Titanium-mediated rearrangement of cyclopropenylmethyl acetates to (E)-halodienes

Article abstract of DOI:10.1039/c0ob01046d

TiCl₄ and TiBr₄ rapidly transform cyclopropenylmethyl acetates to (E)-halodienes via ring-opening to allyl-vinyl cations. DFT calculations suggest that the regioselectivity of the halogenation of this cationic intermediate by [TiX

Full text of DOI:10.1039/c0ob01046d

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 3359
www.rsc.org/obc
PAPER
Titanium-mediated rearrangement of cyclopropenylmethyl acetates to
(E)-halodienes†
Gary Gallego,a Alireza Ariafard,*b,c Kiet Tran,a David Sandoval,a Leera Choi,a Yi-Hsun Chen,a
Brian F. Yates,b Fu-Ming Taoa and Christopher J. T. Hyland*a,b
Received 18th November 2010, Accepted 3rd February 2011
DOI: 10.1039/c0ob01046d
TiCl4 and TiBr4 rapidly transform cyclopropenylmethyl acetates to (E)-halodienes via ring-opening to
allyl-vinyl cations. DFT calculations suggest that the regioselectivity of the halogenation of this cationic
intermediate by [TiX4OAc]- is under thermodynamic control, while the stereoselectivity is governed by
kinetics.
Introduction
The Lewis-acid mediated abstraction of propargylic leaving groups
to provide a propargyl cation is a widely investigated method
for functionalisation of propargyl alcohol-derived substrates.1
Cyclopropenylmethyl acetates are an interesting subclass of
cyclopropene, which due to the p-rich nature of their dou-
ble bond bear some resemblance to propargyl alcohol-derived
substrates.2 Indeed, we have previously reported that p-activation
of these cyclopropenes with Au(I) catalysts, results in stereose-
lective rearrangement to (Z)-acetoxydienes via 2 (Scheme 1).3
Within this context, we proposed that divergent reactivity by
s-activation with oxaphilic Lewis-acids, such as TiX4, may be
brought about. Acetate abstraction by TiX4 from cyclopropenes
1 could potentially generate cyclopropenyl cation 3, which could
undergo ring-opening to cation 4. These two cations can then
potentially be intercepted by X- originating from the TiX4 to yield
halogenated products.4 As both 3 and 4 can react with X- at more
than one electrophilic site, this proposed reaction raises several
chemoselectivity and regioselectivity questions.
In this paper, we demonstrate that upon treatment with TiX4,
cyclopropenes 1 can be transformed regioselectively and stereose-
lectively into (E)-halo-dienes and use DFT calculations to propose
a mechanistic rationale for this reaction.
Scheme 1 Activation of cyclopropenylmethyl acetates with p- and
s-Lewis acidic metals.
Results and discussion
We began our experimental investigations by preparing the re-
quired cyclopropenylmethyl acetates from tribromocyclopropanes
using the method of Baird et al.,5 followed by acetylation of
the resulting alcohols. (Scheme 2). Nitrobenzaldehyde-derived
cyclopropene 1a was chosen as our first substrate as its crystalline
nature would enable product identification by X-ray analysis if
necessary.
aDepartment of Chemistry and Biochemistry, California State University,
Fullerton, CA, 92831, USA. E-mail: chyland@fullerton.edu; Tel: +1 (657)
278-2631
Scheme 2 Synthesis of cyclopropenylmethyl acetates.
bSchool of Chemistry, University of Tasmania, School of Chemistry,
University of Tasmania, Private Bag 75, Hobart TAS, 7001, Australia.
