Study of the reaction of bis- and tris-pyrylium and -thiopyrylium cations with methoxide ion. Electronic effects of the heterocyclic rings as substituents

Article abstract of DOI:10.1021/jo050884a

The bis- and tris-pyrylium and thiopyrylium cations 1-4 were prepared in gram scale by heterocyclization of the corresponding bis- and tris-1,5-pentanediones 6 and 8. Their reaction with CD₃ONa in CD ₃OD was studied by ¹H

Full text of DOI:10.1021/jo050884a

Study of the Reaction of Bis- and Tris-pyrylium and -thiopyrylium
Cations with Methoxide Ion. Electronic Effects of the Heterocyclic
Rings as Substituents
Barbara Branchi,† Giancarlo Doddi,*,‡ and Gianfranco Ercolani*,†
Dipartimento di Scienze e Tecnologie Chimiche, Universita` di Roma Tor Vergata, Via della Ricerca
Scientifica, 00133 Roma, Italy, and Dipartimento di Chimica e CNR-IMC, Universita` di Roma La
Sapienza, P.le Aldo Moro, 2, 00185 Roma, Italy
ercolani@uniroma2.it
Received May 2, 2005
The bis- and tris-pyrylium and thiopyrylium cations 1-4 were prepared in gram scale by
heterocyclization of the corresponding bis- and tris-1,5-pentanediones 6 and 8. Their reaction with
CD3ONa in CD3OD was studied by 1H NMR at -40 °C and at +25 °C. At low temperature,
kinetically controlled mixtures of 2H and/or 4H adducts were detected, whereas at room temperature
the mixtures equilibrated to yield, in all of the cases, the more stable 2H adducts exclusively. A
spectrophotometric study of the reactions with sodium methoxide in methanol was carried out at
25.0 °C with the aim of determining the stepwise equilibrium constants for the addition of MeO-
at the R position of the heteroaromatic rings. The obtained equilibrium constants allowed the
evaluation of the electronic effects of chalcogenopyrylium and 2H-chalcogenopyran subunits as
substituents. Despite the different sensitivity to electronic effects, pyrylium and thiopyrylium rings
have a similar electron-withdrawing effect with a σ+p ∼ 0.8 and a σ+m ∼ 0.5. Apart from the expected
importance of the inductive effect due to the positive charge, the difference between these two
values remarks the importance of the resonance contribution. In contrast both the neutral 2H-
pyranyl and thiopyranyl rings have a negligible effect as substituents, independently of the position,
para or meta, they occupy.
Introduction
known to undergo reversible addition by a number of
nucleophiles to yield neutral adducts.
Dynamic covalent chemistry1 requires reversible reac-
tions to selectively build template-driven assemblies
under thermodynamic control. In the course of our
ongoing research on the physical basis of template effects
and self-assembly,2 we planned to investigate the behav-
ior of covalent dynamical systems based on pyrylium3 and
thiopyrylium4 cations, as these heteroaromatic units are
As a preliminary study, we report here the preparation
of the bis- and tris-pyrylium and thiopyrylium cations
1-4 (see the Abstract graphic) and a quantitative study
of their reaction with methoxide ion in methanol, aimed
at elucidating the electronic effects of chalcogenopyrylium
and chalcogenopyran subunits on addition equilibria.
† Universita` di Roma Tor Vergata.
(3) (a) Balaban, A. T.; Schroth, W.; Fischer, G. Adv. Heterocycl.
Chem. 1969, 10, 241. (b) Balaban, A. T.; Dinculescu, A.; Dorofeenko,
G. N.; Fischer, G. W.; Koblik, A. V.; Mezheritskii, V. V.; Schroth, W.
Pyrylium Salts. Syntheses, Reactions, and Physical Properties. Adv.
Heterocycl. Chem., Suppl. 2 1982. (c) Schroth, W.; Do¨lling, W.; Balaban,
A. T. Houben-Weyl Methoden Org. Chem. 1992, E7b, 755. (d) Balaban,
T. S.; Balaban, A. T. In Science of Synthesis; Thomas, J., Ed; Georg
Thieme Verlag: Stuttgart, 2003; Vol. 14, pp 11-200.
‡ Universita` di Roma La Sapienza.
(1) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K.
M.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 2002, 41, 898.
(2) (a) Doddi, G.; Ercolani, G.; Franconeri, S.; Mencarelli, P. J. Org.
Chem. 2001, 66, 4950. (b) Ercolani, G. J. Phys. Chem. B 2003, 107,
5052. (c) Ercolani, G. J. Am. Chem. Soc. 2003, 125, 16097. (d) D’Acerno,
C.; Doddi, G.; Ercolani, G.; Franconeri, S.; Mencarelli, P.; Piermattei,
A. J. Org. Chem. 2004, 69, 1393. (e) Doddi, G.; Ercolani, G.; Mencarelli,
P.; Piermattei, A. J. Org. Chem. 2005, 70, 3761.
(4) (a) Doddi, G.; Ercolani, G. Adv. Heterocycl. Chem. 1994, 60, 65-
195. (b) Rudorf, W.-D. In ref 3d, pp 649-718.
