Understanding the competitive dehydroalkoxylation and dehydrogenation of ethers with Ti-C multiple bonds

Article abstract of DOI:10.1039/c3sc22195d

The divergent reactivity of a transient titanium neopentylidyne, (PNP)TiC^tBu (A) (PNP = N[2-PⁱPr₂-4- methylphenyl]₂^-), that exhibits competing dehydrogenation and dehydroalkoxylation reaction pathways in the presence of acyclic ethers (Et₂O, ⁿPr₂O, ⁿBu₂O, ^tBuOMe, ^tBuOEt, ⁱPr₂O) is presented. Although dehydrogenation takes place also in long-chain linear ethers, dehydroalkoxylation is disfavoured and takes place preferentially or even exclusively in the case of branched ethers. In all cases, dehydrogenation occurs at the terminal position of the aliphatic chain. Kinetics analyses performed using the alkylidene-alkyl precursor, (PNP)TiCH^tBu(CH ₂^tBu), show pseudo first-order decay rates on titanium (k_avg = 6.2 ± 0.3 × 10^-5 s^-1, at 29.5 ± 0.1°C, overall), regardless of the substrate or reaction pathway that ensues. Also, no significant kinetic isotope effect (k _H/k_D ～ 1.1) was found between the activations of Et₂O and Et₂O-d₁₀, in accord with dehydrogenation (C-H activation and abstraction) not being the slowest steps, but also consistent with formation of the transient alkylidyne A being rate-determining. An overall decay rate of (PNP)TiCH^tBu(CH ₂^tBu) with a t_1/2 = 3.2 ± 0.4 h, across all ethers, confirms formation of A being a common intermediate. Isolated alkylidene-alkoxides, (PNP)TiCH^tBu(OR) (R = Me, Et, ⁿPr, ⁿBu, ⁱPr, ^tBu) formed from dehydroalkoxylation reactions were also independently prepared by salt metatheses, and extensive NMR characterization of these products is provided. Finally, combining theory and experiment we discuss how each reaction pathway can be altered and how the binding event of ethers plays a critical role in the outcome of the reaction.

Full text of DOI:10.1039/c3sc22195d

Chemical Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Understanding the competitive dehydroalkoxylation
and dehydrogenation of ethers with Ti–C multiple
bonds†
Cite this: Chem. Sci., 2013, 4, 2543
Marco G. Crestani, Andras Olasz, Balazs Pinter, Brad C. Bailey, Skye Fortier,
*
Xinfeng Gao, Chun-Hsing Chen, Mu-Hyun Baik and Daniel J. Mindiola
The divergent reactivity of a transient titanium neopentylidyne, (PNP)Ti^CtBu (A) (PNP ¼ N[2-PiPr2-4-
methylphenyl]2ꢀ), that exhibits competing dehydrogenation and dehydroalkoxylation reaction pathways
i
in the presence of acyclic ethers (Et2O, nPr2O, nBu2O, tBuOMe, tBuOEt, Pr2O) is presented. Although
dehydrogenation takes place also in long-chain linear ethers, dehydroalkoxylation is disfavoured and
takes place preferentially or even exclusively in the case of branched ethers. In all cases,
dehydrogenation occurs at the terminal position of the aliphatic chain. Kinetics analyses performed
t
using the alkylidene-alkyl precursor, (PNP)Ti]CHtBu(CH2 Bu), show pseudo first-order decay rates on
titanium (kavg ¼ 6.2 ꢁ 0.3 ꢂ 10ꢀ5 sꢀ1, at 29.5 ꢁ 0.1 ꢃC, overall), regardless of the substrate or reaction
pathway that ensues. Also, no significant kinetic isotope effect (kH/kD ꢄ 1.1) was found between the
activations of Et2O and Et2O-d10, in accord with dehydrogenation (C–H activation and abstraction)
not being the slowest steps, but also consistent with formation of the transient alkylidyne A being
rate-determining. An overall decay rate of (PNP)Ti]CHtBu(CH2 Bu) with a t1/2 ¼ 3.2 ꢁ 0.4 h, across
t
all ethers, confirms formation of A being a common intermediate. Isolated alkylidene-alkoxides,
Received 10th December 2012
Accepted 26th March 2013
(PNP)Ti]CHtBu(OR) (R ¼ Me, Et, nPr, nBu, Pr, Bu) formed from dehydroalkoxylation reactions were also
independently prepared by salt metatheses, and extensive NMR characterization of these products is
provided. Finally, combining theory and experiment we discuss how each reaction pathway can be
altered and how the binding event of ethers plays a critical role in the outcome of the reaction.
