Mesostructured vanadia-alumina catalysts for the synthesis of vitamin K3

Article abstract of DOI:10.1016/j.cattod.2014.12.026

Liquid-phase selective oxidation of 2-methylnaphtalene to the vitamin K₃ was investigated on a series of mesostructured vanadia-alumina catalysts with different V/Al ratios (0.1-1). They were prepared by the co-precipitation of ammonium metavanadate and aluminum nitrate at pH 5 and characterized by XRD, Raman, TG-DTA, elemental analysis and adsorption-desorption isotherms of nitrogen at -196 °C. The influence of the catalyst composition and calcination temperature upon the physico-chemical properties were closely investigated in relation with the catalytic performances. The result of this investigation indicated the catalysts containing a higher vanadium loading and calcined at 300 °C as the most active. Thus, the VAl07 catalyst led to a selectivity in vitamin K₃ of 54% for a conversion of 2-methylnaphtalene of 76%. Such an activity corresponded to mixtures where metavanadate (Raman line at 940 cm^-1) and decavanadate (Raman line at 990 cm^-1) species were the most abundant. No lines of V₂O₅ (around 705 cm^-1) were detected in the catalysts calcined below 600 °C.

Full text of DOI:10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Mesostructured vanadia–alumina catalysts for the synthesis of
vitamin K3
M. Florea∗, R.S. Marin, F.M. Pa˘la˘s¸ anu, F. Neat¸u, V.I. Pârvulescu
University of Bucharest, Faculty of Chemistry, Department of Organic Chemistry, Biochemistry and Catalysis, 4–12 Regina Elisabeta Bvd., Bucharest 030016,
Romania
a r t i c l e i n f o
a b s t r a c t
Article history:
Liquid-phase selective oxidation of 2-methylnaphtalene to the vitamin K3 was investigated on a series of
mesostructured vanadia–alumina catalysts with different V/Al ratios (0.1–1). They were prepared by the
co-precipitation of ammonium metavanadate and aluminum nitrate at pH 5 and characterized by XRD,
Raman, TG–DTA, elemental analysis and adsorption-desorption isotherms of nitrogen at −196 ◦C. The
influence of the catalyst composition and calcination temperature upon the physico-chemical proper-
ties were closely investigated in relation with the catalytic performances. The result of this investigation
indicated the catalysts containing a higher vanadium loading and calcined at 300 ◦C as the most active.
Thus, the VAl07 catalyst led to a selectivity in vitamin K3 of 54% for a conversion of 2-methylnaphtalene of
76%. Such an activity corresponded to mixtures where metavanadate (Raman line at 940 cm−1) and deca-
Received 7 September 2014
Received in revised form
15 December 2014
Accepted 19 December 2014
Available online xxx
Keywords:
Selective oxidation of 2-methylnaphtalene
Mesoporous vanadia–alumina oxides
Synthesis of vitamin K3
vanadate (Raman line at 990 cm−1) species were the most abundant. No lines of V2O5 (around 705 cm−1
were detected in the catalysts calcined below 600 ◦C.
)
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
inorganic waste per 1 kg of target product, and therefore this
process is often cited as an example of a “dirty” fine chemical
industry process [2]. Also, alternative methods have been proposed
to prevent disposal problems of Cr3+ and to improve yields and
selectivity to VK3. Both homogeneous and heterogeneous cata-
lysts were considered in this scope [4,5]. CH3ReO3 combined with
hydrogen peroxide or tert-butyl hydroperoxide [6], Pd(II)-modified
polystyrenesulfonic acid resins [7], metallophthalocyanine com-
plexes [8] and metalloporphyrin [9] or Ti and Fe containing zeolites
[10,11] are among the suggested replacing catalytic systems. In
addition to these catalysts, oxidation with Ce methanesulfonate
followed by a subsequent electrolytic oxidation of reduced cerium
was also suggested [12]. It has also been proposed the use of 2-
methyl-1-naphthol as substrate instead of 2-methylnaphthalene
[13], but this route involves the catalytic methylation of 1-naphthol,
which complicates the technology. Also 2-methyl-1-naphthol is
more expensive than 2-methylnaphthalene.
