A study of the reaction of n-BuLi with Ti(Oi-Pr)4 as a method to generate titanacyclopropane and titanacyclopropene species

Article abstract of DOI:10.1016/j.tet.2013.12.048

The use of the combination of reagents Ti(Oi-Pr)₄/n-BuLi, introduced by the group of J.J. Eisch in 2001, has only found a few applications so far, with sometimes conflicting observations. This article describes a study aimed at clarifying the nature, the stability and the reactivity of the active organometallic species involved. Reactions with CO₂ and other trapping reagents reveal that it is generated within a few minutes at 0 C in THF, where it can be considered to be stable for 30 min. Most of our results are consistent with the expected titanacyclopropane nature of this reagent but some observations suggest that the chemistry at play may be more complicated.

Full text of DOI:10.1016/j.tet.2013.12.048

Accepted Manuscript
A study of the reaction of nBuLi with Ti(OiPr) as a method to generate
4
titanacyclopropane and titanacyclopropene species
Valentin Rassadin, Yvan Six
PII:
S0040-4020(13)01916-9
10.1016/j.tet.2013.12.048
TET 25132
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 8 November 2013
Revised Date: 11 December 2013
Accepted Date: 17 December 2013
Please cite this article as: Rassadin V, Six Y, A study of the reaction of nBuLi with Ti(OiPr) as a method
4
to generate titanacyclopropane and titanacyclopropene species, Tetrahedron (2014), doi: 10.1016/
j.tet.2013.12.048.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
A study of the reaction of nBuLi with Ti(OiPr)4 as a method to
generate titanacyclopropane and titanacyclopropene species
Valentin Rassadin, Yvan Six*
Laboratoire de Synthèse Organique DCSO, UMR 7652 CNRS/Ecole Polytechnique, 91128 Palaiseau Cedex
(France). Fax: (+33) (0)1 6933 5972 ; E-mail: yvan.six@polytechnique.edu
Abstract
The use of the combination of reagents Ti(OiPr)4 / nBuLi, introduced by the group of J. J.
Eisch in 2001, has only found a few applications so far, with sometimes conflicting
observations. This article describes a study aimed at clarifying the nature, the stability and the
reactivity of the active organometallic species involved. Reactions with CO2 and other
trapping reagents reveal that it is generated within a few min at 0 °C in THF, where it can be
considered to be stable for 30 min. Most of our results are consistent with the expected
titanacyclopropane nature of this reagent but some observations suggest that the chemistry at
play may be more complicated.
Keywords: organolithium reagents, titanium, metallacycles, carbon dioxide, Kulinkovich
reaction
Introduction
In 1989, O. G. Kulinkovich et al. reported a new general synthetic method converting
carboxylic esters into cyclopropanols in the presence of Ti(OiPr)4 and several equivalents of a
Grignard reagent.[1,2] Over the years, these initial findings have led to the development of a
whole family of reactions relying on the use of this combination of reagents.[3,4,5,6] The
various transformations that can be carried out under these conditions can be explained by the
formation of titanacyclopropane complexes 1, as initially proposed by Kulinkovich (Scheme
1). These highly reactive complexes 1 are presumably formed after transmetallation reactions
that take place to produce the intermediates 2 and 3 successively. The putative species 1 are
unstable and are therefore almost always generated in the presence of the substrate of the
reaction (e.g. ester, amide, nitrile, alkyne…), by which they are rapidly trapped. Nonetheless,
− 1 −

ACCEPTED MANUSCRIPT
there are a few examples of preparations of solutions of titanacyclopropanes 1 from Grignard
reagents and XTi(OiPr)3 (X = Cl, OiPr), at low temperature (< −30 °C).[7,8]
iPrO
OiPr
M
Ti
OiPr
iPrO
OiPr
R
Ti
iPrO OiPr
iPrOM
R
2
iPrO
OiPr
M
iPrO
OiPr
Ti
R
fast
Ti
iPrOM
R
R
R
3
1
H
R
Scheme 1. Formation of a titanacyclopropane complex 1 from Ti(OiPr)4 and an
organometallic reagent RCH2CH2M used in excess (typically, M = MgCl or MgBr).
In 2001, the group of J. J. Eisch introduced the use of nBuLi in place of the Grignard reagent,
describing the generation of “Ti(OiPr)2”, assumed to be in equilibrium with 1 (R = Et).[9,10]
The experimental procedure involves slow addition of nBuLi (2.0 equiv) at −78 °C to a THF
solution of Ti(OiPr)4 (1.0 equiv) and then stirring overnight with the temperature being
allowed to raise to 25 °C. The published results are consistent with the reliable formation of
around 0.5 equiv of “Ti(OiPr)2” / 1 (R = Et), as evidenced by reactions with nitriles, ketones,
alkenes or alkynes. The conditions used for the synthesis of the reagent and its stability at
room temperature are rather amazing, since using Grignard reagents, 1 is generated within
minutes even at low temperature and is highly unstable, as already mentioned.
It may seem surprising that more than ten years later, only very few applications of such new
chemistry based on nBuLi have appeared in the literature. To the best of our knowledge, they
can be extensively listed as follows: i) Ni-catalysed coupling of aryl iodides with
titanacyclopropene complexes generated from alkynes (group of Y. Tsuji);[11] ii)
intramolecular coupling of carboxylic imides with alkenes to produce cyclic hemiaminals
(group of J. K. Cha);[12] iii) intramolecular coupling of silyl-substituted alkynes with alkenes
(group of A. J. Phillips);[13] iv) generation of azatitanacyclopropane species from imines,
which are then coupled with allylic alkoxides (group of G. C. Micalizio);[14] v) one example
of intermolecular coupling of a lactone with an olefin to produce a ketone (group of A. J.
Phillips).[15]
These extremely interesting pieces of work demonstrate the reality of the potential of the
Ti(OiPr)4 / nBuLi reagent system. However, many of the observations reported in the
corresponding articles are inconsistent with the original claims. First, it is worthy of note that
none of these applications actually use the procedure described further above. In fact, THF or
− 2 −

ACCEPTED MANUSCRIPT
Et2O solutions containing Ti(OiPr)4 and nBuLi are never reported to be kept for more than 30
min at room temperature before another reagent is added. Cha et al. even noted that their
yields dropped when the reagent had been prepared more than 45 min before the substrate was
added.[12]
In order to clarify our knowledge of the chemistry of the organometallic species involved, we
have decided to undertake a careful and rational study, with the hope that it would be useful
for the development of new applications.
Results and discussion
Our starting hypothesis was that the active species was indeed in the form of a
titanacyclopropane 1 (R = Et), possibly as an “ate” complex with lithium isopropoxide,
produced as displayed in Scheme 1 (M = Li). The generation of 1 would not only be
analogous to the reaction of a Grignard reagent with Ti(OiPr)4,[6] but also to the reaction of
nBuLi with Cp2ZrCl2 to generate the Negishi reagent.[16] With a view to provide new evidence
for the 1,2-dimetallated nature of the key intermediate formed, we decided to carry out
experiments involving sequential trapping with carbon dioxide and iodine, as we had
successfully done, a few years ago, in the case of the reagent combination Ti(OiPr)4 /
cC5H9MgCl.[8c] When 2.0 equiv of nBuLi were added over 4 to 5 min to a THF solution of 1.0
equiv of Ti(OiPr)4 at 0 °C, and the resulting mixture was stirred for 4 min at the same
temperature, trapping with CO2 and then I2 led to the formation of around 0.50 equiv of the β-
iodoacids 4a and 4b accompanied by a smaller amount of the acids 5a and 5b (Scheme 2).
This result is consistent with the formation of at least 0.50 equiv of the titanacyclopropane 1
(R = Et) and with its reaction with CO2 to give the corresponding titanalactone intermediates
6a and 6b.
− 3 −

