Degradation reaction of monoethanolamine using TS-1 zeolite as a photocatalyst

Article abstract of DOI:10.1039/a906585g

Photocatalytic properties of TS-1 zeolite were examined via a degradation reaction of monoethanolamine (MEA). When TS-1 was mixed with MEA aqueous solution, the MEA molecules were bonded to the Ti atoms in TS-1 since the latter were coordinatively unsaturated. In fact, the Ti atoms in TS-1 were isolated and coordinated by four oxygen atoms. The change in the coordinative situation around the Ti atom shifted the absorption of ultraviolet (UV) light from about 220 nm, which is intrinsic absorption of TS-1, to a longer wavelength, i.e. around 300 nm. When the suspension mixture of TS-1 and MEA aqueous solution was irradiated by UV light with a wavelength of around 300 nm, photo-excitation of the MEA-adsorbed Ti species took place leading to a degradation of MEA molecules. Although a photocatalytic activity per weight of a specimen was lower for TS-1 than for TiO₂, the activity per Ti atom in TS-1 was higher than that in TiO₂. The strong adsorption of MEA to TS-1 and its microporous structure are attributed to the high activity of the photocatalytic reaction. The use of either TS-1 or TiO₂ as a photocatalyst had no effect on the kind of products and the product ratios.

Full text of DOI:10.1039/a906585g

View Article Online / Journal Homepage / Table of Contents for this issue
Degradation reaction of monoethanolamine using TS-1 zeolite as a
photocatalyst
Takayuki Ban, Satoshi Kondoh, Yutaka Ohya and Yasutaka Takahashi
Department of Chemistry, Gifu University, 1-1 Y anagido, Gifu 501-1193, Japan
Received 13th August 1999, Accepted 26th October 1999
Photocatalytic properties of TS-1 zeolite were examined via a degradation reaction of monoethanolamine
(
MEA). When TS-1 was mixed with MEA aqueous solution, the MEA molecules were bonded to the Ti atoms
in TS-1 since the latter were coordinatively unsaturated. In fact, the Ti atoms in TS-1 were isolated and
coordinated by four oxygen atoms. The change in the coordinative situation around the Ti atom shifted the
absorption of ultraviolet (UV) light from about 220 nm, which is intrinsic absorption of TS-1, to a longer
wavelength, i.e. around 300 nm. When the suspension mixture of TS-1 and MEA aqueous solution was
irradiated by UV light with a wavelength of around 300 nm, photo-excitation of the MEA-adsorbed Ti species
took place leading to a degradation of MEA molecules. Although a photocatalytic activity per weight of a
specimen was lower for TS-1 than for TiO , the activity per Ti atom in TS-1 was higher than that in TiO .
2
2
The strong adsorption of MEA to TS-1 and its microporous structure are attributed to the high activity of the
photocatalytic reaction. The use of either TS-1 or TiO as a photocatalyst had no e†ect on the kind of
2
products and the product ratios.
zeolite to give 6-fold coordinated ones. If a photoexcitation of
the Ti sites modiÐed by organic molecules can lead to a degra-
dation of the organic compounds, an interesting photo-
catalytic reaction can be expected, and TS-1 can be applied as
a photocatalyst with a molecular selectivity.
1
Introduction
Zeolites are crystalline tectoaluminosilicates that have cavities
or channels smaller in size than 2 nm, which are able to recog-
nize, discriminate, and organize molecules with slight di†er-
ences within 0.1 nm. The zeolites are widely used in many
chemical and physical processes such as catalysis reactions,
ion exchanges, separations, and chemical sensing technol-
ogies.1h5 If the zeolites are applied to a photocatalyst, it
would be possible to realize the unique photocatalyst with a
molecular selectivity. However, a modiÐcation of the frame-
work and/or within the channels of zeolites is necessary
because aluminosilicate zeolites are transparent in a wave-
length range of UV-Vis light. Since a synthesis of TS-1 zeolite
by Taramasso et al.,6 who modiÐed the framework of
Semiconductor photocatalysts such as TiO have been
2
extensively investigated and used for the treatment of waste-
water. In the photocatalytic reaction using TiO , electron and
2
hole in conduction and valence bands formed via photoexcita-
tion of TiO contribute to the reduction and oxidation reac-
2
tions, respectively, as shown in Fig. 1(b). On the other hand, in
a photocatalytic reaction using TS-1, charge transfer excita-
tion in the Ti species modiÐed via the adsorption of organic
compounds, that is, Ti4`-O2~ ] Ti3`-O~, is expected to lead
to photocatalytic reactions, as shown in Fig. 1(a). Depending
silicalite-1 (SiO ) via an isomorphic substitution of Si in it for
Ti, many kinds of metallosilicate zeolites containing metal
on whether TS-1 or TiO is used as a photocatalyst, di†erent
2
2
atoms such as Ti, V, Mn, Zn and Fe in the framework have
been synthesized and investigated.7h23 Although TS-1 zeolite
absorbs only UV light with very short wavelength around 220
nm, Anpo and coworkers showed that a decomposition reac-
tion of NO24 and a reduction of CO with H O25h27 were
2
2
photocatalyzed in the presence of TS-1. In the decomposition
of NO, TS-1 led to the formation of N and O and showed a
2
2
high activity, while TiO resulted in the formation of N O
2
2
and O .24 In the reaction of CO and H O, the use of TS-1
led to the formation of methane and methanol, while TiO
2
2
2
2
yielded only methane.25h27 Afterwards, some workers have
studied the photocatalytic reaction using titanosilicate zeolites
or mesoporous materials with larger channels in order to
increase the reaction rate.28h32 However, an application of
metallosilicate zeolite to a photocatalytic treatment of waste-
water has not been examined at all. So far, some workers
investigated the inÑuence of water or NH aqueous solution
3
on the framework structure of titanosilicate zeolite.33h41 Their
investigations via XPS,33 IR,34,35 Raman,34 UV-Vis,34h37 and
XANES measurements38h41 suggested that water or NH
Fig. 1 Anticipated photocatalytic reactions using (a) TS-1 and (b)
3
molecules were adsorbed to 4-fold coordinated Ti atoms in
TiO .