However, the smaller yields and selectivities compared to
the initial chromium-based system represent the most important
drawbacks of the proposed methodologies. Therefore, designing
highly active and selective catalysts for this reaction is still a chal-
lenging goal. In this effort, here we report an attempt to selectively
oxidize 2-MN to VK3 using hydrogen peroxide as oxidant agent and
vanadia–alumina mixed oxides as catalysts (Scheme 2). Vanadium
containing catalysts are still considered among the most promis-
ing catalytic systems for oxidation reactions [14], and one of the
Quinone derivatives play an important role in bio-systems and
are useful intermediates and products of fine chemistry [1]. Besides,
aromatic oxidation is one of the key-reactions for the synthesis of
vitamin building blocks [2], produced in amounts ranging between
1000 and 10,000 t per year [2]. In this context the development of
environmentally benign methods for the production of substituted
quinones is still a challenging goal.
Vitamin K3 (2-methyl-1,4-naphthoquinone, known also as
menadione, VK3) is a synthetic chemical compound sometimes
used as a nutritional supplement. It is widely used as a blood
coagulating agent due to its anti-hemorrhagic effects and is a key
intermediate in the synthesis of the other vitamins of group K.
Hence, it is also classified as a provitamin (Scheme 1).
At industrial scale, vitamin K3 (VK3) is produced via stoichiomet-
ric oxidation of 2-methylnaphtalene (2-MN) by CrO3 in sulphuric
acid and depending on the reaction conditions, the yield to VK3
is rather poor and does not exceed 40–60% [3]. Moreover, the
E-factor (environmental efficiency) is rather unfavorable for the
vitamin K3 production. This method produces about 18 kg of toxic
∗
Corresponding author. Tel.: +0040 213675727; fax: +0040 214100241.
E-mail addresses: mihaela.florea@chimie.unibuc.ro, mihaela.florea@g.unibuc.ro
(M. Florea), vasile.parvulescu@g.unibuc.ro (V.I. Pârvulescu).
http://dx.doi.org/10.1016/j.cattod.2014.12.026
0920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
2
M. Florea et al. / Catalysis Today xxx (2015) xxx–xxx
XRD patterns were recorded with a Shimadzu 7000 powder
˚
diffractometer using Cu K␣ radiation (ꢀ = 1.5418 A). The data acqui-
sition was performed in the range 6–80 2 theta, with a step of
2◦ min−1
.
The chemical composition of the catalysts was determined by
inductively coupled plasma (ICP) atomic emission spectroscopy
using a Thermo Jarrell Ash Iris Advantage equipment. Prior to anal-
ysis the sample was brought into solution by alkali oxidative fusion
using NaOH/Na2O2 and subsequent dissolution with diluted HCl.
Raman spectroscopy was performed using a Horiba spectrom-
eter, equipped with a He–Ne (ꢀ = 633 nm) laser. The spectra were
recorded in the 200–1600 cm−1 range.
Thermogravimetry was performed with a Shimadzu DTG-60
instrument under dry nitrogen flow (50 mL min−1), temperature
range 25–800 ◦C with a heating-up rate of 5 ◦C min−1. Differential
thermal analysis was conducted with a Shimadsu DTA-60 instru-
ment.
Scheme 1. Vitamins of group K.
Elemental analysis was performed on a EuroEA 3000 automated
analyzer. The sample (less than 1 mg) were weighed in tin con-
tainers and were burned in a vertical reactor (oxidation tube) in
the dynamic mode at 980 ◦C in an He flow with the addition of O2
(10 mL) at the instant of sample introduction. Portions of the sam-
ple in tin capsules were placed in the automated sampler, from
which they were transferred to the oxidation tube at regular inter-
vals. The concentration of each element was calculated using the
Callidus program supplied with the analyzer.
Scheme 2. Oxidation pathway for 2-methylnaphthalene.
2.3. Catalytic tests
reasons lies in the variability in geometric and electronic structure
of surface vanadium oxides in the supported vanadium oxides [15].
In this scope a series of V2O5–Al2O3 were prepared following the
co-precipitation method, and calcined at different temperatures.
The characterization of these materials was carried out using pow-
der XRD, isotherms of adsorption of nitrogen at −196 ◦C, Raman
spectroscopy, TG–DTA and elemental analysis.
