Synthesis of highly hindered oxepins and an azepine from bis-trityl carbenium ions: Structural characterisation by NMR and X-ray crystallography

Article abstract of DOI:10.1039/b207108h

Reactions of the bis-carbenium ion 2,2′-bis[bis(p-methoxyphenyl)methyl]biphenyl ditetrafluoroborate with ammonia, n-propylamine and benzylamine have been studied with the aim of developing an acid-labile protecting group for primary amines that masks both

Full text of DOI:10.1039/b207108h

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of highly hindered oxepins and an azepine from bis-trityl
carbenium ions: structural characterisation by NMR and X-ray
crystallography
1
Kieran A. Carey,a William Clegg,b Mark R. J. Elsegood,†b Bernard T. Golding,*a
M. N. Stuart Hilla and Howard Maskill*a
a Department of Chemistry, Bedson Building, University of Newcastle upon Tyne,
Newcastle upon Tyne NE1 7RU
b Chemical Crystallography Unit, Bedson Building, Department of Chemistry,
University of Newcastle upon Tyne, Newcastle upon Tyne, UK NE1 7RU
Received (in Cambridge, UK) 19th July 2002, Accepted 19th September 2002
First published as an Advance Article on the web 7th November 2002
Reactions of the bis-carbenium ion 2,2Ј-bis[bis(p-methoxyphenyl)methyl]biphenyl ditetrafluoroborate with
ammonia, n-propylamine and benzylamine have been studied with the aim of developing an acid-labile protecting
group for primary amines that masks both hydrogen atoms of the NH2 group. Although the parent 5,5,7,7-
tetrakis(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]azepine was isolated and fully characterised, the corresponding
azepines could not be obtained in a pure state from the reactions with the primary amines. The bis-carbenium
ion was prepared from the treatment of 2,2Ј-bis[bis(p-methoxyphenyl)hydroxymethyl]biphenyl with fluoroboric
acid. Attempts to prepare the corresponding diphenyl analogue were unsuccessful using either fluoroboric acid
or boron trifluoride. The only product isolated from these reactions was 5,5,7,7-tetraphenyl-5,7-dihydrodibenzo-
[c,e]oxepin. The oxepin and its phenyl-p-methoxyphenyl and bis(p-methoxyphenyl) analogues were efficiently
obtained by dehydration of the corresponding diol, e.g. 2,2Ј-bis[bis(p-methoxyphenyl)hydroxymethyl]biphenyl,
in the presence of 3 Å molecular sieves or Amberlite IR-200 resin. The compounds 2,2Ј-bis[bis(p-methoxyphenyl)-
hydroxymethyl]biphenyl, (R,R/S,S)-2,2Ј-[bis(phenyl-p-methoxyphenyl)hydroxymethyl]biphenyl, 5,5,7,7-tetraphenyl-
5,7-dihydrodibenzo[c,e]oxepin, 5,5,7,7-tetrakis(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]oxepin and 5,5,7,7-
tetrakis(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]azepine were characterised by crystal structure analyses.
2,2Ј-Bis[bis(p-methoxyphenyl)hydroxymethyl]biphenyl and 5,5,7,7-tetrakis(p-methoxyphenyl)-5,7-dihydro-
dibenzo[c,e]oxepin exhibited dynamic NMR properties, and free energy barriers have been determined.
We have recently described the use of substituted triphenyl-
Introduction
methyl (trityl) groups for the protection of primary and
secondary amines.27 The substituted trityl protecting groups,
however, do not solve the problem posed above, i.e. one N–H
remains unblocked. We therefore considered the possibility that
a suitable bis-trityl dication, which would be regenerated under
acidic conditions,27–29 could be used to mask both hydrogen
atoms of a primary amine. We now present results of an
investigation into the use of bis-trityl dications as their
ditetrafluoroborate salts 1, Fig. 1, as possible new protecting
groups for primary amines. This led to a study of the highly
hindered oxepins 2 and azepine 3, Fig. 1, derived from
these bis-trityl dications. Although a satisfactory procedure
for protecting a primary amine by reaction with 1 was not
achieved, several of the compounds derived from 1 exhibited
interesting structural features which are described herein.
The need for the protection of amino groups is commonplace
in the synthesis of natural products, e.g. amino acids1,2 and
peptides,3,4 polyamines,5–7 and antibiotics.8 Varied protecting
strategies have been described for amino functions, including
the use of N-alkyl (e.g. benzyl), acyl (e.g. acetyl), alkoxycarb-
onyl (e.g. BOC), silyl (e.g. trimethylsilyl),9–11 and metallo
groups (e.g. cobalt() derivatives).12 Most protecting groups for
primary alkylamines (RNH2), however, mask only one of the
hydrogen atoms connected to nitrogen. The remaining N–H is
susceptible to removal by strongly basic reagents, which can
cause problems in multistage synthesis because the derived
nitrogen-based anion is intrinsically reactive as a base and a
nucleophile. Furthermore, the salt formed by deprotonation of
an amine may be insoluble in organic solvents. Various possible
solutions to this problem have been presented, including the
use of phthalimide, dichloro-13 and tetrachlorophthalimides,14
dithiasuccinoyl,15 dimethylmaleoyl,16 diphenylsilyldiethylene
groups,17 and other silicon-containing protecting groups such
as STABASE (tetramethyldisilylazacyclopentane),18 benzo-
stabase19,20 and TEDI (tetraethyldisilaisoindoline).21 Primary
amines have also been protected as N-substituted 2,5-dimethyl-
pyrroles,22 2,5-bis[tris(isopropylsilyl)oxy]pyrrole (BIPSOP)
derivatives,23 triazones,24 and N,N-dibenzylformamidines;25
diprotected primary amines have also been prepared from
3,5-dinitro-1-(p-nitrophenyl)-4-pyridone.26
Results and discussion
Synthesis of diols 4
Diols 4a and 4b, Fig. 2, were prepared by reaction of an aryl
lithium with diphenic acid dimethyl ester.30,31 The yield of 4a
was 73% using phenyllithium freshly prepared from bromo-
benzene and n-butyllithium. The identity of the product was
confirmed by comparison of the 1H and 13C NMR spectra with
those of an authentic sample (kindly supplied by Professor
Toda). The tetraanisyl diol 4b32 was prepared in a similar
manner in 72% yield from p-methoxyphenyllithium, which
was generated in situ from p-bromoanisole and n-butyl-
† Present address: Chemistry Department, Loughborough University,
Loughborough, UK LE11 3TU.
