Visible Light-Gated Cobalt Catalysis for a Spatially and Temporally Resolved [2+2+2] Cycloaddition

Article abstract of DOI:10.1021/jacs.6b08792

The ability to exert spatial and temporal control over a transition-metal catalyst offers diverse opportunities for the fabrication of functional materials. Using an external stimulus such as visible light to toggle a catalyst between an active and dorman

Full text of DOI:10.1021/jacs.6b08792

Communication
pubs.acs.org/JACS
Visible Light-Gated Cobalt Catalysis for a Spatially and Temporally
Resolved [2+2+2] Cycloaddition
Kyle E. Ruhl†,‡ and Tomislav Rovis*,†,‡,§
†Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523, United States
‡Department of Chemistry, Columbia University, New York, New York 10027, United States
S
* Supporting Information
utilizes complexes that are active in the presence of light and
inactive when the light source is removed (Figure 1b).9 This is
ABSTRACT: The ability to exert spatial and temporal
control over a transition-metal catalyst offers diverse
opportunities for the fabrication of functional materials.
Using an external stimulus such as visible light to toggle a
catalyst between an active and dormant state has proven to
be an effective approach for controlled, radical method-
ologies. Outside of radical bond formation, there is a
dearth of evidence that suggests traditional transition metal
catalysis can similarly be controlled with visible light
energy. Many cobalt complexes that catalyze the [2+2+2]
cycloaddition are assisted by UV photolysis, but strict
photocontrolled methods are unattainable due to high
levels of thermally driven reactivity. Herein, we disclose
the first light-controlled, cobalt-catalyzed [2+2+2] cyclo-
addition via a dual cobalt and photoredox catalyst
manifold. We demonstrate the power of this method
with a spatially and temporally resolved technique for
arene formation using photolithography.
the situation with photoredox catalysis, with the caveat that
recent evidence suggests some extent of initiation behavior with
these catalysts but with short lifetimes.10 Extending this control
to new, nonradical bond constructions would greatly expand our
ability to use diverse catalyst methods in myriad applications
(Figure 1c).11
We began our studies by examining [2+2+2] cycloadditions of
three unsaturated coupling partners using earth-abundant Co(II)
salts as precatalysts. High oxidation state Co is typically reduced
in situ with (super)stoichiometric metal additives in an
uncontrolled manner. We speculated that a photoredox catalyst
and a sacrificial organic reductant would lead to the controlled
formation of a catalytically active species. Voltammetry studies
(see SI, Table S1) of phosphine-based cobalt complexes [CoX2·
(PR3)2] indicated many undergo a quasi-reversible reduction
event at potentials ranging from −0.60 to −1.04 V versus Ag/
AgCl. Measurements also indicate CoBr2·(PCy3)2 displays the
highest ability to effectively stabilize low-valent cobalt with
reduction events as −1.04 V [CoII/CoI] and −1.39 V [CoI/Co0]
versus Ag/AgCl. With this knowledge, we envisioned that a
reductive quench of the photocatalyst excited state IrIII* with
sacrificial amine would furnish the reduced IrII species (E1/2[IrIII/
IrII]= −1.30 V vs Ag/AgCl).12 Reduction of CoBr2·(PCy3)2 by
IrII would be thermodynamically favored to generate an active
low-valent cobalt catalyst.
To test this hypothesis and intercept an active cobalt(II)
species for arene formation, we first screened reaction conditions
for the [2+2+2] cycloaddition of diyne 1 and phenylacetylene 2.
In the presence of [Ir(dF-CF3ppy)2(dtbbpy)]PF6 and sub-
stoichiometric amounts of sacrificial reductant diisopropylethyl-
amine (DIPEA), the complex CoCl2·(PCy3)2 delivers arene
product 2 in 74% yield (Figure 2A, entry 1). Exclusion of light or
photocatalyst results in little to no conversion (Figure 2A, entries
2, 3) but removal of the amine reductant only diminishes reaction
efficiency, still providing 3 in 43% yield. In the absence of
sacrificial amine reductant, tricyclohexylphosphine (E1/2 = +0.89
V vs Ag/AgCl) likely serves as a reductant for IrIII* which
decreases the effective ligand concentration (Figure 2A, entry 4).