The catalytic tests were carried out in a two-necked flask
equipped with a magnetic stirrer and a condenser at atmospheric
pressure. In a typical procedure, the substrate (2-MN, 2.4 ␮mole),
0.05 g of catalyst and 5 mL of solvent were loaded into the flask,
and then the appropriate hydrogen peroxide (27–30 wt%) was drop
wise added (0.025 mL min−1) using a liquid pump. After H2O2 was
added totally, the reaction mixture was stirred for different reaction
times.
The concentration of hydrogen peroxide prior to use and during
the reaction was determined iodometrically. The catalytic exper-
iments were carried out under vigorous stirring at temperatures
between 40 and 100 ◦C and reaction times from 0.5 till 24 h. After
completion of 2-MN oxidation and filtering off the catalyst, the
mother liquor was diluted 5 times with acetonitrile and the prod-
ucts were analyzed by high-performance liquid chromatography
(HPLC) on a Thermo Scientific Accela 600 device equipped with a
UV–vis detector and a Mediterranean Sea 18 (C18) column. The
mobile phase containing 80% acetonitrile and 20% water was fed at
a flow rate of 0.6 mL min−1 at 25 ◦C using a two channels-detection
and an injection volume of 3 ␮L. A maximum of absorption was
found for 2-MN at ꢀ = 265 nm and for VK3 at ꢀ = 245 nm.
2. Experimental
2.1. Catalyst preparation
Vanadia–alumina mixed oxides with different V/Al ratios were
prepared by the co-precipitation of aluminum nitrate (0.08 M) and
ammonium metavanadate (0.12 M) solutions at pH 5. Ammonium
metavanadate was firstly dissolved in hot water (60 ◦C) under vig-
orous stirring, and then acidified with nitric acid (progressively
added (drop by drop) until a pH of 3.0). To this solution, the solution
containing aluminum nitrate (0.08 M) was added in one charge, and
the pH changed to 2.5. Then, the precipitation was achieved with
aqueous ammonia (25 wt%) progressively (drop by drop) till a final
pH of 5) and the mixture was stirred for another hour. The slurry
was subsequently centrifuged and the solid was washed several
times with hot water and methanol till the complete removal of
nitrates and ammonia ions. Drying was performed in two steps:
at 60 ◦C for 4 h under vacuum, and continued at 120 ◦C, overnight
under ambient conditions. Finally, the samples were calcined at 300
or 500 ◦C for 5 h, in static air. The catalysts were denoted function
of the V/Al ration and calcination temperature (Table 1).
3. Results and discussion
3.1. Catalysts characterization
Table 1 compiles the theoretical and experimental V/Al ratios,
BET surface areas, and the nitrogen content determined from the
elemental analysis. Surface area varied as function of V/Al ratios
and calcination temperatures. As a general trend, the surface area
increased with the decrease of the V/Al ratio, with the highest
value of 288 m2 g−1 for the VAl01 sample. Fig. 1 shows adsorption
isotherms of type IV, with hysteresis loops specific to a capillary
condensation characteristic to mesoporous materials. As expected,
further calcination at 500 ◦C induced a further decrease of the sur-
face area, which most probably is associated to a sintering process
2.2. Physico-chemical characterization
Surface areas were determined by nitrogen adsorption at
−196 ◦C on a fully computerized Micromeritics ASAP 2020 instru-
ment, using the BET formalism. The catalyst powder was degassed
for 2 h at 150 ◦C and 0.1 Pa before each adsorption measurement.
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
M. Florea et al. / Catalysis Today xxx (2015) xxx–xxx
3
Table 1
Surface area of dried and calcined VAlO catalysts with different V/Al ratios.
V/Al ratio
Surface area (m2 g−1
)
Catalyst
N content in dried
catalysts (%)
Theoretical
Experimental
Dried samples
Calcined at (300 ◦C)
Calcined at (500 ◦C)
VAl01
VAl025
VAl05
VAl07
VAl1
0.1
0.25
0.5
0.7
1.0
0.1
0.25
0.5
0.67
0.93
288
200
149
140
134
278
93
117
115
110
160
71
75
76
70
2.5
1.5
1.2
1.1
0.9
27.68◦, and V2O5 (PDF card 00-053-0538) defined by diffraction
lines at 2 theta = 28.64 and 31.09◦. In addition, small reflections
lines can be observed at 600 ◦C (2 theta = 19.4, 37.2 and 45.6◦)
attributed to ␥-Al2O3, in agreement with the database standard
(PDF card no. 00-010-0425). Similar phases were identified for the
other V/Al compositions, as well.
