Hydroxy-substituted triarylcarbeniuin bromides. Synthesis, structure, derivatization and facile conversion to highly substituted xanthenes

Article abstract of DOI:10.1039/b001032o

The tris(3,5-dialkyl-2-hydroxyphenyl)carbenium bromides (2a) (3-tert-butyl, 5-methyl) and 2b (3,5-di-tert-butyl) have been prepared by the oxidation of the corresponding tris(3,5-dialkyl-2-hydroxyphenyl)methanes with bromine. Compound 2a has been fully ch

Full text of DOI:10.1039/b001032o

View Article Online / Journal Homepage / Table of Contents for this issue
Hydroxy-substituted triarylcarbenium bromides. Synthesis,
structure, derivatization and facile conversion to highly substituted
xanthenes†
1
Maarten B. Dinger and Michael J. Scott*
Department of Chemistry, University of Florida, P.O. Box 117200, Gainesville,
FL 32611-7200, USA. Fax: 352-392-3255; E-mail: mjscott@chem.ufl.edu
Received (in Cambridge, UK) 7th February 2000, Accepted 22nd March 2000
Published on the Web 12th May 2000
The tris(3,5-dialkyl-2-hydroxyphenyl)carbenium bromides (2a) (3-tert-butyl, 5-methyl) and 2b (3,5-di-tert-butyl)
have been prepared by the oxidation of the corresponding tris(3,5-dialkyl-2-hydroxyphenyl)methanes with bromine.
Compound 2a has been fully characterized by a single crystal X-ray study, which shows that the bromide, atypically,
does not bind to the central carbon atom. The materials 2 are reactive, and the trityl alcohol (4a), methyl (7a, 7b) and
phenyl (8a) derivatives can be readily prepared in good yield by reaction of 2 with the appropriate anions; X-ray
structures for 4a, 7b (as its lithium salt) and 8a are reported. In solution, 2a and 4a undergo a facile, uncatalyzed,
controllable intramolecular dehydration to produce the corresponding xanthene derivatives; X-ray structures for
the xanthene carbocation (bromide salt) (3a) and the 9-hydroxy derivative (5a) are also reported.
groups can be derivatized using the well-established method-
Introduction
ology employed for the calixarene systems.1 In all of the com-
Recently, we have been exploring the chemistry of the tris-
(3,5-dialkyl-2-hydroxyphenyl)methanes (1a–1d) and have
found these C3 symmetric compounds to exhibit a pronounced
tendency to adopt a conformation both in solution and in the
pounds characterized to date, the three arms of the molecule all
point up with respect to the central methine hydrogen (Type 1,
Fig. 1a). In the solid state, 1b forms a very weakly associated,
hydrogen-bonded dimer (shortest H ؒ ؒ ؒ O contact 2.368 Å),
solid state wherein the oxygen donor atoms are all held in align-
and although NOE studies suggest the aromatic rings rotate
ment with the proton on the methine group linking the phenolic
freely about the central carbon linker in 1, the rotation ceases
arms.1 Due to the rigidity of this platform, these molecules
when the hydroxy groups are derivatized due to the steric bulk
share many traits with the calix[n]arene macrocycles and seem
of the tert-butyl groups at the periphery of the molecule. Prior
to exhibit similarly diverse chemistry. The peripheral hydroxy
to substitution of the hydroxy groups, however, the NMR data
intimate a second conformation can be adopted by 1 with the
hydroxy groups orientated in a direction opposite to the central
† Further details of the X-ray crystal structure analyses together with
methine hydrogen (Type 2, Fig. 1b). With this dramatic change
the complete numbering schemes used in the X-ray structures are avail-
in conformation, the Type 2 system should exhibit properties
able as supplementary data. For direct electronic access see http://
www.rsc.org/suppdata/p1/b0/b001032o/
markedly different from those of the Type 1 systems previously
Scheme 1
DOI: 10.1039/b001032o
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748
This journal is © The Royal Society of Chemistry 2000
1741

View Article Online
Fig. 1
characterized, and they may serve as interesting ligands for
transition metals. Intent on the synthesis of these molecules, we
pursued the transformation of the methine hydrogen in 1 to
other moieties.
Two methods, in particular, present themselves for derivatiz-
ing the central methine of triarylmethanes. First, the hydrogen
should be sufficiently reactive for it to be deprotonated by
strong bases, such as alkyllithiums, to form carbanions. This
carbanion can, in turn, react with alkylating agents, generating
the substituted product.2 Unfortunately this methodology is not
immediately applicable to 1 since the bases initially deprotonate
the phenolic oxygens to produce aggregates, in which the
methine hydrogen is protected in a cage like structure, blocking
its reaction.3 If the phenolic groups were first protected with an
unreactive moiety (such as a methyl group), this strategy may
succeed, although the subsequent deprotection step(s) may be
sterically difficult and/or often require very harsh conditions.4
The second route, via oxidation of the methine in 1 to produce
the carbocation (carbenium ion), was thus pursued. Herein we
report the results of these studies.
Fig. 2 Structure of 2aؒCHCl3 (30% probability ellipsoids). The com-
plete numbering scheme is depicted in the electronic supplementary
data. Selected bond lengths (Å) and angles (Њ): C(1)–C(2) 1.50(1);
C(1)–C(13) 1.50(1); C(1)–C(24) 1.40(1); C(1) ؒ ؒ ؒ Br 3.34(1); O(1) ؒ ؒ ؒ Br
3.151; O(2) ؒ ؒ ؒ Br 3.132; O(3) ؒ ؒ ؒ Br 3.178; O(1)H ؒ ؒ ؒ Br 2.572;
O(2)H ؒ ؒ ؒ Br 2.458. O(1)–H–Br 124.5; O(2)–H–Br 133.7.
2aؒCHCl3 is shown in Fig. 2, while crystal and data collection
parameters are given in Table 1.
In the solid state, the bromine is not bonded to the central
carbon atom, but resides at a distance of 3.34(2) Å, and forms
closer approaches [3.13(2), 3.14(2) and 3.16(2) Å] to the phen-
olic oxygens. For the structure of triphenylmethyl bromide, an
average distance (3 independent molecules) of 2.009 Å was
measured.7 The absence of a carbon–bromine bond in 2a
almost certainly can be attributed to an electronic effect, since
the steric constraints at the other side of the molecule are such
that a bromine atom could be accommodated. Our inability to
prepare the fluorinated compound, but success in the synthesis
of the methyl- and phenyl-substituted materials (vide infra),
further supports this notion. The carbocation is not completely
planar, with the central carbon C(1) slightly puckered 0.104 Å
above the plane defined by C(2), C(13) and C(24), in contrast
to the average distance of 0.40(2) Å measured for C(1) in the
structure of 1a.8
Although the data were not of sufficient quality to locate
with certainty the absolute positions of the phenolic hydrogen
atoms, small peaks were located in the difference Fourier maps
near the oxygens, and these were assigned as hydrogens and
refined. In two cases, the position for the hydrogen atoms
refined in an orientation directing towards the bromine with
H ؒ ؒ ؒ Br contacts of 2.57(1) and 2.46(1) Å for O(1)H ؒ ؒ ؒ Br and
O(2)H ؒ ؒ ؒ Br, respectively. The third hydrogen refined at a
location 3.07(2) Å distant from the bromine. As in 1b, the aro-
matic rings are canted with respect to the core [C(2), C(13) and
C(24)] by 37.9(4), 40.9(4) and 41.2(3)Њ for the rings bound to
O(1), O(2) and O(3), respectively. Somewhat larger angles of
42.96(4), 46.81(4) and 49.75(4)Њ were measured for the three
rings of the three independent molecules in the structure of 1a,
and these are typical twist angles for triarylmethyl cations.9 Two
very disordered chloroform solvents each with site occupancy
of one half were also located in electron density maps, but these
were not involved in any interactions with 2a.
