Dehydration of Glycerol to Acrolein over Sulfated Iron Oxide Catalysts

Article abstract of DOI:10.1134/S0965544119090111

Abstract: The problem of disposal of glycerol, the main waste product in the production of biodiesel, has been explored. It has been shown that cheap catalysts on the basis of iron(III) oxide can be use for mediating the transformation of glycerol to acro

Full text of DOI:10.1134/S0965544119090111

ISSN 0965-5441, Petroleum Chemistry, 2019, Vol. 59, No. 9, pp. 988–993. © Pleiades Publishing, Ltd., 2019.
Russian Text © The Author(s), 2019, published in Neftekhimiya, 2019, Vol. 59, No. 5, pp. 538–543.
Dehydration of Glycerol to Acrolein
over Sulfated Iron Oxide Catalysts
P. G . M i n gal ev a, A. G. Aslanlya, and G. V. Lisichkina, *
aFaculty of Chemistry, Moscow State University, Moscow, 119234 Russia
*e-mail: Lisich@petrol.chem.msu.ru
Received June 9, 2018; revised April 2, 2019; accepted May 13, 2019
Abstract—The problem of disposal of glycerol, the main waste product in the production of biodiesel, has
been explored. It has been shown that cheap catalysts on the basis of iron(III) oxide can be use for mediating
the transformation of glycerol to acrolein—an important intermediate product of industrial organic synthesis.
The influence of external factors on the process of dehydration of glycerol to acrolein has been examined.
Keywords: disposal of glycerol, acrolein, iron(III) oxide, acid catalysis
DOI: 10.1134/S0965544119090111
As is known, the main process for production of bio- The resulting methyl esters of fatty acids represent biodiesel.
diesel [1] includes transesterification of vegetable oils with The byproduct glycerol is formed in significant amounts,
methanol over a base or acid catalysts as the key step [2]. namely, about 100 kg of glycerol per ton of biodiesel.
R
CH2 OOC
CH OOC
CH2 OOC
CH2 OH
CH OH
CH2 OH
OH−/Η+
R.
3MeOOC
+
R + 3MeOH
R
The main obstacle restraining wide use of biodiesel pressure and temperature) or using catalysts contain-
is the need for the withdrawal of significant areas from ing relatively scarce elements (zirconium, tungsten,
the agricultural use and a high cost [3]. Because of etc.). At the same time, as has been noted above, not
this, it is important to use all the products formed in simply the development of the ways for disposal of
the process of production of biodiesel, in particular, glycerol, but the search for cheaper methods of its dis-
glycerol, fullest possible. However, the amount of posal is of the greatest relevance. Therefore, the
glycerol obtained as the waste in the production of bio- researchers face the task of maximally reducing the
diesel many times exceeds the demand for it. Accord- cost of each stage of the glycerol processing scheme, in
ingly, the development of chemical processes utilizing particular, searching for the cheapest catalysts for this
glycerol as feedstock becomes a relevant task.
processing.
The conversion of glycerol to acrolein occurs over
acid catalysts. One of the cheapest and abundant acid
catalyst is iron(III) oxide additionally treated with an
acid, e.g., sulfuric acid [12]. In this work, we studied
the applicability of several such catalysts to the dehy-
dration reaction of glycerol to acrolein.
One of the most valuable substances synthesized
from glycerol, acrolein, is an important intermediate
product of modern chemical industry [4]. Acrylic acid
and its nitrile and esters required for the production of
various plastics and polymer fibers as well as essential
amino acid D,L-methionine used as a feed additive and
its analogue, 2-hydroxy-4-methylthiobutyric acid
(hydroxy analogue of methionine, HAM), are obtained
from it.
Many works are devoted to the investigation of the
methods of transformation of glycerol to acrolein [5–
11]. Various acid catalysts are used for the execution of
this reaction. However, good yields of acrolein are
generally achieved either under harsh conditions (high
EXPERIMENTAL
Synthesis of Iron(III) Oxide
A 16.9-g portion of FeCl3 · 6H2O was dissolved in
50 mL of distilled water with constant stirring using a
magnetic stirrer. Then a solution of 15.75 g of NaHCO3
in 190 mL of H2O was added dropwise from a drop-
988