ACCEPTED MANUSCRIPT
Scheme 2. Trapping of the organotitanium complex generated from 1 equiv of Ti(OiPr)4 and
2 equiv of nBuLi by CO2 and then I2.
Several explanations could be put forward to account for the formation of the non-
halogenated acids 5a and 5b. The most obvious would be protonolysis caused by moisture in
the CO2 gas used.[17] 6b would certainly react with small amounts of water faster than 6a and
this would account for the higher ratio of the iodinated regioisomers 4a/4b compared to
5a/5b. Alternatively, alkylation of 6a or 6b by residual n-butyllithium would generate a
dialkylated titanium species that could then, similarly to the intermediate 3, lead to 5a/5b and
one molecule of 1 (R = Et). Again, 6b could react faster than 6a in such a process. On the
basis of these hypotheses and of our previous work,[8c] the influence of the relative amounts of
nBuLi used was next studied (Table 1).
Table 1. Trapping, by CO2 and then I2, of the organotitanium species generated from 1 equiv
of Ti(OiPr)4 and various amounts of nBuLi.
Entry
Yield (4a+4b)a
Ratio 4a/4b
73 : 27
78 : 22
80 : 20
76 : 24
Amount (5a+5b)
0.19 equiv
0.10 equiv
0.06 equiv
0.04 equiv
Ratio 5a/5b
63 : 37
60 : 40
63 : 37
60 : 40
n
1
2
3
4
0.29 equiv (29%)
0.48 equiv (48%)
0.47 equiv (59%)
0.35 equiv (71%)
1.5
1.0
0.80
0.50
a Yield calculated on the basis of the theoretical formation of n (n ≤ 1.0) or 1.0 (n ≥ 1.0) equiv
of titanacyclopropane 1 (R = Et).
− 4 −

ACCEPTED MANUSCRIPT
Overall, little variation was observed with respect of the regioselectivies 4a/4b or 5a/5b. The
use of 2.5 equiv of nBuLi, rather than 2.0 equiv, clearly proved to be detrimental, resulting in
a lower yield of the β-iodoacids 4a and 4b and a larger amount of the singly functionalised
compounds 5a and 5b (entry 1 vs entry 2). Conversely, when the quantity of nBuLi
introduced in the reaction system was reduced from 2.0 to 1.8 equiv, the observed amounts of
5a and 5b decreased significantly, without affecting the amounts of the products 4a and 4b
(entry 3 vs entry 2). Using even less nBuLi (1.5 equiv), the by-products 5a and 5b were
further reduced (entry 4). However, this improvement was achieved at the expense of the
quantity of desired products 4a and 4b. It can nonetheless still be considered as advantageous,
giving the best proportion of iodinated acids relative to the simple non-iodinated acid by-
products. Therefore, these last two conditions (1.8 equiv or 1.5 equiv of nBuLi for 1.0 equiv
of Ti(OiPr)4) were selected for the rest of our studies. We believe they ensure a fast formation
of 1 (R = Et) by preventing the concentration of the n-butyltitanium intermediate 2 (R = Et) to
drop close to zero during the addition of nBuLi. They are also likely to minimise the
occurrence of deleterious reactions between 1 (R = Et) and nBuLi. There is also a possibility
that the monoalkylated complex 2 (R = Et), formed along with 1 (R = Et), could react
reversibly with the latter, as displayed in Scheme 3. Such a process could extend the lifetime
of 1 (R = Et) by “storing” it in the form of the proposed structures 7 or 8. It is also worth
pointing out that these hypothetical species 7 and 8, if they exist, might well react themselves
with CO2 and iodine to give the observed iodinated acid products.
nBuLi
(1 equiv)
iPrO
OiPr
Ti
OiPr
iPrO
OiPr
Ti
+ iPrOLi
iPrO OiPr
Et
2 (R = Et)
1 equiv
1 equiv
nBuLi
(0.5 equiv)
iPrO
OiPr
iPrO
OiPr
Ti
OiPr
Ti
+
+ iPrOLi
Et
Et
2 (R = Et)
1 (R = Et)
0.5 equiv
0.5 equiv
1.5 equiv
?
OiPr
OiPr
Et
Et
iPrO
OiPr
(iPrO)3Ti
Ti
Ti
Ti(OiPr)3
or
Et
Et
7
8
0.5 equiv
Scheme 3. Reaction of 1 equiv of Ti(OiPr)4 and 1.5 equiv of nBuLi to generate 1 (R = Et) and
2 (R = Et). Proposed reversible reaction with 7 and/or 8.
− 5 −

ACCEPTED MANUSCRIPT
A short study of the influence of the nature of the solvent next demonstrated that THF is
superior to Et2O, tBuOMe or cC5H9OMe for the production of 4, with the best yield and with
a regioselectivity lying among the best observed (Table 2).[18]
Table 2. Effect of the nature of the solvent on the generation and trapping of the putative
species 1 (R = Et).
Entry
Solvent
THF
Et2O
tBuOMe
cC5H9OMe
Yield (4a+4b)a
Ratio 4a/4b
80 : 20
68 : 32
84 : 16
72 : 28
1
2
3
4
0.47 equiv (59%)
0.37 equiv (46%)
0.31 equiv (39%)
0.30 equiv (38%)
a Yield calculated on the basis of the theoretical formation of 0.80 equiv of titanacyclopropane
1 (R = Et).
In order to assess the stability of the complex formed, a few additional experiments were
carried out, in which CO2 was introduced after different time intervals following the end of
the addition of nBuLi (Table 3). This study demonstrated that the putative species 1 (R = Et)
generated as described above is reasonably stable in THF, at 0 °C, over 20 min, but cannot be
kept at that temperature for 2 h without extensive decomposition (entries 1−3 and 5−6). This
is in sharp contradiction with Eisch’s original reports but in good agreement with Cha’s
observations.[12] Using 1.5 equiv of nBuLi rather than 1.8 equiv, the lifetime of 1 (R = Et) is
somewhat increased (entry 6 vs entry 3). This positive effect is lost when the solvent is
changed from THF to Et2O (entry 8 vs entry 6). Complementary experiments showed that
adding nBuLi at −50 °C, rather than at 0 °C, to the Ti(OiPr)4 solution in THF does not grant
any particular advantage (entry 4 vs entry 1 and entry 7 vs entry 5).
− 6 −

ACCEPTED MANUSCRIPT
Table 3. Assessment of the stability of the putative species 1 (R = Et).
Entry
n
Conditions
Yield (4a+4b)a
Ratio 4a/4b
80 : 20
75 : 25
67 : 33
81 : 19
1
2
3
4
0.80 THF, 0 °C, 4 min
0.80 THF, 0 °C, 23 min
0.80 THF, 0 °C, 2 h
0.47 equiv (59%)
0.44 equiv (56%)
0.080 equiv (10%)
0.45 equiv (56%)
0.80 THF, − 50 to 0 °C, 10 min
5
6
7
0.50 THF, 0 °C, 4 min
0.50 THF, 0 °C, 2 h
0.50 THF, − 50 to 0 °C, 10 min
0.35 equiv (71%)
0.085 equiv (17%)
0.36 equiv (72%)
76 : 24
75 : 25
82 : 18
8
0.50 Et2O, 0 °C, 2 h
0.050 equiv (10%)
70 : 30
a Yield calculated on the basis of the theoretical formation of n mmol of titanacyclopropane 1
(R = Et).
The proposed titanalactone structures 6a and 6b are further supported by trapping with N-
bromosuccinimide (NBS), oxygen gas and deuterium oxide, which delivered the
corresponding β-bromoacids 9a and 9b, β-hydroxyacids 10a and 10b, and β-deuteriated acids
5a-d and 5b-d respectively (Scheme 4).
1. CO2
2. NBS
3. H3O+
Br
CO2H
Br
CO2H
85 : 15
+
Et
Et
54%
9a
9b
1. CO2
2. O2
3. Na2S2O3
OH
CO2H
H3O+
iPrO
OiPr
Ti
CO2H
OH
+
Et
Et
Et
Et
60%
Et
1a
10a
84 : 16
10b
1. CO2
2. D2O
3. H3O+
>85% d
>81% d
D
CO2H
CO2H
83 : 17
D
+
72%
5a d
5b d
− 7 −

