuniform
0
micropores [63]; this was observed at low relative pressures (p/p ). A wider space can be observed
between adsorption and desorption isotherms at low p/p for catalysts, compared with the pure
0
support. This may be due to micropore formation during the catalyst drying process, which
promotes water and methanol evaporation from the catalyst surface. The isotherms of all samples
observed on Fig.6 confirm the SBA-15 support channels remain accessible [57].
(Insert Fig. 7 here)
Table 1 summarizes the textural properties of the catalysts and SBA-15 support. It is widely known
that a high specific surface area is desired for heterogeneous catalysis applications in order to
promote active phase dispersion; having a higher surface area will also promote adsorption and
desorption processes. However, a bigger surface area does not guarantee a better catalytic
behavior. Textural properties obtained for SBA-15 material were: a specific surface area SBET = 583
2
3
m /g, average pore diameter D
P
= 3.81 nm and pore volume V = 0.67 cm /g. All of these data
p
agree with previous reports by other authors [57, 64, 67, 68]. After the metallic Salen complex was
immobilized over the SBA-15 support, no significant changes were observed in average pore sizes,
while specific surface area and pore volume showed a prominent decrease, in the order SBA-15 >
Fe Salen/SBA-15 > Co Salen/SBA-15 > Cu Salen/SBA-15. This may suggest that metal Salen
complex was mainly grafted inside the mesoporous support surface [53, 55]. The wall thickness
may have been increased when the catalyst was anchored over the support, since the complex is
mainly located over the inner surface of mesoporous materials [48]. When the complex is
supported, the pore diameter distribution is narrower than that of pure SBA-15, suggesting that the
anchorage of the complex occurs in the larger pores. The decrease in pore volume and specific
surface area after complex immobilization indicates the complex was grafted inside the mesoporous
silica channels [55, 69].
1
4

3
.1.6 SEM
In Fig. 8 (with scale of 200 nm) hexagonal channels are observed, with an average pore size of 4.4
nm, compared with 3.81 nm calculated from the N physorption analysis. This morphology is
characteristic of SBA-15 materials due their synthesis with the template (surfactant Pluronic P123).
2
(Insert Fig. 8 here)
3.1.7 X-ray photoelectron spectroscopy (XPS)
To get deeper insight into the surface constitution of the catalysts and the oxidation states of the
elements, the C1s, N1s, O1s and M 2p (M= Cu, Co or Fe) binding energies were investigated. For
this purpose, Fig. 9 and Fig. 10 show the high energy resolution XPS spectra of the elements
mentioned. The C 1s spectrum can be deconvoluted in three peaks, for Co Salen/SBA-15 (Fig. 9a)
the peak at 284.7 eV is attributed to both aromatic and aliphatic carbons of the salen ligand [70]; a
peak at 286.5 eV is associated with C ‒ O or C ‒ N and the low intensity band at ca. 289 eV can be
related to a shake-up satellite associated with π ‒ π* transitions in the aromatic carbon rings of the
ligand [70]. The C1s spectrum for Cu Salen/SBA-15 (Fig. 9b) also showed three peaks at similar
positions 284.7, 286.3 and 288.7 eV and as well as 284.5, 286.7 and 289 eV for Fe Salen/SBA-15
(Fig. 9c) respectively. These results suggest that although Salen complexes can represent the
same kind of interactions with different metallic species, the aromatic rings can modify the electron
density in the conjugated π-system in the three different systems leading to a potential charge
increase and causing a shift to the negative region [71].
N1s spectrum can be deconvoluted also into three peaks, the first one is attributed to the binding
energy between metal and nitrogen (M ‒ N) [72] at 399.5, 399.2 and 399.7 eV for Co Salen/SBA-15
(Fig. 9a), Cu Salen/SBA-15 (Fig. 9b) and Fe Salen/SBA-15 (Fig. 9c) respectively. Second peak at
400.8 (Fig. 9a), 400 (Fig. 9b) and 401.7 eV (Fig. 9c) is associated with C = N [73] and the last one
1
5

