A multistable molecular switching system based on an anthracene-extended bis-thiaxanthylidene with three individually addressable states that can be interconverted by electrochemical, thermal, and photochemical reactions is reported. Besides reversible switching between an open-shell diradical- and a closed-shell electronic configuration, our findings include a third dicationic state and control by multiple actuators. This dicationic state with an orthogonal conformation can be switched electrochemically with the neutral open-shell triplet state with orthogonal conformation, which was characterized by EPR. The remarkably stable diradical shows kinetic stability as a result of a significant activation barrier for isomerization to a more stable neutral closed-shell folded geometry. We ascribe this activation barrier of ΔG⧧(293 K) = 25.7 kcal mol-1 to steric hindrance in the fjord region of the overcrowded alkene structure. The folded closed-shell state can be converted back to the diradical state by irradiation with 385 nm. The folded state can also be oxidized to the dicationic state. These types of molecules with multiple switchable states and in particular stable diradicals show great potential in the design of new functional materials such as memory devices, logic gates, and OFETs.
A multistable molecular switching system based on an anthracene-extended bis-thiaxanthylidene with three individually addressable states that can be interconverted by electrochemical, thermal, and photochemical reactions is reported. Besides reversible switching between an open-shell diradical- and a closed-shell electronic configuration, our findings include a third dicationic state and control by multiple actuators. This dicationic state with an orthogonal conformation can be switched electrochemically with the neutral open-shell triplet state with orthogonal conformation, which was characterized by EPR. The remarkably stable diradical shows kinetic stability as a result of a significant activation barrier for isomerization to a more stable neutral closed-shell folded geometry. We ascribe this activation barrier of ΔG⧧(293 K) = 25.7 kcal mol-1 to steric hindrance in the fjord region of the overcrowded alkene structure. The folded closed-shell state can be converted back to the diradical state by irradiation with 385 nm. The folded state can also be oxidized to the dicationic state. These types of molecules with multiple switchable states and in particular stable diradicals show great potential in the design of new functional materials such as memory devices, logic gates, and OFETs.
Artificial molecular
switches and motors play a prominent role
in the transition from molecules with static function to dynamic molecular
systems in contemporary chemistry.[1−3] Taking inspiration from
biology, e.g., the process of vision, it is evident that molecular
switching offers numerous opportunities for the reversible control
of material properties,[4] molecular information
content,[5] or biological function.[6] Switching between redox and geometric states
can be induced by external stimuli including electricity, light, and
heat.[7,8] A molecular switch can typically form either
of two stable isomers, so-called bistability, to control, for example,
magnetic, electronic, and optical properties, supramolecular assembly,
and molecular recognition or complex functions such as catalysis.[9−15]Moving beyond bistability, the design of multiple switchable
molecular
systems that can reach various stable states triggered by distinct
and orthogonal external inputs e.g., photochemical, thermal, and electrochemical,
remains particularly challenging. Herein, we report reversible switching
between three states in a single molecule, including
a highly stable diradical state, controlled through multiple actuators
(electricity, light, and heat). Stable organic diradicals feature
unique electronic properties and therefore show great potential for
applications in molecular electronics such as organic field-effect
transistors (OFETs).[16] Reversible switching
between open-shell diradical (or diradicaloid) and closed-shell electronic
configurations of organic molecules is fascinating both experimentally
and theoretically.[17−20] The system we describe here adds a third state that is interconvertible
with the open-shell diradical state via redox chemistry. We anticipate
that such a system with three independently addressable states will
open doors toward smart, tunable electronics such as memory devices
and logic gates.[21]The systems presented
here are based on bistricyclic aromatic enes
(BAEs),[22] a class of molecules that includes
the overcrowded alkene light-driven molecular motors, which are robust
examples of multistage switchable molecular systems.[23] In BAEs, the steric crowding in the fjord region plays
a key role in the relative stability of various isomers.[22] Two such geometries are described as folded
(F) and twisted (T) (Figure ). Facing the challenge to design orthogonal
switching systems to allow ultimately more complex dynamic and responsive
behavior, bisthiaxanthylidenes, a specific class of BAEs, offer the
opportunity to combine multiple effectuators and states.[12] Bisthiaxanthylidenes undergo geometrical changes
induced by either redox or photochemical reactions. A photochemical-redox
luminescent switch based on bisthiaxanthylidenes was developed earlier
by our group, in which oxidation leads to a stable dicationic state
with a nearly orthogonal (O) orientation of the two tricyclic
aromatic moieties (Figure a).[12]
Figure 1
(a) Reversible switching
of bis-thiaxanthylidenes from our earlier
studies,[12] (b) Tschitschibabins’s
hydrocarbon,[24] (c) DMA-based switch from
Kubo et al.,[18] and (d) our current work
on aromatically extended bis-thiaxanthylidene switches.