DOI: 10.1039/b207108h
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679
This journal is © The Royal Society of Chemistry 2002
2673

View Article Online
reported the preparation from 4b of the ditetrafluoroborate and
diantimony hexachloride salts of dication 1b and analysed their
structures by X-ray crystallography.
Attempts to obtain the methoxy-free analogue 1a from the
corresponding diol 4a were unsuccessful so, following a report
by Olah and co-workers35 on rearrangements induced by protic
acids in related systems, we investigated an alternative pro-
cedure using boron trifluoride.36 This was considered to be
superior to the previously used method in that it could be
performed under anhydrous conditions, under nitrogen, and
the product would precipitate out of solution without the need
for addition of more ether. The procedure was first investigated
using 4,4Ј,4Љ-trimethoxytrityl alcohol (TMTrOH) and trityl
alcohol (TrOH), which gave 99% and 60% yields of the
tetrafluoroborates of the (substituted) trityl cations, respect-
ively. The tetraphenyl diol 4a was subjected to the same
procedure. A green colour was observed on addition of
BF3–diethyl ether and a white compound precipitated out of
1
solution, which was identified as oxepin 2a by H NMR and
Fig. 1
TLC comparison with an authentic sample. The diethyl ether–
BF3 phase was hydrolysed, worked up, and was shown to
contain further oxepin 2a. The total yield of oxepin was about
70% with no isolable salt of the dication. Hart and co-workers
had previously studied reactions of diols related to and includ-
ing 4a and 4b.37 They reported that, when the two hydroxy
groups can achieve the correct proximity, formation of a
relatively stable oxonium ion after the first ionisation pre-
vents formation of the dication from compounds without
carbocation-stabilising substituents. Salts with diphenylmethyl
cationic groups at para positions on the biphenyl system have
been isolated.38
Synthesis of oxepins 2
Fig. 2
The oxepins 2a–c were prepared by dehydrating the correspond-
ing diols, and the ease of these reactions reflects the relative
stabilities of the intermediate cations. Tetramethoxy diol 4b was
simply stirred with 3 Å molecular sieves in dichloromethane at
room temperature for 24 days to give oxepin 2b quantitatively.
This transformation occurred overnight when Amberlite was
used in place of molecular sieves. Diol 4c was converted to
oxepin 2c in 6 days under the same conditions. Reaction of diol
4a by the same procedure gave isolated yields of oxepin 2a
(70%) and starting material (20%) after two weeks, or overnight
by refluxing in toluene with Amberlite, using a Dean–Stark
apparatus.
lithium by a modification of the method of Rapoport and co-
workers.33
The nature of diol 4c did not allow it to be prepared by the
method employed for 4a and 4b, and no synthesis of it had
been previously reported. The compound was prepared
by reaction of 2,2Ј-dilithiobiphenyl (freshly prepared from
2,2Ј-dibromobiphenyl and n-butyllithium) with an excess of
4-methoxybenzophenone. There is, of course, little or no
stereocontrol in the generation of the two stereogenic centres in
1
this reaction. The H NMR spectrum of the crude reaction
mixture showed two methoxy signals of almost equal intensity
indicating that 4c was a 1 : 1 mixture of RR/SS and meso forms.
Synthesis of azepines 3
Synthesis of dications 1
Dication ditetrafluoroborate 1b was reacted with ammonia
according to a literature procedure27 to give 5,5,7,7-tetrakis-
(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]azepine 3a (40%)
and the corresponding diamine, 5a in a 2 : 1 ratio, Scheme
2. The azepine decomposed during attempted separation of
these compounds on silica. Chromatography on Florisil
proved effective in terms of separating efficiency and recovery.
Dication ditetrafluoroborate 1b was also reacted with n-propyl-
amine and benzylamine. 1H NMR spectra of the crude reaction
mixtures indicated the presence of both the desired ring closed
protected amines 3 along with the diamines 5, Scheme 2, in
approximately equal amounts. It was not possible to separate
these compounds by chromatography. The instability of
the bis-tritylated alkylamines on silica is not too surprising in
the light of results published by Maskill and coworkers29 on the
simpler methoxy-substituted N-tritylalkylamines. Attempts to
isolate the azepines by crystallisation from the crude products
also proved unsuccessful.
Tetramethoxy diol 4b was converted into 1b in 98% yield using
the Dauben procedure34 Scheme 1. Salt 1b was also prepared
from oxepin 2b in 57% yield using the same procedure. Soon
after the completion of our work, Suzuki and co-workers32
Given the difficulties of separation of azepines 3 from the
diamines 5, we attempted driving the reaction wholly towards
3 and minimising the formation of diamine by using just
one equivalent of primary amine and a highly hindered base.
Scheme 1 Synthesis of dications.
2674
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679

View Article Online
Fig.
