The first allylation of esters by an allylsilane: One-pot domino synthesis of triallylmethane derivatives

Article abstract of DOI:10.1002/adsc.201400296

We report the first successful allylation of aromatic esters by allyltrimethylsilane. The reaction is mediated by 0.3-1.0equiv. of titanium tetrachloride (TiCl₄) at room temperature and leads in a multi-step, one-pot reaction to quaternary triallylmethane derivatives 5. The most efficient ester possessed two ortho-methoxy substituents on the phenol ring. Molecular modelling revealed the ability of this compound to form a bidentate titanium complex, thus improving the steric accessibility of the ester carbonyl to nucleophilic attack.

Full text of DOI:10.1002/adsc.201400296

FULL PAPERS
DOI: 10.1002/adsc.201400296
The First Allylation of Esters by an Allylsilane: One-Pot Domino
Synthesis of Triallylmethane Derivatives
Chenna Kesava Reddy Bandi,a Anatoly Belostotskii,a and Alfred Hassnera,*
a
Department of Chemistry, Bar-Ilan University, Ramat-Gan, 52900, Israel
Fax : (+972)-3-738-4053; e-mail: hassna@mail.biu.ac.il
Received: March 25, 2014; Published online: June 24, 2014
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201400296.
Abstract: We report the first successful allylation of lecular modelling revealed the ability of this com-
aromatic esters by allyltrimethylsilane. The reaction pound to form a bidentate titanium complex, thus
is mediated by 0.3–1.0 equiv. of titanium tetrachlor- improving the steric accessibility of the ester carbon-
ide (TiCl4) at room temperature and leads in a multi- yl to nucleophilic attack.
step, one-pot reaction to quaternary triallylmethane
derivatives 5. The most efficient ester possessed two Keywords: allylation; allylsilanes; esters; Lewis
ortho-methoxy substituents on the phenol ring. Mo- acids; triallylmethanes
Introduction
Allyl substituents are versatile groups that can be
readily transformed by hydroboration, hydrosilyla-
tion, hydrocyanation, hydroformylation, metathesis,
epoxidation or aziridination into multiple functional
groups, some of them in an enantioselective manner.
Hence, efficient methods leading to polyallyl com-
pounds are of interest not only as building blocks for
the introduction of polyfunctional groups but also in
the synthesis of hyperbranched polymers and den-
drimers.[1]
Scheme 1. Proposed triallylation of esters using allylsilane 2
and a Lewis acid.
Carboxylic esters are readily available starting ma-
terials that could lead to triallylated products. We set
as a desirable goal a one-pot, domino conversion of
carboxylic esters 1 to quaternary triallylmethanes 5
(Scheme 1), by means of a non-metallic allylsilane 2,
using the same Lewis acid in all three steps. Allylsi-
lanes are mild, environmentally friendly reagents and
such a multi-step, atom-efficient process would also
reduce waste by minimizing the volume of solvents.
Of the three steps in Scheme 1, the second one
(3!4) should proceed readily since it is an example
of the facile Sakurai–Hosomi reaction[2] (Scheme 2) in
Scheme 2. Allylation of ketons or aldehydes using allylsilane
2 and a Lewis acid.
which aldehydes or ketones 6 undergo nucleophilic the diallyl alcohol stage. Recently, Batey et al.[4]
attack by allylsilanes in the presence of Lewis acids found that environmentally friendly non-metallic allyl
leading to homoallyl alcohols 7.
trifluoroborates (4 equiv.) were effective in the allyla-
Regarding the first step in Scheme 1, allylsilanes tion of nitriles, proceeding to diallylamines; in the
generally do not react with esters. But carboxylic case of para-methoxycarbonylbenzonitrile the ester
esters can be doubly allylated with metal allyl deriva- group was not affected. No one-step triallylation of
tives[3] of Mg, Al, Zn, Sm and the reaction stops at esters has so far been reported.
Adv. Synth. Catal. 2014, 356, 2661 – 2670
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2661

FULL PAPERS
Chenna Kesava Reddy Bandi et al.
Table 1. Formation of 5 by reaction of 1c with 2 in the pres-
ence of different Lewis acids.
Scheme 3. Allylation of activated alcohols using Lewis acid.
Although on first glance, the third step in Scheme 1
appears problematic, recent studies have shown that
activated alcohols 8 (mainly benzyl or some tertiary)
can be converted to allyl derivatives 9 (Scheme 3) by
allylsilanes using a variety of Lewis acids based on
B,[5] In,[6] Bi,[7] Fe,[8] Al,[9] Ca,[10] Zr,[11] Re,[12] Rh,[13] as
well as Brønsted acids such as Montmorillonites[14]
and PMA (phosphomolybdic acid),[15] either in cata-
lytic or stoichiometric amounts. Recently, we found
that TiCl4 in dichloromethane (DCM) is more effi-
cient than most Lewis acids and promotes, within mi-
nutes, the allylation of benzyl, allyl, as well as of some
tertiary alcohols at room temperature.[16]
Results and Discussion
The first attempts to carry out Scheme 1 were disap-
pointing. In the event, no triallyl product 5 was de-
tected in the reaction of 3.1 equiv. of allyltrimethylsi-
lane 2 in the presence of TiCl4 (1–2 equiv.) at room
temperature within 2 h with para-nitrophenyl ben-
zoate 1a and meta-nitrophenyl benzoate 1b, each pos-
sessing a good nitrophenolate leaving group. The first
promising result was isolation of ca. 15% of 5 on
heating of 1a at reflux for 2 h, though 1b did not
afford 5 on heating.
