Alkali-metal ion catalysis and inhibition in nucleophilic displacement reactions at carbon, phosphorus and sulfur centres. IX. p-Nitrophenyl diphenyl phosphate

Article abstract of DOI:10.1039/b208408b

We report on the catalytic effects by alkali-metal ions in the ethanolysis of p-nitrophenyl diphenyl phosphate, in continuation of our studies on alkali-metal ion catalysis and inhibition in nucleophilic displacement reactions at carbon, phosphorus and sulfur centres. The following selectivity order of catalytic reactivity was observed for nucleophilic displacement at the phosphorus center with p-nitrophenoxide as leaving group: Li⁺ > Na⁺ > K⁺ > Cs⁺ A minor reaction pathway with phenoxide leaving was also found. The k_obs data have been dissected into reaction pathways by free ions (k_EtO) and by ion pairs (k_MOEt) with the latter being dominant, in a 4-membered transition-state. Further analysis is given in terms of initial-state and transition-state stabilization by the alkali-metal ions in terms of the Eisenman model (electrostatic interaction vs. desolvation). Results of ab-initio MO calculations are presented based on interaction between M⁺ and a model bipyramidal phosphorane intermediate and compared with the sulfurane analogue.

Full text of DOI:10.1039/b208408b

View Article Online / Journal Homepage / Table of Contents for this issue
Alkali-metal ion catalysis and inhibition in nucleophilic
displacement reactions at carbon, phosphorus and sulfur centres.
IX.† p-Nitrophenyl diphenyl phosphate‡
Ruby Nagelkerke, Gregory R. J. Thatcher and Erwin Buncel*
Department of Chemistry, Queen’s University, Kingston, ON, Canada K7L 3N6
Received 29th August 2002, Accepted 10th October 2002
First published as an Advance Article on the web 3rd December 2002
We report on the catalytic effects by alkali-metal ions in the ethanolysis of p-nitrophenyl diphenyl phosphate, in
continuation of our studies on alkali-metal ion catalysis and inhibition in nucleophilic displacement reactions at
carbon, phosphorus and sulfur centres. The following selectivity order of catalytic reactivity was observed for
nucleophilic displacement at the phosphorus center with p-nitrophenoxide as leaving group: Liϩ > Naϩ > Kϩ > Csϩ.
A minor reaction pathway with phenoxide leaving was also found. The kobs data have been dissected into reaction
Ϫ
pathways by free ions (kEtO ) and by ion pairs (kMOEt), with the latter being dominant, in a 4-membered transition-
state. Further analysis is given in terms of initial-state and transition-state stabilization by the alkali-metal ions in
terms of the Eisenman model (electrostatic interaction vs. desolvation). Results of ab-initio MO calculations are
presented based on interaction between Mϩ and a model bipyramidal phosphorane intermediate and compared with
the sulfurane analogue.
Previous studies with p-nitrophenyl diphenyl phosphate have
Introduction
included micellar catalyzed hydrolysis by Bunton and co-
Metal ions are known to play an important role in biological
workers,7 by Moss et al.,8 and by Menger and Whitesell,9 which
processes, and especially on the mechanism of phosphoryl
included metallomicelles, while Breslow and coworkers have
transfer reactions.2 The majority of studies have been concerned
reported on the catalytic hydrolysis by tetracoordinated zinc
with the catalytic effects of divalent metal ions, e.g. Mg2ϩ, Ca2ϩ
,
complexes.10
Mn2ϩ, Fe2ϩ, Co2ϩ, Cu2ϩ and Zn2ϩ in the hydrolysis of phosphate
esters.3
Results
The ubiquitous presence of alkali metals in biological
systems4 calls for an examination of their effects on phosphoryl
transfer and such studies were initiated in our laboratory.
Unexpectedly, we found that ethanolysis of the phosphinate
ester, p-nitrophenyl diphenyl phosphinate, 1, exhibited catalysis
by alkali metals, with the selectivity order Liϩ > Naϩ > Kϩ.5
Reaction pathways for alkali-metal ethoxides
The rates of ethanolysis reactions of 3 with LiOEt, NaOEt,
KOEt and CsOEt were determined spectrophotometrically in
anhydrous ethanol at 25 ЊC under pseudo-first-order conditions
with the base in excess. Rate constants, kobs, were calculated
from linear plots of log(Ainf Ϫ At) vs. time where the absorb-
ances (A) were taken at 400 nm corresponding to formation
of p-nitrophenoxide ion. Results are given in Supplementary
Tables S1–S4‡ and shown graphically in Fig. 1.