ACCEPTED MANUSCRIPT
Scheme 4. Trapping of the putative species 1 (R = Et) generated in THF, at 0 °C, from 1
equiv of Ti(OiPr)4 and 1.5 equiv of nBuLi, with CO2 and then NBS, O2 or D2O (yields and
ratios estimated by NMR).
We next examined the engagement of sBuLi instead of nBuLi. Although in principle, with
this organolithium reagent, both titanacyclopropane complexes 1 (R = Et) and 11 could be
generated, only the reaction of the former was evidenced after trapping with CO2 followed by
I2. The formation of the iodoacid 4c was not detected (Table 4). Intriguingly, the ratio of the
two regioisomers 4a and 4b was markedly higher under these conditions than when nBuLi
was employed [entries 1−2 (95 : 5) vs Table 1, entry 4 (76 : 24)]. This pleads against the
hypothesis of the same common intermediate 1 (R = Et) and in favour of the formation of
complexes such as 7 and/or 8 (Scheme 3).[19] Another experiment conducted with tBuLi and
Ti(OiPr)4 gave a single detectable regioisomer, albeit in 17% yield only.[20] With sBuLi, the
use of Ti(OnPr)4 was also investigated. A rather high 4a/4b ratio was observed again (90 :
10), but the yield was poor (entry 3). This was also the case with Ti(OnPr)4 and nBuLi: 19%
yield (85 : 15 selectivity).
Table 4. Trapping by CO2 and then by I2 of the organotitanium species generated from 1.0
equiv of TiX4 (X = OiPr or OnPr) and 1.5 equiv of sBuLi.
Entry TiX4
Yield (4a+4b+4c)a Ratio 4a/4b/4c Amount (5a+5b) Ratio 5a/5b
1
Ti(OiPr)4 0.26 equiv (52%)
Ti(OiPr)4 0.24 equiv (48%)
Ti(OnPr)4 0.065 equiv (13%) 90 : 10 : 0
95 : 05 : 0
94 : 06 : 0
0.07 equiv
0.03 equiv
0.05 equiv
70 : 30
c
2b
3
−
c
−
a Yield calculated on the basis of the theoretical formation of 0.5 equiv of titanacyclopropanes
1 (R = Et) and 11. Reaction repeated with another (new) bottle of sBuLi. Could not be
b
c
determined unambiguously.
Having identified simple and reliable conditions for the generation of an active reagent from
Ti(OiPr)4 and nBuLi, with similar reactivity properties to those that would be expected with 1
(R = Et), we next turned our attention to the transformations of alkynes. Under essentially the
same conditions as described above, alkynes could be converted into the corresponding (Z)-
alkenes, although a longer reaction time, typically 2 h, was necessary to achieve good
conversion (Scheme 5).[21] This is consistent with the formation of a titanacyclopropene
− 8 −

ACCEPTED MANUSCRIPT
species 12, as originally proposed by Eisch et al.[9b] and in analogy with the results obtained
with Grignard reagents.[3,6]
Scheme 5. Generation of putative titanacyclopropene species 12 from alkynes, in THF, at 0
°C, using Ti(OiPr)4 and nBuLi (isolated yields).
Interestingly, starting from 1-phenylpent-1-yne, when the reaction time was extended to 4 h,
the yield changed only slightly, dropping from around 80 to 75%. This suggests that the
intermediate 12 is considerably more stable than 1 (R = Et), in agreement with the article
published by Tsuji et al.[11] To get further insight on the formation of 12, kinetic studies were
performed, in Et2O and in THF (Figure 1). The results clearly show that in both cases, the
formation of 12 is much slower than the generation of 1 (R = Et), which is virtually
instantaneous at the same temperature, 0 °C, as described further above. Moreover, the
complex 12 is formed significantly faster in Et2O than in THF but it also decomposes to a
greater extent over a period of 4 h.
− 9 −

ACCEPTED MANUSCRIPT
Figure 1. Kinetic plots of the reactions of 1-phenylpent-1-yne with 2.2 equiv of Ti(OiPr)4 and
3.3. equiv of nBuLi, at 0 °C, in Et2O and in THF.
Attempted trapping of complex 12 with carbon dioxide or p-anisaldehyde gave only poor
results (Scheme 6), which is in marked contrast with the known reactivity of the analogous
species generated using Grignard reagents.[3a,8a−b]
Scheme 6. Attempted trapping reactions of a putative titanacyclopropene complex, generated
from 1-phenylpent-1-yne, either with CO2 or p-anisaldhyde. In all cases, the major product
observed was (Z)−1-phenylpent-1-ene.
Finally, the reactivities of alkenes, amides and esters were briefly studied. Ligand exchange of
1 (R = Et) with 4-methylstyrene 13 followed by trapping with CO2 gave the corresponding
hydrocarboxylated product 14 in good yield, provided a large excess of 1 (R = Et) was
generated (Scheme 7, top).[22] Indeed, with 1.25 equiv of Ti(OiPr)4 and 2.25 equiv of nBuLi,
the yield dropped to 22%. This is due to incomplete conversion and partial polymerisation of
13. Starting from 4-benzyloxybut-1-ene 15, letting the reaction mixture warm to room
temperature and stirring it for 12 h eventually yielded the ethers 16 and 17 (Scheme 7,
bottom). While the formation of the latter can be explained by 1,2-alkene insertion of 15 into
1 (R = Et) and hydrolysis upon work-up, the presence of 16 can be accounted for by invoking
a ligand exchange of 1 (R = Et) with 15, followed by 1,2-insertion of the same molecule.
Reactions of this kind have been observed before, using Ti(OiPr)4 and Grignard reagents, and
the assigned (R*,R*) relative configurations of 16 and 17 are based on these literature
results.[7a,23] An intramolecular Kulinkovich-de Meijere-type reaction[4] was observed starting
from the N-alkenylamide 18 (Scheme 8, top), proceeding in much lower yield than under the
“classic” conditions.[24] Finally, using the ester 19, the cyclopropanol 20 resulting from a
Kulinkovich-type reaction was formed in low yield (Scheme 8, bottom),[25] in agreement with
earlier observations by Cha et al.[26] The ketone 21 was produced as well, which is
reminiscent of a result reported by Phillips et al., who also synthesised a ketone under similar
conditions and achieved a good yield using a large excess of reagents.[15]
− 10 −

ACCEPTED MANUSCRIPT
(1.0 equiv)
13
0 °C, 20 min
1.
nBuLi (9.0 equiv)
THF, 0 °C
5 min
2. CO2, 0 °C, 25 min
3. H3O+
CO2H
1 (R = Et)
Ti(OiPr)4
14
75%
(6.0 equiv)
1. Ti(OiPr)4 (1.25 equiv)
nBuLi (2.5 equiv)
THF, 0 °C, 60 min
then 20 °C, 12 h
2. H3O+
BnO
BnO
+
BnO
BnO
56%
15
16
77 : 23
17
Scheme 7. Examples of reactions of alkenes using the Ti(OiPr)4 / nBuLi reagent system.
1. Ti(OiPr)4 (1.5 equiv)
nBuLi (3.0 equiv)
THF, 0 °C, 15 min
then 20 °C, 35 min
2. H2O
O
N
N
Bn
Bn
14%
18
MeO2C
OMe
OMe
1.
nBuLi (3.0 equiv)
THF, 0 °C
5 min
(1.0 equiv)
0 to 20 °C
2. H3O+
1 (R = Et)
19
Ti(OiPr)4
(2.0 equiv)
OH
O
OMe
OMe
OMe
OMe
Et
+ Et
14%
13%
20
21
Scheme 8. Examples of reactions of an amide and an ester using the Ti(OiPr)4 / nBuLi
reagent system.
Conclusions
To conclude, we have demonstrated that the formation, from Ti(OiPr)4 and nBuLi, of a
reactive complex with properties similar to those expected for the titanacyclopropane 1 (R =
Et), can be accomplished within a few min in THF at 0 °C, in a highly reliable fashion, as
evidenced by the trapping reactions with CO2 and iodine. The use of lower temperatures is not
necessary and, in our hands, does not provide any advantage. The organotitanium species thus
generated is fairly stable at 0 °C but cannot be kept for more than 30 min without significant
− 11 −

ACCEPTED MANUSCRIPT
decomposition. Although most of our results are in agreement with a titanacyclopropane
structure such as 1 (R = Et), the outcome of the reactions with sBuLi are not fully compatible
with this picture, suggesting that the organotitanium species reacting with CO2 might be other
complexes, perhaps bis-metallated intermediates such as 7 and/or 8.
In the presence of an alkyne, a ligand exchange takes place at 0 °C to generate a putative
titanacyclopropene species 12. This process is slower than the formation of 1 (R = Et) from
Ti(OiPr)4 and nBuLi.[27] The complex 12 is fairly stable in THF at 0 °C, at least for several
hours.
Overall, these reactions are analogous to those observed using Ti(OiPr)4 and Grignard
reagents, the main differences being dramatically enhanced stability and modified, generally
lower, reactivity. These properties could be attributable to a strong coordination of the
titanium centre with lithium isopropopxide generated as a side product, as suggested by Eisch
et al.[10a,28] The effect of the formation of such an “ate” complex would be to diminish
reactivity partly (?) because of the steric requirements of the ligands around the metal and of
reduced Lewis acidity.
Acknowledgements
We are grateful to Michel Levart for the MS analyses of our compounds and to the École
Polytechnique for financial support, in particular for the research grant attributed to V. R.
Funding from the CNRS is also acknowledged. Y. S. especially wishes to thank Prof. S. Z.
Zard for his continuous encouragement and valuable support.
Supporting information
Detail of the method of analysis of the crude products by NMR. NMR data for compounds 4a,
4b, 5a, 5a-d, 5b, 5b-d, 9a, 9b, 10a, 10b, 14, 16, 17, 20, 21 and the alkene products of Scheme
5. Some additional experimental details and complementary results.
− 12 −