Notably, heating the reaction in the absence of light also does not
provide product (Figure 2A, entry 5). Although several cobalt
precatalysts furnish product (Figure 2A, entry 6, 7), the highest
he need for increased control of catalyst activity for
Tmaterials and biological applications has underscored the
importance of spatially and temporally resolved transformations.
This level of control is generally obtained by using an external
stimulus to start or stop a catalytic reaction.1,2 By far the most
ideal external stimulus is visible light, easy to handle and deliver
spatially, inexpensive and widely accessible. The use of light to
activate or inactivate a catalyst is referred to as light-gated
catalysis, and represents a powerful approach to afford exquisite
control over several transformations.3,4
Forty years of cobalt catalysis literature has illustrated the
synthetic utility of these catalysts in assembling arenes via
[2+2+2] cycloadditions. The typical precatalysts used in these
transformations such as Co2(CO)8 and CpCo(CO)2, when
irradiated with ultraviolet light, catalyze the [2+2+2] cyclo-
addition reaction at an accelerated rate.5 Although photo-
dissociation of the labile CO ligand assists reactivity, thermal
ligand dissociation is a competitive pathway which prevents a
purely light-controlled reaction. A viable alternative to carbonyl-
based precatalysts uses in situ reduction of phosphine-based
Co(II) complexes with low-valent metals but affords little
opportunity for external regulation.6 Indeed, examination of the
relevant literature reveals that there are no light-gated catalysts
for the vast majority of common organic transformations.3,7,8
Photocaged metal catalysis involves uncaging and irreversible
release of an active catalyst (Figure 1a). Light-gated catalysis
Received: August 22, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.6b08792
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 1. (a) Light initiation of photocaged catalysis typically involves releasing the caged catalyst. (b) Light-gated catalysis leads to a catalyst being on
under light irradiation and off once the light source is removed. (c) Light-gated process allows control of catalysis to specific times and spaces.
precise temporal control over the entire reaction period (Figure
3). Initial irradiation affords product in 27% yield, but removing
the light source quickly stops catalysis. Reactivity is easily
restored with a second period of irradiation and the reaction
proceeds to give a high yield of arene product. Furthermore, the
precatalysts CoCl2·(PEt3)2 and CoBr2·(dppp) both deliver
Figure 2. (A) Reaction optimization. Standard conditions: 0.20 mmol
diyne, 0.35 mmol alkyne, 0.1 M MeCN. Yield determined by HPLC
analysis using an internal standard. (B) Reaction scope. Reaction scope
under conditions in entry 8, unless otherwise stated. (a) Photocatalyst
removed. (b) Reaction heated to 80 °C. (c) 5% CoBr2(PEt3)2, (d) DCE
as solvent. (e) 50 °C. (f) 1:1 dr.
yield of 92% is obtained with CoBr2·(PCy3)2 using a standard 14
W CFL bulb (Figure 2A, entry 8).
To examine further the impact of light we conducted
experiments under alternating periods of irradiation and
darkness. The parent precatalyst CoBr2·(PCy3)2 shows a distinct
response to light (controlled beyond initiation), which allows
Figure 3. (A) Ligand effect on catalysis. (B) Attempt at thermal
restoration of catalysis.
B
DOI: 10.1021/jacs.6b08792
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 4. (a) Light-gated spatial control of photolithographic modification of polydimethylsiloxane (PDMS). (b) Magnified fluorescence images of
functionalized material with binary masks. (c) Gradient scale photomasks suggest increased light exposure gives rise to increased level of functional
density on the PDMS surface. Plot shows average fluorescence as a function of distance, left to right.
product albeit in low yield. In contrast to CoBr2·(PCy3)2, the
precatalysts form arene product during periods of darkness which
offers no external regulation beyond initiation.
