Y. Yang et al. / Journal of Catalysis 320 (2014) 160–169
161
obtained by microwave synthesis exhibited higher hydrophobicity
[25,26]. Silylation with organosilanes extinguished Si–OH groups
effectively, forming hydrophobic groups. Thus, trimethylsilylated
Ti-MCM-41 and Ti-MCM-48 showed significantly enhanced perfor-
mance in epoxidation reactions [27]. The hydrophobicity and then
the oxidation properties of Ti-MWW, on the other hand, were
improved by a unique reversible structural rearrangement [28].
Recently, NH4F treatment was shown to increase the surface
hydrophobicity of TS-1 [23].
Si–OH and Ti–OH groups both generate weak Brønsted acidity,
which may alter the negative influence on the Ti sites and also
cause undesirable side reactions such as solvolysis ring opening
of epoxides [29]. The addition of salts and controlled ion exchange
with alkali ions effectively enhances the selectivity of epoxides by
reducing opening reaction [30–34]. Also, nitrogen doping via
ammonia treatment removed the acidity of TS-1 and Ti-MWW
effectively [35,36].
reactor by N2 flow, and the Del-M sample was treated at 673 K
for 1 h. The sample was then purged with pure N2 at the same
temperature for 1 h to remove any residual TiCl4 from the zeolite
powder. After cooling to room temperature under N2, the treated
sample was washed with deionized water and dried in air at
353 K overnight. The resultant product, named Ti-M-P (Si/
Ti = 54), was further used as the parent to prepare fluorine-doped
samples.
In typical fluorine doping, the Ti-M-P sample was treated in a
solution containing NH4F (F/Si molar ratio of 0.54) with a solid-
to-liquid ratio of 1 g to 20 g at 423 K for 6 h in a Teflon-lined auto-
clave under magnetic stirring. Methanol or water was employed as
solvent. The fluorinated samples were denoted as F-Ti-M-M for
methanol and F-Ti-M-W for water. To optimize fluorine doping,
the effects of post-treatment temperature (333, 353, 373, 393,
and 423 K) and the amount of added NH4F (F/Si of 0.011, 0.027,
0.081, 0.54 and 0.81) were investigated in these two solvents.
For comparison, alkali metal fluorides (LiF, NaF, KF, CsF) were
also employed as fluorine sources instead of NH4F to prepare F-
Ti-M samples in water and methanol under the optimal conditions
of NH4F treatment. The use of NH4Cl to replace NH4F was also
investigated.
Fundamentally, the essence of titanosilicate-catalyzed oxida-
tion is the primary activation of H2O2 on Ti sites, forming a peroxo
a
intermediate Ti–O –Ob–H, in which the coupled oxygen atoms
have different electrophilicity. Therefore, it is possible to tune the
electric environment of tetrahedral Ti and thus to affect its activity.
As the most electronegative element, fluorine is expected to
change the electropositivity of the elements’ bonds in a neighbor-
hood easily through an electron-withdrawing effect. Fluorine mod-
ification was reported to influence the solid acidity and
hydrophilic–hydrophobic nature of zeolites by reacting with the
terminal Si–OH groups to form Si–F ones [37–39]. However, fluo-
rine species directly incorporated into Ti-Beta were claimed to
retard its catalytic performance [40]. The fluorine species incorpo-
rated into ZSM-5 [41], or UCB-1 [39] were usually connected to the
silicate frameworks as pentacoordinated SiO4/2FÀ units or existed
as extra-framework SiF26À species. On the other hand, tetrahedrally
coordinated SiO3/2F units were reported possibly to exist in fluori-
nated silica glasses [42]. Recently, we reported that the fluorina-
tion of the Ti-MWW framework led to enhanced catalytic
performance of F-Ti-MWW in the epoxidation of various alkenes
[43,44]. This is probably the first example of fluorine species
enhancing the catalytic performance of titanosilicates.
