Synthesis of novel conjugated oligomers for second-order nonlinear optics: Incorporation of a central spacer as a conjugation modulator

Article abstract of DOI:10.1039/a606547c

A new series of second-order nonlinear optical chromophores has been synthesized that consists of a conjugated segment end-capped with an electron acceptor and an electron donor, respectively and a central spacer intended to modulate the electro-optical e

Full text of DOI:10.1039/a606547c

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of novel conjugated oligomers for second-order nonlinear
optics: incorporation of a central spacer as a conjugation modulator
J ian Xin Zhang, Philippe Dubois* and Robert J érôme
Center for Education and Research on Macromolecules (CERM), University of Liège,
Sart-Tilman B6, B-4000 Liège, Belgium
A new series of second-order nonlinear optical chromophores has been synthesized that consists of a
conjugated segment end-capped with an electron acceptor and an electron donor, respectively, and a
central spacer intended to modulate the electro-optical effect. Conjugated chains have been tailored with
trans-vinylene-1,5-thienylene as the building unit and N,N-dimethylamino and nitro groups as the donor–
acceptor pair. Four spacers have been incorporated into the central part of the conjugated oligomers,
which range from saturated to totally unsaturated functions, i.e. from methylene to vinylene units. The
general strategy relies upon two consecutive Wittig or Wittig–H orner reactions between the spacer
precursor and an aromatic phosphonium or phosphonate bearing the strong electron donor and the
acceptor, respectively. Two synthetic pathways have been studied. The first procedure is based on the use
of a symmetric precursor for the spacer. H owever, a reaction byproduct is formed, which must be removed
and decreases the reaction efficiency. The second approach requires an asymmetric precursor for the
spacer, the synthesis of which is a multistep process. In order to evaluate the effect of the spacer, a
completely conjugated oligomer has been prepared by the one-step coupling of two conjugated segments
end-functionalized by the electron donor and the acceptor, respectively.
Accordingly, the NLO systems envisaged in this paper are
Introduction
based on the tailoring of (macro)molecules of a triblock struc-
ture in which two highly polarizable and/or conducting seg-
ments are separated from each other by a spacer. 1,4-Phenylene,
1,5-thienylene and vinylene units have been considered as build-
ing blocks for the conjugated segments, N,N-dimethylamino
and nitro groups have been selected as the acceptor–donor pair.
Several spacers, X in Scheme 1, will be inserted into the conjug-
ated chain, for which electronic conjugation will be modulated
from high delocalization, i.e. vinylene, to a complete break by a
totally saturated group, such as propane-2,2-diyl. The general
structure of the proposed (macro)molecules is shown in
Scheme 1.
Information processing is likely to change significantly in the
near future as the steadily increasing needs for speed and stor-
age capacity of digital data rise. Interest in optical devices is
rapidly growing because of their promise of extremely high
speed processing, transmission and storage of data. The appli-
cation of optical fibres in high speed communication is essen-
tially devoted to a small part of modern optics: linear optics.
Although design and use of optical computers are feasible on
the basis of nonlinear optics (NLO) principles,1 the poor per-
formance of present NLO materials limits this. Inorganic
materials are known for this low NLO response,2 and organic
NLO materials have a low degree of supramolecular organiza-
tion and/or absorb strongly in the UV–VIS region.3
As a result of intensive research efforts, some fundamental
relationships between properties and the molecular structure of
NLO organic materials have emerged. For instance, second-
order NLO chromophores need a push–pull structure, which
means that π–π conjugated segments must be end-
functionalized by strongly electron conjugated donors and
acceptors, respectively. This traditional concept emphasizes a
parallel between the strength of the electron withdrawing and
donating capability of the end-groups and the nonlinearity of
the conjugated molecule. More recently, Marder 4 has stated
that molecular geometry is also a very important factor affect-
ing NLO properties. Although NLO organic materials usually
have a much higher NLO response than inorganic materials,
their optical application is currently limited by a dramatic
absorption red-shifted from the UV to the visible region.5 Zyss6
and Moylan 7 have indicated that some interruption of chain
conjugation, for instance, by the incorporation of a semi-
conjugated spacer, might lead to good NLO properties and low
absorption in this region.
Scheme 1
Very few studies8,9 have dealt with conjugated chains contain-
ing a spacer other than a π–π electronic conjugated bridge.
They have shown that the incorporation of the σ–σ and σ–π
conjugations has no deleterious effect on NLO properties,
whereas it might result in a very high damage threshold work-
ing in the off-resonance mode.10
This type of molecule may be of great interest for improving
our knowledge of the dependence of NLO properties on mol-
ecular structure, without disregarding potential applications.
Indeed, modulation of electronic conjugation should allow
electro-optical properties to be modulated by the application
of an electric field of variable strength. Moreover, the spacer
J. Chem. Soc., Perkin Trans. 2, 1997
1209

View Article Online
within the conjugated chain is expected to improve solubility
in common solvents, to facilitate processing and to displace the
UV absorption far away from the visible light region.
This paper reports the synthesis of the aforementioned mol-
ecules 1a–d with n = 1. A forthcoming paper will deal with the
increase in length of the conjugated system (n = 2,3) and substi-
tution of 1,5-thienylene for the 1,4-phenylene end-groups, all
these modifications being carried out in order to optimize the
NLO response of the basic material (Scheme 1, n = 1).
been proposed that forms the final molecule together with the
central spacer. For instance, the molecule 1d is poorly soluble in
common organic solvents, so that its synthesis according to the
proposed precursor method is quite a problem. The coupling
method, which requires a unique Wittig reaction, is then a
straightforward method to synthesise 1d, as shown in Scheme 2.
Symmetrical spacer precursors
Bis-(5-formyl-2-thienyl) sulfide 2a was prepared according to
Fedorov’s procedure11 (Scheme 3). Di-2-thienyl sulfide 5a was
formylated by metallation with BuLi in tetrahydrofuran (THF)
and finally reacted with dimethylformamide (DMF).
Silane precursor 2b was prepared by the general method of
carbon–silicon bond formation,12 i.e. coupling of organometal-
lic compounds with chlorosilane derivatives (Scheme 3).
