E-Olefins through intramolecular radical relocation

Article abstract of DOI:10.1126/science.aav1610

Full control over the selectivity of carbon-carbon double-bond migrations would enable access to stereochemically defined olefins that are central to the pharmaceutical, food, fragrance, materials, and petrochemical arenas. The vast majority of double-bond migrations investigated over the past 60 years capitalize on precious-metal hydrides that are frequently associated with reversible equilibria, hydrogen scrambling, incomplete E/Z stereoselection, and/or high cost. Here, we report a fundamentally different, radical-based approach.We showcase a nonprecious, reductant-free, and atom-economical nickel (Ni)(I)-catalyzed intramolecular 1,3-hydrogen atom relocation to yield E-olefins within 3 hours at room temperature. Remote installations of E-olefins over extended distances are also demonstrated.

Full text of DOI:10.1126/science.aav1610

RESEARCH
ciated cost (19). By contrast, nickel is nonprecious,
relatively cheap and abundant (~90,000 times
more than Ru), and capable of triggering double-
bond migration, albeit in varying E/Z ratios and
so-far limited generality (20). This reactivity has
been linked to Ni(II)-H as active species.
ORGANIC CHEMISTRY
E-Olefins through intramolecular
radical relocation
We envisioned that a simplified mechanistic
scenario that minimizes co-reagents and inter-
mediates may increase the likelihood for E/Z
control under nonprecious Ni catalysis in double-
bond migrations. Arguably, the simplest scenario
involves the direct delivery of a hydrogen from
position 1 to 3 in the allylic framework. Given
that radicals are not subject to the Woodward–
Hoffmann orbital symmetry restrictions, we hypo-
thesized that a radical-induced 1,3-H shift could
deliver a solution to the E/Z selection challenge.
The intermediacy of high-energy free organic
radicals would need to be avoided in this con-
text, as additional challenges, such as inter-
molecular H atom transfer events for radical
termination and chain propagation, would arise,
and these scenarios are known to deliver mod-
erate selectivities (see Fig. 1A). We envisioned
instead that if the substrate were coupled to a
nonprecious metalloradical complex, such as a
low-coordinate Ni(I) (21)—e.g., via p-coordination—
the resulting metalloradical-substrate complex
might have an inherent driving force for inter-
nal reorganization.
As part of our ongoing program in odd-
oxidation-state metal reactivity, we also studied
the propensity of dimeric metal complexes in
the +1 oxidation state to generate open-shell
monomers (22). Our findings suggest that for
certain complexes of suitable ligand environment,
such as the dinuclear N-heterocyclic carbene
(NHC)–derived Ni(I) dimer 1 (Fig. 2) (23), which
is readily accessible at low cost from commer-
cially available precursors, a double bond or solvent
could function as a trigger to generate an open-
shell metal complex.
Ajoy Kapat, Theresa Sperger, Sinem Guven, Franziska Schoenebeck*
Full control over the selectivity of carbon–carbon double-bond migrations would enable
access to stereochemically defined olefins that are central to the pharmaceutical, food,
fragrance, materials, and petrochemical arenas. The vast majority of double-bond
migrations investigated over the past 60 years capitalize on precious-metal hydrides that
are frequently associated with reversible equilibria, hydrogen scrambling, incomplete E/Z
stereoselection, and/or high cost. Here, we report a fundamentally different, radical-based
approach. We showcase a nonprecious, reductant-free, and atom-economical nickel
(Ni)(I)-catalyzed intramolecular 1,3-hydrogen atom relocation to yield E-olefins within
3 hours at room temperature. Remote installations of E-olefins over extended distances
are also demonstrated.
he carbon–carbon double bond in olefins
serves as a precursor to a rich array of
transformations and is a cornerstone in
the materials, pharmaceutical, agrochemical
arenas, and food industry (1, 2). Its con-
and generates stoichiometric amounts of waste.
Metal catalysis can potentially circumvent these
drawbacks (see below) and may also unleash the
possibility for remote functionalization. As such,
metal-catalyzed alkene isomerizations have had
major societal and industrial impact, e.g., in the
large-scale production of gasoline, nylon, deter-
gents, soaps, coatings, lubricants, food, cosmetics,
rubbers, and fragrances (2).