10.1021/jo050884a CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/01/2005
6422
J. Org. Chem. 2005, 70, 6422-6428

Reactions of Bis- and Tris-pyrylium and -thiopyrylium Cations
SCHEME 1
CHART 1
Results
Synthesis of 1-4. The bispyrylium 1 was prepared
by an improved procedure that proved to be simpler and
slightly more efficient than that reported in the literature
(72% vs 61% total yield):5 Michael addition of 2 equiv of
pinacolone to the bisenone 5, promoted by heterogeneous
sodium amide in toluene, yielded the bispentanedione 6
in 80% yield; this was converted into the bispyrylium
cation 1 in high yield (91%) by oxidation in acetic
anhydride with triphenylcarbenium tetrafluoroborate
generated in situ (Scheme 1).
Since the 1,5-pentanedione substructure may be re-
garded as a synthetic precursor of the thiopyrylium ring
as well, the bispentanedione 6 was also converted into
the bisthiopyrylium cation 2 (yield 42%) by the method
of Strzelecka and Gionis,6 i.e., by reaction in acetic acid
with P4S10 in the presence of lithium perchlorate (Scheme
1).
Analogous procedures were used for the preparation
of the triscations 3 and 4, as illustrated in Scheme 2.
The tris-enone 7 was prepared in 63% yield by NaOH-
catalyzed condensation in aqueous ethanol of benzene-
1,3,5-tricarbaldehyde with 3 equiv of pinacolone. Michael
addition of 3 equiv of pinacolone to the tris-enone 7,
promoted by heterogeneous sodium amide in toluene,
yielded the tris-pentanedione 8 in 77% yield. This was
converted into both the tris-pyrylium cation 3 (yield 90%)
and the tris-thiopyrylium cation 4 (yield 21%) by the
same methods followed for the syntheses of 1 and 2. It
should be noted that the low yield of the thiopyrylium
salts 2 and 4 is mainly due to the purification procedure
that required repeated recrystallizations in order to
remove mixed pyrylium-thiopyrylium byproducts.
1H NMR Study of the Reactions of 1-4 with
Methoxide Ion. We had previously studied the reaction
of a number of simple pyrylium and thiopyrylium cations,
symmetrically substituted in the 2,6-positions, with
methoxide ion in methanol by 1H NMR.7 Besides provid-
ing quantitative data about the effects of the ring-
heteroatom and of substituents in pyrylium and thio-
pyrylium ions, these studies were aimed at gaining a
deeper understanding of cation-anion combination reac-
tions. The reactions of the bis- and tris-cations 1-4 were
studied by the same approach, i.e., by recording 1H NMR
spectra at low temperature (≈ -40 °C) immediately after
the mixing of the reactants, and at room temperature
(≈ 25 °C). At low temperature, the reaction is under
kinetic control, and thus, the observed composition of the
mixture of the adducts reflects the relative magnitude
of the kinetic constants for the nucleophilic attack. In
contrast, at room temperature, the mixture rapidly
attains equilibration, and thus, its composition reflects
the relative stabilities of the adducts. In all cases, the
addition of the solid substrate to an excess CD3ONa
solution in CD3OD thermostated at -40 °C within the
NMR probe led to the immediate disappearance of the
signals of the substrates and the appearance at higher
fields, owing to the neutralization of the positive charges,
of the signals of the corresponding adducts. The reaction
of cations 1 and 2 with excess methoxide ion can give in
principle the bis-pyrans 9-11 and bis-thiopyrans 12-
14, respectively (Chart 1). The 1H NMR technique proved
to be unsuccessful in discriminating between the dia-
stereomers represented by the general structures 9 and
12, probably because the stereogenic centers that are
formed upon addition of methoxide ion to the R positions
of the various heteroaromatic nuclei, are remote from
each other. In contrast constitutional isomers, due to the
attack of either R or γ position, could be easily distin-
guished.
SCHEME 2
J. Org. Chem, Vol. 70, No. 16, 2005 6423

Branchi et al.
CHART 2
SCHEME 3
trolled mixture of 2H- and/or 4H-adducts that equili-
brates at 25 °C to yield, in all of the cases, the more stable
2H adducts exclusively.
Spectrophotometric Study of the Equilibria of
1
1-4 with Methoxide Ion. The H NMR study showed
that at 25 °C the addition of methoxide ion to the cations
1-4 is thermodynamically controlled and occurs exclu-
sively at the R positions. We carried out a spectrophoto-
metric study of the reactions in methanol at 25 °C to
evaluate the corresponding stepwise equilibrium con-
stants. The reactions were monitored by following the
absorbance of the pyrylium or thiopyrylium nucleus, at
a wavelength where it is the only chromophore absorbing,
as a function of the methoxide ion concentration. Since
the required concentrations of methoxide ion to follow
the reactions were very low, suitable buffers or a strong
acid (HClO4) were necessary to modify the pH of the
solutions.
In the case of the bis-pyrylium 1, only the bis-adducts
9 and 10 in a ratio 4:1 were observed at -40 °C. By
raising the temperature at 25 °C, the bis-adduct 10
completely converted into the bis-adduct 9. In contrast,
the addition of the bis-thiopyrylium cation 2 to the
CD3ONa/CD3OD solution at -40 °C led to the immediate
precipitation of a white solid. The precipitate was dis-
solved in CDCl3 at room temperature and then subjected
to 1H NMR analysis. In CDCl3, the equilibration among
the adducts is slow even at room temperature, so the
composition of the isomeric mixture has still not reached
equilibrium. The mixture was constituted by the adducts
12 and 13 in a ratio 10:1 with trace amounts of the
adduct 14. To attain equilibration, it was necessary to
add a small percent of CD3OD and raise the temperature
at about 45 °C. After equilibration, only one product was
observed, namely the bis-adduct 12.