Among the titanosilicates developed so far, Ti-MOR cannot yet
be synthesized directly, but it is readily prepared by a combination
of dealumination and solid–gas reaction with TiCl4 vapor [6]. Ti-
MOR with 12-MR large pores shows the advantages of not using
organic structure-directing agents in preparation and it provides
attractive catalytic performance in ketone ammoximation and aro-
matics hydroxylation [45,46]. Containing dealumination defect
sites such as hydroxyl nests, Ti-MOR potentially has the opportu-
nity to gain higher performance by fluorination.
2.2. Characterization methods
Powder X-ray diffraction (XRD) was employed to check the
structure and crystallinity of the zeolites. The XRD patterns were
collected on a Rigaku Ultima IV diffractometer using Cu Ka radia-
tion (k = 1.5405 Å) at 35 kV and 25 mA in the 2h angle range of
5–35°. Scanning electron microscopy (SEM) was performed on a
Hitachi S-4800 microscope to determine the morphology. The
29Si MAS NMR spectra were measured on a VARIAN VNMRS
400WB NMR spectrometer using the single-pulse method at a fre-
quency of 79.43 MHz, a spinning rate of 3 kHz, and a recycling
delay of 60 s. The chemical shift was referred to Q8M8 ([(CH3)3-
SiO8]SiO12). 19F MAS NMR spectra were acquired at 9.4 T on a
Varian Infinity Plus 400 WB spectrometer using a 2.5 mm HX
MAS probe. The chemical shift was referenced to trifluoroacetic
acid at À76.55 ppm. The Si, Ti, and Al content was determined by
inductively coupled plasma emission spectrometry (ICP) on a
thermo IRIS Intrepid II XSP atomic emission spectrometer after
the samples were dissolved in HF solution. The adsorption iso-
therms of N2 and other gases (water, benzene, cyclohexane, n-hex-
ane, and cyclohexanone) were measured on a BELSORP-MAX
instrument equipped with a precise sensor for low-pressure mea-
surement at 77 K for N2, 298 K for water, benzene, cyclohexane,
and n-hexane, and 312 K for cyclohexanone. The samples were
activated in advance at 573 K under vacuum for at least 4 h. The
In this study, we have investigated in detail the fluorination of
Ti-MOR. The effects of fluorination on the catalytic performance
of F-Ti-MOR depend on the types of oxidation reactions.
UV–visible diffuse reflectance spectra were recorded on
a
Shimadzu UV-2550 spectrophotometer using BaSO4 as a reference.
Infrared spectra were obtained on a Nicolet Nexus 670 FT-IR spec-
trometer in the absorbance mode at a spectral resolution of 2 cmÀ1
.
2. Experimental
To avoid the influence of adsorbed water, the wafer was set in a
quartz IR cell sealed with KBr windows and connected to a vacuum
system, and it was evacuated at 723 K for 1 h before measurement.
2.1. Preparation of materials
The spectra in the region of framework vibration (500–1300 cmÀ1
)
A commercially available H-mordenite (Si/Al = 6.5) was dealu-
minated following previous procedures [6], giving rise to a sili-
ceous sample with a Si/Al molar ratio of 110, denoted as Del-M.
Del-M was used as the starting material for preparing Ti-contain-
ing mordenite, Ti-M. Ti-M samples were postsynthesized through
a solid–gas reaction between Del-M and TiCl4 vapor at elevated
temperature [6]. A Del-M sample (2.0 g) placed in a quartz tube
reactor (diameter 3 cm) was pretreated at 673 K for 2 h in a dry
N2 stream (50 mL minÀ1). Then TiCl4 vapor was brought into the
were recorded using the KBr pellet technique (1.5-cm-diameter
wafer, 6 wt.% diluted in KBr), whereas the spectra in the hydroxyl
stretching region (3100–3900 cmÀ1) were obtained using self-sup-
ported wafers 4.8 mg cmÀ2 thick. X-ray photoelectron spectro-
scopic (XPS) analyses were performed on a Kratos Axis Ultra
spectrometer (Kratos Analytical, UK) equipped with a monochro-
matized aluminum X-ray source. DFT calculations with the B3LYP
hybrid exchange correlation functional and 6–311 g were
employed for all types of atoms [47].