Results and discussion
An examination of the literature reveals oxidations of triaryl-
methanes are readily accomplished by means of nitrosium
tetrafluoroborate.5 Although a rapid reaction was clearly
noted, all attempts to use this methodology with 1a failed,
and only extensive decomposition was observed. Instead, a
procedure was developed based loosely on early research con-
cerning the reaction of bis(2-hydroxyphenyl)phenylmethane
with bromine to give the corresponding pentabromo species.6
Since the positions ortho and para to the hydroxy groups are
blocked in 1, we reasoned this methodology can only lead to
reaction at the central carbon position. Thus, the reaction of
tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)methane 1a with an
excess of bromine produces a rapid intense dark green color-
ation. After 2 hours, and removal of the solvent and excess
bromine, the carbocation 2a was formed in essentially quanti-
tative yield.
Although 2a was completely stable in the solid state and
could be stored for long periods in air at room temperature,
solutions of 2a were moderately unstable, and decomposition
to 9-(3-tert-butyl-5-methyl-2-hydroxyphenyl)-4,5-di-tert-butyl-
2,7-dimethylxanthen-9-ylium bromide (3a) (vide infra) was
observed spectroscopically. An analogous compound derived
from tris(3,5-di-tert-butyl-2-hydroxyphenyl)methane (1b) was
also prepared, yielding the corresponding adduct 2b when
treated with bromine. Unfortunately, solutions of this material
were found to be even less stable than 2a in solution.
Since triphenylmethyl bromide is a colorless solid with the
bromine covalently bound to the central carbon,7 the very
intense coloration for 2 implied the bromine may not be
associated with the central carbon, so we undertook a crystallo-
graphic analysis of the system. Slow evaporation of a chloro-
form solution of 2a at 4 ЊC yielded very dark green blocks
suitable for X-ray structural analysis. The structure of
Solutions of 2 slowly decompose in unreactive organic
solvents, and complete decomposition was effected in one day
at room temperature to give a new, dark orange material. None-
theless, NMR data for 2 could be readily obtained. From a
1
comparison of the H NMR spectrum of 2a with that of the
1742
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748

View Article Online
Table 1 X-Ray data for the presented crystal structures
7b Lithium
2
–
1a
2aؒCHCl3
3aؒCH2Cl2
4aؒ2Me2CO
5a
saltؒ43THF
8aؒ0.8Me2CO
Total reflections
Unique reflections
R(int)
Temperature/K
Chemical formula
Formula weight
Crystal system
Space group
µ(Mo-Kα)/mmϪ1
a/Å
b/Å
c/Å
α/Њ
β/Њ
28036
6209
0.0582
173(2)
19643
6266
18656
5935
0.0752
173(2)
C35H45O2Cl2Br
648.52
Orthorhombic
P212121
1.406
9.3103(5)
12.5249(7)
28.920(2)
—
10563
6651
9535
6746
25905
8046
0.2046
172(3)
C62.67H100.33O7.67Li3
997.27
Monoclinic
P21/c
4408
4263
0.0566
173(2)
0.0631
173(2)
C35H46O3Cl3Br
700.98
Orthorhombic
P21212
1.415
19.085(1)
19.464(1)
9.5618(5)
—
0.0339
173(2)
C40H58O6
634.86
Triclinic
0.0229
173(2)
C34H44O3
500.69
Triclinic
C34H46O3
502.71
Hexagonal
C42.4H54.8O3.8
625.26
Triclinic
¯
¯
¯
¯
P3
P1
P1
P1
0.068
22.3915(7)
—
10.4686(5)
—
—
—
0.073
0.071
0.068
0.066
10.2038(6)
13.5294(8)
14.9405(9)
96.034(1)
94.077(1)
110.752(1)
1905.0(2)
2
11.1823(6)
12.6574(7)
12.8089(7)
103.779(1)
108.435(1)
111.672(1)
1463.1(1)
2
20.984(1)
10.8817(7)
27.237(2)
—
98.486(2)
—
11.956(2)
12.740(2)
13.430(2)
97.043(5)
90.828(4)
105.813(3)
1951.0(5)
2
—
—
3552.0(3)
4
—
—
3372.4(3)
4
γ/Њ
Vc/Å3
4545.5(3)
6
6151.2(7)
4
Z
R1 [I ≥ 2σ(I) data]
wR2 (all data)
0.0480 [4116]
0.1111
0.0777 [5434]
0.2043
0.0471 [5060]
0.1109
0.0676 [4830]
0.1667
0.0487 [5849]
0.1378
0.0882 [3153]
0.2676
0.0738 [2473]
0.2028
1
Fig. 3 Comparison of H NMR (300 MHz) spectra of 1a, 2a and 3a
in CDCl3 solvent. Residual CHCl3 is denoted by *.
starting material 1a (Fig. 3), the most notable change, besides
the absence of the methine proton, was the large upfield shift
(4.43 ppm) of the hydroxy proton signals to 9.22 ppm, possibly
due to the electron deficient delocalized ring system. In the 13
C
Fig. 4 ORTEP diagram of 3aؒCH2Cl2, normal to the xanthene plane.
The complete numbering scheme is depicted in electronic supplemen-
tary data. Selected bond lengths (Å) and angles (Њ): C(1)–C(2) 1.492(5);
C(1)–C(13) 1.397(5); C(1)–C(24) 1.407(5); C(14)–O(2) 1.360(4); C(25)–
O(2) 1.352(4); C(1) ؒ ؒ ؒ Br 3.467(7); O(1) ؒ ؒ ؒ Br 3.161(7); O(1)H ؒ ؒ ؒ Br
2.384(6). O(1)–H–Br 166.2(3); C(13)–C(1)–C(24) 119.8(3); C(14)–O(2)–
C(25) 121.8(3).
NMR spectrum, the central carbon appeared at 195.4 ppm, in
sharp contrast to the value of 42.4 ppm measured for 1a.1 Addi-
tionally, the resonance for the ipso phenolic carbon in 2a was
found at 161.3 ppm, a position 10.2 ppm upfield relative to 1a.
As 2a slowly decomposed, it was clear from 1H NMR
spectroscopy that only one product was formed and that this
new material was no longer C3 symmetric, since only two of
the aromatic rings were equivalent. In addition, the phenolic
hydrogen resonance, located at 8.41 ppm, revealed only one
proton on integration. Addition of methanol directly to solid
2a greatly facilitated the reaction, with complete decomposition
of 2a effected in only a few seconds. In order to determine
the absolute conformation and connectivity of this material, a
crystal structure determination was undertaken (Fig. 4).
