Towards a more sustainable production of triacetoneamine with heterogeneous catalysis

Article abstract of DOI:10.1016/j.molcata.2014.06.023

The acid-catalyzed condensation of acetone and ammonia to directly produce 2,2,6,6-tetramethy-4-piperidone (triacetonamine) was studied under both homogeneous and heterogeneous catalysis. The selectivity to the desired product was affected by the presence of a complex reaction network involving several kinetically parallel and consecutive reactions, leading to by-products such as diacetonealcohol, diacetoneamine, mesityl oxide, acetonine, and 2,2,4,6-tetramethyl-2,5-dihydropyridine. The latter was the most undesired by-product, since its formation was irreversible. Key elements in achieving high selectivity in the direct synthesis of triacetonamine were the molar feed ratio between acetone and ammonia, and the amount of water in the reaction medium; in fact, water was found to play an important role in the transformation of the intermediate acetonine into triacetonamine. Compared with homogeneous catalysis, the selectivity achieved by the use of the heterogeneous H-Y zeolites was controlled by means of a proper selection of the zeolite features. In fact, the use of a highly hydrophilic H-Y zeolite made it possible to achieve the same selectivity as that obtained under homogeneous catalysis conditions, with the additional advantage of using an easily separable and reusable catalyst, which showed no deactivation.

Full text of DOI:10.1016/j.molcata.2014.06.023

Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Towards a more sustainable production of triacetoneamine with
heterogeneous catalysis
Gherardo Gliozzia,b, Lucia Frattinia, Paolo Righia, Fabrizio Cavania,b,∗
a Dipartimento di Chimica Industriale “Toso Montanari”, Università di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
b Consorzio INSTM, Unità di Ricerca di Bologna, Florence, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 6 May 2014
Received in revised form 15 June 2014
Accepted 17 June 2014
Available online 26 June 2014
The acid-catalyzed condensation of acetone and ammonia to directly produce 2,2,6,6-tetramethy-4-
piperidone (triacetonamine) was studied under both homogeneous and heterogeneous catalysis. The
selectivity to the desired product was affected by the presence of a complex reaction network involving
several kinetically parallel and consecutive reactions, leading to by-products such as diacetonealcohol,
diacetoneamine, mesityl oxide, acetonine, and 2,2,4,6-tetramethyl-2,5-dihydropyridine. The latter was
the most undesired by-product, since its formation was irreversible. Key elements in achieving high selec-
tivity in the direct synthesis of triacetonamine were the molar feed ratio between acetone and ammonia,
and the amount of water in the reaction medium; in fact, water was found to play an important role
in the transformation of the intermediate acetonine into triacetonamine. Compared with homogeneous
catalysis, the selectivity achieved by the use of the heterogeneous H-Y zeolites was controlled by means
of a proper selection of the zeolite features. In fact, the use of a highly hydrophilic H-Y zeolite made it pos-
sible to achieve the same selectivity as that obtained under homogeneous catalysis conditions, with the
additional advantage of using an easily separable and reusable catalyst, which showed no deactivation.
© 2014 Elsevier B.V. All rights reserved.
Keywords:
Triacetoneamine
Acetone
H-Y zeolite
Ammonium chloride
Zeolite hydrophilic properties
1. Introduction
6-tetrahydropyrimidine (acetonine, ACTN), carried out at room
temperature, and the reaction of the separated ACTN with acetone
2,2,6,6-Tetramethyl-4-piperidone (triacetonamine, TAA) is the
key intermediate for the synthesis of polymer light stabilizers
HALS (Hindered-Amines-Light-Stabilizers) [1,2]; their ability to
efficiently interrupt polymer degradation by radical scavenging is
based on the sterically hindered amine functional group, which
is able to form stable N-oxides as active intermediates (Denisov
cycle). TAA is also used in the synthesis of drugs, nitroxyl radi-
cals of piperidine, and pyrrolidone derivatives [2]. For instance,
by the oxidation of TAA or its derivatives, the corresponding
tetramethylpiperidine-N-oxide, called TEMPO, is formed; the lat-
ter may be used as a polymerization inhibitor, molecular weight
regulator, or oxidation catalyst.
or water to yield TAA [9,10]. Patent literature describes various cat-
alysts for these reactions, such as homogeneous Lewis and Brønsted
acids (CaCl2, ZnCl2, NH4Cl, AlF3, and BF3, used especially in the
second step of the two-step process) [10–13], sulfonated resins
(Amberlyst® and Nafion®) [14] and p-nitrotoluene [15]. In the
one-pot process, the catalysts used industrially are based on homo-
geneous systems (such as CaCl2, NH4NO3 or NH4Cl), because of
their better performance, although the use of a heterogeneous sys-
tem might offer two advantages: an easier separation of the catalyst
from the reaction medium, and fewer problems related to prod-
uct contamination. On the other hand, the heterogeneous catalyst
should be easily regenerable and show constant performances dur-
ing repeated uses or long time spans in the continuous-flow mode;
in this respect, inorganic solid acids are preferable over organic
resins.
