Towards solar energy storage in the photochromic dihydroazulene-vinylheptafulvene system

Article abstract of DOI:10.1002/chem.201500100

Abstract One key challenge in the field of exploitation of solar energy is to store the energy and make it available on demand. One possibility is to use photochromic molecules that undergo light-induced isomerization to metastable isomers. Here we present efforts to develop solar thermal energy storage systems based on the dihydroazulene (DHA)/vinylheptafulvene (VHF) photo/thermoswitch. New DHA derivatives with one electron-withdrawing cyano group at position 1 and one or two phenyl substituents in the five-membered ring were prepared by using different synthetic routes. In particular, a diastereoselective reductive removal of one cyano group from DHAs incorporating two cyano groups at position 1 turned out to be most effective. Quantum chemical calculations reveal that the structural modifications provide two benefits relative to DHAs with two cyano groups at position 1: 1) The DHA-VHF energy difference is increased (i.e., higher energy capacity of metastable VHF isomer); 2) the Gibbs free energy of activation is increased for the energy-releasing VHF to DHA back-reaction. In fact, experimentally, these new derivatives were so reluctant to undergo the back-reaction at room temperature that they practically behaved as DHA to VHF one-way switches. Although lifetimes of years are at first attractive, which offers the ultimate control of energy release, for a real device it must of course be possible to trigger the back-reaction, which calls for further iterations in the future. Molecular solar thermal energy storage: Derivatives of the dihydroazulene (DHA)/vinylheptafulvene (VHF) photo/thermoswitch with only one cyano group at position 1 were prepared and investigated experimentally and theoretically. Compared to derivatives with two cyano groups, these compounds exhibit larger energy differences between DHA and VHF isomers and hence larger energy storage capacities, while the VHF to DHA back-reaction is on hold (see scheme).

Full text of DOI:10.1002/chem.201500100

DOI: 10.1002/chem.201500100
Full Paper
&
Renewable Resources
Towards Solar Energy Storage in the Photochromic
Dihydroazulene–Vinylheptafulvene System
Martina Cacciarini,[a, b] Anders B. Skov,[a] Martyn Jevric,[a] Anne S. Hansen,[a] Jonas Elm,[a]
Henrik G. Kjaergaard,[a] Kurt V. Mikkelsen,[a] and Mogens Brøndsted Nielsen*[a]
Abstract: One key challenge in the field of exploitation of
solar energy is to store the energy and make it available on
demand. One possibility is to use photochromic molecules
that undergo light-induced isomerization to metastable iso-
mers. Here we present efforts to develop solar thermal
energy storage systems based on the dihydroazulene (DHA)/
vinylheptafulvene (VHF) photo/thermoswitch. New DHA de-
rivatives with one electron-withdrawing cyano group at po-
sition 1 and one or two phenyl substituents in the five-mem-
bered ring were prepared by using different synthetic
routes. In particular, a diastereoselective reductive removal
of one cyano group from DHAs incorporating two cyano
groups at position 1 turned out to be most effective. Quan-
tum chemical calculations reveal that the structural modifi-
cations provide two benefits relative to DHAs with two
cyano groups at position 1: 1) The DHA–VHF energy differ-
ence is increased (i.e., higher energy capacity of metastable
VHF isomer); 2) the Gibbs free energy of activation is in-
creased for the energy-releasing VHF to DHA back-reaction.
In fact, experimentally, these new derivatives were so reluc-
tant to undergo the back-reaction at room temperature that
they practically behaved as DHA to VHF one-way switches.
Although lifetimes of years are at first attractive, which
offers the ultimate control of energy release, for a real
device it must of course be possible to trigger the back-reac-
tion, which calls for further iterations in the future.
Introduction
According to the International Energy Agency (IEA), the sun
could be the biggest source of energy by 2050, producing
over a quarter of the world’s electricity.[1] When it comes to the
engineering of organic molecules for light-harvesting, most of
the focus is on organic solar cells. However, the development
of organic molecules for solar energy storage has attracted in-
creasing interest in recent years.[2] Photochromic molecules
that undergo light-induced isomerizations to higher-energy,
metastable isomers present such targets, which would provide
closed-cycle systems of energy harvesting, storage, and re-
lease. A particularly attractive property would be the release
on-demand, with no environmentally unfriendly oxidation
products involved.[2] The dihydroazulene/vinylheptafulvene
Scheme 1. DHA–VHF isomerization. DHA numbering is also shown.
1 is a yellow photochromic compound, which upon irradiation
turns into the red VHF 2. The back-reaction of this metastable,
higher-energy isomer into DHA 1 is not light-induced but
occurs instead thermally (thermal back-reaction, TBR). Its rate is
solvent dependent and enhanced by treatment with mild
Lewis acids.[4]
Two significant challenges for using photochromic mole-
cules for energy storage are to reach a sufficiently high energy
density and at the same time to halt the TBR. Boulatov et al.[2a]
have estimated an upper limit for the energy density of photo-
chromic molecules of 1 MJkgꢀ1, which would be useful for real
applications. Recent calculations show that the Gibbs free
energy difference of the parent DHA/VHF 1/2 is 27.7 kJmolꢀ1
in vacuum, at the M06-2X/6-311+G(d) level, which corre-
sponds to a Gibbs free energy density of only 0.11 MJkgꢀ1
(energy stored in VHF).[5] The parent system is thus far from
optimum. Our calculations showed, however, that more than
a doubling of the energy density could be obtained by remov-
ing one of the two cyano groups (DHA/VHF 3/4, Figure 1).
Thus, a Gibbs free energy difference between isomers 3 and 4
of 57.7 kJmolꢀ1 was obtained, corresponding to an energy
density of 0.25 MJkgꢀ1. At the same time, calculations showed
pair 1/2 (DHA/VHF, Scheme 1) is
a photo/thermo-switch
system, which, if suitably functionalized, could be a candidate
for a molecular solar thermal energy storage system.[3] DHA
[a] Dr. M. Cacciarini, A. B. Skov, Dr. M. Jevric, A. S. Hansen, Dr. J. Elm,
Prof. Dr. H. G. Kjaergaard, Prof. Dr. K. V. Mikkelsen,
Prof. Dr. M. Brøndsted Nielsen
Department of Chemistry and
Center for Exploitation of Solar Energy, University of Copenhagen
Universitetsparken 5, 2100 Copenhagen (Denmark)
E-mail: mbn@kiku.dk
[b] Dr. M. Cacciarini
Department of Chemistry, University of Florence
via della Lastruccia 3–13, 50019 Sesto F.no (FI) (Italy)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500100.