To our delight, ortho-nitrophenyl benzoate 1c in
the presence of 1 equiv. of TiCl4 at room temperature
furnished 5 within 10 min. in 58% yield, after chroma-
tography. Other Lewis acids, [BF3, SbCl5, FeCl3, InCl3,
[a]
Additional time did not improve the yields.
Isolated yields after purification.
No triallylated prodcut was formed.
[b]
[c]
BiBr3, SnCl2, TiACHTUNGTRENNUNG(O-i-Pr)4], or HOTf were ineffective,
except for AlCl3, which led to 15% of 5 after 0.5 h
(Table 1).
We have further extended this protocol to the reac-
tion of other ortho substituted phenyl esters. Thus,
On the assumption that the positive effect observed pentafluorophenyl benzoate 1n, possessing not only
with 1c may be due to additional coordination of the ortho Fs but also a good leaving group, reacted
ortho nitro function with the TiCl4 oxophile,[17] we smoothly to produce 5 in 70% yield within 1 h. The
chose several esters to study the scope of the reaction most efficient substrate by far proved to be 2,6-dime-
(see Table 2). Simple phenyl benzoate 1d requires thoxyphenyl benzoate 1j, which afforded 5 in 83%
2 equiv. of TiCl4 and 1 h to give compound 5 in low yield within 2 min at room temperature. In another
yield. The reaction with ortho-chlorophenyl benzoate aromatic ester example, 2,6-dimethoxyphenyl ester
1e did not lead to 5 at room temperature within 2 h 10a also reacted faster and gave triallyl product 17 in
but heating at reflux yielded 5 in good yield, while better yield (30 min, 65%) than the corresponding
2,4-dinitrophenyl benzoate 1i, led to 5 (40%) at room phenyl ester 10b (120 min, 35%). In many subsequent
temperature within 30 min. Indeed, ortho-methoxy- cases we used 2,6-dimethoxyphenyl esters. Thus, ester
phenyl benzoate 1f was more efficient than para-me- 11a gave triallyl product 18 in good yield. However,
thoxyphenyl benzoate 1g in providing 5 (55%, in even 2,6-dimethylphenyl benzoate 1k, led to 5 in 49%
30 min vs. 30%, in 60 min)[18] but even ortho-methyl yield after 30 min and more sterically hindered 2,6-di-
benzoate 1h gave the triallylated product in reasona- alkylated phenyl esters (1l and 1m) likewise under-
ble yield.
went triallylation in good yields, indicating that steric
2662
asc.wiley-vch.de
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2014, 356, 2661 – 2670

FULL PAPERS
The First Allylation of Esters by an Allylsilane: One-Pot Domino Synthesis
Table 2. Reactions of esters with allyltrimethylsilane and TiCl4 to form triallylmethanes.[a]
Adv. Synth. Catal. 2014, 356, 2661 – 2670
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2663

FULL PAPERS
Chenna Kesava Reddy Bandi et al.
Table 2. (Continued)
2664
asc.wiley-vch.de
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2014, 356, 2661 – 2670

FULL PAPERS
The First Allylation of Esters by an Allylsilane: One-Pot Domino Synthesis
Table 2. (Continued)
[a]
Standard reaction conditions: ester (1.0 mmol), allyltrimethylsilane 2 (3.1 mmol), TiCl4 (1.0M in CH2Cl2,
1 mmol), and CH2Cl2 (4 mL) at 258C under an inert atmosphere (Ar).
Longer reaction time did not improve the yield.
[b]
[c]
[d]
Isolated yields after purification.
While the reaction of 1e at reflux temperature gave 5 in good yield (1a gave 15% of 5), no improvement in
yield was observed for the reactions 1b, 1f, 1g under refluxing conditions even after 2–3 h.
No product was formed and starting material was recovered.
[e]
effects in the phenol portion of the ester play only
a minor role.[19]
TiCl4, a good oxophile, can coordinate with an ester
carbonyl,[20] enhancing its positive character and facili-
After successfully optimizing the conditions for the tating the first step in Scheme 1, nucleophilic attack
synthesis of triallylmethanes on phenyl esters, we ex- by the allylsilane 2. The second and third steps in the
amined the heterocyclic system 12. Gratifyingly, the cascade (Scheme 1) are reactions which normally take
reaction of 12a, 12b with allyltrimethylsilane 2 fur- place within a few minutes at room temperature,
nished the corresponding triallylated product 19 in keeping in mind that aromatic esters will be preferred
good yields (50, 40%). Finally, the versatility of the substrates since they promote the third step in
reaction was explored with substrates having two Scheme 1. Indeed the most reactive examples
ester groups (13a, 14a, 15a). Both para- and meta-di- (Table 2) were aromatic esters. In the allylation of
ACHTUNGTRENNUNGsubstituted esters (13a, 14a) gave triallylated methane phenolate esters, an ortho substituent (OMe, NO2) ca-
products 20, 21 in good yields (55, 50%), while 15a pable of coordination with TiCl4 has a beneficial
gave diallylated lactone 22 in moderate yield (30%), effect with the di-ortho-methoxy ester 1j being the
as expected due to neighbouring group interaction. most reactive.