E-mail: chris.hyland@utas.edu.au; Fax: +61 3 6226 2858; Tel: + 61 3 6226
To our delight, treatment of p-nitrobenzaldehyde-derived cy-
clopropene 1a with 1.2 equivalents of TiCl4 at - 78 ◦C for
5 min resulted in the formation of (E)-chlorodiene 5a in 72% yield
(Table 1, entry 1). The stereochemistry of 5a was unambiguously
determined by X-ray crystallography.6 It is of note that the NMR
of the crude reaction mixture did not indicate the presence of
cDepartment of Chemistry, Faculty of Science, Central Tehran Branch,
Islamic Azad University, Shahrak Gharb, Tehran, Iran
† Electronic supplementary information (ESI) available: Experimental
section, X-ray crystallographic data and NMR spectra. CCDC reference
number 737881. For ESI and crystallographic data in CIF or other
electronic format see DOI: 10.1039/c0ob01046d
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3359–3363 | 3359
©

View Article Online
Fig. 1 The DFT energy profile for the reaction of cyclopropenylmethyl acetate 1b with TiCl4. The relative free energies and potential energies (in
parentheses) are given in kcal mol-1.9
any of the (Z)-isomer. Short reaction times were particularly
important as much lower yields were obtained if the reaction
◦
was left longer than 5–10 min at - 78 C. These lower yields are
presumably a result of product decomposition under the strongly
Lewis-acidic conditions. A deep purple colour was also noted
upon the addition of TiCl4, pointing towards the formation of a
carbocation intermediate.6
To test the substrate scope of the reaction, a range of cyclo-
propenylmethyl acetates with varying Ar groups were subjected
to the same reaction conditions and in all cases high yields
and complete (E)-selectivity was observed (Table 1, entries 1–7).
The only exception was the p-tolylaldehyde derived cyclopropene
5c, which gave the resulting diene in only 36% yield. We were
delighted to find that TiBr4 could also initiate the transformation
of cyclopropenes 1a–f to (E)-bromodienes 5h–k (entries 8–11,
Table 1).7
Scheme 3 Possible reaction pathways for interception of cations 3 and 4.
Given the four possible regioisomeric products outlined in
Scheme 3, it is remarkable that the reaction is both completely
regio- and stereoselective for (E)-halodienes 5. In order to explain
the high selectivity observed in the reaction we turned our
attention to computational studies (Fig. 1).
3360 | Org. Biomol. Chem., 2011, 9, 3359–3363
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 1 TiX4 mediated synthesis of halodienes 5
the more stable diene 8_Ph_Cl with an activation barrier of about
15.511 (3.3) kcal mol-1 (Fig. 1). Therefore, it can be seen that diene
8_Ph_Cl represents a thermodynamic well in the reaction and
is obtained as the only product due to the reversibility of allene
7_Ph_Cl formation at low temperature. Repeating the calculations
with TiBr4 resulted in a similar outcome. These calculations gave
a Gibbs free energy for 7_Ph_Br of - 21.3 kcal/mol, which is
about 7.3 kcal mol-1 less stable than 8_Ph_Br. Also, the energy
difference between 3_Ph/[TiBr4(OAc)]- and 7_Ph_Br/TiBr3(OAc)
is still small at 6.2 kcal mol-1, which again supports the reversibility
of the reaction 3_Ph + [TiBr4(OAc)]- → 7_Ph_Br + TiBr3(OAc).
To explain the stereochemical outcome of the reaction it is im-
portant to remember that the sterically encumbered [TiCl4(OAc)]-
is the species that delivers a chlorine to the allyl-vinyl cation
3_Ph. Delivery of chloride to the same face as the phenyl ring
is effectively blocked by an unfavorable steric interaction between
the in-plane Ph group and the incoming [TiCl4(OAc)]-. For this
unfavorable attack it was possible to locate the transition state
3TS_Ph_Cl, which is about 6 kcal mol-1 higher than for attack
trans to phenyl. This significant energy difference explains the
complete stereoselectivity observed experimentally. In support
of this proposal is the lowered stereoselectivity observed upon
treatment of dodecylaldehyde-derived cyclopropene 1g with TiCl4
(Scheme 4). This reduced stereoselectivity fits in with our proposal
above and can be ascribed to the lowered steric demand of an alkyl
group compared to an aryl group. This lowered steric demand
is reflected in the earlier transition state 3TS_Et_Cl: the Ti–Cl1
Entry
Substrate
TiX4
Ar
Product
Yield%a
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1f
1a
1b
1d
1e
1f
TiCl4
TiCl4
TiCl4
TiCl4
TiCl4
TiCl4
TiBr4
TiBr4
TiBr4
TiBr4
TiBr4
p-Nitrophenyl
Phenyl
p-Tolyl
5a
5b
5c
5d
5e
5f
5g
5h
5i
72
84
36
84
89
72
68
70
79
70
68
o-Chlorophenyl
o-Bromophenyl
p-Fluorophenyl
p-Nitrophenyl
Phenyl
o-Chlorophenyl
o-Bromophenyl
p-Fluorophenyl
9
10
11
5j
5k
a Isolated yield following chromatography.