T
struction in a selective and stereochemically
defined manner—i.e., E versus Z olefin—is of
utmost importance, as the geometry is ulti-
mately coupled to function. Although numerous
strategies to construct olefins have become es-
tablished textbook knowledge, the E/Z selectivity
is frequently incomplete or comes at the expense
of valuable functionality in these classical ap-
proaches (e.g., Wittig, Julia, Peterson olefinations
or Birch reduction). Mixtures of E and Z isomers
are difficult to separate, however. More modern
catalytic strategies commonly achieve high se-
lectivity through semihydrogenations, requiring
an atmosphere of H2 or stoichiometric amounts
of acid or other H sources (3). Olefin metathesis
catalysts were developed to selectively access
Z-olefins, whereas the E-isomer is accessible in
high selectivity only with certain halogenated
or low-functionality compounds (4, 5). Ring-
opening strategies via C–C cleavage are ele-
gant alternatives to selectively install E-olefins
(6, 7).
The past 60 years of research focused pri-
marily on two mechanisms for double-bond
transposition (11): The vast majority of trans-
formations proceed via metal hydrides. Whereas
the precious-metal hydrides deliver an external
H through closed-shell polar mechanisms (9, 11),
less-precious metal hydrides, such as Co, have
been shown to engage in reversible hydrogen
atom transfer via biradical pairs (Fig. 1A) (12, 13).
These processes make use of external H (that is
present as co-reagent). Because of the revers-
ible nature of these reactions, scrambling of
hydrogens may occur, and control of E/Z se-
lectivity is challenging. The E-selectivity is gen-
erally much higher for the precious-metal hydride
processes that proceed through 1,2-addition,
followed by b-hydride elimination (Fig. 1A) (9). The
latest advance in this area is from Skrydstrup’s
laboratory employing in situ–generated Pd-H at
elevated temperatures (14). The possibility for
hydride-free binuclear Pd-catalyzed isomeri-
zation was recently also reported (15). An al-
ternative mechanistic scenario involves the less
well-described p-allyl mechanism (Fig. 1B) that
proceeds by means of 1,3-addition (11). Examples
include isomerizations involving zirconocenes
(Bu2ZrCp2) (16) and Grotjahn’s highly active Ru
catalyst that delivers E-olefins with excellent
selectivities via a hemilabile ligand that acts
as a base (17).
When we subjected [Ni(m-Cl)IPr]2 dimer 1 to
4-methoxyallylbenzene 2, selective generation
of the corresponding isomerized E-alkene 3 was
observed. After evaluating different reaction pa-
rameters, we found that E-olefin 3 was gen-
erated quantitatively and exclusively (24) within
3 hours at room temperature in chlorobenzene
(Fig. 2A).
Double-bond shifts of allyl precursors con-
stitute a potentially powerful alternative strat-
egy, as the allyl group can be installed readily,
and the desired E-geometry may subsequently
be set in a catalytic and atom-economical trans-
formation (8, 9, 10). This goal cannot be reached
through classical synthetic approaches, as direct
sigmatropic hydrogen shifts are geometrically
inaccessible in the required antarafacial sense.
Although base may be used to mediate double-
bond migration [at ~200°C, as used on an in-
dustrial scale (9)], the E/Z ratios of the resulting
double bonds are modest (~80:20), and the pro-
cess is incompatible with sensitive functionality
To gain further insight, we subjected Ni(I)
catalyst 1 to a mixture of substrate 4 and the
deuterium-labeled 5-d1 (containing >98% deu-
terium enrichment). We did not observe isotopic
scrambling and instead formed 6 and 7-d1 as
the exclusive products, regardless of whether the
reaction was run in deuterated or nondeuterated
solvent (Fig. 2B and fig. S11 to S14). These data
provide strong support that the solvent does not
act as a source of hydrogen/deuterium and that
the 1,3-H shift proceeds exclusively intramolecu-
larly. Our further investigation of the catalytic
reaction with paramagnetic proton nuclear mag-
netic resonance (NMR) and electron paramag-
netic resonance (EPR) spectroscopic analyses
showed clear indications of radicals being
present (see figs. S19 to S22 for details). We
In light of the ever-increasing demands for
greater sustainability, particularly for processes
of societal and industrial relevance, a key chal-
lenge in this arena is to overcome the need for
precious-metal catalysts and/or stoichiometric
additives (and waste) (9, 18). Ru has an Earth
abundance of only 10−7% and a considerable asso-
Institute of Organic Chemistry, RWTH Aachen University,
Landoltweg 1, 52074 Aachen, Germany.