(a) Reversible switching
of bis-thiaxanthylidenes from our earlier
studies,[12] (b) Tschitschibabins’s
hydrocarbon,[24] (c) DMA-based switch from
Kubo et al.,[18] and (d) our current work
on aromatically extended bis-thiaxanthylidene switches.Recently, photochemical and thermal interconversion of stable
closed-shell
and open-shell diradical states of an organic compound (with a concomitant
geometry change) has emerged as a new modality of molecular switching.[18] Organic compounds with an open-shell diradical
resonance contributor in the ground state date back to Tschitschibabin’s
hydrocarbon, reported in 1907 (Figure b).[24,25] Noting the importance of stable
open-shell systems, other organic compounds with diradicaloid ground
states have since been described.[26−29] Bistability between open- and
closed-shell states can be achieved when the activation barriers are
sufficiently high,[18] while low barriers
(with respect to kBT)
lead to thermal equilibration over the states.[30,31] The open-shell diradical configurations are stabilized by a combination
of factors including steric hindrance, delocalization of spin density,
and aromaticity.[32,33] In 2020, Kubo reported an organic
compound capable of controlled switching between open-shell diradical
and closed shell state: this BAE, based on the 9,10-dihydro-10,10-dimethyl-9-anthryl
(DMA) motif, undergoes conversion from a triplet open-shell diradical
state to a neutral closed-shell thermally, which can then be converted
back to the open-shell state photochemically (Figure c).[18]The
anthracene-extended bis-thiaxanthylidene 1 described
here allows for three-state reversible switching (Figure d). In our design the peripheral
thioxanthene motif enables redox switching by stabilizing a dicationic
state (1) by virtue of the
sulfur heteroatom, which enables full conjugation to an “anthracene-like”
delocalized structure facilitated by mesomeric electron donation (an
effect that is entirely lacking in the DMA motif).[12,34] The thioxanthenes also stabilize an open-shell diradical configuration
(1-O) in the neutral form by spin density delocalization.[35,36] Steric crowding in the fjord region caused by the central anthracene
linker induces additional kinetic stabilization to the diradical.
The stable open-shell orthogonal diradical converts thermally to a
lower energy closed-shell folded geometry (1-F), and
irradiation restores the compound to its orthogonal diradical state.
In contrast, the analogous benzene-bridged bis-thiaxanthylidene 2 (Figure d) displays a thermal equilibrium between an open-shell diradical
state (2-O) and a lower energy closed-shell folded state
(2-F), which demonstrates the effect of steric crowding
in the fjord region.
Results and Discussion
Synthetic Procedures
The arene-extended bisthioxanthylium
salts 1 and 2 were prepared with anthracene and benzene spacers,
respectively (Scheme ). Thioxanthen-9-one 3 was iodinated to obtain 4, which was then functionalized with ethoxy groups through
an Ullmann coupling to increase the solubility of these compounds
in organic solvents. Dilithiation of 9,10-dibromoanthracene and 1,6-dibromobenzene
and subsequent reaction with 5 produced diols 7 and 9, respectively, which after dehydration with aqueous
HBF4 yielded the dicationic BF4– salts 1 and 2. Characterization of compound 1 by 1H NMR spectroscopy was
precluded by the presence of traces of paramagnetic species, which
is probably a monoradical impurity that facilitates fast electron
exchange on the NMR time scale and consequent line broadening. The
presence of a monoradical impurity is supported by an EPR signal that
is typical for a doublet species (Figure S8). The salts 1 and 2 were reduced chemically with cobaltocene
to obtain the neutral compounds 1-O and 2-F. The 1H NMR spectrum of compound 1-O (Figure S2) shows paramagnetic shifts and broadening,
consistent with open-shell diradical character. This is not the case
for compound 2-F, which we therefore expect to be in
the lowest energy folded closed-shell state (DFT calculations vide infra), and we denote it as 2-F. Reoxidation
of 1-O and 2-F to their respective dicationic
salts 1 and 2 was achieved with NOBF4 (Figures S2 and S3).