DEHYDRATION OF GLYCEROL TO ACROLEIN
ping funnel. After this, the precipitate was aged by Catalyst Preparation
989
boiling for 1.5 h and left overnight. The precipitate was
rinsed with distilled water six times and centrifuged at
10000 rpm. The obtained sample was dried in an oven
for 5 h at 200°C. After this, the catalyst was ground to
a powder in a mortar. The fine powder fraction was
separated by means of sedimentation in distilled water.
Samples of the Fe2O3 · H2SO4 catalyst were obtained
on the basis of iron oxide supports (Ssp = 53 m2/g and
Vpore = 0.6 mL/g; Ssp = 364 m2/g and Vpore = 0.83 mL/g)
(Tables 1 and 2) by mixing a weighed portion of the
oxide with the diluted acid. The concentration of the
acid was chosen in such a way that the volume of the
added acid would correspond to the pore volume of
the taken amount of the oxide. After this, the catalyst
was dried for 2 h at 160°C, then heated at a rate of
5°C/min to the corresponding temperature, and held
at this temperature for the required time.
Synthesis of Nickel(II) Oxide
An 18-g portion of NiCl2 · 6H2O was dissolved in
50 mL of distilled water with constant stirring using an
electrical stirrer. Then a solution of 11 g of NaOH in
25 mL of H2O was added dropwise using a dropping
funnel. The mixture was continuously stirred for 1 h
and left overnight. To purify the precipitate from salts,
it was rinsed with distilled water five times and centri-
fuged at 10000 rpm. The obtained sample was dried in
an oven for 6 h at 230°C. After this, the catalyst was
ground to a powder in a mortar. The fine powder frac-
tion was separated by means of sedimentation in dis-
tilled water.
Another set of Fe2O3 · H2SO4 catalysts was pre-
pared on the basis of the iron oxide support with Ssp =
364 m2/g and Vpore = 0.83 mL/g (Table 2).
NiO · H2SO4 dehydration catalysts were prepared
on the basis of a nickel oxide support (Ssp = 150 m2/g
and Vpore = 0.8 mL/g) (Table 3). The catalyst prepara-
tion procedure was the same as in the case of iron
oxide supports.
Glycerol Dehydration Procedure
Characterization
Glycerol was added to a catalyst in a flask; the cat-
alyst : glycerol weight ratio was 1 : 20. Then this mix-
ture was distilled in a Claisen flask, and the distillate
was collected. Solid resin was left in the flask after dis-
tillation.
The specific surface area of the samples was deter-
mined on a Sorbtometr-M instrument via thermal
desorption of nitrogen. To calculate the specific sur-
face area, the Brunauer–Emmett–Teller (BET) equa-
tion was used.
The crystal structure of the initial support and pre-
pared catalysts was determined via X-ray diffraction
analysis (XRD) on a DRON-3M instrument (cobalt
Kα radiation, having a wavelength of 0.179021 nm).
Gas Chromatographic Analysis
The analysis was performed on a Kristall-Lyuks gas
chromatograph equipped with a flame ionization
detector. A capillary column with a length of 20 m and
an internal diameter of 0.22 mm was used. The FFAP
stationary phase (polyethylene glycol esterified by
nitroterephthalic acid) was used. The stationary-phase
film thickness was 0.25 µm. The carrier gas was helium.
For the first 2 min, the elution was carried out at 30°C,
and then the column was heated to 200°C at a rate of
20°/min. The duration of one analysis was 15 min.
The obtained products were analyzed in the form of
10% solutions in ethanol. For quantitative measure-
ments, a model mixture consisting of 80 wt % ethanol,
10 wt % glycerol, and 10 wt % acrolein was injected
into the chromatograph with a microsyringe. To deter-
The acid properties of the samples were studied via
temperature programmed desorption of ammonia
(TPD NH3) on a Unisit USGA-101 sorption analyzer.
The sample was calcined in a dried helium flow at 400
or 600°C and then cooled to 60°C. The adsorption of
ammonia was performed for 30 min at 60°C; ammo-
nia was diluted with nitrogen in a ratio of 1 : 1. Phy-
sisorbed ammonia was removed by flushing with dry
helium at 100°C for 1 h. The experiments on TPD
NH3 were performed in the temperature range of 60
to 800°C in a dry helium stream (at a flow rate of
30 mL/min). The heating rate was 8°C/min.
Determination of the total pore volume by the drop mine the characteristics of hydroxyacetone, a 10%
method. Distilled water was added dropwise to a finely solution of hydroxyacetone in ethanol was used. The
powdered precipitate using a pipette while shaking the retention times of glycerol, hydroxyacetone, and acro-
sample and achieving the destruction of the initially lein were determined by the obtained chromatograms,
formed lumps of the wet support. The addition contin- as well as the relative sensitivity of the detector to glyc-
ued until the lumps of the support stopped to be erol, hydroxyacetone, and acrolein in comparison
destroyed by shaking, which gave the evidence of the with ethanol. The sensitivity of the detector to acro-
complete saturation of the pores with water. The pore lein, hydroxyacetone, or glycerol was 0.67, 1.5, or 1.6
volume was judged by the added volume of water.
of the sensitivity to ethanol.
PETROLEUM CHEMISTRY Vol. 59 No. 9 2019