ACCEPTED MANUSCRIPT
Experimental part
General considerations.
Titanium(IV) isopropoxide (VERTEC® TIPT) was purchased from Alfa Aesar, distilled under
reduced pressure and stored under nitrogen for several months. Other commercial reagents
were used as received, without purification. n-Butyllithium solution in hexanes was purchased
from Alfa Aesar, Sigma-Aldrich or Acros and titrated either with diphenylacetic acid[29] or
with menthol in the presence of 1,10-phenanthroline.[30] Tetrahydrofuran and diethyl ether
were purified using a MB SPS-800 solvent purification system (MBRAUN). tert-Butyl
methyl ether and cyclopentyl methyl ether were purchased from Acros or Alfa Aesar and used
as received. Petroleum ether refers to the 40−60 °C fraction. It was purchased from VWR
(Normapur® grade) and used without purification. The syntheses and characterisations of
compounds 15, 18 and the bicyclic aminocyclopropane resulting from the intramolecular
Kulinkovich-de Meijere reaction of 18 are reported elsewhere.[8b,31,32] The methyl ester 19 was
made by esterification of the corresponding commercially available carboxylic acid (cat.
pTSA, MeOH, reflux, 16 h). All the reactions were carried out under a stream of nitrogen or
argon, unless specified otherwise. The temperatures mentioned are measured with a
thermometer introduced into the reaction medium. Concentration under reduced pressure was
carried out using rotary evaporators at 40 °C.
Flash column chromatography was performed on Merck silica gel 60 (40–63 ꢀm). NMR
spectra were recorded with AM 400 and AVANCE 400 Bruker spectrometers (1H at 400
MHz, 13C at 100.6 MHz. Chemical shifts δ are given in ppm, referenced to the peak of
tetramethylsilane, defined at δ = 0.00 (1H NMR), or the solvent peak of CDCl3, defined at δ =
77.0 (13C NMR). Multiplicities are abbreviated as follows: s = singlet, d = doublet, t = triplet,
q = quartet, quint = quintet, sext = sextet, sept = septet, m = multiplet, br = broad. Coupling
constants J are given in Hz.
Generation of 1 (R = Et); trapping with CO2 and then by I2, typical procedure.
nBuLi (2.2 M in hexanes, 1.5 equiv, 3.0 mmol, 1.4 mL) was slowly added to a solution of
Ti(OiPr)4 (1.0 equiv, 2.0 mmol, 0.59 mL) in THF (10 mL), at 0 °C, over 6 min. After 4
additional min at the same temperature, CO2 was bubbled into the solution during 20 min
while the temperature was allowed to increase to room temperature. A solution of iodine (1.5
equiv, 3.0 mmol, 0.76 g) in THF (3.0 mL) was then added. After 2 h of stirring, 10% HCl
aqueous solution (10 mL) was added, followed by saturated Na2S2O3 aqueous solution (3.0
mL). The organic phase was separated and the aqueous layer was extracted with Et2O (3 × 10
mL). The combined organic phases were dried over anhydrous MgSO4, filtered and
− 13 −

ACCEPTED MANUSCRIPT
concentrated in vacuo (P ≥ 25 mbar) to give the crude product. Analysis by NMR (after
addition of 0.25 mmol of 4-(4-methoxyphenyl)-4-oxobutanoic acid used as an internal
standard)[20] revealed that it contained
a
mixture of 3-iodopentanoic and 2-
(iodomethyl)butanoic acids 4a and 4b in an 76 : 24 ratio and in an estimated combined
amount of 0.71 mmol [71% yield based on the theoretical formation of 1.0 mmol of 1 (R =
Et)]. The amount of non-iodinated acids (pentanoic and 2-methylbutanoic acids 5a and 5b)
was estimated to be 0.09 mmol (ratio ≈ 60 : 40).
Note: the stream of CO2 introduced into the reaction flask should be gentle. It may be bubbled
into or passed over the solution. The application of a strong flow of CO2 gives poorer results.
The use of dry ice as the source of CO2 is possible.[17]
Bromoacids 9a and 9b.
nBuLi (2.3 M in hexanes, 1.5 equiv, 3.0 mmol, 1.3 mL) was slowly added to a solution of
Ti(OiPr)4 (1.0 equiv, 2.0 mmol, 0.59 mL) in THF (10 mL), at 0 °C, over 6 min. After 4
additional min at the same temperature, CO2 was bubbled into the solution during 20 min.
NBS (1.5 equiv, 3.0 mmol, 0.53 g) was then added in one portion. The cold bath was removed
and stirring was continued for 2 h. 10% HCl aqueous solution (10 mL) was then slowly added
at 0 °C. The organic phase was separated and the aqueous layer was extracted with Et2O
(3 × 10 mL). The combined organic phases were dried over anhydrous MgSO4, filtered and
concentrated in vacuo (P ≥ 80 mbar) to give the crude product. Analysis by NMR (after
addition of 0.25 mmol of 4-(4-methoxyphenyl)-4-oxobutanoic acid used as an internal
standard)[20] revealed that it contained a mixture of 3-bromopentanoic and 2-
(bromomethyl)butanoic acids 9a and 9b in an 85 : 15 ratio and in an estimated combined
amount of 0.54 mmol [54% yield based on the theoretical formation of 1.0 mmol of 1 (R =
Et)].
Hydroxyoacids 10a and 10b.
nBuLi (2.2 M in hexanes, 1.5 equiv, 3.0 mmol, 1.3 mL) was slowly added to a solution of
Ti(OiPr)4 (1.0 equiv, 2.0 mmol, 0.59 mL) in THF (10 mL), at 0 °C, over 6 min. After 4
additional min at the same temperature, CO2 was bubbled into the solution during 20 min,
followed by O2 during 60 min, still at 0 °C. Saturated Na2S2O3 aqueous solution (3.0 mL) was
then carefully added, followed 10% HCl aqueous solution (10 mL). The organic phase was
separated and the aqueous layer was extracted with Et2O (3 × 10 mL). The combined organic
phases were dried over anhydrous MgSO4, filtered and concentrated in vacuo (P ≥ 80 mbar)
to give the crude product. Analysis by NMR (after addition of 0.25 mmol of 4-(4-
methoxyphenyl)-4-oxobutanoic acid used as an internal standard)[20] revealed that it contained
a mixture of 3-hydroxypentanoic and 2-(hydroxymethyl)butanoic acids 10a and 10b in an 84 :
− 14 −

