Vapour-phase synthesis of 2-methyl- and 4-methylquinoline over BEA* zeolites

Article abstract of DOI:10.1016/j.jcat.2006.02.015

4-Methylquinoline and 2-methylquinoline were synthesized from acetaldehyde and aniline in the gas phase over BEA* zeolite catalysts. High combined yields of 2- and 4-methyl-substituted quinolines were obtained with H-BEA* zeolite and with BEA*-F synthesized in fluoride medium, with 4-methylquinoline being the predominant isomer. Postsynthesis fluorination of the H-BEA* with ammonium fluoride leads to dealumination and formation of extra-framework aluminium fluoride compounds. Product selectivities changed with time over this catalyst, such that 2-methylquinoline became the predominant product. New insight into the reaction mechanism is offered, and previous propositions can be rationalized based on these new results.

Full text of DOI:10.1016/j.jcat.2006.02.015

Journal of Catalysis 239 (2006) 362–368
www.elsevier.com/locate/jcat
Vapour-phase synthesis of 2-methyl- and 4-methylquinoline
over BEA* zeolites
Roald Brosius a, David Gammon b, Frederik Van Laar a, Eric van Steen c,
Bert Sels a,∗, Pierre Jacobs a,∗
a
Centre for Surface Chemistry and Catalysis, Katholieke Universiteit Leuven, Kasteelpark Arenberg 23, 3001 Heverlee, Belgium
b
Department of Chemistry, University of Cape Town, Rondebosch, 7701, South Africa
c
Catalysis Research Unit, Department of Chemical Engineering, University of Cape Town, Rondebosch, 7701, South Africa
Received 24 October 2005; revised 19 January 2006; accepted 22 February 2006
Available online 20 March 2006
Abstract
4-Methylquinoline and 2-methylquinoline were synthesized from acetaldehyde and aniline in the gas phase over BEA* zeolite catalysts. High
combined yields of 2- and 4-methyl-substituted quinolines were obtained with H-BEA* zeolite and with BEA*-F synthesized in fluoride medium,
with 4-methylquinoline being the predominant isomer. Postsynthesis fluorination of the H-BEA* with ammonium fluoride leads to dealumi-
nation and formation of extra-framework aluminium fluoride compounds. Product selectivities changed with time over this catalyst, such that
2-methylquinoline became the predominant product. New insight into the reaction mechanism is offered, and previous propositions can be ratio-
nalized based on these new results.
 2006 Elsevier Inc. All rights reserved.
Keywords: Vapour-phase synthesis; 2-methylquinoline; 4-methylquinoline; BEA* zeolite; Aniline; Acetaldehyde; Fluorination
1. Introduction
diallylamines with Co2(CO)8 as a catalyst in a heteroannula-
tion reaction also yield 2,3-substituted quinolines [10]. Good
yields have also been obtained with a ruthenium–tin complex
catalyzing an amine exchange reaction between anilines and
trialkylamines, followed by Schiff-base dimerisation and het-
eroannulation. With m-toluidine, the 7-substituted isomer is
favoured [11,12]. Ln(OTf)3 and Sc(OTf)3 have been used to
promote three-component coupling reactions between aldehy-
des, amines, and dienes or alkenes. These Lewis acid catalysts
were active in imine formation and successive imino Diels–
Alder-type reactions. High yields of poly-substituted tetrahy-
droquinolines were reported [13,14].
The quest for green alternatives for chemical reactions is
motivated by our growing awareness of their potential im-
pact on the environment. Reactions in the vapor phase with
heterogeneous catalysts offer several advantages over liquid-
phase processes, including easy catalyst recovery, continuous
processing, and absence of acid waste streams. A number
of attempts to synthesize quinolines in the vapor phase have
been successful [15–20]. The industrial relevance of gas-phase
quinoline synthesis was recently demonstrated when McAteer
Nitrogen-containing heterocyclic compounds, such as quino-
line and alkylquinolines, are important intermediates in the
production of pharmaceuticals, herbicides, fungicides, dyes,
etc. [1]. Until now, quinoline has been recovered from coal tar
with sulfuric acid, followed by precipitation with ammonia [1].
Several methods of synthesising quinolines using homogeneous
acid catalysts exist, but they produce toxic waste streams [2].
