H/D exchange, protolysis and oxidation of C3-C5 alkanes in HF-SbF5. σBasicity vs. reactivity of C-H bonds

Article abstract of DOI:10.1039/b201437h

The relative reactivity and basicity of CH bonds in C₃-C₅ alkanes was studied using the strongest deuterated superacid, DF-SbF₅, at various concentrations of SbF₅. In all cases, at concentrations up to approxima

Full text of DOI:10.1039/b201437h

View Article Online / Journal Homepage / Table of Contents for this issue
H/D exchange, protolysis and oxidation of C3–C5 alkanes in
HF–SbF5 . r-Basicity vs. reactivity of C–H bonds
Alain Goeppert and Jean Sommer*
Laboratoire de Physico-Chimie des Hydrocarbures (CNRS UMR 7513), Institut de Chimie,
´
Universite Louis Pasteur, 4 rue Blaise Pascal, F-67070, Strasbourg cedex, France.
Tel: +33 (0)3 90 24 14 85;Fax: +33(0)3 90 24 14 87E-mail: Sommer@chimie.u-strasbg.fr
Received (in Montpellier, France) 7th February 2002, Accepted 26th March 2002
First published as an Advance Article on the web 16th September 2002
The relative reactivity and basicity of CH bonds in C3–C5 alkanes was studied using the strongest deuterated
superacid, DF–SbF5 , at various concentrations of SbF5 . In all cases, at concentrations up to approximatively
18 mol % SbF5 in DF(HF), the predominant reaction, beside minor protolytic cleavage, was rapid isotope
exchange of all C–H bonds via reversible protonation. The relative rate of exchange follows the Olah s-basicity
concept: tertiary > secondary > primary C–H bonds independently of the further reactivity of these bonds. At
higher concentrations of SbF5 the exchange process gives way to increasing ionization of the alkane, first via
protolysis and later via an oxidative process with concomitant reduction of SbF5 .
Acid-catalyzed transformation of hydrocarbons, such as
cracking, isomerization and alkylation, are large-scale indus-
trial processes using solid or strong liquid acids, such as H-zeo-
lites, chlorinated aluminas, sulfuric acid, and hydrogen
fluoride.1,2 The high acidity and/or high temperature compen-
sate the well-known inertness of the starting material. In fact,
since the late 1960s, when Hogeveen3 and Olah4 published
their pioneering work, it is known that saturated hydrocarbons
do react, even at room temperature and below, with various
liquid superacids.5 Now, it is widely accepted that the key step
in these reactions is the formation of reactive carbocations.
The initial step is often ascribed6 to proton attack on a C–H
or C–C bond, following the s-basicity concept developed by
Olah.7 However, protonated alkanes, characterized by a
three-center-two-electron bonded structure with pentacoordi-
nated carbon atoms,8 generally have a lifetime too short to
allow direct observation in superacidic media by NMR. Never-
theless, initial product distribution and H/D exchange studies
suggest the existence of these cations as reaction intermediates
or transition states.8 The theoretical approach of alkane proto-
nation by Mota and colleagues using ab initio and DFT meth-
ods9 seems to indicate that protonated alkanes are not
minimas on the potential energy surface. This was also con-
firmed by experimental and theoretical studies of reversible
methane protonation by Ahlberg et al.10
In contrast, H/D exchange in weaker acids such as
or D2O-exchanged solid acids19 takes place only
in the methyl positions via successive deprotonation of the t-
butyl cation and reprotonation of isobutene. In triflic acid, a
weak superacid with H0 ¼ ꢀ14.1 the same regiospecificity in
H/D exchange was observed.20
Direct hydron exchange via formation of carbonium ions
occurs only in the very strong superacid system DF–SbF5 at
low temperature.