i
t
DOI: 10.1039/c3sc22195d
www.rsc.org/chemicalscience
in order to exploit their reactivity. However, unlike alkanes,
ethers might show divergent reactivity due to the oxygen lone
Introduction
The ether functional group is ubiquitous in industry, being
readily produced on massive scales either by acid-catalyzed
dehydration of alcohols (alkoxy-de-hydroxylation; the preferred
method for linear, symmetric ethers), or by reaction of alkyl
halides with alkali-metal alkoxides (the Williamson-ether
synthesis).1 Commonplace in chemistry, acyclic ethers are
essentially inert substrates comparable to alkanes.2 This feature
explains their ready availability and use as solvents, yet confers
them little importance for chemical transformations. Given our
interest in C–H activation of alkanes, it is possible to conceive
that the same organometallic compounds that cleave C–H
bonds of alkanes can equally be useful in the presence of ethers,
pair of electrons being a better donor than a C–H bond. A
promising alternative could be the direct conversion of ethers to
ether-substituted a-olens via transfer dehydrogenation; an
attractive transformation given the potential that these unsat-
urated products exhibit as monomers in polymerization reac-
tions.3 Unfortunately, a,b-dehydrogenation of ethers is a rare
phenomenon4–7 when compared to a,a0-dehydrogenation of
ethers8,9 or a,b-dehydrogenation of alkanes,10,11 and therefore,
this process constitutes a new paradigm in chemical trans-
formation: ether-substituted a-olens sometimes being chal-
lenging compounds to synthesize. Another important
transformation of ethers is their cracking12 or dehydroalkoxy-
lation to a-olens.13
Templeton and co-workers reported an example of transfer
dehydrogenation of Et2O and THF by a transient (nacnac)Pt(CH3)
(nacnacꢀ ¼ [PhNC(CH3)]2CH) complex, using tert-butyl ethylene
as a hydrogen acceptor.7 Bergman and Grubbs have also reported
iridium based systems that can dehydrogenate Et2O to the vinyl
ether adduct.5,6 To our knowledge, examples of early transition
metal complexes that can dehydrogenate ethers have not been
Department of Chemistry, Indiana University, E Kirkwood Ave., Bloomington, Indiana,
47405, USA. E-mail: mindiola@indiana.edu; Fax: +1 812 855 8300; Tel: +1 812 855
2399
† Electronic supplementary information (ESI) available: Experimental procedures,
spectroscopic data for all new compounds, kinetics data, structural determination
parameters and theoretical protocols. CCDC reference number 914777. For ESI
and crystallographic data in CIF or other electronic format see DOI:
10.1039/c3sc22195d
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2543–2550 | 2543

View Article Online
Chemical Science
Edge Article
reported and developing a system that could convert ethers to a-
olens having an ether moiety would be attractive, especially if
one could understand the mechanism for this conversion and
selectivity behind C–H versus C–O activation. In this work we
report competitive transformations in ethers promoted by a
transient titanium neopentylidyne, (PNP)Ti^CtBu (A) (PNP ¼
N[2-PiPr2-4-methylphenyl]2
) and establish that ethers can be
ꢀ
14
dehydroalkoxylated as well as selectively dehydrogenated at the
terminal position, in close resemblance to what happens with
linear alkanes;15 the extent and preference of one route over the
other being dependent on the chain length of the aliphatic ether
and its degree of ramication. Experimental and theoretical
studies were performed to understand the nature of these
competitive pathways which share common intermediates and
in this light, predict how each pathway can be favored or
disfavored.
Scheme
2
Dehydroalkoxylation and dehydrogenation reactions involving
t
(PNP)Ti]CHtBu(CH2tBu) and BuOMe.
formed in the mixture, using a combination of spectroscopic
techniques (1H NMR spectroscopy and GC-MS). Unfortunately, we
cannot conclusively state by these techniques whether another
organic product, (CH2]CHOEt), was also present given the low
quantity of (PNP)Ti(h2-H2C]CH2)(CH2 Bu) that was produced, as
t
Results and discussion
Dehydroalkoxylation of ethers by a titanium neopentylidyne
well as the reactive nature of this vinyl ether. Regardless, forma-
tion of the organometallic adduct with ethene suggests that a
competitive pathway takes place in the reaction: dehydrogenation
t
Dissolving (PNP)Ti]CHtBu(CH2 Bu) in neat Et2O (room
t
of Et2O to a vinyl ether adduct, (PNP)Ti(h2-H2C]CHOEt)(CH2 Bu),
temperature) over 12 h results in clean (>90%) formation of the
neopentylidene-ethoxide (PNP)Ti]CHtBu(OEt) (1), accompanied
which then likely undergoes exchange with ethene to form
t
(PNP)Ti(h2-H2C]CH2)(CH2 Bu) and free CH2]CHOEt
t
by concomitant evolution of ethene and CH3 Bu (Scheme 1).16
1
Examination of the reaction mixture by H and 31P{1H} NMR
(Scheme 1).
When (PNP)Ti]CHtBu(CH2 Bu) is treated with BuOMe, the
neopentylidene-alkoxide derivative (PNP)Ti]CHtBu(OMe) (2) is
the only titanium complex formed, along with isobutene and
t
t
spectroscopy also revealed formation of an additional titanium-
based product in minor yield (#5%), which we identied to be the
t
ethene complex, (PNP)Ti(h2-H2C]CH2)(CH2 Bu), by spectroscopic
t
CH3 Bu (Scheme 2). Analogous to 1, complex 2 can be
comparison with an authentic sample prepared independently
similarly prepared from (PNP)Ti]CHtBu(OTf) and NaOMe
(although contaminated with some free ligand (PNP)H, as
gauged by 31P{1H} NMR spectroscopy, Scheme 2). Compounds 1
and 2 possess similar NMR spectroscopic signatures: the alkyli-
dene hydrogen at ꢄ11 ppm (1JCH ꢄ 98 Hz) and alkoxide reso-
nances around 4.5 ppm (4.7 and 4.5 ppm for 1, 4.4 ppm for 2;
Table S2†)16 in the 1H NMR spectra. To probe for selectivity of the
t
from (PNP)Ti]CHtBu(CH2 Bu) and ethane or by salt metathesis of
(PNP)Ti]CHtBu(OTf) with LiCH2CH3 (Scheme 1).15 Compound 1
1
was characterized spectroscopically by H, 13C{1H}, and 31P{1H}
NMR spectroscopy, along with other multidimensional tech-
niques (COSY, dqCOSY, HSQC, DEPT),16 and was independently
prepared (albeit marred with ꢄ4 to 6% content of the free ligand
(PNP)H, by 31P{1H} NMR) from (PNP)Ti]CHtBu(OTf) and NaOEt,
in THF (Scheme 1). Vacuum transfer of the volatiles from neat
Et2O activation permitted denite identication of free ethene
dehydroalkoxylation reaction, complex (PNP)Ti]CHtBu(CH2 Bu)
t
was also treated with tBuOEt. Examination of the mixture
revealed almost a 1 : 1 ratio of 1 and (PNP)Ti]CHtBu(OtBu) (3),
in addition to the b-elimination products ethene and isobutene
1
(Scheme 3). No vinyl ether could be detected by H NMR spec-
t
troscopy although formation of (PNP)Ti(h2-H2C]CH2)(CH2 Bu)15
is also observed but in very low yields (<2% based on 31P{1H}
Scheme
1
Dehydroalkoxylation and dehydrogenation reactions involving
Scheme
3
Dehydroalkoxylation and dehydrogenation reactions involving
t
(PNP)Ti]CHtBu(CH2tBu) and Et2O.