ACCEPTED MANUSCRIPT
16 ratio and in an estimated combined amount of 0.60 mmol [60% yield based on the
theoretical formation of 1.0 mmol of 1 (R = Et)].
Deuteriated acids 5a-d and 5b-d.
nBuLi (2.1 M in hexanes, 1.5 equiv, 3.0 mmol, 1.4 mL) was slowly added to a solution of
Ti(OiPr)4 (1.0 equiv, 2.0 mmol, 0.59 mL) in THF (10 mL), at 0 °C, over 6 min. After 4
additional min at the same temperature, CO2 was bubbled into the solution during 40 min. The
reaction mixture was quenched with D2O (2.0 mL). The cold bath was removed and stirring
under N2 was maintained for 20 min. 10% HCl aqueous solution (10 mL) was then added.
The organic phase was separated and the aqueous layer was extracted with Et2O (3 × 10 mL).
The combined organic phases were dried over anhydrous MgSO4, filtered and concentrated in
vacuo (P ≥ 80 mbar) to give the crude product. Analysis by NMR (after addition of 0.25
mmol of 4-(4-methoxyphenyl)-4-oxobutanoic acid used as an internal standard)[20] revealed
that it contained a mixture of 3-deuteriopentanoic and 2-(deuteriomethyl)butanoic acids 5a-d
and 5b-d in an 83 : 17 ratio and in an estimated combined amount of 0.72 mmol [72% yield
based on the theoretical formation of 1.0 mmol of 1 (R = Et)]. The extent of deuterium
incorporation was estimated by NMR to be >85% in 5a-d and >81% in 5b-d.
Generation of 12 and hydrolysis, general procedure.
nBuLi (2.0 M in hexanes, 3.3 equiv, 3.3 mmol, 1.6 mL) was slowly added to a solution of the
alkyne substrate (1.0 equiv, 1.0 mmol) and Ti(OiPr)4 (2.2 equiv, 2.2 mmol, 0.65 mL) in THF
(10 mL), at 0 °C, over 6 min. After 2 h of stirring at the same temperature, 10% HCl aqueous
solution (10 mL) was carefully added. The organic phase was separated and the aqueous layer
was extracted with Et2O (3 × 10 mL). The combined organic phases were dried over
anhydrous MgSO4, filtered and concentrated in vacuo (P ≥ 25 mbar) to give the crude
product. Purification by flash column chromatography on silica gel (Et2O/petroleum ether,
gradient from 0 to 10% typically) yielded the pure (Z)-alkene product.
(Z)-1,2-Diphenylethene.[33] Colourless oil (156 mg, 0.87 mmol, 87%). Rf = 0.35 (petroleum
ether). 1H NMR (CDCl3, 400 MHz) δ 6.57 (2 H, s), 7.11–7.25 (10 H, m). 13C NMR (CDCl3,
100.6 MHz) δ 127.0, 128.2, 128.8, 130.2, 137.2.
(Z)-1-Phenylpent-1-ene.[34] Colourless oil (114 mg, 0.78 mmol, 78%). Rf = 0.4 (petroleum
ether). 1H NMR (CDCl3, 400 MHz) δ 0.95 (3 H, t, J 7.5), 1.49 (2 H, sext, J 7.5), 2.32 (2 H,
tdd, J 7.5, 7.0, 2.0), 5.68 (1 H, dt, J 11.5, 7.0), 6.42 (1 H, dt, J 11.5, 2.0), 7.18–7.37 (5 H, m).
13C NMR (CDCl3, 100.6 MHz) δ 13.8, 23.1, 30.7, 126.4, 128.1, 128.7, 128.8, 133.0, 137.8.
(Z)-1-(4-Methoxy)phenylpent-1-ene.[35] Colourless oil (136 mg, 0.78 mmol, 78%). Rf = 0.1
(petroleum ether). IR (neat)
ν 3005 (w), 2958 (m), 2931 (m), 2871 (w), 2835 (w), 1608 (m),
1
1510 (s), 1463 (m), 1442 (w), 1302 (w), 1250 (s), 1175 (m), 1037 (m), 837 (m) cm–1. H
NMR (CDCl3, 400 MHz) δ 0.93 (3 H, t, J 7.5), 1.47 (2 H, sext, J 7.5), 2.29 (2 H, tdd, J 7.5,
− 15 −

ACCEPTED MANUSCRIPT
7.0, 2.0), 3.77 (3 H, s), 5.56 (1 H, dt, J 11.5, 7.0), 6.34 (1 H, dt, J 11.5, 2.0), 7.03 (4 H,
AA’BB’ system,[36] δA 6.85, δB 7.21, N 9.0, L 8.0, K 5.0 (M could not be measured
accurately)). 13C NMR (CDCl3, 100.6 MHz) δ 13.8, 23.2, 30.7, 55.1, 113.5, 128.2, 129.9,
130.5, 131.4, 158.1. MS m/z (positive CI, NH3) 121, 147, 175, 176, 177 (MH+), 178. MS m/z
(EI) 115, 121, 132, 147, 148, 176 (M+•). HRMS m/z (EI) 176.1199 (M+• C12H16O requires
176.1196).
Carboxylation of p-methylstyrene 13.
nBuLi (1.9 M in hexanes, 9.0 equiv, 9.0 mmol, 4.7 mL) was slowly added to a solution of
Ti(OiPr)4 (6.0 equiv, 6.0 mmol, 1.8 mL) in THF (20 mL), at 0 °C, over 5 min. After 5
additional min at the same temperature, a solution of 4-methylstyrene (1.0 equiv, 1.0 mmol,
0.12 g) in THF (3.0 mL) was slowly added over 2 min. The mixture was allowed to warm to
20 °C over 5 min and after 30 min, CO2 was bubbled into the solution during 20 min. 10%
HCl aqueous solution (10 mL) was then added. The organic phase was separated and the
aqueous layer was extracted with Et2O (3 × 10 mL). The combined organic phases were dried
over anhydrous MgSO4, filtered and concentrated in vacuo (P ≥ 15 mbar) to give the crude
product. Analysis by NMR (after addition of 0.25 mmol of 4-(4-methoxyphenyl)-4-oxo-
butanoic acid used as an internal standard)[20] revealed that it contained extensive amounts of
the acids 5a and 5b and 0.75 mmol of 3-(p-tolyl)propanoic acid 14 (75%).
3-(4-Methylphenyl)propanoic acid (14).[33] 1H NMR (CDCl3, 400 MHz) δ 2.30 (3 H, s),
2.64 (2 H, t, J 8.0), 2.90 (2 H, t, J 8.0), 7.09 (4 H, br s), 10.83 (1 H, br s, CO2H). 13C NMR
(CDCl3, 100.6 MHz) δ 21.0, 30.2, 35.8, 128.2, 129.3, 135.9, 137.1, 179.7.
Reaction of ((but-3-en-1-yloxy)methyl)benzene 15.
nBuLi (2.1 M in hexanes, 2.5 equiv, 5.0 mmol, 2.4 mL) was slowly added to a solution of
Ti(OiPr)4 (1.25 equiv, 2.5 mmol, 0.74 mL) and ((but-3-en-1-yloxy)methyl)benzene (1.0
equiv, 2.0 mmol, 0.32 g) in THF (10 mL), at 0 °C, over 10 min. The reaction mixture was
kept at this temperature for 1 h and then stirred at 20 °C for 12 h. 10% HCl aqueous solution
(10 mL) was then added. The organic phase was separated and the aqueous layer was
extracted with Et2O (3 × 10 mL). The combined organic phases were dried over anhydrous
MgSO4, filtered and concentrated in vacuo (P ≥ 15 mbar) to give the crude product.
Purification by flash column chromatography on silica gel (Et2O/petroleum ether, gradient
from
0
to 5%) yielded fairly pure [(3R*,4R*)-6-benzyloxy-3,4-dimethyl-
hexoxy]methylbenzene following products. 16 (0.14 g), a 58 : 42 mixture of 16 and 17 (68
mg) and pure [(3R*,4R*)-3,4-dimethylhexyloxy]methylbenzene 17 (25 mg). The yields of the
products 16 and 17 are thus calculated to be 43% and 13% respectively.
[(3R*,4R*)-6-Benzyloxy-3,4-dimethylhexoxy]methylbenzene (16). Colourless oil. Rf 0.3
(petroleum ether, UV-active). 1H NMR (CDCl3, 400 MHz) δ 0.79 (6 H, d, J 6.5), 1.38–1.72 (6
H, m), 3.41–3.54 (4 H, m (AB part of an ABXY system), H4), 4.48 (4 H, AB system, δA 4.47,
− 16 −

ACCEPTED MANUSCRIPT
δB 4.49, JAB 12.0), 7.21–7.38 (10 H, m). 13C NMR (CDCl3, 100.6 MHz) δ 14.3, 33.6, 34.5,
69.0, 72.9, 127.4, 127.6, 128.3, 138.6.
[(3R*,4R*)-3,4-Dimethylhexyloxy]methylbenzene (17). Colourless oil. Rf 0.35 (petroleum
1
ether, UV-active). H NMR (CDCl3, 400 MHz) δ 0.76 (6 H, br d, J 7.0), 0.87 (3 H, t, J 7.0),
1.15 (1 H, m), 1.23–1.38 (2 H, m), 1.45 (1 H, m), 1.55–1.72 (2 H, m), 3.43–3.55 (2 H, m (AB
part of an ABXY system)), 4.50 (2 H, br s), 7.23–7.37 (5 H, m). 13C NMR (CDCl3, 100.6
MHz) δ 12.2, 14.0, 14.3, 27.4, 33.2, 34.7, 38.7, 69.2, 72.9, 127.5, 127.6, 128.3, 138.7.
− 17 −

