Intermolecular condensation of ethylenediamine to 1,4-diazabicyclo(2,2,2)octane over H-ZSM-5 catalysts: Effects of Si/Al ratio and crystal size

Article abstract of DOI:10.1016/j.apcata.2010.02.031

The intermolecular condensation of ethylenediamine (EDA) to 1,4-diazabicyclo [2.2.2] octane (DABCO) or triethylenediamine (TEDA) was conducted over H-ZSM-5 catalysts. The effects of reaction parameters, Al content and crystal size on the EDA condensation over H-ZSM-5 have been investigated. The H-ZSM-5 catalyst with a medium Al content (Si/Al = 110) and a small crystal size (ca. 100 nm) showed 99% EDA conversion and afforded a TEDA yield as high as 74% under optimized conditions. The mechanism for the H-ZSM-5-catalyzed condensation of EDA has also been considered. The framework Al-related Br?nsted acid sites were assumed to contribute to selective intermolecular condensation of EDA to TEDA. The primary intermolecular condensation of EDA to piperazine (PIP) took place mainly inside the micropores of the MFI structure, while the secondary condensation of PIP with EDA to TEDA was favored by the acid sites located near the pore entrance and on the external surfaces of crystals.

Full text of DOI:10.1016/j.apcata.2010.02.031

Applied Catalysis A: General 379 (2010) 45–53
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Intermolecular condensation of ethylenediamine to 1,4-diazabicyclo(2,2,2)octane
over H-ZSM-5 catalysts: Effects of Si/Al ratio and crystal size
∗
∗
Yong Wang, Lifang Guo, Yun Ling, Yueming Liu , Xiaohong Li, Haihong Wu, Peng Wu
Shanghai Key Laboratory of Green Chemistry and Chemical Processes, Department of Chemistry, East China Normal University,
North Zhongshan Rd. 3663, Shanghai 200062, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
The intermolecular condensation of ethylenediamine (EDA) to 1,4-diazabicyclo [2.2.2] octane (DABCO)
or triethylenediamine (TEDA) was conducted over H-ZSM-5 catalysts. The effects of reaction parameters,
Al content and crystal size on the EDA condensation over H-ZSM-5 have been investigated. The H-ZSM-5
catalyst with a medium Al content (Si/Al = 110) and a small crystal size (ca. 100 nm) showed 99% EDA
conversion and afforded a TEDA yield as high as 74% under optimized conditions. The mechanism for the
H-ZSM-5-catalyzed condensation of EDA has also been considered. The framework Al-related Brønsted
acid sites were assumed to contribute to selective intermolecular condensation of EDA to TEDA. The
primary intermolecular condensation of EDA to piperazine (PIP) took place mainly inside the micropores
of the MFI structure, while the secondary condensation of PIP with EDA to TEDA was favored by the acid
sites located near the pore entrance and on the external surfaces of crystals.
Received 21 October 2009
Received in revised form 23 February 2010
Accepted 26 February 2010
Available online 6 March 2010
Keywords:
Ethylenediamine
Triethylenediamine
Piperazine
Condensation
H-ZSM-5
© 2010 Elsevier B.V. All rights reserved.
1
. Introduction
because of their outstanding thermal and hydrothermal stability,
suitable acidity, and unique shape selectivity [6–10]. The synthesis
1
,4-Diazabicyclo [2.2.2] octane (DABCO) or triethylenediamine
of TEDA from MEOA over modified MCM-41 catalysts has also
been reported [11]. In addition, it has been patented that Al-free
TS-1 catalyst is capable of catalyzing the addition reaction of EDA
and PIP to TEDA [12]. These previous studies all disclose that
the unique shape selectivity and suitable acidic properties allow
the MFI-type zeolites to be suitable catalysts for intermolecular
condensation of EDA (or with other amino compounds) to TEDA.
Nevertheless, the reaction mechanism for the zeolite catalysts
is still not well addressed. The determining parameters for the
selective formation of TEDA, particularly the correlation between
the nature of zeolite acidity and product distributions, remain
unclear.
In a previous study [13], we verified that TS-1 zeolites were
superior to Ti-MWW, Ti-Beta, Ti-FER and Ti-MOR for the inter-
molecular condensation of EDA to TEDA. TS-1 with a relatively
high Ti content, corresponding to Si/Ti ratio of 30, showed 95% EDA
conversion and a total selectivity of ca. 85% to PIP and TEDA. By
means of ion-exchange experiment and IR spectra investigation,
the active sites for TEDA formation were assumed to be the acidic
Si–OH groups adjacent to the “open” Ti sites. In addition, TS-1 with
a small crystal size is suitable for the secondary condensation of PIP
to TEDA, a reaction involving bulky molecules and requiring open
reaction spaces such as pore entrances and external surfaces. The
findings with TS-1 stimulated us to investigate the possibility of
selective production of TEDA on its aluminosilicate analogue, ZSM-
(
TEDA), is one of the principal catalysts used for producing
polyurethanes from polyisocyanates and polyols [1]. It also acts as
an anti-fade reagent that scavenges the free radicals produced by
excitation of fluorochromes. In addition, TEDA serves as a useful
structure directing agent in zeolite synthesis [2]. The known pro-
cesses for TEDA preparation differ either in the nature of starting
materials or in the types of catalysts. It is possible to synthesize
TEDA using a range of different starting materials: for example, the
raw materials that involve ethylenediamine, monoethanolamine
(
MEOA),
aminoethyethanolamine,
piperazine
(PIP),
or
substituted piperazines such as aminoethylpiperazine, N-(2-
hydroxyethyl)piperazine and N,N-di(2-hydroxyethyl)piperazine,
and also mixtures of the foregoing [3–12]. Recently, researchers
tend to use the mixtures of EDA and MEOA in the process of TEDA
preparation because of the low price and availability of MEOA. On
the other hand, a variety of catalysts are reported to be applicable
to TEDA synthesis. Early routes to synthesize TEDA involve the
vapor-phase condensation of substituted piperazines over alumina
catalysts [3,4], and the liquid-phase reaction in the presence of
aromatic carboxylic acid catalysts [5]. In the recent years, much
attention has been focused on modified ZSM-5 zeolite catalysts
∗ Corresponding authors. Fax: +86 21 6223 2292.