However, the difractogram of the sample calcined at 900 ◦C con-
tains less diffraction lines than the one calcined at 600 ◦C, attributed
to only two phases: AlVO4 and V2O5. Beside this, these patterns
shows a dramatic decrease of the intensity of the V2O5 charac-
teristic lines while the intensity corresponding to diffraction lines
of AlVO4 increased slowly. Thus, the amount of V2O5 which had
formed at the lower temperature was most probably depleted due
to its reaction with Al2O3 followed by the formation of AlVO4, espe-
cially at higher temperatures.
Fig. 1. N2 adsorption–desorption hysteresis loops for vanadia–alumina mixed
oxides catalysts with different V/Al ratios, calcined at 300 ◦C.
The DTA curves shown in Fig. 3 revealed for all catalysts two
endothermic peaks at 90 and 200 ◦C, associated with the removal
of residual water, nitrate and ammonium ions. An exothermic effect
accompanied by a mass loss of around 3% was evidenced for all sam-
ples at higher temperatures. For the sample with the lowest V/Al
ratio (0.1) it was located at 744 ◦C and shifted to lower tempera-
tures with the increase of the V/Al ratio. In correlation to the XRD
analysis these effects were attributed to the formation of AlVO4.
For the samples with V/Al ratios of 0.5, 0.7 and 1 an endothermic
effect at 759 ◦C was also evidenced, which can be assigned to the
melting of V2O5.
Raman spectroscopy is a useful technique providing infor-
mations about the surface vanadium species. According to
Weckhuysen and Keller the increase of the vanadium loading may
lead to different vanadium oxide species in the following order:
orthovanadate (VO4) → pyrovanadate (V2O7) → metavanadate
(VO3)n → decavanadate (V10O28) → vanadium pentoxide (V2O5)
[15]. At low vanadium loadings, isolated four-fold coordinated
mono-vanadate species are usually present on the surface. Higher
vanadium loadings led to identification of lines in the region
1000–900 cm−1 that are attributed to the V O stretch vibration
of polymeric vanadium species or VxOy “clusters” formed on the
surface. The Raman lines below 900 cm−1 account to the vibration
of particles. Data presented in Table 1 indicate that the experi-
mental V/Al molar ratios in the studied catalysts are lower than
the theoretical values for the ratios higher than 0.7, most probably
due to the poor precipitation of vanadium with aluminum in the
starting solution.
Table 1 also presents the nitrogen content in the dried samples.
It decreased with the increase of the V/Al ratio and paralleled the
decrease of the aluminum nitrate content and the decomposition of
ammonium ions in the drying step. After calcination, the nitrogen
was completely removed.
Fig. 2 shows the XRD patterns of VAl07 catalyst at different
calcination temperatures. As a general tendency, all samples are X-
ray amorphous until 500 ◦C. At 600 ◦C two well crystallized phases
were identified as following: AlVO4 phase (PDF card 00-039-0276)
defined by three diffraction lines at 2 theta = 24.42, 25.68 and
Fig. 2. XRD patterns of VAl07 calcined at different temperatures.
Fig. 3. DTA curves for V2O5–Al2O3 with different V/Al ratios.
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
4
M. Florea et al. / Catalysis Today xxx (2015) xxx–xxx
Fig. 4. Raman spectra of VAlO samples prepared with different V/Al ratios (0.1, 0.25,
0.5, 0.7, and 1) and calcined at 300 ◦C.
Fig. 5. Raman spectra of VAl07 at different temperatures.
(940 cm−1) and decavanadate (990 cm−1) species observed for
VAl07 300, is transformed in decavanadate (992 cm−1) species, as
the surface area was reduced from 115 to 76 m2 g−1. Calcination
of polymeric V
may identify three possible active oxygen sites: terminal V O,
binding surface V V polymeric species, and binding V Al,
O V species [15]. Accordingly, Raman analysis
O
O
at 700 ◦C, led to sharp Raman lines at 694, 530, 406 and 281 cm−1
which indicates the formation of crystalline V2O5 particles.