Chem. Eur. J. 2015, 21, 1 – 9
1
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Figure 1. New DHA–VHF derivatives with only one cyano substituent at C1.
that the activation enthalpy of the TBR was increased from
92.3 kJmolꢀ1 (transition state from 2 to 1) to 106.9 kJmolꢀ1
(transition state from 4 to 3), in vacuum, at the PBE0/6-311+
G(d) level, and the TBR would thus be considerably slowed
down for this isomer. This is in line with previous findings that
the TBR relies on the ability to stabilize a negative charge at
C1.[6] The system is thereby attractive as it has the potential to
store more energy without fast release. Few examples with
substituents different from CN at C1 have been described, and
information on their photochromic properties was not report-
ed in detail.[7] However, a “monocyano-DHA” derivative bearing
a 4,5-dihydro-1,3-thiazole ring had a very slow TBR.[8] For these
reasons, we decided to develop a synthetic protocol for mono-
cyano-DHA/VHF molecules. The overall objective is to elucidate
whether this structural modification is useful for solar energy
storage and hence if it would be beneficial to combine it with
other structural modifications to reach ultimately the upper
energy storage limit.
Scheme 2. Attempted synthesis of monocyano-DHAs from monocyano-VHF
precursors. Oxidant system: i) NOBF4, ii) pyridine, with R=H; or i) tritylium
tetrafluoroborate, ii) NEt3, with R=Me.
obtain 4, although isolation of a pure fraction was still not pos-
1
sible. It was identified later by H NMR spectroscopic analysis
of an irradiated solution of DHA 3, obtained through the
second protocol (see below). Besides, no traces of DHA 3 were
detected upon heating 4 for up to 12 h.
Results and Discussion
To enhance the stability, we decided to replace the H at C1
with a methyl group by using the same synthetic strategy
(Scheme 2). The methyl phosphonate 11 was used in
a Horner–Wadsworth–Emmons reaction, providing the corre-
sponding VHF precursor 12 in 65% yield (26:1 E/Z isomers,
ratio evaluated by NMR analysis). The oxidation–elimination
step with tritylium tetrafluoroborate and NEt3, under strictly
degassed conditions, now furnished monocyano-VHF 13 as
a mixture of E/Z isomers in 45% yield with no traces of degra-
dation compounds. Nevertheless, attempts to convert 13 into
the corresponding DHA 14 by heating in 1,2-dichloroethane/
toluene were not successful. Although 13 proved to be more
stable than its analogue 4 and allowed full characterization of
this product, decomposition in solution was detected after
12 h in an NMR tube.
Synthesis
For the synthesis of monocyano-DHA 3, two protocols were in-
vestigated: The first protocol employed the corresponding
VHF 4 as a potential precursor and the second employed the
easily obtainable DHA 1. Although the second protocol was
successful, the first one, in which the VHF had to be construct-
ed from suitable precursors, was more problematic, as outlined
in Scheme 2. The first intermediate 5 was prepared from aceto-
phenone 6 and tropylium tetrafluoroborate 7, following the
procedure reported by Gobbi et al.[9] for the initial step to give
1. Then, a Horner–Wadsworth–Emmons reaction of 5 with the
ethyl phosphonate 8 in the presence of lithium bis(trimethylsi-
lyl)amide (LiHMDS) in THF was performed, providing 9 as a 6:1
mixture of E/Z isomers in 63% yield. The subsequent oxida-
tion/elimination of the VHF precursor 9 to VHF 4 was not
straightforward. The standard procedure[10] with tritylium tetra-
fluoroborate as oxidizing agent and triethylamine as base in
1,2-dichloroethane heated at reflux showed evidence of de-
composition and traces of azulene species, later assigned as
10, whereas 4 was difficult to isolate from the byproducts. Re-
ducing the amount of NEt3 (from excess to stoichiometric),
running the reaction at room temperature, and degassing the
solution excluded the formation of azulene 10 and limited the
decomposition. Replacement of tritylium tetrafluoroborate
with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) pro-
voked complete decomposition. The use of nitrosyl tetrafluoro-
borate as hydride abstractor in MeCN followed by deprotona-
tion with pyridine turned out to be the best conditions to
The second protocol was then followed to prepare 3. When
a solution of DHA 1 in THF was treated with diisobutylalumini-
um hydride (DIBAL-H) at 08C in the dark, no aldehyde forma-
tion was detected. Instead, replacement of a cyano group with
a hydride to give monocyano-DHA 3 occurred in 23% yield
(Scheme 3). Small amounts of azulene were detected. We note
that an analogous anomalous reductive decyanation of an a-
sulfonitrile induced by LiAlH4 was previously reported.[11]
The crude mixture was purified and only one pair of enan-
tiomers of DHA 3 was detected. A large coupling constant
(9.7 Hz) between H8a and H1 was measured in the 1H NMR
spectrum. From structures optimized at the B3LYP/6-311+
G(d,p) level of theory of the four possible isomers combined
with the Karplus Equation, we can assign the isomers to the
&
&
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
2
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
stant to the 1R,8aS/1S,8aR enantiomeric pair and the others to
the 1R,8aR/1S,8aS enantiomeric pair. Crystals of the major
isomer of 16 were grown from dichloroethane/heptane and X-
ray analysis confirmed that, in the major isomer, H1 and H8a
lie on the same side of the ring (Figure 3). The dihedral angle
V H1/C1/C8a/H8a is 2.98, in agreement with the values ex-
trapolated from the coupling constant and the two Ph rings
are twisted out of co-planarity by 62.38. Compound 16 readily
underwent ring opening to VHF 17 upon irradiation. This VHF
was also reluctant to undergo the TBR, but, in fact, photophys-
ical studies seem to indicate that some conversion is
nevertheless obtained after several hours at 708C in
MeCN (see below).
Scheme 3. Direct synthesis of monocyano-DHAs by stereoselective reduc-
tion. a) TBR of VHF 17 to some extent at elevated temperature for a long
period.
From calculations at the CAM-B3LYP/6-311+G(d)
level, the 1R,8aR diastereomer of 3 was found to be
6.02 kJmolꢀ1 lower in free energy than that of the
1R,8aS diastereomer in MeCN. Similarly, for the
1R,8aR diastereomer of 16 compared with the 1R,8aS
one, a free energy difference of 4.44 kJmolꢀ1 was
found in MeCN. This result indicates that the forma-
tion of the 1R,8aS/1S,8aR isomers as major products
is due to kinetic control.