E-mail addresses: ymliu@chem.ecnu.edu.cn (Y. Liu), pwu@chem.ecnu.edu.cn
5
, because the manufacturing cost of ZSM-5 is much lower and its
(
P. Wu).
acidity and morphology are more easily controllable.
0
926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.02.031

4
6
Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
In this study, we have synthesized the H-ZSM-5 zeolites with
run, 1 g of catalyst (18–30 mesh size) was activated at 673 K for
various Al contents and crystal sizes, and we have investigated
their effectiveness and catalytic properties in the intermolecular
condensation of EDA to TEDA. The active sites for the EDA conver-
sion have been correlated to the zeolite acidity. The issues of the
reaction space’s contributions to the consecutive condensation of
EDA to PIP and finally to TEDA have also been considered.
3 h in a flow of N . After cooling to the reaction temperature, the
2
reaction was started by feeding an EDA solution (50 wt% aqueous
−
1
solution) into the reactor at a flow rate of 5 mL h together with
a nitrogen flow of 30 mL min−1. The products were collected in a
receiver cooled with an ice-salt bath. The amounts of the products
were quantified on a Shimadzu 14B gas chromatograph equipped
with an OV-1 column (30 m) and an FID detector, and they were
identified using authentic samples or by GC–MS (Agilent 6890).
2
. Experimental
3. Results and discussion
2.1. Catalyst preparation and characterizations
3.1. A summary of catalysts characterization
ZSM-5 with various Si/Al ratios was hydrothermally synthesized
using sodium aluminate (Al 28.1%, NaOH 54.0%), tetraethylorthosil-
icate (TEOS), and tetrapropylammonium hydroxide (TPAOH, 20%
aqueous solution) by modifying the procedures reported elsewhere
All the zeolites synthesized had the MFI topology and a relatively
high crystallinity, and contained no impurity of any other phases as
detected by XRD patterns (not shown). Table 1 lists the Si/Al ratios,
particle sizes and textural properties of H-ZSM-5 synthesized with
various Al contents when using TEOS as silicon source and TPAOH
as SDA. The samples had reasonably high surface area (Langmuir)
[
14]. Typically, sodium aluminate was first dissolved in an aqueous
solution containing TPAOH as a structure-directing agent (SDA),
which led to a clear solution after stirring at room temperature for
0.5 h. TEOS was then added into the above solution. The resultant
2
−1
in the range of 556–605 m g as determined by N2 adsorption,
which indicated that they were of good quality. The images from
scanning electron micrographs (SEM) revealed that the samples
with a lower Al content, corresponding to a Si/Al >110, were com-
posed of 200 nm spherical crystals (Table 1, Nos. 4–6), while the
samples with a higher Al content had the same crystal morphology
but larger crystal sizes of 450–1000 nm (Table 1, Nos. 1–3). Thus,
the crystal size depended greatly on the Si/Al ratio in the gels even
when using the same Si source and SDA.
mixture was stirred for about 3 h to hydrolyze TEOS completely.
Next, it was further heated at 353 K under stirring to evaporate
ethanol formed, resulting in clear gels with the molar compositions
of 1 SiO :(0.00167–0.02) Al O :0.25 TPAOH:30 H O. Finally, the
2
2
3
2
gels were charged into Teflon-lined stainless autoclaves and crys-
tallized at 443 K for 2 days under autogenous pressure and static
conditions. After the crystallization, the solid product was collected
by centrifugation, washed thoroughly with distilled water, dried at
3
83 K overnight and calcined in air at 823 K for 10 h to burn off the
H-ZSM-5 catalysts with various crystal morphologies and sizes
were synthesized at a Si/Al ratio of 120 by varying the silicon
source and SDA. As shown in Fig. 1, the H-ZSM-5 sample obtained
from fumed silica and TPAOH consisted of the cylinder shaped
crystals of ca. 3 ␮m × 2 ␮m (Fig. 1a), while the sample prepared
from granule silica and n-butylamine showed a plate morphology
of 1 ␮m × 0.5 ␮m × 0.4 ␮m crystals (Fig. 1b). On the other hand,
the samples synthesized from TEOS and TPAOH exhibited similar
spherical crystal morphology, however, their crystal sizes varied
with the TPAOH/Si ratio in the gels. With increasing TPAOH/Si
ratio, the average crystal size tended to decrease from 2 ␮m to
occluded organic species. In addition, the ZSM-5 samples with tun-
able crystal sizes were hydrothermally synthesized from the above
system by varying the TPAOH/Si ratio while keeping a fixed Si/Al
ratio of 120. For control experiments, Silicalite-1 with similar crys-
tal morphology and size was also prepared by this method but in
the absence of sodium aluminate.