TiCl4 proved effective in the one-pot triallylation of
After the first step in the ester allylation cascade in
benzoate esters not only of phenols, but also of alco- which the phenolate is lost, identical chemical struc-
hols (e.g., 1o, 1p) at room temperature in reasonable tures are the reaction participants in both second and
yield. During the reaction of 1c–n, phenolate is elimi- third step allylations of Scheme 1 for all substrates 1.
nated and therefore phenol esters give better yields In this light, the first step involving TiCl4 complexa-
than alcohol esters. Furthermore, we extended our tion should be rate-determining. To better understand
protocol to aliphatic ester 16a, which gave only dially- the impact of molecular structure on the TiCl4-pro-
lated alcohol 23 as a major product (50%, requiring moted reaction of aryl esters 1 with allylsilane 2, we
2 h). This is not surprising, since further allylation focused on the initial stage of this multi-step transfor-
most often requires activated alcohols.
mation and performed theoretical modelling for
Adv. Synth. Catal. 2014, 356, 2661 – 2670
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2665

FULL PAPERS
Chenna Kesava Reddy Bandi et al.
esters 1a, 1c, 1d, 1f, 1h, 1j, 1k, 1n, and 11a, as well as
for 1:1 and 1:2 ester:TiCl4 complexes.[21] These com-
plexes have been computationally located by using
DFT calculations with modelling dichloromethane so-
lution (see Experimental Section and the Supporting
Information).
The computationally located 1:1 ester-TiCl4 com-
plexes have different chemical structures, depending
on the presence of functional groups. Monodentate
C=O–TiCl4 complexes 1aTi, 1cTi, 1dTi, 1fTi, 1hTi, 1jTi,
1kTi, 1nTi, and 11aTi have been located for both non-
functionalized esters 1d, 1h, 1k, and for esters 1a, 1c,
1f, 1j, 1n, 11a with a heteroatom-bearing substituent
(see Figure 1 for 1jTi and the Supporting Information
for others).[22] Methoxy-substituted phenyl esters 1fTi,
1jTi, and 11a provide additional complexes; they are
monodentate MeO-liganded complexes 1fTiꢀOMe, 1jTiꢀ
OMe, and 11aTiꢀ
(chemically isomeric to 1fTi, 1jTi and
OMe
11a, respectively; Figure 1 and the Supporting Infor-
mation), as well as bis Ti complexes 1j2Ti, 1f2Ti and
ꢀ
2O
1j2Ti
(Figure 1 and Supporting Information). In ad-
ꢀ
2O
dition we found bidentate Ti complexes 1fTiꢀ2O, 1jTiꢀ
2O
and 11aTiꢀ (Figure 1 and the Supporting Informa-
2O
tion). Similarly, our calculations have located NO2-li-
ganded Ti complexes 1aTiꢀ and 1cTiꢀ for function-
ON
ON
alized esters 1a and 1c, respectively (Figure 1a in the
Supporting Information). Energy calculations per-
formed at the MP2/6-31+G
G
ꢀ
MeO TiCl4 complexes are slightly more stable than
ꢀ
the C=O TiCl4 ones, while the latter have a signifi-
cantly higher stability than the NO2 TiCl4 com-
plexes.[22]
Molecular geometries of the modelled Ti complexes
indicate that the complexes, in which the carbonyl
oxygen and the ether oxygen both chelate the Ti
atom (cf. 1jTiꢀ2O) are more reactive in the ester allyla-
tion than other Ti complexes. In the monodentate C=
ꢀ
O TiCl4 complexes the carbonyl carbon is severely
shielded from being attacked by a nucleophile. Both
ꢀ
the OPh ring and Ti Cl fragments are exposed near-
orthogonally to the C=CH2 plane of the approaching
ꢀ
allylsilane molecule (Figure 2). By turning the O Ph
ring by ~308 and shifting the TiCl4 molecule from the
Figure 1. Optimized molecular geometries for complexes
ꢀ
O Ph ring plane, chelation of the Ti atom in biden-
1jTi, 1jTiꢀOMe, 1j2Ti 2O, 1fTiꢀ2O and 1jTiꢀ
.
ꢀ
2O
date complexes 1fTiꢀ2O, 1jTiꢀ and 11aTiꢀ opens one
2O
2O
stereoface of the carbonyl fragment. This striking dif-
ference in steric crowding makes evident that these understanding of the structure-reactivity relationship
chelates (and not their monodentate chemical iso- for this allylation (Supporting Information).
mers) are the Ti complexes which undergo the first
Methoxy-substituted esters 1f, 1j, and 11a are capa-
ester allylation. This conclusion successfully explains ble of chemically differently complexing TiCl4
the observed higher reactivity (see above) of phenyl (Figure 1 and Supporting Information). Chelation of
esters bearing ortho-methoxy substituents that are ca- Ti both by the C=O and the ortho-substituent signifi-
pable of complexing TiCl4. This also explains the cantly diminishes steric hindrances for a nucleophilic
much better yield of 5 from ortho-nitro ester 1c than attack, as is shown for the ortho-MeO structures 1fTiꢀ
from the para isomer 1a (Table 2). Calculated
and 1jTiꢀ (Figure 2). Therefore these bidentate
2O
2O
HOMOs and LUMOs for the modelled esters and complexes are the reactive electrophilic substrates in
their complexes with TiCl4 provide further qualitative the first allylation (Scheme 4). In orbital terms, this
2666
asc.wiley-vch.de
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2014, 356, 2661 – 2670

FULL PAPERS
The First Allylation of Esters by an Allylsilane: One-Pot Domino Synthesis
for corresponding alternative monodentate Ti com-
plexes 1fTiꢀ 1jTiꢀ and 1j2Ti, respectively
,
,
OMe
OMe
(Scheme 4 and Supporting Information). Thus, Ti spe-
cies 1fTi, 1jTi and 11aTi are favoured precursors of the
corresponding reactive complexes 1fTiꢀ2O, 1jTiꢀ and
2O
11aTiꢀ .