Scheme 1
Synthesis of Compounds 1, 2, 1-O, and 2-F
X-ray Crystallography
Single-crystal X-ray analysis
of 1 and 2-F revealed
the structures shown in Figure . DFT-optimized structures of 1-O and 2 are shown in Figure . Both the anthracene-bridged dication 1 (Figure a) and the benzene-bridged dication 2 (DFT-optimized structure, Figure b) adopt an almost
orthogonal conformation of the aromatic spacers relative to the thioxanthylium
units with dihedral angles of 80° and 75°, respectively.
In both structures the two thioxanthylium motifs are coplanar. For 1, the C1–C2 and C2–C3
distances are 1.494(3) and 1.407(3) Å, respectively, which shows
that it does not have a quinoidal structure. DFT calculations predict
a similar quasi orthogonal conformation for 1-O similar
to the dications with a dihedral angle of 82° between the anthracene
spacer relative to the thioxanthenes (Figure a). In contrast, 2-F has a double-folded
boat-like structure (Figure b). Within one thiaxanthylidene the fold angle between the
two C6H3OEt rings is 133°. The bond length
of ∼1.38 Å between C5–C8 and C1–C2 and ∼1.35
Å between C6–C7 and C3–C4 are consistent with double-bond
character. On the other hand, single-bond character was noted between
C2–C7, C2–C3, C4–C5, and C5–C6 (∼1.45
Å). This reveals that 2-F has a quinoidal closed-shell
structure in the solid state.
Figure 2
ORTEP drawings of the (a) 1 and (b) 2-F structures. Hydrogen
atoms, counterions,
and solvent molecules are omitted for clarity.
Figure 3
DFT-optimized
structures of the (a) 1-O and (b) 2 structures. Hydrogen atoms are
omitted for clarity.
ORTEP drawings of the (a) 1 and (b) 2-F structures. Hydrogen
atoms, counterions,
and solvent molecules are omitted for clarity.DFT-optimized
structures of the (a) 1-O and (b) 2 structures. Hydrogen atoms are
omitted for clarity.
Electronic Absorption and
Emission Spectroscopy
The
redox couple 1 and 1-O both display UV/vis absorption spectra (Figure a and c, Table ) with absorptions at long wavelength in
the visible region between 500 and 650 nm. These are common bands
for twisted geometries.[12,18,37−39] Thus, it seems likely that the conformation does
not change significantly between 1 and 1-O. The absorption bands for 1 and 1-O between 325 and 400
nm with vibronic splitting are typical for anthracene.[33] In contrast, the spectra of 2 and 2-F (Figure b and d, Table ) are consistent with a major change in geometry. 2-F shows a broad absorption band between 350 and 550 nm.
The absorption bands at shorter wavelength are typical of an increase
in S0–S1 gap expected for a folded structure. Its spectrum
is similar to that observed for the folded anthracene-bridged bis-DMA,
confirming its folded structure in solution,[18] while 2 shows characteristic
absorptions of a twisted conformation.[12] Compound 1 exhibits red emission
(λmax = 641 nm), whereas 1-O is not
luminescent as expected due to its triplet diradical character (vide infra). Furthermore, emission spectroscopy shows that
reduction of 2 to 2-F results in a change in luminescence with λmax shifting
from 682 nm (red) to 569 nm (yellow), respectively.
Figure 4
UV/vis absorption and
normalized emission spectra (dotted lines)
of (a) 1 (6.25 × 10–6 M), (b) 2 (6.25
× 10–6 M), (c) 1-O (2.5 ×
10–5 M), and (d) 2-F (1.25 × 10–5 M) in DCM. Excitation wavelengths: (a) 560 nm, (b)
520 nm, and (d) 450 nm.
Table 1
λmax Values for 1, 2, 1-O, and 2-F
λmax (nm)
UV/vis
emission
(Φb)
(a) 12+
260, 299, 359, 377, 398,
448, 569
641 (0.02)
(b) 22+
253 (Sa), 303, 376, 443, 561
682 (4.0)
(c) 1-O
264, 284, 340 (S),
360,
376, 397, 444, 614
(d) 2-F
255,
391, 468
569
S = shoulder.