Repetitive scanning of the LiOEt reaction solution over
210–450 nm revealed also absorbance increases at 230 nm and
290 nm corresponding to the concurrent release of phenoxide
ion, albeit at a much slower rate. The generation of PhOϪ
similarly followed first-order kinetics with a rate constant ca.
1/250, slower than for formation of p-nitrophenoxide ion. This
secondary process (Scheme 1) was also observed by Breslow10
in the OHϪ hydrolysis of 3 and was not investigated further
in this work.
In contrast, in the ethanolysis of p-nitrophenyl benzene-
sulfonate, 2, Liϩ brought about inhibition of rate and the
catalytic order of the other alkali-metals was Kϩ > Csϩ
Naϩ.6
>
Effect of added complexing agents
The present work is concerned with study of catalytic effects
on the ethanolysis of the phosphate ester, p-nitrophenyl di-
phenyl phosphate, 3. In this case, in addition to nucleophilic
displacement at the phosphorus center with p-nitrophenoxide
as leaving group, a minor reaction pathway with phenoxide
leaving was also identified. The structural change in the phos-
phorus() center, could be expected to have a marked effect on
reactivity. Kinetic analysis has enabled dissection of metal
ion interactions in the ground-state and transition-state, with
comparison among the substrates 1–3.
The effect of added 18-Crown-6 ether to the reaction of 3 with
KOEt is shown in Fig. 2 (see also Table S5‡). The rate decreases
until one equivalent is added, at which point further addition of
complexing agent has no effect. A similar result had been found
in the phosphinate system.5b However these results contrast
with typical SN2 type reactions where metal complexing
agents produce rate acceleration. A rate enhancing effect was
also observed4,6 for the addition of complexing agents to the
reaction of LiOEt with the benzenesulfonate 2 (vide infra), in
contrast with the present work.
† For Part VIII see ref. 1.
Effect of added metal ions
‡ Electronic supplementary information (ESI) available: kinetic data for
the reaction of p-nitrophenyl diphenyl phosphate with various reagents
in differing media. See http://www.rsc.org/suppdata/ob/b2/b208408b/
Added potassium iodide and acetate to the reaction of 3 with
KOEt increased the rate significantly; as was similarly the case
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 6 3 – 1 6 7
163

View Article Online
Scheme 1
Fig. 3 Kinetic data for the reaction of 3 with LiOEt in the presence of
LiCl and LiNO3 and with KOEt in EtOH at 25 ЊC in the presence of
KO(O)CCH3 and KI. (Data shown in Tables S6(a)–(d)).‡
Fig. 1 Kinetic data for the reaction of 3 with LiOEt, KOEt, CsOEt
and KOEt in EtOH and in the presence of excess 18-Crown-6 at 25 ЊC
(data in Tables S1–S5‡).
the two substrates can be had from the reactions of MOEt in
Ϫ
Ϫ
presence of excess 18C6 which yields kEtO (1)/kEtO (3) = 10.
This reactivity difference finds explanation in O P π-back-
bonding in 3, but its absence in 1, which thus results in 1 having
the more electrophilic P center.
The upward curvature in the plots of kobs vs. [MOEt] (Fig. 1),
which was also observed in the reaction of 1,5b contrasts with
expected behavior if free ethoxide were more reactive than the
ion-paired species. In fact, a number of SN2 processes at
saturated carbon centers have been found to be characterized
by downward curvature of the k vs. [Nu] plots,11a–11e which was
interpreted as arising from decreased reactivity of the ion-
paired species as compared to the dissociated anions. The
contrasting behavior in the present system finds explanation in
terms of reactivities of free ions vs. ion pairs toward the phos-
phorus ester E, which can be expressed in terms of Scheme 2
and equations 1 and 2:
Fig. 2 Kinetic data for the reaction of 3 with KOEt in the presence of
18-Crown-6, in EtOH at 25 ЊC. (Data in Table S5‡).
on addition of LiNO3 and LiCl to the LiOEt reaction (Fig. 3,
Table S6‡). However, addition of (n-Bu)4NCl as an inert salt to
the reaction of 3 with KOEt in the presence of 18C6 gave only a
small (ca. 4%) increase in rate (Table S7‡), indicating a small salt
effect. These results confirm the catalytic effect of alkali metal
cations on the ethanolysis of 3 in the order Liϩ > Kϩ.
