Development of 2,2′-iminodiethanol selective production process using shape-selective pentasil-type zeolite catalyst

Article abstract of DOI:10.1246/bcsj.80.1075

Ethanolamines are produced on an industrial scale exclusively by the reaction of ethylene oxide with an aqueous solution of ammonia. The reaction is a typical consecutive reaction with three steps; therefore, it is difficult to produce 2,2′-iminodiethanol (the second product of the reaction) with high selectivity by conventional means. We developed a catalytic 2,2′-iminodiethanol selective production process using a pentasil-type zeolite catalyst modified with rare earth elements. This highly active catalyst was able to recognize the difference at molecular level between 2,2′-iminodiethanol and 2,2′,2″-nitrilotriethanol; 2,2′,2″-nitrilotriethanol was not formed in its micropore. We also succeeded in producing a binderless molded zeolite catalyst that does not form problematic impurities and that has a shape suitable for a fixed-bed reactor. The problem of activity deterioration was overcome by developing a regeneration process using high-temperature and high-density ammonia gas as a rinse medium.

Full text of DOI:10.1246/bcsj.80.1075

Ó 2007 The Chemical Society of Japan
Bull. Chem. Soc. Jpn. Vol. 80, No. 6, 1075–1090 (2007) 1075
Award Accounts
The Chemical Society of Japan Award for Technical Development for 2005
Development of 2,20-Iminodiethanol Selective Production Process
Using Shape-Selective Pentasil-Type Zeolite Catalyst
ꢀ1
Hideaki Tsuneki, Masaru Kirishiki,1 and Tomoharu Oku2
1Strategic Technology Research Center, Nippon Shokubai Co., Ltd., Suita, Osaka 564-8512
2Strategic Technology Research Center, Nippon Shokubai Co., Ltd., Tsukuba 305-0856
Received July 26, 2006; E-mail: hideaki tsuneki@shokubai.co.jp
Ethanolamines are produced on an industrial scale exclusively by the reaction of ethylene oxide with an aqueous
solution of ammonia. The reaction is a typical consecutive reaction with three steps; therefore, it is difficult to produce
2,20-iminodiethanol (the second product of the reaction) with high selectivity by conventional means. We developed a
catalytic 2,20-iminodiethanol selective production process using a pentasil-type zeolite catalyst modified with rare earth
elements. This highly active catalyst was able to recognize the difference at molecular level between 2,20-iminodiethanol
and 2,20,200-nitrilotriethanol; 2,20,200-nitrilotriethanol was not formed in its micropore. We also succeeded in producing a
binderless molded zeolite catalyst that does not form problematic impurities and that has a shape suitable for a fixed-bed
reactor. The problem of activity deterioration was overcome by developing a regeneration process using high-temper-
ature and high-density ammonia gas as a rinse medium.
DEA
1. Introduction
H
N
k2
ð2Þ
ð3Þ
+
H2N
HO
OH
OH
Ethanolamines are commodity chemicals produced at a
rate of about 1:2 ꢁ 106 metric tons per year throughout the
world. They are versatile chemicals that have the properties
of amines and alcohols. Because of the amine functionality,
they are mildly alkaline and react with acids to form salts.
The alcohol functionality permits esterification and gives
hygroscopic properties. Ethanolamines are used in diverse
areas, such as gas sweetening, detergent and specialty cleaner
formulations, intermediates for ethyleneamines, agricultural
chemicals, like herbicides, urethane foam catalysts, pharma-
ceuticals, textile processing aids, metalworking, and oil-well
rust preventatives.
Ethanolamines were prepared in 1860 by Wurtz1 from
ethylene chlorohydrin and aqueous ammonia. In 1897, an
ethanolamine mixture was separated into 2-aminoethanol
(MEA), 2,20-iminodiethanol (DEA), and 2,20,200-nitrilotrietha-
nol (TEA) components by Knorr by fractional distillation.2
Ethanolamines were not available commercially before the
early 1930s, and their commercial importance as intermediates
started growing steadily only after 1945, because of the large-
scale production of ethylene oxide (EO). The synthetic method
for ethanolamines is shown in Eq. 1 to 3, which is a typical
consecutive reaction with 3 steps.
O
HO
H
k3
+
N
TEA
HO
OH
O
N
HO
OH
Today, ethanolamines are produced on an industrial scale
exclusively by the reaction of EO with excess ammonia.3 As
pointed out by Knorr, the reaction of EO with anhydrous am-
monia takes place very slowly and is accelerated by water.4,5
The kinetics for this aqueous ammonia and EO reaction have
been studied by Potter and McLaughln,6 Lowe et al.,7 and Miki
et al.8
The water must be separated from the products, which is an
energy-consuming disadvantage. This has led to an increasing
interest in anhydrous methods using heterogeneous catalysts.
Weibull et al. have reported the possibility of using ion-ex-
change resins. A potential disadvantage with organic ion-ex-
change resins is their lack of stability at high temperatures.
The requirements of having high-temperature stability and of
containing acidic sites are met by many inorganic solid acids,
such as acidic silica–aluminas, zeolites, and acid clays.
All of the reaction steps have about the same activation en-
ergy and show a roughly quadratic dependence of the reaction
rate on the water content of the ammonia–water mixture used.
Therefore, product distribution depends solely on the molar ra-
tio of ammonia to EO and not on the water content, reaction
EO
O
MEA
H
k1
H2N
ð1Þ
+
H
N
OH
H
Published on the web June 13, 2007; doi:10.1246/bcsj.80.1075

1076 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
Table 1. List of Rate Constant Ratios ꢀ, ꢁ in Aqueous Ammonia Solution
ꢀ
ꢁ
CW/mol Lꢂ1
Reference
Potter and McLaughln
Miki et al.
Ahn et al.
6
7.2–0.042CW
21.1–0.42CW
8.51–0.051CW
40
93
105
14
9.36–0.153CW
4
16–0.22CW
16.4–0.25CW
14.81–0.196CW
14
81
100
40
8.51–0.135CW
18–30
7–40
6
8
9
10
37
Park et al.
Stoyanov et al.
0.693
0.411
0.215
—
Johnson and Fred
Our results
38
This work
6–24
temperature, or pressure.11–13 Some homogeneous catalysts,
such as salts of inorganic or organic acids have been proposed
by Green14 and by Watanabe et al.15–20 to control product dis-
tribution.
In the early 1990s, Nippon Shokubai developed a modified
clay catalyst21–24 with high activity and selectivity for MEA
production, but it is very difficult to obtain DEA, which is a
second-step product of a three-step consecutive reaction, selec-
tively. In an apparent move to maximize DEA for use as an in-
termediate for glyphosate herbicides, Nippon Shokubai devel-
oped a process using a zeolite25–28 to bias the product distribu-
tion towards DEA without increasing the TEA. This process
was industrialized in 2003 and has operated prosperously on
a scale of 50000 metric tons.
90
80
70
60
50
40
30
20
10
0
MEA
DEA
TEA
0
5
10
15
NH3/EO molar ratio
2. Conventional Aqueous Ammonia Method
Fig. 1. Calculated product distribution as a function of
molar ratio of NH3 to EO. Reaction rate constant ratio
ꢀ ¼ ꢁ ¼ 5.
2.1 Reaction Kinetics in an Aqueous Ammonia and EO
System. Reaction rate ðr1; r2; r3Þ equations of each step in
reaction scheme are expressed as follows:
r1 ¼ k1CNCE ¼ dðCM þ CD þ CTÞ=dt;
r2 ¼ k2CMCE ¼ k1ꢀCMCE ¼ dðCD þ CTÞ=dt;
r3 ¼ k3CDCE ¼ k1ꢁCDCE ¼ dCT=dt;
ð4Þ
ð5Þ
ð6Þ
2
k3 ¼ k1ꢁ ¼ ð41 þ 4:0CW Þð16 ꢂ 0:22CWÞ 102
ꢃ
expðꢂ11000=RTÞ ¼ k1ð16 ꢂ 0:22CWÞ:
ð11Þ
ꢃ
They have also determined the reaction rate constants of reac-
tions with EO and ethanolamines in anhydrous ethanolamine
or high-concentration aqueous solutions of ethanolamine at
303 K.
where CN, CE, CM, CD, and CT, are concentration of ammonia,
EO, MEA, DEA, and TEA, respectively. k1, k2, and k3 are re-
action rate constants. ꢀ and ꢁ are defined as k2=k1 and k3=k1,
respectively. Let CN0 denote CN þ CM þ CD þ CT. We obtain
Eqs. 7 and 8 by integration of Eqs. 4, 5, and 6.
0
1=2
k2 ¼ ð6:6 þ 6:0CW Þ 10ꢂ3
:
ð12Þ
ð13Þ
ꢃ
ꢀ
ꢁ
ꢀ
0
1=2
CN
CN
k3 ¼ ð16 þ 3:1CW Þ 10ꢂ3
:
ꢃ
ꢂ
CN0
CN0
CM
Equation 12 or 13 is very different from Eq. 10 or 11. It is
thought that the hydroxy group of ethanolamines has catalytic
activity like water.
¼
¼
:
ð7Þ
CN0
1 ꢂ ꢀ
2
3
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
ꢁ
ꢀ
ꢁ
CN
CN
CN
CN
ꢂ
ꢂ
Table 1 shows the reaction constant ratios that we and other
researchers have determined. Solving Eqs. 7 and 8 numerical-
ly, these parameters ðꢀ; ꢁÞ have been calculated from product
distribution. Selectivity for MEA increases when these param-
eters diminish. We also obtained similar results. In the indus-
trially operable range of water concentration, ꢀ and ꢁ were
about 5–7. Figure 1 shows calculated product distribution
of the reactants in the case of ꢀ ¼ ꢁ ¼ 5 as a function of
NH3/EO molar ratio. In the most efficient condition (NH3/
EO = 2–4), three components were formed in similar quanti-
ty. A high NH3/EO ratio was required to obtain a high MEA
yield, and then large amounts of ammonia and water had to be
recycled.
6
6
7
7
C
CN0
CN0
CN0
CD
ꢀ
4
5
N0
ꢂ
:
1 ꢂ ꢀ ꢃ
CN0
1 ꢂ ꢁ
ꢀ ꢂ ꢁ
ð8Þ
Miki et al.8 have studied the reaction kinetics of this system
and have determined the rate constants k1, k2, and k3 as func-
tions of water concentration (CW). According to their results,
k1, k2, and k3 can be expressed as Eqs. 9, 10, and 11:
2
k1 ¼ ð41 þ 4:0CW Þ 102 expðꢂ11000=RTÞ;
ð9Þ
ꢃ ꢃ
k2 ¼ k1ꢀ ¼ ð41 þ 4:0CW Þð7:2 ꢂ 0:042CWÞ 102
2
ꢃ
expðꢂ11000=RTÞ ¼ k1ð7:2 ꢂ 0:042CWÞ;
ð10Þ
ꢃ

