Catalytic performance of nanoscopic, aluminium trifluoride-based catalysts in the synthesis of (all-rac)-α-tocopherol

Article abstract of DOI:10.1002/adsc.200800411

A novel nanoscopic, partly hydroxylated aluminium fluoride was prepared in a sol-gel fluorination synthesis and characterised by IR probe molecules. It was shown that this material has, in contrast to high surface area aluminium fluoride (HSAlF₃) which is one of the strongest Lewis acids, a low amount of strong Lewis acid sites combined with weak and medium strength Br?nsted sites. Both materials were tested in the synthesis of (all-rac)-α-tocopherol through the condensation of 2,3,6- trimethylhydroquinone (TMHQ) with isophytol (IP). The activity data indicate the partly hydroxylated aluminium fluoride to be a very efficient and highly selective catalyst for (all-rac)-α-tocopherol (>99.9%, for an IP conversion of 100%) whereas high surface area aluminium fluoride is a poor catalyst for this reaction.

Full text of DOI:10.1002/adsc.200800411

COMMUNICATIONS
DOI: 10.1002/adsc.200800411
Catalytic Performance of Nanoscopic, Aluminium Trifluoride-
Based Catalysts in the Synthesis of (all-rac)-a-Tocopherol
S. M. Coman,a,b S. Wuttke,b A. Vimont,c M. Daturi,c and E. Kemnitzb,*
a
University of Bucharest, Faculty of Chemistry, Department of Chemical Technology and Catalysis, Bdul Regina Elisabeta
4-12, 030016 Bucharest, Romania
Humboldt-Universitꢀt zu Berlin, Institut fꢁr Chemie, Brook-Taylor-Straße 2, 12489 Berlin, Germany
b
Fax : (+49)-30-2093-7277; e-mail: erhard.kemnitz@chemie.hu-berlin.de
Laboratoire Catalyse et Spectrochimie, UMR 6506, CNRS-ENSICAEN – Universitꢂ de Caen, Boulevard du Marꢂchal
c
Juin, F-14050 Caen Cedex, France
Received: July 3, 2008; Revised: September 16, 2008; Published online: October 16, 2008
Abstract: A novel nanoscopic, partly hydroxylated
aluminium fluoride was prepared in a sol-gel fluori-
nation synthesis and characterised by IR probe mol-
ecules. It was shown that this material has, in con-
trast to high surface area aluminium fluoride (HS-
AlF3) which is one of the strongest Lewis acids, a
low amount of strong Lewis acid sites combined
with weak and medium strength Brønsted sites.
Both materials were tested in the synthesis of (all-
rac)-a-tocopherol through the condensation of
2,3,6-trimethylhydroquinone (TMHQ) with isophy-
tol (IP). The activity data indicate the partly hy-
droxylated aluminium fluoride to be a very efficient
and highly selective catalyst for (all-rac)-a-toco-
pherol (>99.9%, for an IP conversion of 100%)
whereas high surface area aluminium fluoride is a
poor catalyst for this reaction.
flates) catalysts – in biphasic green type systems[2]
have been published. These new catalysts have the
advantage of simple separation from the reaction
mass, the absence of wash-water containing the dead
catalyst and a high purity of (all-rac)-a-tocopherol.
On the other hand, it was recently shown that the un-
wanted benzofuran regioisomers can be successfully
avoided either by using tailored catalysts in which the
active site units are bonded to bulkier polymeric
anions[5e] or by using catalysts in which one of the
functions is to modulate the phytyl carbocation equili-
brium.[5c]
However, in many cases the high purity of the main
product is not only due to the merits of the catalyst
itself but also of the experimental approach, which
unfortunately sometimes becomes complicated, and
therefore less attractive for practical applications.[2,6]
The development of a more effective and practical
catalytic system for an industrial process, that is able
to fulfil all the features needed to avoid the formation
of the by-products but that does not affect the catalyst
performances, is still much sought after. In this re-
spect we recently discovered the unexpected catalytic
perfomances of nanoscopic hydroxylated aluminium
fluoride prepared by a novel sol-gel fluorination syn-
Keywords: Brønsted acidity; heterogeneous cataly-
sis; Lewis acidity; nanoscopic aluminium fluoride;
vitamins
The industrial-scale synthesis of (all-rac)-a-tocopherol thesis.[7] Here we report on this new catalyst, its cata-
through the condensation of 2,3,6-trimethylhydroqui- lytic performances in the synthesis of (all-rac)-a-toco-
none (TMHQ) with isophytol (IP) comes up against pherol, and the comparison to pure but highly Lewis
corrosion caused by acidic media, contamination of acidic aluminium fluoride.