(PNP)Ti]CHtBu(CH2tBu) and BuOEt.
2544 | Chem. Sci., 2013, 4, 2543–2550
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
Table 1 Summarized kinetic data for aliphatic ether activations as well as some
hydrocarbons
Conv. kavg ꢂ 10ꢀ5
t1/2
c
Tavg
Entry Substrate
(ꢃC)
(%)
(sꢀ1
)
s ꢂ 10ꢀ5 (h)
1
tBuOMe
Et2O
29.4 90
29.6 97
29.2 91
29.5 83
29.6 85
29.6 96
29.6 95
5.8
7.0
6.4
5.9
5.0
7.5
6.0
5.7
6.5
6.0
0.2
0.2
0.2
0.5
0.0
0.1
0.5
0.5
0.4
0.5
0.02
3.3
2.7
3.0
3.3
3.9
2.6
3.2
3.4
3.0
3.2
3.9
2a
3a
4
Et2O-d10
nPr2O
5
6
7
8
nBu2O
tBuOEt
iPr2O
tBuOMe:Et2O 29.5 92
9b
10
11
C6D6
Pentane
Cyclohexane
27.0 88
29.7 90
29.6 89
4.97
a
KIE ¼ kH/kD ¼ 1.1(0). Reactions were performed by duplicate runs; s
Fig. 1 Solid state molecular structure of [(PNP)Ti]CHtBu(OtBu)] (3) with 50%
b
c
indicates the normal error (standard deviation). Ref. 13. t1/2
ln(2)/kavg
¼
probability ellipsoids. Asterisks denote symmetry generated atoms.
.
formation of transient A have been previously reported.14 Table 1
summarizes our kinetic data involving various ethers. Measure-
ments in C6D6 (kavg ¼ 6.5 ꢁ 0.4 ꢂ 10ꢀ5 sꢀ1, at 27 ꢃC),14 pentane
NMR spectroscopy). Complex 3 has been fully characterized
including a single crystal structural study (Fig. 1), and samples
have been compared spectroscopically to independently
prepared material from (PNP)Ti]CHtBu(OTf) and NaOtBu
(Scheme 3).16 In the solid state, unfortunately, the tert-butoxide
and neopentylidene ligands of 3 exhibit positional disorder in the
asymmetric unit that results in the overlap of the two groups and
their tert-butyl moieties. The Ti–O1 and Ti–C18 bond lengths are
therefore not reported due to this occupancy disorder. Regard-
less, the solid state structure of 3 clearly conrms formation of an
alkylidene-alkoxide titanium derivative from the dehydroalkoxy-
lation of tBuOEt.
(kavg ¼ 6.0 ꢁ 0.5 ꢂ 10ꢀ5 sꢀ1, at 29.7 ꢃC) and cyclohexane (kavg
¼
4.97 ꢁ 0.02 ꢂ 10ꢀ5 sꢀ1, at 29.6 ꢃC)16 are also included for
comparison purposes.
Based on the measured rates, we propose a reaction mech-
anism for dehydroalkoxylation that differs from prior studies
involving low-valent Zr(II) complexes known to undergo C–O
bond cleavage of the ether group.13,17,18 For Et2O at least, we
t
propose that (PNP)Ti]CHtBu(CH2 Bu) undergoes a-hydrogen
abstraction to form A, followed by 1,2-CH bond addition of the
A separate competition experiment using equimolar amounts
distal group on the ether (Fig. 2, initial step of the blue pathway)
of Et2O and tBuOMe in the presence of (PNP)Ti]CHtBu(CH2 Bu)
t
also showed simultaneous formation of 1 and 2, with approxi-
t
mately the same minor amount of (PNP)Ti(h2-H2C]CH2)(CH2 Bu)
being formed. Vacuum transfer of volatiles conrmed the pres-
ence of ethene and isobutene based on 1H NMR spectroscopy.16
To probe the mechanism of formation of the olen and the
titanium alkoxide complex, we conducted kinetic measurements
as well as an isotopic labeling study using Et2O-d10. It was found
that formation of 1–3 from (PNP)Ti]CHtBu(CH2 Bu) follows
t
pseudo rst order decay with similar rates (kavg ¼ 7.0 ꢁ 0.2, 5.8 ꢁ
0.2 and 7.5 ꢁ 0.1 ꢂ 10ꢀ5 sꢀ1 at 29.5 ꢁ 0.1 C; duplicate runs)
ꢃ
respectively, which is comparable to the reported rate of decay of
t
(PNP)Ti]CHtBu(CH2 Bu) in C6D6 (kavg ¼ 6.5 ꢁ 0.4 ꢂ 10ꢀ5 sꢀ1) at
27 C, in pentane (kavg ¼ 6.0 ꢁ 0.5 ꢂ 10ꢀ5 sꢀ1), and in cyclo-
ꢃ
hexane (kavg ¼ 4.97 ꢁ 0.02 ꢂ 10ꢀ5 sꢀ1) at 29.6 ꢃC (Table 1).14 In
t
addition, conducting the reaction of (PNP)Ti]CHtBu(CH2 Bu) in
Et2O-d10 at 29.2 ꢃC did not change the overall rate of the reaction
(kavg ¼ 6.4 ꢁ 0.2 ꢂ 10ꢀ5 sꢀ1; KIE ¼ 1.099 ꢁ 0.004 ꢄ 1.1, duplicate
runs), thus implying that activation of ether (e.g., C–H activation)
is not the rate determining step and without signicant
secondary isotope effects. These results in addition to the similar
rates for pseudo-rst order decay suggest that a-hydrogen
abstraction, to give the transient alkylidyne intermediate A from
Fig. 2 Computed reaction coordinates for the competing dehydrogenation and
dehydroalkoxylation of Et2O using intermediate A as the starting point. All
its precursor (PNP)Ti]CHtBu(CH2 Bu), is most likely the overall
t
rate-determining step. Detailed kinetic studies arguing for the energies are indicated with only one decimal.20
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2543–2550 | 2545

View Article Online
Chemical Science
Edge Article
to yield an intermediate such as (PNP)Ti]CHtBu(CH2R) (R ¼ formation of critical amounts of both dehydrogenation and dehy-
CH2OEt; see intermediate B in Fig. 2). The latter rapidly droalkoxylation products advocating that as the chain of ether
undergoes b-alkoxide elimination to furnish compound 1 and becomes longer, formation of the former products becomes more
the corresponding olen (black path in Fig. 2). A more general prominent. As a result, the reaction of nPr2O and nBu2O with
mechanism that applies to all the ethers (branched or linear) (PNP)Ti]CHtBu(CH2 Bu) result in the formation of a-olens, pro-
t
involves competing coordination of the ether functionality to A, pene and 1-butene, along with the alkoxides, (PNP)Ti]CHtBu(OnPr)
followed by concerted cleavage of a distal C–H and a C–O bond (4) and (PNP)Ti]CHtBu(OnBu) (6), as well as the dehydrogenated
t
to form the respective alkoxide and a terminal olen (red alkene products, (PNP)Ti(CH2 Bu)(h2-H2C]CHCH2OnPr) (5) and
t
curve in Fig. 2). In addition, the presence of small amounts of (PNP)Ti(CH2 Bu)(h2-H2C]CHCH2CH2OnBu)
(7),
respectively
t
(PNP)Ti(h2-H2C]CH2)(CH2 Bu) from experiments involving (Scheme 4).