2O
The exceptionally high reactivity of ester 1j
(Table 1a in the Supporting Information) becomes ap-
parent from the LUMOꢀMO-nO energy gap which is
significantly lower for dimethoxy structure 1jTi com-
pared to monomethoxy and trimethoxy analogues 1fTi
and 11aTi.[24]
Conclusions
In summary, we have developed for the first time
a synthetic strategy in which aromatic esters are con-
verted to triallylated products by means of allyltrime-
thylsilane and TiCl4 in an efficient and rapid cascade
reaction at room temperature. The successful forma-
tion of bis triallylated products 19 and 20 may be val-
uable for the synthesis of hyperbranched polymers
and dendrimers. Molecular modelling studies indicate
that coordination of ortho-methoxy groups to TiCl4
relieves steric congestion around the carbonyl group
of phenolate esters facilitating allylation.
Figure 2. Steric hindrances (indicated by dotted arcs) in the
nucleophilic attack of the allylsilane (shown schematically
on the right of each structure) on monodentate Ti com-
plexes 1dTi, 1fTi, 1hTi, 1jTi, 1kTi and 1nTi, as well as bidendate
complexes Ti 1fTiꢀ and 1jTiꢀ2O. In the optimized molecular
2O
geometries (shown) of the Ti complexes, bent arrows point
out the carbon atom of the carbonyl unit.
Experimental Section
General Methods
Unless otherwise noted, chemicals were purchased from
commercial suppliers at the highest purity grade available
and were used without further purification. Solvents were
distilled by standard methods. Thin layer chromatography
was performed on Merck pre-coated 0.25 mm silica gel
plates (60F-254) using UV light as visualizing agent and/or
iodine as developing agent. Neutral alumina (Fluka; pH/:
7.0+/- 0.5) was used for column chromatography.
Melting points were recorded on an uncorrected Perfit
Melting Point instrument. IR spectra were recorded on an
FT-IR spectrometer and expressed as wave numbers (cmꢀ1).
1H and 13C NMR spectra were recorded on a Bruker
(300 MHz and 75 MHz) spectrometer. Spectra were refer-
enced internally to the residual proton resonance in CDCl3
(d=7.26 ppm) or with tetramethylsilane (TMS, d=
0.00 ppm) as the internal standard. Chemical shifts (d) were
reported as part per million (ppm) in d scale downfield from
TMS. 13C NMR spectra were referenced to CDCl3 (d=
77.0 ppm, the middle peak). Coupling constants were ex-
pressed in Hz. The following abbreviations were used to ex-
plain the multiplicities: s=singlet, d=doublet, t=triplet,
dt=doublet of triplet, m=multiplet, br=broad. High reso-
lution mass spectra (HR-MS) were obtained on a Bruker
micrOTOFꢁ-Q II mass spectrometer (ESI-MS).
Scheme 4. Formation of bidendate Ti complexes 1fTiꢀ2O, 1jTiꢀ
ꢀ
ꢀ
2O, and 1jTiꢀ from monodentate C=O TiCl4 and MeO
2O
TiCl4 complexes. Grey background indicates favoured mon-
odentate precursors, MO-nO refers to the highest occupied
molecular orbital for which localization includes the O
atoms.
chelation is provided by interaction of the LUMO (lo-
calized on the TiCl4 fragment) and the molecular or-
bital associated with a higher energy lone electron
pair of the carbonyl or ether oxygen (Figure 1a in the
Supporting Information).[23]
The energy gap DELꢀH between the LUMO and
this molecular orbital for complexes 1fTi, 1jTi and
11aTi is 0.038523, 0.033990, and 0.036313 hartrees, re-
Ester compounds (1a–1p, 10a, 10b, 11a, 12a, 12b, 13a, 14a,
spectively. Similar values of DELꢀH are much larger 15a, 16a) were prepared using literature procedures, and for
Adv. Synth. Catal. 2014, 356, 2661 – 2670
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2667

FULL PAPERS
Chenna Kesava Reddy Bandi et al.
known compounds data have been compared with the re-
ported data. Characterization data for new compounds is
given below.
Bis(2,6-dimethoxyphenyl) isophthalate (14a, Table 2):
Yield: 3.65 g (83%); white solid; mp 154–1568C; H NMR
1
(300 MHz, CDCl3): d=3.81 (s, 12H), 6.67 (d, J=8.4 Hz,
4H), 7.18 (t, J=8.4 Hz, 2H), 7.65 (t, J=7.8 Hz 1H), 8.49
(dd, J=7.8, 6.8 Hz, 2H), 9.10 (m, 1H); 13C NMR (75 MHz,
CDCl3): d=56.4, 105. 2, 119.3, 126.7, 129.0, 130.3, 132.6,
135.3, 147.5, 152.7, 164.0; MS: m/z (rel intensity)=461 (100)
[M+Na]+.
Calculations
Non-empirical calculations have been performed using
Gaussian09 package. PCM model (an implicit model of sol-
vation) implemented in this package has been used for mod-
elling dichloronmethane solution. The molecular geometry
of all the modelled Ti complexes has been optimized by em-
ploying DFT calculations at the MPWLYPM1/cc-pVDZ
level; for complexes 1aTi, 1aTiꢀON, 1cTi, 1cTiꢀON, 1jTi, and 1jTiꢀ
OMe, also by using ab initio calculations at the MP2/6–31+G-
Bis(2,6-dimethoxyphenyl) phthalate (15a, Table 2): Yield:
3.2 g (73%); white solid; mp 159–1618C; 1H NMR
(300 MHz, CDCl3): d=3.75 (s, 12H), 6.61 (d, J=8.4 Hz,
4H), 7.12 (t, J=8.4 Hz, 2H), 7.62 (dd, J=9.0, 5.7 Hz, 2H),
8.10 (dd, J=9.0, 5.7 Hz, 2H); 13C NMR (75 MHz, CDCl3):
d=56.4, 105.3, 119.3, 126.5, 129.3, 130.2, 131.5, 132.1, 152.9,
164.6; MS: m/z (rel intensity)=461 (100) [M+Na]+.