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1077
Table 2. Rate Constant Ratios of Various Homogeneous Catalysts in Aqueous Ammonia
Water
content
/wt %
EO
content
/wt %
NH3
content
/wt %
NH3/EO
molar
ratio
Catalyst
/wt %
Catalyst
ꢀ
ꢁ
a)
(NH4)2CO3
(NH4)2CO3
32.6
21.4
22.6
33.3
35.1
35.1
35.1
16.9
25.4
33.7
30.0
14.3
27.7
40.4
55.6
36.2
49.3
40.7
40.2
40.7
62.2
55.0
47.9
51.7
66.0
49.8
15.7
13.0
20.8
11.7
8.8
22.9
11.3
16.1
18.5
14.3
15.8
12.8
15.1
15.1
15.1
16.0
15.1
14.0
3.77
3.52
3.54
3.60
3.83
3.46
3.83
3.50
3.50
3.50
3.52
3.48
3.49
ca. 2
1.37
1.96
2.68
2.89
3.47
2.48
2.71
2.48
2.18
3.35
3.76
3.42
—
4.06
4.68
4.34
4.91
8.39
4.01
4.18
4.27
4.56
4.20
4.75
4.69
NH4Cl
NH4NO3
(NH4)2SO4
NH4Br
10.5
8.8
ClCH2COONH4
10.0
10.0
10.0
10.4
10.0
11.4
CH3COONH4
HCOONH4
CH3CH(OH)COONH4
a) Reynhart33 MEA yield: 66.2 wt %, other data based on Watanabe et al.15–17
2.2 Process Description of a Conventional Aqueous Am-
Table 3. Rate Constant Ratios of Various Homogeneous
Catalysts in Anhydrous System
monia Process. In a conventional process, the reaction takes
place in liquid phase, and the operation pressure must be suf-
ficiently large to prevent vaporization of ammonia at the reac-
tion temperature. In current procedures, the pressures are up to
15 MPa, the reaction temperatures are up to 420 K, and an ex-
cess up to 40 mol of ammonia per 1 mole of EO is used. The
reaction is highly exothermic; the heat of reaction is about
125 kJ per mole of EO. Unreacted ammonia and water, fed
in as a catalyst, are separated from the end products in a dis-
tillation line downstream of the reactor, and are recycled. In
large-scale continuous single-line plants, the requirement for
low energy use (i.e., operation with minimum steam consump-
tion) determines both the transport of heat from the reactor and
the design of the thermally integrated distillation line. The
ethanolamine mixture is worked up in vacuum distillation col-
umns to give pure products.
Product distribution of the three ethanolamines can be con-
trolled by an appropriate choice of the ammonia to EO ratio, as
shown in Fig. 1. A higher DEA or TEA content can also be ob-
tained by recycling MEA or DEA to the reactor or by reacting
them with ethylene oxide in a separate unit. Several improved
processes have been proposed by some producers.29,30 Gibson
et al.31,32 have shown that a high MEA yield is obtained in very
low water content when the reaction is carried out in super-
critical condition. For example, if CW is 1.02 mol Lꢂ1 (i.e.,
the water concentration equals 2.84 wt %), the NH3/EO ratio
is 44, reaction temperature is 347 K, and the reaction pressure
is 20.7 MPa, the product mixture had the following composi-
tion: MEA: 83 wt %, DEA: 15 wt %, TEA: 2 wt %. The disad-
vantage of this process is the high equipment cost due to the
high reaction temperature and pressure.
Concentration
/mol Lꢂ1
Content
/wt %
Catalyst
ꢀ
ꢁ
Water (control)
NH4Cl
6.14
0.97
3.22
1.0
17.2
8.4
24.8
7.8
8.81
8.56
6.00
8.3
8.08
7.42
5.76
8.6
CH3COONH4
3.0
23.4
7.7
6.1
nitrate, chloride, sulfate, and carboxylate. Rate-constant ratios
of these catalysts are shown in Table 2. They have also devel-
oped a process using these catalysts, such as the recovery of
the catalyst (as CO2), purification by distillation, and MEA
or DEA recycle.18–20 In the case of ammonium monochloro-
acetate, ꢀ values decrease (i.e., MEA selectivity increases)
along with an increase of the amount of catalyst. Ammonium
carbonate has especially excellent MEA selectivity, but in
spite of an improvement in the ꢀ, ꢁ values, they did not de-
crease markedly. In addition, this homogeneous catalyst sys-
tem required a large amount of catalyst (catalyst to EO molar
ratio 0.5–2.4) and water.
We examined the catalytic properties of ammonium salts
in anhydrous systems. Ammonium chloride and acetate indi-
cated catalytic activity, and their ꢀ and ꢁ values are listed
in Table 3. The behavior of ꢁ was similar to ꢀ, contrary to
Watanabe’s results. However, for practical use, this system al-
so required a large amount of catalyst, and catalyst recovery
was very difficult and uneconomical.
4. Heterogeneous Catalyst for MEA Selective Production
4.1 Ion-Exchange Resin. Product distribution is controlled
by the NH3/EO ratio in the reaction mixture. If an MEA-rich
composition is desired, the reaction must be carried out with a
greater NH3/EO ratio. When a conventional aqueous ammonia
method is used, a large amount of water as well as unreacted
ammonia must be recycled. Water has a very large latent heat
of vaporization, and thus, an anhydrous reaction system has
been desired. In the 1950s, Weibull described the reaction of
3. Homogeneous Catalyst
Some attempts have been made to avoid the disadvantage of
the aqueous ammonia system using a homogeneous catalyst.
Reynhart has revealed that ethanolamines can be produced
from ammonia or a primary amine and ethylene oxide, and a
salt of a weak acid, such as ammonium carbonate.33 Watanabe
et al.15–17 have proposed ammonium salts, such as carbonate,