and the last one at 402.8 (Fig. 9a), 403 (Fig. 9b) and 403.5 (Fig. 9c) is attributed to the Salen ligand
nitrogen [73].
The XPS spectra in the O1s region show an intense peak around 532 eV which can be
deconvoluted in three peaks. The first is associated to the energy binding between metal and
oxygen (M ‒ O) [70] at 530.5, 531.7 and 532.5 eV as is showed in the O1s spectrums in Fig. 9a, b
,
c for Co Salen/SBA-15, Cu Salen/SBA-15 and Fe Salen/SBA-15 respectively. The second at 532.5
eV for Co Salen/SBA-15 and Cu Salen/SBA-15 catalysts (Fig. 9a, b) and 532.7 eV for Fe
Salen/SBA-15 (Fig. 9c) can be attributed to C‒O binding energy from the metal-Salen complex [74]
and also for Si ‒ O binding energy [75] from the SBA-15 matrix; this is why it is the most intense
peak, since the binding energy between the Si ‒ O bond of silica predominates. The third peak at
533.3, 533.5 and 534 eV (Fig. 9 a,b,c) is due to the O ‒ H in the surface [75]. The binding energy
exhibited in the spectra confirms the formation of the metal Salen type catalysts
The elemental analysis results from the atomic absorption spectroscopy showed a loading (mmol/g)
Cu > Co > Fe which can be related to the anchored metal Salen complex, due to this analysis is
from the bulk and the XPS is a surface analysis, the variation in the intensity of the C1s and N1s
spectra being higher for Cu Salen/SBA-15 catalyst followed by Co Salen/SBA-15 and Fe Salen may
be due to the anchorage of Cu Salen is favored more within the pores of the SBA-15 than on the
surface.
Figure 10 shows the M2p high resolution XPS spectrum for each catalyst where two peaks
associated to 2p 3/2 and 2p1/2 spin-orbit levels can be seen. Co 2p spectrum (Fig. 10a) shows two
peaks from the Co 2p3/2 at 781.5 and Co 2p1/2 at 797 eV attributed to Co ‒ N from Co (II) Salen
complex supported and satellite bands at 787 and 802.3 eV [48, 76-78]. Cu 2p spectrum (Fig. 10b)
exhibits two peaks at 932.7 and 952.5 eV attributed to Cu 2p3/2 and Cu 2p1/2 from Cu (II) Salen
complex supported and low intensity bands around 940 and 945 eV associated to satellites [48, 60].
Fe 2p spectrum (Fig. 10c) shows two deconvoluted peaks attributed to Fe 2p3/2 and Fe 2p1/2 with
a maximum around 711 and 723.5 eV respectively, which shows peaks at 711 and 723.5 eV
associated to Fe (II) while the peaks at 714.3 and 726 eV can be related to Fe (III) species [48, 75,
1
6

7
9]; with satellites at 718.7 and 730.5 eV. Satellite lines and the separation of the 2p 3/2 and 2p1/2
spin-orbit levels often provide valuable information concerning the oxidation state and electronic
structure of metal ions [48]. For the studied catalysts the separation between the peaks that
correspond to 2p 3/2 and 2p1/2 spin-orbit levels were 15.5, 19.8 and 12.5 eV for Co, Cu and Fe
(metal Salen/SBA-15 catalysts) respectively nearly to the reported (from Co (II), Cu (II) and Fe (III)
tetrahydro Salen and Salen complex onto amino-functionalized SBA-15) by other authors [48, 80].
In the other hand, the presence of satellite lines in 2p3/2 and 2p1/2 spin-orbits is indicative of a
high-spin of the metallic center [48].
For metal oxides, the expected binding energies are around 933.2-933.1 eV (in presence of shake
up satellites) for Cu 2p3/2 from CuO, a peak between 933-933.8 eV (in absence of shake up
satellites) for Cu 2p3/2 from Cu
Cu Salen/SBA-15 catalyst studied showed peaks at a lower energy (932.7 and 952.5 eV) . For CoO
and Co species binding energies have been reported between 779.6-780.6 eV and 794.5-796.4
2
O and a peak at 954 eV is related to Cu 2p1/2 [81, 82], while the
3
O
4
eV associated to Co 2p1/2 and 2p3/2 respectively, while a peak at 778.1 eV is due to metallic
cobalt [83, 84]. In contrast to Co 2p peaks and t higher energy (781.5 and 797 eV) from Co
Salen/SBA-15 catalyst. From the Fe 2p results it may be apparent that different species of iron
complex are present.
Finally, a relationship between XPS and other characterization techniques with the activity potential
for each catalyst will be made. The third peak from O1s spectra (attributed to O ‒ H in the surface)
shows a larger area for Cu Salen/SBA-15 catalyst (Figure 9b), compared with Co (Figure 9a) and
Fe (Figure 9c) Salen supported catalysts, whose areas for the same peak are 21 and 29% shorter
than the area for Cu Salen catalyst, respectively. This indicates that more O-H groups are available
on the surface of the Cu Salen/SBA-15 catalyst, which agrees with the observed in N1s spectra,
where the first peak (associated to M ‒ N) presents a higher intensity for Co Salen/SBA-15 catalyst,
which is related to a larger amount of metal Salen complex in the surface. The Fe Salen/SBA-15
catalyst shows a lower amount of complex in the surface (related with a smaller N1s peak intensity),
while the Cu Salen shows the lowest amount of complex from all the catalysts. As stated before,
1
7