990
MINGALEV et al.
Table 1. Conditions of preparation of the catalysts on the basis of iron oxide with a specific surface area of 53 m2/g
Number of acid molecules
Sample no.
Calcination temperature, °C
Calcination time, h
per 1 nm2 of surface
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
10
10
10
5
5
5
5
5
5
5
200
400
600
200
200
200
400
400
400
600
600
600
200
400
600
6
6
6
2
4
6
2
4
6
2
4
6
6
6
6
5
5
3
3
3
Table 2. Conditions of preparation of the catalysts on the basis of iron oxide with a specific surface area of 364 m2/g
Number of acid molecules
per 1 nm2 of surface
Calcination
temperature, °C
Sample no.
Calcination time, h
Catalyst mass, mg
1
2
3
4
5
6
3
3
3
3
6
6
200
600
200
600
200
200
2
6
2
6
2
2
120
120
20
20
120
20
Table 3. Conditions of preparation of the catalysts on the basis of nickel oxide
Number of acid molecules
Sample no.
Calcination temperature, °C
Calcination time, h
per 1 nm2 of surface
1
2
3
4
3.0
3.0
1.5
1.5
400
600
400
600
6
6
6
6
RESULTS AND DISCUSSION
Dehydration over Iron Oxide Catalysts
support samples of with specific surface areas of 53
and 364 m2/g were prepared.
According to the XRD data, all the studied samples
are pure hematite (α-Fe2O3), no other phases have
been detected in any of the cases.
It has been found that the activity and selectivity of
the samples with a high specific surface area is notice-
ably higher, and the yield of acrolein can reach 21%
under otherwise equal conditions. However, it
should be noted that the synthesis of the iron oxide
samples with such a high value of the specific surface
The influence of the specific surface area of catalysts
on the dehydration reaction. To study the character of
the influence of the specific surface area of the catalyst
on the occurrence of the dehydration reaction, two area (364 m2/g) is distinguished by low reproducibil-
PETROLEUM CHEMISTRY
Vol. 59
No. 9
2019

DEHYDRATION OF GLYCEROL TO ACROLEIN
991
Table 4. The influence of the amount of acid on the results of dehydration of glycerol (iron oxide catalysts, Ssp of 53 m2/g)*
Number of acid molecules
per 1 nm2 of surface
Catalyst calcination
temperature, °C
Yd, %
Ya, %
Yh, %
10
10
10
5
5
5
3
3
3
200
400
600
200
400
600
200
400
600
77.0
51.0
65.0
32.0
38.0
68.0
17.0
16.0
20.0
2.3
7.8
0.2
6.2
3.2
2.5
0.2
0.2
0.2
0.3
1.1
0.3
0.3
0
0.3
0.2
0.1
0.1
* Y is the yield of the distillate, Y is the yield of acrolein, and Y is the yield of hydroxyacetone.
d
a
h
Table 5. The influence of the temperature and time of calcination of the iron oxide catalysts on the results of dehydration
of glycerol*
Yd, %
Ya, %
Yh, %
Calcination temperature, °C Calcination time, h
200
200
200
400
400
400
600
600
600
2
4
6
2
4
6
2
4
6
74.0
54.0
32.0
49.0
54.0
38.0
28.0
21.0
680
2.7
0.2
6.2
8.1
11
3.2
4.5
7.5
2.5
0.4
0.2
0.3
1.4
2.3
0
1.2
0.1
0.3
* Y is the yield of the distillate, Y is the yield of acrolein, and Y is the yield of hydroxyacetone.
d
a
h
ity. Because of this, we cannot recommend such sup- acrolein oligomerization products. Because of this,
ports for the glycerol dehydration reaction until repro- further discussion will only concern the samples with
ducible synthesis methods of them are developed. the surface concentration of the acid of five molecules
per 1 nm2 of the support surface.
Further discussion will only concern the samples with
a specific surface area of 53 m2/g, which are character-
ized by high reproducibility.
The influence of the catalyst calcining temperature
and time. The results of the dehydration reaction over
the catalysts prepared under different conditions are
presented in Table 5.
It has been found as a result of the studies that the
dependence of the activity and selectivity of the cata-
lyst on the temperature of calcination passes through a
maximum, i.e., the yield of both the distillate and
acrolein increases. The highest yield is observed at a
calcination temperature of 400°C. For this tempera-
ture, the dependence of the catalyst activity and selec-
tivity also passes through a maximum. The investiga-
The influence of the amount of acid supported on the
catalysts on the yield of acrolein. To find out this influ-
ence, a series of samples with the number of acid mol-
ecules per 1 nm2 of the surface from three to ten was
synthesized. The obtained catalysts were tested in the
dehydration reaction; the results are presented in
Table 4. The time of calcination of the samples was 6 h.
As is seen from the table, samples with the formal
surface concentration of the acid of five molecules per
1 nm2 possess the best properties in almost all the tion of the catalysts using the ammonia TPD method
cases. Samples with a lower concentration of the acid shows that the best yields are achieved for the catalysts
turned out to be almost inactive in all the cases, appar- with a high relative concentration of strong acid sites.
ently, due to the too low acidity. Regarding the sam- The yields over the samples with a high relative con-
ples with a higher acidity, the yield of the distillate is centration of weak acid sites are low. This difference is
also higher in this case, but it mainly contains the apparently due to the fact that the dehydration of glyc-
PETROLEUM CHEMISTRY Vol. 59 No. 9 2019