,
vanadium to support species. Fig. 4 reveals the Raman spectra of
V2O5–Al2O3 mixed oxides with different V/Al ratios. The weak and
broad Raman lines in the 100–1000 cm−1 region are due to the
presence of surface vanadium oxide species on the alumina. Sam-
ples VAl01 300 and VAl025 300 show major Raman line at around
950 cm−1, which is accompanied by weak lines at ∼560, ∼345
and ∼220 cm−1. These lines correspond to vanadium in polymeric
tetrahedral vanadates and are generated by surface metavanadate
(VO3)n species [16,17]. Higher V/Al ratios led to a shift of the major
line to higher wavenumbers and a new Raman line at 985 cm−1 for
VAl05 300 sample and at 1000 cm−1 for VAl07 300 and VAl1 300
samples. This major line is accompanied by a series of broad bands
at around 800, 580, 440 and 320 cm−1. According to the literature
data [17] we can assume that it accounts for the presence of
surface decavandate V10O286−species [16]. Another argument for
the presence of these species can be given with the aid of the
Pourboix diagram [18].
The presence of decavanadate species (HnV10O28) is thus justi-
fied by the preparation conditions where the precipitation pH was
5. Besides, low vanadium loadings correspond to higher aluminum
precursor loadings with a basic character. Accordingly, the basicity
of solution increased leading to the formation of vanadium poly-
meric species, such as (VO3)n metavanadates. With the increase
of the vanadium loading the acidity of the solutions increased thus
leading to the formation of polymeric species such as decavandates.
The presence of V2O5 for higher V/Al ratios (0.7 and 1) cannot be
excluded, and may be associated to the presence of characteristics
Raman lines (287, 310, 490, 540 cm−1).
This result is in agreement with the X-ray diffraction pat-
terns, which presented for the same calcination conditions planes
characteristic to V2O5. It is also important to note that the sam-
ple calcined at 500 ◦C (VAl07 500) showed lines characteristics
to V2O5 nanoparticles, which were not observed in XRD due to
the technique limitation (particles <4 nm in size) [19]. The broad
line centered at around 900 cm−1 denotes the presence of bridging
V
O Al bonds [20] that are formed by condensation of V OH and
Al OH during the co-precipitation step.
3.2. Catalytic behavior
Blank experiments in the absence of catalysts indicated no oxi-
dation of 2-MN. The only identified product was acetamide as a
result of the reaction of acetonitrile with hydrogen peroxide [21]
or by the hydrolysis of acetonitrile under the reaction condition.
However, these results infirm the literature reports indicating the
formation of VK3 even in the absence of a catalytic system [10,11].
On the contrary, 2-MN was oxidized to VK3 on the synthesized
catalysts and in order to optimize the reaction the effect of the cata-
lysts structural parameters and reaction conditions was separately
investigated.
3.2.1. Effect of the V/Al ratio
The oxidation of 2-MN with H2O2 over the catalysts calcined at
300 ◦C is presented in Fig. 6. The main products were VK3 and mena-
diol, but some other complex products, such as those shown in
According to Fig. 4, the samples VAl01 300 and VAl025 300
show a Raman line at 1052 cm−1, which can be assigned most
−
probably to surface residual NO3 and is characteristic only for
V2O5–Al2O3 catalysts with a low vanadium loading. This result is
confirmed also by elemental analysis (see Table 1), which point out
on the decrease of nitrogen percentage with the increase of the
vanadium–aluminum ratio.
The V2O5–Al2O3 dried samples also presented lines assigned to
surface metavanadate and decavanadate species and their relative
concentration depended on the vanadium loading. The influence
of the calcination temperature is presented from Raman spectra of
the VAl07 sample (Fig. 5). Calcination of the samples at elevated
temperatures was accompanied by a decrease of the surface area
(Table 1) that corresponded to an increase of the surface vana-
dium oxide coverage. Different vanadium species were identified
as a function of the vanadium loading and calcination tempera-
ture. After calcination at 500 or 700 ◦C the mixture of metavanadate
Fig. 6. Conversion of 2-MN and selectivity to VK3 and menadiol as a function of
the V/Al ratio (catalysts calcined at 300 ◦C, 50 mg catalyst, 2 mL acetonitrile, 80 ◦C,
substrate/H2O2 ratio 1:5, 24 h).