A final synthetic objective was to develop a more
Figure 2. Possible DHA 3 diastereomers with corresponding Newman projection on C1–
C8a and dihedral angle V. The enantiomeric pair in the box seemed to be formed only
from the reaction shown in Scheme 3.
general method to introduce substituents straightfor-
wardly at C1 from DHA 1. Preliminary results have
enantiomeric pair (1R,8aS) and (1S,8aR), in which the hydride
was inserted from the same side of H8a, presumably for steric
reasons (Figure 2).
Although the yield of the reaction was modest, it furnished
3 directly from 1 and can be prepared on large scale.[10] NMR
experiments were performed to follow the ring opening of 3
to 4 upon irradiation and to confirm the structure of com-
pound 4 obtained through the first protocol. Photolysis of the
NMR sample showed a decrease of the characteristic signals,
H1 and H8a at d=4.71 and 3.33 ppm, and the increase of two
new signals at d=6.05 and 5.23 ppm, indicating protons
linked to sp2 carbons, H1 and H3 (see the Supporting Informa-
tion). Hence, DHA 3 was readily opened upon irradiation to
VHF 4, but again this compound did not return to DHA upon
heating.
With VHF 4 we have achieved a molecule for which the TBR
seems to be completely prevented. To shed further light on
Figure 3. Molecular structure of 16 (major isomer) from X-ray crystallograph-
the influence of having only one CN at C1, we decided to com-
bine this structural motif with a VHF-like molecule, which is
known to undergo an ultrafast ring-closure reaction. Daub and
co-workers[12] have reported the 2,3-diphenyl-substituted DHA
15 (Scheme 3). The corresponding VHF undergoes a very fast
TBR with a relaxation time t of only 20 s in MeCN, which corre-
sponds to a t1/2 of 14 s (for 2 in MeCN, t1/2 =218 min). Subject-
ing DHA 15 to a DIBAL-H reduction gave 16, as a 6:1 mixture
of diastereomers, in 16% yield. Coupling constants of 9.8 and
4.5 Hz between H8a and H1 were measured in the 1H NMR
spectrum of the major and the minor isomer, respectively.
Based on the geometry-optimized structures reported in
Figure 2, we assign the isomers with the large coupling con-
ic analysis.
been obtained through a reductive decyanation reaction, fol-
lowed by trapping of the carbanion with an electrophile fol-
lowing a protocol developed for derivatives of malononitrile.[13]
Although the method is not yet optimized or generalized for
“dicyano-DHAs”, it allowed the synthesis and characterization
of DHA 14 (Scheme 4). DHA 1 was treated with 3 molar equiv-
alents of lithium naphthalenide (LN)[14] in THF at ꢀ788C and
the intermediate was then quenched with methyliodide to
give 14 in 21% yield. Photolysis gave VHF 13, but no TBR was
observed.
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
3
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
being stabilized by a polar solvent such as MeCN. The possibili-
ty for TBR was monitored in time at 25 and 608C, but again no
signs of the ring-closure reaction were evidenced. Addition of
silver triflate (10–200 equivalent) or ZnCl2 in ethereal solution
(200 equiv), was previously reported as successfully enhancing
the TBR of 2 to 1.[4] Yet, this addition did not promote the TBR
of 4. Furthermore, monitoring the spectrum at 608C in pres-
ence of Lewis acids revealed partial decomposition of 4, to
give an azulene, tentatively assigned to 10 (see the Supporting
Information). The TBR was also investigated with VHF 13 in
DCE, both by heating the sample and by addition of Lewis
acids, but no TBR was evidenced.
Scheme 4. Direct synthesis of methylcyano DHA by reductive alkylation.
LN=lithium naphthalenide.
The UV/Vis absorption spectrum of 16 shows an absorption
maximum at 352 nm in DCE (Figure 5), whereas for DHA 15
Figure 4. UV/Vis absorption spectra of DHA 3 (black) and VHF 4 (red) in
MeCN (c, 2.6ꢁ10ꢀ5 m) and DCE (a, 2.8ꢁ10ꢀ5 m) and of DHA 14 (black)
and VHF 13 (red) in DCE (g, 4.7ꢁ10ꢀ5 m).
UV/Vis Spectroscopy and switching studies
The UV/Vis absorption spectra in MeCN (solid lines) and in di-
chloroethane (DCE, dashed lines) of DHA 3 and VHF 4 are
shown in Figure 4. In MeCN, compound 3 shows two main
peaks at 353 and 264 nm (S1–S0 and S2–S0 transitions). The
peak at 353 nm is at the same wavelength as for DHA 1 in
MeCN, thus the characteristic DHA absorption is not influenced
by removal of one CN. This seems plausible as it is not in con-
jugation with the p-system. Irradiation of the sample with
monochromatic light at 353 nm for 10 s intervals and for
a total duration of 150 s resulted in complete conversion to
VHF 4, which has an absorption maximum at 379 nm. This ab-
sorption is significantly blueshifted relative to that of VHF 2
(lmax 470 nm in MeCN), which is in accordance to it having one
less cyano group in conjugation with the p-system. The possi-
bility for 4 to undergo TBR was monitored for 24 h at 258C
and at 608C in MeCN. No signs of ring closure and formation
of DHA were detected, although after 24 h at 608C, small
changes in the spectrum appeared and a decrease in the VHF
absorption of roughly 4%, possibly due to decomposition.
Also in DCE, compound 3 was converted to 4 upon irradia-
tion (Figure 4, dashed lines). In this solvent the absorption
maxima of 3 and 4 are redshifted in comparison to those in
MeCN, revealing a solvatochromic effect (in particular for 4).
Thus, lmax for DHA 3 is at 358 nm and for VHF 4 at 403 nm.
The strong solvatochromism of the VHF absorption signals sig-
nificant charge delocalization in the ground-state structure,
Figure 5. UV/Vis absorption spectra of DHA 16 (black) and VHF 17 (red) in
MeCN (c 6.5ꢁ10ꢀ5 m) and DCE (a, 5.0ꢁ10ꢀ5 m).
a maximum at 340 nm in toluene was reported.[12] The absorp-
tion spectrum after irradiation of 16 at 365 nm showed new
maxima at 276 and 329 nm and a shoulder at 383 nm. Thus,
the spectrum of the generated monocyano-VHF 17 was quite
different from that of VHF 4. Yet, a similar behaviour was re-
ported by Daub and co-workers for the corresponding VHF of
15.[12] Attempts to induce the TBR of 17 upon heating at 508C
in DCE and adding Lewis acids (ZnCl2 and silver triflate) were
not successful. For a solution of the VHF, complete decomposi-
tion to an azulene was recorded after 60 d in deoxygenated
MeCN (see the Supporting Information). However, a partial
TBR, seemingly with some decomposition also, was induced
upon heating at 708C in MeCN for 72 h under strictly anaero-
bic conditions (see the Supporting Information). We cannot
quantify the degree of conversion, but observed that the spe-
cies generated, tentatively assigned to a DHA, was indeed
light-sensitive, and upon irradiation it was returned into a VHF.