In the case of the divalent cations 1 and 2, a two-step
process can be envisaged as shown in Scheme 3.
The absorbance as a function of the methoxide ion
concentration is given by eq 1
csub(ꢀsub + ꢀ2HK1[MeO-])
1 + K1[MeO-] + K1K2[MeO-]2
A )
(1)
where csub is the analytical concentration of the substrate
(1 or 2) and ꢀsub and ꢀ2H are the molar extinction
coefficients of the substrate and of the monoadduct (23
or 24), respectively. It should be noted that since the
absorption at the chosen wavelength is only due to the
The reaction of the tris-cations 3 and 4 is in principle
more complicated because four constitutional isomers
can form, namely the tris-pyrans 15-18 and the tris-
thiopyrans 19-22 (Chart 2).
1
Indeed, at low temperature, the H NMR spectrum of
the tris-pyrylium 3 in CD3ONa/CD3OD solution showed
an indiscernible mixture of isomers. It was only possible
to evaluate the global ratio between 2H- and 4H-pyranyl
rings, equal to 1:5. Upon raising the temperature at 25
°C, the mixture completely converted into a unique
product, namely the tris-pyran 15. In contrast, the
reaction of the tris-thiopyrylium 4 appeared much sim-
pler to decipher: at low temperature the 1H NMR
revealed only the presence of the tris-4H adduct 22,
whereas at 25 °C only the thermodynamically more
stable tris-2H adduct 19 was detected.
In conclusion all the examined cations react with
CD3ONa/CD3OD at -40 °C to yield a kinetically con-
FIGURE 1. Absorbance at λ ) 370 nm of a 2.2 × 10-5
M
solution of 2 as a function of methoxide ion concentration.
(5) Boldt, P.; Bruhnke, D.; Gerson, F.; Scholz, M.; Jones, P. G.; Ba¨r,
F. Helv. Chim. Acta 1993, 76, 1739.
(6) Gionis, V.; Strzelecka, H. Synth. Commun. 1984, 14, 775.
(7) Doddi, G.; Ercolani, G. J. Chem. Soc., Perkin Trans. 2 1986, 271.
Filled and empty circles refer to experiments carried out in
the presence of ClCH2CO2H/ClCH2CO2- and PhCO2H/PhCO2
-
buffers, respectively; the best fit curve is calculated by eq 1.
6424 J. Org. Chem., Vol. 70, No. 16, 2005

Reactions of Bis- and Tris-pyrylium and -thiopyrylium Cations
SCHEME 4
pyrylium or thiopyrylium cromophore, the molar extinc-
tion coefficients ꢀsub and ꢀ2H were expected to be ap-
proximately in a 2:1 ratio; this was successively veri-
fied experimentally. In the case of the bis-pyrylium cation
1, the methoxide ion concentration in solution was
modulated by Cl3CCO2H/Cl3CCO2- and by ClCH2CO2H/
TABLE 1. Equilibrium Constants for the Reaction of
Cations 1-4 with MeO- in MeOH at 25.0 °Ca
cations
log K1
log K2
log K3
1
2
3
4
12.6
9.4
13.1
10.0
10.1
7.2
11.3
8.4
9.7
6.6
-
ClCH2CO2 buffers, while in the case of the bis-thio-
pyrylium cation 2, ClCH2CO2H/ClCH2CO2- and PhCO2H/
a Estimated error is ( 0.2 log units.
-
PhCO2 buffers were used (see the Experimental Sec-
tion). The experimental absorbance as a function of
methoxide ion concentration was fitted by nonlinear
least-squares to eq 1 by considering csub and ꢀsub as known
parameters and ꢀ2H, K1, and K2 as adjustable parameters.
For the sake of illustration, the experimental points and
the calculated curve in the case of the bis-thiopyrylium
cation 2 are reported in Figure 1.
In the case of the trivalent cations 3 and 4, the addition
of methoxide ion occurs in three steps as shown in
Scheme 4.