from the AAS and N2 physisorption analyses, the Cu Salen catalyst showed the highest metal
loading and the lowest specific surface area. This suggests that the Cu Salen complex is anchored
inside the hexagonal channels of the SBA-15 support, and a better dispersion was apparently
achieved with this catalyst, which contributes to a better catalytic performance.
(Insert Fig. 9 here)
(Insert Fig. 10 here)
3
.2 Catalytic activity of metal (Co II, Fe III and Cu II) Salen complex in HMF oxidation under
mild conditions.
3.2.1 Effect of the transition metal of the metal Salen complex on HMF oxidation
.
In order to evaluate the effect of the transition metal present in the catalyst, the catalysts Co Salen,
Cu Salen and Fe Salen supported on SBA-15 were tested. The activity tests with an initial substrate
concentration of 0.4 mM, 0.050 g of catalyst and 100 μL of H
2
O 30 w/V %; at 30 °C were
2
performed. In the Fig. 11 a significant difference with the Co Salen/SBA-15 catalyst can be
observed, which showed a lower performance with an HMF conversion around 40%, while with the
others (Fe Salen/SBA-15 and Cu Salen/SBA-15), a conversion around 100 % was reached. For Cu
Salen/SBA-15 the reaction rate was slightly higher in the early minutes. An explication of the higher
activity, Cu2+ > Fe3+ > Co2+ could be related with the standard electrode potential values (E
0
), being
more positive for Cu2+ with E0 Cu /Cu = + 0.34 V and more negative for Co2+ with E0 Co Co = - 0.29
2+
2+/
V, while for Fe3+ 0 Fe /Fe = - 0.04; this means that Cu2+ is more oxidant than Fe3+ and Co . On the
E
3+
2+
other hand, a higher activity was expected with Cu Salen/SBA-15 catalyst, due to a higher metal
loading and a better dispersion achieved with this metal Salen complex over the support, which
favor its catalytic activity, as previously mentioned in section 3.17. However, differences in the
molecular structure of the different complexes and their nature (depending on the used precursor)
1
8

could also contribute to the catalyst to have a better or poor performance in catalytic activity. Is
worth mentioning that after a physical observation, the Fe Salen/ SBA-15 catalyst was the least
stable, since during the reaction course its yellowish color was lost and it became almost completely
white, similar to the pure support. With the foregoing the Cu Salen/SBA-15 was selected to continue
with the activity test applied in the HMF oxidation. It is considered as a good finding for this reaction,
where metals such as Fe and Cu showed an increased catalytic performance compared with Co
(known for its high activity in catalysis and one of the most studied Salen complexes, although in
homogeneous catalytic reactions), known as toxic and harmful to health and environmental for at
high concentration of exposition.
(Insert Fig. 11 here)
3.2.2 Effect of the H
2
O amount on HMF oxidation with Cu Salen/SBA-15
2
Fig. 12 shows comparative plots to study the effect of the amount of H
Salen/SBA-15. It was observed that the amount of H had a significant effect on the conversion
rate, reaching 100% of HMF conversion using 100 μL of H 30% w/V compared to about 5% with
μL of H 30% w /V. Based on the foregoing, the experimental activity tests were performed
using 100 μL H . It can also be observed that the reaction goes from 0 to 74% of conversion in
the first 50 minutes, later becomes slower.
Insert Fig. 12 here)
2
O
2
, using the catalyst Cu
2
O
2
2
O
2
5
2
O
2
2
O
2
(
3.2.3 Effect of the temperature on HMF oxidation with Cu Salen/SBA-15
Activity tests at different temperatures were performed in order to know its effect in the HMF
oxidation with Cu Salen/SBA-15. These experiments were performed with an initial HMF
concentration of 0.4 mM, 100 μL of H
2
O
2
30 w/V % and 0.050 g of Cu Salen/SBA-15 catalyst,
during 24 h approximately. The temperatures evaluated were 25, 30 and 40 °C as maximum in
1
9