We surmised that the first step in the reaction would be
coordination of TiCl4 to the acetate Ph. Based on this, our DFT
calculations predicted that the adduct 1_Ph_Cl is formed and
subsequently undergoes acetate abstraction via 1TS_Ph_Cl give
2_Ph and TiCl4(OAc)-. This relatively low energy transition state
(21.98 (8.5) kcal mol-1) represents the rate-determining step in
the reaction and provides an indication of why the reaction is so
facile at low temperature.9 Barrierless ring-opening to allyl-vinyl
cation 3_Ph then occurs through 2TS_Ph. This remarkably facile
transformation is due to the associated relief of ring-strain and as
such, allyl vinyl cation 3_Ph should be considered the only reactive
intermediate in the reaction. Indeed, there are examples of simple
˚
bond in 3TS_Et_Cl is ~0.041 A shorter than that in 3TS_Ph_Cl;
1
2
˚
the Cl –C bond in 3TS_Et_Cl is ~0.114 A longer than that in
3TS_Ph_Cl and the C1–C2–C3 bond angle in 3TS_Et_Cl is less
bent than in 3TS_Ph_Cl. This earlier nature of the transition
state 3TS_Et_Cl results in a ~4.2 kcal mol-1 stabilization on the
potential energy surface compared to 3TS_Ph_Cl. As a result, the
reaction affording (Z)-chlorodiene 5l becomes viable under the
reaction conditions and this leads to the lack of stereoselectivity
observed experimentally.
cyclopropenes undergoing ring-opening reactions by metal salts,
such as MgX2,10a Fe(acac)3 NaI10c and Au(I)10d–f
.
10b
At this point it is important to note that calculations showed
that dissociation of TiCl4(OAc)- to generate Cl- is an endergonic
process (7.3 kcal mol-1) with an energy barrier of approximately
17.5 kcal mol-1. As such, it is probable that it is TiCl4(OAc)- rather
than free Cl- that undergoes direct attack on 3_Ph. We can now
consider the regioselectivity of the reaction in order to explain why
the formation of diene 5 is favored over allene 7.
All attempts to locate the transition states for the formation
of 8_Ph_Cl and 7_Ph_Cl were unsuccessful and led to adducts
4_Ph_Cl and 5_Ph_Cl, respectively. This indicates that the po-
tential energy surface with respect to attack of TiCl4(OAc)- on
3_Ph should be very flat and that transformation into 8_Ph_Cl
and 7_Ph_Cl should occur with roughly equal propensity. This
result suggests that the regioselectivity of the reaction is not
determined by kinetic factors and we will demonstrate below
that the thermodynamic factors are responsible for the observed
regioselectivity of halogenation. From Fig. 1, we can see that both
the reactions 3_Ph + [TiCl4(OAc)]- → 8_Ph_Cl + TiCl3(OAc) and
3_Ph + [TiCl4(OAc)]- → 7_Ph_Cl + TiCl3(OAc) are thermody-
namically favourable, although the formation of diene 8_Ph_Cl
is more exothermic than formation of allene 7_Ph_Cl. The
Scheme 4 Reaction of dodecylaldehyde-derived 1g with TiCl4.
The [TiCl4(OAc)]-, which acts as the chloride shuttle in the
reaction has an approximately octahedral structure as illustrated
in Fig. 2. It is interesting to note that there are two chlorines
that can potentially be transferred to the cationic intermediate.
However, calculations revealed that a chlorine trans to chlorine is
more nucleophilic than the chlorine trans to oxygen.