Table 1 summarizes the yields to TAA reported in literature,
using various catalysts and procedures. It is shown that the reaction
of ACTN transformation into TAA – the second step in the two-step
process – can be carried out with a remarkable yield, by using a
Lewis or a Brønsted acid catalyst, depending on the reaction condi-
tions used. The stoichiometry for the transformation of ACTN into
TAA involves the elimination of either NH3 replaced with H2O, or
In literature, various procedures are reported for the synthesis
of TAA, with the most important being: (i) the direct condensation
of acetone and ammonia (Scheme 1) [3–8], and (ii) the two-step
process comprising the synthesis of 2,2,4,4,6-pentamethyl-1,2,5,
∗
Corresponding author at: Dipartimento di Chimica Industriale “Toso Montanari”,
Università di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy.
Tel.: +39 0512093680.
E-mail address: fabrizio.cavani@unibo.it (F. Cavani).
http://dx.doi.org/10.1016/j.molcata.2014.06.023
1381-1169/© 2014 Elsevier B.V. All rights reserved.

326
G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
Table 1
Main catalysts and TAA yield in the two-step and one-pot process.
References
Starting materials
Catalyst
TAA yield (mol%)
95; 95 (wt.%)a
[10]
[11]
[6]
[13]
[3]
[4]
[5]
[14]
[8]
[12,16]
ACTN, acetone
ACTN, acetone (+ water)
ACTN, acetone
BF3; NH4Cl
p-Toluenesulphonic acid salt of ACTN
Acetyl halide
Acid ion exchanger, in continuous fixed-bed reactor
CCl4
91
b
Acetone, NH3
25c
Acetone (+ methanol), NH3
Acetone (+ methanol), NH3
Acetone, NH3
Acetone, NH3
Acetone, NH3
26c
Organotin halides; cyanuric halides
CaCl2·2H2O; NH4Cl; CaCl2·2H2O + NH4Cl
36; 39c
20; 24; 28c
25c
Perfluorinated sulphonic acids (Nafion®
Dimethyl sulfate
)
28c
Acetone, NH3
CaY + NH4NO3
22c
a
wt and mol yield, calculated with respect to initial ACTN, are very similar, because of the very similar MW of ACTN and TAA.
Weight yields higher than 100% were claimed, probably referring to some unspecified TAA salt.
Molar yield, calculated after the multiplication of the number of TAA moles by the factor 3.
b
c
prepared by first exchanging the H-zeolite with Na+ (treatment
with a 2 M NaCl solution at 80 ◦C and pH adjusted at 10–11 with
NaOH); then after filtration and drying at 110 ◦C, Na-zeolites were
treated with 2 M NH4Cl solution overnight and dried again at
110 ◦C.
For reactivity experiments, in a typical batch test, 2.3 g of
acetone and a variable quantity of NH4OH (30 wt%, aqueous solu-
tion) were loaded into a Teflon-capped test tube equipped with
a magnetic stirrer. The mixture was homogenized at room tem-
perature for 2–3 min and then the catalyst, either an ammonium
salt (0.046 g) or a zeolite (0.23 g), was quickly added. The process
was carried out at 65 ◦C while stirring at about 500 rpm. Diffusional
limitations were excluded by means of preliminary experiments,
carried out by varying the mixing speed.
Molar yields were calculated by dividing the concentration of
a defined product by the initial molar concentration of acetone,
normalized with respect to the number of C atoms. Molar selec-
tivities were calculated by the ratio between the yield to a specific
compound and acetone conversion.
A continuous flow reactor with internal recirculation was also
used for short-lifetime experiments; 3.0 g of zeolite in the form of
extrudates (sample H-␤ 303 H/02, delivered by Süd-Chemie), were
loaded, while a make-up feed consisting of acetone and NH4OH
aqueous solution in the desired molar ratio was introduced, and
an analogous volumetric feed containing unconverted reactants
and products was purged out from the reactor. The make-up/purge
flows and the internally recirculated volumetric feed (withdrawn
from the liquid buffer with an average composition equal to that
of the purge stream) were set up in order to develop an overall
reaction contact time of around 30 min.
At the end of the reaction, the mixture was analyzed by means
of a Thermo Focus GC gas-chromatograph equipped with a FID
detector and Agilent HP-5 column, using n-decane as the internal
standard. Analysis conditions were: 50 ◦C for 2 min, 10 ◦C/min up
to 280 ◦C, and 5 min at 280 ◦C.
Scheme 1. TAA synthesis stoichiometry in the one-pot process.
2-iminopropane (dimethylketimine) replaced with acetone. There-
fore, the reaction can be carried out in any of three ways: in the
presence of at least an equimolar amount of water, in the presence
of excess acetone and without water (the preferred procedure), or
with both acetone and water [6,10,11].
The case of the one-pot reaction is different, since several
equilibria establish in the presence of acetone (or diacetonealcohol)
and ammonia, and most reactions involved are equilibrium-
limited. In fact, in most cases, the yield to TAA reported is in the
range of 20–30%. Worthy of note is the outstanding TAA selectivity,
higher than 95%, with a 22% yield: it was obtained by combining
a very large amount of NH4NO3 (used both as a catalyst and as an
ammonia source) with Ca-exchanged HY zeolite catalysts [12,16].
As shown in Scheme 1, water is not needed in the overall reac-
tion stochiometry, and in fact gaseous ammonia is typically used in
industrial processes. On the other hand, a NH4OH aqueous solution
may also be used [16], clearly increasing the process safety, while
the actual role of water on the reaction mechanism, as well as on
yield and selectivity, has not yet been investigated in detail.