Highly crystalline ZSM-5 zeolites with other crystal morpholo-
gies were hydrothermally synthesized either using fumed silica
(
Cab-o-sil) and TPAOH or using granule silica and n-butylamine at
Si/Al ratio of 120. All the Na-form samples were transformed into
H-form by ion-exchange with 1 M HNO3 at 353 K for 24 h and then
calcination in air at 773 K for 5 h.
The catalysts were characterized by various techniques. The
Si/Al ratio was determined by inductively coupled plasma (ICP) on
a Thermo IRIS Intrepid II XSP atomic emission spectrometer after
dissolving the samples in HF solution. The X-ray diffraction (XRD)
patterns were recorded on a Bruker D8 ADVANCE diffractometer
1
[
1
00 nm (Figs. 1c–e), which was consistent with previous reports
14]. It is worth mentioning that the sample crystal size was about
50–250 nm at a TPAOH/Si ratio of 0.5 (Fig. 1f), which is probably
owing to nanoparticle agglutination at a high alkalinity [15].
The H-ZSM-5 samples with various crystal sizes synthesized
using TEOS and TPAOH were further characterized by powder
XRD patterns (Fig. 2). A close inspection of the diffraction pat-
terns reveals that the peak intensities decreased, while the line
widths increased slightly, as the crystal size decreased. Thus, the
broadening and attenuation of the peaks is apparently a result
of decreasing crystal size [16]. Table 2 lists the particle sizes
(
Cu K␣). The N2 adsorption was carried out at 77 K on an Autosorb
Quancachrome (02108-KR-1) instrument. The crystal morphology
and size were measured by scanning electron microscopy (SEM)
on a Hitachi S-4800 microscope. Thermogravimetric analysis (TGA)
was measured on METTLER TOLEDO TGA/SDTA851e in air atmo-
sphere. IR spectra were collected on a Nicolet NEXUS-FTIR-670
spectrometer. The IR spectra of pyridine adsorption were recorded
Table 1
−2
Physicochemical properties of ZSM-5 with various Al contents.
as follows: a self-supported wafer (9.6 mg cm thickness and Ø
cm) was set in a quartz IR cell sealed with CaF2 windows, where
Si/Ala
Average particle
size (nm)
Textural properties
2
No.
b
it was evacuated at 773 K for 2 h before the pyridine adsorption.
The adsorption was carried out by exposing the wafer to a pyri-
dine vapor (1.3 kPa) at 323 K for 0.5 h. The physisorbed pyridine
was then removed by evacuation at 423 K for 1 h. The IR spectra
were collected at room temperature.
PVc (mL g−1
SSAc (m2
−1
g )
)
1
2
3
4
5
6
26
41
59
110
194
252
1000
800
450
200
200
200
0.27
0.24
0.31
0.29
0.31
0.29
556
559
572
605
594
592
2
.2. Condensation of EDA to TEDA
a
Molar ratio determined by ICP.
Evaluated by SEM.
Pore volume (PV) and specific surface area (SSA, Langmuir) measured by N2
b
c
The intermolecular condensation of EDA was carried out in a
down-flow fixed-bed reactor at atmospheric pressure. In a typical
adsorption at 77 K.

Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
47
Fig. 1. SEM images of ZSM-5 samples synthesized at Si/Al ratio of 120 with fumed SiO2 and TPAOH (a), granule SiO2 and n-butylamine (b), TEOS and TPAOH at TPAOH/Si
ratios of 0.1 (c), 0.25 (d), 0.4 (e) and 0.5 (f).
and textural properties of H-ZSM-5 with various crystal sizes.
With increasing TPAOH/Si ratio in the starting gels, the crystal
size tended to decrease, while the pore volume and surface area
increased. Moreover, the external area increased drastically when
the crystal size changed from micro scale particles to nanoparticles.
In summary, we have succeeded in synthesizing the H-ZSM-5
samples with various Al contents and crystal sizes. Their physico-
chemical properties suggested that these samples were qualified
as solid acid catalysts for the intermolecular condensation of EDA.
condensation product of PIP. Secondly, PIP further condensed with
one more EDA molecule to give rise to TEDA. Meanwhile, PIP par-
tially underwent dehydrogenation to form pyrazine, which may
further react with the products from EDA degradation to produce
corresponding derivatives, alkyl pyrazines.
It is known that the acid property of H-ZSM-5 is one of key
parameters which determine the activity and selectivity in acid
catalysis reactions such as hydrocarbon isomerization and the alky-
lation of hydrocarbons [17–20]. Since the acid property is closely
related to the framework Al, a series of H-ZSM-5 catalysts with Si/Al
ratios of 26–252 were thus prepared and applied to the intermolec-
ular condensation of EDA together with Silicalite-1 catalyst. These
samples that were all prepared from TEOS and TPAOH had similar
spherical crystal morphology. The reaction was carried out under
optimized reaction conditions (the details will be given in later sec-
tions). Al-free Silicalite-1 gave 24% EDA conversion and 13% TEDA
selectivity. It means that the neutral silanol groups contained in
Silicalite-1 were not active and selective to the reaction because
of a too weak acidity. Fig. 3 shows the changes of EDA conversion
and product selectivity with time on stream (TOS) for the H-ZSM-5
catalysts with various Al contents. The initial activity and deactiva-
tion rate closely depended on the Si/Al ratio (Fig. 3A). The H-ZSM-5
with medium or high Al content (Si/Al ≤110) gave nearly 100% EDA
3.2. Intermolecular condensation of EDA over H-ZSM-5 catalysts
with various Si/Al ratios
The GC–MS analysis showed that the major products in the
intermolecular condensation of EDA were PIP and TEDA, while the
byproducts were lower aliphatic amines (ethylamine and methy-
lamine) and the derivatives of pyrazine, which was in agreement
with the results in the literature [9,13]. According to the mechanism
reported previously [8–10,13] and the products actually obtained,
a general reaction pathway of intermolecular condensation of EDA
was obtained. As depicted in Scheme 1, the intermolecular con-
densation of EDA was a typical successive reaction. Firstly, an
intermolecular deamination of two EDA molecules led to a primary

4
8
Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
For example, the EDA conversion of H-ZSM-5 (Si/Al = 26) decreased
to 76% at TOS of 4 h. Thus, the deactivation rate of H-ZSM-5 followed
an order of 26 > 41 > 59 > 110 ≈ 194 ≈ 252 in Si/Al ratio.