adducts that equilibrates at 25 °C to yield, in all of the
cases, the more stable 2H-adducts exclusively. This
pattern of reactivity had been previously observed in
simple pyrylium and thiopyrylium cations.7-11 Indeed,
previous studies of the reaction of pyrylium and thio-
pyrylium cations with methoxide ion in methanol showed
that (i) pyrylium ions are more reactive than the corre-
sponding thiopyrylium ions from both a kinetic and a
thermodynamic point of view and (ii) the kinetically
favored product is not always that thermodynamically
more stable. The first observation finds justification in
the higher carbocationic character of the pyrylium ring,
which in turn is due to the higher electronegativity of
oxygen. The second observation indicates that the transi-
tion states are significantly different from the final
products. Indeed, there are evidences suggesting that the
rates of nucleophilic attack are dominated by coulombic
interactions,12 whereas the thermodynamic regioselec-
tivity is dominated by the greater stability of the 2H-
adducts. Indeed, MNDO and AM1 calculations showed
that 2-methoxy-2H-pyrans are much more stable than
the corresponding 4H-isomers because of negative hyper-
conjugative effects between the geminal oxygen atoms.13
This effect is less important in 2-methoxy-2H-thiopyrans
as shown by the lower equilibrium constants for their
formation.8-11
Accordingly, the equation yielding the absorbance as
a function of the methoxide ion concentration is as follows
A )
csub(ꢀsub + ꢀ2HK1[MeO-] + ꢀ2H,2HK1K2[MeO-]2)
1 + K1[MeO-] + K1K2[MeO-]2 + K1K2K3[MeO-]3
(2)
where csub is the analytical concentration of the substrate
(3 or 4) and ꢀsub, ꢀ2H, and ꢀ2H,2H are the molar extinction
coefficients of the substrate, of the mono-adduct (25 or
27), and of the bis-adduct (26 or 28), respectively. In this
case, the molar extinction coefficients ꢀsub, ꢀ2H, and ꢀ2H,2H
were expected to be approximately in a 3:2:1 ratio; this
was successively verified experimentally. In the case of
the tris-pyrylium cation 3, the methoxide ion concentra-
tion in solution was modulated by HClO4 and by
-
-
Cl3CCO2H/Cl3CCO2 and ClCH2CO2H/ClCH2CO2 buff-
ers, while in the case of the tris-thiopyrylium cation 4,
Cl2CHCO2H/Cl2CHCO2-, ClCH2CO2H/ClCH2CO2-, and
PhCO2H/PhCO2- buffers were used (see the Experimen-
tal Section). The experimental absorbance as a function
of methoxide ion concentration was fitted by nonlinear
least-squares to eq 2 by considering csub and ꢀsub as known
parameters and ꢀ2H, ꢀ2H,2H K1, K2, and K3 as adjustable
parameters. The obtained equilibrium constants for the
reaction of cations 1-4 are reported in Table 1.
The equilibrium constants reported in Table 1 provide
the first quantitative data about the thermodynamic
reactivity of polyvalent pyrylium and thiopyrylium cat-
ions and merit to be discussed in some detail. The
sequential addition of methoxide ion to the R positions
(8) Doddi, G.; Ercolani, G. J. Org. Chem. 1986, 51, 4385.
(9) Di Vona, M. L.; Doddi, G.; Ercolani, G.; Illuminati, G. J. Am.
Chem. Soc. 1986, 108, 3409.
(10) Doddi, G.; Ercolani, G. J. Org. Chem. 1988, 53, 1729.
(11) Doddi, G.; Ercolani, G. J. Chem. Soc., Perkin Trans. 2 1989,
1393.
(12) Cerichelli, G.; Doddi, G.; Ercolani, G. Gazz. Chim. Ital. 1988,
118, 291.
(13) Doddi, G.; Ercolani, G.; Mencarelli, P. J. Org. Chem. 1992, 57,
4431.
Discussion
Cations 1-4 react with CD3ONa/CD3OD at -40 °C to
yield a kinetically controlled mixture of 2H- and/or 4H-
J. Org. Chem, Vol. 70, No. 16, 2005 6425

Branchi et al.
CHART 3
of chalcogenopyrylium rings is well differentiated in free
energy terms; i.e., the first addition is significantly more
favored than the second one, and in the case of the tris-
cations 3 and 4, the second addition is also significantly
more favored than the third one. This is principally due
to the fact that the chalcogenopyrylium ring, because of
its positive charge, behaves as an electron-withdrawing
substituent, whereas the neutral 2H-chalcogenopyranyl
ring is expected to be less effective. Thus, in the case of
bis-chalcogenopyrylium ions, for example, the addition
of the first methoxide ion is favored by the electron-
withdrawing effect of the other chalcogenopyrylium ring
not undergoing nucleophilic attack, whereas the addition
of the second methoxide ion is only slightly affected by
the electronic behavior of the 2H-chalcogenopyranyl ring
formed in the previous step. Another effect, although of
secondary importance, favoring the first attack over the
second one, and, in the case of the tris-cations, also the
second attack over the third one, is the statistical factor.
Indeed, considering the bis-chalcogenopyrylium ions, the
first equilibrium is favored by a factor 2 because there
are two equivalent rings that can undergo the attack,
whereas the second equilibrium is disfavored by a factor
2 (or favored by a factor 1/2), because in the reverse
reaction, both the 2H-chalcogenopyranyl rings can re-
lease methoxide ion. In the case of the tris-chalcogeno-
pyrylium ions, for the same reasons, the first equilibrium
is favored by a factor 3, the second equilibrium by a factor
1, and the third equilibrium by a factor 1/3.