Results and discussion
General synthetic strategy
The general strategy relies upon the synthesis of a precursor for
each spacer (2 or 3) that contains two end-groups reactive in the
Wittig reaction, i.e. formyl 2 and/or derivatized acetal 3. Two
successive Wittig-like reactions then have to be carried out to
attach the donor and the acceptor groups to the precursors
(Scheme 2). Benzylic phosphonium salts or phosphonates bear-
MeO
MeO
MeO
MeO
i
CH
CH
Li
S
S
4
ii iii
OHC
X
CHO
S
S
Me
Si
2
+
+
CH2Y
N
YCH2
NO2
OHC
CHO
or
S
S
Me
67–80%
OHC
X
CH(OCH3)2
2b
S
S
3
–X– = –S–
–Si(CH3)2–
–C(CH3)2–
Precursor method
iv
63–69%
OHC
X
CHO
X
S
S
S
S
5a,d
(X = S, CH=CH)
iv: 2 BuLi–DMF
2a,d
X
Me
S
S
N
NO2
iii: H30+
Cl Si Cl
Me
i: BuLi
ii:
1 a – d
Scheme 3
–X– = –CH CH–
Coupling method
Dichlorodimethylsilane was reacted with compound 4 followed
by in situ hydrolysis of the acetal with dilute HCl. It is worth
pointing out that hydrolysis must be carried out carefully
because of the very high sensitivity of the aromatic C᎐Si bond
to high acidity.13
S
CH2Y
N
(E)-1,2-Bis(5-formyl-2-thienyl)ethylene 2d was prepared by
metallation of (E)-1,2-di-2-thienylethylene 5d followed by reac-
tion with dry DMF according to the procedure used for the
synthesis of 2a. This result shows that the double bond (X:
+
OHC
S
᎐CH᎐CH᎐) does not react with BuLi at low temperature, in
᎐
NO2
spite of activation of the conjugation effect. It also appears that
formylation of oligo(vinylenethienylene) should be an efficient
method for attaching a functional end-group and for increasing
the length of conjugated segments by a Wittig-type conden-
sation. This increase in chain length might be of interest for
applications in the electronic and photonic domains, such as
polymer conductors, organic light emitting diodes (LEDs) and
third-order NLO materials.14,15
where Y = P(O)(OEt)2 or P(C6H5)3Br
Scheme 2
ing the donor or acceptor will be considered depending on the
starting materials available and the stability of the reaction
intermediates.
Chromophores from symmetrical spacer precursors
The spacer precursors are just the spacer attached to two
thiophene rings. The symmetrical precursors 2 are substituted
by two identical formyl groups and are easy to prepare. How-
ever, their use in the first Wittig reaction might lead to the
formation of symmetrical byproducts. In order to prevent this
side-reaction from occurring, precursors 3 have been syn-
thesized, that bear one protected formyl group and one alde-
hyde selectively reactive in the first of the two Wittig reactions.
As an alternative approach, a direct coupling reaction has
Chromophores 1a,b were first prepared by two successive Wit-
tig reactions from the symmetrical precursors 2a,b (Scheme 4).
In order to limit the undesired formation of the symmetrical
compounds 7 in the first Wittig reaction, the ylide of 4-
nitrobenzyl(triphenyl)phosphonium bromide was reacted with
a twofold excess of spacer precursors 2a,b. In the case of silane
spacer 2b, the ylide was first prepared by reaction with BuLi,
since in situ generation of these ylides by lithium ethoxide is
responsible for the rupture of the silane bond during the Wittig
1210
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
reaction. The resulting byproduct with the Si᎐OH moiety was
characterized by a broad IR absorption band at 3400 cmϪ1. For
this reason, the ylide was prepared separately using BuLi, which
is less reactive towards silane than alcoholate. Although BuLi
can react with the nitro group as previously mentioned,16 no
side reaction was observed, possibly due to the much lower
pKa value of the phosphonium salt compared to that of
thiophene.17
reported for Wittig reactions in case of stabilized phenyl ylides
(trans content: 50 ± 30%).18
The second Wittig reaction was performed similarly, i.e.
preparation of the ylide, which was then added to the aldehyde
solution. Four geometrical isomers, trans–trans, trans–cis, cis–
trans and cis–cis, were obtained even after purification by
chromatography on silica gel. Characterization of the mixture
1
by H NMR is quite a problem due to the overlapping of the
Although 2 equiv. of 2a,b were used, the expected final prod-
uct was contaminated by ca. 30% of the symmetrical product 7.
It is also worth noting that the Wittig reaction applied to a ylide
stabilized by conjugation is poorly stereoselective.18 Contamin-
ant 7 was separated from 6 by chromatography on a silica gel
column; the ratio of the cis/trans isomers of 6 remained
unchanged. This ratio was measured by 1H NMR spectroscopy
and found to be close to 3:2 in agreement with values currently
signals of the aromatic and unsaturated protons [Fig. 1(a) for
X = ᎐SiMe2᎐].
The chromophore containing a central vinylene spacer 1d is
actually a chromophore without any spacer, since the vinylene
unit does not perturb the electronic conjugation all along the
chain. In this case, the product formed by the first Wittig reac-
tion is of a very limited solubility in common organic solvents,
which prevents further purification and characterization.
2a,b
CH P(Ph)3
THF
NO2
CH2P(Ph)3Br + BuLi
NO2
X
CHO
X
2
+
S
S
S
NO2
NO2
6a,b (25–42%)
7a,b (ca. 30%)
CH
N
(Ph)3P
X
S
S
NO2
N
(51–80%)
I2/xylene
X
S
S
NO2
N
1a,b (56–81%)
Scheme 4
i
ii S
iii K3Fe(CN)6
BuLi
4
(MeO)2CH
Li
S
S
S
S
S
X = –S–
18%
i BuLi
ii (PhSO2)2S
X = –S–
i POCl3MFA
ii CH(OMe)3
X
X = –CMe2,–S–
65–75%
X
CH(OMe)2
S
S
S
S
S
Acetone
H2SO4
8a–c
5a,c
X = –CMe 2
–
76–81%
i BuLi
ii
Me
Cl
X = –SiMe2–
Cl–SiMe2–Cl
S
Me
4
(MeO)2CH
Li
S
9b
BuLi
DMF
X
CH(OMe)2
OHC
X
CH(OMe)2
S
S
S
S
55–91%
3a–c
8a–c
Scheme 5
J. Chem. Soc., Perkin Trans. 2, 1997
1211

View Article Online
with 5-lithiothiophene-2-carbaldehyde dimethyl acetal 4 in
order to reduce the disulfide to sulfide 8a. The yield of the
recovered sulfide 8a is very low (18%), and may result from a
secondary reaction between the acetal and the RSϪ anion
formed as a byproduct.22
Since the disulfide reduction yield is low, a new strategy has
been proposed for the synthesis of the precursor 3a. Sulfide 5a
is first synthesized by reaction of 2-lithiothiophene with
bis(phenylsulfonyl) sulfide as a coupling agent.23 Sulfide 5a can
be converted to the mono-formyl compound by Vilsmeier reac-
tion,11 followed by acetalization of the aldehyde by reaction
with trimethyl orthoformate in the presence of toluene-p-
sulfonic acid with formation of 8a. This second strategy uses
easily controlled reactions and gives quite high yields.