To compare directly the relative basicity vs. reactivity of
each kind of C–H bond, we found it of interest to compare
the H/D exchange and ionization of a series of small alkanes
varying from propane to isopentane in DF–SbF5 , isopentane
being the simplest alkane containing primary, secondary as
well as tertiary C–H bonds. We present here our results based
on 1H, 2H NMR and GC analysis including earlier results
obtained in the case of isobutane and n-butane. In order to
prevent hydride transfers or skeletal rearrangement in carbe-
nium ion intermediates, all reactions were conducted in the
presence of excess carbon monoxide, which converts these ions
into oxocarbenium ions, stable in this superacid under our
experimental conditions.21
16–18
D2SO4
Result and discussion
For alkane activation studies, isobutane has often been used
as a convenient probe molecule as the higher reactivity of its ter-
tiary C–H bond leads to very simple product distribution even
under superacidic conditions.11–13 The groups of both Hogev-
een14 and Olah15 have observed protium-deuterium exchange
of the methine hydrogen during isobutane ionization in a
HF–SbF5–freon system and in ‘‘magic acid’’ at ꢀ78 ꢁC, respec-
tively, and both authors concluded that a substantial exchange
of this type of hydrogen had occurred compared to ionization.
No exchange was observed in the methyl group. More recently,
however, we showed that in DF–SbF5 the protonation site on
isobutane is independent of the further reactivity of the penta-
coordinated carbonium ion. It takes place on both types of
C–H bonds in correlation with the basicity:12 tertiary
C–H > primary C–H. Beside rapid hydron exchange, partial
C–C and C–H bond cleavage also occur, leading to ionization.
Reactivity of propane in DF–SbF5 superacids
Propane has only primary and secondary C–H bonds and
reacts in HF–SbF5 in the presence of CO to give, besides
hydrogen, methane and ethane, isopropyloxocarbenium and
ethyloxocarbenium ions.21 Reacting propane with DF–SbF5
(Scheme 1) leads to protonation of the C–H and C–C bonds
with formation of carbonium ions 1 to 3. C–C bond protolysis
in species 3 leads to the formation of deuterated methane and
ethyl cation, which can undergo hydride transfer with a pro-
pane molecule. Regioselective deuteration in the primary and
secondary positions is respectively obtained with carbonium
ions 1 and 2. In DF–SbF5 mixtures acidity is directly corre-
lated with the SbF5 concentration.22 The product distribution,
especially the ratio isotope exchange/conversion of propane
DOI: 10.1039/b201437h
New J. Chem., 2002, 26, 1335–1339
1335
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

View Article Online
SbF5 . Ionization was measured by quenching the oxocarbe-
nium ions with excess ethanol and neutralizing.
Reactivity of isopentane in DF–SbF5 superacids
The exchange and protolysis pathways for isopentane in DF–
SbF5 are described in Scheme 2. C–H and C–C bonds are pro-
tonated to form the very reactive carbonium ions 4 to 10. C–H
bond protonation leads to species 4 to 7, which yield either
pentylium ions by loss of HD or H/D exchanged isopentane.
Deuteration in the primary positions is obtained via carbo-
nium ions 5 and 7. Secondary and tertiary hydrons are
exchanged via carbonium ions 6 and 4, respectively. Deutero-
lysis of a C–C bond in species 8 to 10 leads to the formation of
deuterated methane and 2-butylium ion with carbonium ion 8
whereas deuterated ethane and 2-propylium ion are generated
via carbonium ion 9. Alternative pathways leading to very
energetic primary carbocations are much slower and generally
negligible when other pathways are available.
Scheme 1 H/D exchange and ionization of propane in DF–SbF5 .
As it was observed with propane, the product distribution
depends on the acidity of the media. Fig. 2 shows the conver-
sion of isopentane as a function of the concentration of SbF5 .
A strong increase in conversion was noted when the concentra-
tion of the Lewis acid increased. At the same time hydrogen
formation was measured (Fig. 2) and it appears that for up
to 20 mol % SbF5 in DF–SbF5 , ionization and hydrogen pro-
duction are parallel in accord with a protolytic ionization
pathway. For higher SbF5 concentrations, however, a change
in the product distribution was noticed: conversion was still
increasing whereas hydrogen production decreased drastically.
The same phenomenon was observed for isobutane12 and can
be rationalized by the presence of uncomplexed SbF5 in HF at
concentrations higher than 18 mol % SbF5 in HF as demon-
strated by 19F NMR studies.23 This free SbF5 can be reduced
during isopentane conversion, producing SbF3 and HF as sug-
gested earlier by Olah,5,7 forming directly carbenium ions by
an oxidative process.