n
Et2O or tBuOEt activations indicates that a separate, competitive
In the case of Pr2O, compounds 4 and 5 were formed in a
dehydrogenation mechanism akin to that of alkanes by A (e.g., 1 : 1 ratio while activation of nBu2O resulted in 6 and 7 in a 1 : 2
ethane, pentane or hexane)15 is also operative for linear ethers ratio as implied by 31P{1H} NMR spectroscopy. Additional
(nal step of the blue path). The absence of vinyl ether type minor unidentiable products were also formed over time in
complexes (namely, (PNP)Ti(h2-H2C]CHOR)(CH2 Bu); R ¼ Et, the latter mixture with consumption of the olen adducts,
t
tBu) in the solution suggests that products of this trans- indicative of their low stability. Overall however, the experi-
formation must undergo facile exchange with ethylene to yield ments show that increasing the chain length of the ether favors
the titanium–ethene adduct 1 and free ether-substituted a- dehydrogenation over dehydroalkoxylation. Worth noting,
olens. Detailed discussion of these reaction pathways is shown complexes 4 and 6 show similar spectroscopic features to the
in Fig. 2 and will be further addressed in this paper.
alkoxide derivatives 1–3 and can be prepared independently and
cleanly from salt metathesis reaction of (PNP)Ti]CHtBu(OTf)
with the corresponding sodium alkoxide (NaOnPr or NaOnBu).16
The olen adducts 5 and 7 exist as a mixture of diastereomers
(Scheme 4). This feature is not unexpected since the analogous a-
Evidence of competing dehydrogenation and
dehydroalkoxylation pathways
To further examine if dehydrogenation was an operative mode
in the aforementioned reactions and whether an olen
adduct could be clearly identied, we explored the reactivity of
t
olen adducts (PNP)Ti(CH2 Bu)(h2-H2C]CHR) (R ¼ nPr or nBu),
which are formed from the dehydrogenation of pentane and
t
hexane by (PNP)Ti]CHtBu(CH2 Bu), exist as a mixtures of dia-
(PNP)Ti]CHtBu(CH2 Bu) with longer-chain linear ethers such
t
stereomers each with their corresponding pair of enantiomers
due to the C2 symmetry of the PNP backbone and one stereogenic
olenic carbon.15
as nPr2O and nBu2O. Activation of the terminal C–H bond of
these substrates would place the ether functionality further away
from the metal site (Scheme 4) and, thus, b-alkoxide elimination
through the black trace shown in Fig. 2 should not interfere
with the dehydrogenation process. Accordingly, treatment of
1
Although complexes 4 and 6 display similar H and 31P{1H}
NMR spectroscopic features to 1–3 (Table S2†),16 complexes 5
and 7 are novel olen adducts of titanium that contain a
pendant ether arm. The latter makes spectroscopic elucidation
less trivial, especially due to the presence of diastereomers.
Additional peak overlap in the 1H NMR spectra rising from the
mixture of dehydrogenation and dehydroalkoxylation products
also complicates assignment. Regardless, 31P{1H} NMR spec-
troscopy is critical for the assignment of these compounds: the
(PNP)Ti]CHtBu(CH2 Bu) with any of these two ethers resulted in
t
2
phosphorus resonances and the magnitude of the JPP of the
PNP backbones in the components of the mixtures being an
extremely useful tool for their characterization. Typically, the
2JPP for titanium-alkoxides is in the order of 48 Hz (Table S2†),16
whereas those of the a-olen-adducts (produced either from
ether or alkane activation) ranges from ꢄ22 to 27 Hz.15
To conclusively establish the formation of such olens, we
resorted to treatment with N2O since oxidation of alkene
adducts derived from alkanes is known to cleanly produce the
t
metastable complex (PNP)Ti]O(CH2 Bu)15 with concomitant
release of the olen. Independently, N2O does not react with the
Ti(IV) alkylidene-alkoxide side-products 4 or 6. Indeed, treat-
ment of a mixture of 4–5 or 6–7 resulted in the immediate and
exclusive consumption of 5 and 7. The titanium-alkoxides
remained unreacted in presence of N2O as expected and
vacuum transfer of the volatiles revealed the existence of the a-
olens 3-propoxy-1-propene and 4-butoxy-1-butene respectively,
based on a combination of GC-MS and 1H NMR spectroscopies
Scheme 4 Dehydroalkoxylation and dehydrogenation of nPr2O and nBu2O by
complex (PNP)Ti]CHtBu(CH2tBu).