Although the H-ZSM-5 catalysts with a high Al content (Si/Al
≤
59) showed an initial conversion of about 100% EDA, they gave
lower TEDA selectivity and higher byproduct selectivity than H-
ZSM-5 (Si/Al = 110) (Fig. 3B–D). This implies that the H-ZSM-5
catalysts with too many Brønsted acid sites may easily cause the
side reaction as well as the coke formation, which leads to a rapid
deactivation and also lowers the TEDA selectivity. As the cata-
lysts deactivated, the PIP selectivity increased gradually with TOS
(Fig. 3C). It seems that the secondary step of the EDA intermolecu-
lar condensation, that is, the reaction of PIP with EDA to TEDA, was
greatly retarded. From the viewpoints of the activity and in particu-
lar the duration and product selectivity, the results inferred that the
H-ZSM-5 catalyst with a medium Al content was the most effective
and stable catalyst for the intermolecular condensation of EDA to
TEDA. The H-ZSM-5 (Si/Al = 110) gave the maximum EDA conver-
sion (99%) and the highest TEDA selectivity (63%) (Fig. 3A and B).
Therefore, the effects of various reaction parameters on the
reaction have been investigated in detail by using the catalyst of
H-ZSM-5 (Si/Al = 110).
Fig. 2. XRD patterns of ZSM-5 samples synthesized with TEOS and TPAOH at
TPAOH/Si ratio of 0.1 (a), 0.25 (b) and 0.4 (c). The Si/Al ratios of the gels was fixed
at.120.
3.3. Effects of reaction parameters on the intermolecular
condensation of EDA over H-ZSM-5 (Si/Al = 110)
conversion at initial stage, whereas those catalysts with lower Al
content showed only ca. 85% EDA conversion because of an insuffi-
cient number of acid sites for EDA condensation. A sharp decrease in
EDA conversion was observed for the ZSM-5 with a high Al content.
3.3.1. Effect of reaction temperature
The reaction temperature exhibited a great effect on the inter-
molecular condensation of EDA over H-ZSM-5 catalyst. As shown
Table 2
Effect of TPAOH/Si ratio on the particle size and textural property of H-ZSM-5.
No.
TPAOH/Si
Si/Ala
Average particle sizeb (nm)
Textual propertiesc
PVc (mL g−1
)
SSAc (m2
g
−1
)
External SSAd (m2
g )
−1
1
2
3
0.10
0.25
0.40
105
110
108
2000
200
100
0.24
0.29
0.40
582
605
623
114
163
214
a
Molar ratio determined by ICP.
Evaluated by SEM.
Pore volume (PV) and specific surface area (SSA, Langmuir) measured by N2 adsorption at 77 K.
b
c
d
External surface area calculated by t-plot method.
Scheme 1. .

Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
49
Fig. 3. The changes of EDA conversion (A) and selectivities of TEDA (B), PIP (C) and other products (D) with the time on stream in the intermolecular condensation of EDA
on H-ZSM-5 catalysts with Si/Al ratios of 26 (a), 41 (b), 59 (c), 110 (d), 194 (e) and 252 (f). Reaction conditions: catalyst, 1 g; pressure, 1 atm; temp., 623 K; EDA feed (50 wt%
−1
−1
aqueous solution), 5 mL h ; N2, 30 mL min .
in Fig. 4, the conversion of EDA increased with increasing tempera-
ture and remained at >99% when the temperature was above 623 K,
whereas the PIP selectivity decreased monotonously and the TEDA
selectivity increased first and then decreased. The TEDA selectivity
reached a maximum value of 63% at 623 K, where the PIP selec-
tivity was 21%, respectively. When the temperature was further
raised to 673 K, EDA was 100% converted, whereas the selectivities
to TEDA and PIP sharply decreased to 41% and 7%, and the forma-
tion of byproducts was clearly enhanced. This can be attributed to
two reasons. First, the Lewis acid sites may be formed as a result
of the dehydroxylation of structural Si(OH)Al groups at high tem-
peratures, which then contributes to the formation of pyrazine
derivatives [21]. Moreover, an elevated temperature makes the
degradation of EDA to aliphatic amines take place more easily.
These two kinds of side reactions then would decrease the selec-
tivities to PIP and TEDA. The reaction at temperatures below 623 K
does not favor the formation of pyrazine and its derivatives, but is
suitable for the formation of the main products of PIP and TEDA [9].
However, too low temperatures are energetically not enough for
converting EDA to PIP and further to TEDA. Consequently, the fore-
going experiments showed that the intermolecular condensation
of EDA to TEDA proceeded most effectively at an optimum reaction
temperature of 623 K.
3.3.2. Effect of concentration of EDA
Fig. 5 shows the effect of the concentration of EDA on the
reaction at 623 K. The EDA conversion was comparably high (>95%)
in the EDA concentration range of 10–70%. The product selec-
tivity varied greatly with the EDA concentration. At 50% of EDA
concentration, the catalyst gave the highest selectivity to TEDA
(63%) as well as the highest total selectivity to PIP and TEDA (79%).