The structure revealed the product to be a 9-arylxanthen-
ylium bromide (3a), resulting from an intramolecular dehydra-
tion involving two of the phenolic rings. The reaction was
surprisingly facile, even at Ϫ20 ЊC. Condensations of this type
have never been observed for the starting methane 1a, which,
like bis(2-hydroxyphenyl)methane,10 has been found to be stable
with respect to cyclodehydration. Typically, 9-substituted
xanthenes are synthesized by the cyclization of ortho-methoxy-
substituted triarylmethanols facilitated by pyridinium chloride
at 205 ЊC,11 or by reaction of xanthone with a Grignard
reagent.12 Furthermore, 9-substituted xanthene derivatives have
been isolated in low yield from certain reactions of various
substituted phenoxymagnesium bromides with triethyl
orthoformate.13 In this particular system, the cyclization was
suggested to proceed via a pathway involving carbocations
analogous to 3a, which were not isolated. The condensation
observed in the present work lends additional validity to the
proposed scheme.
Although a number of 9-arylxanthenylium salts have been
structurally characterized,14 3a represents the first example to
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748
1743

View Article Online
Fig. 6 Structure of 5a (30% probability ellipsoids). The complete
Fig. 5 Structure of 4aؒ2Me2CO (30% probability ellipsoids). The
remaining acetone solvate has been omitted, for clarity. The complete
numbering scheme is depicted in the electronic supplementary data.
Selected bond lengths (Å) and angles (Њ): C(1)–O(4) 1.474(3);
O(1) ؒ ؒ ؒ O(4) 2.605(6); O(2) ؒ ؒ ؒ O(4) 2.649(6); O(4) ؒ ؒ ؒ O(3) 2.601(6);
O(3) ؒ ؒ ؒ O(100) 2.747(6); O(1)H ؒ ؒ ؒ O(4) 1.783(5); O(2)H ؒ ؒ ؒ O(4)
1.848(5); O(4)H ؒ ؒ ؒ O(3) 1.925(5); O(3)H ؒ ؒ ؒ O(100) 2.062(5). O(1)–H–
O(4) 143.9(3); O(2)–H–O(4) 153.5(3); O(4)–H–O(3) 130.0(3); O(3)–H–
O(100) 151.4(3).
numbering scheme is depicted in the electronic supplementary data.
Selected bond lengths (Å) and angles (Њ): C(1)–O(3) 1.458(1);
O(2) ؒ ؒ ؒ O(3) 2.540(2); O(3)H ؒ ؒ ؒ O(2) 1.749(2); C(1)–C(2) 1.458(1);
C(1)–C(13) 1.512(2); C(1)–C(24) 1.509(2); O(1)–C(3) 1.368(2); C(14)–
O(1) 1.388(1); C(25)–O(2) 1.391(1). O(3)–C(1)–C(2) 107.74(9); C(1)–
O(3)–H 148.4(1); C(13)–C(1)–C(24) 111.54(9); C(14)–O(2)–C(25)
120.01(9).
suitable for crystallographic analysis were grown from a
saturated acetone solution of 4a at Ϫ20 ЊC. An ORTEP
diagram of 4a is shown in Fig. 5, while collection and crystal
data is presented in Table 1.
contain a 9-hydroxyphenyl substituent. Unlike the structure of
2a, the core, defined by C(1), C(2), C(13) and C(24) is abso-
lutely planar, with no atom deviating from the least-squares
plane drawn through these atoms by more than 0.008(3) Å for
C(1). In the solid state, the bromine is not situated directly
above C(1), but rather, in reference to Fig. 4 which is drawn
normal to the xanthene plane, it is shifted significantly up, and
slightly to the left, while residing ~3.32 Å above the ring system.
Additionally, the hydroxy hydrogen atom (located and refined)
hydrogen-bonds to the bromine, with a fairly short H ؒ ؒ ؒ Br
contact of 2.384(6) Å. Consistent with the dark coloration of
3a, the xanthene ring system is completely delocalized, as
evidenced by the similar bond lengths and angles of the
atoms involved, though the xanthene ring system is not com-
pletely planar, with a significant twist of 14.8(9)Њ between the
planes defined by C(13)–C(23) and C(24)–C(34). If a com-
pletely planar conformation were realized, there would be a
significant steric interaction between the two xanthene tert-
butyl groups, for which a dihedral angle of 29.1Њ is measured.
The phenolic ring is also twisted relative to the xanthene plane
by 72.5(1)Њ.
From an examination of the structure, it was immediately
apparent that the compound is orientated in a Type 1 fashion.
All four hydroxy hydrogens were located in difference maps and
refined without constraint, and these OHs are all involved in
hydrogen-bonding interactions. Two of the phenolic protons,
O(1)H and O(2)H, form hydrogen-bonds to the central
hydroxy, with O ؒ ؒ ؒ H contacts of 1.783(5) and 1.848(5) Å,
respectively, while the proton on the central hydroxy forms a
weaker bond to O(3) at a distance of 1.925(5) Å. The remain-
ing hydroxy group is hydrogen-bonded to a molecule of
acetone solvent with an O(3)H–O(100) contact of 2.062(5) Å.
The central carbon–oxygen distance [C(1)–O(4)] is typical,
1.474(3) Å; the related compounds tris(2-methoxyphenyl)-
methanol [1.446(2) Å]15 and tris(4-methoxyphenyl)methanol
[1.459(3) and 1.441(3) Å]16 show similar O–C bond lengths. The
aromatic rings are not perpendicular with respect to the core
[defined by C(2), C(13) and C(24)], but instead are canted by
48.5(1), 47.64(8) and 48.8(1)Њ for the rings attached to O(1),
O(2) and O(3), respectively.
Solutions of 4a were unstable and decomposed over the
course of a few days at room temperature to produce the
9-arylxanthene (5a), identified by the loss of C3 symmetry in
the NMR spectrum and by analogy with the spectrum of 3a.
The compound 5a was also subjected to complete characteriz-
ation by a single crystal X-ray analysis (Fig. 6).
As depicted in Fig. 6, the phenolic hydrogen atom interacts
with the central hydroxy group at a distance of 1.749(2) Å, but
the central hydroxy proton does not exhibit any hydrogen
bonding. As for 3a, the two aromatic rings of the xanthene
moiety in 5a are not coplanar, but twisted by 13.80(6)Њ. The
phenolic ring cants relative to the xanthene plane at an angle of
78.67(3)Њ. In contrast to the structure of 3a, no delocalization
of the xanthene group is evident, with C(1)–C(13) and C(1)–
C(24) distances 0.1 Å longer than in 3a. The bonds formed to
Reactivity of tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)-
carbenium bromide (2a)
The carbocation 2a should be a reactive material, more so even
than triphenylmethyl bromide, and therefore may serve as a
useful precursor for the derivatization of the central carbon
atom. With this in mind, a number of different reagents were
added to the material. When 2a, dissolved in chloroform, was
shaken with saturated sodium bicarbonate, the organic phase
rapidly turned from dark green to pale yellow, and the C3 sym-
metric triarylmethanol (4a) was produced. From spectroscopic
data, it was difficult to determine if the preferred conformation
of 4a was Type 1 or Type 2 (Fig. 1), and for this reason, as well
as to examine hydrogen-bonding interactions, the solid-state
structure of the material was determined. Colorless blocks
1744
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748

View Article Online
O(2) are also slightly extended (0.03 Å), but the remaining
structural features are typical.