Fig. 3 shows the deuteration expressed as atom % of pri-
mary, secondary and tertiary C–H bonds in isopentane reacted
with DF–SbF5 as a function of acidity. The shape of the curves
is similar to the one described above for propane or reported
earlier for isobutane in the same acid,12 but gives more infor-
mation. At SbF5 concentrations up to 15 mol %, H/D
exchange in all positions is increasing with acidity, depending
only on the s-basicity of each C–H bond. Nevertheless, proto-
lysis of the tertiary CH bond (monitored by HD production)
occurs at almost the same rate as H/D exchange. Above 15
and up to 18 mol % SbF5 however, protolysis increases in com-
parison with H/D exchange, leading to tertiary pentyl ion from
Fig. 1 H/D exchange and conversion of propane reacted with DF–
SbF5 at various SbF5 concentrations.
via protolysis, depends very much on the acidity of the media
as shown in Fig. 1. This exchange is similar to the one observed
with isobutane and described earlier.12 SbF5 is not only gov-
erning the acidity but also the stabilization of the resulting
cations. When the concentration of SbF5 in DF is increased
from 10 to 18 mol %, the exchange rate increases substantially
but then decreases dramatically at higher SbF5 concentrations
due to predominant ionization with simultaneous reduction of
Scheme 2 H/D exchange and ionization of isopentane in DF–SbF5 .
1336
New J. Chem., 2002, 26, 1335–1339

View Article Online
The total isotope exchange, at various acidities, representing
the number of mol of hydrons exchanged per mol of isopen-
tane collected during the initial 20 min, was calculated by sum-
ming up the measured H/D exchange on each position
expressed in atom %, taking into account that isopentane
has 9 methyl, 2 methylene and 1 methine hydrogens. For
example at 10.7 mol
%
SbF5 : 0.083 ꢂ 9 + 0.115 ꢂ
2 + 13.1 ꢂ 1 ¼ 0.747 + 0.230 + 0.131 ¼ 1.108 mol of protons
exchanged per mol isopentane as illustrated in Table 1. This
total exchange shows also a maximum around 17.5 mol %
SbF5 . The deuterium distribution between primary, secondary
and tertiary positions is, however, changing with acidity.
Whereas the secondary/primary ratio increases from 1.38 at
10.7 mol % SbF5 to 1.65 at 33.8 mol % SbF5 , the tertiary/pri-
mary ratio, first stable at 1.6 upto 14.5 mol % SbF5 , decreases
drastically to reach a minimum at 0.1 for 33.8 mol % SbF5 .
This can be rationalized by the fact that, at higher SbF5 con-
centrations, protolysis and oxidative cleavage of the tertiary
CH bond become the preferred pathways. On the other hand,
with higher acidity, cracking products also increase at higher
SbF5 concentrations as can be seen on Fig. 4 showing the
alkane distribution in the recovered gas phase after reaction
(isopentane is omitted for clarity). Below 15 mol % SbF5 the
amount of cracking products is very low but increases rapidly
with increasing SbF5 concentrations over 15 mol %. Up to 19–
20 mol % SbF5 , propane and n-butane represent an important
part of these cracking products due to the protolytic C–C bond
cleavage from carbonium ions 8 and 9. At higher SbF5 concen-
trations methane and ethane, the least reactive alkanes, are the
predominant products in the gas phase.
Fig. 2 Isopentane conversion compared to hydrogen production.
H/D exchange of primary, secondary and tertiary C–H bonds
in small alkanes: r-basicity vs. r-reactivity
The s-basicity scale as proposed by Olah in the early seventies
was essentially based on the reactivity of the various C–H and
C–C bonds:
tertiary C H > C C bond > secondary C H > primary C H
Theoretical ab initio studies by Mota and colleagues9 provide a
clear insight in the exchange mechanism, which is suggested to
occur not via a bona fide carbonium ion (protonated alkane)
intermediate but rather via a transition state complex (Fig. 5)
in which the amount of positive charge developed on the
alkane depends on the basicity of the counterions, SbFꢀ6 ,
Sb2Fꢀ11 and higher. The reversible protonation of the C–C
bond unfortunately cannot be measured with our experimental
procedure. Nevertheless, protolysis can be avoided when weak
nonoxidizing superacid media are used, such as DF containing
less than 13 mol % SbF5 . However, under our experimental
conditions, below 9 mol % isotope exchange is too slow
to be accurately measured. The results are collected in
Fig. 3 Isopentane in DF–SbF5 : H/D exchange and HD formation
as a function of the acidity.
carbonium ion 4. Above 18 mol % SbF5 , protolysis as well as
H/D decrease again, in contrast with the still increasing con-
version of isopentane and simultaneous reduction of SbF5 .