2
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752
This journal is ( The Owner Societies 1999
5745

View Article Online
reaction products may result because a di†erent reaction
process is involved. Other unique photocatalytic properties
are also anticipated.
In this study, we examined how the adsorption of organic
compounds such as alkanolamine on Ti atom in TS-1 inÑu-
enced the UV-Vis spectra. It was also examined whether the
adsorbed alkanolamine molecules are photocatalytically
decomposed as shown in Fig. 1.
UV-Vis spectra (Hitachi, Model U3500) of liquid specimens
were collected in a transmittance mode using a dual beam.
The specimens were put into a quartz glass cell (10 ] 10 ] 45
mm, Nippon Quartz Glass Co.). An empty cell was used as a
reference.
Infrared (IR) spectra (Perkin Elmer, Model 2000 FT IR) of
solid specimens were measured in a wavenumber range
between 450 and 4500 cm~1 via a KBr method. A very small
amount of the specimen was mixed with 0.15 g of KBr
powder. It was uniaxially pressed at 0.05 t cm~2 under a
reduced pressure of 0.05 mmHg to give a translucent pellet.
The speciÐc surface area (Japan Bel Co., Model BELSORP
2
Experimental procedures
2
.1 Preparation of TS-1
TS-1 zeolite was synthesized from an aqueous solution con-
2
8SA) was measured via a BET method using N -gas adsorp-
2
taining tetraethyl orthosilicate (Si(OEt) ), triethanol-
tion. The specimens were pre-treated at 200 ¡C under 0.05
mmHg for 2 h. An adsorption isotherm was measured at 77 K
in a relative pressure range from 0.15 to 0.40.
4
aminatotitanium i-propoxide (Ti((OC H ) N)(OPri)) and
2
4 3
tetra-n-propylammonium hydroxide (NPrnOH) as a template
agent via a hydrothermal treatment. At Ðrst, 50.00 g of 10
4
wt.% NPrnOH aqueous solution, which contained 25.0 mmol
2.3 Photocatalytic reactions
4
of NPrnOH, was cooled in an ice bath. Then, 12.75 g (61.25
4
A reactant solution with a concentration of MEA of 0.5 mM
was prepared from milli-Q water and MEA. The pH value of
the solution was 9.8. The powder of TS-1 was added into the
solution and an added amount of TS-1 was adjusted to 0.5 g
L~1. The suspension was stirred for 40 min in the dark and
then transferred into a test tube (z 10 mm) made of quartz
glass. A photoirradiation was conducted under stirring using a
light from super high pressure Hg lamp, which was passed
through an optical Ðlter (Toshiba Glass Co., UV-D33S) and
water in order to cut UV light with short wavelength and
infrared radiation, respectively. The UV-Vis spectrum of the
optical Ðlter itself showed a maximum transmittance of 89%
at 324 nm and transmittance larger than 50% in the wave-
length range between 252 and 393 nm. As references, silicalite-
mmol) of Si(OEt) was added into the solution under stirring,
followed by stirring for 1 h at 0 ¡C. In the next step, 0.20 ml of
4
8
0 wt.% Ti((OC H ) N)(OPri) propan-2-ol solution, which
4 3
contained 1.25 mmol of Ti((OC H ) N)(OPri), was added
2
2
4 3
dropwise into the solution, followed by stirring for 4 h at 0 ¡C.
The chemical composition of the solution was adjusted to
TiO : SiO : NPrnOH : H O \ 0.01 : 0.49 : 0.20 : 20, that is,
2
2
4
2
Ti/(Si ] Ti) \ 0.02. The solution was transferred into a teÑon-
lined autoclave and then heat-treated at 170 ¡C under an
autogenous pressure for 1 week. A prepared powder was
separated by centrifugation, washed using 1000 ml of distilled
water and then heat-treated at 600 ¡C for 15 h in a Ñow of air
in order to remove the template agent. Thus prepared powder
was used as a TS-1 specimen. A silicalite-1 specimen was pre-
pared via the same procedure without the Ti source.
1, which has the same crystal structure as TS-1 and a chemical
composition of SiO , and TiO (Degussa P-25), which con-
2
2
sists of anatase and very small amount of rutile, were used.
The photocatalytic decomposition of MEA was also exam-
ined via photoirradiation while bubbling N or O gas. The
2.2 Characterization of the specimens
Crystalline phases in the prepared specimens were examined
via powder X-ray di†raction (XRD) measurements (Rigaku
GeigerÑex RAD-2R) using Cu-Ka radiation monochromated
by graphite. The measurements were conducted with a scan
speed of 2¡ min~1 for a scanning range between 10 and 70¡ in
terms of 2h. Lattice constants were estimated via Rietveld
analysis of XRD pattern recorded for a mixture of the speci-
men and crystalline Si powder for a reference. The XRD
pattern for the estimation of lattice constants was measured in
a step scan mode with a step width of 0.02¡ and a Ðxed time of
2
2
bubbling of N and O was conducted in order to eliminate
2
2
and increase soluble O , respectively. These gases were
bubbled into 6 ml of the suspension containing 0.5 g L~1 of
2
the specimen powder from its bottom using a needle with an
inside diameter of 0.5 mm at a Ñow rate of 2 L min~1. The
bubbling of the gas was carried out for 60 min before the pho-
toirradiation and continued also during the photoirradiation.
The photoirradiation was conducted under stirring using a
light from super high pressure Hg lamp, which was passed
through the above-mentioned optical Ðlter and water.
1
0 s for a scanning range between 20 and 80¡ in 2h. The Riet-
veld analysis was conducted using software of RIETAN 94
written by Izumi.42
The photoirradiated suspensions were Ðltered using a mem-
brane with a pore size of 0.20 lm. The components in the
solutions were characterized via high performance liquid chro-
matography (HPLC). The amine components such as mono-
ethanolamine, glycine (H NCH COOH), and ammonia were
Thermal analysis (TG-DTA measurements, Shimadzu,
Model DT-40) was conducted in order to examine the com-
bustion temperature of the template agent using 20 mg of the
specimen before the heat treatment at 600 ¡C. TG-DTA curves
were recorded in a Ñow of air in a temperature range between
2
2
detected using a separation column for cation analysis (Tosoh
Co., IC-CATION SW model) via a change of conductivity
2
5 and 1000 ¡C at a heating rate of 5 ¡C min~1.