Silane precursor 3b was synthesized in a two-step process.
First, a threefold molar excess of dichlorodimethylsilane was
reacted with 2-lithiothiophene. The monochlorosilane 9b was
then separated from the starting materials and the symmetrical
dithienyl silane by distillation in vacuo. The same coupling reac-
tion was repeated, in which the monochlorosilane 9b is reacted
with 5-lithiothiophene-2-carbaldehyde dimethyl acetal 4 in dry
THF. Although the order of these two reactions might be
reversed, the extreme instability of 5-chlorodimethylsilylthio-
phene-2-carbaldehyde dimethyl acetal 10b in the presence of air
makes this modification of no practical use. Indeed, the acetals
10b are easily prepared (Scheme 6) but they are rapidly
decomposed with formation of HCl.24
Fig. 1 1H NMR spectra of the Wittig reaction products 1b. (a) Before
isomerization, and (b) after isomerization with iodine. Assignment of
the resonance peaks can be found in the Experimental section.
O
O
O
O
+
Li
Cl Si Cl
Si Cl
Since the pure trans-isomer has the highest NLO response,19
it is important to transform the original cis/trans mixture into
completely the trans configuration. To this end, the isomeriza-
tion reaction by iodine deserves interest. This reaction is usually
carried out in nitrobenzene.20,21 Due to the low thermal stability
of the silane derivative, 1b was refluxed in the presence of a
trace of iodine in nitrobenzene for only 5 min. Nevertheless,
decomposition occurs with formation of a hydroxy compound
as proved by IR spectroscopy. Substitution of xylene for nitro-
benzene allows the olefin isomerization to take place after 1 h
of reflux with no evidence of decomposition. Chromophores
containing silane and sulfide spacers 1a, 1b have been isomer-
ized to pure trans-olefins.
S
S
4
10b
deacetalization
Scheme 6
In addition to the spacers X able to modulate the chain con-
jugation, special attention was also paid to a completely satur-
ated spacer that would completely prevent the electronic conju-
gation from being propagated. The methylene group (X =
᎐CH2᎐) was chosen for this purpose.
The general strategy proposed for the preparation of the
methylene precursor consists of the preliminary synthesis of 11
(Scheme 7) followed by metallation with BuLi and reaction of
It has been observed that a high content of trans-isomer 1
can be obtained by isomerization of the crude reaction product
independently of the addition of iodine, in the case where the
aldehyde 6 has been isomerized with iodine just after the first
Wittig reaction. This observation merely indicates that the cata-
lytic amount of iodine has a masked affinity for the olefinic
oligomers, since it is still active beyond the synthesis and purifi-
cation of 6. Although there is no report on the possible effect of
additives on the second-order NLO response, evidence for the
possible effect of iodine on the excited state energy might be
found in the UV spectrum of the silane-containing chromo-
phore 1b, in which an additional small absorption at 650 nm
characteristic of the oligomer doped with iodine is recorded
after isomerization by iodine.
i
+ HCHO
CH2
S
S
S
ii
iii
CH2
CH(OCH3)2
S
S
11
The only way of avoiding the iodine treatment is the direct
synthesis of the all-trans isomer chromophores, which is
reported in the next section, together with the improvement of
the efficiency of the first Wittig reaction.
iv
Li
CH2
CH(OCH3)2
S
S
Synthesis of asymmetrical spacer precursors
CH
Li
CH(OCH3)2
The use of asymmetrical precursors substituted by a protected
formyl group 3a–c instead of the previously used symmetrical
precursors would prevent the final product from being con-
taminated by the symmetric oligomer 7.
S
S
Scheme 7 Reagents: i: HCl ii: POCl3, MFA iii: HC(OMe)3/Hϩ iv: BuLi
The two first steps for the synthesis of the sulfide precursor
3a (Scheme 5) are the same as for the synthesis of 2a, i.e. the
symmetrical counterpart. Di-2-thienyl disulfide was reacted
the lithio derivative with DMF. However, metallation of the
thiophene ring does not occur as desired. Actually, the methyl-
ene protons of 11 are expected to be acidic enough to react
1212
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
with BuLi and form an anion stabilized by the two thienyl sub-
stituents. For comparison, diphenylmethane has a pKa of 35.25
Thus a proton abstraction from the methylene spacer by BuLi
competes the desired metallation of thiophene.
This acidity of the methylene spacer decreases the interest in
any chromophore that contains it. Indeed, an electronic
delocalization may occur with formation of a thiophene qui-
noidal structure (Scheme 8).26 A more suitable spacer for isolat-
hydrolysis conditions must be used. For instance, aldehyde 12b
was obtained in a 96% yield by using pyridinium toluene-p-
sulfonate as catalyst for the hydrolysis.
In contrast to the synthesis of the NLO chromophores from
the symmetrical precursors 2a,b, a Wittig–Horner reaction is
now proposed to attach the nitro acceptor group to the oligo-
mers 12a–c in mainly the trans-configuration. An excess of
phosphonate anion is prepared separately by reaction of parent
phosphonate with BuLi in dry THF. The aldehyde 12a–c is then
transferred into the anion solution. Most often, the chromo-
phore is easily recovered by precipitation of the crude reaction
product into water, while the phosphonate salt remains soluble
in water. In the particular case of chromophore 1b, the silane
spacer could be decomposed by the basic solution of the excess
of phosphonate anions. This explains why the chromophore
containing a silane-type spacer 1b was purified differently, i.e.
first by chromatography on silica gel in order to remove the
phosphonate and other highly polar impurities, and then
by recrystallization from light petroleum–hexane mixture.
All three synthesized NLO chromophores 1a–c have an excel-
lent stereochemical purity, since the cis-isomer cannot be
–H+
CH
CH2
non-conjugated
+H+
S
S
S
S
conjugated
Scheme 8
ing the conjugated system completely could be found in a sub-
stituted methylene group, such as the propane-2,2-diyl group
(X = ᎐CMe2᎐).
2,2-Di-2-thienylpropane 5c was prepared from thiophene
and acetone with H2SO4 as a catalyst.27 Then, classical formyl-
ation and aldehyde protection led to 8c with very good yields
(71 and 92%, respectively) (Scheme 5).