1078 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
a)
Table 4. Catalytic Performance of Cation-Exchange Resin35
Reaction
temperature
/K
Reaction
pressure
/MPa
WHSV
/g gcatꢂ1 hꢂ1
Product distribution/wt %
NH3/EO
ꢀ
ꢁ
molar ratio
MEA
DEA
TEA
40
353b)
373b)
8.1
10.1
12.2
14.7
11.1
11.1
10
13
14
25
—
—
89.5
89.4
85.8
84.3
84.2
91.0
9.9
10.1
12.7
14.1
14.7
8.6
0.6
0.5
1.5
1.6
1.1
0.4
5.50
5.58
7.77
8.81
4.63
5.09
5.27
4.32
10.39
10.07
3.46
393b)
413b)
318–385c)
328–383c)
21
44
4.43
a) EO conversion > 99 mol %. b) Isothermal condition. c) Adiabatic condition.
alkylene oxides and ammonia in an anhydrous batch-wise re-
action system,34 using cation-exchange resins saturated with
hydrogen or non-quaternary ammonium ions and anion-ex-
change resins partly or fully saturated with organic or inorgan-
ic acids as catalysts. In 1968, Weibull et al. proposed a contin-
uous system using cation-exchange resins35 as the heterogene-
ous catalyst for an anhydrous NH3/EO reaction. They list
many types of cation-exchange resins with active groups like
sulfonic, carboxylic, and phosphoric acid groups; basic resins:
polystyrene, copolymers of styrene, and divinylbenzene, poly-
methacrylic acid, and processed coal. In their example, a com-
mercially available sulfonated styrene–divinylbenzene copoly-
mer strong acid-type resin has been used. The reaction is
carried out under elevated pressures (8–14.5 MPa) and reaction
temperatures (298–413 K) using a fixed-bed-type reactor. The
reaction can be carried out under isothermal or adiabatic con-
ditions. Under isothermal conditions, the preferred reaction
temperature range is 348–410 K. Since the reaction is strongly
exothermic, it is necessary to eliminate reaction heat from the
reaction mixture to keep the temperature approximately con-
stant. It is generally preferable to carry out the reaction under
adiabatic or nearly adiabatic conditions, because no cooling
equipment is required. In this case, the reactants are preheated
to 298–348 K before they are introduced into the reactor. Be-
cause of the reaction heat evolved, the reaction temperature
is rapidly increased. The maximum reaction temperature
may not be obtained until the mixture has reached the end of
the catalyst bed. Table 4 shows reaction conditions, product
distribution and kinetic parameters ðꢀ; ꢁÞ. MEA selectivity,
which is represented by ꢀ and ꢁ, of the ion-exchange resin
is similar to aqueous ammonia systems.
Kamemski et al.36 and Stoyanov et al.37 have studied the
catalytic reaction of EO and anhydrous ammonia using an
ion-exchange resin catalyst (wofatit kps-Hþ). This resin is a
more effective catalyst than water (10 g-ion/L of kps have
the same activity as 10 to 15 wt % water). They have derived
the kinetic model for this consecutive reaction at 323–352 K.
According to their study, the conversion of DEA to TEA has
a fractional order (0.25) in DEA, but all other reactions are
first order in each reactant. In addition, a DEA formation path
directly from ammonia and EO has been proposed.
4.2 Inorganic Solid Acid Catalyst. The process using the
cation-exchange resin has been commercialized. However, or-
ganic resin catalysts are not as stable in high-temperature sit-
uations as is desired. Therefore, a more stable and more selec-
tive catalyst is required. Johnson et al.38 have proposed inor-
ganic acid compounds, such as acidic silica–aluminas, zeolites,
molecular sieves, acid clays, or other acid metal oxides. The
preferred catalysts are the silica–aluminas, also known as alu-
minosilicates. These catalysts typically contain a proportion of
aluminum oxide. Grice and Knifton39 have proposed an acidic
clay obtained by treating montmorillonite with heteropolyacid
(e.g., 12-tungstophosphoric acid, 12-tungstosilicic acid, or 12-
molybdophosphoric acid), or with group 3 or 13 metal chlo-
rides (e.g., TiCl4, ZrCl4, or AlCl3). Vamling and Cider40 have
discussed the ability of different forms of zeolites to catalyze
the formation of MEA, DEA, and TEA from EO and ammonia
in the liquid phase, as well as a comparison of zeolite catalysts
with organic ion-exchange resins. They have predicted the
product distribution and the capacity of a full-scale adiabatic
packed-bed reactor for the different catalysts. They have
adapted a simple second-order consecutive reaction model
and built up a reactor math model for scale-up. Table 5 shows
the catalyst, NH3/EO ratio, product distribution, and kinetic
parameters ðꢀ; ꢁÞ. The ꢀ values in Table 5 are less than those
in Table 4. These inorganic catalysts can be expected to be
more stable in high temperatures, but their catalytic perform-
ances do not match that of the organic resin.
Decker et al.41–44 have reported the vapor-phase reaction of
EO and ammonia using various solid catalysts. Table 6 shows
their results with various catalysts. Catalytic activities of these
catalysts in vapor-phase reaction were quite low (EO conver-
sion: 20–30 mol %), and their MEA selectivity is extremely
low (ꢀ value: 30–300). In addition to these disadvantages, con-
siderable side reactions occur, and large amounts of byprod-
ucts (10–60 wt %) form.
4.3 Clay Catalyst Modified with Rare Earth Element.
4.3.1 Catalyst Screening: As mentioned above, past inorgan-
ic acid catalysts do not have enough performance for MEA se-
lective production. We started development of various cata-
lysts without limiting solid acid catalysts, such as heteropoly-
acid (Cs acid salt and multivalent metal salts), phosphoric acid
salts (ꢀ-zirconium phosphate, ꢂ-zirconium phosphate, and
boron phosphate), oxides or mixed oxides (silica–alumina,
titania–silica, titania–zirconia, and activated alumina), magic
acid (nafion and SO4/ZrO2), metal sulfates, solid base (magne-
sia and hydrotalcite), layered clays (montmorillonite and sap-
onite), and zeolite. Figure 2 shows the laboratory scale equip-
ment used in our experiments. When a heat-exchanger-type
reactor was used, unreacted EO concentration was measured
using an online gas chromatograph. On the other hand, when
an adiabatic reactor was used, unreacted EO was scarcely

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1079
Control valve
NH3
High pressure
metering pump
liquid NH3
Sampling
Oil bath
EO
Ethanolamines
Fig. 2. Laboratory scale experimental equipment. Reactor:
ꢃ
10.7 mm ꢁ 250 mmL, catalyst: 5–10 g, heat-exchange-
type reactor: bath temp: 340–400 K, adiabatic-type reac-
tor: reactor inlet temp: 320–350 K, reaction pressure:
12–14 MPa, NH3/EO molar ratio: 5–15, WHSV: 5–15
g gcatꢂ1 hꢂ1
.
detected (EO conversion of about 100%). The catalysts that
show significant high activity worthy to measure MEA selec-
tivity were silica-supported heteropoly iron salts, slica–alu-
mina, layered clay, and pentasil-type zeolite (e.g., ZSM-5,
ZSM-11). Table 7 shows the catalytic performances of these
catalysts. In particular, layered clays, such as montmorillonite
or saponite, showed high MEA selectivity comparable to ion-
exchange resin.
4.3.2 Control of Catalyst Acid–Base Properties: Acid
properties of these catalysts were controlled to improve their
activity and selectivity. Silica–alumina, which is treated with
a base, does not have a strong acidic site. Therefore, catalytic
activity was not expected in a reaction catalyzed by a strong
acid site. However, as shown in Table 7, silica–alumina im-
pregnated with NaOH (see Entry 2) had greater activity and
selectivity than non-treated silica–alumina (see Entry 1). This
means that an active site effective for the reaction may not
be always be a strong acid site. Then, acid catalysts were
modified with various metal elements (Cu, Ni, Zn, Mg, Ca,
Sr, Ba, Al, Fe, La, Y, Yb, and Ce) by ion-exchange methods,
or by impregnation, in order to control their acid–base proper-
ties. As shown in Entries 3, 6, 9, 10, and 12 of Table 7, the
activities of catalysts modified with rare earth elements were
improved as compared with others. Silica–yttrium catalyst
(see Entry 4), which is a weak-base catalyst, showed consider-
able activity.
In order to prepare a heat-resistant catalyst, it is necessary to
calcine the catalyst at a high temperature. When the inter-layer
cation in clay is exchanged by a multivalent metal cation, the
hydrated metal cation becomes a strong acid site.45 Therefore,
if cation-exchanged clay is treated in a temperature high
enough to lose water, a strong acid site, formed with hydrated
cation, is also lost. However, if an active catalyst does not need
a strong acid site, the cation-exchanged layered-clay catalyst
can be calcined at temperatures above 400 K. Table 8 shows
the effect of the calicination temperature of layered-clay modi-
fied with a rare earth element. The activity and selectivity im-
proved as the calcination temperature increased. Figure 3
shows the crystal structure of layered-clay at various calcina-
tion temperatures. XRD analysis showed that the layer struc-
ture was destroyed as the calcination temperature increased.