4
(RR/SS)-2,2Ј-bis(phenyl-p-methoxyphenylhydroxymethyl)-
biphenyl 4c.
the mean literature value of 1.43 Å for a C–OH bond.42 The
C(sp3)–CAr bond lengths are shown in rows 4–6. There are two
sets of three for the bis-trityl diols. It is worth noting that the
two C–anisyl bonds in diol 4c are appreciably shorter than the
C–Ph bonds in the same molecule, although this relationship
does not hold for the tetramethoxy diol 4b, where all such
bonds are the same within experimental error. Also, the rela-
tionship is reversed between TrOH and TMTrOH, i.e. the
C(sp3)–CAr bonds are shorter in the former, but the effect here is
very small. The sums of the bond angles around the tetrahedral
carbon (Σ CAr–C–CAr) and (Σ O–C–CAr) are shown in rows 8–
11. As can be seen, the values for Σ CAr–C–CAr and Σ O–C–CAr
are in good agreement amongst all four compounds, which
indicates that there is no unusual crowding in the bis-trityl
systems compared with the simpler trityl alcohols.
Scheme 2 Synthesis of azepines and diamines.
Golding and coworkers39 had reported that the visible
spectrum of 4-methoxy- and 4,4Ј-dimethoxytrityl tetrafluoro-
borates in acetonitrile remains virtually unchanged when
titrated with 2,6-di-tert-butyl-4-methylpyridine (DTBMP).
Excess DTBMP was added to a solution of the dication 1b in
acetonitrile and the red colour remained. A stoichiometric
amount of n-propylamine was added, but the characteristic red
colour of the dication did not disappear; it only became less
intense, even after addition of more amine and continued
stirring for 4 days. The reaction mixture was worked up as
described below, but NMR analysis of the crude mixture
showed no signals characteristic of the propyl group.
The biphenyl bond lengths in diols 4b and 4c are 1.505 and
1.501 Å, respectively. The biphenyl torsion angles are 84.0 and
83.7Њ, respectively, which are accompanied by intramolecular
hydrogen bonds between the hydroxy groups (there is no inter-
molecular hydrogen bonding). These are compared with similar
compounds, Fig. 5 in Table 2 and, as could be expected, the
X-Ray crystal structures
Diols
Crystals of a single stereoisomer of 4c were obtained by chrom-
atography of the crude reaction mixture followed by crystallis-
ation using the vapour diffusion technique. The crystal
structures of diols 4b and racemic 4c are shown in Fig. 3 and
Fig. 5 Literature diols related to 4.
biphenyl bond length is shorter for the p,pЈ-diols, where there is
less crowding around the biphenyl bond. The presence of the
guest molecules for the literature diols in Table 2 probably
affects their ability to form the intramolecular hydrogen bonds,
and accounts for the torsion angle of the o,oЈ-diols.
Oxepins
The crystal structures of oxepins 2a and 2b are shown in Fig. 6
and Fig. 7, respectively. The salient features are summarised in
Table 3, along with the data for ditrityl ether43 (Tr–O–Tr) for
comparison. The crystal structure of 2a shows one molecule of
toluene (which has been omitted from Fig. 6 for clarity) for
each molecule of oxepin. This is simply solvent of crystallisa-
tion rather than a host–guest relationship and there are no evi-
dent significant intermolecular interactions. The molecule of 2b
has C2 crystallographic symmetry. The C–O bond lengths are
essentially the same in both oxepins and in ditrityl ether, indi-
Fig. 3 2,2Ј-bis[bis(p-methoxyphenyl)hydroxymethyl]biphenyl 4b.
Fig. 4, respectively. Some of the important features are sum-
marised in Table 1 along with relevant data for 4a30 and the
simple trityl alcohols TrOH40 and TMTrOH.41 As can be seen,
results for all five compounds are in good agreement. The
C–OH bond lengths for the bis-trityl diols are comparable with
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679
2675

View Article Online
Table 1 Structural parameters for 4a, 4b, 4c, TrOH and TMTrOH; standard uncertainties are given in parentheses for individual results obtained in
this work, here and in other tables
Compound
4af
4b
4c
TrOH
TMTrOH
Bond length/Å
Bond length/Å
C(sp3)a–OH
C(sp3)a–OH
Mean
1.457
1.451
1.454
1.508 1.538
1.546 1.548
1.532 1.526
—
322.7
322.5
1.429 (2)
1.453 (2)
1.441
1.454 (8)
1.435 (8)
1.445
1.449
—
—
1.514
1.514
1.514
1.514
324.0
—
1.459
—
—
Bond lengthc/Å
Bond lengthc/Å
Bond lengthc/Å
C(sp3)a–Ar
C(sp3)a–Ar
C(sp3)a–Ar
Mean
1.542 (3) 1.545 (3)
1.525 (10) 1.542 (10)
1.525b
1.529b
1.521b
1.525
324.8
—
1.537 (3)b 1.536 (3)b
1.501 (12)b 1.558 (10)
1.537 (3)b 1.545 (3)b
1.482 (11)b 1.539 (11)
—
—
Σ anglesd/Њ
Σ anglesd/Њ
Σ anglese/Њ
Σ anglese/Њ
O–C(sp3)a–CAr
O–C(sp3)a–CAr
CAr–C(sp3)a–CAr
CAr–C(sp3)a–CAr
322.9
324.8
333.7
331.9
322.5
324.7
331.9
334.2
334.1
334.2
323.7
—
331.9
—
a Tetrahedral carbon. b Anisyl group. c C–C bond lengths for the sp2-sp3 bonds in the bis-trityl diols and simple trityl alcohols d The sum of the three
O–C(sp3)–CAr bond angles around the sp3 carbon atoms as a measure of the extent of steric crowding. e The sums of the three CAr–C(sp3)–CAr bond
angles around the sp3 carbon as a measure of the extent of steric crowding. f Complex with acetone.