These methods include conventional routes, including Skraup,
Döbner–von Miller, Gould–Jacobs, Knorr, Beyer, Friedländer,
and Pfitzinger syntheses [3–5]. In recent years several new ap-
proaches have been developed for synthesis in the liquid phase
using transition metal-based Lewis acid catalysts. Ruthenium,
rhodium, palladium, and iron complexes catalyze the formation
of 2,3-substituted quinolines from nitrobenzene and aldehy-
des or alcohols in the presence of CO [6–9]. Conversion of
*
Corresponding authors.
E-mail addresses: bert.sels@agr.kuleuven.ac.be (B. Sels),
pierre.jacobs@agr.kuleuven.ac.be (P. Jacobs).
0021-9517/$ – see front matter  2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.02.015

R. Brosius et al. / Journal of Catalysis 239 (2006) 362–368
363
et al. were awarded a world patent [15]. They reported for-
mation of 8-methylquinoline in good yield from o-toluidine
and a mixture of formaldehyde, acetaldehyde, and water at
450 ◦C over amorphous aluminosilicate and zeolite catalysts.
Zeolites with BEA* topology performed slightly better than ze-
olites with MOR, FER, or MFI topologies. The main quinoline
byproducts are, in order of importance, 3,8-dimethylquinoline,
6,8-dimethylquinoline, and 2,8-dimethylquinoline. Aniline re-
acted with crotonaldehyde over fluor-containing aluminosili-
cate catalyst at 450 ◦C to yield predominantly 4-methylquin-
oline, with 2-methylquinoline as only a minor impurity [20].
Campanati et al. found that the reaction of 2-ethylaniline with
ethyleneglycol over an acid-treated commercial K10 mont-
morillonite catalyst yielded mainly 2-methyl-8-ethylquinoline
[18,19]. A reaction pathway was proposed in which ethyl-
ene glycol is converted to crotonaldehyde through dehydration
and condensation and subsequently undergoes conjugate or
Michael-type addition, followed by electrophilic aromatic sub-
stitution. It is surprising that in this instance the 2-substituted
isomer is formed, in contrast to the 4-substituted isomer
formed as major product by other methods. Michael addition
of 2-ethylaniline to crotonaldehyde is apparently favoured over
an initial imine-forming reaction of aniline with crotonalde-
hyde. The regioselectivity of the reaction of the aniline base
with the α,β-unsaturated aldehyde may be influenced by the
reaction temperature, acidity of the catalyst, alkylation of the
aniline base, and choice of aldehyde reagents that condense to
form the α,β-unsaturated aldehyde. Campanati et al. suggested
that an alternative to the imine route, involving regiospecific
ortho-alkylation of the aniline base with the β-carbon of the
α,β-unsaturated aldehyde, leads to the 4-substituted quinoline.
They argued that a substituted aniline, such as 2-ethylaniline,
would experience steric hindrance at the amine function and
preferentially form the 2-substituted isomer via Michael addi-
tion [18]. But it is not clear why the 2-ethylaniline would not
undergo Friedel–Crafts acylation with α,β-unsaturated alde-
hyde, which would give the 2-substituted isomer. The feasibility
of regiospecific ortho substitution of a secondary aniline with-
out a substituent in the para position has been demonstrated
using boron trichloride as a catalyst and benzene as a sol-
vent [33]. In summary, to date no coherent explanation has been
given for the regioselectivity of the reaction between aniline
bases and aldehyde reagents or their condensation products.
This paper reports the high activity and selectivity of BEA*
zeolite catalysts for the synthesis of methyl-substituted quino-
lines from aniline and acetaldehyde. The influence of fluorina-
tion of the zeolites on the activity and selectivity is addressed.
New insight into the reaction mechanism is gained from these
new results, as well as from recent literature on the synthesis
of heterocyclic nitrogen-containing compounds [34]. Previous
propositions can be rationalized based on our findings.
solution of NH4F by the incipient wetness method and dried
overnight at 60 ◦C. (Another way of introducing F− into the
catalyst is to use HF as the mineralizing agent in the hydrother-
mal synthesis of a zeolite [21].) A BEA*-F zeolite was crys-
tallized from a gel with the following relative molar amounts:
TEAOH:14 TEOS:25 Al:1 H2O2:8.6 H2O:189 HF:14, adapting
a recipe for a Ti-Beta zeolite synthesized in fluoride medium
by replacing Ti with Al [22]. The BEA*-F zeolite crystals
were calcined at 500 ◦C, and the crystallinity was confirmed
by X-ray diffraction. Catalysts were activated at temperatures
above 450 ◦C in dry air. Before catalytic testing, the temper-
ature was lowered and the reactor was flushed with nitrogen.