(
Tosoh Co., Model IC-8010). A 2.5 mM of HNO solution
3
Di†use reÑectance ultraviolet-visible (DR-UV-Vis) spectra
with H O and acetonitrile (CH CN) as solvents, in which
H O : CH CN \ 1 v/v, was used as a mobile phase. The Ñow
2
3
(
Hitachi, Model U4000) were measured in a wavelength range
2
3
between 200 and 1500 nm using a BaSO -lined integration
rate was 0.6 mL min~1. The components of some aldehydes
and carboxylic acids were detected using a separation column
for anion analysis (Tosoh Co., Model OA PAK-A) via changes
of conductivity (Tosoh Co., Model CM-8020) and absorbance
of UV light at 210 nm (TOSOH Co., Model PD-8020). A 0.75
mM of H SO aqueous solution was used as a mobile phase.
4
sphere (z 60 mm). An a-Al O disk was used as a reference.
2
3
For powder specimens, a pellet (z 10 mm) was made from 0.3
g of the specimen at an uniaxial press of 0.1 t cm~2 under a
reduced pressure of 0.05 mmHg, and set to a sample holder
with a hole (z 8 mm). For suspension specimens, the specimen
was put into a quartz glass cell (3 ] 10 ] 45 mm, Nippon
Quartz Glass Co.) for the measurement. All spectra were
2
4
The Ñow rate was 0.8 mL min~1.
reduced
according
to
KubelkaÈMunk
theory;43
3
Results and discussion
F(R) \ (1 [ R)2/2R, where F(R) is the KubelkaÈMunk func-
tion and R is reÑectance. When the KubelkaÈMunk function
is less than 3, a peak area is proportional to a concentration
of an absorbant.43
3
.1 Characterization of a prepared specimen
Just after the hydrothermal synthesis the powder was charac-
terized via TG-DTA measurement. The thermograms showed
5746
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752

View Article Online
an exothermic peak with a large weight loss around 400 ¡C.
This may be assigned to the combustion of the template agent,
NPrnOH. An amount of the template agent in the specimen
4
was estimated at Si
Ti O É 4NPr OH from the weight
9
6~x x 192
4
loss around 400 ¡C, which was consistent with the previous
report.44 No other endothermic or exothermic peaks and
weight losses were observed. Because the template was burned
o† around 400 ¡C all at once, it can not be considered that the
specimen, which was heat-treated at 600 ¡C, contained organic
substances. The prepared specimen was characterized via
XRD, DR-UV-Vis, and IR measurements. All the peaks
observed in the peaks XRD patterns were assigned to MFI-
type zeolite, into which TS-1 and silicalite-1 are classiÐed.
Because no peaks assignable to TiO such as rutile and
2
anatase were observed, the Ti atoms may be homogeneously
incorporated into the zeolite.
Fig. 2(A) and (E) show DR-UV-Vis spectra of the TS-1 and
silicalite-1 specimens, respectively. These Ðgures clearly show
that silicalite-1 had no absorption in the measured wavelength
range, while TS-1 exhibited a peak at 220 nm assignable to an
isolated and 4-fold coordinated Ti, indicating that all Ti
atoms in the specimen were incorporated in the framework of
the zeolite.33,36,45h51 When the spectrum of the TS-1 speci-
men was examined in detail, absorption started at 300 nm and
gradually became stronger with decreasing the wavelength of
the irradiated light. This indicates that a small amount of
Fig. 3 IR spectra of (A) TS-1 and (B) silicalite-1. The arrow shows
absorption around 960 cm~1.
relation between I(960 cm~1) : I(800 cm~1) and the chemical
composition reported by Tarramasso et al.6 and Tuel,55 the
chemical composition of the TS-1 specimen was estimated at
Ti : (Ti ] Si) \ 0.013 to 0.014. The lattice constants of TS-1
were estimated via Rietveld analysis of its XRD pattern:
a \ 2.0097(1) nm, b \ 1.9925(1) nm and c \ 1.3403(1) nm, and
the lattice volume was 5.367(1) nm3, where the numbers in
parentheses show a standard deviation. The chemical com-
position can also be estimated from the unit cell volume.
Based on the relationship between the lattice volume and the
chemical composition that was reported by Tarramasso et al.6
and Tuel,55 it was found that the TS-1 specimen has the
chemical composition of Ti : (Ti ] Si) \ 0.013, that is,
5-fold coordinated Ti species is present in the TS-1 specimen.
The 5-fold coordinated Ti species may not be extra-
framework Ti species and may form via a partial hydration of
intra-framework Ti species by water in air.
It has been well known that the IR spectrum of zeolites
with intra-framework Ti exhibits an absorption band around
9
60 cm~1, which is absent for pure silica zeolites and charac-
Si
Ti
O . The combined results of XRD and FTIR
teristic of metal-substituted zeolites.46,52h54 The peak may be
assigned to a stretch mode of SiO unit perturbed by neighbor
Ti.46,52h54 As clearly shown in Fig. 3, the TS-1 specimen
0.987 0.013 2
measurements clearly indicate that the chemical composition
of the TS-1 specimen is supposed to be Ti : (Ti ] Si) \ ca.
4
0
.013, that is, Si
Ti
O , although a small amount of
exhibited an absorption around 960 cm~1, while that of
silicalite-1 specimen did not, conÐrming the presence of intra-
framework Ti atoms in the TS-1. However, the peak was
shifted to 970 cm~1. This indicates that a small amount of
0.987 0.013 2
H O may also be present. Ti concentration in the prepared
TS-1 specimen was lower than that in the starting solution.
2
Our
preliminary
study
indicated
that
when
Ti((OC H ) N)(OPri) was used as a Ti source, excess Ti was
5-fold coordinated intra-framework Ti species formed in the
2 4 3
removed during the washing process via a complex formation
TS-1 specimen via a partial hydration by water in air. The
intensity ratio of the absorption peak at 960 cm~1 to that
observed at 800 cm~1 in the IR spectrum can be used for the
estimation of the chemical composition of the TS-1 specimen.
