Catalysis by coke deposits: Synthesis of isoprene over solid catalysts

Article abstract of DOI:10.1002/anie.201307083

A help rather than a hindrance: Carbonaceous deposits have been found to play a key role in the selective synthesis of isoprene from formaldehyde and isobutene over solid catalysts. They accumulate on the catalyst surface during the induction period and promote the interaction of the substrates at the steady state. The proposed mechanism (see scheme) shows the way forward for the design of efficient solid catalysts for the synthesis of isoprene. Copyright

Full text of DOI:10.1002/anie.201307083

Angewandte
Chemie
DOI: 10.1002/anie.201307083
Heterogeneous Catalysis
Catalysis by Coke Deposits: Synthesis of Isoprene over Solid Catalysts
Irina Ivanova,* Vitaly L. Sushkevich, Yury G. Kolyagin, and Vitaly V. Ordomsky
The formation of coke deposits on solid catalysts is a problem
of great concern for many catalytic processes, since it leads to
catalyst deactivation. Nevertheless, in some catalytic applica-
tions it can play a positive role in the generation of highly
efficient active sites and in the formation of active “working”
catalysts.[1] The concept of catalysis by carbonaceous deposits
was formulated more than 40 years ago.[2] According to this
concept, the catalytically active surface of a catalyst is formed
during coke deposition, and the active sites of the “working”
catalyst are quite different from those of the fresh catalyst and
involve reactive carbonaceous species, such as surface carbe-
nium ions or radicals. Among the catalytic reactions which
obey this concept are such well-known industrial processes as
“methanol to olefins” (MTO),[1a–c] the skeletal isomerization
of n-butenes,[1d] the hydroisomerization of long-chain n-
alkanes,[1e] and the alkylation of toluene[1f] and naphthale-
ne.[1g] These processes were the subject of a comprehensive
review by Guisnet.[3] A general feature of all these processes
is the existence of the induction period, during which the
active surface species are formed. The identification of such
species, unraveling of the mechanism of the induction period,
and the elucidation of the chemical process that takes place
after the induction period are the main steps required for the
understanding of catalyst deactivation.
Herein, we report on the key role of carbonaceous
deposits in the synthesis of isoprene from formaldehyde and
isobutene over solid catalysts. Today, isoprene is mainly
produced by extraction from the C5 fraction of naphtha steam
cracking, and therefore its output is entirely dependent on
ethylene production. However, ethylene producers are
switching to more efficient technologies for ethylene produc-
tion. This trend will lead to a lack of isoprene on the market;
thus, alternative technologies are required for its synthesis.
The condensation of formaldehyde with isobutene is among
the most attractive alternative technologies for the synthesis
of isoprene.
2) the vapor-phase decomposition of 4,4-dimethyl-1,3-diox-
ane (DMD) to isoprene over solid phosphoric acid catalysts.[4]
This process has many drawbacks involving economical,
technological, and environmental issues, such as low effi-
ciency and selectivity, corrosion, and separation problems.
Therefore, the future challenge in this area is the develop-
ment of a one-step selective process based on solid catalysts.
The solid catalysts studied so far in the condensation of
isobutene with formaldehyde include zeolites,[5] oxides,[6]
sulfates,[7] phosphates,[8] and heteropolyacids.[9] Although
some of these catalytic systems showed rather high initial
yields of isoprene (60–70%) in the direct synthesis,[5–9] they
exhibited rather low selectivity and very short life times owing
due to rapid deactivation because of fast coke formation.
Analysis of the results indicates that for the most active
catalysts, there is usually an induction period before the
activity reaches the steady state.[6–9]
These observations brought us to the hypothesis that
carbonaceous deposits formed during the induction period
may play an important role in the one-step synthesis of
isoprene. In this study, we aimed to clarify the nature of
carbonaceous deposits formed during the induction period
and their role in the mechanism of condensation of formal-
dehyde with isobutene over niobium oxide, which is among
the most active and selective catalysts for the Prins con-
densation.[6c]
To prepare the active niobium oxide catalyst with a high
surface area and high acidity, we subjected a commercial
Nb2O5 sample to redispersion and hydration by a dissolution–
precipitation procedure described by Ushikubo et al.[10] The
catalyst obtained had a surface area of 70 m2 gÀ1 and an acid-
sites content of 240 mmolgÀ1, as determined by measurement
of the temperature-programmed desorption of ammonia.
The conversion of isobutene and formaldehyde over
hydrated niobium oxide was studied in a continuous flow
reactor at 573 K in the presence of excess isobutene (C4H8/
CH2O 7:1; a more detailed description of the preparation and
characterization of the catalyst and experimental techniques
is given in the Supporting Information). The yields of the
main reaction products are plotted versus time on stream
(TOS) in Figure 1.
The formation of isoprene is preceded by an induction
period. During this period, the main reaction products are
carbon monoxide, hydrogen, and methylbutenes (MBs).
Carbon monoxide and hydrogen are formed by the decom-
position of formaldehyde over Nbd+–OdÀ ion pairs.[8b,11] The
rapid decrease in the yield of carbon monoxide with TOS can
be explained either by surface hydration and the transforma-
tion of Lewis acid sites into Brønsted acid sites, which are
responsible for condensation reactions,[8b] or by rapid poison-
ing of the active sites by carbonaceous deposits.
The main industrial process based on this route involves
two steps: 1) the liquid-phase condensation of isobutene with
formaldehyde under the catalysis of aqueous sulfuric acid and
[*] Prof. I. Ivanova, V. L. Sushkevich, Y. G. Kolyagin, V. V. Ordomsky
Department of Chemistry, Lomonosov Moscow State University
Leninskye Gory 1, bld. 3, 119991 Moscow (Russia)
E-mail: iiivanova@phys.chem.msu.ru
Prof. I. Ivanova, Y. G. Kolyagin
A.V. Topchiev Institute of Petrochemical Synthesis
Russian Academy of Science
Leninskiy prospect, 29, 119991 Moscow (Russia)
V. L. Sushkevich, V. V. Ordomsky
“Unisit” LLC
Moscow (Russia)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201307083.
Angew. Chem. Int. Ed. 2013, 52, 12961 –12964
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12961