In the activation procedure for NH4F-treated H-BEA* (NH4F-
BEA*), NH4F was allowed to decompose at 250 ◦C in dry air,
releasing ammonia, after which the activation temperature was
set.
The reaction gas mixture was composed of water, acetalde-
hyde, and aniline in nitrogen carrier in molar ratios of 4:4:1.
The gas feed with water and acetaldehyde and the gas feed with
aniline were led seperately to the reactor vessel so that mix-
ing of the reagents set in shortly above the catalyst bed. The
gas feed was passed over a 5-mL catalytic bed at a GHSV of
900 h−1. Aniline was administered at a rate of 0.2 mL h−1
.
Reaction products were condensed at −78 ◦C and analyzed by
gas chromatography. A blank reaction over inert SiC at 450 ◦C
produced only trace amounts of acetaldehyde anil (AA). Mass
spectrometry analysis confirmed structure assignments. Solid-
state 29Si and 27Al MAS nuclear magnetic resonance (NMR)
spectra of selected samples were recorded on a Bruker AMX-
300 (7 T) spectrometer operating at 79.5 and 104.26 MHz,
respectively. Chemical shifts were referred to tetramethylsilane
and Al(NO3)3·6H2O, respectively.
3. Results
Table 1 provides data on aniline conversion; product se-
lectivities for 2-methylquinoline (2MeQ), 4-methylquinoline
(4MeQ), quinoline, and N-ethylaniline (EA); and combined
yields of quinolines. With BEA* catalyst activated at 500 ◦C,
a yield of 83% was obtained at 450 ◦C after 4.5 h on stream
(HOS). Table 2 lists the best yields reported in literature for
several quinoline synthesis reactions in the gas phase. The gas-
phase reaction of acetaldehyde with aniline gave higher yields
than previously reported gas-phase quinoline synthesis reac-
tions. Remarkably, both 4MeQ and 2MeQ were observed, with
significant selectivity toward 4MeQ, and some quinoline was
also produced. To the best of our knowledge, this is the first
report of both 2- and 4-substituted quinolines as major reac-
tion products. Interestingly, after some time on stream, 4MeQ
was the predominant isomer. These observations rightly raise
questions about the mechanism of formation of these positional
isomers.
2. Experimental
Yields were lower with the NH4F-BEA* catalyst owing to
decreased selectivity for quinolines. At 97% aniline conver-
sion, a 71% combined yield of quinolines was obtained. The
following observations arise from data shown in Table 1. Cata-
lyst activated at 450 ◦C displayed high and stable conversions,
The fluorination of the BEA* zeolite was carried out as fol-
lows. A 10-g sample of a commercial zeolite (PQ) powder with
Si/Al: 21.6 was impregnated with ca. 17 mL of a 2.0 M aqueous

364
R. Brosius et al. / Journal of Catalysis 239 (2006) 362–368
Table 1
Aniline conversion, selectivities for Q, 2MeQ, 4MeQ and EA, yield of quino-
lines in selected experiments
a
b
c
Catalyst,
T
HOS C (%) S (%)
(h)
Yield
(%)
reaction
◦
T
( C)
aniline
pretreatment
2MeQ 4MeQ
Q
EA
H-BEA*,
500 C
450
2
3
4
4.5
0.5
1
2
3
4
4.5
1
2.5
3
3.5
4
4.5
5
5.5
76
90
94
90
66
88
93
96
97
97
62
74
78
82
81
77
72
65
24
18
20
29
23
25
28
34
37
43
27
32
39
43
43
43
46
47
27
36
45
47
48
47
42
37
32
29
55
46
41
38
33
28
27
26
10
13
12
16
5
3
2
2
2
1
6
2
2
2
1
1
1
/
/
/
46
60
72
83
50
66
67
70
69
71
55
59
64
68
◦
/
NH F-
4
428
450
4
4
4
4
5
6
1
4
6
7
BEA*,
◦
450 C
NH F-
4
Fig. 1. Selectivity for 4MeQ (a), 2MeQ (b), EA (c) and Q (d) in the reaction
BEA*,
◦
◦
of aniline with acetaldehyde over NH F-BEA* zeolite at 450 C. The catalyst
was activated at 550 C.
550 C
4
◦
11 62
12 55
13 53
14 47
0
a
C: conversion of aniline.