Φ = quantum yield in %.
UV/vis absorption and
normalized emission spectra (dotted lines)
of (a) 1 (6.25 × 10–6 M), (b) 2 (6.25
× 10–6 M), (c) 1-O (2.5 ×
10–5 M), and (d) 2-F (1.25 × 10–5 M) in DCM. Excitation wavelengths: (a) 560 nm, (b)
520 nm, and (d) 450 nm.S = shoulder.Φ = quantum yield in %.
EPR Spectroscopy
1-O shows a typical EPR
spectrum for a triplet diradical in toluene at 77 K (Figure ). Simulation (EasySpin, see
the SI) yields a g-value
of 2.0023 and zero field splitting parameters D =
183 MHz (65.3 G) and E = 3 MHz. D was used to estimate a spin–spin distance of 7.53 Å
by the point-dipole approximation.[26] This
distance is greater than the distance between the two carbons at the
9-position (5.87 Å), which indicates delocalization of the unpaired
electrons.[26,40] The forbidden Δms = 2 signal is observed at 1722 G (effective g-value of 4.006), the intensity of which increases with
decreasing temperature. The Δms =
2 double integrated intensity (IEPR) vs
1/T (temperature range 70–8.6 K) plot was
fitted with the Bleaney–Bowers equation (eq ), where C is the Curie constant, J is the electron exchange interaction, 2J equals ΔES-T (=+22 ×
10–3 kcal mol–1), and J = +3.9 cm–1 (ferromagnetic coupling).
Figure 5
Top: EPR signal
of compound 1-O (4.4 × 10–4 M)
in toluene at 77 K with inset: forbidden Δms = 2 signal at 40 K. Bottom: Curie plot of 1-O with fit to the Bleaney–Bowers equation (adj. R-square = 0.9939).
Top: EPR signal
of compound 1-O (4.4 × 10–4 M)
in toluene at 77 K with inset: forbidden Δms = 2 signal at 40 K. Bottom: Curie plot of 1-O with fit to the Bleaney–Bowers equation (adj. R-square = 0.9939).From both the paramagnetic 1H NMR and EPR spectral data
we conclude that 1-O has a triplet diradical ground state
with a singlet diradical state slightly higher in energy. As expected,
for compound 2-F a signal was not observed, consistent
with its preferred closed-shell, folded (F) quinoidal
structure (vide supra).
Computational Studies
DFT calculations on 1 and 2 were performed
to further investigate the stabilities
of the folded quinoidal geometries relative to the orthogonal diradical
conformations. Computational investigation at the B3LYP/6-31** level
of theory (for the selection of the computational method, see the SI) of the various possible geometries showed
that the most stable geometry is a folded (F) structure
for the neutral compounds 1 and 2 and an
orthogonal (O) conformation for the dicationic compounds 1 and 2 (Tables S4, S5). More specifically,
the most stable geometry of 1-F is in a doubly anti-folded state that minimizes steric hindrance with the
anthracenyl bridge. For 2-F the steric hindrance is significantly
lower, and it has a syn-folded orientation as its
most stable form. However, the isomerization barriers of the neutral
compounds differ greatly. Reduction of 1 leads to the diradical state 1-O without significant
conformational change. 1-O is a local minimum (ΔE = 0.7 kcal mol–1 relative to 1-F, Figure ) kinetically
trapped with a high barrier (ΔE⧧ = 30.6 kcal mol–1) for isomerization at room temperature
toward the global minimum 1-F. Conversely, as deduced
from electrochemistry (vide infra), the reduction
of 2 does initially lead to 2-O, which subsequently isomerizes to 2-F. The
calculated barrier for the latter process (ΔE⧧ = 9.5 kcal mol–1) supports
the experimental observations. The low barrier and relative energy
differences between 2-O and 2-F provide
a rationale for the shape of the cyclic voltammogram (Figure e), where a relatively fast
equilibrium means that both the orthogonal and folded structures are
present in solution. This isomerization pathway was probed with a
potential energy surface scan (Figures S22 and
S23). The relative energies of the diradical singlets and the
corresponding triplets for 1-O and 2-O are
sufficiently close for population of both states (ΔES-T = −14 × 10–3 kcal
mol–1 for 1-O and ΔES-T = −0.145 kcal mol–1 for 2-O). The low ΔES-T for 1-O is in accordance with EPR data, although experimentally
a triplet ground state was observed. However, the calculated energy
difference is less than the expected error margin, and hence it can
only be concluded that the energy levels are highly similar. For all
diradical species we found that the spin density is mainly located
on the thioxanthene motifs with negligible contribution from the bridging
motifs.