1
detected by H NMR analysis. In spite of a large number of
The last step in the synthesis of the spacer precursors 3a–c
consists of the formylation of the thiophene ring of 8a–c by
BuLi (Scheme 5). The intermediate product in the reaction of
the organometallic compounds with DMF is easily hydrolysed
by water with formation of a strongly basic solution. It is the
reason why a careful neutralization with a dilute aqueous HCl
solution is recommended for precursor 3b, so as to avoid the
risk of breaking the C᎐Si bonds. Precursors 3a–c are finally
isolated by chromatography on silica gel and characterized by
1H NMR spectroscopy (see Experimental).
protons in the aromatic region, all the protons were identified
from the observed chemical shifts and coupling constants. The
experimental chemical shifts and their assignments are given in
the Experimental section.
Synthesis of the chromophore 1d by the coupling method
The NLO chromophore with a central vinylene ‘spacer’ 1d was
synthesized as a standard for a completely conjugated chromo-
phore. The strategy based on the asymmetrical precursor was
first applied to the synthesis of 1d. The starting material was
(E)-1,2-di-2-thienylethylene, which was prepared by the Wittig
reaction of thiophene-2-carbaldehyde with thienyltriphenyl-
phosphonium bromide in the presence of lithium ethoxide as a
catalyst.28 Formylation by 1 mol of BuLi and DMF mainly
forms the mono-substituted aldehyde 13. Whatever the experi-
mental conditions, the acetalization of 13 has repeatedly failed.
Upon the addition of a small amount (0.05 mol%) of an acid,
i.e. toluene-p-sulfonic acid, no reaction occurs even at 50 ЊC for
2 h, although a rapid decomposition of the aldehyde is
observed when the acid amount is increased up to ca. 1 mol%.
Formation of a less stable quinoid-like structure by a proton
addition could favour the occurrence of side reactions (Scheme
10).
Chromophores from asymmetrical spacer precursors
The Wittig–Horner reaction has been proposed to build up the
desired conjugated chains with high trans double bond content.
Nevertheless, due to the exceedingly high stability of the inter-
mediate compound formed by the addition of diethyl 4-
dimethylaminobenzylphosphonate on the aldehyde in the
Wittig–Horner reaction,16 the alkene linking the donor group to
the thiophene-based chromophore must be formed by a clas-
sical Wittig reaction with the phosphonium derivative of the
donor (Scheme 9). Recrystallization has proved to be a very
efficient method for the separation of the trans- and cis-isomers
of compounds 12a–c. Compounds 12a and 12c with the sulfide
and the propane-2,2-diyl spacer, respectively, were easily
recrystallized from ethanol–water, so that the pure trans-
intermediates were easily isolated. For compounds with a silane
spacer 12b, recrystallization was carried out from light petrol-
eum and hexane at a low temperature (Ϫ20 ЊC) after chroma-
tography on silica gel in order to eliminate the triphenylphos-
phine oxide byproduct.
H+
O
+
OH
S
S
S
S
13
Hydrolysis of the acetal is currently performed in an aqueous
HCl solution at pH = 2–3. These conditions are suitable for
chromophores containing stable spacers, such as sulfide and
propane-2,2-diyl. The reaction is very rapid and complete with-
in 5 min. In the presence of the carbon–silicon bonds, mild
Decomposition
Scheme 10
This drawback has been avoided by using the coupling
method illustrated in Scheme 11. 5-[(4-Dimethylaminophenyl)-
O
–
i
N
CH P(Ph)3
(EtO)2P CH
NO2
X
S
CHO
3a–d
S
S
N
+
50–67%
ii
H3O
47–69%
12a–c
X
S
N
NO2
1a–c
Scheme 9
J. Chem. Soc., Perkin Trans. 2, 1997
1213

View Article Online
Table 1 Properties of NLO chromophores with spacers 1a–d
Br(Ph)3PCH2
S
N
Spacer
᎐C(Me)2᎐
᎐Si(Me)2᎐
᎐S᎐
7.1
᎐CH᎐CH᎐
᎐
14
µ/Debye
βH R S
6.1
150
5.0
180
6.4
2980
+
a
218
(10Ϫ30 esu)
CHO
S
λmax/nm
377
376
388
487
NO2
(eV)
(3.29)
(3.30)
(3.19)
(2.54)
15
a
1064 nm, in CHCl3 solution.
61%
LiOEt
mutually reactive segments, one bearing the electron donor
group and the second one the electron acceptor group. The
NLO properties of the novel chromophores were also measured
and the preliminary results have demonstrated the possibility of
modulating the NLO properties by the insertion of central
spacers. The increase in conjugated length by the incorporation
of additional thienylene-(E)-vinylene units on both sides of the
spacer will be the topic of the forthcoming papers.
S
N
S
NO2
1d
Scheme 11
(E)-vinyl]thienylmethyltriphenylphosphonium bromide 14 was
synthesized as an electron donor containing a conjugated seg-
ment, according to the improved reaction of a phosphonium
bromide with an alcohol as previously reported.29 5-[(4-
Nitrophenyl)-(E)-vinyl]thiophene-2-carbaldehyde 1516 was
considered as the reactive counterpart bearing the electron
acceptor segment.
The coupling reaction, which is nothing but a Wittig reac-
tion, has led to the deep-red compound 1d. Due to the long
conjugated chain and the trans double bond in each of the
original conjugated segments, the solubility of 1d in common
organic solvents is limited even for a chromophore containing
the cis-isomer of the central vinylene spacer. Accordingly,
recrystallization from ethanol or diethyl ether cannot be used as
an efficient method for the separation of the cis- and trans-
Experimental
Materials
Thiophene (99ϩ%), thiophene-2-carbaldehyde (98%), BuLi
(2.5  in hexane) and dichlorodimethylsilane (99%) were pur-
chased from Janssen Chimica, Belgium. THF was refluxed over
sodium–benzophenone and distilled just before use. Di-2-
thienyl sulfide,11,23 di-2-thienyl disulfide,11 thiophene-2-carb-
aldehyde dimethyl acetal,16 4-dimethylaminobenzyl(triphenyl)-
phosphonium bromide,29 4-nitrobenzyl(triphenyl)phosphonium
bromide,32 diethyl 4-nitrobenzylphosphonate,33 (E)-1,2-di-2-
thienylethylene28 and bis(phenylsulfonyl) sulfide23 were pre-
pared according to the cited references.