The peak observed at 800 cm~1 remained regardless of the
presence of intra-framework Ti, which may be assigned to a
between Ti species and N(C H OH) .
2 4
3
3.2 Adsorption of various compounds to the Ti atoms in TS-1
In order to apply TS-1 as a photocatalyst, an organic com-
pound that is appropriate for a photocatalytic reaction has to
be selected. As the photocatalytic reaction was carried out
using diluted aqueous solution of an organic compound, it is
necessary for the organic compound to adsorb onto Ti site in
SiÈO symmetric stretching vibration of SiO units in the
4
framework. The peak intensity ratio, I(970 cm~1) : I(800
cm~1), was found to be 0.45 for the TS-1 specimen. Using the
TS-1 more preferentially than H O molecules and the used
2
organic compound has to be small enough in size in order to
enter and di†use into the channel of TS-1 zeolite. The TS-1
specimen was mixed with various neat organic compounds
such as alcohols and alkanolamines and their DR-UV-Vis
spectra were measured. DR-UV-Vis spectra of the obtained
suspensions are shown in Fig. 4, together with that of the
TS-1 specimen itself. Although the spectra of the suspensions
were measured just after the mixing and after aging for
1, 2 and 4 h, the peak intensities and proÐles did not change
with the aging time. All spectra of the suspensions showed
peaks not only around 220 nm but also around 300 nm. Inten-
sities of the absorption peak around 300 nm increased in the
following
order:
MeOH O N(C H OH)
(TEA) \
2
4
3
EtOH O PriOH \ ButOH \ NH
aq. \ H O \ MEA \
3
2
NH(C H OH) (DEA). The spectra of the suspensions con-
2
4
2
taining alkanolamines such as MEA and DEA tended to
exhibit more intense peaks around 300 nm than those con-
Fig. 2 UV-Vis spectra of (A) TS-1, (E) silicalite-1, (F) MEA, their
mixtures; (B) TS-1 ] MEA, and (D) silicalite-1 ] MEA and (C) the
specimen after annealing TS-1 ] MEA at 600 ¡C.
taining alcohols, H O, and aqueous ammonia solution, and
among the alcohol suspensions the alcohols with bulkier alkyl
2
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752
5747

View Article Online
Fig. 5 Schematic diagrams of the formations of 6-fold coordinated
Ti. (a) Adsorption of organic compounds to intra-framework Ti in
TS-1. (b) Dropping Ti out of the framework via an alcoholysis reac-
tion followed by association of the Ti atoms.
coordinatable to a metal atom. TiÈOÈTi bond formation via a
breakage of TiÈOÈSi bond in the framework, as shown in Fig.
5
6
(b), is another possible explanation for the formation of
-fold coordinated Ti units. However, if any, the process is
Fig. 4 DR-UV-Vis spectra of (A) TS-1 and mixtures of TS-1 with (B)
MeOH, (C) EtOH, (D) PriOH, (E) ButOH, (F) H O, (G) NH
MEA, (I) DEA and (J) TEA.
3 aq., (H)
irreversible, and is not consistent with the reversibility in the
spectral change. It was conÐrmed that the formation of 6-fold
coordinated Ti species is attributed to the adsorption of MEA
molecules onto 4-fold coordinated intra-framework Ti atom,
as shown in Fig. 5(a).
The adsorption of MEA to TS-1 was quantitatively exam-
ined using 0.12 to 0.54 mM MEA aqueous solution. After
about 12 mg of the TS-1 specimen has been added into 24 ml
of the MEA solutions with various concentrations, a change
of MEA concentration in the solution with stirring time was
measured. The same experiments were conducted also for
2
groups led to more intense peaks. As mentioned above, the
absorption around 300 nm may be attributed to the adsorp-
tion of organic compounds to the Ti atoms in TS-1. Therefore,
it was found that alcohols with bulky alkyl group and alka-
nolamines were adsorbed to Ti in the framework more
readily, except for TEA. Probably TEA molecules may be
considered too large in molecular size to enter into the
channel of the zeolite. Because the alcohols with a bulkier
alkyl group among the tested ones have stronger basicity, e.g.
silicalite-1 and TiO . For all the specimens, the MEA concen-
tration rapidly decreased for 10 min after the addition of the
2
pK \ 15, 16 and 18 for MeOH, EtOH and ButOH, respec-
specimens and the adsorption of MEA reached an equilibrium
after 40 min. Relations of the equilibrated amounts of
adsorbed MEA per weight of the specimens with the concen-
tration of MEA in the solution, that is, adsorption isotherms
are shown in Fig. 6. The amount of the adsorbed MEA
increased in the following order: silicalite-1 B TiO @ TS-1.
The data were replotted assuming Langmuir adsorption. The
Langmuir adsorption is expressed in the following equation;
a
tively, they may adsorb more readily to Ti atom in TS-1. Con-
sequently, a photocatalytic reaction using TS-1 was carried
out via a degradation of MEA in diluted aqueous solution.
Fig. 2 shows UV-Vis spectra of TS-1, silicalite-1, MEA and
their mixtures. Fig. 2(A) shows that the spectrum of the TS-1
specimen exhibited only absorption around 220 nm, which
2
can be assigned to a charge transfer excitation in TiO tetra-
4
hedra in the framework. On the other hand, the spectra of
C/w \ C/w ] (Kw )~1,
(1)
mixture of TS-1 and MEA showed peaks not only around 220
nm but also around 300 nm, as shown in Fig. 2(B). Fig. 2(E)
and (D) show DR-UV-Vis spectra of silicalite-1, and the
mixture of silicalite-1 and MEA, respectively. Silicalite-1
exhibited no particular absorption in the measured wave-
length range. The mixture of silicalite-1 and MEA exhibited a
peak around 210 nm. The UV-Vis spectrum of MEA shown in
Fig. 2(F) exhibited absorption of UV light with wavelength
shorter than 250 nm. Although the TS-1 suspension in MEA
had a strong absorption around 300 nm, the spectrum of the
specimen which was recovered from the suspension and
annealed at 600 ¡C in air was consistent with that of native
TS-1 as shown in Fig. 2(C), indicating that the spectral change
is reversible. Therefore, it can be concluded that the peak
around 300 nm is attributable not to the framework of MFI-
type zeolite itself and/or its interaction with MEA but to the
Ti atoms in TS-1, especially 6-fold coordinated Ti
units.33,36,45h51 The Ti sites are originally 4-fold coordinated,
but the increase in the coordination number of Ti may be
realized by an adsorption of organic compounds to the Ti
atom, as shown in Fig. 5(a). Because original intra-framework
Ti is coordinatively unsaturated and the tested organic com-
pounds have O and/or N atoms with lone pair to be Ðrmly
=
=
where C is a concentration of a solute in a solution, w is an
amount of solute molecules adsorbed to unit weight of an
adsorbent, w is an amount of adsorbed molecules when a
surface of the adsorbent is completely covered via a mono-
=
Fig. 6 Relation between amounts of MEA adsorbed to the speci-
mens and the concentrations of MEA in the solutions, that is, adsorp-
tion isotherms for TS-1 (open circle), silicalite-1 (closed circle) and
TiO (open square). The amount of the specimens was 0.5 g L~1.