Table 2 Torsion angles and central C–C bond lengths for biphenyls 4a and 6–10
Compound
4a
6a
7b
8c
9d
10d
10d e
10d a
Torsion angle/Њ
Biphenyl bond/Å
78.0
1.501
88.1
1.508
88.3
1.510
89.1
1.514
7.2
1.495
41.2
1.492
37.9
1.486
18.0
1.477
a Complex with acetone. b Complex with butyronitrile. c Complex with cyclohexanone. d p, pЈ-diol; no simple related m,mЈ-diols were found in the
Cambridge Structural Database. e Polymorph.
Table 3 Structural parameters for 2a, 2b and ditrityl ether
Compound
2a
2b
Tr–O–Tr
C–O bond/Å
C–O bond/Å
Mean
1.448 (2)
1.454 (2)
1.451
1.452 (1)b
1.454
1.466
1.460
1.544 1.548
1.553 1.541
1.541 1.533
C(sp3)a–CAr bond/Å 1.535 (3) 1.531 (3)
C(sp3)a–CAr bond/Å 1.541 (3) 1.551 (3)
C(sp3)a–CAr bond/Å 1.546 (3) 1.544 (3)
1.535 (2)b
1.540 (2)b
1.548 (2)b
a Tetrahedral carbon. Rows 1–2, C–O bond lengths for oxepins 2a, 2b
and trityl ether. Rows 4–6, C(sp3)–CAr bond lengths. b Data for both
halves of the C2 symmetric molecule supplied.
Fig. 6 5,5,7,7-Tetraphenyl-5,7-dihydrodibenzo[c,e]oxepin, 2a.
Fig.
8
5,5,7,7-Tetrakis(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]-
azepine, 3a
three compounds in Table 3 are in good agreement with the
corresponding bond lengths in the diols, Table 1.
Fig.
7
5,5,7,7-Tetrakis(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]-
Azepine
oxepin, 2b.
The crystal structure of azepine 3a is shown in Fig. 8 and some
features are summarised in Table 4. The amine hydrogen is dis-
ordered equally over two positions related by a crystallographic
two-fold axis, but only one position is shown. As can be seen in
cating that the crowding in the ortho positions of the biphenyl
system does not lead to unusually long C–O bonds. The C–O
bond lengths and the mean C(sp3)–CAr bond lengths for all
2676
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679

View Article Online
Table 4 Structural parameters for 3a, TrNH2, and TMTrNH2
energy barrier to inversion of oxepin 2b in d5-chlorobenzene by
locating the coalescence temperature of the methoxy signals,
using an established mathematical treatment,48 but there was no
significant change in the methoxy signals between 24 and 120
ЊC. A minimum value of 79 kJ molϪ1 for the free energy barrier
can be estimated, assuming coalescence at 120 ЊC.
Compound
Azepine 3a TrNH2 TMTrNH2
C(sp3)a–N bond length/Å
C(sp3)a–Ar bond length/Å
C(sp3)a–Ar bond length/Å
C(sp3)a–Ar bond length/Å
1.480 (2)b
1.537 (2)b
1.551 (2)b
1.558 (2)b
1.486
1.540
1.538
1.543
1.540
331.9
324.7
1.484
1.537
1.539
1.540
1.539
330.6
326.2
Although there was no significant change in the methoxy
signals of oxepin 2b upon heating, the aromatic region changed
dramatically. The broad signal evident at δ 8.5 at lower temper-
atures disappears, and (amongst other changes) the doublet
at δ 8.1 coalesces into a broad singlet. Attempts to assign the
Mean C(sp3)a–Ar bond length/Å 1.549
Σc CAr–C(sp3)a–CAr/Њ
Σd N–C(sp3)a–CAr/Њ
327.5b
329.3b
a Tetrahedral carbon. b Data for both halves of the molecule. c The sums
of the three CAr–C(sp3)–CAr bond angles around the sp3 carbon as a
measure of the extent of steric crowding. d The sum of the three N–
C(sp3)–CAr bond angles around the sp3 carbon atoms as a measure of
the extent of steric crowding.
1
signals in the H NMR spectrum at room temperature were
unsuccessful, even with the aid of COSY, ROESY, and
HETCOR experiments.
1H spectra of oxepin 2b in deuteriochloroform were also
obtained at 10Њ intervals down to Ϫ50 ЊC. The broad aromatic
signal at δ 8.4 at 23 ЊC separates into a doublet at Ϫ50 ЊC when
two extra distinct signals become visible between δ 6.6–7.0, and
the broad signal in this region at 23 ЊC is no longer present. The
resolved spectrum indicates that conformational interconver-
sions have been frozen out. The signals in the 1H spectrum of 2b
at Ϫ50 ЊC were assigned with the aid of COSY, HETCOR, and
ROESY techniques. Analysis of the aromatic regions for 2b on
warming up from Ϫ50 ЊC showed that 4 signals remain
unaffected and 4 change dramatically. These are assigned to the
two anisyl rings; one set of ring protons are labelled a,aЈ,c,cЈ
and the other b,bЈ,d,dЈ in Fig. 9. Signals of hydrogens a,c
Table 4, the C–N bond lengths in azepine 3a are both 1.480 Å
and are in good agreement with those in tritylamine (TrNH2)
and trimethoxytritylamine (TMTrNH2).44 There are no unusual
bond lengths; this is analogous to what was observed for the
oxepins and trityl alcohols. The C(sp3)–Ar bond lengths (of
which there are two sets) and the sums of the bond angles
around the tetrahedral carbons (two sets of angles for each
carbon, i.e. CAr–C–CAr and N–C–CAr) are also shown in Table
4. Results for all three compounds are in good agreement and
show that there is no unusual crowding in the azepine system
compared to the simpler tritylamines.