the waste-water with acids, and the difficult purifica-
Applying the non-aqueous sol-gel route for inor-
tion of the main product.[1] For example, the forma- ganic fluorides results in the formation of very strong
tion of regioisomeric benzofuran derivatives as by- Lewis acid sites on high surface area aluminium fluo-
products, which are very difficult to separate from the ride (HS-AlF3), comparable to those of SbF5.[8] The
main product, has been frequently reported.[2] Then, HS-AlF3 synthesis consists of two consecutive steps: i)
IP easily dehydrates in the presence of acids[3], form- the sol-gel reaction of aluminium isopropoxide and
ing phytadienes. Last but not least, it is known that non-aqueous HF in dry isopropanol, and ii) the post-
traces of oxygen in the reaction system can lead to fluorination of the resulting xerogel with CHClF2 or
the formation of quinone compounds.[4] Therefore HF.[9] By adding a stoichiometric amount of HF in the
data concerning the use of active and selective hetero- form of its aqueous solution not only Lewis but also
geneous[5] or homogeneous (e.g., rare earth metal tri- Brønsted acid sites can be introduced in the solid cat-
Adv. Synth. Catal. 2008, 350, 2517 – 2524
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2517

S. M. Coman et al.
COMMUNICATIONS
ꢀ
alyst because of partial hydrolysis of the M OR bond tocopherol. The main by-product of the reaction is
resulting in M OH bond formation [Eq. (1)]. The re- phytadiene and, when using an active catalyst like
ꢀ
action of the metal alkoxides with HF results in an AlF3-40, formation of considerable amounts of this
ꢀ
M F bond formation (fluorolysis) and consequently by-product is favoured.
generates Lewis acidity (under coordinated Mn+
For each catalyst, increasing the TMHQ/IP molar
sites). Schematically, the particular synthesis of the ratio from 1/1 to 2/1 leads to a higher yield to toco-
partly hydroxylated aluminium fluoride using this ap- pherol by decreasing the formed by-product amounts.
proach can be depicted as in Eq. (1).
This effect was accompanied by a shift of the maxi-
mum yield from the AlF3-70 sample (in the case of
TMHQ/IP=1/1) to the AlF3-50 sample. Therefore, it
is possible that the excess of TMHQ not only blocks
the dehydration of IP to phytadienes but also regu-
lates the Lewis/Brønsted acid sites ratio on the cata-
lysts surface, by a competitive chemisorption of its
Brønsted sites.
By using different molar ratios of HF to H2O
Since the best catalytic results in the (all-rac)-a-to-
during the sol-gel synthesis different partly hydroxy- copherol synthesis were obtained for the sample pre-
lated aluminium fluorides were obtained and tested in pared with 50% wt aqueous HF (AlF3-50), this
the synthesis of (all-rac)-a-tocopherol. Noteworthy sample was further characterised and compared with
results are shown in Table 1.
the pure Lewis HS-AlF3 sample.
As Table 1 shows, all partly hydroxylated alumini-
HS-AlF3 and AlF3-50 are X-ray amorphous and ex-
um fluorides samples are very active catalysts, the hibit high surface areas connected with mesoporous
conversion of IP being 100% in 1 h of reaction. At pores: 190 m2 g (Dp =65 ꢄ) for AlF3-50 and 250 m2 g
the same time the distribution of the reaction prod- (Dp =68 ꢄ) for HS-AlF3. Nevertheless, the acidic
ucts is different as a function of the aqueous HF solu- properties, for example, the nature, the strengths and
tion concentration used during the catalyst prepara- the densities of the acidic sites, are quite different de-
tion or, in other words – as a function of the Lewis/ pending on the preparation method.