1080 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
Table 6. Catalytic Performance of Various Catalysts in Vapor-Phase Reaction41–44
NH3/EO
Reaction
temp/K
GHSV
/hꢂ1
EO
Product distribution/wt %
Catalysta)
ꢀ
ꢁ
molar ratio
conv/%
MEA
DEA
TEA
others
Active carbon41
41
Al2O3
6
8
10
5
6
6
463
363
463
333
363
333
400
200
200
300
400
400
25
10
47
22
30
30
24
25
42.1
38
36.4
9
15
10
7.2
18
8.1
55
34
5
46.3
17
39.5
23
27
60
4.4
27
16
13
88
149
37
36
38
195
243
342
73
266
29
Silica–alumina41
Zeolite 13X42
Faujasite zeolite43
Pentasil zeolite44
253
a) Catalyst 5 cm3.
Table 7. Activity and Selectivity of Various Catalysts in Liquid-Phase Reactiona)
Bath
temp
/K
EO
conversion
/mol %
WHSV
/g gcatꢂ1 hꢂ1
Entry
Ref.
Catalyst
ꢀ
ꢁ
ꢂ ¼ ꢀ=ꢁ
Cation-exchange
resin (Dowex-50W)
SiO2–Al2O3
SiO2–Al2O3–Na2O
SiO2–Al2O3–Y2O3
SiO2–Y2O3
14.3
361
97.7
5.5
4.9
1.1
1
2
3
4
5
6
7
8
9
6.3
6.2
380
381
373
389
384
359
384
391
360
360
360
353
ꢃ
92.2
96.0
98.0
97.6
91.2
97.0
92.0
88.9
98.1
97.6
94.5
93.9
9.2
6.8
7.6
7.0
7.7
8.6
5.4
5.7
5.0
4.3
5.8
5.3
14.0
8.8
12.0
8.2
6.5
2.4
4.2
5.4
3.7
2.8
4.3
4.6
0.7
0.8
0.6
0.9
1.2
3.6
1.3
1.1
1.4
1.5
1.3
1.2
10.0
11.5
10.4
12.1
10.1
14.5
10.0
11.2
9.9
H-ZSM-5
La-ZSM-5
H-montmorillonite
Na-montmorillonite
La-montmorillonite
Y-montmorillonite
H-saponite
10
11
12
La-saponite
10.6
a) Heat-exchanger-type reactor, inner diameter 10.7 mm , NH3/EO molar ratio 10 to 17, reaction pressure
12 to 14 MPa, oil bath temperature 360 to 390 K, WHSV 5 to 15 g gcatꢂ1 hꢂ1. ꢀ ¼ k2=k1: indicator of MEA
selectivity (the smaller the better). ꢂ ¼ k2=k3 ¼ ꢀ=ꢁ: indicator of DEA selectivity (the larger the better).
Table 8. Effect of Calcination Temperature of Clay Catalyst
on Activity and MEA Selectivitya)
Calcination
temperature
/K
Oil bath
temp
/K
EO
conv
/mol %
Entry
ꢀ
ꢁ
1
2
3
4
393
573
673
773
363
370
367
361
95.4
96.8
93.6
96.0
4.9
4.1
3.6
3.4
3.9
3.4
4.2
3.0
393K
573K
773K
a) Catalyst: Layered-clay modified with rare earth element.
NH3/EO molar ratio: 10 to 12, WHSV: ca. 11 g gcatꢂ1 hꢂ1
.
0
10
20
2θ/deg.
30
40
The layer structure of the catalyst calcined at 773 K collapsed
completely. This means that the active sites are not inter-layer
hydrated cations but are rare earth element cations fixed on
surface ion-exchange site.
Fig. 3. XRD pattern of layered-clay calcined on various
temperatures Cu Kꢀ.
4.3.3 Reactant Diffusion Effect on Activity and Selec-
tivity: A particle diameter effect on activity at replication
examination of ion-exchange resins was observed. Reactant
diffusion possibly affects the apparent catalytic activity and se-
lectivity in the case of fast reaction in a liquid phase, such as
the catalytic EO/NH3 reaction. In the case of layered-clay cat-
alyst modified with a rare earth element and calcined at 773 K,
the catalyst particle diameter effect on activity and selectivity
was investigated. Table 9 shows the EO conversion of cata-
lysts of various sizes. The EO conversion of relatively large-
diameter catalysts (see Entry 3) was quite low even if the re-
action temperature was 14 K higher than the standard condi-
tion. The activity of the Entry 3 catalyst was much smaller
than that of the Entry 1 catalyst. This means that a large-diam-
eter catalyst does not work effectively by limitation of the re-
actant diffusion rate, that is, the catalyst effectiveness factor is

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1081
Table 9. Effect of Particle Size of Catalyst Pellet on Appa-
rent Activitya)
Particle size
/mm
Oil bath
temperature/K
EO conv.
/mol %
ꢀ
ꢁ
Entry
1
2
3
0.149–0.297
0.65–0.35
0.71–1.0
363
372
387
95.4
92.8
87.6
4.3 4.0
4.6 4.3
5.8 5.4
a) Catalyst: Layered-clay modified with rare earth element
calcined at 773 K. NH3/EO molar ratio: 10 to 12, WHSV:
ca. 11 g gcatꢂ1 hꢂ1
.
Fig. 4. An example of molded layered clay catalyst.
Table 10. Impact of Pore Volume of Catalyst on Apparent Activitya)
Pore volume
/cm3 gcatꢂ1
Particle size
/mm
Oil bath
temperature/K
EO conv
/mol %
ꢀ
ꢁ
Entry
1b)
2b)
3b)
4c)
0.08
0.26
0.39
0.38
0.7–1.0
0.7–1.0
0.7–1.0
387
378
373
363
87.6
96.8
97.4
94.1
5.8
4.8
4.0
3.1
5.4
4.3
3.8
3.0
ꢃ
0.3–0.4 ꢁ 2–5L
a) Catalyst: Layered-clay modified with rare earth element calcined at 773 K. NH3/EO molar
ratio: 10 to 12, WHSV: ca. 11 g gcatꢂ1 hꢂ1. b) Pulverized type. c) Extruded type.
much smaller than unity.
pore volume increased. This means that a large pore-volume
catalyst has a large effective diffusion coefficient and a large
catalyst effective factor.
The catalyst effectiveness factor ꢄ is a function of Thiele
modulus46 ꢃ. In the case of spherical particles, this function
is expressed as
Similarly, the diffusion of the reactants and products affects
selectivity. In the case of the ion-exchange resin, ꢀ of 50–100
mesh-sized (0.075–0.150 mm) resin was about 5, while that of
20–50 mesh-sized (0.15–0.8 mm) resin was 6–7. In the case of
a layered-clay catalyst modified with a rare earth element, ꢀ of
a standard catalyst (about 0.2 mm in diameter) was about 3,
while that of a large particle size catalyst (about 2–3 mm)
was 6–7. This reaction is a typical consecutive reaction
(MEA ! DEA ! TEA). Therefore, in a region where effect
diffusion cannot be ignored, the second or third reaction pref-
erably occurs according to a longer residence time47 in the cat-
alyst pellet.
Figure 5 shows the MEA concentration in a reactant mix-
ture in the case of an ion-exchange resin and a layered-clay
catalyst modified with a rare earth element. Selectivity-
improved layer-clay catalysts required only about half the
NH3/EO ratio compared with ion-exchange resin.
ꢀ
ꢁ
3
1
1
ꢄ ¼
ꢂ
:
ð14Þ
ꢃ
tanhðꢃÞ
ꢃ
Thiele modulus ꢃ is defined as
sffiffiffiffiffiffiffiffi
sffiffiffiffiffiffiffiffiffiffiffi
k
k
ꢃ ¼ R
¼ R
ð15Þ
Deff
Dꢅ=ꢆ
k: Reaction rate constant, Deff: effective diffusion coefficient
(¼Dꢅ=ꢆ), D: diffusion coefficient, ꢅ: pore ratio (=pore-vol-
ume/catalyst particle volume), ꢆ: tortuosity factor, R: catalyst
particle radius. It is effective to enlarge the effective diffusion
coefficient in order to enlarge the catalyst effectiveness factor.
Increasing the pore ratio tends to decrease the tortuosity factor.
It is very effective to enlarge the pore ratio. A large pore-vol-
ume catalyst is prepared by the following procedure: (1) add a
pore-forming agent to a precursor and mix them well, (2) mold
the mixed material to a pellet, and (3) calcine the pellet at high
temperatures to remove a pore-forming agent by decomposi-
tion or by combustion.
These pore-forming agents include, for example, various
ammonium salts, such as ammonium nitrate and ammonium
acetate, organic compounds, such as urea, and various poly-
mers and fibers. Water-insoluble organic compounds can be
used in view of their efficiency of pore formation and ease of
molding. Figure 4 shows an example of a molded layered clay
catalyst. The pellet diameter (correspond to R in Eq. 15) was
very small in order to decrease diffusion distance. Table 10
lists the EO conversion of various pore-volume catalysts. The
apparent activities of the catalysts markedly increased as the
5. Heterogeneous Catalyst for DEA-Selective Production
Recently, the demand for DEA as an intermediate for gly-
phosate herbicides has markedly increased. Product distribu-
tion could be controlled by NH3/EO ratios as shown in
Fig. 1. DEA fraction was only 35 wt % at most. If a part of
the formed MEA is recycled to the reaction mixture, the net
DEA fraction may increase, but it has a limitation. Figure 6
shows the calculated net product distribution when a part of
the MEA was recycled to the reaction mixture when the
(NH3 + MEA) to EO molar ratio was 8. Even when equal
amounts of MEA to produced DEA was recycled, the DEA
fraction was only 43 wt %, and the TEA fraction reached
14 wt %. Therefore, a high DEA-selective catalyst is necessary