2,6-Dimethoxyphenyl hexanoate (16a, Table 2): Yield:
ACHTUNGTRENNUNG(d,p) level. In cases of difficult convergence, successful ge-
ometry optimization has been achieved by combining (i)
Newton–Raphson method, (ii) the use of Cartesian coordi-
nates, (iii) computation of force constants at the first point,
and (iv) involvement of the quadratically convergent SCF
procedure [in Gaussian terms, by combining commands
Opt=(Newton, Cartesian, CalcFc) and SCF=QC (or
SCF=XQC, if unsuccessful)]. For initial structures of Ti
complexes, see the Supporting Information.
1
2.2 g (87%); yellow liquid; H NMR (300 MHz, CDCl3): d=
0.9 (t, J=7.1 Hz, 3H), 1.32–1.49 (m, 4H), 1.80 (pent, J=
7.4 Hz, 2H), 2.62 (t, J=7.4 Hz, 2H), 3.88 (s, 6H), 6.60 (dd,
J=6.3, 1.8 Hz, 2H), 7.12 (m, 1H); 13C NMR (75 MHz,
CDCl3): d=14.2, 22.6, 25.0, 31.4, 34.0, 56.3, 105.1, 126.3,
129.1, 152.6, 171.8; MS: m/z (rel intensity)=252 (100) [M]+.
General Procedure for the Preparation of
Triallylmethane Products
General Procedure for the Synthesis of Esters:[25]
To an ice-cold solution of phenol (11 mmol), triethylamine
(11 mmol) in DCM (10 mL) was added dropwise an acyl
chloride (10 mmol). After stirring at room temperature for
30 min and washing with 1M HCl solution (5ꢂ2 mL), the
combined organic layers were dried on anhydrous MgSO4.
The ester product was recrystallized from ethanol-hexane.
2,6-Dimethoxyphenyl 4-methylbenzoate (10a, Table 2):
Yield: 2.4 g (88%); white solid; mp 99–1018C; 1H NMR
(300 MHz, CDCl3): d=2.45 (s, 3H), 3.79 (s, 6H), 6.62 (d,
J=8.5 Hz, 2H), 7.16 (t, J=8.5 Hz, 1H), 7.29 (d, J=8.1 Hz,
2H), 8.16 (d, J=8.3 Hz, 2H); 13C NMR (75 MHz, CDCl3):
d=22.0, 56.4, 105.2, 119.3, 126.5, 126.9, 129.4, 130.0, 130.7,
144.4, 152.9, 164.8; MS: m/z (rel intensity)=295 (100) [M+
Na]+.
2,6-Dimethoxyphenyl 4-methoxybenzoate (11a, Table 2):
Yield: 2.2 g (76%); white solid; mp 110–1128C; 1H NMR
(300 MHz, CDCl3): d=3.79 (s, 6H), 3.87 (s, 3H), 6.66 (d,
J=8.5 Hz, 2H), 6.99 (d, J=8.7 Hz, 2H), 7.15 (t, J=8.5 Hz,
1H), 8.18 (d, J=8.8 Hz, 2H); 13C NMR (75 MHz, CDCl3):
d=55.7, 56.4, 105.1, 113.9, 119.3, 122.0, 126.4, 132.7, 152.9,
164.0; MS: m/z (rel intensity)=311 (100) [M+Na]+.
To a stirred solution of ester (1 mmol) in dry DCM (4 mL),
were added TiCl4 (1.0M sol.ution in DCM, 1 mL) and allyl-
trimethylsilane (3.1 mmol) under an inert atmosphere and
kept at 208C for the stipulated time (see Table 2). Longer
reaction times did not improve the yield. After completion
of the reaction (TLC), the excess of solvent was removed by
rota-vapor and the residue was purified by neutral alumina
column chromatography, using ethyl acetate (1–2%) in hex-
anes as an eluent to give the triallylated products in good to
excellent yields.
(4-Allylhepta-1,6-dien-4-yl)benzene
(5,
Table 2):[26]
1H NMR (300 MHz, CDCl3): d=2.62 (d, J=7.0 Hz, 6H),
5.12–5.24 (m, 6H), 5.64–5.80 (m, 3H), 7.30–7.38 (m, 1H),
7.47 (d, J=4.7 Hz, 4H); 13C NMR (75 MHz, CDCl3): d=
42.1, 43.5, 117.8, 126.0, 126.9, 128.3, 134.8, 145.9; MS (EI,
70 eV): m/z (%)=212 (M+, 23), 171 (86); HR-MS (EI+):
m/z=212.1526, calcd. for C16H20 [M]+: 212.1565.
1-(4-Allylhepta-1,6-dien-4-yl)-4-methylbenzene
(17,
1
Table 2): H NMR (300 MHz, CDCl3): d=2.33 (s, 3H), 2.45
(d, J=7.2 Hz, 6H), 4.98–5.04 (m, 6H), 5.58–5.70 (m, 3H),
7.15 (d, J=8.1 Hz, 2H), 7.20 (d, J=8.3 Hz, 2H); 13C NMR
(75 MHz, CDCl3): d=21.2, 42.2, 43.2, 117.8, 126.8, 129.0,
135.0, 135.4, 142.9.