order to avoid the H
2
O degradation. As can be observed in the Fig. 12, a significant difference is
2
not observed in this range of temperatures with the rate reaction and HMF percent conversion for all
cases the conversion reached nearly 100%. This reveals that the reaction can be carried out at
room temperature and that if temperature changes during the reaction, it will not significantly
change the catalytic performance, at least in the studied range.
(Insert Fig. 13 here)
3.2.4 Kinetic study of HMF oxidation with Cu Salen/SBA-15
In Fig. 14 a change in the concentration was observed during the course of the reaction (around 24
h) for HMF and the identified products DFF, FDCA, HMFCA and FFCA. For the mentioned
experimental tests, the Cu Salen/SBA-15 catalyst was evaluated. A rapid disappearance of HMF
during the first 200 min is observed, to stabilize close to 0 mM. For the products, the maximum was
reached between 50-100 min. After this the concentration dropped; an explanation could be that the
reaction continued to form more oxidized products (unidentified). From the identified products by the
developed analytical method (HPLC), it was observed that before the concentration dropped under
the condition of study (CHMF 0 = 0.4 mM, 100 μL of H
0 °C), the selectivity was higher for FDCA and DFF products; while it was lower for HMFC and
FFCA and very similar between this both last products.
2
O 30 w/V %, 0.050 g of Cu Salen/SBA-15 at
2
3
(Insert Fig. 14 here)
Consequently, it was noted that the reaction under the established conditions was selective to other
products, different to the identified by the HPLC separation method previously developed. However,
the kinetic analysis was continued with the information obtained during the first 100 min of the
reaction, to study the kinetics before the decrease in products concentration; since there is no
monitoring of the unidentified products.
2
0

To observe the behavior of the obtained curves where the product concentrations increase,
reaching a maximum at the same time and not a curve that grows and decays while another
appears according to the literature [85], a mechanism with parallel reactions was proposed as
follows (Fig. 15).
(Insert Fig. 15 here)
The respective rate reaction equations for each compound were proposed as follows.
For HMF, the rate reaction (−풓푯푴푭) is expressed:
풅푪푯푴푭
풓푯푴푭 = −
풅풕
Equation 1
ꢀ
풅푪푯푴푭
=
풌ퟏ푪
+ 풌 푪
+ 풌 푪
+ 풌 푪
퐻푀퐹 4 푯푴푭
푯푴푭
ퟐ
푯푴푭
ퟑ
풅풕
Equation 2
ꢀ
풅푪푯푴푭
Where the negative sign indicates that HMF is disappearing,
is the HMF concentration
풅풕
change with time, 푪푯푴푭 is the HMF concentration at an arbitrary time. The kinetic constants 풌 , 풌 ,
ퟏ
ퟐ
풌 푦 풌 correspond to the HMF reaction to produce DFF, FDCA, HMFCA and FFCA respectively.
ퟑ
4
By grouping the kinetic constants 풌 = 풌 + 풌 + 풌 + 풌 , then remain
ퟏ
ퟐ
ퟑ
ퟒ
ꢀ
풅푪푯푴푭
=
풌푪푯푴푭 , integrating and replacing the initial conditions 풕 = ퟎ
풅풕
and 푪푯푴푭 = ퟎ. ퟒ 풎푴 it was obtained the equation for 푪푯푴푭
.
푪푯푴푭 = ퟎ. ퟒ풆ꢀ풌풕
Equation 3
For DFF, FDCA, HMFCA and FFCA products, differential rate reaction equations remain as follows:
2
1

풅푪푫푭푭
= 풌 푪
ퟏ 푯푴푭
풅풕
풓푫푭푭
=
Equation 4
풅푪푭푫푪푨
= 풌 푪
2 푯푴푭
풅풕
풓푭푫푪푨
=
Equation 5
풅푪푯푴푭푪푨
= 풌 푪
3 푯푴푭
풅풕
풓푯푴푭푪푨
=
Equation 6
풅푪푭푭푪푨
풓푭푭푪푨
=
= 풌 푪
4 푯푴푭
풅풕
Equation 7
Where it can be seen that the sign is positive, since these are products formed. Integrating the
above equations and substituting for 풕 = ퟎ when only HMF is present and the initial concentrations
of the products 푪푫푭푭, 푪푭푫푪푨 ,푪푯푴푭푪푨 and 푪푭푭푪푨 are zero:
ퟎ.ퟒ풌ퟏ
푪푫푭푭
=
(ퟏ − 풆ꢀ풌풕
)
풌
Equation 8
ퟎ.ퟒ풌ퟐ
푪푭푫푪푨
=
(ퟏ − 풆ꢀ풌풕
)
풌
Equation 9
2
2