Raman Spectroscopy
In Figure 6, the solid-state Raman spectra of DHAs 1, 15, and
16 are shown in the CN-stretching region. Both 1 and 15 have
&
&
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
4
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 1. Energies for the VHF to DHA back-reaction based on the lowest-
energy identified conformation of DHAs 1, 3, 15 and 16, and of the corre-
sponding VHFs.[a]
2!1[5]
DG
4!3[5]
DG
“VHF of 15”! 15 17!16
DH
DH
DG
DH
DG
DH
Vacuum ꢀ27.7 ꢀ35.2 ꢀ57.7 ꢀ62.3 ꢀ51.4
ꢀ60.3
ꢀ56.2
ꢀ46.8
ꢀ71.9 ꢀ76.5
ꢀ67.6 ꢀ77.4
ꢀ64.9 ꢀ74.7
CH
ꢀ22.7 ꢀ30.7 ꢀ57.2 ꢀ62.4 ꢀ48.2
ꢀ15.7 ꢀ22.9 ꢀ58.5 ꢀ64.1 ꢀ45.5
MeCN
[a] All values are presented in kJmolꢀ1 at 298 K.
and 4, respectively). As the entropy changes for the VHF to
DHA conversion are negative, the enthalpy changes become
more negative (corresponding to larger heat releases) than the
changes in Gibbs free energy.
Removal of a cyano group significantly increased the storage
capacity of DHA 3 compared with DHA 1.[5] Similarly, a signifi-
cant increase in the storage capacity (ignoring the increased
molecular weight) is observed when including a phenyl group
in position 3. Indeed, the effect is found to be additive with
a DG of ꢀ71.9 kJmolꢀ1 for 16 in vacuum. The relative stability
of 1 and 15 turned out to be slightly dependent on solvent, as
the solvent polarity increases, the stability decreases. This
effect is significantly lower for the monocyano derivatives 3
and 16.
Figure 6. Raman spectra of solid state of 16 (g), 15 (c), and 1 (a) in
the CN-stretching region.
symmetric and asymmetric CN-stretching vibrations, but we
were unable to distinguish them. The CN-stretching vibration
in DHA 1 is observed at 2247 cmꢀ1 with a full width at half
maximum (FWHM) of 11 cmꢀ1. The harmonic symmetric and
asymmetric CN-stretching vibrations in 1 are calculated to be
within 6 cmꢀ1 of each other, hence the band observed at
2247 cmꢀ1 contains both CN-stretching vibrations. The band
observed in the CN-stretching region for 15 at 2247 cmꢀ1 has
a FWHM of 17 cmꢀ1, which also contains both the symmetric
and asymmetric CN-stretching vibrations that are calculated to
be within 8 cmꢀ1 of each other. The substitution of an addi-
tional Ph group to 1 has no effect on the frequency, however,
the width of the band increases by about 6 cmꢀ1 suggesting
that the transitions are split slightly more in 15 than in 1. The
CN-stretching vibration in 16 is observed at 2238 cmꢀ1 and is
thus redshifted relative to the CN-stretching vibrations in 15
and 1. The FWHM of the CN-stretching band in 16 is 11 cmꢀ1
as for 1.
Barriers DGVHF!TS for the TBR from s-cis-VHF to the transition
state are shown in Table 2 for the formation of DHAs 1, 3, 15,
Table 2. Calculated back-reaction barrier DGVHF!TS for conversion of s-cis
VHFs to the corresponding DHAs.[a]
2!1[5]
4!3[5]
“VHF of 15”!15
17!16
Vacuum
CH
MeCN
119.5
115.0
103.4
133.0
128.6
128.8
112.8
113.4
73.8
125.7
125.4
123.3
[a] All values are presented in kJmolꢀ1 at 298 K.
Calculations: Thermodynamic and kinetic properties
and 16 in vacuum, CH, and MeCN. In vacuum a slight decrease
in the barrier of 15 compared with 1 is observed. A similar
effect is seen for 16 compared with 3. Again, the dicyano de-
rivatives 1 and 15 are observed to be significantly solvent de-
pendent, with a decrease in the TBR barrier of almost
40 kJmolꢀ1 in the case of formation of 15 when changing from
vacuum to MeCN for M06-2X. Conversion of the VHF of 15 has
a low barrier in MeCN of 73.8 kJmolꢀ1, in agreement with the
experimental finding of ultrafast switching.[12] However, the
barrier for the monocyano derivatives is seen to be almost un-
affected by solvent polarity (4!3 and 17!16).
To shed further light on the thermochemical properties and ki-
netics of the TBR of the 3-phenyl derivatives and hence the ap-
plicability for use as molecular heat batteries, a computational
study was conducted and the results compared to energies re-
cently calculated for the 1/2 and 3/4 pairs.[5] The storage Gibbs
free energies and storage enthalpies, defined as the relative
energies between the lowest-energy VHF and DHA conforma-
tions for DHAs 1, 3, 15, and 16, were calculated by using the
CAM-B3LYP, M06-2X, and PBE0 functionals with the 6-311+
G(d) basis set in vacuum, cyclohexane (CH), and MeCN (see the
Supporting Information). Table 1 shows the reported data from
M06-2X, the functional which previously was found most opti-
mum.[5] For the VHFs of 15 and 16, the s-cis conformations
were lowest in energy (in agreement with the very fast ring
closure of the VHF of 15), whereas the s-trans conformations
were most stable for the VHFs of 1 and 3 (corresponding to 2
From the calculations alone for the relative Gibbs free ener-
gies and free energy of activation for the TBR, the DHA/VHF
pair 16/17 shows attractive features for applications as a solar
heat battery, with a high storage capacity and a large barrier
for the TBR. The barrier corresponds to a half-life of >10 years
at ambient temperature, but a half-life of about 160 h at 708C.
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
The calculations thus support our experimental results. Indeed,
no TBR was evidenced at ambient temperature, but after 72 h
at 708C, some conversion to a DHA did indeed seem to occur
(likely with some degradation also).