Discussion
Reactivities of alkali-metal ethoxide ion pairs vs. free ions
Notable immediately on comparing the reactivities of the
phosphate ester 3 with the phosphinate ester 1, is the reactivity
Ϫ
order 1 > 3. Direct comparison of the rate constants, kEtO , for
Scheme 2
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 6 3 – 1 6 7
164

View Article Online
kobs = kEtOϪ[EtOϪ] ϩ kMOEt[MOEt]
(1)
(2)
Table 1 Calculated rate constants for free ions and ion pairs in eth-
anolysis of 3 at 25 ЊC
kobs/[EtOϪ] = kEtO ϩ kMOEtKa[EtOϪ]
Ϫ
MOEt/M
kMOEt/sϪ1MϪ1
The plots of kobs/[EtOϪ] vs. [EtOϪ] are generally linear
(R2: Li, 0.998; Na, 0.942; K, 0.987; Cs, 0.999) but the small
intercept values have rather large associated errors, e.g. Fig. 4
LiOEt
NaOEt
KOEt
CsOEt
KOEt/18C6 (Table S7‡)
1.36 0.026
1.03 0.057
0.85 0.035
0.42 0.0008
0.094 0.001
kinetically indistinguishable. There is evidence for participation
by ion pairs in similar systems12 and the following discussion is
based on catalysis occurring by MOEt ion pairs.
Metal ion interactions in ground- and transition-states: dissection
of stabilization energies
The catalytic role of metal ions in the reactions of alkali metal
ethoxides with the compounds under investigation can be
analyzed in terms of differing metal–ion interactions in the
ground- and transition-states. Catalysis by ion pairs will be
assumed, as opposed to the pre-association mechanism, as
stated above.
To investigate the catalytic role of metal ions in the reaction
mechanism, the relative stabilization of the ground-states (GS)
to that of the transition-states (TS) should be considered. We
use a method devised by Kurz14a and in different forms by
Mandolini14b and Tee.14c
Association constants for interactions of alkali metals and
ethoxides (Ka, in Scheme 1) are known (LiOEt, 212 MϪ1;
NaOEt, 102 MϪ1; KOEt, 90 MϪ1; CsOEt, 121 MϪ1),13 and it is
assumed here (for lack of other evidence) that interaction
between Mϩ and E in the ground-state can be neglected. Using
the thermodynamic cycle in Scheme 3 one can then calculate a
corresponding virtual association constant, KЈa, for the TS.
Fig. 4 Ion pairing treatment of the data for the reaction of 3 with
KOEt in EtOH at 25 ЊC. (Data in Table S8‡).
Ϫ
(Table S8‡), for KOEt. A more reliable value of kEtO could be
obtained from plots of kobs/[KOEt] vs. [KOEt] in the presence
of excess 18C6, Fig. 5 (Table S9‡) for KOEt, which yields
Scheme 3
From transition-state theory, one can then derive eqn. (3)
which combines the GS association constants with reaction
rates to obtain association constants KaЈ for the TS. The latter
in turn can be expressed in terms of free energies of stabiliz-
ation, δGts and δGip, with the net catalytic effect δGcat being
given by the difference in free energies of activation between the
catalyzed and uncatalyzed reactions in eqn. 4. This relationship
between GS and TS stabilization is illustrated in Fig. 6.
Fig. 5 Ion pairing treatment of the data for the reaction of 3 with
KOEt in the presence of excess complexing agent. (Data in Table S9‡).
kEtO = 0.0939 0.001 MϪ1sϪ1. Values of kMOEt obtained from
Ϫ
the various plots are collected in Table 1. The overall reactivity
order is LiOEt > NaOEt > KOEt > CsOEt > EtOϪ.