Table 1 H/D exchange in isopentane at various concentration of SbF5 in DF
SbF5/mol % in DFa
10.7
Primary C–H
8.3 (1)
Secondary C–H
11.5 (1.38)
Tertiary C–H
13.1 (1.57)
Total exchange/mol Db
1.108
Exchange/atom %c
(normalized to primary position)
12.4
14.5
17.5
22
12.1 (1)
16.4 (1)
22.9 (1)
11.3 (1)
4.9 (1)
16.2 (1.34)
24.6 (1.50)
31.8 (1.39)
17.3 (1.53)
8.1 (1.66)
7.9 (1.65)
19.3 (1.60)
26.2 (1.60)
21.0 (0.92)
5.8 (0.52)
1 (0.20)
1.606
2.230
2.907
1.421
0.614
0.595
28.6
33.8
4.8 (1)
0.5 (0.1)
a
b
Mol % determined by weight, ꢃ1% of the indicated value. The total exchange represents the number of mol protons exchanged per mol iso-
pentane. The maximum value is 12. Determined by 1H, 2H NMR, ꢃ3% of the indicated value.
c
New J. Chem., 2002, 26, 1335–1339
1337

View Article Online
Table 2 H/D exchange observed in small alkanes reacted with DF–
SbF5
a
Exchange on/atom %c
SbF5/mol %
in DFb
Primary
C–H
Secondary
C–H
Tertiary
C–H
12
12.0
1
16.5
1.38
38.9
1.38
d
e
12.7
12.7
28.1
1
7.7
1
17.5
2.27
12.4
12.1
16.2
19.3
1
1.34
1.60
Fig. 4 Alkane distribution in the recovered gas phase after reaction
of isopentane in DF–SbF5 at various SbF5 concentrations (isopentane
is omitted for clarity).
a
All the numbers in italics are normalized to the exchange in the pri-
b
mary position. Mol % determined by weight, ꢃ1% of the indicated
value. Determined by 1H, 2H NMR, ꢃ3% of the indicated value.
c
d
e
See ref. 24 for complete study. See ref. 12 for complete study.
results are obtained in the condensed phase and that the
exchange occurs via a concerted mechanism in the alkane-acid
complex. This is also in line with the H/D exchange experi-
ments and calculations performed on the methane/acidic zeo-
lite system, albeit at much higher temperatures.25,26 It is
interesting to note that the relative rates, going from 1 for pri-
mary hydrons to 1.60 for tertiary hydrons, show only a very
small difference in comparison with the large difference in reac-
tivity when the bonds undergo protolytic cleavage to generate
primary, secondary or tertiary carbenium ions.
Conclusion
The behavior of small alkanes in nonoxidizing liquid supera-
cids is typical of an acid-base reaction. The H/D exchange
in DF–SbF5 takes place in the protonated alkane-acid com-
plex. At SbF5 concentrations in DF–SbF5 below 15 mol %,
in the presence of carbon monoxide, rapid reversible hydron
exchange occurs on all C–H bonds with negligible competition
of side reactions. It is noteworthy to underline that the relative
rates of the reversible protonation are in the same order of
magnitude of the basicity of the s-bonds, following the con-
cept developed by Olah, and do not follow the large order of
magnitude of the usual relative reactivity of these bonds asso-
ciated with the activation energy needed to generate primary,
secondary and tertiary carbenium ions.
Fig. 5 Geometry of the transition state for protonation in the tertiary
C–H bond of isobutane (H/H exchange) suggested by Mota et al. in
ref. 9a.
Experimental
Experimental procedures
Table 2 for SbF5 concentrations of about 12 mol %. Data for
isobutane and n-butane reported in this table were previously
published.12,24
The secondary/primary ratio is the same for propane, n-
butane and isopentane. The tertiary/primary ratio, however,
is quite different between isobutane and isopentane (2.27 and
1.6, respectively), which may be related to the relative reactiv-
ity of the tertiary C–H bond as discussed above.