For example, this more nucleophilic chloride preferentially
adopts 3TS_Et_Cl, which is about 2 kcal mol-1 lower in energy
than 3¢TS_Et_Cl (Fig. 2). The higher nucleophilicity of the
chlorine trans to chlorine is supported by a longer Ti–Cl bond
length and a greater partial negative charge on Cl of [TiCl4OAc]-.
This results in 3TS_Et_Cl being an earlier transition state than
3¢TS_Et_Cl; for 3TS_Et_Cl the angle C1–C2–C3 is larger and
the Cl–C2 bond length is longer (Fig. 2). We believe that these
low exergonicity for the reaction of 4_Ph + [TiCl4(OAc)]-
→
7_Ph_Cl + TiCl3(OAc) (9.9 kcal mol-1) suggests that the attack
at carbon 4 is reversible (Fig. 1). In such a case, TiCl3(OAc) can
readily abstract the chloride from allene 7_Ph_Cl and regenerate
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3359–3363 | 3361
©

View Article Online
Gibbs free energies in solvent (DG), entropy corrections were
calculated at the B3LYP/BS1 level and added to the solvent
potential energies. We have used the solvent energies throughout
the paper unless otherwise stated.
Materials and methods
Unless otherwise stated all reactions were carried out under an
argon or nitrogen atmosphere in flame-dried glassware. Reactions
were monitored by TLC using glass-backed silica gel XHL plates
(purchased from Sorbtech with UV 254, thickness 250 mm) or
by GC and GC/MS on an Agilent 7890 (FID detector and HP-
5 column 30 m ¥ 0.32 mm 0.25 mm, part number: 19091j-413)
and Varian 3900/Saturn 2100T (ion-trap mass-selective detector
and, FactorFour capillary column VF-5ms, 30 m ¥ 0.25 mm, Part
number: cp8944), respectively. Compounds were purified using
flash column chromatography with Sorbtech silica gel or by radial
R
chromatography with a Chromatotronꢀ. Chromatography plates
R
employed on the Chromatotronꢀ were 1 mm or 2 mm prepared
Fig.
2
B3LYP-optimized structures for 3TS_Ph_Cl, 3TS_Et_Cl,
3¢TS_Et_◦Cl and [TiCl4(OAc)]-. Selected structure parameters are given
from a mixture of calcium sulfate hemihydrate and Aldrich
Silica gel (TLC standard grade, without binder, with fluorescent
indicator). NMR spectra were recorded on a BrukerAvance DRX-
400 (400 MHz) in CDCl3 unless otherwise noted. CDCl3 was
-
˚
in A and . Muliken atomic charges for selected atoms of [TiCl4OAc] are
given in italics.
˚
observations could have potentially important implications for
the design of titanium based catalysts.
stored over K2CO3 or 4 A molecular sieves and was purchased
from Acros. Melting points are uncorrected and measure on
a Thomas Hoover Unimelt capillary melting point apparatus.
IR spectra were recorded on a Thermo electron corporation
Nicolet 380 FT-IR as films on NaCl plates (liquids) or in a NaCl
solution cells (solids). Accurate mass measurements were recorded
at the University of California, Riverside High Resolution Mass
Spectrometry Facility and at the University of California, Irvine.
X-ray analyses were carried out at the W.M. Keck Foundation
Center for Molecular Structure in California State University,
Fullerton. nBuLi was purchased from Strem and titrated against
diphenylacetic acid prior to use. Other reagents were commercially
available and used without further purification.
Conclusions
In conclusion, we have developed a stereoselective synthesis of (E)-
2-chloro-1,3-dienes and (E)-2-bromo-1,3-dienes and accounted
for the mechanism of the reaction using DFT calculations. These
calculations show that TiX4-mediated ring-opening is extremely
facile and that the regioselectivity of the subsequent halogenation
operates under thermodynamic control while the stereoselectivity
of halide addition is kinetically controlled. It should be noted
that dienes of this type have not received much attention due
to a surprising scarcity of methods for their preparation.12 Only
recently have Barluenga and coworkers reported an elegant
and simple synthesis of 2-chloro-1,3-dienes by selective Suzuki
coupling of 1,1-dichloroethylene.13 Work is currently underway on
further developing the reactivity of cyclopropenylmethyl acetate-
derived allyl-vinyl cations.