Syntheses
2-(2,4,6-Cycloheptatriene-1-yl)-1-phenylethanone (5): Acetophe-
none 6 (8.90 g, 77.0 mmol) and acetic acid (2.4 mL) were added to
a solution of 7 (12.0 g, 70.0 mmol) in MeOH (120 mL). The mixture
was stirred under Ar for 6 h at RT. A second portion of 7 (4.0 g,
22.0 mmol) was added, and after stirring for an additional 8 h,
a visible color change from light yellow to dark brown had oc-
curred. Then, water (200 mL) was added and the mixture was ex-
tracted with Et2O (3ꢁ100 mL). The combined organic layers were
dried over Na2SO4 and the crude product was purified by dry
column vacuum chromatography (SiO2, CH2Cl2 in heptane, 5%
steps from 0 to 75%, 20 mL fractions) to give the product 5
Conclusion
The preparation of monocyano-VHF/DHA derivatives was ac-
complished, either by building up the molecules from simple
precursors or by direct modification of the parent DHA 1 by re-
ductive decyanation followed by treatment with an electro-
phile; only the latter method gave access to the DHAs. These
monocyano-DHAs retained the ability to undergo ring opening
to the VHF forms upon irradiation. Replacement of a cyano
group with a hydrogen or methyl group inhibits, however, the
back-reaction. High activation energy barriers were confirmed
by calculations (lifetimes of years) and are in principle a desira-
ble outcome, allowing for control of energy release, together
with the fact that calculations show that the energy storage
capacity is increased significantly by having only one cyano
group instead of two such groups. Yet, an obvious challenge is
to trigger the back-reaction in a suitable manner, which is the
focus of future studies.
1
(10.13 g, 63%) as a dark-red oil. H NMR (500 MHz, CDCl3): d=7.96
(dd, J=8.3, 1.0 Hz, 2H), 7.57 (tt, J=7.1, 1.1 Hz, 1H), 7.47 (dd, J=
8.3, 7.1 Hz, 2H), 6.66 (t, J=2.9 Hz, 2H), 6.24–6.19 (m, 2H), 5.30 (dd,
J=6.1, 2.9 Hz, 2H), 3.35 (m, 2H), 2.51 ppm (m, 1H); 13C NMR
(125 MHz, CDCl3): d=198.86, 137.10, 133.25, 131.15, 128.75, 128.22,
125.47, 125.31, 41.63, 34.91 ppm (5 carbon atoms missing due
to overlap); HRMS (ES): m/z calcd for C15H14ONa+: 233.09369
[M+Na]+; found: 233.09372; elemental analysis calcd (%) for
C15H14O: C 85.68, H: 6.71; found: C 85.35, H: 6.45.
E/Z-4-(Cyclohepta-2,4,6-triene-1-yl)-3-phenylbut-2-enenitrile (9):
Lithium bis(trimethylsilyl)amide (LiHMDS, 1m solution in toluene,
29.9 mL, 29.9 mmol) was added to a solution of 8 (5.29 g, 4.83 mL,
29.9 mmol) in dry THF (80 mL) under Ar at 08C. The mixture was
stirred at 08C for 30 min, then a solution of 5 (5.78 g, 27.5 mmol)
in THF (30 mL) was added. After stirring for 3 h at RT, water
(200 mL) was added and the mixture was extracted with CH2Cl2
(2ꢁ100 mL). The organic phase was washed with water (100 mL)
and dried over MgSO4. The crude dark-red oil was purified by dry
column vacuum chromatography (SiO2, toluene in heptane, 5%
steps, from 0 to 100%, Rf (Z isomer): 0.34, Rf (E isomer): 0.56 in tolu-
ene) to give 9 (6:1 E/Z mixture), as a yellow oil (4.0 g, 63%). A pure
fraction of the E isomer was obtained as a yellow solid, which al-
lowed its full characterization. M.p.: 71–728C (E isomer); 1H NMR
(500 MHz, CDCl3): d=7.50–7.30 (m, 5H), 6.57 (t, J=3.1 Hz, 2H),
6.21–6.07 (m, 2H), 5.59 (s, 1H), 5.22 (dd, J=5.8, 3.2 Hz, 2H), 3.23
(d, J=7.9 Hz, 2H), 1.84 ppm (m, 1H); 13C NMR (125 MHz, CDCl3):
d=163.08, 137.55, 131.09, 130.21, 129.03, 126.44, 125.62, 124.41,
117.56, 97.89, 37.79, 36.70 ppm (5 carbon atoms missing due to
overlap); HRMS (MALDI, dithranol): m/z calcd for C17H16N+:
234.12773 [M+H]+, for C17H15Na+: 256.10967 [M+Na]+; found:
234.12785 [M+H]+, 256.10983 [M+Na]+; elemental analysis calcd
(%) for C17H15N: C 87.52, H: 6.48, N 6.00; found: C 87.56, H: 6.31, N:
5.95.
Experimental Section
General procedures
All reactions were carried out under an argon atmosphere, using
HPLC-grade solvents. THF was distilled over Na/benzophenone.
Acetonitrile was dried over activated 3 ꢂ molecular sieves. All han-
dling of DHAs was done in the dark, with flasks and columns wrap-
ped in tin foil. TLC was carried out on commercially available pre-
coated plates (Silica 60) with fluorescence indicator P254. Spectro-
photometric measurements were carried out in a 1 cm path length
cuvette at 258C, unless otherwise stated. Spectrophotometric anal-
ysis of the ring-opening reaction was conducted by irradiating a so-
lution of DHA (conc. 2–7ꢁ10ꢀ5 m) in a cuvette with a 150 W Xenon
arc lamp, equipped with a monochromator (the chosen wave-
length was the longest-wavelength absorption maximum) or by
using a 365 nm lamp. The TBR was attempted by heating the
sample in a cuvette in a Peltier unit in the UV/Vis spectrophotome-
ter. NMR spectra were acquired on a 500 MHz Bruker instrument
equipped with a cryoprobe. All chemical shift values in 1H and
13C NMR spectra are referenced to the residual solvent peak (CDCl3
dH =7.26 ppm, dC =77.16 ppm, C6D6 dH =7.16 ppm, C6D6 dC =
128.06 ppm, CD3CN dH =1.94 ppm, dC =118.26 ppm). High-resolu-
tion mass spectrometry (HRMS) was performed by using either
electrospray ionization (ESI) or matrix-assisted laser desorption ioni-
zation (MALDI); FT-ICR=Fourier transform ion cyclotron resonance.