Figure 6
Schematic and graphic representation of the thermal isomerization
pathway for the neutral geometries 1-O to 1-F and 2-O to 2-F based on DFT calculations
(B3LYP/6-31G**). Hydrogens are omitted for clarity.
Figure 7
Cyclic voltammograms of (a) 1, (b) 2, (c) 1-O, (d) 2-F (0.1 V s–1), (e) 2-F at different scan rates, and (f) 2-F at −50
°C (0.1 V s–1) recorded in DCM with 5 ×
10–4 M analyte and 0.1 M TBAPF6.
Schematic and graphic representation of the thermal isomerization
pathway for the neutral geometries 1-O to 1-F and 2-O to 2-F based on DFT calculations
(B3LYP/6-31G**). Hydrogens are omitted for clarity.Cyclic voltammograms of (a) 1, (b) 2, (c) 1-O, (d) 2-F (0.1 V s–1), (e) 2-F at different scan rates, and (f) 2-F at −50
°C (0.1 V s–1) recorded in DCM with 5 ×
10–4 M analyte and 0.1 M TBAPF6.
Electrochemistry
Compounds 1 and 2 were
characterized electrochemically by cyclic voltammetry in dichloromethane
(DCM) at rt, and both show reversible electrochemical switching (Figure a and b, Table ). For 1 splitting of the redox waves was seen
in DCM, whereas in MeCN (Figure S9) a single
reversible wave was observed. This difference is due to the better
solvation of the radical cation in MeCN after the first reduction
relative to DCM.[41] In solvents with poor
donor properties, such as DCM, the singly occupied molecular orbital
(SOMO) can be delocalized over the entire molecule, stabilizing the
radical cation.[42] Therefore, in DCM, the
second reduction of 1 is shifted
to more negative potentials. The radical cation of 2 is less stabilized, and only a single reversible
wave is observed independent of solvent. Cyclic voltammograms of 1-O and 2-F do not show hysteresis (Figure c and d, Table ). This suggests that 1-O has a structure similar to 1. The open-shell diradical 1-O indeed has an almost
orthogonal conformation. In contrast, it is proposed that the folded
closed-shell 2-F is in equilibrium with the orthogonal
open-shell diradical 2-O. This explains why 2-F exhibits a quasi-reversible cyclic voltammogram where the rapid
prior equilibrium between 2-F and 2-O precedes
the oxidation of 2-O to 2.[31,43,44] Hysteresis
is not observed, as 2-O and 2 have similar conformations. At lower scan rate there
is more time (within the time frame of the electrochemical experiment)
for the equilibrium to shift toward 2-O, and the voltammogram
shows increasing electrochemical reversibility (Figure e). Cyclic voltammetry at −50 °C
shows the direct oxidation of 2-F to 2 (Figure f, Table ).
At this temperature the 2-O state is not populated, and
the large geometrical change from the folded 2-F to the
orthogonal 2 is responsible
for the hysteresis.
Table 2
Peak- and Half-Wave
Potentials for 1, 2, 1-O, and 2-F
Ep,ox
Ep,red
E1/2
(a) 12+
–0.47/–0.36
–0.42/–0.54
–0.39/–0.50
(b) 22+
–0.29
–0.33
–0.31
(c) 1-O
–0.46/–0.34
–0.40/–0.52
–0.49/–0.37
(d) 2-F
–0.12
–0.35
–0.24
(e) 2-F (−50 °C)
0.14
–0.41
Thermal Equilibrium between 2-F and 2-O
Density functional theory (DFT) calculations
indicated
that the orthogonal diradical 2-O is thermally accessible.
Variable-temperature NMR (VT-NMR) spectroscopy studies of 2-F (Figure a) show
that the resolved signals at 25 °C become increasingly broadened
with heating to 80 °C, consistent with the population of the
excited triplet diradical state (2-O). Returning to 25
°C recovers the original spectrum as expected from the low activation
barrier for the 2-O to 2-F conversion. These
results are consistent with our electrochemical data and show that
there is a thermal equilibrium between 2-F and 2-O.