When the EDA concentration was lowered, the selectivity of PIP
decreased greatly owing to the formation of byproducts, although
the selectivity of TEDA remained high. At higher concentrations
of EDA (90%), the conversion of EDA and the TEDA selectivity
decreased. Meanwhile, the effluent became very thick because of
a high concentration of heavy products with high boiling points,
and the catalyst deactivated more rapidly. The steam is assumed
Fig. 4. The effect of reaction temperature on the intermolecular condensation of
EDA on H-ZSM-5 (Si/Al = 110). Other reaction conditions: see Fig. 3.

5
0
Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
reactions take place more easily, leading to a relatively high selec-
tivity to byproducts such as lower aliphatic amines and pyrazine
derivatives.
3.4. Effects of crystal size on the intermolecular condensation of
EDA
Since the EDA reaction involves the reaction steps of inter-
molecular condensation of linear diamine molecules to monocyclic
intermediates and further to more bulky multicyclic products, the
crystal size which determines the diffusion and adsorption dis-
tance of the reactant and product molecules is presumed to exhibit
a great effect on the activity of catalyst and product distribution.
Thus, we have synthesized the H-ZSM-5 catalysts with different
crystal sizes by varying silicon source and SDA, but fixing the Si/Al
ratio at 120. As shown in Table 3, these H-ZSM-5 catalysts (Si/Al
ratios of 105–112) then had very comparable Al content. They all
showed high EDA conversation (>98%) independent of the crys-
tal size. However, the TEDA selectivity was influenced greatly by
the crystal size. The sphere-shaped H-ZSM-5 with 0.1 ␮m crystal
size gave the highest TEDA selectivity (75%) (Table 3, No. 5). When
the H-ZSM-5 samples with similar spherical morphology increased
in crystal size from an average size of 0.1 to 2 ␮m, the selectivity
to TEDA dropped from 75% to 45% (Table 3, Nos. 3–6). In addition,
the cylinder and plate-shaped H-ZSM-5 samples with larger crystal
sizes gave lower selectivity to TEDA (Table 3, Nos. 1–2). It is notice-
able that the selectivities to PIP and byproducts decreased with
decreasing crystal size. This indicated that the decrease of crystal
size was advantageous to the adsorption and diffusion, which made
the primary condensation of EDA to PIP and the secondary conden-
sation of PIP with EDA to TEDA proceed smoothly. Therefore, it is
preferable to prepare the H-ZSM-5 catalysts with small crystal sizes
in order to achieve a high EDA conversion and high selectivities to
PIP and TEDA.
Fig. 5. The effect of EDA concentration on the intermolecular condensation of EDA
on H-ZSM-5 (Si/Al = 110). Other reaction conditions: see Fig. 3.
to function as a refreshing reagent by washing out off the basic
product molecules adsorbed strongly on the acid sites inside
zeolite channels or on crystal surfaces. This is then helpful for
reducing the coke formation on the catalytic active sites. Thus, a
moderate concentration of EDA is preferable to achieve a high EDA
conversion and high selectivity to PIP and TEDA.
3.3.3. Effect of contact time
The effect of contact time on the reaction has been investigated
to investigate the primary products in the EDA condensation. The
contact time expressed as catalyst weight to flow (W/F) ratio was
varied in a wide range. The reactions were carried out by varying
the catalyst amount and the flow of carrier gas at 623 K by using an
−
1
3.5. Investigation into the active sites for the intermolecular
condensation of EDA over H-ZSM-5 catalyst
EDA (50 wt% aqueous solution) feed rate of 5 mL h . As shown in
Fig. 6, the conversion of EDA and the selectivity to TEDA increased
with increasing W/F ratio, while the selectivity to PIP decreased.
Apparently, the dependence of product distribution on contact time
clearly indicated that PIP was the primary product in the reac-
tion, and the formation of TEDA was a result of consumption of
PIP and EDA, which further supported the possible reaction path-
ways (Scheme 1). Meanwhile, a long contact time made the side
Fig. 3 shows that the EDA conversion and the TEDA selectivity
depended closely on the Al content of H-ZSM-5. The results implied
that the catalytic active sites for the EDA condensation to TEDA
were related to the framework Al-related acid sites. In solid acid-
catalyzed reactions, the framework Al species of H-ZSM-5 can serve
as Brønsted acid sites [17,18].
IR spectroscopy has been adopted to investigate the acidity of
H-ZSM-5. Fig. 7 shows the IR spectra in the region of hydroxyl
stretching vibration of H-ZSM-5 with different Al contents and
Silicalite-1 after dehydration under vacuum at 773 K. Silicalite-1
−
1
showed bands at 3725 and 3500 cm (Fig. 7a), which are attributed
to the terminal silanol on internal surfaces and to the hydrogen
bonded silanol nests, respectively [22]. This implied the presence
of a relatively high concentration of defect sites in Silicalite-1 [23].
When Al was introduced into the MFI framework, the intensity of
−
1
−1
the 3725 and 3500 cm bands decreased and the 3741 cm band
attributed to the terminal silanol on the external surface of zeo-
lite crystals gradually became sharp (Fig. 7b–d), which means the
incorporation of Al tends to mend these defects to reduce their
number [24]. In addition, the incorporation of framework Al devel-
−
1
oped a new band at 3610 cm due to the bridging hydroxyl groups
in H-ZSM-5 [25]. When the Al content was low, i.e. at Si/Al = 252,
−
1
the 3610 cm -band was almost unobservable (Fig. 7b). However,
the band was obvious at medium or high Al content and the inten-
sity increased gradually with increasing Al content (Fig. 7c–d). In
addition, H-ZSM-5 (Si/Al = 41) with a high Al content showed a
−
1
weak band at 3663 cm assigned to the hydroxyl related to extra
framework Al [25].