2,6-Dimethoxyphenyl
thiophene-2-carboxylate
(12b,
Table 2): Yield: 2.1 g (79%); white solid; mp 112–1148C;
1H NMR (300 MHz, CDCl3): d=3.80 (s, 6H), 6.64 (d, J=
8.4 Hz, 2H), 7.17–7.19 (m, 2H), 7.64 (dd, J=3.8, 1.1 Hz,
1H), 8.01 (dd, J=2.6, 1.1 Hz, 1H); 13C NMR (75 MHz,
CDCl3): d=56.4, 105.2, 126.7, 128.1, 128.8, 132.9, 133.4,
134.9, 152.9, 160.2; MS, m/z (rel intensity)=265 (50) [M+
H]+, 193 (100).
1-(4-Allylhepta-1,6-dien-4-yl)-4-methoxybenzene
(18,
Table 2):[1e] 1H NMR (300 MHz, CDCl3): d=2.44 (d, J=
7.2 Hz, 6H), 3.81 (s, 3H), 4.98–5.08 (m, 6H), 5.50–5.70 (m,
3H), 6.87 (dd, J=8.9, 6.6 Hz, 2H), 7.23 (dd, J=8.9, 6.7 Hz,
2H); 13C NMR (75 MHz, CDCl3): d=42.2, 42.9, 55.4, 113.6,
113.8, 117.8, 118.6, 127.9, 134.9, 138.0, 157.7.
Bis(2,6-dimethoxyphenyl) terephthalate (13a, Table 2):
Yield: 3.8 g (86%); white solid; mp 205–2078C; 1H NMR
(300 MHz, DMSO-d6): d=3.83 (s, 12H), 6.87 (d, J=8.4 Hz,
4H), 7.31 (t, J=8.7 Hz, 2H), 8.37 (s, 4H); 13C NMR
(75 MHz, DMSO-d6): d=57.3, 106.3, 128.0, 129.0, 131.0,
131.2, 131.6, 134.2, 153.1, 164.1; MS: m/z (rel intensity)=461
(100) [M+Na]+.
2-(4-Allylhepta-1,6-dien-4-yl)thiophene (19, Table 2):
1H NMR (300 MHz, CDCl3): d=2.43 (d, J=7.4 Hz, 6H),
5.00–5.10 (m, 6H), 5.60–5.67 (m, 3H), 6.79 (d, J=3.4 Hz,
1H), 6.93 (m, 1H), 7.18 (d, J=5.1 Hz, 1H); 13C NMR
(75 MHz, CDCl3): d=43.3, 43.4, 117.9, 123.0, 123.5, 126.2,
133.9, 152.0; MS (EI, 70 eV): m/z (%)=248 (M+, 27), 177
(100).
2668
asc.wiley-vch.de
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2014, 356, 2661 – 2670

FULL PAPERS
The First Allylation of Esters by an Allylsilane: One-Pot Domino Synthesis
1,4-Bis(4-allylhepta-1,6-dien-4-yl)benzene (20, Table 2):[1e]
1H NMR (300 MHz, CDCl3): d=2.43 (d, J=7.2 Hz, 12H),
4.96–5.01 (m, 12H), 5.47–5.60 (m, 6H), 7.23 (s, 4H).
13C NMR (75 MHz, CDCl3): d=42.1, 43.3, 117.8, 126.6,
134.9, 135.0, 143.2.
1,3-Bis(4-allylhepta-1,6-dien-4-yl)benzene (21, Table 2):
1H NMR (300 MHz, CDCl3): d=2.42 (d, J=7.0 Hz, 12H),
4.94–5.00 (m, 12H), 5.45–5.57 (m, 6H), 7.07–7.10 (m, 2H),
7.20–7.23 (m, 2H); 13C NMR (75 MHz, CDCl3): d 42.2, 43.8,
117.8, 124.3, 125.7, 127.9, 134.9, 145.4;
5699; b) using Mg or Al reagents, see: K. H. Shen, C.-
W. Kuo, C.-F. Yao, Tetrahedron Lett. 2007, 48, 6348–
6351; c) H. Tanaka, S. Nakahata, H. Watanabe, J. Zhao,
M. Kuroboshi, S. Torii, Inorg. Chim. Acta 1999, 296,
204–207; d) using Sm reagents: Z. Li, Y. Zhang, Tetra-
hedron 2002, 58, 5301–5306.
[4] T. R. Ramadhar, J. Bansagi, R. A. Batey, J. Org. Chem.
2013, 78, 1216–1221.
[5] Using boron derivatives, see: a) J. A. Cella, J. Org.
Chem. 1982, 47, 2125–2130; b) M. Rubin, V. Gevorgy-
an, Org. Lett. 2001, 3, 2705–2707; c) G. W. Kabalka,
M. L. Yao, S. Borella, J. Am. Chem. Soc. 2006, 128,
11320–11321.
[6] Using indium derivatives, see: a) M. Yasuda, T. Saito,
M. Ueba, A. Baba, Angew. Chem. 2004, 116, 1438–
1440; Angew. Chem. Int. Ed. 2004, 43, 1414–1416; b) S.
Kim, C. Shin, A. Pae, H. Koh, M. Chang, B. Chung, Y.
Cho, Synthesis 2004, 1581–1584; c) T. Saito, M. Yasuda,
A. Baba, Synlett 2005, 1737–1739; d) T. Saito, Y. Nishi-
moto, M. Yasuda, A. Baba, J. Org. Chem. 2006, 71,
8516–8522; e) T. Saito, Y. Nishimoto, M. Yasuda, A.
Baba, J. Org. Chem. 2007, 72, 8588–8590.
