CHEMCATCHEM
FULL PAPERS
As the Brønsted active sites are the same for ethanol dehydra-
tion and m-xylene isomerization, we were able to determine
the TOF in m-xylene isomerization: 1.4ꢀ10ꢀ3 sꢀ1 siteꢀ1. This
value is 22 times lower than obtained over the BAS of a USY
zeolite without extraframework aluminum.
Part of ð1 0 0Þ uncovered ½%ꢂ ¼
ð6Þ
EtOH dehydration sitesꢀBAS
EtOH dehydration sites of alumina
in which the EtOH dehydration sites of pure alumina are fixed
at 0.83 nmꢀ2 and the fraction of BAS of the total number of
ethanol dehydration sites (Table 6):
Experimental Section
TGA experiments were performed with a Mettler Toledo apparatus
(TGA/SDTA851e) according to the procedure of Kwak et al.[4] Before
the measurements, the samples were pretreated in a flow of
Fraction of BAS in EtOH dehydration sites ½%ꢂ ¼
ð7Þ
BAS
EtOH dehydration sites
helium (60 mLminꢀ1) at 200 and 5008C, respectively, 108Cminꢀ1
,
for 2 h to eliminate physisorbed water from the surface and/or to
activate sites. Samples were cooled to RT. Helium was passed
through a saturator containing ethanol at RT for approximately
30 min, and the increase in the weight of the sample was deter-
mined. Weakly bonded ethanol molecules were evacuated by
means of a helium purge at RT for 1 h. TPD of ethanol was per-
formed at a rate of 108Cminꢀ1 until 4008C. The intensity of the de-
sorption features was measured by calculation of the first deriva-
tive of weight loss during TPD with respect to temperature, and
normalized by the surface area of the sample. The release of etha-
nol and ethylene was confirmed by mass spectrometry. The
amount of dehydrated ethanol determined from the TGA results
was calculated as follows:
These data prove the preferential deposition of the first sili-
con species, at an early stage of the grafting, under relatively
mild synthesis conditions, on the (100) surface of g-alumina
(Table 6, entries 1–7): the degree of coverage of the (100) sur-
face is always higher than that of the whole alumina surface.
There are virtually no BAS on pure alumina. On SA(7/anh),
most of the dehydration sites are still on the (100) surface.
SA(12/anh) and SA(17/12eqW) reveal that some of the dehy-
dration sites on the (100) surface are still present while some
BAS appear. On SA(17/anh), SA(17/4eqW), and SA(17/CVD250),
the (100) surface is completely covered.
Ethanol dehydrated ½nmꢀ2ꢂ ¼
Nonselective grafting of silicon atoms on the facets of
g-alumina by CVD-4008C
ð8Þ
Apeak ethylene
ethanol retained after purge ꢃ
Aethylene þ Apeak ethanol
The process of grafting differs in the case of CVD-4008C. De-
spite the similar coverage of the alumina surface to that of
SA(17/anh), SA(17/4eqW), or SA(17/CVD250), shown by the
amount of ethanol bound to the surface (Figure 3), and consis-
tent with our previous data,[19] SA(17/CVD400) (Table 6, entry 9)
has a smaller number of BAS and strong LAS. The IR spectrum
shows a nonpreferential decrease in the alumina OH bands
(Figure 6); the position of the ethanol dehydration peak is simi-
lar (although less intense) to that of g-alumina. This indicates
that, at high grafting temperature, the facets of g-alumina are
nonselectively grafted by the silicon species. By calculating the
percentage of the free (100) surface (Table 6), we confirm that
the value (26%) is close to the overall percentage of the free
alumina surface (35%). The remaining sites on the (100) sur-
face allow the dehydration of ethanol in a similar way to that
on g-alumina.
in which Apeak i is the area of the peak corresponding to the desorp-
tion of species i when calculating the derivative of the weight loss
of ethanol during TPD.
The m-xylene isomerization test consisted of reacting a catalyst
bed of 0.5 g with gaseous m-xylene (0.6 cm3 hꢀ1). Catalysts were
first preheated in an air flow at 3508C, followed by xylene conver-
sion at 3508C in a flow of nitrogen. Analysis of the products was
performed by on-line gas chromatography by means of a Hewlett
Packard 6850 apparatus, equipped with a flame ionization detector
(FID) and an FFAP column. m-Xylene was isomerized to p- and o-
xylene or disproportionated to toluene and isomers of trimethyl-
benzene. In the latter case, the reaction of trimethylbenzene with
toluene or xylene may also lead to the formation of isomers of
xylene by a bimolecular mechanism (Supporting Information, Fig-
ure S3). The rate of m-xylene conversion measured after the first
10 min of reaction is given in Equation (9):
Rate ½molecule hꢀ1 nmꢀ2ꢂ ¼
Conclusions
species
½
=ð species
þ
Þꢂ
X
X
Am-xylene
Ai
Ai
NC atoms i
NC atoms i
8
ð9Þ
Silicon grafting on g-alumina takes place preferentially on the
(100) surface if grafting occurs under mild conditions. When
this whole surface is covered, the (110) surface is grafted. At
almost complete coverage of the (110) surface, silicon is graft-
ed on the previously deposited silicon species. Lewis acid sites
(LAS) on the uncovered g-alumina (100) surface are the active
sites for ethanol dehydration; these sites are hydroxylated or
dehydroxylated forms of AlVIꢀOH or (AlVI)3ꢀOH. Brønsted acid
sites (BAS) and strong LAS are created by silicon grafting on
the (110) surface. These BAS are responsible for m-xylene iso-
merization and are also active in the dehydration of ethanol.
i
i
flow rate ꢃ densitym-xylene
Mm-xylene ꢃ mcatalyst
NA
ꢃ
ꢃ
SBET
in which Ai is the area of the GC peak of species i (i=toluene, p-
xylene, o-xylene, trimethylbenzenes) (6.02ꢀ1023 molꢀ1), flow rate is
in [cm3 hꢀ1], NA is the Avogadro number, and SBET in [nm2 gꢀ1] the
surface area of the catalyst.
IR spectra were recorded by a Nexus Fourier transform instrument.
Samples (self-supported pellets of compressed powder) were pre-
treated in situ under vacuum (ꢁ10ꢀ5 mbar) for 10 h at 4508C
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 3644 – 3656 3655