The 1H NMR spectrum of 5a is similar to that of 3a, except
for a large shift upfield for the 2-hydroxyphenyl proton from
8.41 in 3a to 10.29 ppm in 5a, likely due to the strong hydrogen-
bonding interaction between this proton and the central
hydroxy group. For comparison, this resonance can be found at
7.15 ppm for 4a and 4.79 ppm for 1a. Lacking a hydrogen bond,
the central hydroxy resonance is further downfield at 3.22 ppm;
the central hydroxy in 4a appears at 6.67 ppm. In the 13C NMR
spectra, the central carbon resonances occur at 87.7 and 76.7
ppm for 4a and 5a, respectively. The IR spectrum of 5a exhibits
two main stretches at 3529 and 3295 cmϪ1, and these bands may
be assigned to the 9-hydroxy and phenolic hydroxy groups,
respectively, based on the hydrogen-bonding criterion.
In an attempt to add a methoxy group to the carbocation,
2a was reacted with an excess of sodium methoxide, but this
reaction cleanly converted 2a to the 9-aryl-9-methoxyxanthene
derivative (6a). In all likelihood, the triarylmethoxymethane
formed as an intermediate in this reaction, however this
material was neither isolated nor detected in the procedure.
Presumably, in a manner analogous to 5a, the phenolic hydroxy
hydrogen bonds to the methoxide, since the resonance for this
proton is found at almost the same position (10.54 ppm) as 5a
in the 1H NMR spectrum, and the infrared spectrum exhibits a
2
–
Fig. 7 Structure of 7b lithium saltؒ43THF (20% probability ellips-
oids). The non-coordinating THF solvate, the tert-butyl methyl groups,
and the THF hydrogens have been omitted for clarity. The triarylmeth-
ane fragment is drawn with solid lines, while THFs are represented by
open bonds. The complete numbering scheme is depicted in the elec-
tronic supplementary data. Selected bond lengths (Å) and angles (Њ);
C(1)–C(44) 1.568(8); C(1)–C(2) 1.547(8); C(1)–C(16) 1.575(9); C(1)–
C(30) 1.555(9); Li(1)–O(1) 1.76(1); Li(2)–O(2) 1.91(1); Li(2)–O(3)
1.97(1); Li(3)–O(2) 1.86(1); Li(3)–O(3) 1.87(1); Li(1)–O(101) 1.89(1);
Li(2)–O(201) 2.14(1); Li(2)–O(401) 1.96(1); Li(3)–O(301) 1.92(1);
Li(1) ؒ ؒ ؒ C(30) 2.36(1); Li(1) ؒ ؒ ؒ C(31) 2.48(2); Li(3) ؒ ؒ ؒ C(2) 2.23(1);
Li(3) ؒ ؒ ؒ C(15) 2.53(1). O(1)–Li(1)–O(101) 116.4(8); O(2)–Li(2)–O(201)
113.4(6); O(2)–Li(2)–O(401) 117.3(7); O(2)–Li(3)–O(301) 114.3(6);
O(3)–Li(3)–O(301) 109.4(6); Li(2)–O(2)–Li(3) 80.1(6); Li(2)–O(3)–
Li(3) 78.2(6); Li(1)–O(1)–C(3) 145.0(6); Li(3)–O(2)–C(17) 128.6(5);
Li(2)–O(2)–C(17) 112.9(6); Li(2)–O(3)–C(31) 110.5(6); Li(3)–O(3)–
C(31) 130.0(6).
similar OH stretch at 3243 cmϪ1
.
We also explored the synthesis of the fluorine analog of 1a,
both by the standard methodology, and by the reaction of 2a
with hydrogen fluoride. Normally, triarylmethanes can be
readily fluorinated in a one-pot reaction with nitrosium
terafluoroborate and pyridinium poly(hydrogen fluoride)
(PPHF).5 However, when 1a was reacted in this manner, the
mixture turned dark brown and a large number of products
were formed. Since oxidation of triphenylmethane with NOBF4
produced a carbocation, we reasoned that 2a should be a suit-
able substitute, so a large excess of PPHF was added to a
dichloromethane solution of 2a. After work-up, which involved
neutralizing the PPHF with sodium bicarbonate solution, 4a
was identified as the major product, with a small amount of 5a
also present. Steric constraints do not seem to be sufficient as to
prevent fluorine from covalently binding to the central carbon,
but the hydroxy groups may preclude the addition of the very
electronegative atom by rendering the central carbon electron
deficient.
readily decomposed to a variety of uncharacterized compounds
in only a few minutes. For these reasons, crystallization
attempts of the initially formed lithium salts prepared in the
synthesis of 7 were carried out. A small number of crystals were
eventually isolated after slow evaporation of solvent from a
THF solution of the lithium salt of 7b, and this compound was
characterized by X-ray crystallography (Fig. 7).
Addition of a hydrophobic moiety at the central carbon was
thought to possibly induce inversion to a Type 2 system, and a
suitable procedure to substitute various groups at this position
was designed. Treatment of triphenylmethyl chloride with
Grignard reagents readily produced various 1,1,1-triphenyl-
alkanes or -aryls,17 and 4,8,12-trioxa-4,8,12,12c-tetrahydrodi-
benzo[cd,mn]pyrene (trioxatricornan) derivatives can be pre-
pared by reaction of the corresponding tetrafluoroborate salt
with methyllithium.18 Although the compounds 2 have reactive
hydroxy groups, we anticipated that the addition of at least four
equivalents of a metal alkyl or aryl should allow for reaction at
the central carbon. Accordingly, when 2a or 2b, dissolved in
ether, were reacted with six equivalents of methyllithium, the
dark green solutions slowly turned yellow–orange. After
addition of water, and separation, yellow residues were col-
lected, containing the corresponding 1,1,1-tris(3,5-dialkyl-2-
hydroxyphenyl)ethanes (7a and 7b).
Inspection of the structure immediately reveals that the con-
formation of the organic fragment now closely resembles that
of Type 2. In sharp contrast, the lithium salt of 1b is a hexa-
nuclear complex, comprising of two organic fragments adopt-
ing a Type 1 (Fig. 1a) conformation, with the two methine
hydrogens pointing directly at one another in the cage-like
structure.3 As seen in Fig. 7, two of the aromatic rings, with the
corresponding hydroxy groups O(2) and O(3), are canted
down by 64.4(2) and 65.3(3)Њ, respectively, with respect to the
central methyl group [C(44)], while the remaining ring is canted
just slightly up [10.9(3)Њ]. Concomitant with this observation,
while Li(2) and Li(3) show approximately tetrahedral coordin-
ation environments by bridging adjacent hydroxy groups, Li(1)
cannot quite reach a second oxygen and forms only two bonds:
to O(1) and a ligating THF. Additionally, two lithium atoms
exhibit significant interactions with aromatic ring bonds; Li(1)
interacts with the C(30)–C(31) bond at 2.36(1) and 2.48(2) Å,
and Li(3) interacts with the C(2)–C(15) linkage at 2.23(1) and
2.53(1) Å, respectively. Together with these interactions, Li(1)
and Li(3) each coordinate a single THF molecule, while the
lower steric requirements of Li(2) allow the binding of two
THFs, although the Li(2)–O(201) distance is longer than the
others at 2.14(1) Å. The Li(1)–O(1) distance at 1.76(1) Å is very
short, and is comparable to the shortest distance previously
1
As discussed above, H NMR NOE studies have suggested
the aromatic rings in 1a and 1b freely rotate about the central
carbon linker. NOE experiments similarly revealed that the
rings in 7a and 7b rotate, and the preferred conformation (Type
1 versus Type 2, Fig. 1) for 7 was unknown. Unlike 1a and
1b, the compounds 7a and 7b were, somewhat surprisingly,
extremely soluble in all organic solvents, and exhaustive
attempts at crystallization by evaporation were futile. Addi-
tionally, 7b proved to be unstable in non-alcoholic solvents and
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748
1745

View Article Online
the bond angle data reveals the geometry about the central
carbon linker is almost perfectly tetrahedral, while the bond
length values show the C–Ar distances are approximately equal
and average to 1.56(2) Å, typical for tetraaryl substituted
carbons.20 Relative to the plane drawn through C(2), C(16) and
C(30), the rings are canted at different angles; 32.5(2) and
37.7(2)Њ for the two rings with hydroxys [O(1) and O(3)] pointing
up with respect to the phenyl group, while the ring whose
hydroxy groups points down [O(2)] is considerably closer to
perpendicular, forming an angle of 78.7(2)Њ. Somewhat surpris-
ingly, no intra- or inter-molecular hydrogen-bonding inter-
actions are present between any of the hydroxy groups.