ퟎ.ퟒ풌ퟑ
풌
푪푯푴푭푪푨
=
ꢁퟏ − 풆ꢀ풌풕
ꢂ
Equation 10
ퟎ.ퟒ풌ퟒ
푪푭푭푪푨
=
ꢁퟏ − 풆ꢀ풌풕
ꢂ
풌
Equation 11
After calculating the integrated equations for each concentration, these were used to make the
fitting using Matlab software (version R2016a) using the Curve Fitting tool. Fig. 16, shows the
concentration curves versus time for HMF and the identified products from the catalytic oxidation in
presence of H
2
O and the Cu Salen/SBA-15 catalyst with CHMF 0 = 0.4 mM at 30°C. The points
2
represent the experimental data and the lines the first fitting with the proposed kinetic model
(
-
(
Equations 3 and 8-11).
Insert Fig. 16 here)
Fig. 16, shows that a good fitting for the products experimental data was obtained. For HMF,
although has the same behavior that the experimental data, this does not pass through all the data
as in the case of products (DFF, FDCA, HMFCA and FFCA).
The Table 2 bellow present the kinetic constants 푘 , 푘 , 푘 , 푘 푦 푘; obtained from the model fit.
1
2
3
4
It is notable that the kinetic constant 풌 ≠ 푘 + 푘 + 푘 + 푘 , this because they are giving
1
2
3
4
additional reactions in parallel, then 풌 = 푘 + 푘 + 푘 + 푘 + 푘 . Where 푘 would be a kinetic
1
2
3
4
푥
푥
constant from contributions from additional constants due to the reactions that occur for the
formation of products that are not able to identify at the moment and according to the literature
could be happening something similar to what is shown in Fig. 17 [28]. The reason why the
2
3

experimental data suggests that the products are being formed through parallel reactions from the
oxidation of HMF of the alcohol and aldehyde group of the HMF molecule might be due to the
oxidation steps from DFF and FFCA oxidation are occurring faster and then is possible to observe
the FDCA and FFCA products from the first minutes of the reaction.
(Insert Fig. 17 here)
This is an explanation of why a better fit of the experimental data (CHMF against time) was not fully
achieved, because additional information about other reactions occurring (which were not used in
the deduction of the proposed mechanism) would be needed.
Table 3 shows a summary of HMF % conversion and % selectivity to the identified products DFF,
FFCA, FDCA and HMFCA for the different activity conducted tests.
There is a synergistic effect between the Salen/SBA-15 and Fe Salen/SBA-15 catalysts with the
H
2
O , due to the conversion percentage achieves almost 100% compared with 40% in catalyst
2
absence, but it is important to denote that the Fe Salen/SBA-15 was dissolved and acted in
homogenous reaction. The selectivity of the identified products DFF and FDCA were higher than
the others in all the cases using 100 μL of H
2
O . However, for Cu and Fe, FDCA conversion was
2
the highest. This can be explained due to FDCA having two carboxylic acid groups; this molecule is
the maximum oxidation state of HMF. It is also at the end of the sequential reactions, which means
the environment was oxidant enough to promote a high conversion of DFF and FFCA to FDCA.
The molecule with the least conversion was HMFCA, which is expected due to the difficult of having
oxidizing conditions that can both oxidate the aldehyde group of HMF completely to carboxylic acid,
but leave the alcohol group intact. This would suggest that the studied reaction system was favored
by oxidation of the alcohol side of the HMF molecule to form DFF (aldehyde) followed by FFCA that
is converted rapidly to FDCA (Fig. 18). Oxidation of the aldehyde side of the HMF molecule to form
HMFCA is not as promoted as previously explained, and therefore only few FFCA and FDCA could
have been obtained through this mechanism (Fig. 19).
2
4

(
Insert Fig. 18 here)
Insert Fig. 19 here)
(
3.3 Stability test of Cu Salen/SBA-15 catalyst
Before using the catalyst in the reaction, it was stirred for 15 h in deionized water, it was then
recovered and dried at 80 °C for 15 h, and this was called mechanical treatment. All further tests
(already discussed in this manuscript) were performed with the catalysts after mechanical
treatment. For stability tests, spectroscopic analysis of atomic absorption and DR-UV-vis were
performed for the catalyst before and after reaction (this catalyst was recovered and dried following
the same process described above).Recycle tests were performed using an HMF concentration of
0.4 mM, a 30% H
2
O volume of 100 L and 0.05 g of the Cu (II) Salen/SBA-15 catalyst, with
2
atmospheric pressure. Each cycle lasted for 24 h.
3.3.1 Atomic absorption spectroscopy (AAS)
From the AAS, a copper loading of 0.14 mmol Cu/ g catalyst and 0.024 mmol Cu/g catalyst were
obtained for the catalysts before and after reaction, respectively. This indicates that leaching of the
active species is happening, which was confirmed by the presence of copper in the reaction
solution. Consequently, it could be asked, what is acting as a catalyst? Are the active species in the
solid, liquid or in both phases? For every experiment, once the samples were centrifugated and the
separated liquid was taken for HPLC evaluation, repeated measurements at different times up to
about 24 hours, showed no changes in concentration. Therefore, we can conclude that despite
leaching, the solid catalyst (supported metal Salen complex) was identified as the active species in
the reaction and not the species in solution, discarding the latter as catalytically active.
3.3.2 Recycle tests
Fig. 20 shows the results from the recycle tests performed for the reaction. A slight decrease in the
HMF conversion is present, with a difference of only 2.7% between the first and final cycle.
2
5