Comparing the equilibrium constants of the bis-
pyrylium cation 1 with those of the bis-thiopyrylium
cation 2, as well as those of the tris-pyrylium cation 3
with those of the tris-thiopyrylium cation 4, it is evident,
as expected, the higher reactivity of pyrylium with
respect to thiopyrylium. However it is useful to go deep
in the analysis of the electronic effects by comparing the
obtained results with those previously obtained for the
series of 2,6-di-tert-butyl-4-arylpyrylium, 29-35,8 and
thiopyrylium 36-42 (Chart 3).9
As shown in Figure 2, the equilibrium constants for
the addition of methoxide ion to the R position of cations
29-42 are linearly correlated to the Brown σ+ values,14
with the following parameters: pyrylium series, slope (F)
) 2.55, intercept (log K0) ) 10.2, corr coeff. (r2) ) 0.987;8
FIGURE 2. Plots of log K for the addition of MeO- to the R
position of pyrylium ions 29-35 (filled circles) and thio-
pyrylium ions 36-42 (empty circles).
thiopyrylium series, slope (F) ) 1.99, intercept (log K0)
) 7.4, corr coeff (r2) ) 0.985.9
By correcting the equilibrium constants K1 of the bis-
cations in Table 1, for the statistical factor 2, and
interpolating from the corresponding σ+ correlation, one
obtains the value of σ+ ) 0.81 for the 2,6-di-tert-
p
butylpyrylium ring and of σ+ ) 0.84 for the 2,6-di-tert-
p
butylthiopyrylium ring. The data reveal that both pyr-
ylium and thiopyrylium moieties behave as good electron-
withdrawing substituents, whose effect is comparable to
that of the nitro group (σ+p ) 0.79).14 The effect is mainly
due to the positive charge on the two heteroaromatic
rings, with a negligible effect of the nature of the ring
heteroatom.
Analogously, by correcting the equilibrium constants
K2 of the bis-cations in Table 1, for the statistical factor
1/2, and interpolating from the corresponding σ+ correla-
tion, one obtains the value of σ+ ) 0.06 for the 2,6-di-
p
tert-butyl-2-methoxy-2H-pyranyl and σ+ ) 0.03 for the
p
2,6-di-tert-butyl-2-methoxy-2H-thiopyranyl, showing that
both rings are substantially electron-neutral.
A similar approach can be applied to the tris-cations
by correcting the equilibrium constant K1 for the statisti-
cal factor 3 and considering the effect of the two chalco-
genopyrylium rings in the meta position as additive. In
this case, one obtains upon interpolation the value of σ+
m
) 0.47 for the 2,6-di-tert-butylpyrylium and σ+ ) 0.53
m
for 2,6-di-tert-butylthiopyrylium. Also in this case the
effect of the two rings is the same within the experimen-
tal errors. It is noteworthy, however, that σ+ is signifi-
p
cantly higher than σ+m (about 0.3 units), indicating that
a significant component of the electron-withdrawing
effect exerted in the para position by the chalcogeno-
pyrylium ring is due to resonance. This behavior is quite
unusual for electron-withdrawing substituents as gener-
ally the difference σ+ - σ+ does not exceed 0.1 unit.14
p
m
By correcting the equilibrium constant K3 of the tris-
cations for the statistical factor 1/3, and by considering
the effect of the two 2H-chalcogenopyrans rings in the
meta position as additive, one obtains upon interpolation
the value of σ+ ) -0.01 for the 2,6-di-tert-butyl-2-
m
(14) Smith, M. B.; March, J. March’s Advanced Organic Chemistry,
5th ed.; Wiley: New York, 2001; p 370.
methoxy-2H-pyranyl and σ+m ) -0.09 for the 2,6-di-tert-
6426 J. Org. Chem., Vol. 70, No. 16, 2005

Reactions of Bis- and Tris-pyrylium and -thiopyrylium Cations
4,4′-(1,4-Phenylene)bis(2,6-di-tert-butylpyrylium) Bis-
tetrafluoroborate (1). The bis-pentanedione 6 (2.5 g, 5 mmol)
and triphenylmethanol (2,86 g, 11 mmol) were dissolved in
73 mL of acetic anhydride by gently warming. To this mixture
was added dropwise under stirring a solution of 50% HBF4
(4.04 g, 23 mmol) in 48 mL of acetic anhydride, prepared by
adding the HBF4 in drops to the ice cooled acetic anhydride.
The reaction mixture was then kept at 50 °C for 30 m. After
the mixture was cooled to room temperature, diethyl ether was
added until complete precipitation of 1, which was filtered,
washed with diethyl ether, and purified by dissolution in the
least amount of dichloromethane followed by reprecipitation
with diethyl ether: 2.8 g of white crystals were obtained (yield
91%). Mp: 280-282 °C dec (lit.5 mp 270-272 °C dec). 1H NMR
(CD3CN): δ 1.60 (s, 36 H), 8.26 (s, 4H), 8.39 (s, 4H). 13C NMR
(CD3CN): δ 28.5, 40.3, 118.0, 131.8, 139.0, 167.0, 188.6. UV-
vis (CH3CN): log ꢀ (λmax ) 350 nm) ) 4.62. Anal. Calcd for
C32H44O2B2F8 (634.31): C, 60.59; H, 6.99. Found: C, 60.33;
H, 6.82.