1
The H NMR spectrum of 8a is entirely consistent with the
crystallographic observations. The two hydroxy resonances
pointing up with respect to the phenyl ring, although inequiv-
alent, appear at similar chemical shifts of 5.17 and 5.09 ppm,
while the remaining hydroxy resonance occurs at 4.38 ppm.
Congruent with this asymmetry, six Ar-H peaks are observed
for the substituted rings, with three peaks for the Ar-CH3
groups. All five protons on the introduced phenyl group are
inequivalent, appearing as two multiplets.
In order to ascertain if the hydroxy substituted rings in 8a
rotate at elevated temperature, a sample of 8a was dissolved in
d8-toluene and spectra were recorded at various temperatures.
The two hydroxy resonances (appearing in toluene at 5.60 and
5.58 ppm) coalesced into a single signal at ~50–60 ЊC with a
chemical shift of 5.53 ppm, consistent with the compound
adopting mirror symmetry at this temperature. At higher tem-
peratures (80–108 ЊC), a spectrum consistent with a molecule
possessing C3 symmetry emerged. A single, very broad hydroxy
resonance was observed at 5.33 ppm, together with only two
Ar-H signals from the three substituted aromatic rings, and
single Ar-CH3 (2.05 ppm) and tert-butyl (1.40 ppm) reson-
ances. Based on these data, the aromatic rings appear to rotate
at higher temperatures, and the compound should therefore be
able to act as a system with Type 2 geometry.
Fig. 8 Structure of 8aؒ0.8Me2CO (30% probability ellipsoids). The
acetone solvates and tert-butyl methyl groups have been omitted. The
complete numbering scheme is depicted in the electronic supplementary
data. Selected bond lengths (Å) and angles (Њ): C(1)–C(2) 1.589(6);
C(1)–C(13) 1.558(7); C(1)–C(24) 1.556(6); C(1)–C(35) 1.545(6). C(2)–
C(1)–C(13) 106.5(4); C(2)–C(1)–C(24) 110.7(3); C(2)–C(1)–C(35)
111.4(4); C(13)–C(1)–C(24) 111.7(4); C(13)–C(1)–C(35) 111.0(4);
C(24)–C(1)–C(35) 105.7(4).
measured for [Li3(µ2-OAr)] (where ArOH = 2,6-diphenyl-3,5-di-
tert-butylphenol) at 1.775(7) Å.19 The remaining Li–OAr
distances are much longer (1.85–1.97 Å) and typical for
lithium–phenoxide bond lengths.19
While the C(1)–aryl and C(1)–C(44) bond lengths are typical
for triphenylalkyl systems,20 C(1) adopts a substantially dis-
torted tetrahedral geometry, possibly due to significant strain
exerted by the lithium binding motif. Maximum deviations of
98.8(5) and 115.4(5)Њ for C(44)–C(1)–C(2) and C(2)–C(1)–
C(30) angles, respectively, were measured.
Finally, 2a was reacted with lithium aluminium hydride to
readily regenerate 1a. This methodology may have utility for
incorporation of a deuterium label into the methine moiety
of 1.
The spectroscopic data for 7a and 7b are consistent with a
C3 symmetric compound, and remarkably similar to the
corresponding starting materials 1a and 1b. A single hydroxy
resonance can be found at 5.13 and 5.01 ppm for 7a and 7b,
respectively, compared with 4.79 and 4.83 ppm for 1a and 1b. A
more significant downfield shift is observed in the 13C NMR for
the central carbon, which in 7a and 7b resonated at 49.6 and
50.4 ppm, while occurring at 42.4 and 42.5 ppm for 1a and 1b,
respectively. Comparison of the infrared spectral data demon-
strated that 7a and 7b exhibit higher (33 and 32 cmϪ1, respec-
tively) frequencies for the OH stretches relative to 1a and 1b.
The carbocation 2a was also treated with phenyllithium in a
Conclusions
The title compounds have been prepared and fully character-
ized. These compounds are reactive and of utility for the deriv-
atization of the central carbon atom in the triarylmethanes 1.
The hydroxyphenyl rings in 2 render the carbocation electron
deficient, so more electronegative atoms (or groups), such as the
halogens, fail to bind. Hydroxide, methyl and phenyl deriv-
atives, however, can be readily prepared and these materials
may possess very interesting ligand properties of value in
coordination chemistry; reactions to this end are currently
being investigated. Additionally, the hydroxy groups in 2 and 4,
can give rise to an uncatalyzed cyclodehydration, giving the
9-hydroxyphenyl substituted xanthene compounds.
1
manner analogous for the synthesis of 7. H NMR analysis
confirmed the incorporation of the phenyl moiety, although all
three substituted aromatic rings were now rendered inequiv-
alent, with their associated hydroxy resonances appearing at
three different chemical shifts. The latter observation implies
rotation of the aromatic rings occurs at a rate slower than the
timescale of the NMR experiment. Since the absolute geometry
of the compound was unknown, full characterization by X-ray
crystallography was undertaken. Fig. 8 displays a side-on
view of the structure of 8a, while Table 1 gives collection and
refinement data.
Clearly the compound is extremely sterically crowded,
although perhaps slightly less so than the methyl derivatives 7.21
In order to accommodate the phenyl group, one of the aromatic
rings has rotated 180Њ relative to the C(2), C(16), C(30) plane,
but the remaining hydroxys still point up with respect to the
added phenyl group. The phenyl is effectively locked in posi-
tion by the two hydroxy substituents, and in an unsymmetrical
fashion, rendering the two substituted rings non-equivalent.
Unlike the structure of the lithium salt of 7b, examination of
Experimental
Unless otherwise stated, all manipulations were carried out
under an inert atmosphere of N2 on a vacuum line using stand-
ard Schlenk techniques. All solvents used were dried and dis-
tilled prior to use. NMR spectra were recorded on a Varian
VXR 300 MHz spectrometer at 299.95 and 75.43 MHz for the
proton and carbon channels, respectively, using CDCl3 solvent,
unless otherwise stated. IR spectra were recorded as KBr discs
on a Bruker Vector 22 instrument at a resolution of 2 cmϪ1
.