Tropylium tetrafluoroborate 7 and tritylium tetrafluoroborate were
prepared following reported procedures[10] as was a solution of
lithium naphthalenide (LN).[14] Raman spectra were recorded with
a RFS 100 FT-Raman instrument (Bruker), fitted with a 1064 nm
Nd:YAG laser. The spectra were recorded with a 4 cmꢀ1 resolution,
1000 scans and 350 mW laser power. OriginPro 8.6 software was
used to produce the plots. Single crystal structure was determined
using a Bruker D8-Venture diffractometer using MoKa radiation.
Data were reduced using Apex and solved using Olex2 v. 1.2.
CCDC 1042291 (16) contain the supplementary crystallographic
4-(Cyclohepta-2,4,6-trien-1-ylidene)-3-phenylbut-2-enenitrile (4):
A 0.582m solution of NOBF4 in dry MeCN (3.9 mL, 4.6 mmol) was
slowly added at ꢀ208C to a solution of 9 (0.529 g, 2.3 mmol of 6:1
E/Z mixture) in dry MeCN (20 mL) under Ar, during which the solu-
tion changed color from light yellow to dark orange. The reaction
mixture was stirred at ꢀ208C for 30 min, then it was diluted with
dry CH2Cl2 (20 mL) and pyridine (0.18 mL, 4.6 mmol) was added.
The reaction was stirred at ꢀ208C for 30 min and then poured into
water. The crude mixture was extracted with CH2Cl2 (3ꢁ50 mL) and
the organic phase washed with aq. HCl (1m, 50 mL) and dried over
Na2SO4. The crude residue was purified by dry column vacuum
chromatography (SiO2, EtOAc in heptane, 0.5% steps from 0 to
20%, 10 mL fractions, Rf (4): 0.43 in 20% EtOAc/Heptane) to give 4
&
&
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
contaminated by decomposition products. 1H NMR (500 MHz,
CDCl3, selected signals from irradiation of a solution of 3): d=6.35
(ddd, J=11.9, 2.4, 1.0 Hz, 1H), 6.05 (s, 1H), 6.03–5.97 (m, 1H), 5.95–
5.88 (m, 2H), 5.73–5.67 (m, 1H), 5.66–5.57 (m, 1H), 5.23 ppm (d, J=
4-(Cyclohepta-2,4,6-trien-1-yl)-2-methyl-3-phenylbut-2-enenitrile
(12): Lithium bis(trimethylsilyl)amide (LiHMDS, 1m solution in tolu-
ene, 6.3 mL, 6.3 mmol) was added to a solution of diethyl 2-cya-
noethylphosphonate 11 (1.196 g, 6.3 mmol) in dry THF (20 mL)
under Ar at 08C. The mixture was stirred at 08C for 30 min, then
a solution of 5 (1.1957 g, 5.7 mmol) in THF (10 mL) was added.
After stirring for 3 h at RT, water (50 mL) was added and the mix-
ture extracted with CH2Cl2 (3ꢁ50 mL). The organic phase was
washed with water (50 mL) and dried over MgSO4. The crude dark-
red oil was purified by flash column chromatography (SiO2, 75%
toluene in heptane, Rf =0.42) to give 12 as 26:1 E/Z mixture as
1
1.0 Hz, 1H). H NMR (500 MHz, CD3CN): d=7.52–7.35 (m, 5H), 6.40–
6.36 (m, 1H), 6.09–5.99 (m, 1H), 5.99–5.90 (m, 3H), 5.82–5.79 (m,
1H), 5.74–5.66 (m, 1H), 5.50 ppm (s, 1H); 13C NMR (125 MHz,
CD3CN): d=159.51, 147.03, 140.65, 138.95, 134.71, 133.64, 132.92,
131.00, 130.86, 130.72, 129.94 (2ꢁCH), 128.34 (2ꢁCH), 120.42,
118.80, 95.29 ppm.
1
2-Phenyl-1,8a-dihydroazulene-1-carbonitrile (3): To a solution of
DHA 1 (256 mg, 1.0 mmol) in dry THF (25 mL) shielded from light,
DIBAL-H (1m solution in heptanes, 2 mL) was slowly added at 08C
under Ar. The reaction mixture was stirred at 08C for 2 h, then aq.
HCl (1m, 50 mL) was added and the mixture stirred for 10 min at
RT. The crude product was extracted with CH2Cl2 (2ꢁ100 mL) and
the organic phase dried over Na2SO4. The crude residue was puri-
fied by flash column chromatography (SiO2, 70% toluene in hep-
tane, Rf (3): 0.48 (20% EtOAc/heptane)) to give 3 (60 mg, 26%) as
a yellow oil. 1H NMR (500 MHz, CDCl3): d=7.59–7.56 (m, 2H, Ph),
7.44–7.40 (m, 2H, Ph), 7.38–7.41 (m, 1H, Ph), 6.74 (d, J=2.0 Hz, 1H,
H-3), 6.51 (A part of ABMX system, JAB =6.4 Hz, JAM =4.6 Hz, 1H, H-
5 or H-6), 6.42 (B part of ABMX system, JAB =6.4 Hz, JBX =4.7 Hz, 1H,
H-6 or H-5), 6.26–6.21 (m, 2H, H-4 and H-7), 5.80 (dd, J=9.8,
3.9 Hz, 1H, H-8), 4.71 (dd, J=9.8, 2.0 Hz, 1H, H-1), 3.33 ppm (ddt,
J=9.8, 3.9, 2.1 Hz 1H, H-8a); 13C NMR (125 MHz, CDCl3): d=143.83,
143.04, 132.99, 130.99, 130.19, 129.94, 129.24, 129.04, 126.89,
126.03, 121.28, 118.51, 118.12, 41.66, 40.52 ppm (2 carbon atoms
missing due to overlap); HRMS (MALDI, dithranol): m/z calcd for
C17H14N+: 232.11208 [M+H]+, C17H13NNa+: 254.09402 [M+Na]+;
found: 232.11229 [M+H]+, 254.09430 [M+Na]+; elemental analysis
calcd (%) for C17H13N: C 88.28, H: 5.67, N 6.06; found: C 87.40, H:
5.42, N: 6.25.
a light-yellow oil (0.91 g, 65%). H NMR (500 MHz, CDCl3): d=7.40–
7.30 (m, 3H), 7.09–7.04 (m, 2H), 6.61–6.58 (m, 2H), 6.20–6.15 (m,
2H), 5.22 (dd, J=5.6, 3.6 Hz, 2H), 3.13 (m, 2H), 1.85 (s, 3H),
1.64 ppm (m, 1H); 13C NMR (125 MHz, CDCl3): d=157.02, 137.66,
131.01, 128.66, 128.40, 127.43, 125.33, 124.76, 119.88, 107.87, 41.12,
37.25, 17.91 ppm (5 carbon atoms missing due to overlap); HRMS
(MALDI, dithranol): m/z calcd for C18H18N+: 248.14338 [M+H]+,
C18H17NNa+: 270.12532 [M+Na]+; found: 248.14335 [M+H]+,
270.12529 [M+Na]+.