Figure 8
VT-NMR (500 MHz) spectroscopy studies of compounds (a) 2-F and (b) 1-O in toluene-d8.
VT-NMR (500 MHz) spectroscopy studies of compounds (a) 2-F and (b) 1-O in toluene-d8.
Switching between 1-O and 1-F
DFT calculations predict
that 1-F is more stable than 1-O, but it
was not observed after the chemical reduction
of 1. We speculated that the
as-synthesized 1-O was kinetically trapped. VT-NMR studies
on the paramagnetic compound 1-O (Figure b) showed resolved signals, which are assigned
to 1-F, that emerged as the temperature was increased
to 100 °C. Cooling to 25 °C shows that the signals persist,
consistent with a high activation barrier for interconversion between 1-F and 1-O. Full conversion to 1-F was achieved by holding the solution at 100 °C until its 1H NMR spectrum at rt shows highly resolved signals. 1-O was converted in toluene at 90 °C to isolate and
characterize 1-F. Single crystals of 1-F were grown and confirmed a double anti-folded geometry
(Figure a). Within
one thiaxanthylidene the fold angle between the two C6H3OEt rings is 129°. A quinoidal structure was confirmed
by the double-bond character of the C5–C8 and C1–C2
bonds (∼1.35 Å) and the single-bond character of the C4–C5,
C5–C6, C2–C7, and C2–C3 bonds (∼1.50 Å).
Its UV/vis absorption spectrum supports the hypothesis that a geometrical
change from orthogonal to folded occurred as the longest wavelength
absorption is shifted to 450 nm (Figure b). The emission spectrum of 1-F has a maximum at 474 nm. Thus, by reversible switching between the 1, 1-O, and 1-F states (vide infra) the luminescence can be changed
between red, nonluminescent, and blue, respectively. The activation
barrier for the 1-O to 1-F transition was
determined by UV/vis absorption spectroscopy in toluene. The decay
in absorbance was monoexponential, and rates were used in an Eyring–Polanyi
plot (Figure c). Eyring–Polanyi
analysis yielded the free energy of activation ΔG⧧(293 K) = 25.7 ± 0.2 kcal mol–1. The activation barrier from 1-O to 1-F rationalizes why the reduction of 1 leads to trapping in the 1-O state, as can be
seen from the cyclic voltammograms of the 1/1-O couple. The barriers are related
to the increased steric hindrance in the fjord region, and this behavior
is not seen for the 2O/2F couple. This restricted
movement in the fjord region is also reflected in the onset of the
oxidation potential of 1-F at 0.5 V (relative to the
onset of the oxidation potential of 1-O at −0.6
V) (Figure d). The
lack of chemical reversibility for the oxidation of 1-F is the result of the fast (within the time frame of the electrochemical
experiment) chemical step 1-F → 1. This chemical
step leads to the observation of the following small reduction waves: 1-F → 1-F followed by 1-F → 1-F at peak potentials
0.7 V (shoulder) and 0.6 V, respectively. Oxidation of 1-F (or 1-O) leads to 1, which can subsequently be reduced at −0.39 and −0.50
V to 1-O in the return sweep (from 1 to −1 V).
This hysteresis in the cyclic voltammogram is caused by a geometrical
change.[18,45] Irradiation at 385 nm leads to photochemical
switching from 1-F to 1-O (Figure e). In toluene a band at 615
nm assigned to 1-O is observed. Heating at 90 °C
results in this absorption band disappearing over 1 h. Repeated cycles
of irradiation and heating confirm the excellent reversibility of
this process not showing fatigue. Additional evidence for the photochemical
back reaction of 1-F to 1-O was provided
by EPR spectroscopy. Irradiation at 385 nm of the EPR-silent 1-F leads to the emergence of signals associated with 1-O (Figure f).
Figure 9
(a) ORTEP drawing of the 1-F structure (hydrogen atoms
are omitted for clarity), (b) UV/vis absorption and normalized emission
spectrum (λexc = 350 nm) of 1-F (2.5
× 10–5 M) in DCM including λmax values (quantum yield = 6.9%), (c) Eyring–Polanyi plot of
the thermal conversion of 1-O to 1-F in
toluene monitored at 615 nm obtained by fitting to the linearized
form of the Eyring equation (adj. R-square = 0.9985).