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
M. Florea et al. / Catalysis Today xxx (2015) xxx–xxx
5
Fig. 8. Influence of temperature on the catalytic performances of the VAl07 300
catalyst (50 mg catalyst, 2 mL acetonitrile, substrate/H2O2 ratio 1:5, 24 h).
indicating the fact that the crystallization process already began at
this temperature. Indeed, from more sensitive Raman technique it
was observed that the spectrum of the sample calcined at 500 ◦C
contains lines characteristics to V2O5 that were not evidenced in
the X-ray diffraction pattern of this sample.
It results that the highly dispersed metavandate and decavana-
date species present in amorphous XRD phase favor the selective
oxidation of 2-MN to VK3. Further decrease of the dispersion
increases the selectivity but exhibit a negative effect on the con-
version.
Scheme 3. Other oxidation products identified in 2-MN oxidation.
Scheme 3, were confirmed by GC–MS and H NMR analysis. As a gen-
eral tendency, the selectivity to VK3 increased with the vanadium
loading, achieving a maximum of 54% for the samples with a V/Al
ratio of 0.7. Higher vanadium loadings led to side products such as
2,3-epoxy-2-methyl-1,4-naphthoquinone and 3,3ꢀ-dimethyl-[1,1ꢀ]
binaphthylidene-4,4ꢀ-dione and as a consequence, the selectiv-
ity in VK3 decreased (32%, for VAl1 300 catalyst). These results,
in correlation with the Raman characterization, suggest that the
transformation of 2-MN is favored by the presence of decavana-
date species. They were well revealed for the VAl07 calcined at
300 ◦C. However, a further increase in the polymerized vanadate
species like for V/Al ratios higher than 0.7 is not favorable for the
reaction leading to a decrease of the yields in VK3. At the end of
each reaction, H2O2 was almost converted, largely exceeding the
amount required by the reaction stoichiometry as a consequence
of an over oxidation and of a non-selective decomposition to water
and oxygen. The efficiency of H2O2 was around 35% for the VAlO7
sample and decreased with the decreasing of the V/Al ratio.
3.2.3. Effect of the reaction temperature
Fig. 8 shows the performances of the investigated catalysts in
the range of temperature 40–100 ◦C. After 60 ◦C, hydrogen perox-
ide start to be decomposed non-selectively [22]. It appears that
the in situ generation of oxygen is benefice for both the con-
version and selectivity [23,24]. Higher temperatures than 80 ◦C
generates an additional activation of oxygen leading in the presence
of the investigated catalysts to side reactions, i.e. to over-oxidation
to 2,3-epoxy-2-methyl-1,4-naphthoquinone or coupling reactions
(Scheme 3), with a direct effect on the decrease of the selectivity
to VK3. The participation of the radicals leading to side reactions is
more probable for temperatures lower that 60 ◦C [25].
3.2.2. Effect of the calcination temperature
According to Fig. 7, within the investigated range of calcina-
tion temperatures, 300 ◦C resulted in the best catalytic activity and
selectivity to VK3 (54%). According to X-ray diffraction this calci-
nation temperature is leading to an amorphous structure, merely
constituted from surface metavanadate and decavandate species
(see above Raman results). Higher calcination temperatures where
the formation of V2O5 started, led to an important increase of selec-
tivity to vitamin K3, but with the price of a drastic decrease of the
conversion (8% for the sample calcined at 700 ◦C).
Comparative tests with commercial V2O5 led also to a very low
conversion (5%) and a high selectivity in VK3. XRD results also
proved the formation of V2O5 for the samples calcined at 700 ◦C,
thus explaining the low conversion of VAl07 700. A low conver-
sion (4%) was also determined for the catalyst calcined at 500 ◦C,
3.2.4. Effect of the substrate/H2O2 ratio
Data shown in Fig. 9 indicate that the catalytic performances
are influenced by the substrate/H2O2 ratio. When the reaction was
carried out with a substrate/H2O2 ratio of 1:5 a maximum conver-
sion of 76% was measured corresponding to a selectivity inVK3 of
54%. A higher substrate/H2O2 ratio (i.e. 1:3) led to a drastic decrease
of the conversion till 17% that was accompanied by a decrease in
the selectivity to VK3. Interestingly enough, when increasing the
substrate/H2O2 ratio to 1:7 and 1:9, respectively, the conversion
remained unchanged (44%), but the selectivity to VK3 decreased
to 26%. Thereby, an excessive content in oxidant leads to a fur-
ther oxidation of VK3 to advanced oxidation products issued from
Fig. 7. Influence of the calcination temperature on the catalytic performances
(50 mg catalyst VAl07, acetonitrile, substrate/H2O2 ratio 1:5, 24 h).