2546 | Chem. Sci., 2013, 4, 2543–2550
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
(Scheme 4).15 We were however, unable to unambiguously detect
It is worth noting that unlike iPr2O, when a weakly donating
t
(PNP)Ti]O(CH2 Bu) due to its known thermal instability.14 ether lacking b-hydrogens such as O(SiMe3)2 is used, C–H acti-
Exposure of 4–5 or 6–7 to ethene was also attempted with the idea vation steps involving a-hydrogen abstractions, tautomerization,
t
of promoting olen exchange, since (PNP)Ti(H2C]CH2)(CH2 Bu) a-hydrogen abstraction, and 1,2-CH bond addition overrule the
is known to be one of the most stable olen derivatives15 and dehydroalkoxylation giving rise to the degenerate alkylidene
t
prior observations with Et2O and BuOEt showed this product alkyl, (PNP)Ti]CHSiMe2OSiMe3(CH2SiMe2OSiMe3) along with
t
to form by exchange of the ether-substituted a-olen in two equiv. of CH3 Bu (Scheme 5).14 In context, one possible
t
(PNP)Ti(h2-H2C]CHOEt)(CH2 Bu) (Scheme 1, vide supra). explanation for this result is that a binding event in addition to
t
Indeed, formation of (PNP)Ti(h2-H2C]CH2)(CH2 Bu) was detec- the presence of accessible, internal C–H bonds, are both critical
ted in both mixtures aer exposure to ethene (ꢄ1 atm),16 although for the b-alkoxide elimination pathway to take place.
the reactions are complicated byformation ofother sideproducts,
presumably given the reactive nature of the ether substituted
Computational evidence for competing dehydrogenation and
dehydroalkoxylation pathways
a-olens. In context, our results validate that it is possible for a
complex with formula, (PNP)Ti(h2-H2C]CH(CH2)nOR)(CH2 Bu)
t
(n ¼ 0, 1, 2; R ¼ Et, nPr, nBu) to undergo exchange with ethene in
To gain insight into the mechanisms leading to b-alkoxide
elimination and dehydrogenation in the aforementioned reac-
order to release the ether substituted a-olen. Unfortunately, we
tions we examined several pathways for the transformations of
have yet to develop an independent synthesis of 5 or 7 in order to
Et2O and nPr2O using DFT calculations.16 Fig. 2 depicts the three
fully characterize these species spectroscopically.
t
Monitoring the decay of (PNP)Ti]CHtBu(CH2 Bu) in nPr2O
most plausible mechanistic scenarios for the competing dehy-
and nBu2O by 31P NMR spectroscopy at ꢄ29 C reveals the
ꢃ
drogenation and dehydroalkoxylation of Et2O and in Fig. 3, we
present the three most likely reaction coordinates involving a
following average rate constants, kavg: 5.9 ꢁ 0.5 ꢂ 10ꢀ5 and 5.0 ꢁ
0.0 ꢂ 10ꢀ5 sꢀ1 at 29.5 ꢃC (Table 1), again suggestive of formation
of A being overall rate-determining.16 As a result, the development
n
longer ether chain such as Pr2O.16
The dehydrogenation of Et2O (blue trace) follows a mechanism
analogous to that of ethane to ethene:15 intermediate A and free
ether rst form a s-complex (A-EtOEt), followed by the 1,2-addi-
tion of the terminal C–H bond across the reactive Ti^C linkage
(A-EtOEt-TS). The latter gives rise to the alkyl-neopentylidene inter-
mediate (B),19 which nally allows the formation of the dehydro-
t
of 4–5 and 6–7 from (PNP)Ti]CHtBu(CH2 Bu) and the ethers
nPr2O and nBu2O respectively, must involve competing routes for
C–O and C–H bond activations leading into dehydroalkoxylation
and dehydrogenation via the common intermediate A. The latter is a
recurrent theme for long alkanes or as in this case, long-chain
linear ethers. Also, since 4–5 or 6–7 do not interconvert under
genated vinyl ether adduct (PNP)Ti(h2-H2C^CHOEt)(CH2 Bu) (D)
t
normal conditions or with mild heating (65 C),16 these results
ꢃ
through a concerted metal mediated b-hydrogen migration step
(B-TS). The latter step is associated with a relatively high barrier of
validate the conclusion that the products originate from two
independent routes stemming from the same intermediate. To
29.8 kcal molꢀ1
.
t
further explore this idea, we treated (PNP)Ti]CHtBu(CH2 Bu)
The generic dehydroalkoxylation pathway in Fig. 2 (red trace)
also begins with an adduct involving the alkylidyne A. However,
in this adduct the ether binds to the titanium center with its
i
with a sterically hindered ether such as Pr2O, since dehydroge-
nation was predicted to be disfavored due to the steric hindrance
t
of the isopropyl groups. Indeed, when (PNP)Ti]CHtBu(CH2 Bu)
i
was treated with Pr2O, only (PNP)Ti]CHtBu(OiPr) (8) and pro-
pene (1H NMR) were formed, therefore suggesting the dehydro-
genation pathway to be completely shut down (Scheme 5). The
t
i
rate of decay of (PNP)Ti]CHtBu(CH2 Bu) in Pr2O at 29.6 ꢃC
(kavg ¼ 6.0 ꢁ 0.5 ꢂ 10ꢀ5 sꢀ1) is comparable with all the previous
ones for different ethers and even alkanes (Table 1), thus con-
rming that dehydroalkoxylation is denitely dependent on the
formation of A.
Fig. 3 Computed reaction profiles for the competing dehydrogenation and
dehydroalkoxylation of nPr2O using intermediate A as the starting point. All
energies are indicated with only one decimal.20
Scheme 5 Reactivity of (PNP)Ti]CHtBu(CH2tBu) with iPr2O and O(SiMe3)2.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2543–2550 | 2547

View Article Online
Chemical Science
Edge Article
t
oxygen (A-OEt2). A b-hydrogen transfer from the ethyl moiety to the reaction of (PNP)Ti]CHtBu(CH2 Bu) with Et2O, could
the neopentylidyne ligand concomitant with a concerted C–O undergo exchange with ethene. Our ndings by DFT
bond breaking yields product 1 and free ethene, traversing the suggest that compound D can undergo exchange with ethene
transition state (A-OEt2-TS) with a relative free energy in solu- via
a
dissociative process involving an
S
¼
1 system
tion of 21.0 kcal molꢀ1. Apart from the latter pathway having a (PNP)Ti(CH2 Bu). Notably, it was also found that the ethene
t
lower barrier with respect to dehydrogenation, the product of adduct (PNP)Ti(h2-H2C]CH2)(CH2 Bu) is ꢀ9.6 kcal molꢀ1 lower
t
dehydroalkoxylation, compound 1 is signicantly more ther- in energy than D (Fig. S67†).16 Therefore, ethene exchange in D
modynamically stable than the dehydrogenated product D by as well as in the ether substituted olen complex 5 and 7 should
nearly 40 kcal molꢀ1
In the cases when the alkyl chain of the reacting ether
.
be all thermodynamically favorable steps.
i
t
contains two carbons, that is R ¼ Et, Pr or Bu, these two
pathways are accompanied by a third possible side-reaction
(black trace in Fig. 2) that also leads to 1 and free ethene. From
the C–H activated intermediate B, the putative quasi-octahedral
intermediate C can also be formed by a simple chelation of the
ether moiety. Structure C is a prerequisite for the simultaneous
breaking of the O–C and Ti–C bonds, by which ethene is
extruded from the complex and 1 is formed. Since the latter
pathway involves only two moderate barriers of about 18 kcal
molꢀ1, one for the C–H bond activation and one for the ethene
extrusion step, this pathway provides the energetically most
advantageous channel for the transformation of ethers in the
case of R ¼ Et, Pr or Bu. In addition, Fig. 2 illustrates why
dehydrogenation is not competitive when ethers with short
alkyl chain are involved in the activation: the barriers corre-
sponding to C-TS and B-TS differ by 12 kcal molꢀ1, with the
former being lower, thereby favoring formation of 1 over D in
line with our experimental results.20
Conclusions
In conclusion, we have shown a new mechanistic scenario for
the dehydroalkoxylation and dehydrogenation of ethers. By
studying a series of linear and branched ethers we concluded
the following: (i) the alkylidyne functionality in A binds ethers
and preferentially activates the O–CH2CH2R hydrogen as
opposed to the more exposed OMe group,21 (ii) binding of the
ether to A via its oxygen most likely initiates the dehy-
droalkoxylation pathway, (iii) for short chain ethers dehy-
droalkoxylation can also occur aer one of the terminal C–H
bonds was split by the Ti^CtBu functionality, (iv) the longer the
linear ether chain the more unfavourable the dehydroalkoxy-
lation becomes, whereas the process described in (ii) becomes
non-operative, and (v) a-hydrogen migration is the last alter-
native to other pathways and it is only favoured if the substrate
has no b-hydrogens or no available binding sites.22
i
t
Note that generation of 1 from B, i.e. the black pathway
shown in Fig. 2, is not operative when the alkyl chain of the ether is
longer than two carbons. In these cases traversing C-TS would
result in the formation of a high-energy radical species, where
the terminal carbon holds the unpaired electrons (Schemes S1
and S2†).16 Thus dehydrogenation and dehydroalkoxylation
become more equally competitive for ethers with longer alkyl
chains mostly because the stepwise mechanism leading to 1,
shown in black trace, is the least energetically accessible for
nPr2O because it costs a penalty of 25.9 kcal molꢀ1 for 1,2-CH
addition to take place at the b-position of the ether as indicated
by A-PrOPr-TSb (Fig. 3). When compared to Fig. 2, the barrier
corresponding to the O-bound ether, A-OEt2-TS, is slightly
higher (A-OnPr2-TS, red trace), whereas the barrier of B-TS
becomes lower in energy (15.9 kcal molꢀ1) with longer alkyl
chains. In other words, the red and blue traces shown in Fig. 3
become more competitive according to our calculations.20
Maximal barriers of 23.6 and 27.5 kcal molꢀ1 were calculated for
the dehydroalkoxylation and dehydrogenation of nPr2O,
respectively, which is in good agreement at least qualitatively
(within computational error) with the experimental observa-
tions that 4 and 5 are formed in almost equal amounts. Also
noteworthy is the similarity in the transition state energies
involving the O-bound ether and the b-1,2-CH bond addition of
the ether across the alkylidyne linkage (DDG‡ ¼ 2.3 kcal molꢀ1),
which suggest that either pathway is viable to propose in the
dehydroalkoxylation/dehydrogenation of linear ethers.20
Acknowledgements
Financial support of this research was provided by the National
Science Foundation (CHE-0848248 and CHE-0645381 to D.J.M.),
and National Science Foundation American Competitiveness in
Chemistry Fellowship (CHE-1137284 to S.F.). M.G.C. acknowl-
edges CONACYT for a postdoctoral fellowship.
Notes and references
1 W. Heitmann, G. Strehlke and D. Mayer, Ullmann's
Encyclopedia of Industrial Chemistry, Wiley-VCH, Federal
Republic of Germany, 6th edn, 2003, vol. 12, p. 505;
L. Karas and W. J. Piel, Kirk-Othmer Encyclopedia of
Chemical Technology, ed. J. I. Kroschwitz and M. Howe-
Grant, John Wiley & Sons, USA, 4th edn, 1994, vol. 9, p. 860.
2 For Et2O, the energies of the a- and b-C–H bonds are 89 and
112 kcal molꢀ1. For butane, the internal and terminal C–H
bonds are 97 and 98 kcal molꢀ1
K. A. Miller, J. M. Bartolin, R. M. O'Neill, R. D. Sweeder,
T. M. Owens, J. W. Kampf, M. M. Banaszak Holl and
N. J. Wells, J. Am. Chem. Soc., 2003, 125, 8986.
3 C. Decker and D. J. Decker, J. Macromol. Sci., Part A: Pure
Appl.Chem., 1997, 34, 605.
4 A. C. Atesin, N. A. Ray, P. C. Stair and T. J. Marks, J. Am. Chem.
Soc., 2012, 134, 14682; Y. Ogiwara, T. Kochi and F. Kakiuchi,
Org. Lett., 2011, 13, 3254.