3,3-Diallylisobenzofuran-1(3H)-one
(22,
Table 2):[4]
1H NMR (300 MHz, CDCl3): d=2.61–2.78 (m, 4H), 5.01 (s,
2H), 5.05–5.07 (m, 2H), 5.46–5.60 (m, 2H), 7.36 (d, J=
7.7 Hz, 1H), 7.48 (t, J=7.9 Hz, 1H), 7.64 (td, J=7.7, 7.3 Hz,
1H), 7.84 (d, J=7.6 Hz, 1H); 13C NMR (75 MHz, CDCl3):
d=42.9, 88.5, 120.6, 121.8, 125.9, 127.1, 129.3, 130.9, 134.1,
151.8, 170.1.
4-Allyloct-1-en-4-ol (23, Table 2):[27] 1H NMR (300 MHz,
CDCl3): d=0.90 (t, J=6.7 Hz, 3H), 1.27–1.46 (m, 8H), 2.21
(d, J=7.5 Hz, 4H), 5.08–5.14 (m, 4H), 5.77–5.90 (m, 2H);
13C NMR (75 MHz, CDCl3): d=14.3, 22.9, 32.6, 39.4, 44.0,
73.8, 118.9, 134.1.
[7] Using bismuth derivatives see: a) K. D. Surya, A. G.
Richard, Tetrahedron Lett. 2005, 46, 8345–8350;
b) G. G. K. S. N. Kumar, K. K. Laali, Org. Biomol.
Chem. 2012, 10, 7347–7355.
[8] Using iron derivatives, see: a) G. W. Kabalka, M. L.
Yao, S. Borella, L. K. Goins, Organometallics 2007, 26,
4112–4114; b) J. Han, Z. Cui, J. Wang, Z. Liu, Synth.
Commun. 2010, 40, 2042–2046.
Acknowledgements
We are grateful to the India-Israel Fund for a fellowship to
C.K.R.B. and to the Marcus Center for Medicinal and Phar-
maceutical Chemistry at Bar-Ilan University for partial sup-
port of this research.
[9] S. J. Mahoney, T. Lou, G. Bondarenko, E. Fillon, Org.
Lett. 2012, 14, 3474–3477.
[10] V. J. Meyer, M. Niggemann, Eur. J. Org. Chem. 2011,
3671–3674.
References
[1] Dendrimers are widely used in various medicinal and
pharmaceutical applications. See: a) D. K. Smith, F.
Diederich, Chem. Eur. J. 1998, 4, 1353–1361; b) W.
Maes, W. Dehaen, Eur. J. Org. Chem. 2009, 4719–4752;
c) H. Hagiwara, M. Sekifuji, N. Tsubokawa, T. Hoshi,
T. Suzuki, Chem. Lett. 2009, 38, 926–927; for allyl func-
tionalized dendrimers, see: d) R. Meijboom, J. R. Moss,
A. T. Hutton, T.-A. Makaluza, S. F. Mapolle, F. Waggie,
M. R. Domingo, J. Organomet. Chem. 2004, 689, 1876–
1881; e) S. Nlate, L. Plault, F.-X. Felpin, D. Astruc,
Adv. Synth. Catal. 2008, 350, 1419–1424; f) J. S. Lindsey,
J. Org. Chem. 2005, 70, 7972–7978; g) V. Sartor, S.
Nlate, J. L. Fillaut, L. Djakovitch, F. Moulines, V. Mar-
vaud, F. Neveu, J. C. Blais, J. F. Letard, D. Astruc, New
J. Chem. 2000, 24, 351–370; h) V. Martinez, J.-C. Blais,
D. Astruc, Organometallics 2004, 23, 861–874.
[2] a) H. Sakurai, A. Hosomi, M. Kumada, J. Org. Chem.
1969, 34, 1764–1768; b) H. Sakurai, Pure Appl. Chem.
1982, 54, 1–22; c) I. Fleming, J. Dunogues, R. Smithers,
Org. React. 1989, 37, 57–575; d) H. Urabe, F. Sato,
Ti(IV) Lewis acids, in: Lewis Acids in Organic Synthe-
sis, (Ed.: H. Yamamoto), Vol. 2, Wiley-VCH, Wein-
heim, 2000, Chap. 2, pp 653–798; e) Esters remain unaf-
fected during the Sakurai reaction, see: P. A. Robert-
son, J. A. Katzenellenbogen, J. Org. Chem. 1983, 48,
5288–5302.
[11] G. V. M. Sharma, K. L. Reddy, P. S. Lakshmi, R. Ravi,
A. C. Kunwar, J. Org. Chem. 2006, 71, 3967–3969.
[12] Y. Kuninobu, E. Ishii, K. Takai, Angew. Chem. 2007,
119, 3360–3363; Angew. Chem. Int. Ed. 2007, 46, 3296–
3299.
[13] G. Onodera, E. Yamamoto, S. Tonegawa, M. Iezumi,
R. Takeuchi, Adv. Synth. Catal. 2011, 353, 2013–2021.
[14] Using Montmorillonite derivatives, see: a) K. Moto-
kura, N. Nakagiri, T. Mizugaki, K. Ebitani, K. Kaneda,
J. Org. Chem. 2007, 72, 6006–6015; b) J. Wang, Y.
Masui, M. Onaka, Tetrahedron Lett. 2010, 51, 3300–
3303.
[15] S. T. Kadam, H. Lee, S. S. Kim, Appl. Organomet.
Chem. 2010, 24, 67–70.
[16] A. Hassner, C. R. Bandi, Synlett 2013, 1275–1279.