S: product selectivity; 2MeQ: 2-methylquinoline, 4MeQ: 4-methylquin-
b
oline, Q: quinoline, EA: N-ethylaniline.
c
Sum of yields of 2MeQ, 4MeQ and Q.
whereas catalyst activated at 550 ◦C showed only temporarily
high activity. At the onset of the reaction, the selectivity for
4MeQ was twice that for 2MeQ regardless of catalyst pretreat-
ment. After 4.5 HOS, 2MeQ became the predominant isomer.
Note that the selectivity for EA also increased with time. The
evolution of selectivity with time at 450 ◦C on NH4F-BEA*
catalyst activated at 550 ◦C is shown in Fig. 1. Because the
H-BEA* catalyst does not display this peculiar behaviour, the
evolution of product selectivities is probably related to grad-
ually changing catalyst properties provoked by fluorination.
Consequently, this catalyst was submitted to a more thorough
investigation.
Solid-state 29Si and 27Al MAS NMR spectra were recorded.
Fig. 2 shows 27Al MAS NMR spectra of the BEA* zeolite af-
ter treatment with NH4F (1), the NH4F-BEA* catalyst after
calcination at 550 ◦C and hydration (2), and the NH4F-BEA*
catalyst after activation at 550 ◦C and subsequent reaction at
27
Fig. 2. Al MAS NMR spectra of NH F-BEA* zeolite after impregnation
4
◦
◦
with NH F and drying at 60 C (1), after subsequent calcination at 550 C and
4
◦
hydration (2) and after NH F treatment followed by activation at 550 C and
subsequent reaction at 450 C (3).
4
◦
Table 2
Overview of gas phase quinoline syntheses reported in literature
Catalyst
SiO /Al O /F
(84/15/1) wt%
Acid-treated
montmorillonite
Amorphous
Reagents (rel. molar amounts)
Aniline:crotonaldehyde (1.5:1)
T
Product
4MeQ
GHSV
Yield
Ref.
[20]
reaction
◦
−1a
b
450 C
200 h
3000 h
1000 h
540 h
11%
2
2 3
◦
−1
−1
−1
c
2-Ethylaniline:ethyleneglycol (1:2)
330 C
2MeQ
Q
41%
[18]
[15]
[17]
◦
b
Ortho-toluidine:acetaldehyde + formaldehyde:H O (1:4:4)
470 C
54%
2
SiO /Al O
2 3
2
d
◦
b
CuO–ZnO/Al O
2
Aniline:glycerol (1:2)
425 C
Q
65%
3
−1
a
−1
given by Uebel et al. [20].
Estimated based on the value of 0.09 mol
L
h
aldehyde
catalyst
b
c
d
Calculation based on aniline base conversion.
Calculation based on ethyleneglycol conversion.
Air is used as carrier gas instead of nitrogen.

R. Brosius et al. / Journal of Catalysis 239 (2006) 362–368
365
bined selectivities for 2MeQ, 4MeQ, and quinoline reach-
ing 90% was obtained after 4.5 HOS at 450 ◦C. Activation
at 475 ◦C resulted in a considerable delay in catalyst activ-
ity and decreased selectivity. With this catalyst, the ratio of
2MeQ/4MeQ was consistently <1. It is intriguing to note that
in contrast to the NH4F-treated BEA* zeolite, the selectivity
for 4MeQ was stable on BEA*-F synthesized in the presence of
fluoride. Zeolites were very poor in defect sites when the miner-
alizing species were fluoride ions, whereas high-silica zeolites
made in basic medium with hydroxyl ions displayed a substan-
tial number of framework defects. It is known that SiO4/2F−
units balance the positive charge of organic structure-directing
agents, whereas in zeolites synthesized in basic medium, hy-
droxyl groups are structurally integrated to balance the pos-
itive charge [21]. Calcination removes the organic molecules
and fluoride anions occluded in the solids [23,24]. In contrast,
postsynthesis treatment with NH4F may give rise to the for-
mation of Si–F and/or Al–F linkages during activation of the
zeolite [26,27].