Dynamic NMR studies
1
The H NMR spectrum of the tetramethoxy diol 4b has two
striking features. First, there are two signals corresponding to
the methoxy groups at δ 3.7; this is ascribed to the two methoxy
groups being in different environments due to restricted rota-
tion around the biphenyl central bond. Secondly, there is an
upfield signal at δ 6.0, which is ascribed to hydrogens in the 3,3Ј-
positions of the biphenyl system lying in the shielding region
1
of the anisyl rings. Similarly, the H NMR spectrum of diol
4c is complicated, because there is enantiomerism due to the
restricted rotation (as for 4b) as well as two stereogenic sp3 car-
bons in the molecule. The 1H NMR spectrum of 4c shows two
signals for the hydroxy groups and two upfield doublets near
δ 6.0. In contrast, the spectrum for 4b shows only one doublet
in this region and only one hydroxy signal.
Fig. 9 Signal assignments for oxepin 2b.
remain unchanged on warming from Ϫ50 to 23 ЊC, because this
ring cannot freely rotate, due to unfavourable steric interactions
with the hydrogen on C-3 of the biphenyl system. The hin-
drance to rotation invoked in this interpretation is supported by
low level molecular modelling using the Hyper-Chem Lite pro-
gram. By locating the coalescence temperature of the c,cЈ inter-
conversion, the free energy barrier to rotation of the a,c ring
has been determined as 67 1 kJ molϪ1 at 56 ЊC. If we make the
usual assumption that the entropy of activation is zero, this
is the value of the enthalpy barrier of the conformational
change.
1
The H NMR spectrum of oxepin 2b also shows two meth-
oxy signals attributed to the anisyl rings occupying pseudo-axial
and pseudo-equatorial positions on the seven-membered oxepin
ring, as seen in the crystal structure (Fig. 7). The less different
methoxy signals in the diol 4b, compared with 2b, could be due
to the same phenomenon in a nine-membered ring only weakly
held by a hydrogen bond.
The aromatic region of the spectrum of 2b also shows a con-
siderably wider range of chemical shifts than for the diol 4b,
although the doublet seen at δ 6.0 for 4b is not present for 2b.
The difference in the aromatic regions of the proton spectra
between the alcohols and the oxepins is believed to be due to the
conformationally more restricted oxepins fixing protons in dif-
ferent positions, which are exchangeable by rotations in the less
restrained diols. The size of the ortho substituents causes a sub-
stantial barrier to rotation around the biphenyl bond, which
endows the alcohols with enantiomerism of the type first rec-
ognised by Christie and Kenner.45 The oxygen bridge in the
oxepins adds another dimension to their stereochemistry, and
Wittig and Leo46 were the first to identify molecular asymmetry
in such compounds. In principle, it is possible to interchange
pseudo-axial and pseudo-equatorial aryl groups by a ring flip
mechanism with a concomitant rotation of the biphenyl bond.
A number of other dibenzo[c,e]oxepins showing this phenom-
enon have been reported.47 We attempted to measure the free
Conclusions
This paper describes an attempt to develop a new protecting
group for primary amines, which enables both hydrogen atoms
of the NH2 group to be masked. The strategy adopted using
the bis-carbenium ion 2,2Ј-bis[bis(p-methoxyphenyl)methyl]-
biphenyl as its ditetrafluoroborate was an extension of that
successfully employed to protect primary amines using the
dimethoxy- and trimethoxy-trityl groups.27 However, although
the bis-carbenium ion 2,2Ј-bis[bis(p-methoxyphenyl)methyl]-
biphenyl reacted with ammonia to give the parent 5,5,7,7-
tetrakis(p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]azepine,
N-substituted azepines were not obtained from n-propylamine
and benzylamine. The failure of this approach probably derives
from the severe steric interactions engendered in such highly
hindered systems.
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679
2677

View Article Online
product as a yellow solid (0.17 g, 0.27 mmol, 40%), which was
Experimental
crystallised from ethyl acetate–petrol to give white crystals; mp
266–268 ЊC; δH (CDCl3) 6.3–7.2 (m, 24H, Ar), 3.67 (s, 6H,
OCH3), 3.78 (s, 6H, OCH3), 5.28 (s, 1H, NH).
n-Butyllithium was titrated with diphenylacetic acid before use.
Benzylamine and n-propylamine were purified according to lit-
erature procedures.49 Acetonitrile was pre-dried over potassium
carbonate and distilled from calcium hydride; tetrahydrofuran
was pre-dried by refluxing over sodium wire and was distilled
from lithium aluminium hydride. Amberlite refers to Amberlite
IR 200 ion exchange resin. Thin layer chromatography was per-
formed using TLC aluminium plates pre-coated with silica gel
(Kieselgel 60 F254, 0.2 mm). Silica gel (Kieselgel 60) was used for
2,2Ј-Bis(phenyl-p-methoxyphenylhydroxymethyl)biphenyl 4c
n-Butyllithium (1.9 mol dmϪ3, 2.6 cm3, 4.9 mmol) was slowly
added to an ice-cold solution of 2,2Ј-dibromobiphenyl (0.70g,
2.2 mmol) in THF (25 cm3), and the mixture was stirred for 1 h
at 0 ЊC. p-Methoxybenzophenone (1.0 g, 4.7 mmol) in THF
(25 cm3) was added dropwise and the mixture was stirred for 1 h
at room temperature. After refluxing the reaction overnight,
solvents were evaporated, saturated aqueous ammonium chlor-
ide was added to the cooled residue, and the mixture was
extracted with ether (3 × 20 cm3). The combined ether phase
was dried (MgSO4), filtered, and evaporated. The residual
orange oil was adsorbed onto alumina and chromatographed
(Al2O3, ethyl acetate–petrol–Et3N, 15 : 85 : 1) to give a
colourless oil, (0.55 g, 58%). This was crystallised from
dichloromethane–petrol to give the title compound as colour-
less crystals; mp 234 ЊC; δH (200 MHz, CDCl3) 3.76 (3H, s,
OCH3), 3.82 (3H, s, OCH3), 4.44 (1H, s, OH), 4.47 (1H, s, OH),
5.99 (2H, d, Ar), 6.10 (2H, d, Ar), 6.65–7.35 (24H, m, Ar).