Brønsted acid site ratio. Therefore, for a ratio of
Recently, using CO-IR spectroscopy, we showed
THMQ/IP=1/1 the yield to tocopherol decreases in that pure Lewis acid HS-AlF3 displays a large number
the order: AlF3-70>AlF3-60>AlF3-50>AlF3-40>> of very strong Lewis acidic centres and just a few
>HS-AlF3 (Y=0%). Accordingly with these results, weak Lewis centres at the surface.[8b] Furthermore,
it seems that the synthesis needs a certain Lewis/ water exposed to this material at room temperature
Brønsted acid site distribution, with certain strengths. has been found to adsorp non-dissociatively at the
However, in the case of the AlF3-40 sample (Table 1, surface forming layers with strong hydrogen bonds to
entry 8), the number of the Brønsted acid sites is too other water molecules. In this way the very strong
high and responsible for the decreased selectivity to Lewis acid sites become blocked by water, and some
Brønsted acid sites with moderate strength are gener-
ated. The CO experiments clearly show that these
OH groups are situated at the surface of the solid and
do not destroy the bulk structure at ambient tempera-
ture. However, heating of HS-AlF3 up to 573 K under
Table 1. The influence of the different aluminium fluorides
samples on the yield to (all-rac)-a-tocopherol.[a]
a certain partial pressure of water results in partial
hydrolysis.
Catalyst
Al:HF:H2O
molar ratio
TMHQ/IP
molar ratio
Yield to
tocopherol [%]
CO-IR spectra of the novel AlF3-50 sample
(Figure 1) show two bands: an intensive band at
2172 cmꢀ1 (assigned to Brønsted acid sites) and a
band at 2207 cmꢀ1 (assigned to the Lewis acid sites)
that only developed as a small feature. Therefore, the
Lewis/Brønsted ratio is quite small. On the other
hand, the lutidine-IR spectra of the AlF3-50
(Figure 2) display three bands at 1581 cmꢀ1 (physisorb-
[b]
HS-AlF3
AlF3-70
AlF3-70
AlF3-60
AlF3-60
AlF3-50
AlF3-50
AlF3-40
AlF3-40
1:3:0
1/1
1/1
2/1
1/1
2/1
1/1
2/1
1/1
2/1
0
1:3:1.3
1:3:1.3
1:3:2
90.1
94.3
89.3
91.7
84.7
>99.9
71.4
93.5
1:3:2
1:3:3.6
1:3:3.6
1:3:5
AHCTUNGTRENNUNG
ed species), 1595 cmꢀ1 (assigned to the n8a mode) and
1:3:5
1604 cmꢀ1 (assigned to the n8b mode), corresponding
to weakly coordinated species. In addition, the spectra
display another two bands at 1655 cmꢀ1 and 1631 cmꢀ1
(assigned to the n8a mode and n8b mode, respectively)
corresponding to protonated species, which are only
[a]
Reaction conditions: 152 mg or 304 mg TMHQ+0.4 mL
IP (1/1 or 2/1 molar ratio), 6 mL solvent [propylene car-
bonate/heptane=50/50 (v/v)], 50 mg of catalyst, 1008C,
reaction time=1 h, C=100%.
[b]
C=70%.
2518
asc.wiley-vch.de
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2517 – 2524

Catalytic Performance of Nanoscopic, Aluminium Trifluoride-Based Catalysts
COMMUNICATIONS
Figure 1. IR spectra of AlF3-50 recorded at 100 K after the adsorption of increasing CO doses (a!g).
Figure 2. IR difference spectra after adsorbtion of lutidine on AlF3-50 sample which was before activated at 373 K (a) equi-
librium pressure, and evacuated (b) at room temperature, (c) 323 K and (d) 373 K.
observed for medium and/or strong Brønsted acid sample can be represented schematically and in a
sites.[10]
very simplified manner as shown in Scheme 1.
The strength of these sites also becomes evident
during evacuation of the sample at different tempera-
tures and the persistency of the protonated species in
comparison with the coordinated species (Figure 2).