ACCEPTED MANUSCRIPT
Graphical abstract
1
Kulinkovich, O. G.; Sviridov, S. V.; Vasilevskii, D. A.; Prityskaya, T. S. Zh. Org. Khim.
1989, 25, 2244–2245; Russ. J. Org. Chem. 1989, 25, 2027–2028.
2 Kulinkovich, O. G.; Sviridov, S. V.; Vasilevski, D. A. Synthesis 1991, 234.
3 Alkyne substrates (Sato generation of dialkoxytitanacyclopropene species), initial reports: a)
Harada, K.; Urabe, H.; Sato, F. Tetrahedron Lett. 1995, 36, 3203−3206; b) Nakagawa, T.;
Kasatkin, A.; Sato, F. Tetrahedron Lett. 1995, 36, 3207–3210; c) Urabe, H.; Hata, T.; Sato, F.
Tetrahedron Lett. 1995, 36, 4261−4264.
4
Amide substrates (de Meijere variation of the Kulinkovich reaction), selected references: a)
Chaplinski, V.; de Meijere, A. Angew. Chem. 1996, 108, 491–492; Angew. Chem. Int. Ed.
Engl. 1996, 35, 413–414; b) de Meijere, A.; Kozhushkov, S. I.; Savchenko, A. I. J.
Organomet. Chem. 2004, 689, 2033–2055; c) de Meijere, A.; Chaplinski, V.; Winsel, H.;
Kordes, M.; Stecker, B.; Gazizova, V.; Savchenko, A. I.; Boese, R.; Schill (née Brackmann),
F. Chem. Eur. J. 2010, 16, 13862–13875.
5
Nitrile substrates (Bertus and Szymoniak variation of the Kulinkovich reaction), selected
references: a) Bertus, P.; Szymoniak, J. Chem. Commun. 2001, 1792–1793; b) Bertus, P.;
Szymoniak, J. J. Org. Chem. 2002, 67, 3965–3968; c) Bertus, P.; Szymoniak, J. Synlett 2007,
1346–1356.
6 Selected reviews: a) Kulinkovich, O. G.; de Meijere, A. Chem. Rev. 2000, 100, 2789–2834;
b) Sato, F.; Urabe, H.; Okamoto, S. Chem. Rev. 2000, 100, 2835–2886; c) Wolan, A.; Six, Y.
Tetrahedron 2010, 66, 15–61; d) Wolan, A.; Six, Y. Tetrahedron 2010, 66, 3097–3133; e)
Cha, J. K.; Kulinkovich O. G. in Organic Reactions, Vol. 77 (Ed.: S. E. Denmark), Wiley,
New York, 2012, Chapter 1.
7 a) Averbuj, C.; Kaftanov, J.; Marek, I. Synlett 1999, 1939–1941; b) Mizojiri, R.; Urabe, H.;
Sato, F. Tetrahedron Lett. 1999, 40, 2557–2560; c) Mizojiri, R.; Urabe, H.; Sato, F. J. Org.
Chem. 2000, 65, 6217–6222.
8 Contributions from our group: a) Six, Y. J. Chem. Soc. Perkin Trans. 1 2002, 1159–1160; b)
Six, Y. Eur. J. Org. Chem. 2003, 1157–1171; c) Cadoret, F.; Retailleau, P.; Six, Y.
Tetrahedron Lett. 2006, 47, 7749–7753.
9
a) Eisch, J. J. J. Organomet. Chem. 2001, 617−618, 148−157; b) Eisch, J. J.; Gitua, J. N.;
Otieno, P. O.; Shi, X. J. Organomet. Chem. 2001, 624, 229−238; c) Eisch, J. J.; Owuor, F. A.;
Otieno, P. O. Organometallics. 2001, 20, 4132−4134.
10
Other relevant references by the same group: a) Eisch, J. J.; Gitua, J. N. Eur. J. Inorg.
Chem. 2002, 3091−3096; b) Eisch, J.J.; Gitua, J. N. Organometallics 2003, 22, 24−26; 1170;
c) Eisch, J. J.; Gitua, J. N.; Doetschman, D. C. Eur. J. Inorg. Chem. 2006, 1968−1975.
− 18 −

ACCEPTED MANUSCRIPT
11 Obora, Y.; Moriya, H.; Tokunaga, M.; Tsuji, Y. Chem. Commun. 2003, 2820−2821.
12 Santra, S.; Masalov, N.; Epstein, O. G.; Cha, J. K. Org. Lett. 2005, 7, 5901−5904.
13 Barbour, N. M.; Phillips, A. J. ARKIVOC 2008, xvi, 110−118.
14 Tarselli, M. A.; Micalizio, G. C. Org. Lett. 2009, 11, 4596–4599.
15 McCabe, J.; Phillips, A. J. Tetrahedron 2013, 69, 5710−5714.
16 Review: Negishi, E. Dalton Trans. 2005, 827–848.
17
Two alternative procedures were employed: a) CO2 gas from an AirLiquide cylinder,
introduced directly (industrial quality, containing 200 ppm H2O); or b) CO2 gas produced by
the sublimation of dry ice chunks and dried through a pad of MgSO4 and CaCl2. With freshly
delivered dry ice, both methods gave comparable results but when the dry ice had been stored
in a freezer for several days, the relative amounts of the non-halogenated products 5a and 5b
increased and the yields of the fully functionalised products 4a and 4b were lower.
18
The reaction mixtures looked different depending on the nature of the solvent employed.
For instance, when the experiments were conducted in THF, the reaction mixtures were
typically brown after the addition of nBuLi and then darkened to become black-brown. In
Et2O, although no difference of solubility was noticed, the colour had already turned black-
brown at the end of the addition of nBuLi.
19
This idea is further supported by experiments conducted using MeLi and nBuLi
sequentially, leading to much lower 4a / 4b selectivity. See the supporting information for
detail.
20 See the supporting information for detail.
21 Starting from terminal alkynes, cyclotrimerisation products are observed. The results of our
on-going studies dealing with this transformation will be published shortly.
22
Slow addition of the alkene onto a solution of excess amounts of 1 (R = Et) is expected to
minimise insertion of another molecule of this alkene into the titanacyclopropane resulting
from the ligand exchange reaction.
23 a) Lee, J. C.; Sung, M. J.; Cha, J. K. Tetrahedron Lett. 2001, 42, 2059–2061; b) de Meijere,
A.; Williams, C. M.; Kourdioukov, A.; Sviridov, S. V.; Chaplinski, V.; Kordes, M.;
Savchenko, A. I.; Stratmann, C.; Noltemeyer, M. Chem. Eur. J. 2002, 8, 3789–3801.
24 Lee, J.; Cha, J. K. J. Org. Chem. 1997, 62, 1584–1585.
25 The main product of this reaction was the isopropyl ester resulting from a transesterification
reaction of the substrate 19 (42% yield).
26
In a footnote of the article cited in ref. 12, these authors mention experiments carried out
with esters with the same reagent system, giving the corresponding cyclopropanols in
15−20% yield.
27 This is in contradiction with the claims found in ref. 10b.
28
A couple of additional experiments, conducted with Ti(OiPr)4 and Grignard reagents,
suggest that the presence of added LiCl does not improve the stability of the
titanacyclopropane species.
29 Kofron, W. G.; Baclawski, L. M. J. Org. Chem. 1976, 41, 1879–1880.
30 Lin, H.-S.; Paquette, L. A. Synth. Comm. 1994, 24, 2503–2506.
31
Augustowska, E.; Boiron, A.; Deffit, J.; Six, Y. Chem. Commun. 2012, 48, 5031–5033
(supporting information).
32
Madelaine, C.; Six, Y.; Buriez, O. Angew. Chem. 2007, 119, 8192–8195; Angew. Chem.
Int. Ed. 2007, 46, 8046–8049 (supporting information).
33 Commercially available compound.
34 Bellucci, G.; Chiappe, C.; Cordoni, A. Tetrahedron: Asym. 1996, 7, 197–202.
35 Delmas, M.; Le Bigot, Y.; Gaset, A.; Gorrichon, J. P. Synth. Commun. 1981, 11, 125–132.
36 Günther, H. Angew. Chem. 1972, 84, 907–920; Angew. Chem. Int. Ed. 1972, 11, 861–874.
− 19 −