1
column chromatography. The H NMR spectra were recorded
on a Bruker AC-200E (200 MHz) spectrometer for routine
work, a Bruker WM-300 WB (300 MHz) spectrometer for high
temperature work, and a Jeol JNM-LA500 FT-NMR (500
MHz) for low temperature work. Residual proton signals from
the deuteriated solvents were used as references. 13C Spectra
were recorded on a Bruker AC-200E (50.3 MHz), the 13C signal
from the deuteriated solvent being used as a reference. Infrared
spectra were recorded on a Nicolet 20-PC Fourier Trans-
form IR spectrophotometer. Mass spectra [electron impact
(EI) mode] were recorded on a Kratos MS80 RF spectro-
meter. Combustion analysis results are averages of two
determinations.
5,5,7,7-Tetrakis( p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]-
oxepin 2b
Preparations
Procedure 1. 2,2Ј-Bis[bis(p-methoxyphenyl)hydroxymethyl]-
biphenyl (0.20 g, 0.31 mmol) was dissolved in dichloromethane
(5 cm3) and stirred with 3 Å molecular sieves at room temper-
ature. After 24 days conversion to the oxepin was complete by
TLC. The mixture was filtered and solvents evaporated to give
the title product (0.13 g, 66%, mp (recryst. CH2Cl2–MeCN)
272–275 ЊC; C, 80.92; H, 5.56; C42H36O5 requires C, 81.25; H,
5.85%; δH (500 MHz; CDCl3, Ϫ50 ЊC) 3.55 (6H, s, OCH3), 3.77
(6H, s, OCH3), 6.26 (2H, d), 6.38 (2H, d), 6.43 (2H, d), 6.51
(2H, d), 6.55 (2H, d), 6.67 (2H, d), 6.81 (2H, m), 6.99 (2H, d),
7.02 (2H, d), 7.46 (2H, d), 8.41 (2H, d); δC (500 MHz; CDCl3,
Ϫ50 ЊC) 55.13, 55.31, 85.69, 111.43, 113.15, 126.58, 127.65,
128.27, 129.08, 129.55, 139.12, 140.63, 141.22, 143.16, 157.26,
158.07; m/z 620 (Mϩ, 61%), 513 (Mϩ Ϫ107, 67), 497 (Mϩ Ϫ123,
100), 378 (Mϩ Ϫ242, 95) 135 (Mϩ Ϫ485, 95).
Procedure 2. 2,2Ј-Bis[bis(p-methoxyphenyl)hydroxymethyl]-
biphenyl (0.20 g, 0.31 mmol) was dissolved in dichloromethane
(5 cm3) and stirred with Amberlite ion exchange resin at room
temperature overnight. The mixture was filtered and solvents
removed (rotary evaporator) to give a white solid (0.192 g,
99%); mp (recryst. CH2Cl2–MeCN) 272–274 ЊC. This was
confirmed to be the title compound by comparison with an
authentic sample (mp, TLC, and 1H NMR).
2,2Ј-Bis[bis(p-methoxyphenyl)hydroxymethyl]biphenyl 4b
p-Bromoanisole (1.40 cm3, 10.9 mmol) was added over 5 min to
a solution of n-butyllithium (2.2 mol dmϪ3 in THF, 5 cm3,
11 mmol) in THF (10 cm3) at Ϫ78 ЊC. The mixture was stirred
for a further 10 min at Ϫ78 ЊC then dimethyl biphenyl-2,2Ј-
dicarboxylate (0.50 g, 1.9 mmol) in THF (5 cm3) was added.
The solution was stirred for 2 h at Ϫ78 ЊC, allowed to come to
room temperature, then stirred overnight before solvents were
evaporated and aqueous ammonium chloride (20 cm3, 0.25 mol
dmϪ3) was added. The mixture was extracted with dichloro-
methane (3 × 10 cm3) and the combined extract was dried
(sodium sulfate), filtered, and evaporated to dryness under
reduced pressure. The product was obtained by recrystallisation
using dichloromethane–petrol (0.90 g, 1.4 mmol, 75%); mp
222–223 ЊC, lit.,32 226–228 ЊC; δH (CD3CN), 6.70–6.77 (m, 12H,
Ar), 6.95–7.19 (m, 10H, Ar), 3.74 (s, 6H, OCH3), 3.75 (s, 6H,
OCH3), 5.99–6.04 (m, 2H, Ar), 4.37 (s, 2H, OH).
2,2Ј-Bis[bis( p-methoxyphenyl)methyl]biphenyl
ditetrafluoroborate 1b (from 4b)
Tetrafluoroboric acid (48% w/w aq., 0.50 cm3, 3.9 mmol) was
added dropwise to a cooled solution of 5,5,7,7-tetrakis(p-
methoxyphenyl)-5,7-dihydrodibenzo[c,e]oxepin (0.090 g, 0.14
mmol) in acetic anhydride (25 cm3). The bright red mixture was
stirred for 15 min then ether (140 cm3) was added, whereupon
the product precipitated as deep red crystals, which were filtered
at the pump, rinsed with ether, and dried under high vacuum
(0.060g, 57%).