Although it cannot be clearly proven, these sites are
probably due to surface OH groups which are formed
during the synthesis. It is worth mentioning that such
surface hydroxy groups are observed and are well es-
tablished for aluminas.[11–13]
Scheme 1. Here, the symbol (&) indicates a vacancy (Lewis)
site – Al undercoordinated and accessible site for a nucleo-
Based on the CO- and lutidine-IR spectroscopic re-
sults, the acidic species on the surface of the AlF3-50 philic reactant.
Adv. Synth. Catal. 2008, 350, 2517 – 2524
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2519

S. M. Coman et al.
COMMUNICATIONS
Scheme 2. The synthesis product distribution by the condensation of TMHQ 1 and IP 2 with HS-AlF3 and AlF3-50 (X=con-
version, Y=yield). Reaction conditions: 152 mg TMHQ+0.4 mL IP (1/1 molar ratio), 6 mL solvent [propylene carbonate/
heptane=50/50 (v/v)], 1008C, reaction time=1 h.
As we have shown, when employing HS-AlF3 and 100% in less than 1 h), the activity of the pure but
AlF3-50 as catalysts in the synthesis of (all-rac)-a-to- strong Lewis acidic HS-AlF3 sample was rather low
copherol through the condensation of 2,3,6-trimethyl- (70% conversion of IP after 1 h). Moreover, while the
hydroquinone (TMHQ, 1) with isophytol (IP, 2), the AlF3-50 sample was very selective to (all-rac)-a-toco-
conversion of IP and the product distribution are pherol, in the presence of the HS-AlF3 sample the
completely different depending on the nature of the synthesis stops at the phytylhydroquinone intermedi-
catalyst (Scheme 2). Therefore, in comparison with ate (PHQ, 3). On the other hand, even though only in
the very active AlF3-50 sample (conversion of IP was small amounts, by-products were also identified in the
2520
asc.wiley-vch.de
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2517 – 2524

Catalytic Performance of Nanoscopic, Aluminium Trifluoride-Based Catalysts
COMMUNICATIONS
reaction product. Some of the observed intermediates condensation of TMHQ with IP, the availability and
and by-products, which are also known from litera- proportion of the different carbocations, for example,
ture,[14] are presented in Scheme 2. To the best of our i-phytil, n-phytil, formed by the interaction of IP with
knowledge, the quinone 7 was not claimed until now the catalyst, seems to be the key step for the high re-
in the literature for this synthesis. To elucidate its gioselectivity to the desired (all-rac)-a-tocopherol.
structure, after the product separation, the compound
The AlF3-50 sample is highly selective towards the
was characterised by 2D-NMR spectroscopy (1H-1H formation of (all-rac)-a-tocopherol which is a result
COSY NMR and 1H-13C HSQC NMR techniques). of the medium Brønsted acids sites formed already
Both spectra are depicted in Figure 3.
during the synthesis of the catalyst. Moreover, the
The fact that the AlF3-50 sample catalyses the con- catalytic leaching tests did not show any solubility,
version of IP faster than HS-AlF3 is a clear indication meaning it behaves as a real heterogeneous catalyst
that the pure strong Lewis acid sites are not essential under these reaction conditions. Drying the used cata-
in the first step. It is well known that, in Friedel– lysts in N2 flow, at 1008C for 12 h made possible the
Crafts alkylations using alcohols as alkylating reagent, full restoration of its initial activity and selectivity.
pure Lewis acids show only low or no activity if they
As we showed, further optimisation of the reaction
are not activated by the inadvertent presence or addi- conditions in presence of the AlF3-50 sample im-
tion of small concentrations of co-catalysts or promot- proved the yield to (all-rac)-a-tocopherol. Therefore,
ers, such as water. These co-catalysts interact with the with an excess of TMHQ (TMHQ/IP=2/1, molar
Lewis acids to generate Brønsted sites.[15] Focusing on ratio) it was possible to obtain (all-rac)-a-tocopherol
the HS-AlF3 sample, if exposed to moisture, this ma- with >99.9% yield after only 1 h of reaction
terial was found to adsorb water non-dissociatively at (Table 1). In order to check the reproducibility of
room temperature and to form some Brønsted sites these results on a larger scale, the synthesis was per-
with moderate strength.[8b] The material was handled formed in a glass vial with a capacity of 500 mL in-
in an air atmosphere for a short time, therefore, it is stead of 8 mL and using 62,5 mmol of TMHQ and IP.