1082 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
100
90
80
70
60
Fig. 7. Pentasil-type zeolite frame work structure and mo-
lecular structure of ethanolamines.
10
20
30
NH3/EO molar ratio
Fig. 5. MEA concentration in produced amines as a func-
tion of molar ratio of ammonia to EO. : Layered-clay
Table 11. Ethanolamines Adsorption to Various Zeolites in
Aqueous Solution
ꢃ
modified with rare earth element 0.3–0.4 mm ꢁ 2–5
mmL, : Dowex-50W 20–50 mesh (0.3–0.8 mm).
Adsorbed MEA Adsorbed TEA MEA
Catalyst
/meq gꢂ1
5.18
/meq gꢂ1
4.95
/TEA
Ion-exchange resin
Dowex 50W
H-Y-zeolite
H-Mordenite
H-L-zeolite
80
1.05
MEA
2.33
2.13
2.35
0.60
2.08
0.89
1.24
0.10
1.12
2.39
1.90
6.00
60
H-ZSM-5
40
DEA
amount of MEA or TEA to ion-exchange resin, having no mi-
cropore, was nearly equal to its exchange capacity. No differ-
ence between MEA and TEA was observed. The ratio of the
adsorption amount of MEA and TEA was varied with the
zeolite micropore structure. The ratio in Y-zeolite that has a
relatively large pore was close to unity. The ratio in mordenite
or L-zeolite was markedly larger than unity. In the case of
ZSM-5, having a small pore diameter (0.55 nm) close to the
MEA molecular diameter (0.47 nm), the ratio of the adsorption
amount of MEA and TEA was very large (6). This means that
ZSM-5 can recognize the difference between MEA and TEA.
5.2 Implementation of Shape Selectivity. There are three
possible types of restriction in shape selectivity:48 (1) reactant
restriction, (2) product restriction, and (3) transition-state re-
striction. DEA-selective catalysts must have shape selectivity
by transition-state restriction. Figure 8 shows the reaction
model in zeolite micropores. An MEA molecule can react with
EO to form DEA in a narrow micropore, but DEA cannot react
with EO to form TEA by transition-state restriction. ZSM-5,
which can recognize the difference between MEA and TEA,
is expected to realize this shape selectivity. Reaction data of
H-ZSM-5 is shown in Table 7, Entry 5. Against expectations,
H-ZSM-5 showed comparatively low activity and a low DEA
selectivity indicator ꢂ (1.2). On the other hand, La-ZSM-5 (see
Entry 6), which was prepared by La ion exchange, showed
high activity and comparatively high ꢂ (3.6). Table 12 shows
the effect of lanthanide-ion-exchange amount to DEA selectiv-
ity. Along with lanthanide-ion-exchange amount, ꢁ decreased
significantly, that is, formation of TEA was depressed. In ad-
dition, ꢂ drastically increased up to 50. The reason why mere
20
TEA
0
0
20
40
60
80
100
Recycled MEA/Produced DEA /wt%
Fig. 6. Calculated net product distribution when MEA
recycled in conventional process. (NH3 + MEA)/EO
ratio = 8, ꢀ ¼ ꢁ ¼ 5.
so that formation of TEA is restrained and formation of DEA
is increased.
5.1 Molecular Recognition Ability of Zeolite. The molec-
ular diameter of a branched molecule (TEA) is 1 nm and that
of a straight molecule (MEA and DEA) is 0.47–0.49 nm. There
is a significant difference between the sizes of these molecules.
The concept of molecular sieves is the separation technology
utilization of this difference between sizes in molecular order.
The micro-reaction field in a zeolite micropore can be effective
for selective DEA formation if the zeolite is able to recognize
the difference between ethanolamines. Figure 7 shows the pen-
tasil-type zeolite framework structure and molecular structure
of ethanolamines. DEA, having small molecular diameter, is
able to enter the zeolite micropore, but TEA, having a branch-
ing structure, is not. Adsorption examinations of MEA and
TEA to various zeolites were made to confirm this expectation.
Table 11 shows the saturated adsorption amount of ethanol-
amines (MEA and TEA) to various zeolites. Adsorption

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1083
100
80
60
40
20
0
10
MEA can react with EO
8
B
←
EO
DEA
MEA
6
DEA
4
DEA cannot react with EO in the narrow zeolite micro pore
MEA
2
×
TEA
0
DEA
0
5
10
15
20
25
MEA conc. in Reaction mixture /wt%
Fig. 8. Reaction model in the zeolite micro pore.
100
80
60
40
20
0
10
8
Table 12. Impact of Lanthanide-Ion-Exchange Amount on
DEA Selectivitya)
A
MEA
Entry
La/Al
atomic ratio
Time on
stream/h
ꢀ
ꢁ
ꢂ
6
1
2
3
0.017
2
24
2
24
2
7.81
7.69
8.47
8.33
8.70
8.26
0.61
1.53
0.21
0.59
0.20
0.56
12.8
5.0
4
DEA
0.1
39.9
14.1
50.1
14.7
ꢃ
2
0.12
TEA
24
0
a) Catalyst: ZSM-5 Si/Al = 28. Reaction condition 10 mm
adiabatic reactor, NH3/EO molar ratio 15. Reaction pressure
12 MPa, WHSV 10 g gcatꢂ1 hꢂ1. Reactor inlet temperature
343 K, EO conversion about 100%.
5
10
15
20
NH3/EO molar ratio
Fig. 9. Initial catalytic performance. A: impact of NH3/EO
molar ratio, B: impact of MEA concentration in reaction
mixture (NH3/EO = 10). Catalyst La-ZSM-5, reaction
H-ZSM-5 does not show DEA selectivity is that a non-selec-
tive self-catalytic reaction by the hydroxy group of ethanol-
amines occurs significantly, because of its low activity, and
this cancels its shape selectivity.
Reaction rate constants k1, k2, and k3 in Eqs. 4, 5, and 6 are
considered as follows:
CO is the concentration of the hydroxy group defined as CM
2CD þ 3CT
k1 ¼ k1c þ k1nc þ kOCO;
pressure 10 MPa, WHSV 10 hꢂ1
.
Figure 9B shows the net product distribution of ethanol-
amines when a part of produced MEA is recycled to the reac-
tion mixture. When the MEA concentration in the reaction
mixture was 10 wt % (recycled MEA/produced DEA ratio:
70%), the DEA fraction reached more than 60 wt %, and the
TEA fraction was only 0.9 wt %. The DEA fraction of Fig. 9B
by comparison with that of Fig. 6 shows that the MEA recycle
effect of the DEA-selective process is much larger than that of
a conventional process.
A multivalent metal cation, such as lanthanide, is strongly
hydrated in aqueous solution, and its hydrated ion diameter
is large. It is difficult to exchange a small cation in the narrow
micropore of ZSM-5 with a large cation. Therefore, the ion-
exchange amount of a catalyst that is prepared by ion-ex-
change operation in an aqueous solution was small (Entry 1
in Table 5). Jia et al.49 and Rauscher et al.50 have suggested
that the ion-exchange amount of zeolite can be improved by
a solid-state process. It seems that a bare cation is relatively
small and can diffuse easily into the micropore of the zeolite
at high temperatures, above 700 K. The following experiment
was executed to reveal that ion exchange can occur in solid
state. First, a mechanical mixture of LaCl3 and NH4-ZSM-5
(Si/Al ratio: 28) was carefully ground in a mortar for several
minutes. The La/Al atomic ratio in the mechanical mixture
was 0.114. The ground mixture was then transferred to a por-
þ
ð16Þ
ð17Þ
ð18Þ
k2 ¼ k1ꢀ ¼ k1cꢀc þ k1ncꢀnc þ ꢀOkOCO;
k3 ¼ k1ꢁ ¼ k1cꢁc þ k1ncꢁnc þ ꢁOkOCO;
ꢀ ¼ ðk1cꢀc þ k1ncꢀnc þ ꢀOkOCOÞ=ðk1c þ k1nc þ kOCOÞ; ð19Þ
ꢁ ¼ ðk1cꢁc þ k1ncꢁnc þ ꢁOkOCOÞ=ðk1c þ k1nc þ kOCOÞ; ð20Þ
ꢂ ¼ ðk1cꢀc þ k1ncꢀnc þ ꢀOkOCOÞ
=ðk1cꢁc þ k1ncꢁnc þ ꢁOkOCOÞ:
ð21Þ
The subscripts c, nc, and O of each parameter denote intrinsic
catalytic, non-catalytic, and self-catalytic (OH), respectively.
If k1c (reaction rate constant of a catalyst) is much smaller than
kO (reaction rate constant of self-catalytic reaction), then the
apparent ꢂ value is far from the true value. Figure 9A shows
tangible product distribution (the right axial scale is 10 times
that of the left). The DEA fraction at the same NH3/EO molar
ratio was larger than that of Fig. 1 (aqueous NH3 method) and,
at the same time, the TEA fraction was extremely small. This
result reveals that this catalyst has shape selectivity.