4,4′-(1,4-Phenylene)bis(2,6-di-tert-butylthiopyr-
ylium) bisperchlorate (2). (CAUTION: organic perchlorates
are potentially explosive and must be handled with all neces-
sary precautions). A mixture of the bis-pentanedione 6 (1.8 g,
3.6 mmol), phosphorus pentasulfide (4.80 g, 10.8 mmol), and
lithium perchlorate (4.60 g, 43.2 mmol) in 35 mL of acetic acid
was refluxed for 3 h under stirring. The resulting mixture was
filtered, and the solid residue, after being washed with hot
acetic acid, was eliminated. The cooled filtrate was then added
with diethyl ether until complete precipitation of a solid that
was redissolved in the least amount of CH3CN and repre-
cipitated with diethyl ether to yield 1.77 g of a gray solid. The
1H NMR spectrum of this material was consistent with the
expected bis-thiopyrylium cation and the corresponding
pyrylium-thiopyrylium cation in a 10:1 molar ratio. Double
recrystallization of the crude solid from CH2Cl2 provided 1.05
g of 2 as yellow crystals. Mp: 280-282 °C dec. 1H NMR
(CD3CN): δ 1.70 (s, 36H), 8.29 (s, 4H), 8.81 (s, 4H). 13C NMR
(CD3CN): δ 31.3, 43.3, 131.8, 131.9, 141.5, 160.8, 186.7. UV-
vis (CH3CN): log ꢀ (λmax ) 368 nm) ) 4.65. HRMS: m/z )
246.1438 (M)2+. Anal. Calcd for C32H44O8S2Cl2 (691.72): C,
55.56; H, 6.41; S, 9.27. Found: C, 55.61; H, 6.54; S, 9.10.
1,1′,1′′-Benzene-1,3,5-triyltris(4,4-dimethylpent-1-en-3-
one) (7). Benzene-1,3,5-tricarbaldehyde (0.50 g, 3.1 mmol) was
added to a stirred mixture of pinacolone (1.85 g, 18.5 mmol),
EtOH (3.8 mL), and H2O (0.8 mL). After heating to 40 °C, a
solution of NaOH (0.037 g, 93 mmol) in H2O (2 mL) was added
dropwise. After 3 h, the precipitate was filtered, washed with
H2O and EtOH, and recrystallized from EtOH to yield 0.79 g
of 7 as white crystals (yield 63%). Mp: 172-174 °C. 1H NMR
(CDCl3): δ 1.24 (s, 27 H), 7.15 (d, J ) 15.5 Hz, 3H), 7.71 (d, J
) 15.5 Hz, 3H), 7.71 (s, 3H). 13C NMR (CDCl3): δ 26.2, 43.3,
122.3, 128.9, 141.3, 136.3, 203.8. IR (CHCl3): νmax(CdO) 1683
cm-1, νmax(CdC) 1609 cm-1. Anal. Calcd for C27H36O3 (408.58):
C, 79.37; H, 8.88. Found: C, 79.29; H, 9.00.
5,5′,5′′-Benzene-1,3,5-triyltris(2,2,8,8-tetramethyl-
nonane-3,7-dione) (8). A solution of 7 (0.57 g, 1.4 mmol) in
4 mL of toluene was added to a stirred suspension of sodium
amide (0.41 g, 10 mmol) and pinacolone (0.84 g, 8.4 mmol) in
4 mL of toluene, under a dry nitrogen atmosphere, at room
temperature. After 48 h, 8 mL of water was added and the
resulting mixture extracted with CHCl3. The organic extracts
were dried over CaCl2 and evaporated to dryness to yield a
crude solid. Recrystallization from EtOH gave 0.76 g of pure
8 as white crystals (yield 77%). Mp: 154-156 °C. 1H NMR
(CDCl3): δ 0.96 (s, 54 H), 2.71 (m, 12H), 3.65 (m, 3H), 6.83 (s,
3H). 13C NMR (CDCl3): δ 26.1, 35.7, 42.5, 43.9, 125.1, 144.3,
213.6. IR (CHCl3): νmax(CdO) 1704 cm-1. Anal. Calcd for
C45H72O6 (709.06): C, 76.23; H, 10.23. Found: C, 76.19; H,
10.42.
butyl-2-methoxy-2H-thiopyranyl, showing, also in this
case, that both of the rings are substantially electron-
neutral.
It is interesting to note that the constants K2 of the
tris-cations provide a useful test for the assumption of
additivity of the electronic effects of the groups in the
meta position. Indeed, by this assumption, the equilib-
rium constant K2 can be calculated from the σ+ of the
m
cationic 2,6-di-tert-butylchalcogenopyrylium ring [(σ+
) ]
m cat
and the σ+ of the 2,6-di-tert-butyl-2-methoxy-2H-chal-
m
cogenopyranyl ring [(σ+
m)2H], as shown in eq 3.
+
+
log K2 )F[(σm
)cat + (σm
)
2H] + log K0
(3)
The calculated values of log K2 for the tris-pyrylium 3
() 11.4) and the tris-thiopyrylium 4 () 8.3) cations are
in excellent agreement with the corresponding experi-
mental values reported in Table 1, thus supporting the
assumption of additivity of electronic effects.
Conclusions
The reaction of bis- and tris-pyrylium and thiopyrylium
cations with methoxide ion in methanol is a useful
reaction to probe the effect of the two chalcogenopyrylium
rings as substituents in the para and meta position,
respectively. Despite the different sensitivity to electronic
effects of pyrylium and thiopyrylium rings, as measured
by the different F values (see Figure 2), the two rings
have a similar electron-withdrawing effect with a σ+
∼
p
0.8 and a σ+m ∼ 0.5. Apart from the expected importance
of the inductive effect due to the positive charge, the
difference between these two values remarks the impor-
tance of the resonance contribution. In contrast, both the
neutral 2H-pyranyl and thiopyranyl rings have a negli-
gible effect as substituents, independent of the position,
para or meta, they occupy.