(insert Figure 20 here)
3.3.3 DR-UV-vis spectroscopy
Fig. 21 illustrates the DR-UV-vis spectrum of the Cu Salen/SBA-15 catalyst before and after being
used in the reaction and also the corresponding spectrum to the pure support SBA-15 for
comparative purposes. Fig. 21b shows the characteristic Cu Salen/SBA-15 catalyst bands, already
discussed before. Fig. 21c shows the Cu Salen/SBA-15 catalyst after reaction, where the same
catalyst bands are observed, but with a marked intensity decrease.. Therefore, and although these
catalysts showed good activity and recyclability, it is still necessary to continue the study of such
catalysts to improve their stability.
(insert Figure 21 here)
CONCLUSIONS
Salen catalyst complexes of Fe (III) and Cu (II) supported on SBA-15 were the most active in the
oxidation of HMF using H
2
O as an oxidizing agent; however, the Fe (III) presented low stability, so
2
more studies are needed to have a better understanding of its catalytic performance. Cu
Salen/SBA-15 catalyst, although present leaching, showed the best activity due to the highest metal
loading (Cu Salen complex) and mainly because of a better dispersion of the active phase. H
2
O
2
played an important role in the reaction, since it significantly affects the conversion and selectivity.
However, it has a synergistic effect with the catalyst which was found when the conversion did not
exceed 40% when only carried out with H
2
O alone, while when both the catalyst and the peroxide
2
were present, a conversion close to 100% was reached. Under the reaction conditions, the HMF
oxidation was more selective towards unidentified products; it is presumed that this occur due to the
furan ring rupture. With the decrease in the amount of H
2
O and therefore HMF conversion, the
2
selectivity to DFF, FDCA, FFCA and HMFCA products increased. A good fit to the identified
products (DFF, FDCA, HMFCA and FFCA) with the proposed mechanism with parallel reactions
was achieved, however, for the HMF the adjustment was not perfect due to the lack of information
2
6

about other formed products. However, differences in the molecular structure and transfer energy
could also affect their performance, which may be related to that observed in the DR-UV-Vis
spectra, where the different bands of electronic transitions π-π *, n- π * and metal-N (center of the
ligand) are well-defined for the copper complex, while for the other catalysts are given overlaps. The
behavior of the curves from the experimental data of the identified products (DFF, FDCA, HMFCA
and FFCA) indicates that their formation is being carried out by parallel reactions or the DFF and
FFCA oxidation are very fast reaction steps. On the other hand, a good fit to the experimental data
on the concentration of HMF was not achieved, since more information of the other occurring
reactions is required. It is suggested that the products were not identified until now are a result of
the rupture of the furan ring. Metal Salen complex Co, Cu and Fe catalysts showed a higher
selectivity towards the formation of DFF compared to HMFCA, suggesting that is more favorable the
oxidation of alcohol group in HMF molecule that oxidation of the aldehyde group in the other side of
the molecule under the conditions here described. Metal Salen complex supported should remain a
case study in catalysis since they have the advantage of being active at atmospheric pressure with
green oxidizing agents and at room temperature.
ACKNOWLEDGEMENTS
This work was funded by Facultad de Ciencias Químicas, UANL. Daniela Xulú Martínez Vargas is
very grateful to the National Council of Science and Technology (CONACYT) of Mexico for its
support and for the scholarship awarded to the Postgraduate of Science Focused in Sustainable
Processes at Universidad Autónoma de Nuevo León registered in the National Quality Graduate
Program from the CONACYT key 000597. Daniela Xulú Martínez Vargas and Ladislao Sandoval
Rangel now work at Tecnológico de Monterrey, Monterrey Campus, Research group in Energy and
Climate Change. The authors thank to Dr. Carlos D. Garcia and to University of Texas at San
Antonio (UTSA) for the facilities provided.
REFERENCES
[1] S. Albonetti, A. Lolli, V. Morandi, A. Migliori, C. Lucarelli, F. Cavani, Conversion of
5
-hydroxymethylfurfural to 2,5-furandicarboxylic acid over Au-based catalysts:
2
7