Dashed lines indicate 95% confidence intervals, (d) cyclic voltammogram
of 1-F, Ep,ox1 = 0.66 V (shoulder)
and Ep,ox2 = 0.75 V (5 × 10–4 M in DCM/0.1 M TBAPF6/0.1 V s–1), (e)
cycles of irradiating (1 × 10–3 M) with 385
nm for 10 min (white areas) and heating at 90 °C for 1 h (gray
areas) while monitoring the absorbance at 615 nm, and (f) EPR spectra
of 1-F (5.3 × 10–4 M) in toluene
at 77 K: before irradiation (red) and after irradiation with 385 nm
at rt for 10 min (black).
(a) ORTEP drawing of the 1-F structure (hydrogen atoms
are omitted for clarity), (b) UV/vis absorption and normalized emission
spectrum (λexc = 350 nm) of 1-F (2.5
× 10–5 M) in DCM including λmax values (quantum yield = 6.9%), (c) Eyring–Polanyi plot of
the thermal conversion of 1-O to 1-F in
toluene monitored at 615 nm obtained by fitting to the linearized
form of the Eyring equation (adj. R-square = 0.9985).
Dashed lines indicate 95% confidence intervals, (d) cyclic voltammogram
of 1-F, Ep,ox1 = 0.66 V (shoulder)
and Ep,ox2 = 0.75 V (5 × 10–4 M in DCM/0.1 M TBAPF6/0.1 V s–1), (e)
cycles of irradiating (1 × 10–3 M) with 385
nm for 10 min (white areas) and heating at 90 °C for 1 h (gray
areas) while monitoring the absorbance at 615 nm, and (f) EPR spectra
of 1-F (5.3 × 10–4 M) in toluene
at 77 K: before irradiation (red) and after irradiation with 385 nm
at rt for 10 min (black).
Conclusion
Here we report two arene-extended bis-thiaxanthylidines
in which
the introduction of an anthracene spacer allowed for redox switching
between an orthogonal dicationic (1) and a diradical state (1-O), while thermal and
photochemical switching allows modulation between 1-O and the folded state 1-F. The barrier between the neutral
orthogonal open-shell triplet (1-O) and the folded closed-shell
(1-F) form is high (ΔG⧧(293 K) = 25.7 kcal mol–1), showing that 1-O is a remarkably stable diradical. The reverse reaction
from the closed-shell to the open-shell form has a higher activation
barrier (ΔG⧧ = 31.3 kcal
mol–1 (from DFT)). The resultant three stable states
of this system show that the reduction of 1 leads to the kinetically trapped 1-O, which can be switched thermally to the more stable folded geometry 1-F. Irradiation of 1-F switches it back to the
open-shell 1-O, while oxidation produces 1. For the benzene analogue, redox switching
between an orthogonal dicationic (2) and a diradical state (2-O) was observed. A
thermal equilibrium was found between the neutral more stable folded
closed-shell (2-F) and orthogonal open-shell (2-O) forms with a low activation barrier from the transient 2-O to 2-F. At low temperature 2-F can be
directly oxidized to 2. We
ascribe the difference in activation barrier between 2-O to 2-Frelative to1-O to 1-F to the increased steric bulk in the fjord region.
Overall, this means a robust three-state switch has been developed
where the dication 1, diradical 1-O, and 1-F states can be individually addressed
and interconverted by electricity, temperature control, or irradiation.
This opens opportunities for the exploration of stable multistate
photo- and electrochemical-responsive materials and devices.
Authors: Gabriel E Rudebusch; José L Zafra; Kjell Jorner; Kotaro Fukuda; Jonathan L Marshall; Iratxe Arrechea-Marcos; Guzmán L Espejo; Rocío Ponce Ortiz; Carlos J Gómez-García; Lev N Zakharov; Masayoshi Nakano; Henrik Ottosson; Juan Casado; Michael M Haley Journal: Nat Chem Date: 2016-05-23 Impact factor: 24.427
Authors: Fan Xu; Lukas Pfeifer; Stefano Crespi; Franco King-Chi Leung; Marc C A Stuart; Sander J Wezenberg; Ben L Feringa Journal: J Am Chem Soc Date: 2021-04-08 Impact factor: 15.419