*Corresponding author. Email: franziska.schoenebeck@rwth-aachen.de
Kapat et al., Science 363, 391–396 (2019)
25 January 2019
1 of 5

RESEARCH
| REPORT
therefore next performed a radical clock exper-
iment and subjected catalyst 1 to cis-diphenyl-
cyclopropenyl olefin 8. The stereochemical integrity
of 8 is highly sensitive to transient radical char-
acter (25). Although free organic phenylcyclo-
propylcarbinyl radicals are known to undergo
rapid ring opening (26), this is not necessarily
expected for the Ni(I)-induced radical charac-
ter at the olefin. Our calculations suggest the
opening to be endergonic and hence revers-
ible under Ni(I) coordination. Our 1H-NMR
spectroscopic analysis indicated that 28% of
the trans-olefin (trans-8) was formed in a
mixture with the cis-isomer, which is a strong
indication of radical intermediates. All our ad-
ditional tests with probe 8 and alternative
metal species that are known to isomerize
double bonds or could potentially be derived
from Ni(I) dimer 1 [i.e., Ni(II) or Ni(0)] did not
show cis-to-trans isomerization (see Fig. 2B,
fig. S18, and table S1).
computational studies (27). The bulky Ni(I)
dimer was found to be too shielded to allow
for direct reactivity with the alkene. Instead, a
substrate- or solvent-assisted homolytic cleavage
of the dimer to a substrate-Ni(I) monomer com-
plex is favored. This generates a metal-substrate
complex with overall radical character (Fig. 2C).
A relatively facile H atom transfer to the Ni
center then takes place, generating a shallow
intermediate that rapidly relocates the H atom
to the organic moiety so as to generate the more
stable Ni-substrate radical complex. No move-
ment of the metal occurs while the H atom
migrates. The pathway leading to the E-olefin is
calculated to be favored (by DDG‡ = 6.6 kcal/mol)
over the Z-pathway. These findings are in line
with the excellent E-selectivity, EPR-activity,
and the observed lack of H/D scramblings, as
no stable intermediates are generated and,
therefore, no sufficient time is available for ex-
change processes. This mechanism is therefore
in stark contrast to established metal-hydride
reactivity.
With the fundamentally different mechanism
established, we set out to explore the generality
of the process. Our isomerization shows a wide
scope with short reaction times and excellent
selectivities (Fig. 3). We succeeded in efficiently
isomerizing the methoxyarene-derived anethole,
eugenol, and safrole in excellent yield and
E-selectivity (>98:2) at room temperature to the
corresponding trans-products (3, 6, 9 to 11).
These substrate classes are in considerable in-
dustrial demand, with anethole alone being
produced on a million-ton scale per year.
Our protocol is efficient toward substrates
with electron-withdrawing functional groups
such as ketone or ester (14, 15), the pharmaceu-
tically relevant trifluoromethyl, and potentially
coordinating amine groups (16, 17). Functional
groups that are frequently reactive in typical
precious-metal catalysis—e.g., nitrile (20), bromide
(13), or boronic ester (23)—were also tolerated
(Fig. 3). These groups are of utmost value for
further orthogonal manipulations and build-up
of molecular complexity. Multiple migrations
With these experimental data in strong sup-
port of a radical mechanism, we turned to
Fig. 1. Prevalence of E-olefins and isomerization-based methods to access them. Top: Importance of E-alkene in pharma, food, and fragrance
industries and synthesis (36–38). Middle (A and B) Known mechanisms for metal-catalyzed isomerization. (C) This work. EU, European Union.
Kapat et al., Science 363, 391–396 (2019)
25 January 2019
2 of 5

RESEARCH
| REPORT
can also be executed in the same molecule (25).
Moreover, selected oxygen-containing substrates
(14, 34) reacted in higher yields in the presence
of a catalytic amount of Ti(OiPr)4.
The migration proceeds equally selectively
and efficiently at a 1-g scale (with 13), and the
product is amenable to straightforward puri-
fication. A quantitative conversion is seen in
only 3 hours at room temperature at 5 mol %
catalyst loading (28); however, a lowering to
1 to 2 mol % is also possible but requires longer
time at room temperature (24 hours). The
E-selectivity remains unchanged and is inde-
pendent of the catalyst loading. Moreover, for
those substrates that are liquids, the reactions
may also be run solvent-free. In this context,
we isomerized neat estragole (a naturally oc-
curring liquid) to the industrially produced
anethole 3 in >99:1 E/Z selectivity [>99% con-
version, 96% yield] in 16 hours using 5 mol %
of catalyst. Considering that solely the nonpre-
cious catalyst and the substrate are required,
the process should have considerable potential
also in an industrial context.
Fig. 2. Ni(I)-catalyzed 1,3-H atom shift. (A) Test reaction. (B) Experimental mechanistic studies. (C) Computational studies. Free-energy diagram in
kcal/mol, calculated at CPCM(chlorobenzene) M06L/def2-TZVP//wB97XD/6-31G(d)(SDD) at room temperature (27). cat., catalyst; TS, transition state.
Kapat et al., Science 363, 391–396 (2019)
25 January 2019
3 of 5

RESEARCH
| REPORT
Fig. 3. Scope of 1,3-H atom shift. Conditions: 1 (5 mol %), olefin (1.0 equiv.) in chlorobenzene (0.4 M) (except 30 in CH2Cl2) (24), *10 mol % Ti(OiPr)4,
†48 hours, ‡10 mol % catalyst, §20 mol % catalyst. ¶Small amount (≤5%) of inseparable alternative isomer was also detected. #Yield determined by
1H-NMR against added internal standard. conv., conversion; equiv., equivalent; rt, room temperature.
Another challenge in double-bond migrations
is application to higher substitution patterns,
such as disubstituted and cyclic olefins (29). Our
Ni(I) radical-based isomerization proved to be
efficient to transform acyclic 1,1-disubstituted
olefins (27) and to afford various cyclic olefin
derivatives containing functional groups (-Cl, -OMe,
-OTf) or heteroatoms (34) with high yield and
exclusive regioselectivity (28, 29, 32 to 34).
Only for 30 and 31 could small amounts (≤5%)
of an alternative isomer be detected by gas
chromatography–mass spectrometry (GC-MS).
Although chlorobenzene generally proved to be
an excellent solvent, it may be inconvenient for
the preparation of volatile products, such as 30.
Kapat et al., Science 363, 391–396 (2019)
25 January 2019
4 of 5

RESEARCH
| REPORT
Fig. 4. Chain walking. Conditions: 0.4 M olefin in chlorobenzene (or *in CH2Cl2), 5 mol % 1 (†10 mol %). ‡Small amount (≤5%) of inseparable alternative
1
internal isomer was also detected. §Yields were determined by H-NMR against added internal standard (24).
We found dichloromethane to be a suitable al-
ternative in this case, allowing 30 to be isolated
in 83% yield.
Boy Cornils, Wolfgang A. Herrmann, Matthias Beller, Rocco
Paciello, Eds. (Wiley, 2017), pp. 1365–1378.
3. A. Fürstner, J. Am. Chem. Soc. 141, 11–24 (2019).
4. T. T. Nguyen et al., Science 352, 569–575 (2016).
5. T. S. Ahmed, R. H. Grubbs, J. Am. Chem. Soc. 139, 1532–1537
(2017).
6. A. Masarwa et al., Nature 505, 199–203 (2014).
7. P. H. Chen, B. A. Billett, T. Tsukamoto, G. Dong, ACS Catal. 7,
1340–1360 (2017).
.27. T. Sperger, I. A. Sanhueza, I. Kalvet, F. Schoenebeck,
Chem. Rev. 115, 9532–9586 (2015).