Optimization of active phase and metal–support interaction, Applied Catalysis B:
Environmental, 163 (2015) 520-530.
[2] F. Neațu, R.S. Marin, M. Florea, N. Petrea, O.D. Pavel, V.I. Pârvulescu, Selective
oxidation of 5-hydroxymethyl furfural over non-precious metal heterogeneous catalysts,
Applied Catalysis B: Environmental, 180 (2016) 751-757.
[
3] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Phase Modifiers Promote Efficient
Production of Hydroxymethylfurfural from Fructose, Science, 312 (2006) 1933-1937.
4] L. Hu, G. Zhao, W. Hao, X. Tang, Y. Sun, L. Lin, S. Liu, Catalytic conversion of
[
biomass-derived carbohydrates into fuels and chemicals via furanic aldehydes, RSC
Advances, 2 (2012) 11184-11206.
[
5] D.M. Alonso, J.Q. Bond, J.A. Dumesic, Catalytic conversion of biomass to biofuels,
Green Chemistry, 12 (2010) 1493-1513.
6] M. Vohra, J. Manwar, R. Manmode, S. Padgilwar, S. Patil, Bioethanol production:
Feedstock and current technologies, Journal of Environmental Chemical Engineering, 2
[
(2014) 573-584.
[7] N.-T. Le, P. Lakshmanan, K. Cho, Y. Han, H. Kim, Selective oxidation of 5-
hydroxymethyl-2-furfural into 2,5-diformylfuran over VO2+ and Cu2+ ions immobilized
on sulfonated carbon catalysts, Applied Catalysis A: General, 464–465 (2013) 305-312.
[8] D.X. Martíneza Vargas, J. Rivera De La Rosa, S. Arif Iyoob, C.J. Lucio-Ortiz, F.J.
Cerino Córdoba, C.D. Garcia, Phenol oxidation by air using a Co (II) Salen complex
catalyst supported on nanoporous materials: synthesis, characterization and kinetic analysis,
Applied Catalysis A: General, 506 (2015) 44-56.
[
9] G.W. Huber, S. Iborra, A. Corma, Synthesis of Transportation Fuels from Biomass:ꢀ
Chemistry, Catalysts, and Engineering, Chemical Reviews, 106 (2006) 4044-4098.
10] A. Corma, S. Iborra, A. Velty, Chemical Routes for the Transformation of Biomass
into Chemicals, Chemical Reviews, 107 (2007) 2411-2502.
11] L. Bing, Z. Zhang, K. Deng, Efficient One-Pot Synthesis of 5-(Ethoxymethyl)furfural
[
[
from Fructose Catalyzed by a Novel Solid Catalyst, Industrial & Engineering Chemistry
Research, 51 (2012) 15331-15336.
[
12] S.P. Simeonov , J.A.S. Coelhoꢀ, C.A.M. AfonsoꢀAn Integrated Approach for the
Production and Isolation of 5-Hydroxymethylfurfural from Carbohydrates, ChemSusChem,
(2012) 1388-1391.
13] A.D. Sutton, F.D. Waldie, R. Wu, M. Schlaf, L.A. ‘Pete’ Silks, J.C. Gordon, The
hydrodeoxygenation of bioderived furans into alkanes, Nat Chem, 5 (2013) 428-432.
14] R.-J. van Putten, J.C. van der Waal, E. de Jong, C.B. Rasrendra, H.J. Heeres, J.G. de
5
[
[
Vries, Hydroxymethylfurfural, A Versatile Platform Chemical Made from Renewable
Resources, Chemical Reviews, 113 (2013) 1499-1597.
[15] B. Liu, Z. Zhang, K. Lv, K. Deng, H. Duan, Efficient aerobic oxidation of biomass-
derived 5-hydroxymethylfurfural to 2,5-diformylfuran catalyzed by magnetic nanoparticle
supported manganese oxide, Applied Catalysis A: General, 472 (2014) 64-71.
[
16] T. Werpy, G. Peterson, Top Value Added Chemicals from Biomass, U.S.Department
of Energy (DOE), Springfield, VA, (2004) 22161.
17] B. Kamm, Production of Platform Chemicals and Synthesis Gas from Biomass,
Angewandte Chemie International Edition, 46 (2007) 5056-5058.
18] K.T. Hopkins, W.D. Wilson, B.C. Bender, D.R. McCurdy, J.E. Hall, R.R. Tidwell, A.
[
[
Kumar, M. Bajic, D.W. Boykin, Extended Aromatic Furan Amidino Derivatives as Anti-
Pneumocystis carinii Agents, Journal of Medicinal Chemistry, 41 (1998) 3872-3878.
2
8