Typical procedure for preparation of (E)-chlorodienes
The cyclopropenyl acetate (0.17 mmol, 1.0 equiv.) was weighed
into a flame-dried flask purged with argon and CH2Cl2 (3.4 ml)
added. The solution was then cooled to -78 ◦C with a dry-
ice acetone bath and TiCl4 (0.17 ml of a 1.0 M solution in
CH2Cl2) added dropwise. After 5 min the reaction was quenched
by the addition of MeOH–H2O (5 ml of a 1 : 1 mixture) and
allowed to warm to room temperature. The organic layer was
separated and the aqueous extracted with CH2Cl2 (2 ¥ 5 ml). The
combined organics were dried (MgSO4) and evaporated to dryness.
Experimental
Computational details
Gaussian 0914 was used to fully optimize all the structures reported
in this paper at the B3LYP level of density functional theory.15
The 6-311G(d) basis set was chosen to describe titanium.16 The 6-
31G(d) basis set was used for other atoms. This basis set combina-
tion will be referred to as BS1. Frequency calculations were carried
out at the same level of theory as for structural optimization.
To further refine the energies obtained from the B3LYP/BS1
calculations, we carried out single point energy calculations for
all the structures with the larger 6-311+G(2d,p) basis set (BS2).
The solvation energies were calculated using BS2 on gas phase
optimized geometries with the CPCM solvation model17 using
dichloromethane as a solvent. To estimate the corresponding
R
Purification on SiO2 (pentane on a Chromatotronꢀ) yielded the
pure diene 3e.
(E)-1-chloro-2-(2-chloro-3,4-dimethylpenta-1,3-dienyl)benzene
3e was obtained as a clear yellow oil (72 mg, 84%). IR 1650,
1623, 1468, 1440, 1374, 1148, 1129, 1052, 1035, 946, 929, 806, 749
cm-1.1H NMR (400 MHz, CDCl3) d 7.38–7.35 (m, 1H), 7.26–7.24
(m, 1H), 7.25–7.11 (m, 2H), 6.90 (s, 1H), 1.91–1.90 (m, 3H), 1.65
(d, J = 8.0, 1.0 Hz, 3H), 1.52–1.51 (m, 3H). 13C NMR (100 MHz,
CDCl3) d 138.4, 134.1, 133.6, 133.1, 128.7, 128.4, 126.4, 125.1,
125.0, 115.5, 21.6, 20.0, 17.0. HRMS-CI (m/z): (M + H)+ calcd
for C13H14Cl2, 240.0473; found 240.0466.
3362 | Org. Biomol. Chem., 2011, 9, 3359–3363
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
7 Substrates where the ring gem-dimethyl substituents are replaced with
hydrogens have so far proved unstable to the reaction conditions.
8 The real activation energy is expected to be significantly lower than
calculated because the entropic contributions to the Gibbs free energies
in the calculations are overestimated: (a) H. Tamura, H. Yamasaki,
H. Sato and S. Sakaki, J. Am. Chem. Soc., 2003, 125, 16114; (b) M.
Sumimoto, N. Iwane, T. Takahama and S. Sakaki, J. Am. Chem. Soc.,
2004, 126, 10457; (c) S. Sakaki, T. Takayama, M. Sumimoto and M.
Sugimoto, J. Am. Chem. Soc., 2004, 126, 3332; (d) A. A. C. Braga, G.
Ujaque and F. Maseras, Organometallics, 2006, 25, 3647.
9 The activation barrier for the (OAc abstraction using
MP2/BS2//B3LYP/BS1 and including the solvent effect is calculated
as 14.9 (1.6) kcal mol-1, which further supports the ease of this process.
10 (a) Y. Wang and H. W. Lam, J. Org. Chem., 2009, 74, 1353; (b) Y. Wang,
E. A. F. Fordyce, F. Y. Chen and H. W. Lam, Angew. Chem., Int. Ed.,
2008, 47, 7350; (c) S. M. Ma, J. L. Zhang and Y. J. Cai, J. Am. Chem.