28. We did not specifically examine the exact reaction time
necessary to reach full conversion for each substrate. Grotjahn
and co-workers observed several cases of rapid (<30 min)
reactions at room temperature with their Ru catalyst.
.29. X. Liu et al., J. Am. Chem. Soc. 140, 6873–6882 (2018).
30. H. Sommer, F. Juliá-Hernández, R. Martin, I. Marek, ACS Cent.
Sci. 4, 153–165 (2018).
31. H. Schwarz, Acc. Chem. Res. 22, 282–287 (1989).
32. F. Juliá-Hernández, T. Moragas, J. Cornella, R. Martin, Nature
545, 84–88 (2017).
Our mechanistic analysis suggested that a
double-bond migration over more than one bond
should also be feasible. Migrations of double
bonds over extended carbon skeletons are more
generally termed chain walking (30, 31) and con-
stitute a strategy in relation to remote function-
alizations (32, 33). Upon subjection of catalyst
1 to various substrates (Fig. 4), double-bond
migrations occurred efficiently and with exclusive
E-selectivity in 3 hours at room temperature,
showing that E-olefins can also be installed re-
motely. Whereas for the two-, three-, or four-
carbon chain walks (35 to 38, 40) the benzylic
olefin was formed in high yield (with ≤5% of
alternative internal isomers according to unca-
librated GC-MS analysis), when moving to longer
chains (>4 carbons, i.e., 41, 42) or nonaromatic
driving forces (39), significant amounts of al-
ternative internal double-bond isomers were
observed (34). Whereas Pd-catalysis such as
[(Cl)(H)Pd(II)(PtBu3)2] (14) primarily yields one-
bond isomerizations, E-selective chain walks are
otherwise achieved under Ru catalysis (35) (with
similar or lower driving force) (see fig. S23).
8. A. Vasseur, J. Bruffaerts, I. Marek, Nat. Chem. 8, 209–219
(2016).
9. M. Hassam, A. Taher, G. E. Arnott, I. R. Green,
W. A. L. van Otterlo, Chem. Rev. 115, 5462–5569 (2015).
10. G. Hilt, ChemCatChem 6, 2484–2485 (2014).
11. E. Larionov, H. Li, C. Mazet, Chem. Commun. (Camb.) 50,
9816–9826 (2014).
33. L. Lin, C. Romano, C. Mazet, J. Am. Chem. Soc. 138,
10344–10350 (2016).
34. We also examined the isomerization of a mixture of E/Z internal
double bond isomers, i.e., a 46:54 mixture of (E)- and (Z)-but-2-
en-1-ylbenzene. Using 5 mol % over 24 hours delivered (E)-but-1-
en-1-ylbenzene in 87% conversion and >98:2 E/Z selectivity.
35. D. B. Grotjahn, C. R. Larsen, J. L. Gustafson, R. Nair, A. Sharma,
J. Am. Chem. Soc. 129, 9592–9593 (2007).
36. www.statista.com/markets/412/topic/456/pharmaceutical-
products-market/.
37. S. K. Sharma, V. K. Srivastava, R. V. Jasra, J. Mol. Catal. Chem.
12. S. W. M. Crossley, F. Barabé, R. A. Shenvi, J. Am. Chem. Soc.
136, 16788–16791 (2014).
13. G. Li et al., J. Am. Chem. Soc. 138, 7698–7704 (2016).
14. D. Gauthier, A. T. Lindhardt, E. P. K. Olsen, J. Overgaard,
T. Skrydstrup, J. Am. Chem. Soc. 132, 7998–8009 (2010).
15. E. H. P. Tan, G. C. Lloyd-Jones, J. N. Harvey, A. J. J. Lennox,
B. M. Mills, Angew. Chem. Int. Ed. 50, 9602–9606 (2011).
16. E. Negishi, J. P. Maye, D. Choueiry, Tetrahedron 51, 4447–4462
(1995).
17. C. R. Larsen, D. B. Grotjahn, J. Am. Chem. Soc. 134,
10357–10360 (2012).
18. M. Mayer, A. Welther, A. Jacobi von Wangelin, ChemCatChem 3,
1567–1571 (2011).