, respectively. See:
Lastly, we examined whether the vinyl ether complex D,
which we propose is formed in very small quantities in
5 B. A. Arndtsen and R. G. Bergman, Science, 1995, 270, 1970.
2548 | Chem. Sci., 2013, 4, 2543–2550
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Edge Article
Chemical Science
6 M. T. Whited, Y. Zhu, S. D. Timpa, C.-H. Chen, B. M. Foxman, 11 For relevant examples of C–H activation using early transition
O. V. Ozerov and R. H. Grubbs, Organometallics, 2009, 28,
4560.
metals,
emphasizing
alkane
metathesis
reactions:
´
D. Soulivong, S. Norsic, M. Taouk, C. Coperet, J. Thivolle-
7 N. M. West, P. S. White and J. L. Templeton, J. Am. Chem.
Soc., 2007, 129, 12372.
Cazat, S. Chakka and J.-M. Basset, J. Am. Chem. Soc., 2008,
´
130, 5044; C. Thieuleux, A. Maraval, L. Veyre, C. Coperet,
8 Examples of d8 metals in dehydrogenation reactions of
ethers and close derivatives. M. T. Whited and
R. H. Grubbs, Organometallics, 2009, 28, 161; M. T. Whited
and R. H. Grubbs, Acc. Chem. Res., 2009, 42, 1607;
M. T. Whited and R. H. Grubbs, J. Am. Chem. Soc., 2008,
130, 5874; P. E. Romero, M. T. Whited and R. H. Grubbs,
Organometallics, 2008, 27, 3422; M. T. Whited and
R. H. Grubbs, Organometallics, 2008, 27, 5737;
M. T. Whited and R. H. Grubbs, J. Am. Chem. Soc., 2008,
130, 16476; J. S. Owen, J. A. Labinger and J. E. Bercaw, J.
D. Soulivong, J.-M. Basset and G. J. Sunley, Angew. Chem., Int.
´
Ed., 2007, 46, 2288; D. Soulivong, C. Coperet, J. Thivolle-
Cazat, J.-M. Basset, B. M. Maunders, R. B. A. Pardy and
´
G. J. Sunley, Angew. Chem., Int. Ed., 2004, 43, 5366; C. Coperet,
O. Maury, J. Thivolle-Cazat and J.-M. Basset, Angew. Chem.,
´
Int. Ed., 2001, 40, 2331; V. Vidal, A. Theolier, J. Thivolle-Cazat
and J.-M. Basset, Science, 1997, 276, 99. For recent reviews on
alkane metathesis: V. N. Cavaliere and D. J. Mindiola, Chem.
Sci., 2012, 3, 3356; V. N. Cavaliere, B. F. Wicker and
D. J. Mindiola, Adv. Organomet. Chem., 2012, 60, 1.
Am. Chem. Soc., 2006, 128, 2005; E. Carmona, M. Paneque 12 X. Gao and S. D. R. Johnston, US Pat., 7,700,817.
and M. L. Poveda, Dalton Trans., 2003, 4022; 13 C. A. Bradley, L. F. Veiros and P. J. Chirik, Organometallics,
M. W. Holtcamp, L. M. Henling, M. W. Day, J. A. Labinger
and J. E. Bercaw, Inorg. Chim. Acta, 1998, 270, 467;
M. W. Holtcamp, J. A. Labinger and J. E. Bercaw, J. Am.
2007, 26, 3191; C. A. Bradley, L. F. Veiros, D. Pun,
E. Lobkovsky, I. Keresztes and P. J. Chirik, J. Am. Chem.
Soc., 2006, 128, 16600.
Chem. Soc., 1997, 119, 848; H. F. Luecke, B. A. Arndtsen, 14 J. A. Flores, V. N. Cavaliere, D. Buck, G. Chen, M. G. Crestani,
P. Burger and R. G. Bergman, J. Am. Chem. Soc., 1996, 118,
2517.
B. Pinter, M.-H. Baik and D. J. Mindiola, Chem. Sci., 2011, 2,
1457; A. R. Fout, B. C. Bailey, D. Buck, H. Fan, J. C. Huffman,
M.-H. Baik and D. J. Mindiola, Organometallics, 2010, 29,
5409; A. R. Fout, J. Scott, D. L. Miller, B. C. Bailey,
J. C. Huffman, M. Pink and D. J. Mindiola, Organometallics,
2009, 28, 331; B. C. Bailey, A. R. Fout, H. Fan,
J. Tomaszewski, J. C. Huffman and D. J. Mindiola, Angew.
Chem., Int. Ed., 2007, 46, 8246; A. R. Fout, B. C. Bailey,
J. Tomaszewski and D. J. Mindiola, J. Am. Chem. Soc., 2007,
129, 12640; B. C. Bailey, H. Fan, J. C. Huffman, M.-H. Baik
and D. J. Mindiola, J. Am. Chem. Soc., 2007, 129, 8781;
B. C. Bailey, A. R. Fout, H. Fan, J. C. Huffman, J. B. Gary,
M. J. A. Johnson and D. J. Mindiola, J. Am. Chem. Soc., 2007,
129, 2234; B. C. Bailey, J. C. Huffman and D. J. Mindiola, J.
Am. Chem. Soc., 2007, 129, 5302; B. C. Bailey, H. Fan,
J. C. Huffman, M.-H. Baik and D. J. Mindiola, J. Am. Chem.
Soc., 2006, 128, 6798; B. C. Bailey, H. Fan, E. W. Baum,
J. C. Huffman, M.-H. Baik and D. J. Mindiola, J. Am. Chem.
Soc., 2005, 127, 16016.
9 E. Carmona, M. Paneque, L. L. Santos and V. Salazar, Coord.
Chem. Rev., 2005, 249, 1729; G. Ferrando-Miguel,
J. N. Coalter, H. Gerard, J. C. Huffman, O. Eisenstein and
K. G. Caulton, New J. Chem., 2002, 26, 687; D.-H. Lee,
J. Chen, J. W. Faller and R. H. Crabtree, Chem. Commun.,
2001, 213; C. Slugovc, K. Mereiter, S. Tromenko and
E. Carmona, Angew. Chem., Int. Ed., 2000, 39, 2158;
J. N. Coalter, G. Ferrando and K. G. Caulton, New J. Chem.,
2000, 24, 835.
10 For reviews on iridium catalyzed alkane dehydrogenation of
hydrocarbons. M. C. Haibach, S. Kundu, M. Brookhart and
A. S. Goldman, Acc. Chem. Res., 2012, 45, 947;
R. H. Crabtree, Top. Organomet. Chem., 2011, 34, 1;
C. Jongwook and A. S. Goldman, Top. Organomet. Chem.,
2011, 34, 139; G. E. Dobereiner and R. H. Crabtree, Chem.
Rev., 2010, 110, 681; D. Morales-Morales, Iridium
Complexes in Organic Synthesis, ed. L. A. Oro and C. Claver,
Wiley-VCH, New York, 2009, p. 325; A. S. Goldman and 15 V. N. Cavaliere, M. G. Crestani, B. Pinter, M. Pink,
R. Ghosh, in Handbook of C–H Transformations
Applications in Organic Synthesis, ed. G. Dyker, Wiley-VCH,
–
C.-H. Chen, M.-H. Baik and D. J. Mindiola, J. Am. Chem.
Soc., 2011, 133, 10700.
New York, 2005, p. 616; A. S. Goldman and K. I. Goldberg, 16 See ESI.†
Activation and Functionalization of C–H Bonds, ed. K. I. 17 For additional selected studies on b-alkoxide elimination
Goldberg and A. S. Goldman, vol. 885, ACS Symposium
Series, 2004, p. 1; K. Krogh-Jespersen, M. Czerw and
A. S. Goldman, J. Mol. Catal. A: Chem., 2002, 189, 95;
H.-J. Fan and M. B. Hall, J. Mol. Catal. A: Chem., 2002, 189,
111; E. Odile and R. H. Crabtree, New J. Chem., 2001, 25,
665; R. H. Crabtree, J. Chem. Soc., Dalton Trans., 2001,
2437; J. A. Maguire, W. T. Boese, M. E. Goldman and
A. S. Goldman, Coord. Chem. Rev., 1990, 97, 179. For
additional relevant reviews, consult: J.-M. Basset,
reactions using Zr(IV) compounds. R. Hara, Y. Ura, S. Huo,
K. Kasai, N. Suzuki and T. Takahashi, Inorg. Chim. Acta,
2000, 300–302, 741; H. Ito, T. Nakamura, T. Taguchi and
Y. Hanzawa, Tetrahedron, 1995, 51, 4507; A. F. Houri,
M. T. Didiuk, Z. Xu, N. R. Horan and A. H. Hoveyda, J. Am.
Chem. Soc., 1993, 115, 6614; A. H. Hoveyda and J. P. Morken,
J. Org. Chem., 1993, 58, 4237; G. D. Cuny and S. L. Buchwald,
Organometallics, 1991, 10, 363; S. Karlsson, A. Hallberg and
S. Gronowitz, J. Organomet. Chem., 1991, 403, 133;
S. L. Buchwald, R. B. Nielsen and J. C. Dewan,
Organometallics, 1988, 7, 2324; G. Erker, U. Dorf, J. L. Atwood
and W. E. Hunter, J. Am. Chem. Soc., 1986, 108, 2251.
´
C. Coperet, D. Soulivong, M. Taouk and J. Thivolle Cazat,
Acc. Chem. Res., 2010, 43, 323; J. A. Labinger and
J. E. Bercaw, Nature, 2002, 417, 507.
This journal is ª The Royal Society of Chemistry 2013
Chem. Sci., 2013, 4, 2543–2550 | 2549

View Article Online
Chemical Science
Edge Article
18 For examples of low-valent metals undergoing 20 DFT methods might be signicantly overestimating some of
intermolecular C–O oxidative addition. S. Kundu, J. Choi,
D. Y. Wang, Y. Choliy, T. J. Emge, K. Krogh-Jespersen and
A. S. Goldman, J. Am. Chem. Soc., ASAP article online, DOI:
10.1021/ja312464b; J. Choi, Y. Choliy, X. Zhang, T. J. Emge,
K. Krogh-Jespersen and A. S. Goldman, J. Am. Chem. Soc.,
these barriers due to several reasons including tunneling
effects. In addition, since solvent molecules are activated,
there might be an overestimate of kinetic entropy loss upon
association when applying statistical thermodynamics on
the gas phase reaction.
2009, 131, 15627; J. B. Bonanno, T. P. Henry, 21 We cannot refute the possibility of the MeO group being
t
D. R. Neithamer, P. T. Wolczanski and E. B. Lobkovsky, J.
Am. Chem. Soc., 1996, 118, 5132; C. A. Tolman, S. D. Ittel,
A. D. English and J. P. Jesson, J. Am. Chem. Soc., 1979, 101,
1742; S. D. Ittel, C. A. Tolman, A. D. English and
J. P. Jesson, J. Am. Chem. Soc., 1978, 100, 7577.
reversibly activated, since treatment of (PNP)Ti]CHtBu(CH2 Bu)
with Me2O results in a myriad of products none of
which can be isolated in pure form. A computed barrier
of 20.7 kcal molꢀ1 indicates that the activation of the
C–H bond of methyl group in
a molecule such as
19 Attempts to observe formation of intermediate B in solution
were attempted unsuccessfully. Follow-up of all neat ether
activation reactions by 1H and 31P NMR spectroscopy,
CH3OCH2CH3 is indeed likely. However such barrier is
ꢄ3 kcal molꢀ1 higher than activation of the C–H bond
of ethyl group. In addition, no b-alkoxide elimination or
b-H abstraction is possible aer the OMe is activated.
t
showed only consumption of (PNP)Ti]CHtBu(CH2 Bu)
with concomitant formation of the dehydroalkoxylation or 22 1,2-CO bond addition is observed with (PNP)Ti^CtBu and
dehydrogenation products (see ESI†).
CH3OC6F5. See ref. 14.
2550 | Chem. Sci., 2013, 4, 2543–2550
This journal is ª The Royal Society of Chemistry 2013