Isopentane. CO was bubbled in isopentane at ꢀ10 ꢁC in
order to obtain an alkane partial pressure of approximately
175 Torr and a 1 : 3 molar ratio of alkane–carbon monoxide.
This mixture was bubbled for 20 min at a rate of 4 mL minꢀ1
at atmospheric pressure through 1 mL of DF–SbF5 superacid
in a KelF reactor at ꢀ10 ꢁC.
We suggest that in the absence of side reactions such as pro-
tolysis and cracking, the rate of exchange following the
sequence: tertiary C–H > secondary C–H > primary C–H is
in accordance with the relative s-basicity or proton acceptor
ability of these bonds. The relative enthalpies calculated for
1-H-isobutonium and 2-H-isobutonium show a much larger
difference (17 kcal molꢀ1) than expected for the H/D exchange
reactions.a This can be rationalized both by the fact that our
Propane. The alkane and carbon monoxide (1 : 3 molar
ratio) mixture was bubbled for 30 min at a rate of 4 mL minꢀ1
at atmospheric pressure through 1 mL of DF–SbF5 superacid
in a KelF reactor at ꢀ10 ꢁC.
In both cases the gaseous products were analysed by GC and
further condensed at ꢀ119 ꢁC for 1H and 2H NMR analysis to
determine the deuterium content. The superacid reaction
mixture containing the oxocarbenium ions was neutralized
1338
New J. Chem., 2002, 26, 1335–1339

View Article Online
by mixing with excess ethanol bicarbonate mixture at ꢀ78 ꢁC.
The ethanol solution of the resulting esters was used for GC
analysis after addition of a given amount of cyclohexanone
as internal standard.
3
4
5
6
H. Hogeveen and G. F. Bickel, J. Chem. Soc., Chem. Commun.,
1967, 635.
G. A. Olah and R. H. Schlosberg, J. Am. Chem. Soc., 1968, 90,
2726.
G. A. Olah, S. K. Prakash and J. Sommer, Superacids, Wiley-
Interscience, New York, 1985.
`
(a) A. Corma, P. J. Miguel and V. A. Orchilles, J. Catal., 1994,
145, 171; (b) A. Corma, J. Planelles, J. Sandoz-Marin and F.
Thomas, J. Catal., 1985, 93, 30; (c) W. O. Haag and R. M.
Dessau, Proceedings 8th International Congress on Catalysis,
Dechema, Frankfurt am Main, 1984, vol. 2, p. 305.
G. A. Olah, Angew. Chem., Int. Ed. Engl., 1973, 12, 173.
(a) G. A. Olah, S. K. Prakash, R. E. Williams, L. D. Field and
K. Wade, Hypercarbon Chemistry, Wiley-Interscience, New York,
1987; (b) G. A. Olah, K. K. Laali, Q. Wang and G. K. S. Prakash,
Onium Ions, Wiley Interscience, New York, 1998.
Expression of the H/D exchange
To express the amount of hydrons that have been exchanged at
the different carbon atoms of the alkane we generally use the
unit atom % referring to the amount of H exchanged by D
to the initial amount of H times 100.
To calculate the total amount of protons exchanged by D in
the alkane we sum up all the exchange values obtained for each
position in atom % multiplied by the number of hydrons in this
position. Example for propane in DF–SbF5 , 12 mol % SbF5 :
exchange in the methyl groups : 12 atom % (of each primary
hydron) and in the methylene group : 16.5 atom % (of each sec-
ondary position). Total exchange in mol % per mol propane:
6 ꢂ 12% + 2 ꢂ 16.5% ¼ 105 mol % H exchanged for D in
propane.
7
8
´
´
(a) P. M. Esteves, G. G. P. Alberto, A. Ramırez-Solıs and C. J. A.
Mota, J. Am. Chem. Soc., 1999, 121, 7345; (b) P. M. Esteves, C. J.
9
´
A. Mota, A. Ramırez-Solıs and R. Hernandez-Lamoneda, J. Am.
Chem. Soc., 1998, 120, 3213; (c) C. J. A. Mota, P. M. Esteves, A.
´
´
´
´
´
Ramırez-Solıs and R. Hernandez-Lamoneda, J. Am. Chem. Soc.,
1997, 119, 5193; (d ) N. Okulik, N. M. Peruchena, P. M. Esteves,
C. J. A. Mota and A. Jubert, J. Phys. Chem. A, 1999, 103, 8491.
10 P. Ahlberg, A. Karlsson, A. Goeppert, S. O. Nilsson Lill, P. Dine´r
and J. Sommer, Chem. Eur. J., 2001, 7, 1936.