1084 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
Table 13. Content of Impurities Formed by Binder-Contain-
ing Zeolitea)
2.5
La-O-La
2
Entry Catalystb)
Content/wt %
La-O
MEA DEA TEA Impurity 1 Impurity 2
1.5
1
1
2
3
A
B
C
67.8 27.1 2.0
70.6 28.8 0.4
69.3 30.2 0.5
1.8
—
—
0.3
—
—
a) NH3/EO = 15, WHSV 10 g gcatꢂ1 hꢂ1
,
reactor inlet
0.5
0
temperature 343 K, reaction pressure 12 MPa. b) Catalyst A:
La-ZSM-5 (commercially available) Al2O3 binder 30 wt %,
B: La-ZSM-5 (commercially available) without binder, C:
Binderless zeolite (NSC La-ZSM-5).
0
0.1
0.2
0.3
0.4
Distance /nm
Fig. 10. La K-edge FT-EXAFS spectra (SPring 8 BL19B2).
: La2O3, solid line: La-ZSM-5.
HO
H
N
1
OH
HO
O
N
OH
HO
O
2
Fig. 11. Impurities formed by binder-containing zeolite.
celain dish and calcined in a muffle furnace at 823 K for 5 h.
The resulting sample was washed four times with deionized
water. It was finally dried overnight at 393 K. The La/Al atom-
ic ratio in the final sample, determined by ICP analysis, was
0.11. The catalysts in Table 12 were prepared by a solid-state
ion-exchange procedure adopting this concept.51
Figure 10 shows La K-edge FT-EXAFS spectra of La-ZSM-
5 compared with La2O3. The peak assigned as La–O–La exist-
ed in La oxide spectra, but this peak was not observed in
La-ZSM5 spectra. This suggests that La ion in ZSM-5 is iso-
lated and is fixed to an ion-exchange site without aggregation
to form oxide.52
Fig. 12. Some examples of molded binderless zeolites.
To prevent the formation of impurities, a molded zeolite
catalyst without a binder is required. Matsukata et al. have pro-
posed a new zeolite synthetic method called ‘‘dry gel conver-
sion.’’56,57 Precursor gel containing a silica source, alkali, and a
structure-directing agent (SDA) was contacted with high-tem-
perature steam in an autoclave in order to crystallize to zeolite.
This method was helpful to formulate the preparation of mold-
ed binderless zeolite. Shimizu et al. have also proposed solid
state zeolite transformation method.58,59
5.3 Development of Molded Binderless Zeolite. General-
ly, a high silica zeolite, such as pentasil-type zeolite, has poor
plasticity and formability. Therefore, it is very difficult to mold
without an inorganic binder. A commercially available molded
zeolite usually contains 20 to 40 wt % binder, such as alumina.
Impurities (1 and 2 in Fig. 11) were formed using a catalyst
made of a binder-containing zeolite. Entry 1 of Table 13
shows an example of impurities contents, and Entry 2 shows
that no impurities are formed by the zeolite without binder.
It seems that the impurities are formed on the basic-site of
binder. Boiling points of these impurities were high and close
to the boiling point of TEA. The temperature of the distillation
column must be low to prevent coloring. The distillation col-
umn must be operated in a very high vacuum, such as 100
to 300 Pa, and its pressure drop must be very small. To reduce
its pressure drop, the (packed) height of the column must be
low. It is difficult to separate them from TEA using such a
low-distillation column because of the separation ability short-
age. The formation of these glycol ether-type impurities has
been encountered in conventional aqueous ammonia methods.
Washington et al.53,54 have proposed adding carbon dioxide to
the reaction mixture in order to reduce these impurities. Add-
ing carbon dioxide is not preferred in an anhydrous system.55
We prepared molded binderless zeolite as follows.60,61
Using the new solid-state crystallization (NSC) method, first,
aluminum source, alkali, and SDA were impregnated into
the molded silica support. Then, this precursor was dried
and transferred into an autoclave. Finally, high-pressure and
high-temperature steam was fed to the autoclave, and the pre-
cursor was crystallized to form molded zeolite. The obtained
molded zeolite was robust enough for practical use.
As shown in Fig. 12, when molded silica of various shapes
was used as a raw material, its whole body crystallized, main-
taining its original shape. Figure 13 shows the XRD pattern
of NSC ZSM-5. Sharp and strong diffraction lines, which
were assigned to ZSM-5, showed good crystallinity. Figure 14
shows 29Si and 27Al MAS NMR spectra of NSC ZSM-5 and
of commercially available ZSM-5. These spectrums show that
NSC zeolite had few aluminum atoms out of lattice and Q3-
type silicon atoms. Figure 15 shows the morphology of NSC
zeolite observed by FE-SEM. NSC ZSM-5 had sub-microm-
eter-sized minute crystals and good crystallinity. Figure 16
shows the morphology of molded binderless zeolite. For the
same reason as discussed above about the clay catalyst, the
catalyst diameter must be small. After crystallization, zeolite

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1085
beads were treated with ammonium nitrate to exchange so-
dium ion, and then they were calcined to decompose ammoni-
um ion and SDA. Finally, proton of the zeolite was exchanged
with La ion by a solid-state ion-exchange procedure. The cat-
alytic performance of the NSC La-ZSM-5 catalyst was shown
in Entry 3 of Table 13. Impurities were not detected. The NSC
method can be applied to the synthesis of a zeolite other than
ZSM-5, using various SDAs. Table 14 shows various crystal-
lization products using various precursors and SDAs. In partic-
ular, ꢁ zeolite can be crystallized easily. Scale-up of the NSC
method was comparatively easy, and large-scale industrializa-
tion of zeolite synthesis was accomplished.
6. Process Development Using a Heterogeneous Catalyst
6.1 Adiabatic Reactor with Partially Vaporized Ammo-
nia. The developed catalysts were calcined at temperatures
above 800 K. Therefore, the catalysts have higher heat resist-
ance than that of ion-exchange resin (allowable temperature
limit: 393 K). Selectivities of both catalysts (clay and zeolite)
were improved; thus, the reaction can be executed at high EO
concentrations (i.e., low NH3/EO molar ratio). In this case, the
adiabatic temperature rise is a severe problem. The heat ca-
pacity (Cp) of the reaction mixture estimated using a simple
mixing rule disagrees with an observed value (see Table 15).
10
20
30
40
2θ /deg
Fig. 13. XRD pattern of NSC zeolite (ZSM-5). X-ray: Cu Kꢀ.
Fig. 14. 29Si and 27Al MAS NMR spectra of NSC zeolite (ZSM-5).
Fig. 15. Morphology of NSC zeolite (ZSM-5) crystal (FE-
SEM).
Fig. 16. An example of a molded binderless zeolite (ZSM-5).

1086 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
Table 14. Various Precursors and Crystallization Products of NSC Zeolite Method
Si/Na
atomic ratio
Si/SDA
molar ratio
Si/Al
atomic ratio
SDA
Condition
Product
NPr4OH
NPr4OH
NBu4OH
MeNEt3OH
NEt4OH
30
8
16
28
23
20
30
20
10
10
5
—
16
24
36
30
—
453 K, 8 h
453 K, 8 h
443 K, 32 h
453 K, 48 h
443 K, 20 h
443 K, 144 h
Silicate-1
ZSM-5
ZSM-11
ZSM-12
ꢁ
NMe4OH
5
NU-1
Table 15. Effect of EO Concentration on Heat Capacitya)
EO
Heat capacity/J gꢂ1 degꢂ1
Calculated value
by mixing rule
NH3/EO
molar ratio
Reactor inlet–outlet
temperature/K
concentration
/mol %
Observed
NH3
30.0
22.0
17.5
14.5
0
373.2
—
6.11
5.86
5.78
5.69
5.61
3.24
4.36
5.40
6.44
350.2–385.1
345.1–393.1
330.1–395.4
318.7–396.1
4.82
4.15
3.77
3.68
a) Reaction pressure: 12 MPa, heat capacity: average value of reactor inlet and outlet.
Control valve
(5)
Cooler
CW
8
(4)
EO
7
6
(1)
(2)
NH3
MEA
5
4
(6)
Produced
ethanol amines
3
1
3
(3)
5
EO concentration
/mol%
Recycle NH3
7
Temperature /K
ꢃ
Fig. 17. Model flow sheet of a pilot plant. Reactor: 67 mm
2000 mmL, catalyst 4 to 6 L, with thermal insulation mate-
rial wounded up. Heat liberation is compensated by heat-
ing.
Fig. 18. Heat capacity of reaction mixture as a function of
EO concentration and temperature, pressure: 12 MPa.
effects of EO concentration.
Vamling et al.40 have reported an experimental equation of Cp
(22) as a function of pressure (P) and temperature (T).
Figure 18 shows the heat capacity calculated following
Eq. 23 at 12 MPa pressure. The heat capacity decreased mark-
edly as the EO concentration increased. Figure 19 shows the
adiabatic reaction temperature rise as a function of NH3/EO
molar ratio calculated using Eq. 23. The maximum tempera-
ture of the reactor was limited to under ꢅ420 K by-product
qualities, such as coloring.
There are several means of suppressing temperature rise:
(1) lower EO concentration in reaction mixture of reactor inlet,
(2) use multiple reactors and split EO feed, and (3) use an heat-
exchange-type reactor. These means have the strong disad-
vantage of increasing equipment cost.
We developed a process that can be operated in high EO
concentrations and the temperature increase in the reaction
mixture can be controlled.62 A part of the ammonia is vapor-
ized, and the reaction heat in an adiabatic reactor is eliminated
by the heat of the ammonia vaporization. The reaction pressure
Cp ¼ Cp;a þ Cp;b=jðCp;cð13 MPaÞ þ dCp;cdPðP ꢂ P0Þ ꢂ Tj; ð22Þ
P0: 13 106 Pa; Cp;a, Cp;b, Cp;cð13 MPaÞ, dCp;cdP: parameter. This
ꢃ
equation does not involve an EO concentration term, but the
EO concentration affects Cp considerably. The relationship be-
tween temperature rise and EO concentration was observed by
using the pilot plant shown in Fig. 17. The reactor of this plant
was highly insulated, and heat loss was compensated by heat-
ing to realize nearly adiabatic conditions. Heat capacity Cp is
expressed as a function of EO concentration (CE), pressure,
and temperature (see Eq. 23):
ð13 MPaÞ
Cp ¼ Cp;a þ a expðc CEÞ þ Cp;b=jðCp;c
ꢃ
a, b, and c are additional parameters to Eq. 22 to explain the
ꢃ
þ b CE þ dCp;cdPðP ꢂ P0Þ ꢂ Tj;
ð23Þ
ꢃ