992
MINGALEV et al.
Table 6. The influence of the temperature and time of calcination of the catalysts based on the nickel oxide support (Ssp =
150 m2/g and Vpore = 0.8 mL/g) on the results of dehydration of glycerol*
Number of acid molecules per 1 nm2
Yd, %
Ya, %
Yh, %
Calcination temperature, °C Calcination time, h
3.0
3.0
1.5
1.5
400
600
400
600
6
6
6
6
30.0
31.0
33.0
50.0
2.7
6.9
4.0
4.9
0.2
0.6
0.4
0.5
* Y is the yield of the distillate, Y is the yield of acrolein, and Y is the yield of hydroxyacetone.
d
a
h
erol requires quite strong acid sites [4]; however, the
side reactions (of the polymerization type) may well be
accelerated by weak acid sites.
Dehydration over Nickel-Containing Catalysts
According to the XRD data, all the studied samples
are pure bunsenite (α-NiO), no other phases were
detected in any of the cases.
It has also been found that the presence of iron sul-
fate on the surface of the catalyst is very important as
well. This compound was not detected by XRD; how-
ever, it can be assumed that the decomposition of this
compound is responsible for the peak in the TPD
spectra at temperatures of about 600°C. The yields of
the desired product (as well as hydroxyacetone) over
the catalysts free from iron sulfate are extremely low
despite the presence of a sufficient number of strong
acid sites in the sample.
To explain this effect, it can be assumed that the
active sites of the catalyst are Brønsted sites, and iron
sulfate acts as a Lewis acid in this case forming donor–
acceptor complexes with the glycerol molecule.
Apparently for steric reasons., iron sulfate complexes
with the terminal OH groups of glycerol molecules are
predominantly formed. This structure facilitates the
production of acrolein because its formation results
from the protonation of a glycerol molecule by the
middle hydroxyl, which remains free in such com-
plexes. Note that glycerol is a stronger Lewis base than
acrolein. Because of this the formation of the com-
plexes of glycerol with Lewis acid sites should preclude
the adsorption of acrolein on the catalyst surface and,
hence, impede the undesirable side reactions with the
participation of acrolein. To verify this assumption,
samples on the basis of nickel oxide were prepared and
studied (see below).
The influence of the acid concentration and calcina-
tion temperature on the catalyst activity and selectivity.
To study the influence of the concentration of the acid
on the dehydration reaction, two series of catalysts
with the amount of the acid of 3 and 1.5 molecules on
the surface were prepared (Table 6). The catalyst with
the amount of the acid on the surface of 3 g/nm2
turned out to be much more selective for acrolein. It
has also been found that increasing the concentration
of the acid in the catalyst composition leads to a sharp
decrease in the distillate yield. The samples of the dis-
tillate obtained in this case were not subjected to chro-
matographic analysis in view of their small amount.
To reveal the influence of the specified factors on
the occurrence of the reaction, two sets of experiments
were performed (see Table 6), one set at 400°C and the
other at 600°C. It was found in these experiments that
the activity and selectivity of the catalyst increases
with an increase in the calcination temperature.
It was very important to get an answer to the ques-
tion about the role of Lewis acid sites on the surface.
As has been mentioned above, the presence of iron
sulfate on the surface is the prerequisite for the occur-
rence of the desired reaction over the iron oxide cata-
lysts. This fact can be explained by the assumption that
iron sulfate acts as a Lewis acid in this case (see above).
To verify this assumption, catalysts on the basis of
nickel oxide doped with sulfuric acid were studied
using TPD of ammonia. Nickel is a chemical analogue
of iron; these elements are very close in their chemical
properties. However, nickel is the most stable in the
divalent rather than the trivalent state, in contrast to
iron. The radius of the Ni2+ ion is larger than the
radius of the Fe3+ ion (0.069 versus 0.064 nm), and the
charge is smaller. Hence, nickel(II) sulfate is a notice-
ably weaker Lewis acid than iron(III) sulfate.
It should also be noted that the yields of hydroxy-
acetone over our samples are quite low—they are
always lower than the yields of acrolein. Probably, this
is associated with the instability of hydroxyacetone in
an acidic medium; it can condense according to the
aldol–crotonic type acting as both the carbonyl and
methylene components. Acrolein, however, can only
act as the carbonyl component in this reaction; hence,
it requires third-party molecules for the condensation
by this route.
Thus, the formation of acid sites with the required
It has been found that the number and strength of
strength on the surface requires the calcination of the acid sites on the nickel oxide catalysts are approxi-
samples, however, not very severe.
mately the same as those of the iron oxide catalysts.
PETROLEUM CHEMISTRY
Vol. 59
No. 9
2019