very probable that some water had already been ad- This experiment yielded almost identical results.
sorbed from the atmosphere and generated some
In summary, we can conclude that the novel fluori-
Brønsted acidity. Additionally, each molecule of phy- nation sol-gel method has been successfully applied to
tylhydroquinone (PHQ, 3) formed generates one mol- create reproducible a ꢅbi-acidicꢆ Lewis/Brønsted nano-
ecule of water, and this may also be adsorbed non-dis- scopic aluminium fluoride, by altering the structure of
sociatively on the Lewis sites provoking supplementa- conventional pure Lewis acid aluminium fluoride
ry Brønsted acid sites. Therefore moisture from air (HS-AlF3) to a partly hydroxylated one. The compari-
and the generated water may act as a co-catalyst in son of both materials in the condensation of 2,3,6-tri-
the first step of the condensation process.
methylhydroquinone (TMHQ) with isophytol (IP)
In contrast, AlF3-50 already contains surface OH showed crucial differences in the reaction mixture
groups representing medium Brønsted acid sites which are the result of the different acid sites (Lewis
which obviously catalyse the first step of the synthesis vs. Brønsted) of these materials. Although the surface
of (all-rac)-a-tocopherol.
structure/morphology of these catalytically active sys-
In the case of HS-AlF3, the reaction stops at a low tems have not been resolved with IR experiments so
degree of conversion but provides a high selectivity to far, it can be stated that the difference in the catalytic
PHQ, 3, the Lewis and the created Brønsted acid sites properties can either be due to the different origin of
ꢀ
of this material are unable to transform PHQ into Brønsted sites (coordinated water vs. Al OH species)
(all-rac)-a-tocopherol. Furthermore, the very strong or to the strength of the Brønsted sites present on the
acidic centres of HS-AlF3 obviously adsorb some at- samples. In order to be more certain about these
mospheric oxygen on the surface if the material is points, a deeper investigation of both the solids and
handled in air resulting in the formation of quinones the reaction itself are required.
6 and 7.
At this stage, it is evident that the high activity in
Surprisingly, irrespective of the nature of the cata- combination with an exciting high selectivity is obvi-
lyst, the synthesis took place with a total regioselec- ously the synergistic effect of the presence of both
tivity as long as no benzofuran derivatives, 5 were de- Brønsted and Lewis sites. However, there are several
tected in the reaction products. Generally speaking, points which still need a more comprehensive investi-
the regioselectivity of a catalytic reaction depends on gation. Thus, it might be speculated what additional
the structure of the molecule, on the structure of the impact orginate from these surface Brønsted and
catalyst, and on the reaction conditions, these factors Lewis sites: i) phytil carbocations may be formed at
being highly interdependent. In other words, a com- Brønsted sites in the first reaction step, (ii) reaction
prehensive interpretation of regioselectivity requires water may be boundd at Lewis sites, thus acting as de-
that the substrate-catalyst couple should be consid- hydrating sites, and (iii) not only the presence of sur-
ered as a supramolecular system. Particularly in the face Brønsted sites of optimum medium strength
Adv. Synth. Catal. 2008, 350, 2517 – 2524
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2521

S. M. Coman et al.
COMMUNICATIONS
Figure 3. The 2D-NMR spectra (a. H-1H COSY NMR; b. H-13C HSQC NMR) for quinone 7 [eluent: CH2Cl2 1H NMR: d=
1
1
5.27 ppm (s, 2H)].
2522
asc.wiley-vch.de
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2517 – 2524

Catalytic Performance of Nanoscopic, Aluminium Trifluoride-Based Catalysts
COMMUNICATIONS
After the reaction, the catalyst was separated from the two-
phase solvent mixture, the heptane phase (containing the to-
copherol) was separated from the propylene carbonate
phase and the solvent was removed under vacuum to give a
crude product.
might be beneficial but also a certain topological ar-
rangement at these catalysts may provide a high selec-
tivity towards formation of the desired phytyl carbo-
cations resulting in high regioselectivity to the desired
(all-rac)-a-tocopherol. All these are topics of ongoing
investigations.