2
5748
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752

View Article Online
layer adsorption, and K is an equilibrium constant of Lang-
muir adsorption. Fig. 7 shows the Langmuir plots for TS-1,
silicalite-1 and TiO specimens, all of which exhibited linear
2
dependence. Therefore the adsorption of MEA to the speci-
mens followed Langmuir adsorption. The intercepts and
slopes in Fig. 7 show the 1/Kw and 1/w , respectively.
=
=
Therefore, the equilibrium constants of adsorption, K, were
4
.1, 1.7, and 3.8 L mmol~1 for TS-1, silicalite-1 and TiO
2
specimens, respectively. The values of w were 0.29, 0.14, and
=
0
.10 mmol g~1 for TS-1, silicalite-1 and TiO specimens,
2
respectively. SpeciÐc surface areas were also measured via
BET adsorption of N gas, which were 421, 395, and 72 m2
2
g~1, for TS-1, silicalite-1, and TiO specimens, respectively.
2
The TS-1 specimen showed slightly larger surface area than
silicalite-1 because TiÈO bond is longer than SiÈO. The values
Fig. 8 Changes of MEA concentration with irradiation time of UV
of K and w for TS-1 were much larger than those for
light in the photocatalytic reactions using TS-1 (open circle), silicalite-
=
1
(closed circle) and TiO (open square) as the photocatalyst and in
silicalite-1, being in good agreement with the concept shown
2
the case without the powder specimen (closed square).
in Fig. 4(a) that indicating that the intra-framework Ti atoms
o†ers an additional adsorption sites of MEA. Although the
value of K for TS-1 was only slightly larger than that for
TiO , the speciÐc surface area and w for the TS-1 specimen
men, the relation of log([MEA] : [MEA]) with t was linear in
0
the measured time range and the rate constant was estimated
2
=
were much larger than those for the TiO
the former is a microporous material.
2 specimen because
at 5.3 ] 10~3 min~1. On the other hand, a linear relation was
not obtained for the other specimens, because the photo-
catalytic reaction was accompanied by some reactions such as
an oxidative reaction of alcohol through aldehyde to carbox-
ylic acid and breakage reactions of CÈC or CÈN bonds, as
shown later. Thus, the reaction rate at an initial stage was
estimated from slopes in a time range from 0 to 20 min in Fig.
3
.3 Photocatalytic degradation reaction of MEA using TS-1
zeolite
A photocatalytic reaction using the TS-1 specimen as a photo-
catalyst was examined using 0.5 mM of MEA aqueous solu-
tion. An amount of the TS-1 specimen was Ðxed to 0.5 g L~1.
UV light with the wavelength around 300 nm was irradiated.
The same experiments were conducted for the suspensions
containing the silicalite-1 or TiO
9
. The values of k for TiO , silicalite-1 and the MEA solution
2
without powder were estimated at 6.7 ] 10~2, 2.2 ] 10~3,
and 1.6 ] 10~3 min~1, respectively. The degradation of a
small amount of MEA probably occurred even without a
powder catalysis, because the wavelength range of intrinsic
absorption of MEA is slightly overlapped with that of the
light transmitted through the used optical Ðlter. Although the
addition of silicalite-1 caused the decomposition of slightly
larger amount of MEA than in the case without powders,
silicalite-1 had practically no activity as a photocatalyst. As
obviously seen in Figs. 8 and 9, the photoirradiation of the
2 specimens and the MEA
solution without TS-1 powder.
The changes of the MEA concentration with irradiation
time are shown in Fig. 8. For the MEA solution without
powders, the photoirradiation for 320 min led to only slight
decrease in the MEA concentration from 0.5 to 0.41 mM.
However, in the presence of the silicalite-1, TS-1 and TiO
specimens, the MEA concentration decreased from 0.46 to
2
MEA suspension containing the TS-1 or TiO specimen led to
0.35 mM, from 0.41 to 0.06 mM, and from 0.47 to 0.00 mM,
2
much larger and faster decrease of the MEA concentration
respectively. A rate of the decrease in MEA concentration was
estimated assuming that the degradation reaction of MEA is
Ðrst order. When the reaction is Ðrst order, the concentration
than for the above-mentioned other specimens, indicating that
TS-1 is an e†ective photocatalyst similar to TiO . As it is
2
natural to consider, the Ti atoms in TS-1 played an important
of
MEA,
[MEA],
follows
the
equation
of
role in the photocatalytic reaction, activities of the photo-
catalytic reactions per mole of Ti atoms in the specimens were
compared although the activity per weight of the specimen
log([MEA] : [MEA]) \ kt, where [MEA] and k are the
0
0
starting concentration of MEA and the rate constant, respec-
tively. A sum of the concentration of MEA in the solution and
the amount of MEA adsorbed to the powder specimen which
was deduced from Figs. 6 and 7 was used as [MEA],
assuming a negligible adsorption of the reaction products to
was clearly higher for TiO than for TS-1. The amounts of Ti
2
the specimen powders. The plots of log([MEA] : [MEA]) vs.