Fig. 6 Free energies involved in the uncatalyzed and metal ion
The mechanism we suggest for catalysis by MOEt ion pairs
is one where the metal cation coordinates to the phosphoryl
oxygen concertedly with nucleophilic attack by ethoxide in a
4-membered transition-state. Alternatively, one could consider
pre-equilibrium attachment of Mϩ followed by nucleophilic
attack. However these two alternatives can be shown to be
catalyzed reactions of substrate S with alkali metal ethoxides, MOEt.
Ϫ
Ka = kMOEtKa/kEtO
(3)
(4)
∆Gcat = ∆Gc≠ Ϫ ∆Gu≠ = δGts Ϫ δGip
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 6 3 – 1 6 7
165

View Article Online
Table 2 Free energies of metal ion stabilization of ethoxide (δGip),
transition states (δGts), and the net catalytic effect (∆Gcat) for the reac-
tion of ethoxide with 1, 2 and 3 at 25 ЊC in ethanol. All values are in kJ
molϪ1
δGip
δGts
∆Gcat
Ion
EtOϪ
3
1
2
3
1
2
Liϩ
Naϩ
Kϩ
Ϫ13.3
Ϫ11.5
Ϫ11.2
Ϫ11.9
Ϫ19.7
Ϫ17.2
Ϫ16.4
Ϫ15.4
Ϫ21.2
Ϫ17.6
Ϫ15.1
Ϫ13.5
Ϫ12.1
Ϫ14.0
Ϫ15.0
Ϫ14.9
Ϫ6.4
Ϫ5.7
Ϫ5.2
Ϫ3.5
Ϫ7.9
Ϫ6.1
Ϫ3.9
Ϫ3.0
ϩ1.2
Ϫ2.5
Ϫ3.8
Ϫ3.0
Csϩ
The δGcat, δGip and δGts data for 1, 2 and 3 are presented in
Table 2. It is seen that for the phosphorus based esters 1 and 3,
all the alkali metal ions stabilize the TS to a greater extent than
the GS and thus act as catalysts (∆Gcat < 0). While both esters
exhibit the same trends in stabilization, the phosphinate 1
shows a greater range of values from Liϩ, the most effective
catalyst, to Csϩ, the least effective catalyst (∆∆Gcat = 4.9 kJ
molϪ1) than the phosphate ester 3 (∆∆Gcat = 2.9 kJ molϪ1).
This is interpreted to signify that the phosphinate ester is
more sensitive to the nature of the alkali-metal ion than
the phosphate ester. For the sulfonate ester 2, Liϩ acts as an
inhibitor with ∆Gcat > 0, i.e. association in the GS is stronger
than with the TS. Also, in the case of 2, Kϩ showed the greatest
value for ∆Gcat, having the strongest association with the TS,
i.e. a selectivity order very different from that observed for
the two phosphorus esters.
Fig. 7 Plot of δGts versus the reciprocal of the crystal radius of the
metal ion for the reactions of 3 with alkali metal ethoxides in ethanol at
25 ЊC.
In examining the above contrasting selectivities, an interest-
ing relationship becomes apparent with selectivity patterns
of alkali-metal cations for ion-exchange resins and glass
electrodes which has been discussed elegantly by Eisenman.15
Two opposing factors were considered to govern the strength
of the interaction between a cation and a fixed ionic group:
electrostatic interactions engendered in approach of ions of
opposite charge, and the energy expended in rearranging
solvent molecules around the ions to enable close contact, i.e.
the difference in solvation energies of the ions. The overall
Fig.
8
Structures and charge distributions of sulfurane and
phosphorane intermediates from ab initio calculations.
free energy change for interchange of cations 1 and 2 (∆GЊ )
½
at a fixed anionic site will be given by the differences in
electrostatic interactions and solvent rearrangement terms as
shown in eqn. 5 where e is the electronic charge, r1, r2 and rA
are the ratii of cations 1 and 2 and of the anion, respectively,
and ∆G1, ∆G2 solvation energies of cations 1 and 2.
∆GЊ = [e2/(rA ϩ r2)] Ϫ [e2/(rA ϩ r1)] Ϫ (∆G2 Ϫ ∆G1) (5)
½
One would predict that for an anionic group with localized
charge, or high field strength, the electrostatic term would be
dominant over the solvation energy term, allowing the ions to
come in direct contact. Thus smaller cations would be bound
most strongly, leading to the selectivity order Liϩ > Naϩ > Kϩ >
Csϩ. One also predicts a linear relationship between ∆Gts(Mϩ)
and the inverse of the crystal radius (Fig. 7) as was also reported
for the phosphinate system 1.