Despite the industrial importance of this compound, very little
information is available in literature on the mechanism of TAA for-
mation and the role of the main reaction parameters. We decided
to investigate this reaction under both homogeneous and hetero-
geneous catalytic conditions; the final aim was that of designing
a heterogeneous catalyst for the one-pot synthesis of TAA, using a
NH4OH aqueous solution as the ammonia source, and a properly
selected and fully reusable solid acid catalyst, without the need for
homogeneous co-catalysts.
3. Results and discussion
2. Experimental
3.1. Determination of the reaction scheme in homogeneous
catalysis
The following reactants and products were used for the reaction:
Acetone (Chromasolv® Sigma–Aldrich), Ammonium Hydrox-
ide 28–30 wt% (Sigma–Aldrich), Ammonium Chloride (99.5%+
Sigma–Aldrich). The zeolites were provided by TOSOH: HSZ-
330HUA (HY zeolite SAR 6) 584 m2/g (from catalogue 550 m2/g),
HSZ-390HUA (HY zeolite SAR 200) 814 m2/g (from catalogue
750 m2/g), and HSZ-360HUA (HY zeolite SAR 15, from catalogue
550 m2/g). The H-␤ zeolite was supplied by Süd-Chemie (sample
SN308 H/00, SAR 150). The ammonium-Y zeolite (SAR 15) was
We first carried out
a
kinetic study at 65 ◦C, using an
acetone-to-ammonia molar feed ratio equal to 4.0 and a catalyst-
to-acetone weight ratio equal to 0.02, with a reaction time
ranging from 0 to 120 h. Ammonium hydroxide (30 wt% aqueous
solution) was used as
a source of ammonia, and ammo-
nium chloride as the catalyst. The reaction products, shown in
Scheme 2, were: diacetonealcohol (DAA), diacetoneamine (DAAM),
mesityl oxide (MO), acetonine (ACTN), triacetoneamine (TAA), and

G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
327
shown in Fig. 1 suggests that it forms by consecutive transformation
of TAA via water elimination, even though a direct transformation
of ACTN via ammonia elimination cannot be excluded. Indeed, the
dehydration of TAA may proceed through a series of steps which
involves (a) ring opening of TAA to a enol/imine form; (b) tau-
tomerization to the keto/enamine form; (c) ring closure to a cyclic
“aldiminol” (the equivalent of a cyclic aldol); and (d) dehydration,
which under basic conditions may proceed via an E1cB mecha-
nism (Scheme 4). The last step, as is known in aldol condensations,
may well be irreversible, thus explaining the continuous increase
of TMDP selectivity.
The overall reaction scheme, as inferred from kinetic experi-
ments, is shown in Scheme 5. It is worth noting that the reaction
network consists of several reversible reactions, with the exception
of formation of TMDP. This means that all reaction intermediates
may potentially be converted into the desired TAA while, on the
other hand, TMDP is a waste compound, since it cannot be trans-
formed back into TAA. In other words, the key feature of the process
is not the selectivity to the desired product (TAA), but the selectiv-
ity to the undesired TMDP, which should be as low as possible.
Therefore, the key element for an efficient synthesis of TAA is the
kinetic control over the consecutive and irreversible reaction of
TMDP formation.
Scheme 2. Reaction products of the acid-catalyzed condensation of acetone and
ammonia.
2,2,4,6-tetramethyl-2,5-dihydropyridine (TMDP). The results of the
reactivity experiments are shown in Fig. 1.
Since we fed an excess of acetone compared to the stoichio-
metric requirement, acetone conversion was not complete, but
reached a stable value of 45–47%; indeed, the conversion of ammo-
nia reached its maximum value already a few hours after the
mixing of reactants and, correspondingly, acetone conversion also
reached its maximum value. Thereafter, for longer reaction times,
changes involved the distribution of products only. In this regard,
the only kinetically primary products were DAAM and ACTN; in fact,
the extrapolated selectivity to nil conversion was approximately
20% for DAAM and 55% for ACTN. However, the ACTN and DAAM
selectivity rapidly declined while the conversion rose, with a con-
comitant increase in TAA selectivity. In short, after a rapid first step
where the condensation of acetone and ammonia leads to ACTN and
to DAAM, under the conditions used (T 65 ◦C), the two compounds
convert into TAA. The transformation of ACTN into TAA may occur
either with the insertion of water and elimination of ammonia, or,
in the presence of acetone, with acetone insertion and elimination
of 2-iminopropane. Scheme 3 shows an hypothetical mechanism
for interconversion between the two compounds.
3.2. Effect of the acetone–ammonia molar feed ratio
The effect of the amount of ammonia in the reaction medium
was also investigated. The reaction kinetics was studied at 65 ◦C,
with a catalyst-to-acetone weight ratio equal to 0.02. Two more
acetone-to-ammonia molar feed ratios (2.5 and 10.0) were tested.
As shown in the plots of Fig. 2, the final acetone conversion was
greatly affected by the amount of ammonia fed in the process. In
the case of the acetone–ammonia molar ratio 2.5 (Fig. 2, top), the
conversion reached 60% in a few hours, with high selectivity to
the primary product ACTN (70%). On the other hand, ACTN was
converted with low selectivity into TAA, at the same time foster-
ing TMDP formation. Opposite results were achieved with a high
acetone-to-ammonia molar feed ratio (10.0) (Fig. 2, bottom). In this
case, because of the large excess of acetone used, the conversion of
acetone was quite low (about 30%), with good selectivity to the pri-
mary product ACTN (40%), but the subsequent transformation of the
latter was much more selective to TAA, if compared to the reaction
at high ammonia concentration (lower acetone-to-ammonia molar
ratio).