19. Market price on 20 August 2018 at Infomine.com:
Ru (~8200 $/kg), Ni (~13 $/kg)
20. H. J. Lim, C. R. Smith, T. V. RajanBabu, J. Org. Chem. 74,
4565–4572 (2009).
21. C.-Y. Lin, P. P. Power, Chem. Soc. Rev. 46, 5347–5399
(2017).
22. A. B. Dürr, H. C. Fisher, I. Kalvet, K.-N. Truong, F. Schoenebeck,
Angew. Chem. Int. Ed. 56, 13431–13435 (2017).
23. B. R. Dible, M. S. Sigman, A. M. Arif, Inorg. Chem. 44,
3774–3776 (2005).
24. The literature does not always clearly specify how E/Z ratios
were determined. We observe the E-isomer exclusively by
1H-NMR for the vast majority of compounds; at best, a trace of
the Z-isomer is observed in some cases, which quantifies to
1 to 2% in (uncalibrated) GC-MS analysis.
25. E. Cahard et al., Angew. Chem. Int. Ed. 51, 3673–3676
(2012).
26. S.-Y. Choi, P. H. Toy, M. Newcomb, J. Org. Chem. 63,
8609–8613 (1998).
245, 200–209 (2006).
38. “HERA Risk Assessment of Isoeugenol 2005”.
ACKNOWLEDGMENTS
We thank K. Keisers for assistance in EPR measurements.
Funding: We gratefully acknowledge RWTH Aachen University and
the European Research Council (ERC-637993) for financial
support. Author contributions: A.K. and S.G. performed the
experiments. T.S. performed the computational studies. All authors
analyzed the data and contributed to the preparation of the
manuscript. F.S. wrote the manuscript. Competing interests: A
patent application has been submitted by RWTH Aachen University
for this reaction, with A.K. and F.S. as inventors. Data and
materials availability: All experimental and computational data
are available in the supplementary materials.
Overall, our protocol combines operational
simplicity, ease of purification, sustainability
(no additional reagents, nonprecious metal, po-
tential for solvent-free reactivity), and scalability
with functional group tolerance, short reaction
times, and mild conditions.
SUPPLEMENTARY MATERIALS
www.sciencemag.org/content/363/6425/391/suppl/DC1
Materials and Methods
Figs. S1 to S24
Tables S1 and S2
Movies S1 and S2
REFERENCES AND NOTES
1. R. C. Larock, Comprehensive Organic Transformations: A Guide
to Functional Group Preparations (VCH, ed. 2, 1999).
2. C. R. Larsen, D. B. Grotjahn, “The value and application of
transition metal catalyzed alkene isomerization in industry”
in Applied Homogeneous Catalysis with Organometallic
Compounds: A Comprehensive Handbook in Four Volumes,
References (39–86)
20 August 2018; accepted 17 December 2018
10.1126/science.aav1610
Kapat et al., Science 363, 391–396 (2019)
25 January 2019
5 of 5

E-Olefins through intramolecular radical relocation
Ajoy Kapat, Theresa Sperger, Sinem Guven and Franziska Schoenebeck
Science 363 (6425), 391-396.
DOI: 10.1126/science.aav1610
Olefin shuffle for just a nickel
Controlling the geometry of carbon-carbon double bonds is a central component of chemical manufacturing. One
useful trick is to shift hydrogen atoms around to interconvert C=C isomers selectively. However, this approach typically
requires precious metals. Kapat et al. now report that more-abundant nickel can catalyze rapid conversion of terminal
olefins into internal olefins with high selectivity for trans geometry. The odd-electron nickel complex relies on a radical
mechanism to shuttle hydrogen to the terminal carbon from the saturated carbon adjacent to the double bond.
Science, this issue p. 391
ARTICLE TOOLS
http://science.sciencemag.org/content/363/6425/391
SUPPLEMENTARY
MATERIALS
http://science.sciencemag.org/content/suppl/2019/01/23/363.6425.391.DC1
REFERENCES
This article cites 81 articles, 1 of which you can access for free
http://science.sciencemag.org/content/363/6425/391#BIBL
PERMISSIONS
http://www.sciencemag.org/help/reprints-and-permissions
Use of this article is subject to the Terms of Service
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science is a registered trademark of AAAS.