29
Fig. 3. Si MAS NMR spectra of NH F-BEA* zeolite after treatment with
4
◦
NH F (1), after subsequent calcination at 550 C and hydration (2) and after
NH F treatment followed by activation at 550 C and subsequent reaction at
450 C (3).
4
◦
4
◦
Table 3
Aniline conversion, selectivities for Q, 2MeQ, 4MeQ and EA, yield of quino-
lines in selected experiments
4. Discussion
a
b
c
Catalyst,
T
HOS C (%) S (%)
(h)
Yield
(%)
reaction
4.1. The effect of fluorination
◦
T
( C)
aniline
pretreatment
2MeQ 4MeQ
Q
EA
BEA*-F,
550 C
450
1
94
98
98
18
30
66
88
9
16
21
4
10
15
18
44
43
56
18
37
45
37
34
18
15
19
33
14
8
/
/
/
/
/
/
/
82
75
90
7
24
49
The NMR spectra of the NH4F-BEA* catalyst show that ex-
posure to the reaction gas mixture at 450 ◦C after activation
at 550 ◦C led to the formation of extra-framework fluorinated
aluminium compounds as well as tetrahedral Al3+ species pre-
sumably containing Al–F bonds, as evidenced by the decreased
chemical shift of both the tetrahedral and the octahedral Al3+
species due to the increased proximity of the F atom. The effects
of fluorination have received ample attention in the literature.
On reaction of molecular fluorine with H-ZSM-5 at room tem-
perature, nonframework fluorinated species weakly attached
to the lattice are formed [25]. In our case, on NH4F impreg-
nation, octahedral Al3+ species were formed. On H-ZSM-5,
NH4F impregnation followed by calcination at temperatures
above 500 ◦C leads to the formation of Si–F bonds and to dealu-
mination and formation of extra-framework aluminum species
[26,28]. There is a general concensus that fluorination pro-
vokes dealumination of the zeolite framework regardless of the
fluorination method used [29]. Furthermore, fluorination may
lead to decreased pore volume and specific surface area [30].
H-BEA* zeolite modified by aqueous solutions of HF exhibits
an increased ratio of Lewis acidity to Brønsted acidity and of
strong sites to weak sites but with diminished total acidity [30].
On alumina, the number of Lewis acid sites increases signif-
icantly with fluorine addition [31]. Depending on the degree
of hydration and the fluorine content, various species, such as
aluminium hydroxyfluorides or hydrated aluminium trifluoride,
can be found on the catalyst surface.
◦
2.5
4.5
1
1.5
3
BEA*-F,
450
◦
d
475 C
d
d
4
55
a
C: conversion of aniline.
S: product selectivity; 2MeQ: 2-methylquinoline, 4MeQ: 4-methylquin-
b
oline, Q: quinoline, EA: N-ethylaniline.
c
Sum of yields of 2MeQ, 4MeQ and Q.
Other products included 2ProQ (2-propylquinoline) up to 8%, fully or
d
partially hydrogenated products of the condensation of aniline with C2, C4,
C6, etc., unsaturated aldehydes, α-anilino-butanaldehyde anil and β-anilino-
butanaldehyde anil and traces of unidentified products.
450 ◦C (3). Comparison of 29Si MAS NMR spectra shown in
Fig. 3 reveals that the signal attributed to the Q4(1Al) species at
approximately −103 ppm was lower for those samples exposed
to calcination or to activation at 550 ◦C followed by reaction.
Therefore, it can be concluded that dealumination occurred af-
ter pretreatment at high temperature. On NH4F treatment of the
BEA* zeolite, a very intense line due to sixfold-coordinated
Al3+ species appeared at approximately 0 ppm. The spectrum
of the sample calcined at 550 ◦C shows a peak due to tetrahe-
dral Al3+ species on top of a very broad and flat band between
60 and −60 ppm, as demonstrated by the rise in the baseline.
The line due to an AlVI resonance almost completely disap-
peared. After subsequent prolonged exposure to the reaction
gas at 450 ◦C, broad bands due to AlIV and AlVI resonances
appeared with decreased chemical shift in the 27Al MAS NMR
spectrum.
There are a number of plausible explanations for the evolu-
tion of the selectivities for 4MeQ and 2MeQ over NH4F-BEA*.
The first of these assumes that the observations are related to
the catalyst acidity. Over a fluor-containing amorphous alumi-
nosilicate catalyst, the gas-phase reaction of aniline with cro-
tonaldehyde resulted in 4MeQ at 450 ◦C [20]. The presence of
Table 3 gives aniline conversion, product selectivities, and
combined quinoline yields on BEA*-F catalyst. With BEA*-F
activated at 550 ◦C, an aniline conversion of 98% with com-

366
R. Brosius et al. / Journal of Catalysis 239 (2006) 362–368
water was not mentioned, and we thus assume that Lewis acid
sites prevailed on this catalyst. It may then be suggested that
Lewis acidity and the formation of 4MeQ are somehow related.
Given the considerable amount of water in our experiments, it
seems unlikely that a significant amount of coordinatively un-
saturated Al3+ species exhibiting Lewis acidity remained. The
acidity associated with extra-framework fluorinated aluminium
compounds in the presence of water was probably of the Brøn-
sted type. Perhaps the decreasing selectivity for 4MeQ in the
experiment with NH4F-BEA* catalyst could be due to deple-
tion of the appropriate Lewis acid sites as a consequence of
catalyst hydration and dealumination.
Another explanation assumes an important role for the struc-
tural parameters of the zeolite catalyst rather than the acidity.
Shape selectivity may favour one isomer over the other. Gradual
dealumination of the NH4F-BEA* catalyst may lead to partial
pore blocking or to micropore size restrictions due to extra-
lattice aluminium compounds. The less bulky isomer 2MeQ
then would be favoured over 4MeQ. This can be observed after
4.5 HOS over the NH4F-BEA* catalyst. BEA*-F zeolite cata-
lyst, on the other hand, presumably has very few lattice defects
and displays high and stable selectivity for 4MeQ formation.
In this case the zeolite would not impose restrictions on the
synthesis of the more bulky isomer 4MeQ. We prefer this sec-
ond explanation, which is supported by the fact that H-BEA*
and BEA*-F catalysts displayed similar product selectivities
(compare Table 1, line 4 and Table 3, line 3). In conclusion,
we believe that shape selectivity caused by pore size restriction
determines the distribution of products over the NH4F-BEA*
catalyst.
4.2. The reaction mechanism
Fig. 4 provides a tentative overview of plausible reaction
pathways leading to 2MeQ and 4MeQ. In pathways 1 and 2, an
imine (Schiff base) was formed from aniline and acetaldehyde,
whereas in pathways 3 and 4, aniline reacted with crotonalde-
hyde, the product of aldol condensation of two acetaldehyde
units. In pathway 3, an imine was formed, whereas in path-
way 4, conjugate or Michael-type addition of aniline to croton-
aldehyde occurred. In this section we present evidence in favour
of and against these propositions.
At first sight, observations reported in the literature seem
to suggest that a higher reaction temperature gives preference
to 4MeQ formation [20], whereas lower temperatures favour
2MeQ [18]. This seems to be in line with the heats of forma-
tion of 4MeQ (262.3 kJ mol−1) and 2MeQ (93.6 kJ mol−1) [18].
In several of our experiments, the temperature was lowered to
350 ◦C. Besides low aniline conversions, increased selectivity
for 2MeQ compared with 4MeQ was indeed observed. At the
same time, the byproduct EA was also found in large quan-
tities at 350 ◦C. This suggestion finds confirmation in Fig. 1,
which shows that at 450 ◦C over NH4F-BEA*, the selectivity
for EA increased as 4MeQ decreased with HOS. Therefore, it
can be suggested that AA, giving EA after hydrogen transfer, is
a reaction intermediate rather than a mere byproduct. This casts
doubt on the assumption that the α,β-unsaturated aldehyde is
a reaction intermediate in the formation of 4MeQ. Forrest et
al. demonstrated the viability of a mechanism for the liquid-
phase synthesis of 2MeQ from aniline and acetaldehyde that
does not involve formation of the α,β-unsaturated aldehyde
Fig. 4. Schematic overview of reaction pathways leading to formation of 2MeQ or 4MeQ.

R. Brosius et al. / Journal of Catalysis 239 (2006) 362–368
367
Fig. 5. Schematic representation of the ortho-alkylation route to formation of 4MeQ.