ACCEPTED MANUSCRIPT
A study of the reaction of nBuLi with Ti(OiPr)4 as a method to
generate titanacyclopropane and titanacyclopropene species
Valentin Rassadin, Yvan Six*
Laboratoire de Synthèse Organique DCSO
UMR 7652 CNRS/Ecole Polytechnique, 91128 Palaiseau Cedex (France). Fax: (+33) (0)1 6933 5972
*E-mail: yvan.six@polytechnique.edu
Supporting information
I. Method of analysis of the crude products by NMR
The two following detailed examples are representative of the way the crude products were
analysed.
I.1. Example 1 (Scheme 2 of the article): trapping by CO2 and then by I2 of the
organotitanium species generated from Ti(OiPr)4 (2.0 mmol) and nBuLi (4.0 mmol).
a) 4-(4-Methoxyphenyl)-4-oxobutanoic acid (0.25 mmol) was added to the crude product.
b) The resulting mixture was analysed by 1H and 13C NMR spectroscopy.
c) Information extracted from the 1H NMR spectrum:
• The integral of the signal at 2.85 ppm was set at 0.50 [4-(4-methoxyphenyl)-4-
oxobutanoic acid, CH2CO2H].
• The amount of 3-iodopentanoic acid 4a was estimated on the basis of the integral of the
signal at 3.05 ppm (CH2CO2H). In this example: about 0.76 mmol.
− 1 −

ACCEPTED MANUSCRIPT
• The amount of pentanoic acid 5a was estimated on the basis of the integral of the signal
at 2.35 ppm (CH2CO2H). In this example: about 0.11 mmol.
• When the amount of 2-(iodomethyl)butanoic acid 4b was not too low, it could also be
estimated, on the basis of the integral of the signal at 2.67 ppm (CHCO2H). In this
example: approximately 0.21 mmol.
d) Information extracted from the 13C NMR spectrum:
• The ratio of 3-iodopentanoic acid 4a and 2-(iodomethyl)butanoic acid 4b was estimated
on the basis of the intensity of the signal at 33.4 ppm (CH3CH2) vs the intensity of the
signal at 25.9 ppm (CH3CH2) [or, alternatively, vs the intensity of the signal at 4.0 ppm
(CH2I)]. In this example: about 78 : 22.
• The ratio of pentanoic acid 5a and 2-methylbutyric acid 5b was estimated on the basis of
the intensity of the signal at 26.7 ppm (CH3CH2CH2) vs the intensity of the signal at 26.5
ppm (CH3CH2). This ratio was sometimes difficult to determine, especially when the
amounts of the two acids were very low. There is no such problem in this example: the
ratio is estimated to be approximately 60 : 40.
e) The total amount of iodinated acids (4a + 4b) could then be deduced, as well as the total
amount of non-iodinated acids (5a + 5b). In this example: about 0.97 mmol and 0.20 mmol
respectively.
1H NMR spectrum of the crude product.
▲ reference. ▲ 3-iodopentanoic acid 4a. ▲ 2-(iodomethyl)butanoic acid 4b. ▲ pentanoic acid 5a.
− 2 −

ACCEPTED MANUSCRIPT
13C and DEPT NMR spectra of the crude product.
● reference. ● 3-iodopentanoic acid 4a. ● 2-(iodomethyl)butanoic acid 4b.
● pentanoic acid 5a. ● 2-methylbutyric acid 5b. ● solvents: THF, diethyl ether.
I.2. Example 2: hydrolysis, after 60 min, of the organotitanium species generated in THF
from 1-phenylpent-1-yne (1.0 mmol), Ti(OiPr)4 (2.2 mmol) and nBuLi (3.3 mmol).
a) A rather concentrated sample of the crude product# was prepared and analysed by 1H NMR
spectroscopy.
#
Around 100 mg in 0.50 mL of CDCl3. This is to minimise the integral of the residual
CHCl3 peak.
b) Information extracted from the 1H NMR spectrum:
• The integral of the aromatic proton signals, in the range 7.00−7.60 ppm, was set at 5.00.
• The amount of (Z)-1-phenylpent-1-ene was estimated on the basis of the integral of the
signal at 6.41 ppm (PhCHCH). In this example: about 0.49 mmol.
• The ratio of (Z)-1-phenylpent-1-ene and starting 1-phenylpent-1-yne was estimated on
the basis of the integrals of the signals at 2.31 and 2.38 ppm respectively (CH3CH2CH2).
In this example: about 57 : 43.
c) The amount of starting 1-phenylpent-1-yne could then be deduced. In this example:
approximately 0.37 mmol.
− 3 −

ACCEPTED MANUSCRIPT
1H NMR spectrum of the crude product (left part).
▲ reference. ▲ (Z)-1-phenylpent-1-ene.
1H NMR spectrum of the crude product (right part).
▲ reference. ▲ (Z)-1-phenylpent-1-ene. ▲ 1-phenylpent-1-yne. ▲ solvents: THF, diethyl ether
− 4 −

ACCEPTED MANUSCRIPT
II. Characteristic NMR data
The data provided here relates to the compounds appearing in the main article as well as to
other molecules mentioned in the present document.
II.1. Reference compound used for the NMR analyses of the crude products.
4-(4-Methoxyphenyl)-4-oxobutanoic acid1
1H NMR (CDCl3, 400 MHz)2 2.80 (2 H, t, J 6.5, H2), 3.28 (2 H, t, J 6.5, H3), 3.87 (3 H, s,
H9), 6.94 (2 H, dm,† J 9.0, H7), 7.97 (2 H, dm,† J 9.0, H6), 9.93 (1 H, br s, CO2H). 13C NMR
(CDCl3, 100.6 MHz)2 28.2 (C2), 32.8 (C3), 55.5 (C9), 113.8 (C7), 129.4 (C5), 130.3 (C7),
163.6 (C8), 178.3 (C1), 196.5 (C4).
†The signals of the protons of the aromatic ring appear as an AA’BB’ system but are
described here as two doublets of multiplets, for simplification purposes.
II.2. Compounds that were not isolated.
3-Iodopentanoic acid (4a)
1H NMR (CDCl3, 400 MHz) δ 1.04 (3 H, t, J 7.0, H1), 1.73–1.86 (2 H, m, H2), 3.05 (2 H, d,‡
13
J 7.0, H4), 4.31 (1 H, dtd, J 8.0, 7.0, 5.0, H3), 9.93 (1 H, br s, CO2H). C NMR (CDCl3,
100.6 MHz) δ 14.1 (C1), 29.3 (C3), 33.4 (C2), 45.5 (C4), 175.7 (C5).
‡This signal appears as a doublet in relatively dilute solutions and as an ABX system in more
concentrated solutions.
1− Commercially available compound.
2− The signals are described as they appear on the NMR spectra of the crude product of the reaction depicted in
Scheme 2 of the article. Thus, the chemical shifts reported here may differ slightly from the chemical shifts
measured from an authentic sample.
− 5 −

ACCEPTED MANUSCRIPT
2-(Iodomethyl)butanoic acid (4b)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 0.97 (3 H, t, J 7.5, H1), 2.67 (1 H, tdd, J
7.5, 6.0, 5.5, H3), 3.32 (2 H, AB part of an ABX system, A 3.28, B 3.36, JAB 10.0, JAX 5.5,
13
JBX 7.5, H4), 9.93 (1 H, br s, CO2H). C NMR (CDCl3, 100.6 MHz) δ 4.0 (C4), 11.1 (C1),
25.9 (C2), 49.3 (C3), 178.6 (C5).
Pentanoic acid (5a)1
1H NMR (CDCl3, 400 MHz)2 0.92 (3 H, t, J 7.5, H1), 1.37 (2 H, sext, J 7.5, H2), 1.62 (2 H,
13
quint, J 7.5, H3), 2.35 (2 H, t, J 7.5, H4), 9.93 (1 H, br s, CO2H). C NMR (CDCl3, 100.6
MHz)2 13.7 (C1), 22.1 (C2), 26.7 (C3), 33.8 (C4), 179.9 (C5).
3-Deuteriopentanoic acid (5a-d)
1H NMR (CDCl3, 400 MHz) δ 0.92 (3 H, t, J 7.5, H1), 1.37 (2 H, quint, J 7.5, H2), 1.60 (2 H,
13
quint, J 7.5, H3), 2.34 (2 H, d, J 7.5, H4), 7.79 (1 H, br s, CO2H). C NMR (CDCl3, 100.6
MHz) δ 13.5 (C1), 22.0 (C2), 26.3 (t, J 19.5, C3), 33.6 (C4), 179.3 (C5).
2-Methylbutanoic acid (5b)1
1H NMR (CDCl3, 400 MHz)2 0.94 (3 H, t, J 7.5, H1), 1.17 (3 H, d, J 7.0, H4), 1.60 (2 H, AB
part of an ABX3Y system, A 1.50, B 1.71, JAB 13.5, JAX 7.5, JAY 6.5, JBX 7.5, JBY 7.0, H2),
13
2.40 (1 H, quintd, J 7.0, 6.5, H3), 9.93 (1 H, br s, CO2H). C NMR (CDCl3, 100.6 MHz)2
11.5 (C1), 16.3 (C4), 26.5 (C2), 40.8 (C3), 182.8 (C5).
2-(Deuteriomethyl)butanoic acid (5b-d)
13C NMR (CDCl3, 100.6 MHz) δ 11.4 (C1), 16.0 (t, J 19.5, C4), 26.4 (C2), 40.7 (C3), 182.3
(C5).
− 6 −

ACCEPTED MANUSCRIPT
3-Bromopentanoic acid (9a)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 1.07 (3 H, t, J 7.5, H1), 2.95 (2 H, d, J
7.0, H4)‡, 4.30 (1 H, dtd, J 8.0, 7.0, 4.5, H3), 8.49 (1 H, br s, CO2H). 13C NMR (CDCl3, 100.6
MHz) δ 12.0 (C1), 31.8 (C2), 43.7 (C4), 51.8 (C3), 173.7 (C5).
‡This signal appeared as a doublet in relatively dilute solutions and as an ABX system in more
concentrated solutions.
2-(Bromomethyl)butanoic acid (9b)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 3.54 (2 H, AB part of an ABX system,
13
A 3.50, B 3.59, JAB 10.5, JAX 5.5, JBX 8.0, H4), 8.49 (1 H, br s, CO2H). C NMR (CDCl3,
100.6 MHz) characteristic signals: δ 11.2 (C1), 36.8 (C4), 49.2 (C3), 176.9 (C5).
3-Hydroxypentanoic acid (10a)3
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 0.96 (3 H, t, J 7.5, H1), 2.47 (2 H, AB
part of an ABX system, A 2.43, B 2.52, JAB 16.0, JAX 9.0, JBX 3.5, H4), 3.95 (1 H, dddd, J
9.0, 7.0, 5.5, 3.5, H3), 7.41 (1 H, br s, CO2H). 13C NMR (CDCl3, 100.6 MHz) δ 9.8 (C1), 29.4
(C2), 40.7 (C4), 64.5 (C3), 175.8 (C5).
2-(Hydroxymethyl)butanoic acid (10b)4
13C NMR (CDCl3, 100.6 MHz) characteristic signals: δ 11.7 (C1), 21.6 (C2), 49.0 (C3), 62.6
(C4), 178.4 (C5).
3− Capon, R. J.; Dunlop, R. W.; Ghisalberti, E. L.; Jefferies, P. R. Phytochemistry 1983, 22, 1081−1183.
4− Guazzelli, G.; De Grazia, S.; Collins, K. D.; Matsubara, H.; Spain, M.; Procter, D. J. J. Am. Chem. Soc. 2009,
131, 7214−7215 (supporting information).
− 7 −

ACCEPTED MANUSCRIPT
3-Deuterio-3-methylbutanoic acid (see section IV.2)
1H NMR (CDCl3, 400 MHz) δ 0.97 (6 H, s, H1), 2.18 (2 H, s, H3), 7.00 (1 H, br s, CO2H).
13C NMR (CDCl3, 100.6 MHz) δ 22.3 (C1), 25.6 (C2), 43.1 (C3), 176.7 (C4). The signal of
C2 could not be directly observed because it was buried into the peak of THF. However, its
chemical shift was deduced from an HMBC 2D experiment, where it correlated with both H1
and H3.
3-Deuterio-3-methyl-pentanedioic acid (see section IV.2)
1H NMR (CDCl3, 400 MHz) δ 1.05 (3 H, s, H1), 2.33 (4 H, AB system, δA 2.24, δB 2.42, JAB
15.5, H3), 7.00 (2 H, br s, CO2H). 13C NMR (CDCl3, 100.6 MHz) δ 19.7 (C1), 26.9 (C2), 40.5
(C3), 175.7 (C4). The signal of C2 was not directly observed but its chemical shift was
deduced from an HMBC 2D experiment, showing a correlation with H1 and H3 at the
chemical shift indicated.
3-(4-Methylphenyl)propanoic acid (14)1
1H NMR (CDCl3, 400 MHz) δ 2.30 (3 H, s, H8), 2.64 (2 H, t, J 8.0, H2), 2.90 (2 H, t, J 8.0,
H3), 7.09 (4 H, br s, H5, H6), 10.83 (1 H, br s, CO2H). 13C NMR (CDCl3, 100.6 MHz) δ 21.0
(C8), 30.2 (C3), 35.8 (C2), 128.2, 129.3 (C5, C6), 135.9, 137.1 (C4, C7), 179.7 (C1).
(E)-1-[2-(3,4-Dimethoxyphenyl)ethyl]-2-ethyl-
cyclopropanol (20)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 0.08 (1 H, dd, J 6.0, 5.5, H4a), 0.97 (3
13
H, t, J 7.5, H1), 1.94 (1 H, dddd, J 14.5, 10.0, 6.0, 1.0, H3). C NMR (CDCl3, 100.6 MHz)
characteristic signals: δ 13.7 or 13.8 (C1), 19.1 (C4), 22.6 (C3), 27.6 (C2), 32.0 (C7), 36.4
(C6), 58.7 (C5).
− 8 −

ACCEPTED MANUSCRIPT
1-(3,4-Dimethoxyphenyl)heptan-3-one (21)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 0.88 (3 H, t, J 7.5, H1), 2.38 (2 H, t, J
7.5, H4). 13C NMR (CDCl3, 100.6 MHz) characteristic signals: δ 13.7 or 13.8 (C1), 22.1 (C2),
25.7 (C3), 29.2 (C7), 42.6 (C4), 44.3 (C6), 210.5 (C5).
5-Benzyloxypentanoic acid 5 (see section IV.5)
1H NMR (CDCl3, 400 MHz) δ 1.62−1.78 (4 H, m, H3−H4), 2.37 (2 H, t, J 7.0, H2), 3.49 (2
13
H, t, J 6.0, H5), 4.50 (s, 2 H), 7.23−7.37 (5 H, m, H8−H10), 10.51 (1 H, br s, CO2H). C
NMR (CDCl3, 100.6 MHz) δ 21.5, 29.0, 33.8 (C2−C4), 69.7 (C5), 72.9 (C6), 127.6 (C10),
127.7, 128.4 (C8, C9), 138.3 (C7), 180.4 (C1).
(E)-2-Ethyl-1-methylcyclopropanol (see section IV.6)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 0.05 (1 H, dd, J 6.0, 5.0, H4a), 0.81 (1
H, dd, J 10.0, 5.0, H4b), 0.98 (3 H, t, J 7.5, H1), 1.41 (3 H, s, H6).
Note: the (E) relative configuration was assigned on the basis of the NMR data reported for
the closely related compound (E)-2-butyl-1-methylcyclopropanol.6
2-Hexanone1 (see section IV.6)
1H NMR (CDCl3, 400 MHz) characteristic signals: δ 0.91 (3 H, t, J 7.5, H1), 2.14 (3 H, s,
H6), 2.43 (2 H, t, J 7.5, H4).
5− Börjesson, L.; Csöregh, I.; Welch, C. J. J. Org. Chem. 1995, 60, 2989−2999.
6− a) Chaplinski, V.; Winsel, H.; Kordes, M.; de Meijere, A. Synlett 1997, 111−112; b) Kulinkovich, O. G.;
Kananovich, D. G. Eur. J. Org. Chem. 2007, 2121−2132.
− 9 −

ACCEPTED MANUSCRIPT
II.3. Isolated products.
(Z)-1,2-Diphenylethene1
Colourless oil. Rf 0.35 (petroleum ether, UV-active). 1H NMR (CDCl3, 400 MHz) δ 6.57 (2 H,
s, H1), 7.11–7.25 (10 H, m, H3–H5). 13C NMR (CDCl3, 100.6 MHz) δ 127.0 (C5), 128.2,
128.8 (C3–C4), 130.2 (C1), 137.2 (C2).
1H NMR spectrum (400 MHz, CDCl3).
− 10 −