0
the reaction time t are shown in Fig. 9. For the TS-1 speci-
Fig. 9 Changes of MEA concentration with irradiation time, t, of
UV light in the photocatalytic reactions using TS-1 (open circle),
silicalite-1 (closed circle) and TiO (open square) as the photocatalyst
2
and in the case without the powder specimen (closed square).
Assuming that the degradation reaction of MEA is the Ðrst order,
Fig. 7 Langmuir plots for adsorption of MEA to the specimens:
ln([MEA] : [MEA]) vs. t was plotted. The slopes of the displayed
0
TS-1 (open circle), silicalite-1 (closed circle) and TiO (open square).
lines show the rate constants.
2
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752
5749

View Article Online
atoms in the suspensions were 0.108 and 6.26 mmol L~1 for
The following are possible as the reductive reactions, that is,
proton and dissolved oxygen are reduced via the interaction
with the excited state of the Ti(IV)-ligand complex;
TS-1 and TiO specimens, respectively. Thus, the decomposi-
2
tion rates per mole of Ti in the specimen were 49.1 and 10.6 L
mol~1 min~1 for TS-1 and TiO specimens, respectively. Con-
2
H` ] e~ ] 1/2H C,
(2)
(3)
sequently, the activity of the reaction per mole of Ti in the
2
TS-1 specimen is about 5 times larger than that in the TiO2
specimen. The high activity of the TS-1 specimen may be
attributed to adsorption of larger amount of MEA to TS-1 as
well as the microporous structure which enables all Ti atoms
O ] e~ ] O ~.
2
2
If reaction (3) has a major contribution to the reduction reac-
tion, a bubbling of O gas should lead to an increase of the
2
reaction rate of MEA while that of N gas should retard it.
in TS-1 to act as reaction sites although for TiO only the Ti
2
2
Changes of MEA concentration with periods under photoirra-
atoms on the particle surface can do so.
diation when O and N gases were bubbled are shown in
The change of concentrations of the products with the irra-
diation time for the TS-1 specimen is shown in Fig. 10. The
products were glycine (NH CH COOH), glycolic acid
2
2
Fig. 11. It indicates that the bubbling of O gas resulted in an
2
increase of the reaction rate while the bubbling of N gas
2
2 2
CH (OH)COOH), acetic acid (CH COOH), formic acid
slightly decreased it. Even under N gas bubbling, that is, in
(
(
2
the case without the presence of soluble O , the photo-
2
2
3
CHOOH), formamide (CHONH ) and ammonia (NH ). It
2
3
catalytic reaction did proceed. Consequently, not only reac-
was not only the oxidation of MEA to glycine but also the
breakage of CÈC and CÈN bonds to occur via the photo-
catalytic reactions. The concentrations of glycolic acid and
ammonia increased more rapidly up to 80 min than the other
tion (3) but also other reactions, which may be related to
reaction (2), may contribute as the reductive reaction, that is,
the presence of soluble O is not always needed for the photo-
2
catalytic reaction and the reductive reaction of H` to H may
products. While the concentration of NH increased rather
2
3
act as the reductive reaction in the photocatalytic reaction if
O is not present. Since the reaction rate of the reduction of
O is faster than that of H`, the presence of O may facilitate
the photocatalytic reaction. The e†ects of N or O gas bub-
steadily with the photoirradiation time, that of glycolic acid
began to decrease at a later stage, that is, 80 min. Glycine
began to be formed after an irradiation period of 40 min and
continued to increase up to 320 min. Thus, at Ðrst the break-
age of CÈN bond occurs rapidly to give glycolic acid and
ammonia, gradually followed by the oxidation of MEA to
glycine, which is an oxidation of alcohol to carboxylic acid. In
2
2
2
2
2
bling on the product distribution were also examined. The
typical results for the bubbling of N are shown in Fig. 12.
2
Glycolic acid concentration steadily increased with time and
no formamide and formic acid were found in the product, dif-
the case of TiO the products were the same as in the case of
2
the TS-1 specimen. However, the product distribution and its
time dependence were somewhat di†erent from those of TS-1.
The concentrations of the products and their change up to 20
min were very similar to those for TS-1 up to 320 min. For a
photoirradiation period longer than 20 min, the concentration
of glycine increased up to 80 min and then decreased. The
solution, which was photoirradiated for 80 min, contained
MEA, ammonia, and glycine and Ðnally that for 320 min con-
tained only ammonia. In both cases, the concentration of C
atoms in the suspensions largely decreased although that of N
atoms was almost constant. This may be attributed to the
removal of C component from the suspension via a formation
of CO .
2
The inÑuence of bubbling of O and N gases on photo-
2
2
catalytic degradation reaction of MEA was investigated in
order to examine the reductive reaction. The bubbling of the
gas was carried out for 60 min before the photoirradiation
and also maintained during the photoirradiation. The pho-
toirradiation was conducted using UV light with the wave-
length around 300 nm through the optical Ðlter and water.
Fig. 11 Changes of concentration of MEA with irradiation time of
UV light in the photocatalytic reaction using TS-1 while bubbling O
2
(
open square) and N (closed circle) gases. The changes in the cases of
2
no gas-bubbling (open circle) and no irradiation while bubbling O
2
(
closed square) are also shown.
Fig. 10 Changes of concentrations of products with irradiation time
of UV light in the photocatalytic reaction using TS-1 as the photo-
catalyst. The products are ammonia (open square), glycolic acid
Fig. 12 Changes of concentrations of products with irradiation time
of UV light in the photocatalytic reaction using TS-1 while bubbling
(
(
cross), glycine (open triangle), acetic acid (closed circle), formamide
closed square) and formic acid (closed triangle). The change of MEA
N gas. The products are ammonia (open square), glycolic acid (cross),
2
glycine (open triangle) and acetic acid (closed circle). The change of
concentration is also shown (open circle).
MEA concentration is also shown (open circle).
5750 Phys. Chem. Chem. Phys., 1999, 1, 5745È5752

View Article Online
(5) Both reductions of dissolved oxygen and H` might con-
tribute to the photocatalytic reaction. A decomposition by
O ~ formed via the reduction of dissolved oxygen may con-
2
tribute to a breakage of the CÈC bond.