Fig. 3 summarizes the above considerations, while comparing
the product selectivity resulting from different molar feed ratios at
similar conversions (results taken from Figs. 1 and 2); selectivity
to TAA achieved when the maximum conversion of acetone was
reached is also reported. This figure clearly shows an important
conclusion: the reactant feed ratio greatly affected process selec-
tivity and is an important parameter for TAA production. In fact,
it seems more convenient to carry out the process at low conver-
sion (that is, at a high acetone–ammonia molar ratio) by recycling
the reaction mixture after the separation of TAA, than to use low
feed ratios (e.g. close to the stoichiometric ratio 3), with the aim of
fostering acetone conversion. The former conditions would ensure
greater amounts of TAA in the outgoing stream with lower separa-
tion costs.
Moreover, TAA selectivity dropped at long reaction times,
whereas TMDP selectivity continued to increase throughout the
entire experiment. After 120 h reaction time, TMDP was the pre-
vailing product; therefore, it is possible to assume that reactions
leading to TMDP formation are irreversible. In other words, after
a proper reaction time, TMDP would be the unique product of the
process. In regard to the formation of TMDP, the selectivity trend
3.3. The role of water in homogeneous catalysis
According to our results, the key point for the control of TAA
selectivity is limit the rate of formation of the undesired TMDP com-
pared to the rate of TAA formation (Scheme 5). From the reaction
stoichiometries, one can infer that water may play an important
role in the control of selectivity, since it is involved in the key
Fig. 1. Acetone conversion and selectivity to products in the presence of NH4Cl
catalyst in function of reaction time. Reaction conditions: acetone/ammonia mol.
ratio 4.0, catalyst/acetone wt ratio 0.02, T 65 ◦C. Symbols: (ꢀ) Acetone conversion
(%). Selectivity (%) to: (᭹) TAA, (ꢁ) ACTN, (ꢂ) TMDP, (ꢀ) DAA, ( ) MO, and (×)
DAAM.

328
G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
Scheme 3. An hypothetical mechanism for interconversion of TAA and ACTN.
Scheme 4. An hypothesis about the mechanism for the irreversible transformation of TAA into TMDP.
reaction of ACTN transformation into TAA; indeed, this is the only
reaction, amongst all direct reactions shown in Scheme 5, where
water is a reactant, in all other cases being a co-product. Even in the
case of DAAM transformation into TAA, water is eliminated during
the condensation reaction; however, this reaction contributes less
to TAA formation than that starting from ACTN, due to the lower
initial DAAM selectivity. According to patent literature, the ACTN
transformation into TAA may be carried out either without water
in the presence of acetone, or without acetone in the presence of
water [6,10,11]. On the other hand, the role of water in the one-
pot process has not yet been clearly demonstrated. Fig. 4 shows a
comparison between the reaction carried out under the usual con-
ditions (acetone-to-ammonia molar feed ratio 4.0, T 65 ◦C) and the
same reaction conducted in the presence of anhydrous sodium sul-
phate as the dehydrating agent; it is worth noting that the amount
of salt added was calculated so as to completely remove water from
the reaction medium.
Fig. 4 demonstrates that the biggest difference between the two
experiments concerned the yields to ACTN and TAA. In fact, water
removal decreased the rate of ACTN transformation to TAA (the sum
of yields to the compounds remained the same): a clear indication
that the formation of TAA was accelerated in the presence of water.
On the other hand, the yield of TMDP was not affected so much;
in fact, its formation does not involve a contribution from water.
In conclusion, water is essential for promoting the formation of
TAA, and for obtaining a higher TAA/TMDP selectivity ratio. We also
carried out an experiment adding twice as much water than that
already present when the NH4OH/H2O solution was used; however,
we did not observe any significant effect, thus indicating that the
amount of water available from ammonia solution is sufficient to
observe the selectivity enhancement effect.
3.4. The role of ammonium: acid catalysis in alkaline
(NH4OH/H2O) reaction medium
The ammonia source used for catalytic experiments was NH4OH
in aqueous solution (the Kb value for NH3 hydration with water
dissociation is as low as 1.85 × 10−5). In order to better understand
the role of the ammonium ion in the salt used as the homogeneous
catalyst, specific tests were carried out under the following con-
ditions: acetone-to-ammonia molar ratio 4.0, temperature 65 ◦C,
catalyst-to-acetone ratio wt. 0.02, while using various catalysts.
As shown in Table 2, the blank test (reaction with only acetone
and NH4OH) gave little conversion of acetone with traces of DAA
only (aldol condensation in alkaline conditions), and small amounts
of ACTN. On the other hand, however, no reaction occurred when
only acetone and NH4NO3/H2O were used (the latter with the same
molar amount as for the ammonium hydroxide used previously).
+
This means that the co-presence of both NH3 and NH4 is needed
to carry out this reaction. It is important to note that the Ka value
for ammonium ion in water is equal to 5.6 × 10−10; therefore NH4
+
is a weak Brønsted acid. When the reaction was carried out under
basic conditions, achieved by means of NaOH 50 wt%, there was
a low acetone conversion, with a formation of DAA and traces of
Scheme 5. Reaction scheme of TAA synthesis by direct reaction between acetone and ammonia.