Experimental Section
Materials and Instrumentation. Solvents (A.R. grade)
were distilled before use. Deuterated solvents for NMR
spectroscopy and other chemicals (A.R. grade) were used as
received. Stock solutions of sodium methoxide were prepared
by dissolving the appropriate amount of clean sodium in either
methanol or methanol-d4 under an argon atmosphere. 1H and
13C NMR spectra were recorded on a 300 MHz spectrometer.
Chemical shifts are given in ppm from TMS as internal
standard. Melting points are uncorrected. 1,1′-(1,4-Phenylene)-
bis(4,4-dimethylpent-1-en-3-one)5 (5) and benzene-1,3,5-tri-
carbaldehyde15 were prepared by literature procedures.
5,5′-(1,4-Phenylene)bis(2,2,8,8-tetramethylnonane-3,7-
dione) (6). A solution of 5 (3.3 g, 11 mmol) in 25 mL of toluene
was added to a stirred suspension of sodium amide (2.15 g, 55
mmol) and pinacolone (4.4 g, 44 mmol) in 40 mL of toluene,
under a dry nitrogen atmosphere, at room temperature. After
2 h, water (40 mL) was added and the resulting mixture
extracted with CHCl3. The organic extracts were dried over
CaCl2 and evaporated to dryness to yield a crude solid.
Recrystallization from EtOH gave 4.40 g of pure 6 as white
crystals (yield 80%). Mp: 204-206 °C (lit.5 mp 210 °C). 1H
NMR (CDCl3): δ 0.99 (s, 36 H), 2.75 (m, 8H), 3.70 (m, 2H),
7.09 (s, 4H). Anal. Calcd for C32H50O4 (498.75): C, 77.06; H,
10.10. Found: C, 76.92; H, 10.06.
4,4′,4′′-Benzene-1,3,5-triyltris(2,6-di-tert-butylpyr-
ylium) Trisperchlorate (3). (CAUTION: Organic perchlo-
rates are potentially explosive and must be handled with all
(15) Fourmigue´, M.; Joahnssen, I.; Boubekeur, K.; Nelson, C.; Batali,
P. J. Am. Chem. Soc. 1993, 115, 3752.
J. Org. Chem, Vol. 70, No. 16, 2005 6427

Branchi et al.
the necessary precautions). The tris-pentanedione 8 (0.17 g,
0.24 mmol) and triphenylmethanol (0.206 g, 0.79 mmol) were
dissolved in 5.3 mL of acetic anhydride by gently warming.
To this mixture was added dropwise under stirred a solution
of 60% HClO4 (0.278 g, 1.66 mmol) in 3.5 mL of acetic
anhydride, prepared by adding the HClO4 in drops to the ice-
cooled acetic anhydride. The reaction mixture was kept at 50
°C for 30 min. After the mixture was cooled to room temper-
ature, diethyl ether was added until complete precipitation of
3, which was filtered, washed with Et2O, and purified by
dissolution in the least amount of dichloromethane followed
by reprecipitation with diethyl ether: 0.215 g of white needles
were obtained. Mp: 290-292 °C dec. 1H NMR (CD3CN): δ 1.64
(s, 54 H), 8.49 (s, 6H), 8.92 (s, 3H). 13C NMR (CD3CN): δ 28.5,
40.5, 118.8, 135.9, 137.3, 166.4, 189.0. UV-vis (CH3CN): log
1H NMR spectral data of the detected species are as follows:
9: δ 1.01 (s, 18H), 1.24 (s, 18H), 5.44 (s, 2H), 5.46 (s, 2H),
7.54 (s, 4H).
10: 2H-pyran moiety, δ 1.01 (s, 9H), 1.24 (s, 9H), 5.39 (s,
1H), 5.44 (s, 1H); 4H-pyran moiety, δ 1.22 (s, 18H), 4.65 (s,
2H), 7.37-7.49 (m, 4H).
12: δ 1.09 (s, 18H), 1.30 (s, 18H), 5.52 (s, 2H), 6.32 (s, 2H),
7.44 (s, 4H).
13: 2H-thiopyran moiety, δ 1.10 (s, 9H), 1.31 (s, 9H), 5.50
(s, 1H), 6.30 (s, 1H); 4H-thiopyran moiety, δ 1.27 (s, 18H), 5.46
(s, 2H), 7.35-7.43 (m, 4H).
14: δ 1.25 (s, 36H), 7.31 (s, 4H); note that H-3 and H-5
signals are hidden.
15: δ 1.03 (s, 27H), 1.25 (s, 27H), 5.43 (d, J ) 1.4 Hz, 3H),
5.44 (d, J ) 1.4 Hz, 3H), 7.46 (s, 3H).
18: δ 1.20 (s, 54H), 4.59 (s, 6H), 7.39 (s, 3H).
ꢀ (λmax) 306 nm) ) 4.95. HRMS: m/z ) 375.2116 (M + ClO4)2+
.
Anal. Calcd for C45H63O15Cl3 (950.35): C, 56.87; H, 6.68.
Found: C, 56.50; H, 7.03.
19: δ 1.10 (s, 27H), 1.32 (s, 27H), 5.55 (s, 3H), 6.34 (s, 3H),
7.40 (s, 3H).