Crystal structure determinations‡
Crystals were examined on a Stoe-Siemens four-circle diffract-
ometer with Mo-Kα radiation (λ = 0.71073 Å) for 4b and with
Cu-Kα radiation (λ = 1.54184 Å) for 4c, 2a, 2b and 3a. Data
were collected at 160 K, with on-line profile fitting.50 Intensity
decay ranged from 0 to 5%. Azimuthal-scan absorption correc-
tions were applied. Structure solution was by automatic direct
methods, and refinement with full-matrix least-squares.51 Crys-
tal data and other information are given in Table 5. Hydrogen
atoms were placed in ideal positions and constrained with a
riding model.
In the structure of 4c, one of the aromatic substituents is
disordered over two orientations, with 54 : 46% occupancy;
restraints were applied to geometry and displacement param-
eters for these atoms. The crystal structure of 2a is a 1 : 1
5,5,7,7-Tetrakis( p-methoxyphenyl)-5,7-dihydrodibenzo[c,e]-
azepine 3a
Ammonia (ca. 5 cm3) was condensed into a flask containing
2,2Ј-bis[bis(p-methoxyphenyl)methyl]biphenyl ditetrafluoro-
borate (0.53 g, 0.68 mmol) at Ϫ78 ЊC under nitrogen; the
mixture was stirred for 2 h at Ϫ78 ЊC, then triethylamine (4 cm3)
was added. The excess of ammonia evaporated overnight and
the residue was partitioned between aqueous sodium hydroxide
(0.5 mol dmϪ3, 5 cm3) and dichloromethane (10 cm3). The
aqueous phase was extracted twice more with dichloromethane,
then the combined organic phase was dried (MgSO4), filtered,
and evaporated to dryness. The residue was chromatographed
(Florisil, ethyl acetate–petrol, 10 : 90 ϩ1% Et3N) to give the title
‡ CCDC reference numbers 190257–190261. See http://www.rsc.org/
suppdata/p1/b2/b207108h/ for crystallographic files in .cif or other
electronic format.
2678
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679

View Article Online
Table 5 Crystallographic data for compounds 4b, 4c, 2a, 2b, and 3a
Compound
4b
4c
2a
2b
3a
Molecular formula
Mr
C42H38O6
638.7
C40H34O4
578.7
C38H28OؒC7H8
592.7
C42H36O5
620.7
C42H37NO4
619.7
Crystal system
Space group
Triclinic
P1
Monoclinic
P21/n
Monoclinic
P21/n
Monoclinic
I2
Monoclinic
I2
¯
a/Å
b/Å
c/Å
α/Њ
10.162(4)
12.856(5)
13.729(5)
74.83(3)
71.95(2)
85.00(2)
1645.9(11)
2
13.017(3)
16.298(6)
15.554(4)
13.6161(18)
14.295(2)
16.520(2)
9.656(3)
9.089(3)
18.142(6)
9.8515(11)
8.9880(10)
17.9926(19)
β/Њ
110.699(18)
95.178(18)
91.24(2)
91.406(4)
χ/Њ
U/Å3
3086.7(15)
4
1.245
3202.2(8)
4
1.229
1591.8(9)
2
1.295
1592.7(3)
2
1.292
Z
Dcalc/g cmϪ3
1.289
Reflections measured
Unique reflections
Rint
Number of parameters
R (F, F 2 > 2σ)
Rw (F 2, all data)
Goodness of fit on F 2
Max, min electron density/ÅϪ3
7044
5799
0.0348
440
0.0455
0.1184
1.016
7745
3881
0.0861
469
0.0840
0.2787
1.023
8465
5563
0.0470
417
0.0533
0.1500
1.054
5382
2680
0.0250
216
0.0245
0.0646
1.027
3900
2785
0.0228
220
0.0388
0.1051
1.067
ϩ0.67, Ϫ0.26
ϩ0.16, Ϫ0.24
ϩ0.33, Ϫ0.28
ϩ0.13, Ϫ0.12
ϩ0.21, Ϫ0.17
22 S. P. Breukelman, G. D. Meakins and M. D. Tirel, J. Chem. Soc.,
Chem. Commun., 1982, 800.
23 S. F. Martin and C. Limberakis, Tetrahedron Lett., 1997, 38, 2617.
24 S. Knapp, J. J. Hale, M. Bastos and F. S. Gibson, Tetrahedron Lett.,
1990, 31, 2109.
25 S. Vincent, S. Mons, L. Lebeau and C. Mioskowski, Tetrahedron
Lett., 1997, 38, 7527.
26 E. Matzumura, M. Ariga, Y. Tohda and T. Kawashima, Tetrahedron
Lett., 1981, 22, 757.
toluene solvate. Compounds 2b and 3a are isostructural; for
both of them, a crystallographic two-fold rotation axis passes
through the ring O or N atom, so that the asymmetric unit is
half a molecule, leading to two-fold disorder of the amine
hydrogen atom on either side of the rotation axis in 3a. These
two crystal structures are non-centrosymmetric and chiral; the
absolute structure parameter52 was refined to values of 0.06(15)
for 2b and 0.2(2) for 3a (ideal value is zero) on the basis of the
small degree of anomalous scattering.
27 A. P. Henderson, J. Riseborough, C. Bleasdale, W. Clegg, M. R. J.
Elsegood and B. T. Golding, J. Chem. Soc., Perkin Trans. 1, 1997,
3407.
28 C. Bleasdale, B. T. Golding, W. H. Lee, H. Maskill, J. Riseborough
and E. Smits, J. Chem. Soc., Chem. Commun., 1994, 93.
29 M. C. López, J. C. Martinez, I. Demirtas, H. Maskill and E. Stix,
Organic Reactivity (Tartu), 1997, 31, 71.
30 F. Toda, A. Kai, R. Toyotaka, W.-H. Yip and T. C. W. Mak, Chem.
Lett., 1989, 1921.
Acknowledgements
We thank the EPSRC for a studentship (KAC) and for equip-
ment funding (WC).