Fig. 6. The effect of contact time on the intermolecular condensation of EDA on
H-ZSM-5 (Si/Al = 110). Other reaction conditions: see Fig. 3.

Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
51
Table 3
a
The results of intermolecular condensation of EDA over H-ZSM-5 catalysts with various crystal sizes .
Si/Alb
Crystalc
External surfaced (m2
g
−1
)
EDA conv. (%)
Sel. (%)
PIP
No.
Morphology
Average size (␮m)
TEDA
Otherse
1
2
3
4
5
6
105
106
105
110
108
112
Plate
Plate
Sphere
Sphere
Sphere
Sphere
4 × 3 × 2
102
128
114
163
214
156
98.9
99.3
98.4
98.7
99.1
99.2
35.4
26.5
25.1
16.0
11.7
20.0
32.5
53.4
45.3
63.2
74.5
61.8
32.1
20.1
29.6
20.8
13.8
18.2
1 × 0.5 × 0.4
2
0.2
0.1
0.15
a
Reaction conditions: see Fig. 3.
Determined by ICP.
Given by SEM.
External surface area calculated by t-plot method.
Mainly lower aliphatic amines and the derivatives of pyrazine.
b
c
d
e
Pyridine adsorption was carried out to investigate further the
3.6. Investigation into the reaction space for intermolecular
condensation of EDA over H-ZSM-5 catalyst
acidity of the H-ZSM-5 samples with various Al contents. Fig. 8
compares the spectra of the samples after removing physisorbed
pyridine species by evacuation at 423 K. Al-free Silicalite-1 showed
As shown in Table 3, the catalytic properties were strongly influ-
enced by the crystal size of H-ZSM-5. The H-ZSM-5 catalyst with
100 nm crystal size gave the highest selectivity to TEDA (75%) as
well as the highest total selectivities to PIP and TEDA (86%) (Table 3,
No. 5). Reducing crystal size improved the condensation of EDA to
PIP, especially, it enhanced effectively the consecutive condensa-
tion of PIP with EDA to TEDA. In addition, the external surface area
of H-ZSM-5 measured by t-plot method increased with decreas-
ing the crystal size (Table 3). Thus, the external surface is closely
related to the catalytic activity in the intermolecular condensation
of EDA since the reaction involves steps of small reactant molecules
to monocyclic intermediate and further to more bulky multicyclic
product.
−1
weak bands at 1600 and 1446 cm (Fig. 8a), which are associated
with hydrogen bonded pyridine [26]. In addition, the band at 1546,
−1
1
446 and 1490 cm characteristic of Brønsted acidity, Lewis acid-
ity and both of acidities, respectively, were absent in the spectra
of Silicalite-1, suggesting that it did not contain strong acid sites
but neutral sinaols [27]. With increasing Al content, the band at
−1
1
546, and 1490 cm increased in intensity gradually, whereas the
−1
band at 1454 cm showed almost no change (Fig. 8b–d). We thus
deduce that H-ZSM-5 contained not only stronger but also more
Brønsted acid sites in comparison to with Silicalite-1 [28].
Correlating the behavior EDA condensation (Fig. 3) to the zeolite
acidity (Fig. 8), we can infer that the active sites for the inter-
molecular condensation of EDA are those framework Al species
showing Brønsted acidity. In the case of EDA condensation on TS-1,
a maximum Ti incorporation is advantageous to achieve high EDA
conversion and TEDA yield [13]. The hydroxyl groups developed as
a result of Ti incorporation are of weak to moderate strength, but
the acidity of the framework Al in H-ZSM-5 is much stronger. Thus,
to convert EDA to TEDA selectively on the solid acid catalyst of H-
ZSM-5, it is desirable to control the acid amount or Al content to a
medium extent. This would avoid catalyst deactivation as well as
side reactions like decomposition and deep condensation.
Fig. 9 shows the IR spectra in the region of hydroxyl stretching
vibration of H-ZSM-5 with different crystal sizes after dehydration
at 773 K. Though the spectra of these samples differed in shape and
baseline slope, which is associated to the light scattering depen-
dent on the crystal size [29], all samples showed the same strong
−
1
bands at 3741, 3725 and 3500 cm , which are ascribed to the ter-
minal silanol on external/internal surface and the hydrogen bonded
silanol nests, respectively. In addition, they also showed a weak
−
1
band at 3610 cm assigned to the framework Al. With decreasing
Fig. 7. IR spectra in hydroxyl stretching region of dehydrated Silicalite-1 (a) and
Fig. 8. IR spectra after pyridine desorption at 423 K for Silicalite-1 (a) and H-ZSM-5
H-ZSM-5 samples with Si/Al ratio of 252 (b), 110 (c) and 41 (d).
samples with Si/Al ratios of 252 (b), 110 (c) and 41 (d).

5
2
Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
Fig. 10. IR spectra after pyridine desorption at 423 K for the H-ZSM-5 samples syn-
thesized at TPAOH/Si ratios of 0.1 (a), 0.25 (b) and 0.4 (c). The Si/Al ratio in the gels
was fixed at 120.