.
Angewandte
Communications
Figure 2. Variation in the coke content of the Nb2O5 catalyst with TOS
and DTG curves obtained after 20, 50, and 300 min on stream.
the differential thermogravimetric (DTG) peaks shifted
towards higher temperatures, which points to the trans-
formation of soft coke into hard coke.[12]
Figure 1. Yield of products plotted versus TOS for the reaction of form-
aldehyde with isobutene over hydrated Nb2O5 (isobutene/formalde-
More detailed information on the nature of coke accu-
mulated during various periods of the reaction was obtained
by treating the used catalysts with CH2Cl2 and analyzing the
extracted carbonaceous deposits by GC–MS (Figure 3).[13] In
the early stages of the reaction (20 min on stream), the
carbonaceous deposits contained mainly ethers, esters, and
acetals with a chain length of 3–7 carbon atoms and
substituted cyclopentenones and cyclohexenones containing
7–9 carbon atoms. At longer reaction times (50 min on
stream), the ester and ketone content decreased drastically,
whereas ethers were still present. Finally, at the steady state
hyde 7:1, weight hourly space velocity (WHSV): 3.8 ggÀ1 hÀ1
,
T=573 K). 3-MB-1=3-methyl-1-butene, 2-MB-1=2-methyl-1-butene,
2-MB-2=2-methyl-2-butene.
The main by-products observed during the steady-state
period were 2,2,4-trimethylpentenes, methylcyclopentadiene,
4,4-dimethyl-1,3-dioxane, 2-methylbutanal, 4-methyldehy-
dropyranes, and 3,4,4-trimethylcyclohexene. 2,2,4-Trimethyl-
pentenes are formed by the oligomerization of isobutene,
whereas all other compounds are products of formaldehyde
and isobutene condensation
followed by isomerization,
cyclization, and dehydration
reactions. The main reaction
pathways leading to these
products have been described
elsewhere.[8b]
To elucidate the reasons
for the induction period in
isoprene formation, we inves-
tigated the accumulation of
carbonaceous deposits over
the niobium oxide catalyst.
For this purpose, the catalytic
experiments were interrupted
after various periods of time,
and the coke accumulated on
the catalyst was analyzed by
the thermogravimetric (TG)
technique. We found that the
carbonaceous deposits were
mostly formed during the
induction period (Figure 2).
This result indicates the key
role of coke in the selective
synthesis of isoprene. The
nature of coke also changed
during the induction period:
Figure 3. GC–MS analysis of coke deposits accumulated on Nb2O5 after 20, 50, and 300 min on stream.
The C8–C16 compounds indicated in the top spectrum are olefin and polyene species. The signal marked
with an asterisk is due to a contaminant in the solvent.
12962 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12961 –12964