4-(Cyclohepta-2,4,6-trien-1-ylidene)-2-methyl-3-phenylbut-2-ene-
nitrile (13): Tritylium tetrafluoroborate (0.242 g, 0.733 mmol) was
added to a solution of 12 (0.1648 g, 0.66 mmol) in dichloroethane
(10 mL) under Ar, and the mixture was heated at 808C for 1 h. The
mixture was cooled to 08C and diluted with toluene (10 mL). Then,
NEt3 (0.073 g, 0.0725 mmol) was added slowly over 30 min, at
which point the color changed to dark red. After stirring at 08C for
1 h, the reaction was concentrated to dryness. The resulting dark-
red residue was purified by flash column chromatography (SiO2,
80% toluene in heptane) under N2 atm and using degassed sol-
1
vents, giving 13 (73 mg, 45%). H NMR (500 MHz, CDCl3): d=7.44–
7.31 (m, 4H), 7.19–7.14 (m, 1H), 6.21 (dd, J=12.0, 1.7 Hz, 1H), 6.14
(s, 1H), 5.95–5.75 (m, 3H), 5.48–5.30 (m, 2H), 1.84 ppm (s, 3H);
13C NMR (125 MHz, CDCl3): d=153.89, 144.06, 140.86, 137.52,
133.25, 132.67, 132.08, 129.16, 129.13, 129.04 (2xCH), 128.67 (2ꢁ
CH), 128.57, 123.42, 120.63, 104.57, 17.97 ppm; HRMS (MALDI, di-
thranol): m/z calcd for C18H16N+: 246.12773 [M+H]+, C18H15NNa+:
268.10967 [M+Na]+, C18H15NK+: 284.08361 [M+K]+; found:
246.12792 [M+H]+, 268.10991 [M+Na]+, 284.08386 [M+K]+.
2,3-Diphenyl-1,8a-dihydroazulene-1-carbonitrile (16): DIBAL-H
(1m solution in heptanes, 4.2 mL) was slowly added at 08C under
Ar to a solution of DHA 15 (700 mg, 2.1 mmol) in dry THF (50 mL)
shielded from the light. The reaction mixture was stirred at 08C for
2.5 h, then aq. HCl (1m, 50 mL) was added and the mixture stirred
for 10 min at RT. The crude product was extracted with CH2Cl2 (2ꢁ
50 mL) and the organic phase dried over Na2SO4. The crude residue
was purified by flash column chromatography under N2 atm (SiO2,
60% toluene in heptane, Rf-16: 0.26) to give 16 (105 mg, 16%) as
1
a yellowish solid and 5:1 mixture of isomers (detected by H NMR
2-Phenyl-1,8a-dihydroazulene-1-methyl-1-carbonitrile (14):
A
spectroscopy). A pure sample of the major isomer was isolated for
full characterization. M.p. 212–2158C; 1H NMR (500 MHz, CDCl3):
d=7.35–7.32 (m, 3H), 7.23–7.20 (m, 3H), 7.17–7.13 (m, 4H), 6.50 (A
part of ABMX system, JAB =11.0 Hz, JAM =6.0 Hz, 1H, H-5), 6.45 (B
part of ABMX system, JAB =11.0 Hz, JBX =5.8 Hz 1H, H-6), 6.29 (ddd,
J=9.9, 5.8, 2.0 Hz, 1H, H-7), 5.88 (d, J=6.0 Hz, 1H, H-4), 5.84 (dd,
J=9.9, 4.2 Hz, 1H, H-8), 4.95 (d, J=9.8 Hz, 1H, H-1), 3.30 ppm (ddt,
J=9.8, 4.2, 2.0 Hz, 1H, H-8a); 1H NMR (500 MHz, CDCl3, selected sig-
0.66m stock solution of LN (2 mL, 1.32 mmol) was added to a de-
gassed solution of 1 (110 mg, 0.43 mmol) in dry THF (10 mL) under
Ar at ꢀ788C. After 1 h at ꢀ788C, MeI (0.15 mL, 2.41 mmol) was
added and the solution allowed to reach RT overnight. Then, aq.
NH4Cl (10 mL) was added. The mixture was diluted with Et2O
(10 mL) and the organic phase separated and dried over NaSO4.
The crude residue was purified by flash column chromatography
under N2 atm (SiO2, 60% toluene in heptane, Rf (14): 0.18) to give
14 (22 mg, 21%) as a yellow oil. 1H NMR (500 MHz, CDCl3): d=
7.76–7.72 (m, 2H), 7.44–7.37 (m, 2H), 7.38–7.32 (m, 1H), 6.61 (s,
1H), 6.54 (dd, J=11.2, 6.2 Hz, 1H, H-6 or H-5), 6.40 (dd, J=11.2,
6.0 Hz, 1H, H-5 or H-6), 6.23–6.18 (m, 2H, H-7 and H-4), 5.79 (dd,
J=9.9, 4.0 Hz, 1H, H-8), 2.91 (dt, J=4.0, 2.0 Hz, 1H, H-8a),
1.77 ppm (s, 3H); 13C NMR (125 MHz, CDCl3): d=148.96, 140.95,
133.17, 131.03, 129.77, 129.75, 128.98 (2ꢁCH), 126.87 (2ꢁCH),
126.50, 121.91, 121.87 (CN), 118.53, 52.66 (C-8a), 49.06 (C1),
27.28 ppm (Me). HRMS (ES): m/z calcd for C18H15NNa+: 268.10967
[M+Na]+; found: 268.11035.
nals for the minor isomer): d=6.37 (A part of ABX system, JAB
=
11.0 Hz, JAX =6.0 Hz, 1H, H-6), 6.20 (ddd, J=10.0, 6.0, 1.5 Hz, 1H, H-
7), 5.92 (d, J=6.7 Hz, 1H, H-4), 5.46 (dd, J=9.9, 3.5 Hz, 1H, H-8),
4.27 (d, J=4.5 Hz, 1H, H-1), 3.52 ppm (m, 1H, H-8a); 13C NMR
(125 MHz, CDCl3): d=144.82, 142.60, 139.65, 134.27, 133.63, 131.20,
129.99, 129.41, 129.18, 128.91, 128.51, 128.46, 128.37, 126.87,
125.44, 121.16, 118.44, 118.42, 42.18, 40.84 ppm (3 carbon atoms
missing due to overlap); HRMS (ES): m/z calcd for C23H17NNa+:
330.12532 [M+Na]+; found: 330.12621; elemental analysis calcd
(%) for C23H17N: C 89.87, H: 5.57, N 4.56; found: C 89.77, H: 5.66, N:
4.66.