UV–Vis spectra were recorded on a Varian Cary 50 instrument
using CHCl3 as the solvent. Melting points were determined
using a capillary melting point apparatus and are uncorrected.
The tris(3,5-dialkyl-2-hydroxyphenyl)methanes 1a (3-tert-butyl-
1746
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748

View Article Online
5-methyl) and 1b (3-di-tert-butyl) were prepared following
acetone solvent was rapidly lost after removal of the mother
previously reported procedures.1
liquor, while the remaining acetone solvate is more tightly held
and not lost even under vacuum. Solutions of 4a slowly
decompose to 6a on standing.
Mp 133–134 ЊC (Found C, 76.9; H, 9.2. C34H46O4ؒMe2CO
requires C, 77.0; H, 9.1%); νmax/cmϪ1 3383m br, 3362m br (OH);
δH 7.15 (s, 3H, Ar-OH), 7.13 (d, J 2.4 Hz, 3H, Ar-H), 6.67 (s,
1H, COH), 6.10 (d, J 2.4 Hz, 3H, Ar-H), 2.12 (s, 9H, CH3),
1.41 (s, 27H, tBu); δC 152.2 (CArOH), 137.9, 128.42, 128.40,
128.3, 127.0 (Ar-C), 87.3 (COH), 34.6 [C(CH3)3], 29.9
[C(CH3)3], 21.0 (Ar-CH3).
Tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)carbenium bromide
(2a)
In a 100 ml Schlenk flask, 1a (500 mg, 1.0 mmol) was dissolved
in dry CCl4 (50 ml), and cooled to 0 ЊC with an ice bath. Excess
Br2 (~0.2 ml) was added, rapidly turning the solution dark
green. The reaction was stirred at 0 ЊC for 2.5 h, after which
time the solvent and unreacted Br2 were removed under
vacuum, to leave a dark metallic green powder of 2a. NMR
analysis of the material obtained by this route indicated it
was of sufficient purity for further reactions. Solutions of 2a
decompose to 3a upon standing for a few hours.
9-(3-tert-Butyl-5-methyl-2-hydroxyphenyl)-4,5-di-tert-butyl-2,7-
dimethylxanthen-9-ol (5a)
Decomposes without melting (130 ЊC); λmax/nm 639 (ε/dm3
Compound 2a (100 mg, 0.172 mmol) was dissolved in ether in a
separating funnel. Excess H2O and Et3N were added, and the
mixture was vigorously shaken until the solution was almost
colorless. The organic layer was collected and dried (MgSO4),
filtered, and the solvent removed under reduced pressure. The
residue was dissolved in diethyl ether, and a small quantity of
MeCN was added. The ether was allowed to evaporate deposit-
ing large colorless hexagons of 5a (67 mg, 78%).
molϪ1 cmϪ1 4 300), 435 (8 800), 381 (7 300), 277 (14 200); νmax
/
cmϪ1 3436m v. br, 3143w br (OH); δH 9.22 (s, 3H, OH), 7.46 (d,
J 2.4 Hz, 3H, Ar-H), 6.52 (d, J 2.4 Hz, 3H, Ar-H), 2.17 (s, 9H,
CH3), 1.41 (s, 27H, tBu); δC 195.4 (Cϩ), 161.3 (CArOH), 141.6,
139.4, 135.3, 130.6, 130.3 (Ar-C), 35.1 [C(CH3)3], 29.4
[C(CH3)3], 20.5 (Ar-CH3).
Tris(3,5-di-tert-butyl-2-hydroxyphenyl)carbenium bromide (2b)
Mp 168–169 ЊC (Found C, 81.7; H, 9.6. C34H44O3 requires C,
81.8; H, 9.25%); νmax/cmϪ1 3529m, 3295m br (OH); δH 10.29 (s,
1H, Ar-OH), 7.41 (d, J 2.4 Hz, 2H, xanthene Ar-H), 7.28 (d,
J 2.4 Hz, 3H, xanthene Ar-H), 6.98 (d, J 2.4 Hz, 1H, phenol
Ar-H), 6.14 (d, J 2.4 Hz, 1H, phenol Ar-H), 3.22 (s, 1H,
COH), 2.33 (s, 6H, xanthene CH3), 2.07 (s, 3H, phenol CH3),
1.69 (s, 18H, xanthene tBu), 1.57 (s, 9H, xanthene tBu); δC 151.5
(phenol CArO), 146.1 (xanthene CArOH), 137.3, 137.1, 132.1,
130.9, 129.0, 128.6, 127.4, 126.8, 126.5 (Ar-C), 76.7 (COH),
35.1 [xanthene C(CH3)3], 35.0 [phenol C(CH3)3], 31.0 [xanthene
C(CH3)3], 29.7 [phenol C(CH3)3], 20.94 (xanthene CH3), 20.88
(phenol CH3).
In a manner analogous to that described for 2a, 1b (500 mg,
0.795 mmol) was reacted with excess Br2. The reaction
proceeded more rapidly than that of 2a, and a shorter reaction
time of 1.5 h was used. Longer reaction times led to significant
decomposition of the product.
Decomposes without melting (120 ЊC); λmax/nm 633 (ε/dm3
molϪ1 cmϪ1 5 700), 437 (13 000), 385 (8 000), 277 (15 500); νmax
/
cmϪ1 3426m v. br, 3109w br (OH); δH 9.04 (s, 3H, OH), 7.74 (d,
J 2.4 Hz, 3H, Ar-H), 6.68 (d, J 2.4 Hz, 3H, Ar-H), 1.43 (s,
27H, tBu), 1.20 (s, 27H, tBu); δC 197.1 (Cϩ), 160.5 (CArOH),
143.4, 141.3, 136.4, 131.7, 130.6 (Ar-C), 35.4, 34.3 [C(CH3)3],
30.9, 29.6 [C(CH3)3].
9-(3-tert-Butyl-5-methyl-2-hydroxyphenyl)-4,5-di-tert-butyl-2,7-
dimethyl-9-methoxyxanthene (6a)
9-(3-tert-Butyl-5-methyl-2-hydroxyphenyl)-4,5-di-tert-butyl-2,7-
Compound 2a (578 mg, 1.0 mmol) was dissolved in ether and
NaOMe (300 mg, 5.55 mmol) was added. The solution slowly
turned orange, then gray, and finally pale green at completion
(2 h). The excess NaOMe was filtered off, and the solvent was
removed under vacuum to leave a green–gray residue. The
material was purified by crystallization from MeOH–pentane
to give colorless blocks of 6aؒC5H12 (460 mg, 90%).