Angewandte
Chemie
(300 min on stream), the carbonaceous deposits were mostly
long-chain alkenes containing 8–16 carbon atoms.
saturated and unsaturated carbon atoms of branched olefins
and polyenic species.[15,18] Thus, the initially formed surface
alkoxides and ethers are decomposed into alkenes, which
undergo further reactions that lead to long-chain oligomeric
carbonaceous deposits.
The 1H–13C cross-polarization magic-angle spinning
(CPMAS) NMR spectrum obtained for the catalyst after
300 min on stream confirms the GC–MS results on the nature
of the carbonaceous deposits (Figure 4a). The peaks at d =
30.2 and 20.3 ppm can be attributed to saturated carbon
The decomposition of surface alkoxides at the early stages
of the reaction may account for the formation of 3-methyl-1-
butene (Scheme 1), which was observed as the only primary
Scheme 1. Mechanism of early stages of surface reactions.
product during the induction period in the experiments in
a continuous flow reactor. 3-Methyl-1-butene can undergo
isomerization to 2-methyl-1-butene and 2-methyl-2-butene
(Scheme 1) and oligomerize to give long-chain olefins, as
confirmed by the analysis of reaction products (Figure 1) and
carbonaceous deposits at longer reaction times (Figures 3 and
4).
Finally, heating to 573 K resulted in complete trans-
formation of the reactants into olefinic coke, which was very
similar to that observed in the experiments in the continuous
flow reactor (Figure 4a,b). This coke is composed of C8–C16
branched unsaturated hydrocarbons, which are most probably
protonated at the acidic sites of niobium oxide and can initiate
the reaction of formaldehyde and isobutene to enable
selective isoprene formation.
Figure 4. 1H–13C CPMAS NMR spectra obtained for the Nb2O5 catalyst
a) after 300 min on stream in a continuous flow reactor and b) during
the course of in situ experiments at various reaction temperatures.
Signals marked with asterisks are due to side bands.
atoms of branched aliphatic compounds, whereas the intense
broad peak at d = 128.9 ppm is due to olefinic carbon
atoms.[14,15] Thus, all the results obtained suggest that on the
fresh catalyst surface, oxygen-containing carbonaceous
deposits are formed, whereas at longer reaction times these
compounds are converted into branched polyenic coke, which
is most probably responsible for selective isoprene formation
at longer times on stream.
The results obtained can be further rationalized by the
reaction pathway in Scheme 2. In the first step, the interaction
of formaldehyde with the carbenium ion of active coke leads
To gain deeper insight into the mechanism of active-coke
formation, we studied the initial stages of the reaction by in
situ 1H–13C CPMAS NMR spectroscopy (Figure 4b).
[13C]H2O was used as the labeled reactant. The first major
peaks observed at d ꢀ 77.4 and 68.1 ppm after the coadsorp-
tion of [13C]H2O and unlabeled isobutene can be assigned to
surface alkoxides and ethers.[15] The peaks observed are
different from those detected previously during the interac-
tion of formaldehyde[16] or isobutene[17] individually with
acidic catalysts, which suggests that they arise from the
interaction of both reactants with the catalyst surface. The
observation of surface alkoxides and ethers during the initial
steps of the in situ experiments is in line with the results of the
GC–MS analysis of coke deposits formed during the induc-
tion period in a continuous flow reactor (Figure 3).
Scheme 2. Proposed reaction pathway for the formation of isoprene.
to carbocation I, which further interacts with isobutene to
form carbocation II. Further decomposition of carbocation II
yields isoprene and water and results in restoration of the
active site.
In conclusion, the key role of unsaturated branched
carbonaceous deposits in the synthesis of isoprene over solid
catalysts has been demonstrated. The mechanistic pathway
involves the accumulation of unsaturated branched surface
species during the induction period, followed by their
interaction with formaldehyde and isobutene through an
addition–cleavage mechanism to promote selective isoprene
formation. The mechanism proposed shows the way forward
for further catalyst improvement. In particular, the analysis of
An increase in the reaction temperature led to the
disappearance of the signals at d ꢀ 77.4 ppm and then at d
ꢀ 68.1 ppm and the appearance of peaks in the spectral
regions of d = 20–30 and 120–140 ppm corresponding to
Angew. Chem. Int. Ed. 2013, 52, 12961 –12964
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12963