11 J. Sommer, J. Bukala, S. Rouba, R. Graff and P. Ahlberg, J. Am.
Chem. Soc., 1992, 114, 5884.
NMR measurements
12 J. Sommer, J. Bukala, M. Hachoumy and R. Jost, J. Am. Chem.
Soc., 1997, 119, 3274.
13 J.-C. Culmann and J. Sommer, J. Am. Chem. Soc., 1990, 112,
4057.
14 H. Hogeveen, C. J. Gaasbeck and A. F. Bickel, Recl. Trav. Chim.
Pays-Bas, 1969, 88, 703.
15 (a) G. A. Olah, Y. Halpern, J. Shen and Y. K. Mo, J. Am. Chem.
Soc., 1971, 93, 1251; (b) G. A. Olah, Y. Halpern, J. Shen and Y.
K. Mo, J. Am. Chem. Soc., 1973, 95, 4960; (c) G. A. Olah, Y. K.
Mo and J. A. Olah, J. Am. Chem. Soc., 1973, 95, 4939.
16 J. W. Otvos, D. P. Stevenson, C. D. Wagner and O. Beeck, J. Am.
Chem. Soc., 1951, 73, 5741.
17 D. P. Stevenson, C. D. Wagner, O. Beeck and J. W. Otvos, J. Am.
Chem. Soc., 1952, 74, 3269.
18 J. Sommer, A. Sassi, M. Hachoumy, R. Jost, A. Karlsson and
P. Ahlberg, J. Catal., 1997, 171, 391.
1H and 2H NMR spectra were recorded on a Bruker AM
400 (400 MHz) spectrometer. Quantitative and qualitative
deuterium content was calculated by comparison of the 1H
and H NMR spectra recorded after addition of an adequate
amount of freon-113 (CF2ClCClF2) solution of a CD3Cl–
CHCl3 mixture used as internal standard.
2
Gas chromatography
The analysis of hydrocarbons were performed on a Girdel 300
with FID detector using a packed Hayesed R column (Ø ¼
1/800, l ¼ 2 m). Helium was used as carrier gas. The concentra-
tion of H2 (HD in our case) was determined on an Intersmat
˚
IGC 112M equipped with a 5 A molecular sieve. Argon was
employed as carrier gas. The results were computed on a Delsi
Instrument ENICA recorder integrator.
19 (a) J. Sommer, M. Hachoumy, F. Garin and J. Barthomeuf, J.
Am. Chem. Soc., 1994, 116, 5491; (b) J. Sommer, M. Hachoumy,
´
F. Garin, J. Barthomeuf and J. Vedrine, J. Am. Chem. Soc., 1995,
117, 1135; (c) J. Sommer, D. Habermacher, M. Hachoumy, R.
Jost and A. Reynaud, Appl. Catal. A, 1996, 146, 193; (d ) J. Som-
mer, R. Jost and M. Hachoumy, Catal. Today, 1997, 38, 309.
20 A. Goeppert, B. Louis and J. Sommer, Catal. Lett., 1998, 56, 43.
21 J. Sommer and J. Bukala, Acc. Chem. Res., 1993, 26, 370.
22 R. Jost and J. Sommer, Rev. Chem. Intermed., 1988, 9, 171.
23 J.-C. Culmann, M. Fauconnet, R. Jost and J. Sommer, New J.
Chem., 1999, 23, 863.
Acknowledgements
Financial support of the Loker Hydrocarbon Institute, USC,
Los Angeles, is gratefully acknowledged.
24 J. Sommer, J. Bukala and M. Hachoumy, Res. Chem. Intermed.,
1996, 22, 753.
References
1
2
H. Pines, The Chemistry of Catalytic Hydrocarbon Conversion,
Academic Press, New York, 1981.
G. A. Olah and A. Molnar, Hydrocarbon Chemistry, Wiley-Inter-
science, New York, 1995.
25 G. J. Kramer, R. A. van Santen, C. A. Emeis and A. K. Nowak,
Nature, 1993, 363, 529.
26 J. M. Vollmer and T. N. Truong, J. Phys. Chem. B, 2000, 104,
6308.
New J. Chem., 2002, 26, 1335–1339
1339