Soc., 2003, 125, 12954; (d) J. T. Bauer, M. S. Hadfield and A-L. Lee,
Chem. Commun., 2008, 6405; (e) Z. B. Zhu and M. Shi, Chem.–Eur. J.,
2008, 14, 10219; (f) C. K. Li, T. Zeng and J. B. Wang, Tetrahedron Lett.,
2009, 50, 2956.
Acknowledgements
G. Gallego was supported by the Cal State Fullerton
MARC U*STAR Program grants 5T34GM008612-12 and
5T34GM008612-13. Acknowledgement is made to the Donors
of the American Chemical Society Petroleum Research Fund for
partial support of this research. Financial support by California
State University, Fullerton is also gratefully acknowledged. The
acquisition of an NMR spectrometer was funded by NSF CHE-
0521665. We also thank the Australian Research Council for
funding, and the National Computational Infrastructure (NCI)
and the Tasmanian Partnership for Advanced Computing (TPAC)
for provision of computing.
Notes and references
11 The energy difference between 7_Ph_Cl and 4_Ph_Cl using MP2 theory
is computed as 11.1 (- 1.1) kcal mol-1, which further supports the
reversibility of the reaction.
12 For previous syntheses of 2-halo-1,3-butadienes, see: (a) J. Pornet,
Tetrahedron Lett., 1981, 22, 453; (b) S. Ma and G. Wang, Tetrahedron
Lett., 2002, 43, 5723; (c) A. Horva´t and J-E. Ba¨ckvall, J. Org. Chem.,
2001, 66, 8120.
1 For a reviews on propargyl substitution, see: (a) K. M. Nicholas, Acc.
Chem. Res., 1987, 20, 207; (b) J. Detz, H. Hiemstra and J. H. van
Maarseveen, Eur. J. Org. Chem., 2009, 6263.
2 Relatively little work on cyclopropenylmethyl acetates has been re-
ported. For example, see: S. Simaan, A. Masarwa, E. Zohar, A.
Stranger, P. Bertus and I. Marek, Chem.–Eur. J., 2009, 15, 8449.
3 E. Seraya, E. Slack, A. Ariafard, B. Yates and C. J. T. Hyland, Org.
Lett., 2010, 12, 4768.
4 There are numerous examples of nucleophilic chloride originating from
TiCl4 intercepting cations. For select examples, see: M.-L Yao, T. R.
Quick, Z. Wu, M. P. Quinn and G. W. Kabalka, Org. Lett., 2009, 11,
2647; R. Mahrwald, S. Quint and S. Scholtis, Tetrahedron, 2002, 58,
9847.
5 The synthesis of the cyclopropenylmethyl acetates used in this work is
detailed in the ESI† using the method of Baird and coworkers: M. S.
Baird, H. H. Hussain and W. Nethercott, J. Chem. Soc., Perkin Trans.
1, 1986, 1, 1845.
13 J. Barluenga, P. Moriel, F. Aznar and C. Valde´s, Adv. Synth. Catal.,
2006, 348, 347.
14 M. J. Frisch et al. Gaussian 09, revision A.02; Gaussian, Inc.:
Wallingford, CT, 2009.
15 (a) C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785;
(b) B. Miehlich, A. Savin, H. Stoll and H. Preuss, Chem. Phys. Lett.,
1989, 157, 200; (c) A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
16 The 6-311G(d) and 6-31G(d) basis sets have been shown to perform
well for titanium complexes, see: (a) L. T. Dulatas, S. N. Brown, E.
Ojomo, B. C. Noll, M. J. Cavo, P. B. Holt and M. M. Wopperer, Inorg.
Chem., 2009, 48, 46556; (b) S. N. Brown, E. T. Chu, M. W. Hull and B.
C. Noll, J. Am. Chem. Soc., 2005, 127, 16010; (c) Y. Gong, M. Zhou, S.
X. Tian and J. Yang, J. Phys. Chem. A, 2007, 111, 6127.
6 James Qiang Zhang and Katherine Kantardjieff, CmolS, California
State University, Fullerton are gratefully acknowledged for structure
determination of 5a (CCDC 737881).
17 V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102, 1995.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3359–3363 | 3363
©