.
Angewandte
Communications
[5] a) E. Dumitriu, V. Hulea, I. Fechete, A. Auroux, J.-F. Lacaze, C.
Guimon, Appl. Catal. A 1999, 181, 15 – 28; b) E. Dumitriu, D.
Trong On, S. Kaliaguine, J. Catal. 1997, 170, 150 – 160.
[6] a) A. Krzywicki, T. Wilanowicz, S. Malinowski, React. Kinet.
Catal. Lett. 1979, 11, 399 – 403; b) M. Ai, J. Catal. 1987, 106, 280 –
286; c) Y. Too, T. Nakayama, JP62 – 174032, 1986.
[7] Z. Dang, J. Gu, L. Yu, C. Zhang, React. Kinet. Catal. Lett. 1991,
43, 495 – 500.
[8] a) V. V. Ordomsky, V. L. Sushkevich, I. I. Ivanova, RU 2446138,
2012; V. V. Ordomsky, V. L. Sushkevich, I. I. Ivanova, RU
2421441, 2011; b) V. L. Sushkevich, V. V. Ordomsky, I. I. Iva-
nova, Appl. Catal. A 2012, 441 – 442, 21 – 29.
[9] V. V. Ordomsky, V. L. Sushkevich, I. I. Ivanova, RU 2412148,
2011.
[10] T. Ushikubo, T. Iizuka, H. Hattori, K. Tanabe, Catal. Today 1993,
16, 291.
[11] N. Ballarini, F. Cavani, L. Maselli, A. Montaletti, S. Passeri, D.
Scagliarini, C. Flego, C. Perego, J. Catal. 2007, 251, 423 – 436.
[12] a) F. Bauer, H. G. Karge, Mol. Sieves 2007, 5, 249 – 364; b) Z. R.
Finelli, C. A. Querini, N. S. Figoli, R. A. Comelli, Appl. Catal. A
1999, 187, 115 – 125.
[13] M. Guisnet, P. Magnoux, Appl. Catal. 1989, 54, 1 – 27.
[14] E. Breitmaier, W. Voelter, Carbon-13 NMR Spectroscopy: High-
Resolution Methods and Applications in Organic and Biochem-
istry, VCH, New York, 1990, p. 515.
coke deposits obtained on different acidic catalysts will
enable the relationship between the type of active coke
formed and the selectivity of the catalyst to be established.
This information will provide a key to the design of an
efficient solid catalyst for the synthesis of isoprene.
Received: August 12, 2013
Published online: October 15, 2013
Keywords: active coke · heterogeneous catalysis · isoprene ·
.
Prins condensation · solid-state NMR spectroscopy
[1] a) I. Dahl, S. Kolboe, Catal. Lett. 1993, 20, 329 – 336; b) W. Song,
J. Haw, J. Nicholas, C. Heneghan, J. Am. Chem. Soc. 2000, 122,
10726 – 10727; c) U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T.
Janssens, F. Joenses, S. Bordiga, K. Lillerud, Angew. Chem. 2012,
124, 5910 – 5933; Angew. Chem. Int. Ed. 2012, 51, 5810 – 5831;
d) P. Andy, N. Gnep, M. Guisnet, E. Benazzi, C. Travers, J. Catal.
1998, 173, 322 – 332; e) J. Martens, W. Souverijns, W. Verrelst, R.
Parton, F. Froment, P. Jacobs, Angew. Chem. 1995, 107, 2726 –
2728; Angew. Chem. Int. Ed. Engl. 1995, 34, 2528 – 2529; f) Z.
Da, P. Magnoux, M. Guisnet, Appl. Catal. A 1999, 182, 407 – 411;
g) I. Ferino, R. Monaci, E. Rombi, V. Solinas, P. Magnoux, M.
Guisnet, Appl. Catal. A 1999, 183, 303 – 309.
[15] I. I. Ivanova, Yu. G. Kolyagin, Chem. Soc. Rev. 2010, 39, 5018 –
5050.
[2] A. P. Rudenko, “On the Role of Coke Deposits in Organic
Catalysis”, Actual Problems of Physical Chemistry, Moscow
1968.
[3] M. Guisnet, J. Mol. Catal. A 2002, 182 – 183, 267 – 382.
[4] a) Y. Hamamoto, A. Mitsutani, US3221075, 1965, b) A. Mitsu-
tani, US3284533, 1966; c) K. Halcour, P. Losaker, W. Schwerdtel,
W. Swodenk, DE 2347841, 1975; d) K. Halcour, P. Losaker, W.
Schwerdtel, W. Swodenk, DE2044623, 1970.
[16] M. Sefcik, J. Am. Chem. Soc. 1979, 101, 2164 – 2170.
[17] a) N. Lazo, B. Richardson, P. Schettler, J. White, E. Munson, J.
Haw, J. Phys. Chem. 1991, 95, 9420 – 9425; b) T. Horvath, M.
Seiler, M. Hunger, Appl. Catal. A 2000, 193, 227 – 236.
[18] I. Ivanova, Yu. Kolyagin, V. Ordomsky, E. Asachenko, E.
Pasynkova, Yu. Pirogov, J. Mol. Catal. A 2009, 305, 47 – 53.
12964 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12961 –12964