Fig. 9. IR spectra in hydroxyl stretching region of dehydrated H-ZSM-5 samples
synthesized at TPAOH/Si ratios of 0.1 (a), 0.25 (b) and 0.4 (c). The Si/Al ratio in the
gels was fixed at 120.
pores of MFI structure (ca. 0.55 nm). Thus, the condensation of PIP
with EDA is presumed to take place mainly on the acid sites near
the pore entrance and on the external surface.
crystal size, that is, increasing TPAOH/Si ratio in the gels, the band at
−1
3
741 cm due to external sinaols tended to be sharp and intensive,
To further verify the importance role of pore entrance and outer
surface of catalyst crystal in the condensation of PIP with EDA
to TEDA, we conducted the intermolecular condensation of EDA
over H-ZSM-5 with or without removal of the structure-directing
agent occluded in the pores. The difference between the calcined
sample and unclacined sample is that the former has the open
pore structure that makes the active sites available to the reac-
tion, whereas only the active sites near the pore entrance and on
the external surface were exposed to the reaction in the latter cat-
alyst. This was evidenced clearly by the TPA content and specific
surface (Table 4). Considering that the decomposition tempera-
ture for TPA in the pores is higher than 573 K [30], we carried out
the EDA condensation at a lower temperature of 523 K to avoid
thermal decomposition of TPA. The fraction of PIP in the reaction
mixture was 19.8% when the reaction was conducted over calcined
H-ZSM-5 catalyst (Table 4, No. 1). It decreased to 2.4% when uncal-
cined H-ZSM-5 catalyst was used (Table 4, No. 2). This is because
the uncalcined H-ZSM-5 catalyst has only the active sites near
pore entrance and on the external surface, the number of which
is insufficient to convert EDA into PIP. This result suggests that the
intermolecular condensation of EDA to PIP occurs mainly inside the
pores. Reasonably, a very low TEDA fraction (0.3%) was obtained in
the reaction over uncalcined H-ZSM-5 catalyst when the primary
condensation of EDA proceeds to a low level.
−
1
and it became a dominant band in the region of 3750–3650 cm
.
This is in agreement with the enlargement of external area (Table 2).
After the adsorption and desorption of pyridine at 423 K (Fig. 10),
−1
the intensity of the 1546 cm -bands ascribed to Brønsted acid sites
was very similar for the samples synthesized at the same Al con-
tent but different TPAOH/Si ratio, which means that the total acid
amount in these H-ZSM-5 catalysts was almost independent of the
crystal size.
In order to make the primary condensation of EDA to PIP
and the secondary condensation of PIP with ED to TEDA both
move on smoothly, it is desirable to reduce the crystal size
and then to increase the external surface area. On basis of the
analysis by Gaussian 2003 software, the average molecular sizes
are 0.31 nm × 0.54 nm for EDA, 0.27 nm × 0.48 nm for PIP and
0.46 nm × 0.48 nm × 0.48 nm for TEDA. Therefore, we assume that
the first intermolecular condensation of EDA to PIP takes place
mainly inside the medium pores of the MFI structure. With respect
to secondary condensation of PIP with EDA to TEDA, the 10-MR
channels of H-ZSM-5 would impose mass transfer limitations for
the TEDA molecules which have a polycyclic shape and a relatively
rigid structure. Moreover, TEDA may be formed via a bulky interme-
diate of aminoethylpiperazine (0.27 nm × 0.48 nm × 0.7 nm), which
is not easily accommodated inside the 10-membered ring (MR)
Table 4
a
The results of intermolecular condensation of EDA or mixture of EDA and PIP over ZSM-5 catalysts .
SSAb (m2
g
−1
)
TPAc (wt%)
Feed
Product fraction (%)
No.
Cat.
EDA
PIP
TEDA
Othersd
1
2
3
4
H-ZSM-5 (calcined)
H-ZSM-5 (uncalcined)
HZSM-5 (calcined)
623
154
623
154
0
13.2
0
50% EDA
50% EDA
25% EDA + 25% PIP
25% EDA + 25% PIP
66.9
96.5
23.1
36.1
19.8
2.4
54.0
58.5
8.6
0.3
18.7
4.0
4.7
0.8
4.2
1.4
H-ZSM-5 (uncalcined)
13.2
a
Reaction conditions: temp., 523 K; others see Fig. 3.
Specific surface area (Langmuir) measured by N2 adsorption at 77 K.
Measured by TGA.
b
c
d
Mainly lower aliphatic amines and the derivatives of pyrazine.

Y. Wang et al. / Applied Catalysis A: General 379 (2010) 45–53
53
With the purpose of clarifying this issue, we carried out the
References
condensation of PIP with EDA (PIP:EDA = 1:1) over the above two
catalysts. The uncalcined H-ZSM-5 catalyst showed a TEDA frac-
tion of 4.0% in the reaction mixture (Table 4, No. 4). Although
this value was lower than that obtained with the calcined sam-
ple (Table 4, No. 3), it was much higher than 0.3% shown by the
uncalcined sample in EDA condensation (Table 4, No. 2). This indi-
cated the active sites near pore entrance and on external surface
had an important role in the condensation of PIP with EDA to
TEDA, a reaction requiring open spaces. From the above results,
we deduce that the first intermolecular condensation of EDA to PIP
takes place mainly inside the medium pores of ZSM-5, and then
the constitutive condensation of PIP with EDA to TEDA may occur
on the acid sites near the pore entrances and/or on the external
surface.
[
[
1] Ullmann’s Encyclopedia of Industrial Chemistry, vol. A2, VCH, 1998, p. 16.