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1087
160
140
120
100
80
100
10
1
γ
α
60
40
β
20
0.1
0
0
10
20
30
40
50
0
0.2
0.4
0.6
0.8
1
NH3/EO molar ratio
Relative catalytic activity
Fig. 19. Adiabatic reaction temperature rise as a function of
NH3/EO ratio.
Fig. 21. Apparent selectivity change model along with
deterioration. Plot of equation 19(ꢀ), 20(ꢁ), 21(ꢂ) as a
faction of relative k1c calculated when k1c=kOCO ¼ 50,
ꢀc ¼ 8, ꢁc ¼ 0:03.
450
400
350
100
Maximam temperature at
adiabatic condition
L5
Catalyst bed
L4
B:16MPa
temperature
Liquid phase condition
L3
L2
t5
t4
L1
L0
t3
Time on t
stream
2
∆T
300
450
0
t1
100
Maximam temperature at
adiabatic condition
t0
←
inlet
catalyst bed
outlet
A:10MPa
Fig. 22. Change of temperature profile along with catalyst
deterioration.
400
350
300
NH3 Vaporized condition
conversion (e.g., 440 K at an inlet temperature of 323 K). In
the case of A, at a reaction pressure of 10 MPa, vaporization
of ammonia began when the temperature reached the boiling
point of the reaction mixture, and temperature increased to
about 410 K (30 K lower than A) very slowly.
0
6.2 Deterioration and Regeneration of Zeolite Catalyst.
The data of the elapsed time of 24 h in Table 12 showed cata-
lyst deterioration. Conversions of EO stayed at almost 100%,
but DEA selectivity decreased. It became clear, by examining
temperature distribution in the catalyst bed in detail, that there
is activity deterioration from an entrance part. Decrease in
activity may be caused by occlusion of micropores as a result
of a zeolite catalyst. The reason why selectivity decreases
along with activity deterioration is that non-catalytic reaction
without selectivity becomes a dominant position in the activi-
ty-deteriorated catalyst bed. When intrinsic catalytic reaction
rate constant k1c in Eq. 21 becomes small, ꢂ (DEA-selectivity
indicator) becomes small (see Fig. 21). The same phenomenon
occurs in the catalyst-unpacked part of the reactor63 in the case
of MEA recycle.
0
0.2
0.4
0.6
0.8
1
Relative Catalyst Bed Length
Fig. 20. Model temperature and EO conversion profile in
catalyst bed. NH3/EO molar ratio: 12, reactor inlet tem-
perature: 323 K, reaction pressure: A: 10 MPa, B: 16 MPa,
solid line: temperature profile (left axis), dotted line: EO
conversion profile (right axis).
is controlled to reduce the boiling point of reaction mixture.
Furthermore, even if the reaction heat rises as the reaction pro-
ceeds, ammonia vaporization by the reaction heat and the tem-
perature rise of the catalyst bed is restrained.
Figure 20 shows the calculated profile of temperature and of
EO conversion in the catalyst bed, at an NH3/EO ratio of 12.
In the case of B in Fig. 20, ammonia did not vaporize at the
reaction pressure of 16 MPa, and the catalyst bed temperature
increased with generated reaction heat, proportional to EO
Development of a method to regenerate deactivated catalyst
for practical use in a process using zeolite is extremely impor-
tant. Figure 22 shows a change of model temperature profile in

1088 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
an adiabatic reactor along with catalyst deterioration. Initially
(t0), the temperature increased at the entrance part of the reac-
tor, because the catalyst had not yet deteriorated. The catalyst
deteriorated from the entrance part as the reaction progresses
(t0 ! t5), and the temperature of the deteriorated part of cata-
lyst bed did not increase. Then, the peak of the temperature
profile moved to the latter part (L0 ! L5) of the catalyst
bed. It is necessary to regenerate a catalyst, because EO con-
version and DEA selectivity decrease within a certain time.
The deactivated catalyst did not discolor much and had the
organic substance deposited in it, but the phenomenon did
not involve coking, which is often the cause of catalyst deac-
tivation. When coking occurs, the catalyst becomes black, and
the organic substance becomes a carbon-like deposit. Such a
substance can be removed only by oxidation at a temperature
above 600 K. However, the deactivation of the catalyst in this
reaction is not caused by coking; therefore, the catalyst can be
regenerated by passing it through the catalyst rinse medium at
a high temperature, which is more than the reaction tem-
perature, to decompose or extract the deposited organic sub-
stance.64 The selection of the medium is very important. When
an inert gas, such as nitrogen, is used at temperatures higher
than the reaction temperature, the organic substance becomes
a carbon-like deposit. Solvent, supercritical fluid and high-
density gas can be used. It is preferable not to bring in a
new substance that is not used in the original process. Ammo-
nia was used as a rinse medium was examined, because it is a
stable substance used in the process. Ammonia reached its gas
phase at a temperature higher than the reaction temperature.
However, when it is pressurized at a high pressure, like the
reaction pressure, its density is considerably larger, so it is
an effective medium to regenerate the catalyst.65
used in regeneration, and this recovered ammonia served as the
raw material for the reaction. Plural reactors are necessary for
concurrent reaction and regeneration. After regeneration, the
flow channel is changed to the reaction system, and the reac-
tion is restarted.
The developed process is made up from the following
unique technologies. (1) zeolite catalyst, which is able to rec-
ognize the difference between DEA and TEA; (2) catalyst
modification with rare earth element to improve activity dras-
tically; (3) binderless molded zeolite; (4) catalyst regeneration
using ammonia; and (5) reaction heat elimination using latent
heat of vaporization in an adiabatic reactor.
The authors thank Mr. Shindo, Mr. Morishita, Mr. Arita,
Mr. Takeda, Mr. Hashimoto, and Mr. Yokonuki for their con-
tinued support in catalyst and process development. The au-
thors also thank the staff of the technical development and
manufacture departments in the Chidori plant for their support
in the industrialization of this process, and the staff of the
catalyst manufacture department in the Himeji plant for the
development and production of the commercial catalyst.
Figure 23 shows the change in transition of the temperature
peak in the catalyst bed along with the reaction and regenera-
tion cycles. The catalyst activity was recovered from the dete-
rioration in spite of repeating reaction and regeneration cycles.
Figure 24 shows a model process flow sheet with a regenera-
tion step. The reaction and regeneration steps were switched
at intervals of several days. High boiling-point substances
weeping from the catalyst were eliminated from the ammonia
8
7
6
5
4
Number of
3
2
reaction
1
Time on stream
Fig. 23. Transition of temperature peak position with cata-
lyst regeneration treated with ammonia at high tempera-
ture under high pressures.
EO
Recovered NH3
NH3
Purification
system
NH3
Recovered NH3
MEA
Fig. 24. Model process flow sheet. Reaction and NH3 recovery system.