As the photocatalytic degradation reaction of MEA occurred
via the photoexcitation of the MEA-adsorbed Ti species, a
realization of the photocatalytic reaction with a molecular
selectivity is expected. The investigations of the photocatalytic
reaction with a molecular selectivity are in progress and their
results will be presented elsewhere.
References
1
G. A. Ozin, A. Kuperman and A. Stein, Angew. Chem. Int. Ed.
Fig. 13 Speculated reaction schemes of (a) decomposition of MEA
Engl., 1989, 28, 359.
into NH and glycolic acid and (b) oxidation of MEA to glycine.
3
2
3
J. M. Thomas, Angew. Chem. Int. Ed. Engl., 1988, 27, 1673.
G. D. Stucky and J. E. MacDougall, Science (Washington, D.C.),
ferent from the case of no N -bubbling. In the presence of
1
990, 247, 669.
2
soluble O , O ~ may be formed, which can oxidatively
4
5
6
M. E. Davis, Acct. Chem. Res., 1993, 26, 111.
2
2
decompose an organic compound, which would contribute to
the formations of formamide and formic acid, that is, a break-
age reaction of CÈC bonds. For the breakage of CÈC bonds,
J. M. Newsam, Science (Washington, D.C.), 1986, 231, 1093.
M. Taramasso, G. Perego and B. Natori, U. S. Patent, 4410501,
1
983.
7
A. Thangaraj, R. Kumar, S. P. Mirajjkar and P. Ratnasamy, J.
Catal., 1991, 130, 1.
the presence of soluble O may be necessary. In the absence of
2
soluble O , the photocatalytic reactions such as the oxidation
8
9
J. S. Reddy and R. Kumar, J. Catal., 1991, 130, 440.
T. Tatsumi, K. Yanagisawa, K. Asano, M. Nakamura and H.
Tominaga, Stud. Surf. Sci. Catal., 1994, 83, 417.
A. Tuel and Y. B. Taarit, Appl. Catal., 1993, 102, 69.
T. Tatsumi, K. Yuasa and H. Tominaga, J. Chem. Soc., Chem.
Commun., 1992, 1446.
A. Tuel, Zeolites, 1995, 15, 236.
P. R. H. P. Rao, A. V. Ramaswamy and P. Ratnasamy, J. Catal.,
1992, 137, 225.
2
of MEA to glycine and the decomposition of MEA into
ammonia and glycolic acid may occur but the Ðssion of CÈC
bonds may not take place.
1
1
0
1
In the photodegradation reaction of MEA, the three types
of reactions occurred; (1) the breakage of CÈN bond, (2) oxi-
dation of alcohol to carboxylic acid and (3) the breakage of
CÈC bonds. Fig. 13 shows the reaction schemes speculated for
reactions (1) and (2). When amine in MEA was datively
bonded to intra-framework Ti in TS-1 to give 6-fold coordi-
nated Ti species, charge transfer (CT) transition in the dative
bond of TiÈN may lead to the breakage of the CÈN bond in
1
2
3
1
14 A. V. Ramaswamy, S. Sivasanker and P. Ratnasamy, Micro-
porous Mater., 1994, 2, 451.
1
1
5
6
A. Tuel and Y. B. Taarit, Zeolites, 1994, 14, 18.
T. Sen, M. Chatterjee and S. Sivasanker, J. Chem. Soc., Chem.
Commun., 1995, 207.
G. I. Panov, G. A. Sheveleva, A. S. Kharitonov, V. N. Roman-
ikov and L. A. Vostrikova, Appl. Catal., 1992, 82, 31.
MEA, resulting in the formation of NH and glycolic acid.
1
7
3
This reaction began to occur immediately after the photoirra-
diation. On the other hand, when alcohol in MEA was dati-
vely bonded to intra-framework Ti, CT transition in the
dative bond of TiÈO may lead to the oxidation of alcohol in
MEA to carboxylic acid, resulting in the formation of glycine.
This reaction occurred after an induction period. As men-
18 S. Bordiga, R. Buzzoni, F. Geobaldo, C. Lamberti, E. Giamello,
A. Zecchina, G. Leofanti, G. Petrini, G. Tozzola and G. Vlaic, J.
Catal., 1996, 158, 486.
1
2
2
9
0
1
Y. S. Ko and W. S. Ahn, Microporous Mater., 1997, 9, 131.
A. Tavolaro, J. T herm. Anal., 1996, 47, 171.
A. P. Singh and T. Selvam, J. Mol. Catal. A: Chem., 1996, 113,
489.
tioned above, O ~, which was formed via the reduction of
2
soluble O , may be attributed to the breakage of the CÈC
22 P. S. Raghavan, V. Ramaswamy, T. T. Upadhya, A. Sudalai, A.
2
V. Ramaswamy and S. Sivasanker, J. Mol. Catal. A: Chem., 1997,
bond although details of the reaction were not obvious.
1
22, 75.
2
3
B. Rakshe, V. Ramaswamy, S. G. Hegde, R. Vetrivel and A. V.
4
Conclusions
Ramaswamy, Catal. L ett., 1997, 45, 41.
24 Y. Ichihashi, H. Yamashita and M. Anpo, Abstrct 11th Int.
Zeolite, Conf., Seoul, 1996, p. 129.
A photocatalytic degradation reaction of MEA was examined
using TS-1 zeolite as a photocatalyst. The following conclu-
sions were drawn:
2
2
5
6
M. Anpo and K. Chiba, J. Mol. Catal., 1992, 74, 207.
M. Anpo, H. Yamashita, Y. Ichihashi and S. Ehara, J. Electro-
anal. Chem., 1995, 396, 21.
(
1) A mixing of TS-1 with MEA led to a change of coordi-
native situations around Ti atom, that is, a coordination
number of Ti was changed from 4 to 6 via an adsorption of
MEA to Ti. The change in the coordination number of Ti
shifted the absorption by a charge transfer excitation from 220
to 300 nm.
2
7
H. Yamashita, Y. Ichihashi, M. Harada, G. Stewart and M. A.
Fox., J. Catal., 1996, 158, 97.