DEHYDRATION OF GLYCEROL TO ACROLEIN
993
However, the yields of acrolein and hydroxyacetone
over these catalysts are noticeably lower than in the
case of the iron oxide catalysts. Thus, it can be con-
cluded that the presence of Lewis acid sites on the sur-
face of the catalyst in addition to strong Brønsted acid
sites is indeed the prerequisite for the successful
occurrence of the reaction. In addition, it has been
shown that the yield of acrolein is higher for the sam-
ple containing a greater number of strong acid sites
(calcination at 600°C), which is in agreement with the
results obtained for the iron oxide catalysts.
The presence of iron sulfate on the surface of iron
oxide catalysts leads to an increase in the product
yield. In its absence, the selectivity of the catalysts is
extremely low despite the presence of strong acid sites.
REFERENCES
1. F. R. Ma, M. A. Hanna, and E. K. Wilson, Chem. Eng.
News 70, 1 (1999).
2. L. C. Meher, D. V. Sagar, and S. N. Naik, Renew. Sust.
Energy Rev. 10, 248 (2006).
3. V. S. Arutyunov and G. V. Lisichkin, Usp. Khim. 86,
777 (2017).
In summary, sulfated iron(III) oxide exhibits activ-
ity in the dehydration reaction of glycerol. The yield of
acrolein over the reproducibly obtained samples is
about 11 wt %. It can be increased by increasing the
specific surface area of the catalyst. However, there is
the problem of low reproducibility of the synthesis of
iron oxide samples with a high specific surface area.
This difficulty is due to the fact that the texture of iron
oxide depends on a great number of synthesis param-
eters the influence of which has been little studied.
4. E. Tsukuda, S. Sato, R. Takahashi, and T. Sodesawa,
Catal. Commun. 8, 1349 (2007).
5. S. Niu, Y. Zhu, H. Zheng, et al., Chin. J. Catal. 32, 345
(2011).
6. Y. T. Kim, K.-D. Jung, and E. D. Park, Appl. Catal., A
393, 275 (2011).
7. S. M. Danov, A. S. Esipovich, A. S. Belousov, and
A. E. Rogozhin, Khim. Prom-st. Segodnya, No. 2, 15
(2014).
8. L.-Z. Tao, S.-H. Chai, Y. Zuo, et al., Catal. Today 158,
It has been found that increasing the concentration
of the acid on the surface leads to a growth in the activ-
ity of the catalyst in the glycerol dehydration reaction;
however, the selectivity for the desired product drops
in this case. It has been shown that the presence of
strong acid sites (apparently, of the Brønsted type) on
the catalyst surface is important for the dehydration of
glycerol to acrolein, whereas the presence of weak acid
sites leads to a decrease in the selectivity of the cata-
lysts. The formation of strong acid sites on the surface
requires the calcination of the catalyst.
310 (2010).
9. S. Ramayya, A. Brittain, C. De Almeida, et al., Fuel 66,
1364 (1987).
10. A. Takanori and Y. Masayuki, JP Patent Appl.
No. 2009292773 (2009).
11. A. S. de Oliveira, S. J. S. Vasconcelos, J. R. de Sousa,
et al., Chem. Eng. J. 168, 765 (2011).
12. O. V. Krylov, Heterogeneous Catalysis (Akademkniga,
Moscow, 2004) [in Russian].
Translated by E. Boltukhina
PETROLEUM CHEMISTRY Vol. 59 No. 9 2019