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
7
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[4] C. R. Parker, C. G. Tortzen, S. L. Broman, M. Schau-Magnussen, K. Kilsꢃ,
M. B. Nielsen, Chem. Commun. 2011, 47, 6102–6104.
Acknowledgements
[5] S. T. Olsen, J. Elm, F. E. Storm, A. N. Gejl, A. S. Hansen, M. H. Hansen, J. R.
Nikolajsen, M. B. Nielsen, H. G. Kjaergaard, K. V. Mikkelsen, J. Phys. Chem.
A 2015, 119, 896–904.
The University of Copenhagen, Carlsbergfondet, and the
Danish Center for Scientific Computing are acknowledged for
financial support.
[6] S. L. Broman, M. Jevric, M. B. Nielsen, Chem. Eur. J. 2013, 19, 9542–9548.
[7] a) J. Daub, S. Gierisch, U. Klement, T. Knçchel, G. Maas, U. Seitz, Chem.
Ber. 1986, 119, 2631–2646; b) J. Daub, S. Gierisch, T. Knochel, E. Salbeck,
G. Z. Maas, Z. Naturforsch. B 1986, 41b, 1151–1160; c) T. Toda, K. Saito, T.
Mukai, Tetrahedron Lett. 1972, 13, 1981–1984; d) H. E. Zimmermann,
C. J. Samuel, J. Am. Chem. Soc. 1975, 97, 4025–4036; e) H. Prinzbach,
H.-J. Herr, H. Regel, Angew. Chem. Int. Ed. Engl. 1972, 11, 131–133;
Angew. Chem. 1972, 84, 113–115; f) H. Prinzbach, H.-J. Herr, Angew.
Chem. Int. Ed. Engl. 1972, 11, 135–136; Angew. Chem. 1972, 84, 117–
118.
Keywords: electrocyclic reactions
· energy conversion ·
molecular devices · photochromism · renewable resources ·
quantum chemistry
[1] IEA, 2014, Technology Roadmap: Solar Photovoltaic Energy and Tech-
nology Roadmap: Solar Thermal Electricity, OECD/IEA, 2014 Edition,
Paris.
[8] M. Cacciarini, E. A. Della Pia, M. B. Nielsen, Eur. J. Org. Chem. 2012,
6064–6069.
[2] a) T. J. Kucharski, Y. Tian, S. Akbulatov, R. Boulatov, Energy Environ. Sci.
´
2011, 4, 4449–4472; b) K. Moth-Poulsen, D. Coso, K. Bçrjesson, N. Vino-
[9] L. Gobbi, P. Seiler, F. Diederich, Helv. Chim. Acta 2001, 84, 743–776.
[10] S. L. Broman, S. L. Brand, C. R. Parker, M. ꢂ. Petersen, C. G. Tortzen, A.
Kadziola, K. Kilsꢃ, M. B. Nielsen, ARKIVOC 2011, 2011, 51–67.
[11] a) J.-M. Mattalia, A. Samat, M. Chanon, J. Chem. Soc. Perkin Trans.
1 1991, 1769–1770; b) J.-M. Mattalia, C. Marchi-Delapierre, H. Hazimeh,
M. Chanon, ARKIVOC 2006, 2006, 90–118.
kurov, S. Meier, A. Majumdar, K. P. C. Vollhardt, R. A. Segalman, Energy
Environ. Sci. 2012, 5, 8534–8537; c) K. Moth-Poulsen in Organic Synthe-
sis and Molecular Engineering (Ed.: M. B. Nielsen), Wiley, Hoboken, 2014,
pp. 179–196; d) T. J. Kucharski, N. Ferralis, A. M. Kolpak, J. O. Zheng,
D. G. Nocera, J. C. Grossman, Nat. Chem. 2014, 6, 441–447.
[3] a) J. Daub, T. Knçchel, A. Mannschreck, Angew. Chem. Int. Ed. Engl. 1984,
23, 960–961; Angew. Chem. 1984, 96, 980–981; b) T. Mrozek, A. Ajaya-
ghosh, J. Daub in Molecular Switches (Eds: B. L. Feringa), Wiley-VCH,
Weinheim, 2001, pp. 63–106; c) M. B. Nielsen, S. L. Broman, M. ꢂ. Pe-
tersen, A. S. Andersson, T. S. Jensen, K. Kilsꢃ, A. Kadziola, Pure Appl.
Chem. 2010, 82, 843–852; d) M. Cacciarini, S. L. Broman, M. B. Nielsen,
ARKIVOC 2014, 2014, 249–263; e) S. L. Broman, M. B. Nielsen, Phys.
Chem. Chem. Phys. 2014, 16, 21172–21182.
[12] a) T. Mrozek, H. Gçrner, J. Daub, Chem. Commun. 1999, 1487–1488; b) T.
Mrozek, H. Gçrner, J. Daub, Chem. Eur. J. 2001, 7, 1028–1040.
[13] T.-Y. Tsai, K.-S. Shia, H.-J. Liu, Synlett 2003, 97–101.
[14] H.-J. Liu, X. Shang, Tetrahedron Lett. 1998, 39, 367–370.
Received: January 10, 2015
Published online on && &&, 2015
&
&
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
8
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
&
Renewable Resources
M. Cacciarini, A. B. Skov, M. Jevric,
A. S. Hansen, J. Elm, H. G. Kjaergaard,
K. V. Mikkelsen, M. Brøndsted Nielsen*
&& – &&
Molecular solar thermal energy stor-
age: Derivatives of the dihydroazulene
(DHA)/vinylheptafulvene (VHF) photo/
thermoswitch with only one cyano
group at position 1 were prepared and
investigated experimentally and theo-
retically. Compared to derivatives with
two cyano groups, these compounds
exhibit larger energy differences be-
tween DHA and VHF isomers and hence
larger energy storage capacities, while
the VHF to DHA back-reaction is on
hold (see scheme).
Towards Solar Energy Storage in the
Photochromic Dihydroazulene–
Vinylheptafulvene System
Chem. Eur. J. 2015, 21, 1 – 9
www.chemeurj.org
9
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