[19] M. Del Poeta, W.A. Schell, C.C. Dykstra, S. Jones, R.R. Tidwell, A. Czarny, M. Bajic,
M. Bajic, A. Kumar, D. Boykin, J.R. Perfect, Structure-In Vitro Activity Relationships of
Pentamidine Analogues and Dication-Substituted Bis-Benzimidazoles as New Antifungal
Agents, Antimicrobial Agents and Chemotherapy, 42 (1998) 2495-2502.
[20] A.S. Benahmed-Gasmi, P. Frere, M. Jubault, A. Gorgues, J. Cousseau, B. Garrigues,
Proceedings of the International Conference on Science and Technology of Synthetic
Metals (ICSM'92) 2,5-bis(1,4-dithiafulven-6-yl) substituted furans, thiophenes and N-
methyl pyrroles as precursors for organic metals, Synthetic Metals, 56 (1993) 1751-1755.
[
21] D.T. Richter, T.D. Lash, Oxidation with dilute aqueous ferric chloride solutions
greatly improves yields in the ‘4+1’ synthesis of sapphyrins, Tetrahedron Letters, 40 (1999)
735-6738.
22] X. Tong, Y. Ma, Y. Li, Biomass into chemicals: Conversion of sugars to furan
derivatives by catalytic processes, Applied Catalysis A: General, 385 (2010) 1-13.
23] J. Ma, Z. Du, J. Xu, Q. Chu, Y. Pang, Efficient Aerobic Oxidation of 5-
6
[
[
Hydroxymethylfurfural to 2,5-Diformylfuran, and Synthesis of a Fluorescent Material,
ChemSusChem, 4 (2011) 51-54.
[24] C. Moreau, M.N. Belgacem, A. Gandini, Recent Catalytic Advances in the Chemistry
of Substituted Furans from Carbohydrates and in the Ensuing Polymers, Topics in
Catalysis, 27 (2004) 11-30.
[
25] C. Méalares, A. Gandini, Polymeric schiff bases bearing furan moieties 2. Polyazines
and polyazomethines, Polymer International, 40 (1996) 33-39.
26] A.S. Amarasekara, D. Green, L.D. Williams, Renewable resources based polymers:
Synthesis and characterization of 2,5-diformylfuran–urea resin, European Polymer Journal,
5 (2009) 595-598.
27] H. Choudhary, S. Nishimura, K. Ebitani, Metal-free oxidative synthesis of succinic
[
4
[
acid from biomass-derived furan compounds using a solid acid catalyst with hydrogen
peroxide, Applied Catalysis A: General, 458 (2013) 55-62.
[
28] S. Li, K. Su, Z. Li, B. Cheng, Selective oxidation of 5-hydroxymethylfurfural with
H2O2 catalyzed by a molybdenum complex, Green Chemistry, 18 (2016) 2122-2128.
29] P.A. Son, S. Nishimura, K. Ebitani, Synthesis of levulinic acid from fructose using
Amberlyst-15 as a solid acid catalyst, Reaction Kinetics, Mechanisms and Catalysis, 106
[
(2012) 185-192.
[30] J. Nie, J. Xie, H. Liu, Efficient aerobic oxidation of 5-hydroxymethylfurfural to 2,5-
diformylfuran on supported Ru catalysts, Journal of Catalysis, 301 (2013) 83-91.
31] M.L. Ribeiro, U. Schuchardt, Cooperative effect of cobalt acetylacetonate and silica in
[
the catalytic cyclization and oxidation of fructose to 2,5-furandicarboxylic acid, Catalysis
Communications, 4 (2003) 83-86.
[32] W. Partenheimer, Vladimir V. Grushin, Synthesis of 2,5-Diformylfuran and Furan-2,5-
Dicarboxylic Acid by Catalytic Air-Oxidation of 5-Hydroxymethylfurfural. Unexpectedly
Selective Aerobic Oxidation of Benzyl Alcohol to Benzaldehyde with Metal=Bromide
Catalysts, Advanced Synthesis & Catalysis, 343 (2001) 102-111.
[
33] P. Verdeguer, N. Merat, A. Gaset, Oxydation catalytique du HMF en acide 2,5-furane
dicarboxylique, Journal of Molecular Catalysis, 85 (1993) 327-344.
34] M. Kröger, U. Prüße, K.-D. Vorlop, A new approach for the production of 2,5-
[
furandicarboxylic acid by in situ oxidation of 5-hydroxymethylfurfural starting from
fructose, Topics in Catalysis, 13 (2000) 237-242.
2
9