G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
329
Table 2
Acetone conversion (%) and yield to products (%) in function of base source and catalyst.
Base, catalyst
Conv.
TAA yield
ACTN yield
TMDP yield
Others yield
NH4OH, no
NH4OH, NH4Cl
NH4NO3, no
5
45
0
0
22.5
0
4
7.5
0
0
8.5
0
1
6
0
NH4OH, NH4NO3
42
5
21
0
5
0
8
0
8
5
NaOH, noa
NaOH, NH4NO3
12
0
0
0
12
Reaction conditions: temperature 65 ◦C, catalyst/acetone wt ratio 0.02, reaction time 6 h, acetone/ammonia mol. ratio 4.0. Others: MO + DAA + DAAM.
a
NaOH 50 wt% replacing ammonium hydroxide.
Fig. 3. Comparison of catalytic behaviour for different acetone/ammonia molar
ratios at the maximum acetone and ammonia conversion (top), and for similar ace-
tone conversions (bottom), in the presence of NH4Cl catalyst. Other conditions as in
Figs. 1 and 2.
Fig. 2. Acetone conversion and selectivity to products in the presence of NH4Cl cat-
alyst and with different acetone/ammonia molar ratios in function of reaction time.
Top: acetone/ammonia mol. ratio 2.5. Bottom: acetone/ammonia mol. ratio 10.0.
Reaction conditions: T 65 ◦C, catalyst/acetone wt ratio 0.02. Symbols: (ꢀ) Acetone
conversion (%). Selectivity (%) to: (᭹) TAA, (ꢁ) ACTN, (ꢂ) TMDP, (ꢀ) DAA, ( ) MO,
and (×) DAAM.
the attack of ammonia. Moreover, the co-presence of stable acid
sites (ammonium ion) and ammonia makes possible the occur-
rence of several reactions at the very beginning of the process.
Indeed, the formation of various aldol condensation compounds
MO only. After the addition of a catalytic amount of an ammonium
salt, the conversion of acetone increased (from 5% to 12%), with
higher DAA and MO yields. Thus the ammonium ion shows the cat-
alytic activity for the aldol condensation of acetone also, even in the
presence of a strong base. In conclusion, it is important to note that
the ammonium ion is the only acid specie that exists in an ammo-
nia environment. In other words, whichever Brønsted acid is used
as the reaction catalyst, it immediately reacts with ammonia, thus
generating in situ the corresponding ammonium cation (a weaker
acid species). Table 2 also shows that using either NH4Cl or NH4NO3
led to quite similar results, evidencing a negligible anion effect for
ammonium salts of strong acids.
Since the main primary product of the reaction was ACTN, it is
reasonable to assume the presence of a mechanism involving a very
reactive intermediate species, e.g. 2-iminopropane (dimethylke-
timine). The latter may be formed by the direct condensation of
acetone and ammonia. In this case, the ammonium ion acts by
coordinating the carbonylic moiety and making it more prone to
Fig. 4. Effect of water removal by anhydrous sodium sulfate in the presence of
NH4Cl catalyst. Reaction conditions: T 65 ◦C, acetone/ammonia mol. ratio 4.0, cata-
lyst/acetone wt ratio 0.02, and reaction time 6 h.

330
G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
Fig. 5. Acetone conversion in function of reaction time. Symbols: catalyst HY-6 (ꢁ),
HY-15 (♦), HY-200 (ꢀ), and NH4Y-15 (ꢀ). Reaction conditions: acetone/ammonia
mol. ratio 4.0, catalyst/acetone wt ratio 0.10, T 65 ◦C.
3.5. Heterogeneous catalysis with zeolites
Scheme 6. Aldol condensation of acetone and dimethylketimine (or the corre-
sponding enolic forms).
One of the drawbacks of the homogeneous catalysis is the diffi-
culty of catalyst recovery; complete catalyst recovery is extremely
important if the catalyst is expensive and/or toxic. In TAA synthe-
sis, ammonium salts such as NH4Cl are not expensive, but must be
separated and processed after reaction, with the aim of avoiding
product contamination. Therefore, an easily separated, heteroge-
neous catalyst would be an important achievement.
As shown in Table 1, heterogeneous catalysts have also been
claimed for the one-pot synthesis of TAA, based mainly on organic
sulfonated resins. The only report on the use of an inorganic solid
acid deals with an attempt of combining the Lewis acid properties of
Ca2+ cation in a Y zeolite, and the efficient homogeneous properties
of the ammonium cation [12,16]. In the latter case, the use of a very
large amount of ammonium salt made it possible to operate under
mild temperature conditions (those typically used for the synthesis
of ACTN, the first step in the two-step process), thus achieving a
remarkably high selectivity to TAA, but still with the low TAA yield
typically obtained in this reaction.