H. Tsuneki et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007) 1089
29 H. Hammer, W. Reutemann, Eur. Pat. Appl. EP 673,920,
1995; U.S. Patent 5,545,757, 1996.
References
1 A. Wurtz, Justus Liebigs Ann. Chem. 1860, 114, 51; 1862,
122, 226.
30 S. B Willis, Jr., J. D. Henry, U.S. Patent 4,355,181,
1982.
2
3
L. Knorr, Ber. Dtsch. Chem. Ges. 1899, 32, 729.
M. R. Edens, J. F. Lochary, Kirk-Othmer Encyclopedia of
31 M. Ahmed, J. R. Nelson, C. A. Gibson, Eur. Pat. Appl. EP
99,416, 1984; U.S. Patent 4,845,296, 1989.
Chemical Technology, John Wiley & Sons, New York, 1991,
Vol. 2, pp. 8–9.
32 C. A. Gibson, M. Ahmed, J. R. Nelson, Eur. Pat. Appl. EP
100373, 1984; U.S. Patent 4,847,418, 1989.
4
3152.
5
G. Natta, E. Mantian, J. Am. Chem. Soc. 1952, 74,
P. Ferrero, F. Berbe, L. Flamme, Bull. Soc. Chim. Belg.
33 A. F. A. Reynhart, U.S. Patent 2,186,392, 1938.
34 B. J. G. Weibull, Swedish Patent 158,167, 1957.
35 B. J. G. Weibull, L. U. F. Folke, S. O. Lindstrom, U.S.
Patent 3697598, 1972.
1947, 56, 349.
6
25, 405.
C. Potter, R. R. McLaughln, Can. J. Res., Sect. B 1947,
36 D. Kamemski, R. Boeva, N. Stoyanov, I. Taseva, God.
Vissh. Khim-Tekhnol. Inst., Burgas 1984, 19, 53.
37 A. Stoyanov, R. Boeva, S. Kotov, Zh. Prikl. Khim
(Leningrad) 1983, 56, 1966.
7
8
A. Lowe, D. Buttler, E. Meade, Brit. 1956, 760, 215.
M. Miki, T. Ito, M. Hatta, T. Okabe, Yukagaku 1966, 15,
215.
9
38 L. F. Johnson, Jr., L. Fred, U.S. Patent 4,438,281, 1984.
39 J. N. Grice, J. F. Knifton, U.S. Patent 4,939,301, 1990.
40 L. Vamling, L. Cider, Ind. Eng. Chem. Prod. Res. Dev.
1986, 25, 424.
41 U. Decker, J. Glietsch, I. Koch, K. Bergk, DD Patent,
298,633, 1992; Chem. Abstr. 1992, 117, 5932.
42 U. Decker, J. Glietsch, I. Koch, K. Bergk, K. Heinz, I.
Paleschke, W. Schwieger, DD Patent, 298,634, 1992; Chem.
Abstr. 1992, 117, 497.
J. Przondo, Z. Kossonski, I. J. Onkowska, Prizem. Chem.
1974, 53, 757; Y. S. Ahn, T. Y. Lee, T. K. Chang, C. Chung, S.
Ihm, J. KIChE 1980, 18, 393.
10 K. Park, T. Park, K. Byung, Hwahak Konghak 1985, 23,
409.
11 G. Schulz, W. Goetze, Chem. Ing. Technol. 1968, 40,
446.
12 R. Dahlinger, W. Goetze, G. Schulz, U. Sonksen, DE-OS
¨
1972, 1768335.
13 W. Goetze, P. Wolf, G. Schulz, H. Luedemann, U.S. Patent
3723530, 1973.
43 U. Decker, J. Glietsch, I. Koch, K. Bergk, K. Heinz, I.
Paleschke, W. Schwieger, DD Patent, 298,635, 1992; Chem.
Abstr. 1992, 117, 496.
14 M. J. Green, Eur. Pat. Appl. EP 171,893, 1986.
15 Y. Watanabe, H. Sugawara, K. Suzuki, M. Matsumura,
Jpn. Kokai Tokkyo Koho 94 199,746, 1994; Chem. Abstr. 1994,
121, 300467.
16 Y. Watanabe, K. Suzuki, H. Sugawara, Jpn. Kokai
Tokkyo Koho 94 247,909, 1994; Chem. Abstr. 1994, 122,
30954.
17 Y. Watanabe, K. Suzuki, H. Sugawara, Jpn. Kokai
Tokkyo Koho 94 247,910, 1994; Chem. Abstr. 1994, 122,
30955.
44 U. Decker, J. Glietsch, I. Koch, K. Bergk, K. Heinz, I.
Paleschke, W. Schwieger, DD Patent, 298,636, 1992; Chem.
Abstr. 1992, 117, 495.
45 Y. Izumi, K. Urabe, A. Onaka, Shokubai 1985, 27, 27.
46 E. W. Thiele, Ind. Eng. Chem. 1939, 31, 916.
47 A. Wheler, Adv. Catal. 1991, 3, 249.
48 S. M. Csicsery, Pure Appl. Chem. 1986, 58, 841.
49 C. Jia, P. Beaunier, P. Massiani, Microporous Mater.
1998, 24, 69.
50 M. Rauscher, K. Kesore, R. Monnig, W. Schwieger, A.
¨
18 Y. Watanabe, M. Matsumura, H. Sugawara, T. Okawa, K.
Suzuki, Jpn. Kokai Tokkyo Koho 95 48,327, 1995; Chem. Abstr.
1995, 122, 290331.
19 Y. Watanabe, M. Matsumura, H. Sugawara, T. Okawa, K.
Suzuki, Jpn. Kokai Tokkyo Koho 95 70,004, 1995; Chem. Abstr.
1995, 122, 290326.
20 Y. Watanabe, M. Matsumura, T. Okawa, K. Suzuki, Jpn.
Kokai Tokkyo Koho 95 70,005, 1995; Chem. Abstr. 1995, 123,
9065.
21 A. Moriya, H. T. Suneki, Jpn. Kokai Tokkyo Koho 95
126,228, 1995; Chem. Abstr. 1995, 123, 9064.
22 A. Moriya, H. Tsuneki, Jpn. Kokai Tokkyo Koho 95
173,114, 1995; Chem. Abstr. 1995, 123, 9064.
23 A. Moriya, H. Tsuneki, Shokubai 1996, 38, 91.
24 H. Tsuneki, Shokubai 1998, 40, 304; Chem. Abstr. 1998,
129, 244844.
25 H. Tsuneki, A. Moriya, H. Baba, Jpn. Kokai Tokkyo Koho
99 122,681, 1999; Chem. Abstr. 1999, 131, 201526.
26 H. Tsuneki, M. Kirishiki, T. Oku, H. Shindo, Y.
Hashimoto. H. Yokonuki, Shokubai 2004, 43, 84.
27 H. Tsuneki, Recent Development of Zeolite Catalysts, ed.
by T. Tatsumi, Y. Nishimura, CMC Publishing, Tokyo, 2004,
p. 173.
Tißler, T. Turek, Appl. Catal., A 1999, 184, 249.
51 M. Kirishiki, H. Tsuneki, F. Morishita, M. Urano, Jpn.
Kokai Tokkyo Koho 2002 028,492, 2002; Chem. Abstr. 2002,
136, 120219.
52 N. Saito, SPring8 Report, 2004, A0788-RI-TU BL19B2.
53 S. J. Washington, T. F. Grant, U.S. Patent 5,334,763,
1994.
54 S. J. Washington, T. F. Grant, WO 9,424,089, 1994.
55 D. M. Nygaard, R. M. Diguilio, U.S. Patent, 6,063,965,
2000.
56 P. R. H. P. Rao, M. Matsukata, Chem. Commun. 1996,
1441.
57 M. Matsukata, N. Nishiyama, K. Ueyama, Microporous
Mater. 1996, 7, 109.
58 S. Shimizu, Y. Kiyozumi, F. Mizukami, R. M. Mihayli,
H. K. Beyer, Adv. Mater. 1996, 8, 759.
59 S. Shimizu, Y. Kiyozumi, F. Mizukami, Chem. Lett. 1996,
403.
60 T. Oku, H. Tsuneki, 86th CATSJ Meeting, 2000, Abstr.,
No. 4E19.
61 T. Oku, H. Tsuneki, Jpn. Kokai Tokkyo Koho 2001
58,817, 2001.
62 H. Tsuneki, A. Moriya, Jpn. Kokai Tokkyo Koho 97
52,869, 1997; Chem. Abstr. 1997, 126, 263841.
28 H. Tsuneki, Shokubai 2005, 47, 196.

1090 Bull. Chem. Soc. Jpn. Vol. 80, No. 6 (2007)
AWARD ACCOUNTS
63 H. Tsuneki, Jpn. Kokai Tokkyo Koho 2000 288,378, 2000;
Chem. Abstr. 2000, 133, 298003.
64 H. Tsuneki, M. Kirishiki, Y. Arita, Y. Hashimoto, T. Oku,
H. Shindo, M. Urano, Jpn. Kokai Tokkyo Koho 2001 151,740,
2001; Chem. Abstr. 2001, 135, 5375.
65 H. Tsuneki, T. Oku, M. Urano, F. Morishita, Jpn. Kokai
Tokkyo Koho 2001 314,771, 2001; Chem. Abstr. 2001, 135,
349268.
Hideaki Tsuneki was born in 1953 in Fukushima, Japan. After graduating Tokyo Institute of
Technology in 1976, he entered the graduate school of the same university. After he received
a master’s degree from the Tokyo Institute of Technology in 1978, he joined Nippon Shokubai
Chemical Industry. He was engaged in studies on methacrylic acid synthesis catalyst in 1978
to 1985, in development of ethyleneimine production process and catalyst from 1985 to 1990,
and development of 2-aminoethanol and 2,20-iminodiethanol production process and catalyst
from 1991 to 2003. In 1992 he received the 41st Chemical Society of Japan Award for Technical
development, and the Catalyst Society of Japan Award in Technical Division. Now, he is a Senior
Technical Adviser of R&D division and a General Manager of Strategic Technology Research
Center of Nippon Shokubai Co., Ltd.
Masaru Kirishiki was born in 1963 in Ishikawa, Japan. After graduating Toyama University in
1986, he entered the graduate school of the same university. After he received a master’s degree
from Toyama University in 1988, he joined Nippon Shokubai Chemical Industry. He was engaged
in studies on acrylic acid synthesis catalyst from 1988 to 1990, in development of secondary
alcohol ethoxide production process and catalyst from 1990 to 1999 and development of dimethyl
carbonate production process and catalyst in 1992 to 1994, and development of 2,20-iminodi-
ethanol production process and catalyst from 1999 to 2003. Now he is a Research Manager of
Strategic Technology Research Center of Nippon Shokubai Co., Ltd.
Tomoharu Oku was born in 1966 in Kitakyushu, Japan. After graduating Tokyo University of
Science in 1989, he entered the Interdisciplinary Graduate School of Science and Engineering
at the Tokyo Institute of Technology (Titech). After he received a master’s degree from Titech,
he joined Nippon Shokubai Co., Ltd. in 1992. He was engaged in studies on metal catalysts for
chemical synthesis from 1992 to 1994, and various solid acid–base catalyses in supercritical fluids
from 1994 to 1997, and solid-phase synthesis of zeolite in 1998, and development of the 2,20-
iminodiethanol production process and catalyst from 1999 to 2000. In addition he joined on
temporary to the NEDO’s project of ‘‘Development of Technology to Reduce the Burden on
the Environment Using Supercritical Fluid’’ in Titech from 2000 to 2002. Now he is a Research
Manager of Strategic Technology Research Center of Nippon Shokubai Co., Ltd.