Fig. 9. Influence of substrate/H2O2 ratio on the catalytic performances of the
VAl07 300 catalyst (50 mg catalyst, 2 mL acetonitrile, 80 ◦C, 24 h).
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
6
M. Florea et al. / Catalysis Today xxx (2015) xxx–xxx
Fig. 10. Influence of the solvent on catalytic performances using the VAl07 300
catalyst (50 mg catalyst, substrate/H2O2 ratio 1:5, 80 ◦C, 24 h).
Scheme 4. Type of vanadium entities as a function of the V/Al ratio.
efficient catalyst in this study was considered a preparation proce-
dure in which vanadium and aluminum were co-precipitated.
Scheme 4 shows the evolution of the vanadium entities in the
investigated catalysts as a function of the V/Al ratio. This evolution
was already verified in oxidation of o-xylene to phtalic anhydride
on vanadia-based catalysts [26], where the catalytic properties
demonstrated that the surface vanadium oxide phase is respon-
sible for the overall catalytic activity and selectivity, while the
crystalline V2O5 phase exhibits only a minimal effect. This is also
in concordance to recent reports of Wachs [14] showing that the
surface vanadates are the catalytic active sites for oxidation reac-
tions and that both isolated and polymeric surface vanadia species
exhibit the same specific catalytic activity, compared with larger
crystalline V2O5 nanoparticles that tend to be less active.
Comparing the hydrated and dehydrated (i.e. uncalcined versus
calcined) samples, one can observe that the catalytic activity is dras-
tically affected by the thermal treatment. The most active are the
samples calcined at 300 ◦C, as is presented in Fig. 7, where coexists a
mixture of metavanadate (Raman line at 940 cm−1) and decavana-
date (Raman line at 990 cm−1) species.
Fig. 11. Leaching tests (run (I) 50 mg of VAl07 300 catalyst, substrate/H2O2 ratio
1:5, 2 mL acetonitrile, 80 ◦C, 24 h; run (II) separated solution from run (I) without
catalyst, 80 ◦C, 24 h.
the methyl group oxidation. On the other side, the formation of
larger amounts of water may be a retarding factor since water will
participate in a competitive adsorption on the catalysts active cen-
ters hindering the formation of vitamin K3. It appears thus that an
optimal substrate/H2O2 ratio corresponds to a value of 1:5.
3.2.5. Influence of the solvent
Fig. 10 depicts the influence of various solvents on the catalytic
oxidation of 2-MN over the investigated V2O5–Al2O3 catalysts.
While in acetonitrile VAl07 provided a high conversion and selec-
tivity to VK3, in 1,4-dioxane or methanol the performances were
rather poor. It is however difficult to correlate these performances
with the physical properties of the investigated solvents. It appears
that oxygenated solvent concurs with the substrate to the interac-
tion of the active sites. This process occurs without any additional
oxidation of the solvent and only with the oxidation to menadione.
5. Conclusions
The results of this study indicate the possibility to enhance both
conversion and selectivity in the synthesis of vitamin K3 through
the selective oxidation on heterogeneous catalysts. Particularly for
the vanadia–alumina catalysts, the oxidation of 2-MN is favored
by the co-existence of deca- and metavanadate species, with an
optimal V/Al ratio of 0.7.
3.2.6. Leaching tests
Leaching of the vanadium species in liquid phase oxidations is a
common problem questioning the heterogeneous behavior of the
catalysts. In the present experiments leaching was checked by run-
ning successive reactions as shown in Fig. 11. In the typical leaching
checking procedure the catalyst was separated from the reaction
mixture, and the mother liquor was allowed to react further with-
out catalyst (run II). The conversion of the 2-MN and selectivity of
vitamin K3 after run II were 76% and 41%, respectively. According
to these results the leaching was insignificant.
Acknowledgements
The authors acknowledge UEFISCDI for the financial support
through the project PN-II-TE105/2011.
References
[1] O.A. Kholdeeva, O.V. Zalomaeva, A.N. Shmakov, M.S. Melgunov, A.B. Sorokin, J.
Catal. 236 (2005) 62–68.