2] A. Belhekar, T.K. Das, K. Chaudhari, S.G. Hegde, A.J. Chandwadka, Stud. Surf. Sci.
Catal. 113 (1998) 195–201.
[3] H.G. Bosche, K. Baer, K. Schneider, DE Patent 2,442,929 (1976).
[
[
[
4] I. Hudea, M.B. Werner, R. Czifra, RO Patent 85,563 (1984).
5] G. Engemann, G. Spielberger, US Patent 3,080,371 (1963).
6] J.G. Santiesteban, H.-X. Li, J.N. Armor, US Patent 5,741,906 (1998).
[7] J.N. Armor, J.G. Santiesteban, H.-X. Li, US Patent 5,756,741 (1998).
[
[
8] N. Srinivas, D. Venu Gopal, B. Srinivas, S.J. Kulkarni, M. Subrahmanyam, Micro-
porous Mesoporous Mater. 51 (2002) 43–50.
9] Y. Bhat, J. Das, S. Ali, B.D. Bhatt, A.B. Halgeri, Appl. Catal. A: Gen. 148 (1996)
L1–L6.
[
[
10] R. Anand, B.S. Rao, Catal. Commun. 3 (2002) 479–486.
11] M. Selvaraj, B.R. Min, Y.G. Shul, T.G. Lee, Microporous Mesoporous Mater. 74
(
2004) 157–162.
[12] M. Frauenkron, B. Stein, US patent 6,562,971 B2 (2003).
13] Y. Wang, Y.M. Liu, X.H. Li, H.H. Wu, M.Y. He, P. Wu, J. Catal. 266 (2009) 258–267.
14] A.E. Persson, B.J. Schoeman, J. Sterte, J.E. Otterstedt, Zeolites 15 (1995) 611–619.
15] R. Van Grieken, J.L. Sotelo, J.M. Menendez, J.A. Melero, Microporous Mesoporous
Mater. 39 (2000) 135–147.
[
[
[
4
. Conclusions
[
16] O.A. Fouad, R.M. Mohamed, M.S. Hassan, I.A. Ibrahim, Catal. Today 116 (2006)
8
2–87.
H-ZSM-5 zeolite is an effective catalyst in the intermolecular
[
17] G.P. Babu, M. Santra, V.P. Shiralkar, P. Ratnasamy, J. Catal. 100 (1986) 458–465.
condensation of EDA to TEDA, giving 99% EDA conversion and 74%
selectivity to TEDA when it is synthesized to have a small crys-
tal size (100 nm) as well as a medium Al content corresponding to
Si/Al ratio of 110. It is the Brønsted acid sites related to the frame-
work Al in H-ZSM-5 that contribute to the condensation of EDA.
A small crystal size of H-ZSM-5 benefits more the secondary con-
densation of PIP to TEDA, a bulky reaction which requires open
reaction spaces such as pore entrances and external surfaces, than
the primary condensation of EDA to PIP.
[18] Y. Sugi, Y. Kubota, K. Komura, N. Sugiyama, M. Hayashi, J.-H. Kim, G. Seo, Appl.
Catal. A 299 (2006) 157–166.
[
19] C. Jiménez, F.J. Romero, R. Roldán, J.M. Marinas, J.P. Gómez, Appl. Catal. A: Gen.
49 (2003) 175–185.
2
[
20] Z.R. Zhu, Q.L. Chen, Z.K. Xie, W.M. Yang, C. Li, Microporous Mesoporous Mater.
88 (2006) 16–21.
[
[
[
21] P.R. Reddy, M. Subrahmanyam, S.J. Kulkarni, Catal. Lett. 54 (1998) 95–98.
22] E. Astorino, J.B. Peri, R.J. Willey, G. Busca, J. Catal. 157 (1995) 482–500.
23] A. Zecchina, S. Bordiga, G. Spoto, L. Marchese, G. Petrini, G. Leofanti, M. Padoan,
J. Phys. Chem. 96 (1992) 4991–4997.
[
[
[
[
[
[
[
24] L.M. Kustov, V.B. Kazansky, S. Beran, L. Kubelkova, P. Jiru, J. Phys. Chem. 91
(
1987) 5247–5251.
25] A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Petrini, G. Leofanti, M. Padovan,
C. Otero Arean, J. Chem. Soc. Faraday Trans. 88 (1992) 2959–2967.
26] R. Buzzoni, S. Bordiga, G. Ricchiardi, C. Lamberti, A. Zecchina, G. Bellussi, Lang-
muir 12 (1996) 930–940.
27] C. Pazé, S. Bordiga, C. Lamberti, M. Salvalaggio, A. Zecchina, G. Bellussi, J. Phys.
Chem. B 101 (1997) 4740–4751.
Acknowledgements
We gratefully acknowledge the NSFC of China (20925310,
2
0873043), the Science and Technology Commission of Shang-
hai Municipality (09XD1401500, 08JC1408700), 973 Program
2006CB202508), 863 Program (2007AA03Z342, 2008AA030801),
28] F. Bonino, A. Damin, S. Bordiga, C. Lamberti, A. Zecchina, Langmuir 19 (2003)
2155–2161.
(
29] T. Armaroli, L.J. Simon, M. Digne, T. Montanari, M. Bevilacqua, V. Valtchev, J.
Patarin, G. Busca, Appl. Catal. A: Gen. 306 (2006) 78–84.
30] G. Debras, A. Gourgue, J.B. Nagy, G. De Clippeleir, Zeolites 5 (1985) 377–383.
and the Shanghai Leading Academic Discipline Project (B409). Y.W.
thanks Ph.D. Program Scholarship Fund of ECNU 2008.