On the other hand, the sulfonate system, with charge
delocalized over two oxygens in the TS, would correspond to
a low electric-field strength situation and here the solvation
term would be dominant over the electrostatic term. A linear
relationship between ∆Gts (Mϩ) and the inverse of the solvated
(Stokes) radius of Mϩ is followed.
Fig. 9 Relative energies (kJ molϪ1) for stabilization of sulfurane and
phosphorane intermediates by Liϩ and Naϩ and geometries of metal
ion complexes: sulfurane (diequatorial); phosphorane (apical-
equatorial). The structures presented are the lowest energy stationary
points from exploration of the conformational space and different
coordination geometries. Relative energies include ZPE and were
calculated at B3LYP/6-31ϩG*//HF/3-21ϩG(*).
31ϩG*//HF/3-21ϩG* level, are presented in Fig. 9. It is
apparent that, in the gas phase, there is weaker stabilization
in the sulfurane species compared to the phosphorane on
interaction with the cations. In the Eisenman scheme, this could
lead to interaction with solvated cations in the sulfonate TS
compared to bare cations for the phosphorane TS. The calcu-
lational results are hence condordant with our observations of
selectivities of alkali-metals in the reactions under investigation
as discussed herein.
In parallel theoretical studies transition-state models were
optimized at the ab initio HF/3-21 ϩ G* level as penta-coordin-
ate intermediates resulting from attack of OHϪ on phospho-
16
rane H3PO2 and sulfurane H2SO3. The optimized trigonal
bipyramidal structures along with charge distributions are
shown in Fig. 8. Relative energies for stabilization of the
sulfurane and phosphorane species by Liϩ and Naϩ ions
and the geometries of the metal ion complexes, at the B3LYP/6-
Experimental
p-Nitrophenyl diphenyl phosphate was prepared according to
literature procedures (10) and fully characterized. Alkali-metal
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 6 3 – 1 6 7
166

View Article Online
J. Canceill, L. Lacombe and J.-M. Lehn, J. Chem. Soc., Perkin Trans.
ethoxide solutions were prepared as previously and were stored
under N2 in the refrigerator. Base concentrations were deter-
mined by titration with standard HCl. Salts used were of
reagent quality. 18-Crown-6 ether (Aldrich) was recrystallized
from acetonitrile and dried over P2O5 in vacuo. Bu4NCl was
recrystallized from acetone/ether and dried in vacuo at 60 ЊC.
Kinetics were performed using a Varian Cary 3 or a Hewlett
Packard 8452A diode array spectrophotometer equipped with a
HP 89075C Multicell transport and thermostatted cell holders.
Pseudo-first-order conditions with base at least 20 times greater
than substrate concentration were used. Typically reactions
were followed for 4–10 half-lives and kobs calculated using a
linear regression of the equation
2, 1995, 417; (d ) J. J. R. Fraústo da Silva and R. J. P. Williams,
The Biological Chemistry of the Elements, Clarendon Press, Oxford,
1991.
5 (a) E. Buncel, E. J. Dunn, R. A. B. Bannard and J. P. Purdon,
J. Chem. Soc., Chem. Commun., 1984, 162; (b) E. J. Dunn and
E. Buncel, Can. J. Chem., 1989, 67, 1440; (c) M. J. Pregel, E. J. Dunn
and E. Buncel, Actualidades de Fisico-Quimica Organica, 1991, 110.
6 (a) E. Buncel and M. J. Pregel, J. Chem. Soc., Chem. Commun., 1989,
1566; (b) M. J. Pregel, E. J. Dunn and E. Buncel, Can. J. Chem.,
1990, 68, 1846; (c) M. J. Pregel and E. Buncel, J. Org. Chem., 1991,
56, 5583; (d ) M. J. Pregel and E. Buncel, J. Chem. Soc., Perkin Trans.
2, 1991, 307; (e) M. J. Pregel, E. J. Dunn and E. Buncel, J. Am.