In any case, the catalytic efficiency combined with
practical advantages (e.g., no external inert atmos-
phere and no azeotropic removal of water) of the cat-
alytic experimental set-up make this catalyst system
appealing even in the commercial stage.
The crude product was analysed by HPLC (column – EC
125/4.6 Nucleosil 120–5 C18; eluent: acetonitrile; flow rate:
0.8 mLminꢀ1
; wavelength: 280 nm; volume of sample:
15 mL) chromatography and H, 13C NMR (Bruker AV 400
spectrometer, in CDCl3 solvent and Me4Si as internal stan-
dard) spectroscopy. (all-rac)-a-Tocopherol: 1H NMR
(400 MHz, CDCl3): d=0.90 (m, 12H, CH3), 1.0–1.9 (m,
26H, CH3, CH2, CH), 2.10 and 2.17 (both s, 9H, CH3-Ar),
2.60 (t, 2H, C4H2, J (H, H)=6.7 Hz), 4.3 (s, 1H, OH).
For compound characterisation the crude product was
separated by column chromatography on silica gel (eluent:
CH2Cl2) and analysed by 1H-1H COSY NMR and 1H-13C
HSQC NMR (400 MHz, CDCl3).
1
Experimental Section
Catalyst Preparation
For comparison, similar tests were done in the presence
Nanoscopic partly hydroxylated aluminium fluoride was pre-
pared from aluminium isopropoxide using a one-pot sol-gel
fluorination method as follows: 9.73 g (48 mmol) of com-
mercially available aluminium isopropoxide (Aldrich, 98+
%) were dissolved in 300 mL isopropyl alcohol and reacted
with a stoichiometric amount of approximately 40%, 50%,
60% or 70% hydrofluoric acid and the products are named
in Table 1 as AlF3-40, AlF3-50, AlF3-60 and AlF3-70. After
aging for 12 h, the resulting gel was dried under vacuum at
room temperature. The formed solid product was then dried
further under vacuum at 708C, for 5 h and used as obtained
in this way.
of HS-AlF3 prepared through
a two-step fluorination
method. The preparation of the sample is described in detail
elsewhere.[7]
Trimethylhydroquinone (97 wt%) and isophytol (95 wt%)
were purchased from Acros Organics. The other reagents
(analytical grade) were obtained from Merck.
Acknowledgements
Dr. S. M. Coman is a fellow of Alexander von Humboldt
Foundation. The authors kindly acknowledge I. Hartwich
and I. Buchem for preparing the catalysts.
Catalyst Characterisation
The XRD pattern was recorded with an XRD-7 Seiffert-
FPM diffractometer equipped with a Cu-Ka radiation
source. Textural properties were determined using N2 at
77 K for adsorption by means of a Micromeritics ASAP
2020 instrument. IR spectra were recorded on a Nicolet
Magna 550 spectrometer equipped with a liquid nitrogen
cooled MCT detector in transmission mode and purged by a
dry nitrogen flow. The resolution was 4 cmꢀ1 and 128 scans
were co-added for each spectrum. The cell for IR analysis
consisted of vitreous silica and was equipped with CaF2 win-
dows. The cell was connected to a vacuum line and a glass
injection loop with a known volume (2.15 cm3). CO (>
99.997% pure; supplied by Air Liquide, France) was provid-
ed from a gas balloon, dried with liquid N2, and dosed step-
wise on the sample. The detailed procedures of NH3-TPD
are given elsewere.[7]
References
[1] T. Netscher, Vitamins and Hormones 1996, 76, 155.
[2] W. Bonrath, C. Dittel, L. Giraudi, T. Netscher, T.
Pabst, Cat. Today 2007, 121, 65.
[3] M. B. Tarabrin, G. G. Shestakov, R. P. Evstigneeva, E.
Yu. Bulychev, Pharm. Chem. J. 1984, 18, 572.