Elevated temperatures are inefficient at restoring active
catalyst in the absence of light. Subjecting an arrested reaction
to 80 °C results in a modest increase in reactivity but the reaction
stalls.13 Exposure of a cooled solution to ambient light results in
complete restoration of reactivity (Figure 3B). After finding this
unique ligand effect and its impact on reaction control, we began
to explore the utility of light-gated arene formation.
One leading method to construct arene-rich materials uses
arene radicals generated by the reduction of aryl diazonium
salts.14 Although the high reactivity of arene radicals offers many
opportunities for bond formation, it consequently leads to
undesired side reactions. In most examples, diazonium salt
reduction lacks spatial resolution and requires specialized
equipment.15 Similarly, etching material from a uniform self-
assembled monolayer can also be done with AFM or X-ray
technology.16 We envisioned using photolithographic techni-
ques to control arene formation spatially via light-controlled
cobalt catalysis.
Supporting Information, Figure S2). In addition to binary
photolithography, we wanted to further explore the reaction
control at the surface through grayscale printing. Reports from
Hawker and co-workers demonstrate that the functional density
of ATRP is directly proportional to the amount of light
exposure.18a−c Indeed, conducting the spatially resolved
[2+2+2] cycloaddition with a grayscale mask suggests the extent
of arene functionalization on the surface correlates with the
amount of light exposure (Figure 4C).
The most surprising aspect is the role that light and
photocatalyst play in regulating catalysis. We propose, thus,
that the Ir photocatalyst plays two distinct roles. Confirmed by
electrochemical measurements, the photocatalyst acts as a
reductant to deliver a catalytically active low valent cobalt
species (i, Scheme 1).19−21 Cyclic voltammetry (CV) studies in
the presence of diyne substrate suggest oxidative cyclization
occurs favorably at Co(0) to deliver a cobaltacyclopentadiene
Scheme 1. Proposed Mechanism
We chose polydimethylsiloxane (PDMS)17 as a support and
began our studies with the installation of a diyne-bearing
trichlorosilane to provide an alkyne-rich PDMS substrate. This
functionalized PDMS is then treated with a solution containing
precatalyst, photocatalyst, alkyne and reductant. The solution is
liberally pipetted onto the PDMS substrate and a photomask is
placed in direct contact. A blue LED is then used to irradiate the
solution (Figure 4A). After irradiation, the PDMS is transferred
to a solution of THF where excess reagents are removed from the
PDMS with successive washes. Imaging the PDMS surface with a
fluorescence microscope reveals the [2+2+2] cycloaddition
occurs with excellent spatial resolution (Figure 4B). Control
experiments in the absence of alkyne functionalization or cobalt
precatalyst indicate little to no background fluorescence (see
C
DOI: 10.1021/jacs.6b08792
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
observed by CV (ii, Scheme 1). After an oxidation22 of ii by the
photocatalyst excited state to the octahedral intermediate iii,
alkyne coordination provides intermediate iv.23 Insertion and
subsequent reductive elimination from v generates a Co(I)
species which is reduced to the active cobalt(0) species by the IrII
reduced state.24 Thus, the unique features inherent to photo-
redox catalysis allow the mechanism to access every oxidation
step at Co between 0 and 3.
(5) (a) Agenet, N.; Buisine, O.; Slowinski, F.; Gandon, V.; Aubert, C.;
Malacria, M. Org. React. 2007, 68, 1−302. (b) Vollhardt, K. P. C. Angew.
Chem., Int. Ed. Engl. 1984, 23, 539−556. (c) Gandon, V.; Aubert, C.;
Malacria, M. Chem. Commun. 2006, 2209−2217.