31 H. W. Underwood and E. L. Kochmann, J. Am. Chem. Soc., 1924,
46, 2069.
32 T. Suzuki, J.-I. Nishida and T. Tsuji, Angew. Chem., Int. Ed. Engl.,
1997, 36, 1329.
33 F. Berrée, K. Chang, A. Cobas and H. Rapoport, J. Org. Chem.,
1996, 61, 715.
34 H. J. Dauben, L. R. Honnen and K. M. Harmon, J. Org. Chem.,
1960, 25, 1442.
References
1 R. J. Cherney and L. Wang, J. Org. Chem., 1996, 61, 2544.
2 K. M. Sliedregt, A. Schouten, J. Kroon and R. M. J. Liskamp,
Tetrahedron Lett., 1996, 37, 4237.
3 P. Sieber and B. Riniker, Tetrahedron Lett., 1987, 48, 6031.
4 S. Matysiak, T. Böldicke, W. Tegge and R. Frank, Tetrahedron Lett.,
1998, 39, 1733.
5 J. S. Bradshaw, K. E. Krakowiak and R. M. Izatt, Tetrahedron, 1992,
48, 4475.
6 C. Goulaouic-Dubois, A. Guggisberg and M. Hesse, Tetrahedron,
1995, 51, 12573.
7 C. Jentgens, R. Hofman, A. Guggisberg, S. Bienz and M. Hesse,
Helv. Chim. Acta, 1997, 80, 966.
8 K. C. Nicolau and E. J. Sorensen, Classics in Total Synthesis:
Targets, Strategies, Methods, Wiley-VCH, New York, 1996.
9 T. W. Green and P. G. M. Wuts, Protective Groups in Organic
Synthesis, J. Wiley & Sons Inc., New York, 1991.
10 P. J. Kocienski, Protecting Groups, Thieme, Stuttgart, 1994.
11 J. F. W. McOmie, Protective Groups in Organic Chemistry, Plenum
Press, London, New York, 1973.
12 C. Dalby, C. Bleasdale, W. Clegg, M. R. J. Elsegood, B. T. Golding
and R. J. Griffin, Angew. Chem., Int. Ed. Engl., 1993, 32, 1696.
13 M. Lergenmuller, Y. Ito and T. Ogawa, Tetrahedron, 1998, 54, 1381.
14 J. S. Debenham, R. Rodebaugh and B. Fraser-Reid, J. Org. Chem.,
1997, 62, 4591.
35 G. A. Olah, Q. Liao, J. Casanova, R. Bau, G. Rasul and G. K. Surya
Prakash, J. Chem. Soc., Perkin Trans. 2, 1998, 2239.
36 H. D. Bartfield and W. Flitsch, Chem. Ber., 1973, 106, 1423.
37 H. Hart, T. Sulzberg, R. H. Schwendeman and R. H. Young,
Tetrahedron Lett., 1967, 8, 1337.
38 H. Volz and M. J. Volz de Lecea, Tetrahedron Lett., 1966, 7, 4675.
39 C. Bleasdale, S. B. Ellwood and B. T. Golding, J. Chem. Soc., Perkin
Trans. 1, 1990, 803.
40 G. Ferguson, J. F. Gallagher, C. Glidewell, J. N. Low and S. N.
Scrimgeour, Acta Crystallogr., Sect. C, 1992, 48, 1272.
41 G. Ferguson, C. Glidewell and I. L. J. Patterson, Acta Crystallogr.,
Sect. C, 1996, 52, 420.
42 J. March, Advanced Organic Chemistry, 4th ed., Wiley-Interscience,
New York, 1992.
43 C. Glidewell and D. C. Liles, Acta Crystallogr., Sect. B, 1978, 34, 696.
44 I. Demirtas, Ph.D thesis, Newcastle University, Newcastle upon
Tyne, 1997.
45 G. H. Christie and J. H. Kenner, J. Chem. Soc., 1922, 121, 614.
46 G. Wittig and M. Leo, Chem. Ber., 1931, 64b, 2395.
47 A. Rosowsky, in Heterocycles, vol. 26, Seven-Membered Heterocycles
Containing Oxygen and Sulfur, Wiley-Interscience, New York, 1972.
48 J. Sandström, Dynamic NMR Spectroscopy, Academic Press,
London, 1982.
15 M. Planas, E. Bardaji, K. J. Jensen and G. Baranay, J. Org. Chem.,
1999, 64, 7281.
16 M. R. E. Aly, J. C. Castro-Palomino, E.-S. I. Ibrahim, E.-S. H.
El-Ashyr and R. R. Schmidt, Eur. J. Org. Chem., 1998, 2305.
17 B. M. Kim and J. H. Cho, Tetrahedron Lett., 1999, 40, 5333.
18 S. Djuric, J. Venit and P. Magnus, Tetrahedron Lett., 1981, 22, 1787.
19 R. P. Bonar-Law, A. P. Davis and B. J. Dorgan, Tetrahedron Lett.,
1990, 31, 6721.
20 R. P. Bonar-Law, A. P. Davis and B. J. Dorgan, Tetrahedron Lett.,
1990, 31, 6725.
21 A. P. Davis and P. J. Gallagher, Tetrahedron Lett., 1995, 36, 3269.
49 D. D. Perrin and W. L. F. Armarego, Purification of Laboratory
Chemicals, Pergamon Press, 3rd ed., 1988.
50 W. Clegg, Acta Crystallogr., Sect. A, 1981, 37, 22.
51 G. M. Sheldrick, SHELXTL version 5.1, Bruker AXS Inc.,
Madison, Wisconsin, USA, 1997.
52 H. D. Flack, Acta Crystallogr., Sect. A, 1983, 39, 876.
J. Chem. Soc., Perkin Trans. 1, 2002, 2673–2679
2679