1
isomers. H NMR shows two unsaturated signals with a 12 Hz
coupling constant that corresponds to a cis:trans ratio of 2:3.
Characterization and purification
The new compounds were characterized by 1H NMR and
FTIR. NMR spectra were recorded in CDCl3 using tetrameth-
ylsilane (TMS) as a reference with Bruker AM400 apparatus; J
values in Hz. IR spectra were recorded with a Perkin-Elmer FT-
1600 spectrometer, and samples were prepared by solvent cast-
ing on a KBr crystal. Purification was achieved by chroma-
tography on silica gel: 0.5 g samples were eluted through a col-
umn filled with 25 g of silica gel (Merck 63–200 µm ref. 7734).
The final purity was ascertained by thin-layer chromatography
(TLC) and high performance liquid chromatography (HPLC)
(Waters 600 with detector 486).
Preliminary analysis of NLO properties
The NLO response of the four synthesized chromophores was
measured in CHCl3 solution by hyper-Rayleigh scattering
(HRS)30 (Table 1). The range of β modulation by the spacer
effect varies from 150 × 10Ϫ30 to 218 × 10Ϫ30 esu, which suggests
that the first hyperpolarizability can be improved by choosing a
proper spacer with a moderate red-shift in UV absorption. The
large difference in β values between, e.g. the sulfide-spacer-based
chromophore and the completely conjugated chromophore
(more than ten times higher than the sulfide spacer) indicates
that the NLO properties could be modulated by spacers. The
four chromophores under investigation have higher β values
than that of p-nitroaniline (PNA, βH R S = 23 × 10Ϫ30 esu30). The
βH R S value of the NLO chromophore ‘without spacer’ (X:
General procedure for preparation of acetals
The aldehyde (0.5 g) was added to trimethyl orthoformate (10
ml) at room temp. A catalytic amount of toluene-p-sulfonic acid
monohydrate (5 to 10 mg) was then added slowly. The reaction
medium was neutralized with Na2CO3 after 1 h reaction at the
same temperature. The acetal was recovered by distillation
under vacuum within high yields (>90%).
᎐CH᎐CH᎐) is very high. This effect may be explained by the
᎐
complete conjugation that increases the hyperpolarizability,
and by the resonance mode31 since the doubled frequency of
the laser used for measurements (1064–532 nm) is rather close
to the chromophore absorption (λmax 487 nm). The NLO prop-
erties will be discussed in detail in another paper, in parallel
with theoretical calculations according to the INDO/CI-SOS
methods.
General procedure of formylation via metallation
This formylation procedure has been reported.16 The thiophene
derivatives were dried by azeotropic distillation (× 3) with pre-
viously dry toluene under N2. Dry THF was added as solvent
and cooled down to Ϫ60 ЊC. BuLi (0.8 ) was transferred to the
reaction medium with a syringe. Then dry DMF or N-
formylpiperidine (200% based on thiophene derivatives) was
added to the solution at 0 ЊC. After 1 h, the aldehyde was
recovered by precipitation into water. Due to their high molecu-
lar mass, the synthesized molecules could not be purified by
distillation in vacuo. Chromatography on silica gel was success-
fully used for a purification method of the precursors usually by
using an ethyl acetate–hexane (1:4 v/v) mixture as eluent.
Conclusions
Various synthetic strategies all based on a Wittig-type conden-
sation reaction have allowed four NLO chromophores to be
synthesized with a high stereoselectivity. In this respect,
superiority of the asymmetric precursor method must be out-
lined. Thanks to their solubility in common organic solvents,
separation of the cis–trans-isomers 1a–c, as prepared by a Wit-
tig reaction, was successfully carried out by a selective recrystal-
lization except for the completely conjugated chromophore 1d,
the solubility of which is intrinsically limited. This chromo-
phore 1d was synthesized in a one-step coupling reaction of two
General synthesis of the symmetrical precursors by Wittig
reactions (Scheme 4)
The phosphonium salt (0.015 mol) and the precursor aldehyde
1214
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
(0.01 mol) were separately dried by azeotropic distillation with
dry toluene. The precursor was then dissolved in dry THF and
the phosphonium salt dispersed in the same solvent. BuLi
(0.015 mol) was first added to the phosphonium salt at Ϫ30 to
Ϫ40 ЊC. The temperature was allowed to rise to room temp. for
30 min, before the aldehyde solution in THF was added. After 2
h, the solvent was removed and the crude product was purified
by chromatography on silica gel with ethyl acetate–hexane (1:4
v/v) as eluent.
above (yield: 55%). δH 9.74 (1 H, s), 7.57 (1 H, d), 7.27 (1 H, d),
7.04 (2 H, d), 5.60 (1 H, s), 3.37 (6 H, s), ν/cmϪ1 3086, 2934,
2829, 1666, 1414, 1223, 1179, 1094, 1055, 973, 801, 667.
5-Formyl-2-thienyl-(5Ј-dimethoxymethyl-2Ј-thienyl)-
dimethylsilane 3b
Thiophene (2.6 g, 0.031 mol), distilled from CaH2, was metal-
lated with BuLi (2.5 , 12.5 ml) in dry THF (65 ml) under N2
at 0 ЊC for 30 min. The solution was added dropwise to
dichlorodimethylsilane (12 g, 0.09 mol) under vigorous stirring
at room temp. The reaction was very rapid and followed by
distillation of solvent and unreacted reagents under reduced
pressure. The monochlorosilane 9b was finally purified by
distillation (37 ЊC, 0.1 mmHg): yield, 58%.
The second Wittig reaction in the preparation of the NLO
chromophores from the intermediate 6 was performed by the
same procedure.
Isomerization by iodine
A small pellet of iodine (<1 mg) was added at 130 ЊC to a xylene
solution (20 ml) containing 0.6 g of the sample to be isomer-
ized. After 1 h reflux, the mixture was allowed to cool to room
temp. and the isomerized sample was recrystallized from hex-
ane (100 ml). The product was filtered and dried. Yields were in
the range of 56–81%.
5-Lithiothiophene-2-carbaldehyde dimethyl acetal (0.025
mol) in THF (60 ml) was added to the monochlorosilane 9b (4.5
g, 0.025 mol) at Ϫ78 ЊC. The temperature was first increased to
room temp., and the solution was then poured into water and
extracted with diethyl ether. The ethereal phase was washed
with water and dried over anhydrous Na2SO4. After solvent
removal, the acetal 8b was distilled in vacuo (105 ЊC, 0.1
mmHg): yield, 70%. δH 7.60 (1 H, d), 7.20 (2 H, d), 7.16 (1 H, d),
7.09 (1 H, d), 5.65 (1 H, s), 3.37 (6 H, s), 0.62 (6 H, s), ν/cmϪ1
3062, 2930, 2885, 1530, 1400, 1251, 1210, 1187, 1090, 1061, 981,
832, 786, 671.
Compound 3b was prepared by the standard formylation
procedure via metallation of 8b (final yield: 91%). δH 9.91 (1 H,
s), 7.79 (1 H, d), 7.36 (1 H, d), 7.20 (1 H, d), 7.13 (1 H, d), 5.65
(1 H, s), 3.36 (6 H, s), 0.663 (6 H, s), ν/cmϪ1 3060, 2957, 2830,
2732, 1674, 1512, 1425, 1345, 1254, 1214, 1182, 1094, 1060, 996,
834, 808, 782, 668.
Bis(5-formyl-2-thienyl) sulfide 2a
Di-2-thienyl sulfide 5a (4.0 g, 0.0202 mol) was formylated by
metallation with BuLi (2.5 , 18 ml) in THF (120 ml) and then
reaction with 4.0 ml of dry DMF. Ethyl acetate–hexane (1:1)
was used as eluent for chromatography on silica gel. The
diformyl compound was isolated in 63% yield. δH 9.83 (2 H, s),
7.66 (2 H, d), 7.28 (2 H, d), ν/cmϪ1 3080, 2821, 2733, 1668, 1650,
1412, 1220, 1196, 1066, 989, 801, 749, 665.
Bis(5-formyl-2-thienyl)dimethylsilane 2b
A solution of 5-lithiothiophene-2-carbaldehyde dimethyl acetal
was prepared by metallation of thiophene-2-carbaldehyde
dimethyl acetal (15 g, 0.095 mol) with BuLi (2.5 , 38 ml) in dry
THF at Ϫ78 ЊC. To this was added 5.7 g (0.0475 mol) of freshly
distilled dichlorodimethylsilane at Ϫ78 ЊC. The temperature
was allowed to increase to room temp. for 30 min. Then, the
solution of the acetal was hydrolysed with 0.1  HCl at pH 4–5
for 5 min. The product was recovered by precipitation in water
and purified by chromatography on silica gel with 1:1 ethyl
acetate–hexane as eluent. (Yield, 67%). δH 9.96 (2 H, s), 7.82 (2
H, d), 7.39 (2 H, d); 0.724 (6 H, s), ν/cmϪ1 3089, 2959, 2845,
1664, 1511, 1426, 1254, 1056, 992, 813, 790, 760, 669.
2-(2-Thienyl)-2-(5Ј-dimethoxymethyl-2Ј-thienyl)propane and 2-
(5-formyl-2-thienyl)-2-(5Ј-dimethoxymethyl-2Ј-thienyl) propane
8c and 3c
The same procedure as used for the synthesis of 3a (method 2)
was implemented with 2,2-di-2-thienylpropane instead of di-2-
thienyl sulfide.
8c: yield, 92%, δH 7.31 (1 H, d), 6.89 (1 H, d), 6.86 (2 H, m),
6.73 (1 H, d), 5.54 (1 H, s), 3.35 (6 H, s), 1.83 (6 H, s), ν/cmϪ1
3072, 2969, 2829, 1682, 1478, 1348, 1186, 1095, 1053, 975, 904,
850, 802, 698.
3c: yield, 70%, δH 9.82 (1 H, s), 7.59 (1 H, d), 6.97 (1 H, d),
6.89 (1 H, d), 6.80 (1 H, d), 5.56 (1 H, s), 3.36 (6 H, s), 1.85 (6 H,
s), ν/cmϪ1 3103, 2972, 2801, 1667, 1527, 1446, 1366, 1216, 1038,
850, 811, 760, 700, 666.
(E)-1,2-Bis(5-formyl-2-thienyl)ethene 2d
This compound was prepared from (E)-1,2-di-2-thienylethene
by metallation, followed by reaction with DMF as reported
for 2a (yield, 69%). δH 9.89 (2 H, s), 7.69 (2 H, d), 7.24 (2 H, s),
7.22 (2 H, d), ν/cmϪ1 3082, 1652, 1456, 1229, 1046, 949, 808,
668.
Wittig reaction for the synthesis of intermediates 12a–c
The Wittig reaction previously used for the synthesis of the
symmetrical precursors 2a–c was considered. A typical prepar-
ation of the intermediates 12a–c was as follows for the particular
case of the silane spacer 12b. The precursor 3b (0.5 g, 0.0016
mol) and 4-dimethylaminobenzyl(triphenyl)phosphonium
bromide (0.002 34 mol, 1.0 g) were separately dried in vacuo at
ca. 40 ЊC for 30 min., and then added to 80 ml of dry THF.
BuLi (2.5 , 1.0 ml) was added to the phosphonium suspension
at Ϫ30 to Ϫ40 ЊC, and the ylide solution formed was slowly
warmed to room temp. for 30 min. The solution of 3b was
transferred to the ylide solution with vigorous stirring. After 2
h, the solvent was removed and the crude product was purified
by chromatography on silica gel with ethyl acetate–hexane (1:4
v/v) as eluent. Final hydrolysis in acetone (40 ml) containing
pyridinium toluene-p-sulfonate (50 mg in 10 ml water) for 2 h at
room temp. released the almost completely essentially trans-
aldehyde 12b. Compound 12b was finally purified by recrystal-
lization from hexane–light petroleum (EtOH–H2O for 12a,c) at
Ϫ20 ЊC (yield: 47%).
5-Formyl-2-thienyl 5Ј-dimethoxymethyl-2Ј-thienyl sulfide 3a
Method 1. Di-2-thienyl disulfide (1.74 g, 7.6 mmol), previously
dried by azeotropic distillation with toluene, was added to a
solution of 5-lithiothiophene-2-carbaldehyde dimethyl acetal 4
(0.0106 mol) in THF at Ϫ78 ЊC. The mixture was warmed to
room temp., stirred for 3 h, and then poured into water. Com-
pound 8a was recovered by extraction with diethyl ether,
washed with water, dried over anhydrous Na2SO4 and finally
distilled in vacuo (125 ЊC, 0.1 mmHg): yield, 18%.