Having established the crucial role of the ammonium cation in
the reaction, we decided to use commercial zeolites, both in the acid
form and after the exchange of protons with ammonium cations
(see details in Section 2 regarding the procedure for the prepara-
tion of the exchanged Y zeolite). In fact, we hypothesized that, in
the presence of NH3/H2O, the rapid generation of the ammonium
cation might promote the in situ development of active sites with
the desired acid strength.
by reaction with the corresponding enolic or enaminic forms is
possible, as summarized in Scheme 6. DAAM and DAA were the
most stable aldol condensation compounds, a fact that made it
possible to obtain them as reaction products, whereas the very
reactive imino species immediately underwent consecutive trans-
formations by reaction with either acetone or 2-iminopropane, and
were not isolated as intermediate compounds. The driving force of
the consecutive reaction was the formation of the cyclic ACTN. In
this case, too, the role of the ammonium ion is probably crucial in
catalyzing ACTN formation, as proposed in the hypothetical mech-
anism shown in Scheme 7. A catalytic test carried out by using DAA
as reagent, in the presence of ammonium hydroxide and ammo-
nium salt catalysts, confirmed this hypothesis because the main
product obtained was acetone (40% yield), with lower amounts of
other compounds (DAAM, ACTN, TAA, and others). This means that
the starting point for obtaining N-containing compounds is ace-
tone, and not DAA, although the latter may be used as a reactant,
because it may reversibly yield acetone.
Fig. 5 compares the conversion based on reaction time, achieved
for H-Y zeolites having different silica-to-alumina ratios (SAR). It
is possible to notice that the activity of the H-Y zeolites was not
much affected by the SAR; moreover, the Y zeolite with SAR equal
to 15 showed a similar activity when used either in the proton or in
the ammonium form. All these data suggest that the rate-limiting
step of the process might be either the intraparticle diffusion of
reactants or the counterdiffusion of products, due to the very high
reaction rate compared to the diffusion rate. On the other hand, it
must be remembered that the acidity of the H-Y zeolites in basic
(NH3) aqueous medium is levelled off due to the presence of ammo-
nia and water. Nevertheless, the quick generation of the ammonium
cation in acidic zeolites ensured a rapid reaction kinetics, similar to
that shown with the pre-prepared ammonium form of the same
zeolite.
On the other hand, the distribution of products was affected by
catalyst features, as shown in Table 3, which compares the selectiv-
ity to the main products (ACTN, TAA and TMDP), for similar values
of acetone conversion. Fig. 6 details the effect of reaction time on the
selectivity to the various products, for HY-6 and HY-200 samples. In
this case also, as for the NH4Cl catalyst, the primary product ACTN
Scheme 7. Hypothetical intermediates and mechanism in ACTN formation.

G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
331
Table 3
Acetone conversion (%) and selectivity to products (%) for experiments conducted
with zeolites.
Catalyst
Reaction time (h)
Conv.
TAA
ACTN
TMDP
HY-6
HY-15
HY-200
H␤-150
NH4Y-15
6, 17
6, 17
6, 17
6
36, 40
38, 43
35, 42
56
15, 27
14, 22
12, 19
4
42, 26
42, 24
42, 25
56
5, 14
13, 25
12, 28
5
6
44
10
72
6
Reaction conditions: temperature 65 ◦C, catalyst/acetone wt ratio 0.10, reaction time
6 h, acetone/ammonia mol. ratio 4.0 (except H␤, mol ratio 2.5).
was converted into TAA. The selectivity to TAA and TMDP was con-
siderably affected by the SAR: in particular, the formation of TMDP
was fostered by the HY-200 catalyst. As also shown in Table 3, an
increase in the SAR led to a decreased selectivity to TAA and a cor-
responding higher selectivity to TMDP. The effect shown may be
interpreted considering that the higher the SAR in the H-Y zeolite is,
the lower its hydrophilicity, and vice versa [17–21]. Therefore, the
behaviour shown is in accordance with the homogeneous catalysis
data: in the more hydrophobic HY-200 zeolite, the water molecules
necessary to convert ACTN into TAA are less available within the
zeolite pores, while the competitive reaction of TMDP formation is
kinetically more facilitated than that of TAA formation.
With regard to the performance of the ammonium-Y, a dif-
ference can be seen from the comparison between HY-15 and
NH4Y-15 (Table 3); despite their similar activity (Fig. 5), the consec-
utive transformation of the intermediately formed ACTN into TAA
and then TMDP was hindered, compared to the H-Y zeolite.
Fig. 7. Acetone conversion and selectivity to products as functions of time with HY-
6 (Top) and HY-200 (Bottom) zeolites. Reaction conditions: acetone/ammonia mol.
ratio 10.0, catalyst/acetone wt ratio 0.10, T 65 ◦C. Symbols: (ꢀ) Acetone conversion
(%). Selectivity (%) to: (᭹) TAA, (ꢁ) ACTN, (ꢂ) TMDP, (ꢀ) DAA, ( ) MO, and (×)
DAAM.
These data indicate that the generation of the Z− NH4 ion pair,
+
the formation of the protonated dimethylketimine (occurring by
Table 4
Acetone conversion (%) and selectivity to main products (%) with the HY-6 zeolite.
T (◦C), t (h)
Conv.
TAA
ACTN
TMDP
65, 6
20
19
17
17
48
40
14
29
11
14
49
36
10
14
8
45, 24
25, 72
25, 24a
6
Reaction conditions: catalyst/acetone wt ratio 0.10, acetone/ammonia mol. ratio
10.0.
a
Catalyst 50 wt%.
incorporation of the ammonium cation from the zeolite, a reaction
whose rate may be a function of the acid strength) [22,23] and the
further reaction of the adsorbed imine into the key intermediate
compound, ACTN, are all only marginally affected by the acidity
features. Conversely, the further transformation of ACTN (which
diffuses into the liquid phase and then undergoes consecutive reac-
tions) may be affected both by zeolite features and by the nature of
the active site.