(6) (a) Gandeepan, P.; Cheng, C.-H. Acc. Chem. Res. 2015, 48, 1194−
1206. (b) Slowinski, F.; Aubert, C.; Malacria, M. Adv. Synth. Catal. 2001,
343, 64−67.
(7) Prier, C. K.; Rankic, D. A.; MacMillan, D. W. C. Chem. Rev. 2013,
113, 5322−5363.
In conclusion, we have identified a cobalt catalyst that enables
light-controlled arene assembly using visible light as the external
stimulus. This photochemically gated process provides unique
opportunities for external regulation of catalysis, illustrated with
the photolithographic patterning of PDMS.
(8) (a) Tellis, J. C.; Primer, D. N.; Molander, G. A. Science 2014, 345,
433−436. (b) Zuo, Z.; Ahneman, D.; Chu, L.; Terrett, J.; Doyle, A. G.;
MacMillan, D. W. C. Science 2014, 345, 437−440. (c) Terrett, J. A.;
Cuthbertson, J. D.; Shurtleff, V. W.; MacMillan, D. W. C. Nature 2015,
524, 330−334.
(9) Hecht (ref 3) loosely defines light-gated catalysis to include photo-
caged and photoswitchable catalysis, the latter presumably requiring a
light-stimulus to achieve both on and off states. We prefer to borrow the
definition of light-gated as used in the ion-channel literature which states
that the channel (or catalyst) is open or on with light and off or closed in
its absence. It specifically excludes photocaged examples from the
definition; see: Banghart, M. R.; Volgraf, M.; Trauner, D. Biochemistry
2006, 45, 15129−15141.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.6b08792.
■
S
Procedures and compound characterization (PDF)
(10) Cismesia, M. A.; Yoon, T. P. Chem. Sci. 2015, 6, 5426−5434.
(11) For examples of spatially and temporally controlled transition
metal catalysis, see: (a) Davis, J. J.; Bagshaw, C. B.; Busuttil, K. L.;
Hanyu, Y.; Coleman, K. S. J. Am. Chem. Soc. 2006, 128, 14135−14141.
(b) Buchner, M. R.; Bechlars, B.; Ruhland, K. J. Organomet. Chem. 2013,
744, 60−67.
AUTHOR INFORMATION
Corresponding Author
*tr2504@columbia.edu
ORCID
■
(12) Roth, H. G.; Romero, N. A.; Nicewicz, D. A. Synlett 2016, 27,
714−723.
Tomislav Rovis: 0000-0001-6287-8669
Present Address
(13) We interpret the modest increase in catalytic activity as a thermal
disproportionation of inactive Co catalyst (likely Co(I) to Co(II) and
Co(0)).
§Department of Chemistry, Columbia University, New York,
New York 10027, United States
(14) Aryl diazonium salts. In New coupling agents in polymer and surface
science; Chehimi, M. M., Ed.; Wiley-VCH: Weinheim, 2012.
Funding
NSFCHE-1339674.
(15) Kirkman, P. M.; Guell, A. G.; Cuharuc, A. S.; Unwin, P. R. J. Am.
̈
Notes
Chem. Soc. 2014, 136, 36−39.
The authors declare the following competing financial
interest(s): A provisional patent has been filed on this
technology.
(16) Koh, S.-J. Nanoscale Res. Lett. 2007, 2, 519−545.
(17) (a) Miller, P. J.; Matyjaszewski, K. Macromolecules 1999, 32,
8760−8767. (b) Semsarzadeh, M. A.; Abdollahi, M. J. Appl. Polym. Sci.
2013, 123, 2423−2430.
ACKNOWLEDGMENTS
(18) (a) Fors, B. P.; Poelma, J. E.; Menyo, M. S.; Robb, M. J.;
Spokoyny, D. M.; Kramer, J. W.; Waite, J. H.; Hawker, C. J. J. Am. Chem.