Method 2. Di-2-thienyl sulfide 5a (3.1 g, 0.016 mol) was mixed
with POCl3 (3.12 g, 0.02 mol) and N-methylformanilide (2.75 g,
0.02 mol) at 0 ЊC. After a slight release of HCl, the solution was
heated to room temp. and stirred overnight. The viscous mix-
ture was quenched by adding ice and extracted with diethyl
ether. The extracted product was washed and dried, and the
diethyl ether removed. The aldehyde was purified by distillation
(yield: 82%). The acetalization procedure is the same as
described above. Compound 8a was obtained in 91% yield by
distillation in vacuo (126 ЊC, 0.1 mmHg).
5-Formyl-2-thienyl 5-[(4-dimethylaminophenyl)-(E)-vinyl]-2-
thienyl sulfide 12a: yield, 60%. δH 9.74 (1 H, s), 7.58 (1 H, d),
7.35 (2 H, d), 7.22 (1 H, d), 7.04 (1 H, d), 6.95 (1 H, d), 6.91 (1
H, d), 6.86 (1 H, d), 6.69 (2 H, d), 3.00 (6 H, s), ν/cmϪ1 3080,
Compound 8a was reacted with BuLi and DMF to produce
the precursor 3a following the standard procedure detailed
J. Chem. Soc., Perkin Trans. 2, 1997
1215

View Article Online
2885, 2806, 1651, 1601, 1526, 1418, 1359, 1227, 1200, 1063, 949,
810, 752, 666.
bond in the middle of the molecule). ν/cmϪ1 2950, 1593, 1509,
1363, 1338, 1229, 1182, 947, 931, 865, 811, 793, 746, 688.
5-Formyl-2-thienyl-5-[(4-dimethylaminophenyl)-(E)-vinyl]-2-
thienyldimethylsilane 12b: yield, 47%. δH 9.95 (1 H, s), 7.80 (1
H, d), 7.39 (2 H, d), 7.33 (1 H, d), 7.20 (1 H, d), 7.12 (1 H, d),
7.09 (1 H, d), 7.07 (2 H, d), 6.90 (1 H, d), 3.01 (6 H, s), 0.67 (6
H, s), ν/cmϪ1 3054, 2954, 1673, 1605, 1524, 1425, 1358, 1252,
1202, 1061, 990, 948, 833, 806, 781, 761.
2-(5-Formyl-2-thienyl)-2-[5-[(4-dimethylaminophenyl)-(E)-
vinyl]-2-thienyl]propane 12c: yield, 69%. δH 9.83 (1 H, s), 7.60 (1
H, d), 7.32 (2 H, d), 6.99 (1 H, d), 6.95 (1 H, d), 6.78 (3 H, m),
6.68 (1 H, d), 2.97 (6 H, s), 1.86 (6 H, s), ν/cmϪ1 3070, 2969,
2802, 1667, 1604, 1520, 1445, 1359, 1217, 947, 809, 668.
Acknowledgements
The authors are indebted to the ‘Ministère de la Région Wal-
lonne’ for a research grant (Matériaux nouveaux à architecture
intelligente) and the ‘Fonds National de la Recherche Scienti-
fique (FNRS)’ for financial support. P. D. is Research Associate
of the Belgian FNRS. They also thank the ‘Services des
Affaires Scientifiques Techniques et Culturelles’ (Pôles d’At-
traction Interuniversitaires: Polymères). They warmly thank
Professor A. Persoons, Dr E. Hendrickx and Dr S Houbrechts
(KUL) for the HRS measurements.
Wittig–Horner reaction for the synthesis of NLO chromophores
1a–c: (Scheme 9)
References
As a representative example, a typical procedure for the syn-
thesis of the silane-containing chromophore 1b is as follows.
The intermediate 12b (0.48 g, 0.0012 mol) and diethyl 4-
nitrobenzylphosphonate (0.50 g, 0.0018 mol) were dried separ-
ately under vacuum and dissolved in dry THF (50 ml × 2).
BuLi (1.3 , 1.3 ml) was added dropwise to the phosphonate
solution at Ϫ78 ЊC for 5 min and the solution was then slowly
warmed to room temp. (about 30 min). Compound 12b in THF
was transferred into the anion solution at room temp. After 2 h,
the solvent was removed and the product was purified by chro-
matography on silica gel with ethyl acetate–hexane (1:4 v/v) as
eluent. Compound 1b was finally recrystallized from the same
solvent mixture at Ϫ20 ЊC (yield, 67%).
5-[(4-Nitrophenyl)-(E)-vinyl]-2-thienyl 5-[(4-dimethylamino-
phenyl)-(E)-vinyl]-2-thienyl sulfide (1a): yield, 61% (Found: C,
63.64; H, 4.87; N, 5.55; S, 19.48. Calc: C, 63.65; H, 4.52; N,
5.71; S, 19.6%). δH 8.18 (2 H, d), 7.53 (2 H, d), 7.33 (2 H, d), 7.26
(1 H, d), 7.14 (1 H, d), 7.07 (1 H, d), 6.99 (1 H, d), 6.93 (1 H, d),
6.85 (1 H, d), 6.83 (1 H, d), 6.82 (1 H, d), 6.67 (2 H, d), 2.98 (6
H, s). ν/cmϪ1 3010, 2919, 2850, 1594, 1506, 1361, 1340, 1226,
1110, 1060, 945, 812, 792, 748, 688.
1 D. F. Eaton, G. R. Meredith and J. S. Miller, Adv. Mater., 1991, 3,
564; A. Gandini and C. Méalares, Trends Polym. Sci., 1994, 2, 127.
2 C. R. Moylan, R. J. Twieg, V. Y. Lee, S. A. Swanson, K. M.
Bettereon and D. R. Miller, J. Am. Chem. Soc., 1993, 115, 12 599.
3 G. A. Lindsay, Trends Polym. Sci., 1993, 1, 138.
4 S. R. Marder, L. T. Cheng, B. G. Tiemann, A. C. Friedli, M.
Blanchard-Desce, J. W. Perry and J. Skindhoj, Science, 1994, 263,
511.
5 C. Dehu, F. Meyers and J. L. Brédas, J. Am. Chem. Soc., 1993, 115,
6198.
6 J. Zyss, in Conjugated Polymeric Materials: Opportunities in
Electronics, Optoelectronics, and Molecular Electronics (NATO
E-182), eds. J. L. Brédas and R. R. Chance, Kluwer Academic
Publishers, Dordrecht, 1989, p. 545.
7 C. R. Moylan, R. D. Miller, R. J. Twieg, K. M. Betterton, V. Y. Lee,
T. J. Matray and C. Nguyen, Chem. Mater., 1993, 5, 1499.
8 G. Mignani, A. Kämer, G. Pucetti, I. Ledoux, G. Soula and J. Zyss,
Mol. Eng., 1991, 1, 11.
9 D. Hissink, H. J. Bolink, J.-W. Eshuis, G. G. Malliaras and
G. Hadziioannou, Polym. Prepr., 1993, 34(1), 721.
10 I. Ledoux, J. Zyss, A. Jutand and C. Amatore, Chem. Phys., 1991,
150, 117.
11 B. P. Fedorov and F. M. Stoyanovich, Zh. Obshch. Khim., 1963, 33,
2251.
12 T. J. Barton and P. Boudjouk, in Silicon-Based Polymer Science, A
Comprehensive Resource, eds. J. M. Zeigler and F. W. G. Fearon,
ACS, Washington DC, 1990, p. 3.
5-[(4-Nitrophenyl)-(E)-vinyl]-2-thienyl{5-[(4-dimethyl-
aminophenyl)-(E)-vinyl]-2-thienyl}dimethylsilane (1b): yield,
67% (Found: C, 65.12; H, 5.94; N, 5.42; S, 12.29. Calc: C,
65.08; H, 5.46; N, 5.42; S, 12.41%). δH 8.19 (2 H, d), 7.57 (2 H,
d), 7.41 (1 H, d), 7.35 (2 H, d), 7.25 (1 H, d), 7.23 (1 H, d), 7.20
(1 H, d), 7.06 (2 H, m), 6.98 (1 H, d), 6.91 (1 H, d), 6.69 (2 H, d),
2.98 (6 H, s), 0.65 (6 H, s). ν/cmϪ1 3017, 2953, 1621, 1604, 1514,
1358, 1338, 1067, 990, 947, 830, 808, 779.
13 E. Hengge and S. Waldhör, Monatsh. Chem., 1974, 105, 671.
14 D. Clery, Science, 1994, 263, 1700.
15 F. Würthner, F. Effenberger, R. Wortmann and P. Krämer, Chem.
Phys., 1993, 173, 305.
16 J. X. Zhang, Ph.D. Thesis, University of Liège, Liège, 1995.
17 R. R. Fraser, T. S. Mansour and S. Savard, Can. J. Chem., 1985, 63,
3505.
2-{5-[(4-Nitrophenyl)-(E)-vinyl]-2-thienyl}-2-5-[(4-dimethyl-
aminophenyl)-(E)-vinyl]-2-thienyl}propane (1c): yield, 50%
(Found: C, 69.68; H, 6.11; N, 5.61; S, 12.89. Calc: C, 69.57; H,
5.64; N, 5.60; S, 12.81%). δH 8.18 (2 H, d), 7.52 (2 H, d), 7.31 (3
H, m), 6.98 (1 H, d), 6.96 (1 H, d), 6.84 (1 H, d), 6.82 (1 H, d),
6.78 (2 H, m), 6.76 (1 H, d), 6.68 (2 H, d), 2.97 (6 H, s), 1.85 (6
H, s). ν/cmϪ1 3072, 2967, 2802, 1620, 1603, 1589, 1336, 1109,
946, 863, 809, 746, 688.
18 B. E. Maryanoff and A. B. Reitz, Chem. Rev., 1989, 89, 863.
19 D. R. Kanis, M. A. Ratner and T. J. Marks, J. Am. Chem. Soc., 1992,
114, 10 338.
20 A. Arcoria, S. Fisichella, G. Scarlate and M. Torre, J. Heterocycl.
Chem., 1973, 10, 643.
21 P. Ruggli and A. Staub, Helv. Chim. Acta, 1937, 20, 37.
22 J. March, Advanced Organic Chemistry, Reactions, Mechanisms, and
Structures, 4th edn., Wiley, New York, 1992.
23 F. De Jong and M. J. Janssen, J. Org. Chem., 1971, 36, 1645.
24 T. H. Chan, M. A. Brook and T. Chaly, Synthesis, 1983, 203.
25 R. R. Fraser and T. S. Mansour, Tetrahedron Lett., 1986, 27, 331.
26 F. Würthner, F. Effenberger, R. Wortmann and P. Krämer, Chem.
Phys., 1993, 173, 305.
27 M. Sy and M. Maillet, Bull. Soc. Chim. Fr., 1965, 1495.
28 G. Kossmehl, M. Härtel and G. Manecke, Makromol. Chem., 1970,
131, 15.
29 J. X. Zhang, P. Dubois and R. Jerome, Synth. Commun., 1996, 26,
3091.
30 K. Clays and A. Persoons, Rev. Sci. Instrum., 1992, 63, 3285.
31 D. R. Kanis, M. A. Ratner and T. J. Marks, Chem. Rev., 1994, 94,
195.
Wittig reaction for the synthesis of 1-{5-[(4-nitrophenyl)-(E)-
vinyl]-2-thienyl-2-5-[(4-dimethylaminophenyl)-(E)-vinyl]-2-
thienylethene (1d)
5-[(4-Nitrophenyl)-(E)-vinyl]thiophene-2-carbaldehyde (0.31 g,
0.0012 mol) and 5-[(4-dimethylaminophenyl)-(E)-vinyl]thienyl-
methyl(triphenyl)phosphonium bromide (0.8 g, 0.001 37 mol)
were added to 50 ml of dry THF. Lithium ethoxide (0.4 ; 5.0
ml) was then added to the vigorously stirred suspension at room
temp. After 2 h, the product was recovered by precipitation with
ethanol, and washed with diethyl ether: yield, 61% (Found: C,
69.90; H, 5.53, N, 5.84; S, 12.76. Calc: C, 69.39; H, 4.99; N,
5.78; S, 13.23%). δH 8.19 (2 H, d), 7.57 (2 H, d), 7.34 (2 H, d),
7.33 (1 H, d), 7.16 (1 H, d), 7.09 (2 H, d), 7.02 (1 H, d), 7.00–
6.80 (ca. 4 H, m), 6.68 (2 H, d), 2.99 (6 H, s), 6.58 (ca. 0.4 H, d, J
11.9) and 6.48 (ca. 0.4 H, d, J 11.8) associated with cis-double
32 F. Kröhnke, Chem. Ber., 1950, 83, 291.
33 D. H. Wadsworth, O. E. Schupp, E. J. Sens and J. A. Ford, Jr., J. Org.
Chem., 1965, 30, 680.
Paper 6/06547C
Received 23rd September 1996
Accepted 24th February 1997
1216
J. Chem. Soc., Perkin Trans. 2, 1997