28 K. Kosuge and P. S. Singh, Chem. L ett., 1999, 9.
2
3
3
9
0
1
S. G. Zhang, Y. Fujii, H. Yamashita, K. Koyano, T. Tatsumi and
M. Anpo, Chem. L ett., 1997, 659.
Y. Ichihashi, H. Yamashita and M. Anpo, Stud. Surf. Sci. Catal.,
(
2) Larger amount of MEA was adsorbed to TS-1 than to
1
997, 105, 1609.
silicalite-1 and TiO . This may be attributed to the adsorp-
M. Anpo, H. Yamashita, K. Ichihashi, K. Ikeue, Y. Fujii, Y. Ich-
ihashi, S. G. Zhang, D. R. Park, S. Ehara, S.-E. Park, J.-S. Chang
and J. W. Yoo, Stud. Surf. Sci. Catal., 1998, 114, 177.
S. G. Zhang, Y. Ichihashi, H. Yamashita, T. Tatsumi and M.
Anpo, Chem. L ett., 1996, 895.
2
tion of MEA to the intraframework Ti.
(
3) A photoirradiation of UV light with the wavelength of
3
3
2
3
around 300 nm to a mixture of MEA aqueous solution and
TS-1 caused a photocatalytic degradation of MEA. Although
an activity of the reaction per weight of the specimen was
L. L. Noc, D. T. On, S. Solomykina, B. Echchahed, F. Be land, C.
C. dit Moulin and L. Bonneviot, Stud. Surf. Sci. Catal., 1996, 101,
611.
34 A. Zecchina, S. Bordiga, C. Lamberti, G. Ricchiardi, D. Scarano,
lower for TS-1 than for TiO , that per mole of Ti atom in
2
TS-1 was higher than that in TiO .
(
2
4) No di†erences in products and changes of their concen-
G. Petrini, G. Leofanti and M. Mantegazza, Catal. T oday, 1996,
3
2, 97.
trations were observed in the photocatalytic reactions. At Ðrst,
a breakage of the CÈN bond in MEA occurred rapidly
3
5
S. Bordiga, S. Coluccia, C. Lamberti, L. Marchese, A. Zecchina,
F. Boschrini, F. Bu†a, F. Genoni, G. Leofanti, G. Petrini and G.
Vlaic, J. Phys. Chem., 1994, 98, 4125.
resulting in the formation of NH and glycolic acid. At a
3
second stage, an oxidation of MEA to glycine occurred grad-
36 L. Marchese, T. Maschmeyer, E. Gianotti, S. Coluccia and J. M.
ually.
Thomas, J. Phys. Chem. B, 1997, 101, 8836.
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752
5751

View Article Online
3
7
8
S. Bordiga, F. Geobaldo, C. Lamberti, A. Zecchina, F. Bos-
cherini, F. Genoni, G. Leofanti, G. Perrini, M. Padonvan, S.
Geremia and G. Vlaic, Nucl. Instrum. Methods Phys. Res., Sect.
B, 1995, 97, 23.
S. Bordiga, F. Boscherini, S. Coluccia, F. Genoni, C. Lamberti, G.
Leofanti, L. Marchese, G. Petrini, G. Vlaic and A. Zecchina,
Catal. L ett., 1994, 26, 195.
C. Lamberti, S. Bordiga, A. Zecchina, G. Vlaic, G. Tozzola, G.
Petrini and A. Carati, J. Phys. IV , 1997, 7, C2-851.
T. Maschmeyer, F. Rey, G. Sankar and J. M. Thomas, Nature
47 Y. L. Kim, R. L. Riley, M. J. Huq, S. Salim, A. E. Le and T. E.
Mallouk, Mater. Res. Soc. Symp. Proc., 1991, 233, 145.
48 S. Klein, B. M. Weckhuysen, J. A. Marteno, W. F. Maier and P.
A. Jacobs, J. Catal., 1996, 163, 489.
3
49 S. V. Lysak and B. Y. Sukhaoevskii, Dokl. Acad. Nauk SSSR,
1969, 188, 1299.
50 M. R. Boccuti, K. M. Rao, A. Zecchina, G. Leofanti and G.
Petrini, Stud. Surf. Sci. Catal., 1989, 48, 133.
51 G. Petrini, A. Cesana, G. deAlberti, F. Genoni, G. Leofanti, M.
Padovan, G. Paparatto and P. RoÐa, Stud. Surf. Sci. Catal., 1991,
68, 761.
52 A. Zecchina, G. Spoto, S. Bordiga, M. Padovan, G. Leofanti and
G. Petrini, Stud. Surf. Sci. Catal., 1991, 65, 671.
53 D. Scarano, A. Zecchina, S. Bordiga, F. Geobaldo, G. Spoto, G.
Patrini, G. Leofanti, M. Pandovan and G. Tozzola, J. Chem. Soc.,
Faraday T rans., 1993, 89, 4123.
54 G. Bellussi, A. Carati, M. G. Clerici, G. Maddinelli and R. Minili,
J. Catal., 1992, 133, 220.
55 A. Tuel, Stud. Surf. Sci. Catal., 1997, 105, 261.
3
4
4
4
4
4
4
4
9
0
1
2
3
4
5
6
(
L ondon), 1995, 378, 159.
T. Blasco, M. A. Camblor, A. Corma and J. Pe

rez-Pariente, J.
Am. Chem. Soc., 1993, 115, 11806.
F. Izumi, in T he Rietveld Method, ed. R. A. Young, Oxford Uni-
versity Press, Oxford, 1993, ch. 13.
G. Kortu
Ž m, in ReÑections Spectroskopie, Springer-Verlag, Berlin,
1
969.
H. van Koningsveld, H. van Bekkum and J. C. Jansen, Acta
Crystallogr., 1987, B43, 127.
G. Grubert, M. Wark, N. I. Jaeger, G. Sculz-Eklo† and O. P.
Tkachenko, J. Phys. Chem. B, 1998, 102, 1665.
A. Zecchina, G. Spoto, S. Bordiga, A. Ferrero, G. Petrini, M.
Pandovan and G. Leofanti, Stud. Surf. Sci. Catal., 1991, 69, 251.
Paper 9/06585G
5752
Phys. Chem. Chem. Phys., 1999, 1, 5745È5752