[17] a) H. Hanawa, S. Kii, N. Asao, K. Maruoka, Tetrahe-
dron Lett. 2000, 41, 5543–5546; b) H. Hanawa, D. Ura-
guchi, S. Konishi, T. Hashimoto, K. Maruoka, Chem.
Eur. J. 2003, 9, 4405–4413; these authors had shown
that bidentate coordination of aldehydes promotes
their reaction with allyltributyltin. We surmised that,
besides coordination of TiCl4 with the ester carbonyl
group, additional coordination with an appropriate
neighbouring substituent might increase the electro-
philic character of the ester carbonyl and facilitate nu-
cleophilic attack by the allylsilane.
[3] Diallylation of esters, a) using Zn reagents: Y.-J. Wei,
H. Ren, J.-X. Wang, Tetrahedron Lett. 2008, 49, 5697–
Adv. Synth. Catal. 2014, 356, 2661 – 2670
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2669

FULL PAPERS
Chenna Kesava Reddy Bandi et al.
[18] Except in the case of 1e, heating under reflux tempera-
ture did not lead to significant improvement in the
yield of 5.
[19] The reactivity of sterically encumbered 2,6-disubstitut-
ed phenolates may be due in part to decompression on
loss of the phenolate after initial attack of the allylsi-
lane on the ester.
[20] Esters can complex with TiCl4, see: a) N. J. Clayden, S.
Holmes, P. J. V. Jones, J. Chem. Soc. Chem. Commun.
1988, 1289–1291; b) G. Maier, H. O. Kalinowski, C.
Weber, M. Henrich, Magn. Reson. Chem. 1995, 33,
290–296.
[21] Esters of carboxylic acids and TiCl4 usually form com-
plexes of 1:1 composition: a) O. Saied, B. Bachand,
J. D. Wuest, Can. J. Chem. 1998, 76, 490–497; b) G.
Maier, H.-O. Kalinowski, C. Weber, M. Henrich, Magn.
Reson. Chem. 1995, 33, 290–296.
the 0.207–0.209 nm range, while the range of the length
ꢀ
of the O Ti bond (equatorial) from the MeO fragment
ꢀ
is 0.245–0.247 nm. The C=O -TiCl4 complex 1jTi is
lower in energy by 0.9 kcalmolꢀ1 than its chemical
ꢀ
ꢀ
isomer 1jTiꢀ
with an MeO TiCl4 bond. The NO2
OMe
TiCl4 complex 1cTiꢀ has a 4.2 kcalmolꢀ1 higher energy
ON
ꢀ
than the related C=O TiCl4 complex 1cTi (Supporting
Information).
[23] This molecular orbital is an analogue of the HOMO
for simplest AlkC(=O)X or AlkOAlk compounds
(non-bonding O-centered orbital nO). It interacts with
the TiCl4-localized LUMO of the monodentate com-
plex when liganding the Ti atom to the second oxygen.
[24] However, the DELꢀH differences for 1fTi and 11aTi
(Scheme 4) do not explain the relative reactivity of the
corresponding esters (Table 2). Since the frontier orbi-
tal approximation does not take into account steric fac-
tors that appear when reacting molecules or molecular
fragments draw closer, we also considered structure-
specific details of molecular geometry. The distance be-
tween the Ti-liganding oxygens is quite similar in che-
lates 1fTiꢀ and 11aTiꢀ (0.275 and 0.279 nm, respec-
[22] Modelled monodentate complexes have trigonal bipyr-
amidal geometry of the Ti-centered molecular unit, and
ꢀ
the oxygen ligand occupies an apical position. O Ti
bonds between the carbonyl oxygen of esters with TiCl4
are shorter than such bonds between the ether or nitro
group oxygen with this Lewis acid. The calculated
2O
2O
ꢀ
tively). However, the shorter interatomic O O non-
bonging distance in complex 1fTi (0.321 nm) vs. that in
11aTi (0.362 nm) facilitates intramolecular chelation and
leads to more efficient conversion of 1fTi vs. 11aTi into
ꢀ
length of the C=O TiCl4 bond in the modelled com-
plexes is within the range of 0.209–0.213 nm, while the
ꢀ
length of the O Ti bond (apical) is 0.237, 0.224 and
ꢀ
0.227 nm in the MeO TiCl4 complex 1jTiꢀ
and in the
the related reactive complexes 1fTiꢀ2O and 11aTiꢀ .
OMe
2O
NO2-liganded Ti complexes 1aTiꢀ and 1cTiꢀON, respec-
tively. The O Ti (apical) bond in bis-Ti complexes
[25] C. K. Lee, J. S. Yu, H. J. Lee, J. Heterocycl. Chem. 2002,
39, 1207–1217.
[26] V. Martinez, J.-C. Blais, D. Astruc, Organometallics
2004, 23, 861–874.
ON
ꢀ
1f2Ti 2O, 1j2Ti
and 1j2Ti is elongated to a length of
ꢀ
ꢀ
2O
0.247 to 0.249 nm. Bidentate complexes 1fTiꢀ and 1jTiꢀ
2O
have the Ti atom of somewhat distorted bipyramidal
[27] a) J. Barluenga, C. Rubiera, J. R. Fernꢃndez, J. Florez,
M. Yus, Synthesis 1987, 819–821; b) R. L. Snowden,
S. M. Linder, B. L. Muller, K. H. Schulte-Elte, Helv.
Chim. Acta 1987, 70, 1858–1878.
2O
square geometry with both oxygen ligands occupying
ꢀ
equatorial positions. Values of the length of the O Ti
bond (equatorial) from the carboxyl fragment are in
2670
asc.wiley-vch.de
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2014, 356, 2661 – 2670