Shown in Table 3 are also the results obtained with a H␤ zeolite.
The peculiarity of this zeolite was its high selectivity to ACTN, and a
low selectivity to consecutive by-products, despite the high acetone
conversion. Thus, this zeolite type would be more suited to use for
the first step of the two-step process, where the first step is aimed
at the production of ACTN.
We finally used the more advantageous acetone–ammonia
molar feed ratio (equal to 10.0); results for HY-6 and HY-200 are
shown in Fig. 7. This shows that not only was the rate of ACTN
transformation into consecutive products greatly accelerated with
HY-6 (Fig. 7, top) as compared to HY-200 (Fig. 7, bottom), but the
selectivity ratio between TAA and TMDP was also notably greater.
The effect of the reaction temperature on the catalytic behaviour
of the HY-6 zeolite is shown in Table 4; experiments were con-
ducted under conditions aimed at achieving comparable values of
Fig. 6. Acetone conversion and selectivity to products in the presence of HY-6 (top)
and HY-200 (bottom) zeolites in function of reaction time. Reaction conditions: ace-
tone/ammonia mol. ratio 4.0, catalyst/acetone wt ratio 0.10, T 65 ◦C. Symbols: (ꢀ)
Acetone conversion (%). Selectivity (%) to: (᭹) TAA, (ꢁ) ACTN, (ꢂ) TMDP, (ꢀ) DAA,
(
) MO, and (×) DAAM.

332
G. Gliozzi et al. / Journal of Molecular Catalysis A: Chemical 393 (2014) 325–332
acetone conversion, i.e. using longer reaction times when lower
temperatures were used. The most remarkable effect was a higher
selectivity to ACTN, and a correspondingly lower selectivity to TAA
(that to TMDP being less affected), when the reaction temperature
was decreased. This agrees with literature claims that in order to
limit consecutive transformations occurring on ACTN, and lastly to
obtain high selectivity to this compound (the first step in the two-
step process), it is necessary to use low reaction temperatures, at
least with homogeneous catalysts.
formation of TMDP was low, but a strong interaction of TMDP with
the zeolite would have caused a catalyst deactivation over time.
Conversely, the zeolite maintained its activity over time, and there
was no change in colour. It is also possible that the continuous
flow of the reaction mixture over the fixed-bed allowed a rapid
desorption of products (an event which, conversely, is more hin-
dered when the reaction is carried out in the batch reactor), thus
keeping the catalyst surface clean.
Figs. 2 (bottom) and 7 (top) make it possible to compare the cat-
alytic behaviour shown by NH4Cl and by the HY-6 zeolite, which
was the one showing the best performance among the heteroge-
neous systems investigated. When homogeneous catalysis is used,
the TAA-to-TMDP selectivity ratio can be tuned by optimizing
the acetone-to-ammonia molar feed ratio, whereas in heteroge-
neous catalysis with zeolites the effect of this parameter is not
straightforward, since the effective concentration within the zeo-
lite pores is greatly affected by zeolite characteristics. However,
our results demonstrate that by combining the best acetone-to-
ammonia molar feed ratio of 10.0, as inferred from homogeneous
catalysis experiments, with the best zeolite (HY-6), it is possible
to obtain a catalytic performance which is very similar to that
observed with the homogeneous system (Fig. 2, bottom). Indeed,
it is shown that when a low acetone/ammonia ratio was used, the
catalytic behaviour was clearly better with homogeneous catalysis
than with HY zeolites. However, when the high feed ratio – i.e. the
one leading to the best selectivity to TAA under homogeneous catal-
ysis conditions – was used, the catalytic behaviour was still worse
with the HY-200 zeolite than with the homogeneous system, but
became similar to the latter one when the HY-6 zeolite was used.
With regard to the reusability of the zeolite catalyst, we noted
that when the formation of TMDP was low, after reaction the cata-
lyst appeared slightly yellowish, and could be reused without any
post-treatment, showing no variation in catalytic behaviour. Con-
versely, when the amount of TMDP formed was non-negligible, the
catalyst turned red, because of the surface adsorption of TMDP, and
in this case a washing treatment with acetone was needed before
reusing the catalyst; for instance, with the HY-6 zeolite showing
9.6 0.3 TAA% yield in the first use (Table 4), the same catalyst
gave 9.3 0.3% in the second and 9.8 0.3% in the third use.