[4] a) B. O. Linn, C. H. Shunk, E. L. Wong, K. Folkers, J.
Am. Chem. Soc. 1963, 85, 239; b) J. Greene, D.
McHale, J. Chem. Soc. 1965, 5060; c) K. A. Parker,
T. L. Mindt, Org. Lett. 2001, 3, 3875.
[5] a) F. Schager, W. Bonrath, J. Catal. 1999, 1, 282;
b) R. F. Childs, A. Mika-Gibala, J. Org. Chem. 1982,
47, 4207; c) A. Heidekum, M. A. Harmer, W. F. Hçlder-
ich, J. Catal. 1998, 176, 260; d) S. M. Coman, G. Pop, C.
Stere, V. I. Parvulescu, J. El Haskouri, D. Beltrꢈn, P.
Amorꢉs, J. Catal. 2007, 251, 388; e) A. Hasegawa, K.
Ishihara, H. Yamamoto, Angew. Chem. Int. Ed. 2003,
42, 5731.
[6] K. Ishihara. M. Kubota, H. Yamamoto, Synlett 1996,
1045.
[7] E. Kemnitz, S. M. Coman, S. Rꢁdiger, S. Wuttke, Eur.
Pat. Appl. EP 07 020 498.7, 2007.
[8] a) E. Kemnitz, U. Groß, St. Rꢁdiger, S. C. Shekar,
Angew. Chem. Int. Ed. 2003, 42, 4251; b) T. Krahl, A.
Vimont, G. Eltanany, M. Daturi, E. Kemnitz, J. Phys.
Chem. C 2007, 111, 18317.
Synthesis of (all-rac)-a-Tocopherol
In a typical procedure, 152 mg (1 mmol) TMHQ, 304 mg
(2 mmol) TMHQ or 9.5 g (62.5 mmol) TMHQ was dissolved
in 6 mL or 375 mL of solvent [heptane:propylene carbonate
(50:50)] in a glass vial (standard capacity of 8 mL or
500 mL) equipped with a magnetic stirrer. To this mixture
0.4 mL (1 mmol) or 25 mL (62.5 mmol) of IP and 50 mg
(20.1ꢇ1017 acid centres/m2 – AlF3-H; 18.4ꢇ1017 acid centres/
m2 – HS- AlF3 calculated from NH3-TPD) or 3.1 g of AlF3-
H of catalyst were added. After this, the vial was closed, im-
mersed in an oil bath with a temperature of 1008C, and the
charged mixture was stirred (1250 rpm) for 30–1200 min.
Adv. Synth. Catal. 2008, 350, 2517 – 2524
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2523

S. M. Coman et al.
COMMUNICATIONS
[9] S. Rꢁdiger, G. Eltanany, U. Groß, E. Kemnitz, J. Sol-
Gel Sci. Techn. 2006, 41, 299.
[10] L. Oliviero, A. Vimont, J.-C. Lavalley, F. R. Sarria, M.
Gaillard, F. Maugꢂ, Phys. Chem. Chem. Phys. 2005, 7,
1861.
[11] H. Knçzinger; P. Ratnasamy, Catal. Rev. Sci. Eng. 1978,
17, 31.
[12] C. Morterra, G. Magnacca, Cat. Today 1996, 27, 497.
[13] D. T. Lundie, A. R. McInroy, R. Marshall, J. M. Win-
field, P. Jones, C. C. Dudman, S. F. Parker, C. Mitchell,
D. Lennon, J. Phys. Chem. B 2005, 109, 11592.
[14] S. Wang, W. Bonrath, H. Pauling, F. Kienzle, J. Super-
critical Fluids 2000, 17, 135, and references cited there-
in.
[15] M. C. Clark, C. M. Smith, D. L. Stern, J. S. Beck, in:
Handbook of Heterogeneous Catalysis. Second, Com-
pletely Revised and Enlarged Edition, (Eds.: G. Ertl, H.
Knçzinger, K. Schꢁth, J. Weitkamp), Vol. 7, Wiley-
VCH, Weinheim, 2008.
2524
asc.wiley-vch.de
ꢃ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2517 – 2524