22: δ 1.24 (s, 54H), 5.34 (s, 6H), 7.15 (s, 3H).
4,4′,4′′-Benzene-1,3,5-triyltris(2,6-di-tert-butylthio-
pyrylium) Trisperchlorate (4). (CAUTION: Organic per-
chlorates are potentially explosive and must be handled with
all necessary precautions). A mixture of the tris-pentanedione
8 (0.5 g, 0.7 mmol), phosphorus pentasulfide (1.41 g, 3.18
mmol), and lithium perchlorate (1.35 g, 12.7 mmol) in 11 mL
of acetic acid was refluxed for 3 h under stirring. The resulting
mixture was filtered, and the solid residue, after washing with
hot acetic acid, was eliminated. The cooled filtrate was added
with diethyl ether until complete precipitation of a solid that
was redissolved in the least amount of CH3CN and reprecipi-
UV-vis Spectrophotometric Determination of the
Equilibrium Constants. Determination of the equilibrium
constants was carried out by spectrophotometric titrations of
the given substrate at 25.0 °C in 10-mm quartz cuvettes.
Typically, 2 mL of a methanol solution of the chalcogeno-
pyrylium ion in the range (1-3) × 10-5 M and containing the
acid component of the given buffer in excess (from 1 × 10-3 to
2 × 10-2 M) was added with successive aliquots of a 0.1-0.4
M methanol solution of sodium methoxide by microliter
syringes. After each addition, allowance was made for thermal
equilibration, and then the absorbance value was recorded and
corrected for dilution. The actual methoxide ion concentration
after each addition was calculated by the composition of the
buffer, the known pKa of the buffer acid, and the methanol
autoprotolysis constant (pKMeOH ) 16.92 at 25 °C).16 The acids
used as buffers and their pKa in methanol are the following:
trichloroacetic acid (4.98);17 dichloroacetic acid (6.41);17 chloro-
acetic acid (7.7);18 benzoic acid (9.1).18 Only in one case, namely
substrate 3 at the lower pH values, were the required
methoxide ion concentrations in the cuvette attained by direct
addition of aliquots of a methanol solution of the strong acid
HClO4. The absorbance measurements were carried out at a
fixed wavelength, λ, where either the pyrylium or thiopyrylium
nucleus is the only chromophore absorbing. The value of the
molar extinction coefficients of the substrates, ꢀsub, at this
wavelength was previously determined by UV-vis spectra
carried out in the presence of 0.05 M HBF4. The values of λ in
nm and log ꢀsub for the substrates 1-4 are the following: 1,
354 (4.61); 2, 370 (4.66); 3, 320 (5.79); 4, 340 (6.75). The values
of the absorbances vs the corresponding methoxide ion con-
centrations were fitted by either eq 1 or 2 as further detailed
in the Results.
1
tated with diethyl ether to yield 0.53 g of gray solid. The H
NMR spectrum of this material was consistent with the
expected tris-thiopyrylium cation and the corresponding bis-
thiopyrylium-pyrylium cation in a 6:1 molar ratio. Triple
recrystallization of the crude solid from ethyl alcohol provided
0.147 g of pure 4 as yellow crystals. Mp: 286-288 °C dec. 1H
NMR (CD3CN): δ 1.72 (s, 54H), 8.70 (s, 3H), 8.99 (s, 6H). 13
C
NMR (CD3CN): δ 31.2, 43.4, 132.6, 134.8, 140.4, 160.0, 186.9.
UV-vis (CH3CN): log ꢀ (λmax) 322 nm) ) 4.88. HRMS: m/z
) 399.1804 (M + ClO4)2+. Anal. Calcd for C45H63O12S3Cl3
(998.53): C, 54.13, H, 6.36; S, 9.63. Found: C, 53.95; H, 6.43;
S, 9.53.
1H NMR Measurements. Low-temperature spectra were
recorded just after the addition of the substrate (ca. 0.02 mmol
of solid sample) to a ca. 0.5 mL of a 0.5 M solution of CD3ONa
in CD3OD, thermostated at ∼-40 °C in the NMR probe. Apart
from the case of the substrate 2 that showed the precipitation
of a white solid upon addition to the CD3ONa/CD3OD solution,
the 1H NMR spectra were also recorded at ∼25 °C after
equilibration of the reaction mixtures was attained.
In the case of the substrate 2, the precipitate was separated
by filtration, washed with CD3OD, dissolved in 0.5 mL of
CDCl3 at room temperature, and subjected to 1H NMR
analysis. The pattern of the signals was attributed to the
kinetically controlled formation of the adducts. After equilibra-
tion of the mixture, attained by adding 0.01 mL of CD3OD and
warming the solution up to 45 °C, a second 1H NMR spectrum
was recorded.
JO050884A
(16) Koskikallio, J. Suom. Kemistil. B 1957, 30, 111.
(17) Ritchie, C. D.; Virtanen, P. O. I. J. Am. Chem. Soc. 1972, 94,
4963.
(18) Clare, B. W.; Cook, D.; Ko, E. C. F.; Mac, Y. C.; Parker, A. J. J.
Am. Chem. Soc. 1966, 88, 1911.
Details on the composition of the reaction mixtures under
both kinetic and thermodynamic control are reported in the
Results.
6428 J. Org. Chem., Vol. 70, No. 16, 2005