Additional experiments conducted in a fixed-bed, continuous-
flow reactor, with recirculation of the liquid phase and continuous
make-up of the fresh solution containing both acetone and ammo-
nium hydroxide aqueous solution, offered a better assessment of
the absence of any catalyst deactivation. In this case, the reaction
was carried out at 45–50 ◦C, to minimize evaporation phenomena
causing gas accumulation problems in the reactor head, while keep-
ing the recirculation rate – referring to the purge/make-up flow rate
– very high in order to permit a 35% acetone conversion. The cata-
lyst was the H␤, in the form of extrudates; we chose this catalyst
because of its peculiar behaviour (Table 3), with high selectivity
to ACTN. The reaction was carried out for 10 h. At the beginning,
during the first few hours of reaction, the concentration of the var-
ious products increased in the recirculated liquid, because it was
necessary to reach a steady-state concentration of all components
in the reactor hold-up. After approximately 2 h, the steady-state
behaviour of the reactor was reached, and the composition of
the recirculated liquid remained constant (because the amount
of the product withdrawn within the purge stream equalled the
amount generated by the reaction), providing no catalyst deacti-
vation phenomenon occurred. In fact, under the conditions used,
the prevailing product was ACTN, and we registered a constant
34.0 0.2% ACTN yield, 2.2 0.1% DAA yield, 0.4% TAA yield, 0.15%
MO yield, and 0.1% TMDP yield over a reaction time of 10 h. The
4. Conclusions
In this work, the direct synthesis of 2,2,6,6-tetramethy-4-
oxopiperidone (triacetonamine, TAA) from acetone and ammo-
nium hydroxide aqueous solution was investigated in the liquid
phase. The reaction network includes parallel and consecutive reac-
tions, most of which are reversible. This means that the key element
is the minimization of the consecutive (and irreversible) reaction
leading to the undesired 2,2,4,6-tetramethyl-2,5-dihydropyridine
(TMDP) formation.
The acetone/ammonia ratio notably affected the process perfor-
mance; the best results in terms of selectivity to TAA were obtained
by using the higher acetone/ammonia ratio, with the homogeneous
NH4Cl catalyst. Water played an important role in the selectivity,
especially in the ratio between TAA and TMDP.
HY zeolites with different features (SAR ratio and thus different
hydrophilicity characteristics) led to different behaviours. The best
selectivity was obtained by combining the more hydrophilic zeo-
lite HY-6 (with lower SAR ratio) with the higher acetone–ammonia
molar feed ratio. This behaviour was quite similar to that obtained
with the homogeneous catalyst, but presented the further advan-
tage of an easily separable and reusable catalyst.
Acknowledgment
INSTM is acknowledged for the PhD grant to Gherardo Gliozzi.
References
[1] A.M. Chalmers, D.R. Randell, US Patent 3759926 (1973), Ciba-Geigy Co.
[2] J. Kirchhoff, F. Kraushaar, Polym. Polym. Compos. 8 (2000) 245–254.
[3] H. Wiezer, G. Norwy, H. Häberlein, US Patent 4356308, Hoechst Akt (1982).
[4] Y. Nakahara, N., Kubota, B., Hirai, T. Haruna, US Patent 4418196, Adeka Argus
Chem. Co (1983).
[5] G. Cantatore, P. Cassandrini, US Patent 4536581, Ciba-Geigy S.p.A (1985).
[6] T. Haruna, A. Nishimura, K. Sugibuchi, US Patent 4661597 (1987), Adeka Argus
Chem. Co.
[7] W. Kruse, J.F. Stephen, US Patent 4734502 (1988), ICI Americas Inc.
[8] M. Julius, D. Hermeling, H. Siegel, W. Harder, US Patent 5856494 (1999), BASF
Akt.
[9] B.A. Oude Alink, US Patent 3904625 (1975), Petrolite Co.
[10] I. Orban, H. Lind, H. Brunetti, J. Rody, US Patent 3953459 (1976), Ciba-Geigy Co.
[11] K. Murayama, S. Morimura, T. Yoshioka, T. Kurumada, US Patent 3959298
(1976), Sankyo Co. Ltd.
[12] Y.C. Son, S.L. Suib, R.E. Malz Jr., in: D. Morrell (Ed.), Catalysis of Organic Reac-
tions, vol. 89, Marcel Dekker Inc., CRC Press, New York, 2003, pp. 559–564.
[13] I. Orban, E. Troxler, US Patent 4275211 (1981), Ciba-Geigy Co.
[14] D. Taylor, B. Milligan, US Patent 4831146 (1989), Air Products and Chemicals
Inc.
[15] A. Wu, W. Yang, X. Pan, Synth. Commun. 26 (1996) 3565–3569.
[16] R.E. Malz Jr., Y.C. Son, S.L. Suib, US Patent 6646127 (2003), Uniroyal Chem. Co.
and Univ. Connecticut.
[17] J.A. Martens, D.E. De Vos, C.E.A. Kirschhock, P.P. Pescarmona, B. Sels, I.F.J. Vank-
elecom, Top. Catal. 52 (2009) 1119–1130.
[18] M.J. Climent, A. Velty, A. Corma, Green Chem. 4 (2002) 565–569.
[19] T. Okuhara, Chem. Rev. 102 (2002) 3641–3665.
[20] M. Bolognini, F. Cavani, L. Dal Pozzo, L. Maselli, F. Zaccarelli, M. Armandi, E.
Garrone, Appl. Catal. A 272 (2004) 115–124.
[21] M. Armandi, B. Bonelli, E. Garrone, M. Ardizzi, F. Cavani, L. Dal Pozzo, L. Maselli,
R. Mezzogori, G. Calestani, Appl. Catal. B 70 (2007) 585–596.
[22] J. Kotrla, D. Nachtigallová, V. Bosácˇek, J. Nováková, Phys. Chem. Chem. Phys. 1
(1999) 2613–2620.
[23] T. Xu, J. Zhang, J.F. Haw, J. Am. Chem. Soc. 117 (1995) 3171–3178.