first (Fig. 4, pathway 2) [32]. The gas-phase aniline–acrolein
reaction gave poor yields [15]. In addition, McAteer et al. re-
ported that the aniline–crotonaldehyde reaction yielded an ex-
cess of 4MeQ over 2MeQ at 450 ◦C. In a recent study of the re-
action of o-toluidine (oTOL) with 13C-enriched formaldehyde
and acetaldehyde, the 13C enhancement in individual positions
of the heterocyclic ring of resulting quinoline was analysed by
1H and 13C NMR spectroscopy, and it was found that 13C was
incorporated almost exclusively in the C-2 position, not in the
C-4 position, of the heterocyclic ring. This finding suggests
that a mechanism involving condensation of formaldehyde with
oTOL to form the Schiff base methanimine intermediate that re-
acted further with acetaldehyde was favoured over Michael ad-
dition of oTOL to acrolein (arising from aldol condensation of
acetaldehyde and formaldehyde), because the former resulted in
incorporation of 13C at the C-2 position. Therefore, Schiff base
condensation of acrolein with oTOL is not in keeping with the
observed 13C incorporation measurements [34]. These results
are consistent with our observation that the imine of aniline
and acetaldehyde, AA, is possibly an intermediate in the for-
mation of 4MeQ. Therefore, based on the correlation of the
selectivity for 4MeQ and EA in the reaction over NH4F-BEA*
catalyst and on the above evidence, we propose that 4MeQ is
formed according to pathway 1 rather than pathway 3 in Fig. 4.
AA undergoes a hydride shift to form the corresponding enam-
ine, which subsequently reacts with acetaldehyde followed by
intramolecular electrophilic aromatic substitution, resulting in
ring closure. Thus, 4MeQ is synthesized in a stepwise manner.
The alternative route involving the Schiff base dimerisation
reaction as shown in pathway 2 proceeds via a very bulky inter-
mediate that can in principle give rise to both 2MeQ and 4MeQ.
However, it is unlikely to be a major pathway in the confined
space of a microporous shape-selective catalyst, such as NH4F-
BEA*, after 4.5 h, where formation of 2MeQ over 4MeQ is
favoured. Thus, formation of 2MeQ seems to arise from the
Michael addition of aniline to the α,β-unsaturated aldehyde
crotonaldehyde, followed by a ring-closing electrophilic aro-
matic substitution (pathway 4).
If the ortho-alkylation route, touched on in Section 1, is to
be taken at all seriously, one might envisage that 4-methyl-
quinoline is formed (as shown in Fig. 5) through ortho-alkyl-
ation of aniline or its derived imine (I) with acetaldehyde to
give the hydroxyethylated derivatives (B and G), followed by
sequences of dehydrations, H-shifts, and intramolecular acid-
catalysed reactions. There is no precedent for this mechanism in
the literature, however. This mechanism would require a fairly
strong Lewis acid site, the presence of which has not been
convincingly demonstrated in BEA* zeolite or NH4F-BEA*,
particularly not in wet conditions. Therefore, it makes sense to
not overestimate the importance of this pathway.
In conclusion, it seems that the higher temperatures favour
the imine route with stepwise formation of 4MeQ, whereas
the lower temperature route involves Michael addition to the
α,β-unsaturated aldehyde, leading to 2MeQ preferentially.
5. Conclusion
BEA* zeolite displayed high activity for the synthesis of
both 2MeQ and 4MeQ from aniline and acetaldehyde. A 90%
combined yield of 2MeQ, 4MeQ, and quinoline was ob-
tained with BEA*-F zeolite catalyst synthesized in fluoride
medium. The gas-phase reaction of aniline with acetaldehyde
produced 4MeQ, as well as considerable amounts of 2MeQ,
which had not been reported previously. Postsynthesis fluo-
rination of the BEA* zeolite provoked dealumination during
the quinoline synthesis reaction. The catalyst changes were re-
flected in a transient product distribution with opposite trends
for 4MeQ on the one hand and 2MeQ and N-ethylaniline on
the other hand. We propose that the dealumination leads to nar-
rowing of the micropores, allowing preferential formation of

368
R. Brosius et al. / Journal of Catalysis 239 (2006) 362–368
the less bulky isomer 2MeQ. High temperatures favour forma-
tion of 4MeQ in a stepwise manner starting with formation of
the imine from aniline and acetaldehyde. Lower temperatures
favour the Michael addition of aniline to crotonaldehyde, re-
sulting in 2MeQ formation.
[12] C.S. Cho, B.H. Oh, D.T. Kim, T.-J. Kim, S.C. Shim, Bull. Korean Chem.
Soc. 24 (2003) 1026.
[13] S. Kobayashi, H. Ishitani, S. Nagayama, Chem. Lett. (1995) 423.
[14] H. Ishitani, S. Kobayashi, Tetrahedron Lett. 37 (1996) 7357.
[15] C.H. McAteer, R.D. Davis, J.R. Calvin, World patent WO9703051, 1997.
[16] P.R. Reddy, K.V. Subba Rao, M. Subrahmanyam, Catal. Lett. 56 (1998)
155.