[2] W. Bonrath, T. Netscher, Appl. Catal., A: Gen. 280 (2005) 55–73.
[3] R.A. Sheldon, J. Dakka, Catal. Today 19 (1994) 215–245.
[4] S. Biella, G.L. Castiglioni, C. Fumagalli, L. Prati, M. Rossi, Catal. Today 72 (2002)
43–49.
[5] N. Lebrasseur, G.-J. Fan, M. Oxoby, M.A. Looney, S. Quideau, Tetrahedron 61
(2005) 1551–1562.
[6] W.A. Hermann, J.J. Haider, R.W. Fischer, J. Mol. Catal. A: Chem. 138 (1999)
115–121.
[7] S. Yamaguchi, M. Inone, S. Enomoto, Chem. Lett. (1985) 827–828.
[8] C. Pergrale, A.B. Sorokin, C.R. Acad. Sci. Paris, Ser. IIc: Chim. 3 (2000) 803–810.
[9] R. Song, A. Sorokin, J. Bernadou, B. Meunier, J. Org. Chem. 62 (1997) 673–678.
4. General assessments
Generally, bulk vanadium oxides are not proper to be used in
industrial processes, as they present poor thermal stability and
mechanical strength [15]. Therefore in many selective oxidation
reactions, vanadium oxide is deposited on oxides such as Al2O3,
ZrO2, SiO2 or TiO2. The surface structures of supported vanadia
on these oxides are quite different to that of bulk. To achieve an
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026

G Model
CATTOD-9409; No. of Pages7
ARTICLE IN PRESS
M. Florea et al. / Catalysis Today xxx (2015) xxx–xxx
7
[10] O.A. Anunziata, A.R. Beltramone, J. Cussa, Appl. Catal., A: Gen. 270 (2004) 77–85.
[11] O.A. Anunziata, L.B. Pierella, A.R. Beltramone, J. Mol. Catal., A: Chem. 149 (1999)
255–261.
[20] N. Magg, B. Immaraporn, J.B. Giorgi, T. Schroeder, M. Baumer, J. Dobler, Z. Wu,
E. Kondratenko, M. Cherian, M. Baerns, P.C. Stair, J. Sauer, H.J. Freund, J. Catal.
226 (2004) 88–100.
[12] J. Skarzewski, Tetrahedron 40 (1984) 4997–5000.
[13] K.I. Matveev, V.F. Odyakov, E.G. Zhizhina, J. Mol. Catal., A: Chem. 114 (1996)
151–160.
[21] Y. Ogata, Y. Sawaki, Tetrahedron 20 (1964) 2065–20168.
[22] G. Zi, D. Chen, B. Li, Z. Li, X. Luo, J. Zhang, L. Li, J. Wang, Catal. Commun. 49 (2014)
10–14.
[14] I.E. Wachs, Dalton Trans. 42 (2013) 11762–11769.
[15] B.M. Weckhuysen, D.E. Keller, Catal. Today 78 (2003) 25–46.
[16] G. Deo, F.D. Hardcastle, M. Richards, I. Wachs, Am. Chem. Soc., Div. Polym. Chem.
34 (3) (1989) 529–534.
[17] G. Deo, I. Wachs, J. Phys. Chem. 95 (1991) 5889–5895.
[18] C.F. BaesJr, R.E. Mesmer, The Hydrolysis of Cations, Wiley, New York, NY, 1970.
[19] I.E. Wachs, C.A. Roberts, Chem. Soc. Rev. 39 (2010) 5002–5017.
[23] S. Narayanan, K.V.V.S.B.S.R. Murthy, K. Madhusudan Reddy, N. Premchander,
Appl. Catal., A: Gen. 228 (2002) 161–165.
[24] D.V. Eremin, L.A. Petrov, Russ. J. Appl. Chem. 82 (5) (2009) 866–870.
[25] G. Strukul, F. Somma, N. Ballarini, F. Cavani, A. Frattini, S. Guidetti, D. Morselli,
Appl. Catal., A: Gen. 356 (2009) 162–16626.
[26] B. Grzybowska-Swierkosz, Appl. Catal., A: Gen. 157 (1997) 263–310.
Please cite this article in press as: M. Florea, et al., Mesostructured vanadia–alumina catalysts for the synthesis of vitamin K3, Catal.
Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.026