Chem. Soc., 1991, 113, 3545.
7 (a) C. A. Bunton and L. Robinson, J. Org. Chem., 1969, 34, 773; (b)
C. A. Bunton and L. G. Ionescu, J. Am. Chem. Soc., 1973, 95, 2912.
8 R. A. Moss, K. W. Alevis and J.-S. Shin, J. Am. Chem. Soc., 1984,
106, 2651.
ln(Ainf Ϫ At) = kobst ϩ ln(Ainf Ϫ Ao).
9 (a) F. M. Menger and L. G. Whitesell, J. Am. Chem. Soc., 1985, 107,
707; (b) F. M. Menger, L. H. Gan, E. Johnson and D. H. Durst,
J. Am. Chem. Soc., 1987, 109, 2800.
Kinetic runs were performed generally in replicate; kobs values
are presented as Supplementary Material.‡
10 S. H. Gellman, R. Petter and R. Breslow, J. Am. Chem. Soc., 1986,
108, 2388.
Acknowledgements
11 (a) C. C. Evans and S. Sugden, J. Chem. Soc., 1949, 270; (b)
J. R. Bevan and C.B. Monk, J. Chem. Soc., 1956, 1396; (c) P. Jones,
P. Harrison and L. Wynne-Jones, J. Chem. Soc., Perkin Trans. 2,
1979, 1679; (d ) P. Beronius and L. Pataki, Acta. Chem. Scand., Ser.
A, 1979, A33, 675; (e) N. N. Lichtin and K. N. Rao, J. Am. Chem.
Soc., 1961, 83, 2417.
12 (a) A. Papoutsis, G. Papanastasiou, D. Jannakoudakis and
C. Georgulis, J. Chim. Phys., 1985, 82, 913; (b) J. R. Jones, Prog.
React. Kinet., 1973, 7, 1.
Support of this research by grants from the Natural Sciences
and Engineering Research Council of Canada (NSERC) is grate-
fully acknowledged. R. N. is the recipient of a Postgraduate
Scholarship from NSERC.
Footnotes and References
13 J. Barthel, G. Bader and M. Raach-Lenz, Z. Phys. Chem., 1973, 84,
100.
1 R. Nagelkerke, M. J. Pregel, E. J. Dunn, G. R. J. Thatcher and
E. Buncel, Org. React. (Tartu), 1995, 102, 11.
2 (a) A. Fersht, Enzyme Structure and Mechanism, ed. W. H. Freeman
and company, New York, 2nd edn., 1985; (b) G. R. J. Thatcher and
R. Kluger, Adv. Phys. Org. Chem., 1989, 25, 99; (c) J. Chin, Acc.
Chem. Res., 1991, 24, 145.
3 (a) R. Breslow, Adv. Enzymol., 1986, 58, 1; (b) D. Herschlag and
W.P. Jencks, J. Am. Chem. Soc., 1987, 109, 4665.
4 (a) L. Stryer, Biochemistry, ed. W. H. Freeman and company, New
York, 3rd edn., 1988; (b) T. M. Fyles, T. D. James and K. C. Kaye,
J. Am. Chem. Soc., 1993, 115, 12315; (c) M. J. Pregel, L. Jullien,
14 (a) J. L. Kurz, J. Am. Chem. Soc., 1963, 85, 987; (b) R. Cacciapaglia
and L. Mandolini, Chem. Soc. Rev., 1993, 22, 221; (c) O. S. Tee,
Adv. Phys. Org. Chem., 1994, 29, 1.
15 D. Eisenman, Biophys. J. Suppl., 1962, 2, 259.
16 (a) M. J. Pregel, E. J. Dunn, R. Nagelkerke, G. R. J. Thatcher and
E. Buncel, Chem. Soc. Rev., 1995, 24, 445; (b) R. Nagelkerke,
Ph.D. Thesis, Queen’s University, 1993; (c) A. M. P. Borrajo,
J. F. Gal, P. C. Maria, M. Decouzon, D. C. Ripley, E. Buncel and
G. R. J. Thatcher, J. Org. Chem., 1997, 62, 9203.
17 W. M. Gulick and D. H. Geske, J. Am. Chem. Soc., 1966, 88, 2928.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 6 3 – 1 6 7
167