Acknowledgments
[17] B.M. Reddy, I. Ganesh, J. Mol. Catal. A 151 (2000) 289.
[18] M. Campanati, P. Savini, A. Tagliani, A. Vaccari, Catal. Lett. 47 (1997)
247.
[19] M. Campanati, A. Vaccari, O. Piccolo, Catal. Today 60 (2000) 289.
[20] H.J. Uebel, K.K. Moll, M. Mühlstädt, J. Prakt. Chem. 312 (1970) 263.
[21] H. Koller, A. Wölker, L.A. Villaescusa, M.J. Díaz-Cabañas, S. Valencia,
M.A. Camblor, J. Am. Chem. Soc. 121 (1999) 3368.
[22] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martinez,
J.A. Perdigón-Melón, S. Valencia, J. Phys. Chem. B 102 (1998) 75.
[23] M.A. Camblor, A. Corma, S. Valencia, J. Matter. Chem. 8 (1998) 2137.
[24] H. Koller, A. Wölker, H. Eckert, C. Panz, P. Behrens, Angew. Chem. Int.
Ed. Engl. 36 (1997) 2823.
[25] N.A. Sánchez, J.M. Saniger, J.-B. d’Espinose de la Caillerie, A.L. Blu-
menfeld, J.J. Fripiat, Microporous Mesoporous Mater. 50 (2001) 41.
[26] R. Le Van Mao, T.S. Le, M. Fairbairn, A. Muntasar, S. Xiao, G. Denes,
Appl. Catal. A 185 (1999) 41.
[27] R.B. Borade, A. Clearfield, J. Chem. Soc., Faraday Trans. 91 (1995)
539.
[28] S.K. Pur, R.G. Bryant, Zeolites 16 (1996) 118.
[29] A.G. Panov, V. Gruver, J.J. Fripiat, J. Catal. 168 (1997) 321.
[30] M. Han, C. Xu, J. Lin, Y. Liang, E. Roduner, Catal. Lett. 86 (2003) 81.
[31] M. Moreno, A. Rosas, J. Alcaraz, M. Hernández, S. Toppi, P. Da Costa,
Appl. Catal. A 251 (2003) 369.
[32] T.P. Forrest, G.A. Dauphinee, W.F. Miller, Can. J. Chem. 47 (1969) 2121.
[33] T. Sugasawa, T. Toyoda, M. Adachi, K. Sasakura, J. Am. Chem. Soc. 100
(1978) 4842.
This work is supported by the National Research Foundation
of South Africa and the Flemish government (Bilateral Scien-
tific Cooperation BIL 04/53). B.S. and P.A.J. thank the Belgian
government for financial support (IUAP project on supramole-
cular chemistry and catalysis).
References
[1] Ullman’s Encyclopedia of Industrial Chemistry, sixth ed., Wiley–VCH,
New York, 1999, Electronic Release.
[2] R.H.F. Manske, M. Kulka, in: R. Adams (Ed.), Organic Reactions, vol. 7,
Wiley, New York, 1953, p. 59.
[3] Z.H. Skraup, Ber. Dtsch. Chem. Ges. 13 (1880) 2086.
[4] O. Döbner, W. von Miller, Ber. Dtsch. Chem. Ges. 14 (1881) 2816.
[5] R.G. Gould, W.A. Jacobs, J. Am. Chem. Soc. 61 (1939) 2890.
[6] Y. Watanabe, N. Suzuki, Y. Ohsuji, Tetrahedron Lett. 22 (1981) 2667.
[7] Y. Watanabe, N. Suzuki, Y. Tsuji, S.C. Shim, T. Mitsudo, Bull. Chem. Soc.
Jpn. 55 (1982) 1116.
[8] Y. Watanabe, Y. Tsuji, J. Shida, Bull. Chem. Soc. Jpn. 57 (1984) 435.
[9] Y. Watanabe, K. Takatsuki, S.C. Shim, T. Mitsudo, Y. Takegami, Bull.
Chem. Soc. Jpn. 51 (1978) 3397.
[10] J. Jacob, W.D. Jones, J. Org. Chem. 68 (2003) 3563.
[11] C.S. Cho, B.H. Oh, J.S. Kim, T.-J. Kim, S.C. Shim, Chem. Commun.
(2000) 1885.
[34] J.R. Calvin, R.D. Davis, C.H. McAteer, Appl. Catal. A 285 (2005) 1.