Mp 180–182 ЊC (decomp.) (Found C, 82.1; H, 10.1. C35H47O3ؒ
C5H12 requires C, 81.7; H, 10.1%); νmax/cmϪ1 3421w v. br, 3243m
(OH); δH 10.54 (s, 1H, Ar-OH), 7.25 (d, J 2.4 Hz, 2H, xanthene
Ar-H), 7.22 (d, J 2.4 Hz, 3H, xanthene Ar-H), 6.92 (d, J 2.4
Hz, 1H, phenol Ar-H), 6.07 (d, J 2.4 Hz, 1H, phenol Ar-H),
2.93 (s, 3H, COCH3), 2.29 (s, 6H, xanthene CH3), 2.01 (s, 3H,
phenol CH3), 1.63 (s, 18H, xanthene tBu), 1.53 (s, 9H, xanthene
tBu); δC 151.4 (phenol CArO), 148.2 (xanthene CArOH),
137.1, 136.7, 131.9, 129.6, 128.5, 127.6, 126.7, 126.3, 122.6
(Ar-C), 83.3 (COCH3), 50.3 (COCH3), 35.0 [xanthene/phenol
C(CH3)3], 30.9 [xanthene C(CH3)3], 29.7 [phenol C(CH3)3], 21.0
(xanthene CH3), 20.9 (phenol CH3).
dimethylxanthen-9-ylium bromide (3a)
Compound 2a (252 mg, 0.433 mmol) was reacted with MeOH
(solvent) at room temperature to rapidly (10 seconds) produce a
dark orange solution. The solvent was removed under vacuum
to leave an orange residue. Crystallization from a vapor diffu-
sion of ether into a saturated CH2Cl2 solution to gave 3aؒ
CH2Cl2 (224 mg, 92%).
Decomposes without melting (210 ЊC) (Found C, 71.7; H,
1
––
7.7. C34H43O2Brؒ10CH2Cl2 requires C, 71.6; H, 7.6%); λmax/nm
489 (ε/dm3 molϪ1 cmϪ1 3 600), 378 (22 400), 277 (39 600); νmax
/
cmϪ1 3434m v. br (OH); δH 8.41 (s, 1H, OH), 8.14 (d, J 2.4 Hz,
2H, xanthene Ar-H), 7.79 (d, J 2.4 Hz, 3H, xanthene Ar-H),
7.39 (d, J 2.4 Hz, 1H, phenol Ar-H), 6.75 (d, J 2.4 Hz, 1H,
phenol Ar-H), 2.53 (s, 6H, xanthene CH3), 2.37 (s, 3H, phenol
CH3), 1.78 (s, 18H, xanthene tBu), 1.48 (s, 9H, xanthene tBu);
δC 172.4 (Cϩ), 156.0 (xanthene CArO), 151.0 (phenol CArOH),
142.7, 139.6, 139.4, 137.9, 130.9, 129.2, 128.1, 127.8, 126.4,
121.0 (Ar-C), 35.4 [xanthene C(CH3)3], 34.9 [phenol C(CH3)3],
30.7 [xanthene C(CH3)3], 29.3 [phenol C(CH3)3], 21.5 (xanthene
CH3), 20.6 (phenol CH3).
1,1,1-Tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)ethane (7a)
An ether solution (40 ml) of 2a (501 mg, 0.861 mmol) was
cooled to Ϫ78 ЊC (acetone–dry ice) and MeLi (3.20 ml of 1.6 M
in ether, 5.12 mmol) was added. The solution was maintained at
Ϫ78 ЊC for 1 h and was then slowly allowed to warm to room
temperature. During this time, the dark colored solution
became pale yellow. Water was slowly added and the ether layer
subsequently separated, and dried with MgSO4. The solvent
was removed under reduced pressure to leave a yellow residue,
which was subjected to column chromatography, eluting with
hexanes, to give 7a as a pale yellow powder (267 mg, 60%). The
material collected from the column was pure by NMR,
although the crystalline material (hexanes Ϫ20 ЊC) is colorless.
Tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)methanol (4a)
In a separating funnel, 2a (100 mg, 0.172 mmol) was dissolved
in CHCl3 (25 ml). Saturated aqueous NaHCO3 (30 ml) was
added to the dark green solution and the resulting mixture
was vigorously shaken until the CHCl3 layer was almost color-
less (~5 minutes). The organic layer was collected and dried
with MgSO4, filtered, and the solvent removed under reduced
pressure to leave a pale green residue. A small quantity (~1 ml)
of pentane was added and the white insoluble material was
collected, washed and dried, and recrystallised from acetone to
yield pure 4aؒ2Me2CO (85 mg, 78%). One of the molecules of
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748
1747

View Article Online
Mp 168–170 ЊC (decomp.); m/z (FAB): 516.3507 (Mϩ,
C35H48O3 requires 516.3603); νmax/cmϪ1 3540m (OH); δH 7.11 (d,
J 2.4 Hz, 3H, Ar-H), 6.81 (d, J 2.4 Hz, 3H, Ar-H), 5.13 (s, 3H,
OH), 2.29 (s, 3H, CCH3), 2.22 (s, 9H, Ar-CH3), 1.34 (s, 27H,
tBu); δC 151.9 (CArOH), 139.1, 129.8, 129.7, 128.0, 126.2 (Ar-
C), 49.6 (CCH3), 34.9 [C(CH3)3], 29.7 [C(CH3)3], 25.3 (CCH3),
21.3 (Ar-CH3).
The structure showed the molecules adopt the same general
conformation, hydrogen-bonded dimers, seen previously in the
structure of 1b.1 For the structures 2a, 3a and 8a the disordered
CHCl3, CH2Cl2 and acetone solvates, respectively, were
removed (“squeezed”) from the data by use of the Platon for
Windows software program.22 The residual electron density
removed corresponds to two half site-occupancy CHCl3 mole-
cules for 2a, one full occupancy CH2Cl2 molecule for 3a, and a
single 0.8 site occupancy acetone for 8a. Structural and refine-
ment data for the complexes are presented in Table 1. ORTEP
diagrams were drawn using the ORTEP-3 for Windows
package.23
1,1,1-Tris(3,5-di-tert-butyl-2-hydroxyphenyl)ethane (7b)
In a procedure analogous to that described for 7a, 2b (502 mg,
0.710 mmol) was reacted with MeLi (2.7 ml of 1.6 M in ether,
4.32 mmol). The product was very unstable, decomposing over
only a few hours in non-alcoholic solvents. This reactivity hin-
dered attempts to purify the compound by chromatography or
crystallization. However, the crude material obtained (381 mg,
83%) was pure by 1H NMR spectroscopy.
Acknowledgements
We thank the University of Florida, the National Science
Foundation (CAREER Award 9874966) and the donors of
the American Chemical Society Petroleum Research Fund for
funding this research. Support from the Research Corporation
is also gratefully acknowledged. We would also like to express
our gratitude to Dr Khalil Abboud at the University of Florida
for granting access to X-ray instrumentation.
Mp 104–108 ЊC (decomp.); m/z (FAB): 642.5045 (Mϩ,
C44H66O3 requires 642.5012); νmax/cmϪ1 3535m (OH); δH 7.33 (d,
J 2.4 Hz, 3H, Ar-H), 6.91 (d, J 2.4 Hz, 3H, Ar-H), 5.01 (s, 3H,
OH), 2.37 (s, 3H, CCH3), 1.36 (s, 27H, tBu), 1.18 (s, 27H, tBu);
δC 151.7 (CArOH), 142.9, 138.3, 129.8, 124.0, 122.8 (Ar-C),
50.4 (CCH3), 35.3, 34.5 [C(CH3)3], 31.4, 29.9 [C(CH3)3], 25.3
(CCH3).
Tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)phenylmethane (8a)
References
In a procedure analogous to that described for 7a, PhLi (3.6 ml
of 2 M in cyclohexane–ether, 7.16 mmol) was added to an
ether solution of 2a (694 mg, 1.19 mmol) at Ϫ78 ЊC. After
workup, a bright yellow, oily residue was collected. Addition of
MeCN (1 ml), with stirring, precipitated crude 8a as a cream
solid, which was collected by filtration and dried. The crude
material was subsequently recrystallized by slow evaporation of
a saturated acetone solution to produce colorless crystals of
8aؒ0.8Me2CO (332 mg, 45%). The stoichiometry of the acetone
1 M. B. Dinger and M. J. Scott, Eur. J. Org. Chem., in press.
2 W. Moene, M. Vos, M. Schakel, F. J. J. de Kanter, R. F. Schmitz and
G. W. Klumpp, Chem. Eur. J., 2000, 6, 225; H. W. Gibson, S.-H. Lee,
P. T. Engen, P. Lecavalier, J. Sze, Y. X. Shen and M. Bheda, J. Org.
Chem., 1993, 58, 3748.
3 M. B. Dinger and M. J. Scott, Inorg. Chem., 2000, 39, 1238.
4 T. W. Greene and P. G. M. Wuts, Protective Groups in Organic
Synthesis, 3rd edn., Wiley & Sons, 1999, p. 249.
5 G. A. Olah, J. G. Shih, B. P. Singh and B. G. B. Gupta, J. Org. Chem.,
1983, 48, 3356.
6 Th. Zincke, Ann., 1939, 363, 246.
7 A. Dunand and R. Gerdil, Acta Crystallogr., Sect. B, 1984, 40,
59.
8 The structure of 1a was also determined for the purposes of
comparison with the derivatives presented herein. The bond length
and angle data as well as the general conformation of the structure
were found to be analogous to those of 1b, discussed elsewhere.1
9 B. Kahr, J. E. Jackson, D. L. Ward and S.-H. Jang, Acta Crystallogr.,
Sect. B, 1992, 48, 329.
10 C. A. Buehler, D. E. Cooper and E. O. Scrudder, J. Org. Chem.,
1943, 8, 316.
11 J. C. Martin and R. G. Smith, J. Am. Chem. Soc., 1964, 11, 2252.
12 E. Weber, N. Dörpinghaus and I. Csöregh, J. Chem. Soc., Perkin
Trans. 2, 1990, 2167.
13 G. Casiraghi, G. Casnati, M. Cornia and V. D’Azeglio, Tetrahedron
Lett., 1973, 679; G. Casiraghi, G. Casnati, M. Cornia and
A. Pochini, Gazz. Chim. Ital., 1978, 108, 79.
14 C.-M. Liu, R.-G. Xiong, X.-Z. You and W. Chen, Acta Chem.
Scand., 1998, 52, 883; H. Wang, R.-G. Xiong, C.-M. Liu, H.-Y.
Chen, X.-Z. You and W. Chen, Inorg. Chim. Acta, 1997, 254, 183;
A. Rieker, A. Moosmayer, W. Winter, N. Neumayer and W. Hiller,
Z. Kristallogr., 1997, 212, 161; J.-P. Dubost, J.-M. Leger,
J.-C. Colleter, P. Levillain and D. Fompeydie, C. R. Seances Acad.
Sci., Ser. 2, 1981, 292, 965.
15 I. L. J. Patterson, C. Glidewell and G. Ferguson, Acta Crystallogr.,
Sect. C, 1998, 54, 1970.
16 G. Ferguson, C. Glidewell and I. L. J. Patterson, Acta Crystallogr.,
Sect. C, 1996, 52, 420.
1
solvate was determined both by H NMR integration and by
structural analysis.
Mp 195–197 ЊC (melts with decomposition) (Found C, 81.6;
H, 9.2. C40H50O3ؒ0.8Me2CO requires C, 81.4; H, 8.85%); νmax
/
cmϪ1 3529m, 3295m br (OH); δH 7.45–7.28 (m, 4H, Ar-H), 7.19
(d, J 2.4 Hz, 1H, Ar-H), 7.17 (d, J 2.4 Hz, 1H, Ar-H), 7.14 (d,
3
J 2.4 Hz, 1H, Ar-H), 7.09 (d, J(H,H) 7.20 Hz, 3H, Ar-H),
7.07 (d, J 2.4 Hz, 3H, Ar-H), 6.73 (d, J 2.4 Hz, 3H, Ar-H),
6.72 (d, J 2.4 Hz, 3H, Ar-H), 5.17 (s, 1H, OH ‘up’), 5.09 (s, 1H,
OH ‘up’), 4.38 (s, 1H, OH ‘down’), 2.22 (s, 3H, Ar-CH3), 2.20
(s, 9H, Ar-CH3), 2.19 (s, 9H, Ar-CH3), 1.30 (s, 9H, tBu), 1.27
(s, 18H, tBu); δC 152.61, 152.58, 151.7 (CArOH), 140.2, 139.6,
139.2, 138.3, 133.7, 129.7, 129.4, 129.25, 129.22, 129.0, 128.9,
128.8, 128.7, 128.4, 128.3, 128.2, 128.1, 127.7, 127.3 (Ar-C),
60.1 (Ar3CPh), 34.94, 34.86, 34.7 [C(CH3)3], 29.5, 29.3
[C(CH3)3], 21.4, 21.3, 21.2 (Ar-CH3).
Tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)methane (1a)
Tris(3-tert-butyl-5-methyl-2-hydroxyphenyl)carbenium brom-
ide 2a (25 mg, 0.043 mmol) was dissolved in ether and LiAlH4
(100 mg, excess) added. The solution immediately turned
almost colorless. Water was carefully added dropwise, followed
by 6 M HCl. After extraction with ether, drying with mag-
nesium sulfate, and evaporation of the solvent, MeOH (1 ml)
was added to the solid residue and swirled. The insoluble
material was filtered to give 1a (15 mg, 69%) as a white powder
which exhibited the same spectral data as an authentic sample.1
17 M. Gomberg and L. H. Cone, Chem. Ber., 1906, 39, 2957.
18 M. Lofthagen, R. VernonClark, K. K. Baldridge and J. S. Siegel,
J. Org. Chem., 1992, 57, 61.
19 J. S. Vilardo, P. E. Fanwick and I. P. Rothwell, Polyhedron, 1998, 17,
769.
20 Cambridge Crystallographic Database: D. A. Fletcher, R. F.
McMeeking and D. Parkin, J. Chem. Inf. Comput. Sci., 1996, 36,
746; F. H. Allen and O. Kennard, Chem. Des. Autom. News, 1993, 8,
31.
X-ray crystallography‡
The structure of 1a revealed the presence of three independent
molecules in the unit cell, each residing on a three-fold position.
21 S. P. Verevkin, J. Chem. Eng. Data, 1999, 44, 557.
22 A. L. Spek, Acta Crystallogr., Sect. A, 1990, 46, C-34.
23 L. J. Farrugia, J. Appl. Crystallogr., 1997, 565.
‡ CCDC reference number 207/420. See http://www.rsc.org/suppdata/
p1/b0/b001032o/ for crystallographic files in .cif format.
1748
J. Chem. Soc., Perkin Trans. 1, 2000, 1741–1748