Soc. 2013, 135, 14106−14109. (b) Poelma, J. E.; Fors, B. P.; Meyers, G.
F.; Kramer, J. W.; Hawker, C. J. Angew. Chem., Int. Ed. 2013, 52, 6844−
6848. (c) Adzima, B. J.; Tao, Y.; Kloxin, C. J.; DeForest, C. A.; Anseth, K.
S.; Bowman, C. N. Nat. Chem. 2011, 3, 258−261. (d) For a recent
review, see: Barner-Kowollik, C.; Goldmann, A. S.; Schacher, F. H.
Macromolecules 2016, 49, 5001−5016.
■
This research was conducted in connection with the Catalysis
Collaboratory for Light-activated Earth Abundant Reagents (C-
CLEAR), which is supported by the National Science
Foundation and the Environmental Protection Agency through
the Network for Sustainable Molecular Design and Synthesis
program (NSFCHE-1339674). We thank Rachel Feeney and
Charles S. Henry (CSU) for technical assistance and Travis
Bailey (CSU) for access to instrumentation.
́
(19) Buriez, O.; Labbe, E.; Perichon, J. J. Electroanal. Chem. 2006, 593,
99−104.
(20) Voltammetry on CoBr2·(PCy3)2 also suggests Co0 is likely
generated via initial reduction to CoI and subsequent bimolecular
disproportionation to deliver a Co0 and CoII species.
REFERENCES
■
(1) (a) Treat, N. J.; Fors, B. P.; Kramer, J. W.; Christianson, M.; Chiu,
C.-Y.; Read de Alaniz, J.; Hawker, C. J. ACS Macro Lett. 2014, 3, 580−
584. (b) Theriot, J. C.; Lim, C.-H.; Yang, H.; Ryan, M. D.; Musgrave, C.
B.; Miyake, G. M. Science 2016, 352, 1082−1086. (c) Dadashi-Silab, S.;
Tasdelen, M. A.; Yagci, Y. J. Polym. Sci., Part A: Polym. Chem. 2014, 52,
2878−2888. (d) Perkowski, A. J.; You, W.; Nicewicz, D. A. J. Am. Chem.
Soc. 2015, 137, 7580−7583.
(21) For evidence of Co0 activity in the [2+2+2], see: Chiusoli, G. P.;
Pallini, L.; Terenghi, G. Transition Met. Chem. 1984, 9, 360−362.
(22) The irreversibility of this intermediate observed by CV is
potentially due to a rapid chemical event (coordination or insertion of
alkyne) following oxidation. See SI, Figure S3.
(23) For octahedral cobaltacyclopentadiene, see: Bouayad, A.;
Dartiguenave, M.; Menu, M.-J.; Dartiguenave, Y. Organometallics
1989, 8, 629−637.
(24) Alternatively, a bimolecular disproportionation reaction could
regenerate active intermediate i and a reducible cobalt(II) species.
(2) Leibfarth, F. A.; Mattson, K. M.; Fors, B. P.; Collins, H. A.; Hawker,
C. J. Angew. Chem., Int. Ed. 2013, 52, 199−210.
(3) Stoll, R. S.; Hecht, S. Angew. Chem., Int. Ed. 2010, 49, 5054−5075.
(4) (a) Blanco, V.; Leigh, D. A.; Marcos, V. Chem. Soc. Rev. 2015, 44,
5341−5370. (b) Neilson, B. M.; Bielawski, C. W. ACS Catal. 2013, 3,
1874−1885. (c) Teator, A. J.; Lastovickova, D. N.; Bielawski, C. W.
Chem. Rev. 2016, 116, 1969−1992. (e) Shanmugam, S.; Xu, J.; Boyer, C.
J. Am. Chem. Soc. 2015, 137, 9174−9185. (d) Trotta, J. T.; Fors, B. P.
Synlett 2016, 27, 702−713.
D
DOI: 10.1021/jacs.6b08792
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX