Aromaticity-Controlled Energy Storage Capacity of the Dihydroazulene-Vinylheptafulvene Photochromic System

Article abstract of DOI:10.1002/chem.201601190

Photochemical conversion of molecules into high-energy isomers that, after a stimulus, return to the original isomer presents a closed-cycle of light-harvesting, energy storage, and release. One challenge is to achieve a sufficiently high energy storage capacity. Here, we present efforts to tune the dihydroazulene/vinylheptafulvene (DHA/VHF) couple through loss/gain of aromaticity. Two derivatives were prepared, one with aromatic stabilization of DHA and the second of VHF. The consequences for the switching properties were elucidated. For the first type, sigmatropic rearrangements of DHA occurred upon irradiation. Formation of a VHF complex could be induced by a Lewis acid, but addition of H₂O resulted in immediate regeneration of DHA. For the second type, the VHF was too stable to convert into DHA. Calculations support the results and provide new targets. We predict that by removing one of the two CN groups at C-1 of the aromatic DHA, the heat storage capacity will be further increased, as will the life-time of the VHF. Calculations also reveal that a CN group at the fulvene ring retards the back-reaction, and we show synthetically that it can be introduced regioselectively.

Full text of DOI:10.1002/chem.201601190

DOI: 10.1002/chem.201601190
Full Paper
&
Energy Storage and Conversion
Aromaticity-Controlled Energy Storage Capacity of the
Dihydroazulene-Vinylheptafulvene Photochromic System
Anders B. Skov,[a] Søren Lindbæk Broman,*[a] Anders S. Gertsen,[a] Jonas Elm,*[b]
Martyn Jevric,[a] Martina Cacciarini,[a, c] Anders Kadziola,[a] Kurt V. Mikkelsen,*[a] and
Mogens Brøndsted Nielsen*[a]
Abstract: Photochemical conversion of molecules into high-
energy isomers that, after a stimulus, return to the original
isomer presents a closed-cycle of light-harvesting, energy
storage, and release. One challenge is to achieve a sufficiently
high energy storage capacity. Here, we present efforts to
tune the dihydroazulene/vinylheptafulvene (DHA/VHF)
couple through loss/gain of aromaticity. Two derivatives
were prepared, one with aromatic stabilization of DHA and
the second of VHF. The consequences for the switching
properties were elucidated. For the first type, sigmatropic re-
arrangements of DHA occurred upon irradiation. Formation
of a VHF complex could be induced by a Lewis acid, but ad-
dition of H2O resulted in immediate regeneration of DHA.
For the second type, the VHF was too stable to convert into
DHA. Calculations support the results and provide new tar-
gets. We predict that by removing one of the two CN
groups at C-1 of the aromatic DHA, the heat storage capaci-
ty will be further increased, as will the life-time of the VHF.
Calculations also reveal that a CN group at the fulvene ring
retards the back-reaction, and we show synthetically that it
can be introduced regioselectively.
Introduction
1,8a-Dihydroazulene-1,1-dicarbonitriles (DHAs) are photochro-
mic molecules,[1] and this property was recently exploited for
light-induced conductance switching in molecular electronics
devices.[2] Thus, DHA 1a is converted quantitatively into the
isomeric vinylheptafulvene (VHF) 1b upon irradiation
(Scheme 1), which, in time, returns through a thermally in-
duced back-reaction to the more stable DHA. Enhancing the
energy difference between two isomers that are convertible by
light in at least one direction has also attracted interest as
a means of storing solar energy.[3] The DHA/VHF system is only
photoactive in one direction (with a high quantum yield of
55% for 1a in MeCN[4]), rendering it particularly attractive for
[a] A. B. Skov, Dr. S. L. Broman, A. S. Gertsen, Dr. M. Jevric,
Prof. Dr. M. Cacciarini, Prof. Dr. A. Kadziola, Prof. Dr. K. V. Mikkelsen,
Prof. Dr. M. B. Nielsen
Department of Chemistry, University of Copenhagen
Universitetsparken 5, 2100 Copenhagen Ø (Denmark)
E-mail: slb@kiku.dk
kmi@chem.ku.dk
mbn@chem.ku.dk
Scheme 1. Structures of the original DHA/VHF system 1a/1b with atom
numbering and new benzo-fused systems. Rings that change from aromatic
(red color) to nonaromatic (blue color) are highlighted.
[b] Dr. J. Elm
University of Helsinki, Department of Physics
Gustaf Hꢀllstrçmin katu 2a, 00014 Helsinki (Finland)
E-mail: jonas.elm@helsinki.fi
[c] Prof. Dr. M. Cacciarini
University of Florence, Department of Chemistry
Via della Lastruccia 3–13, 50019 Sesto Fiorentino (FI) (Italy)
this purpose. Energy storage by the meta-stable VHF 1b corre-
sponds, however, only to 27.7 kJmolꢀ1 or 0.11 MJkgꢀ1 [5]
.
This
Supporting information for this article can be found under http://
dx.doi.org/10.1002/chem.201601190.
value is far from the optimum value of 1 MJkgꢀ1 [3c]
,
which cor-
Chem. Eur. J. 2016, 22, 1 – 10
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
responds to the energy difference of the unsubstituted norbor-
nadiene–quadricyclane (NB/QC) system.[6] Unsubstituted nor-
bornadiene is, however, characterized by a quantum yield of
only 9% for the photoisomerization.[6]
The relative DHA/VHF stability can be tuned in several ways.
For example, it is influenced by solvents because the more
polar VHF is stabilized more in polar solvents, thereby reducing
the energy difference.[5] Substitution can also change the rela-
tive stability; placing a phenyl group at C-3 or exchanging one
of the cyano groups on C-1 with hydrogen leads to an approx-
imate doubling of the DHA/VHF energy difference in vacuo.[5]
We became interested in exploring other ways of tuning the
energy difference and here show that combining the DHA/VHF
switching event with loss or gain in aromaticity can lead to
dramatic changes in energy differences and implications for
the switching properties. Previously, the properties of the met-
acyclophanediene–dihydropyrene photoswitch were strongly
tuned by benzannulation.[7] Furthermore, Yang et al.[8] found
that including the ethene bridge in dithienylethene photo-
switches as part of an aromatic system significantly changes
their stabilities relative to the non-benzenoid dihydrodithieno-
benzene isomers. We designed two derivatives of the DHA/
VHF system in which a benzene ring fused to the DHA core
would maintain its aromaticity when connected between posi-
tion C-3 on the DHA and the ortho-position of the phenyl (2a;
Scheme 1), but lose it when connected at the 6,7-position (3a).
Ring-opening of DHA 2a to VHF 2b destroys the aromaticity
of the fused ring, thereby presumably increasing the DHA/VHF
energy difference. The opposite situation is the case for the
3a/3b pair, and, in consequence, the DHA/VHF energy differ-
ence should be reduced. Here, we present a synthesis of these
two compounds, the further functionalization of 2a, and a de-
tailed computational study of the energetics associated with
the conversions in comparison to the couple 4a/4b that was
recently prepared[9] and to the NB/QC couple.
Scheme 2. Synthesis of DHA 2a. HMDS=hexamethyldisilazane.
NOBF4 =nitrosyl tetrafluoroborate. TBAF=tetrabutylammonium fluoride.
TsOH=p-toluene sulfonic acid. TBDMS=tert-butyldimethylsilyl.
benzene, this compound was converted into the desired unsa-
turated DHA 2a in 67% yield. Alternatively, the desilylation–
elimination of 8a could conveniently be achieved in one-pot
by treatment with TsOH in refluxing benzene to furnish prod-
uct 2a in 71% yield.
The other target molecule, VHF 3b, was obtained in three
steps from the known[11] 7H-benzo[7]annulene 10 (Scheme 3).
The benzotropylium species 11[12] was formed by hydride ab-
straction using tritylium tetrafluoroborate and was then treat-
ed with 12[13] in the presence of Et3N to furnish 13 and its re-
gioisomer 14. The two isomers could be separated by chroma-
tography and characterized for analytical purposes (see the
Supporting Information). However, simply treating the mixture
with tritylium tetrafluoroborate followed by Et3N allowed us to
isolate target molecule 3b in an overall yield of ca. 4% after
chromatographic purification. In another synthesis, we ob-
tained the product 3b in a yield of 19% from benzotropylium
tetrafluoroborate 11. The VHF corresponding to oxidation of
14 was not isolated during work-up; it thus seems that 14 ex-
Results and Discussion
Synthesis
In an initial attempt to prepare DHA 2a, we subjected the cor-
responding dihydronaphtho derivative 4a to oxidation or radi-
cal bromination–elimination protocols, but none of these
methods were successful. Instead, the synthesis of 2a was ach-
ieved in six steps (Scheme 2) from the carefully chosen, known
a-tetralone 5 starting material (after an optimization of its syn-
thesis allowing preparation at 13 g scale without use of chro-
matography).
A Knoevenagel condensation with malononitrile using an
AcOH/HMDS buffer gave 6 in 97% yield. Treatment of 6 with
tropylium tetrafluoroborate in the presence of Et3N gave 7 in
98% yield. A two-step oxidation using NOBF4 followed by pyri-
dine gave the VHF, which immediately underwent ring-closure
to form 8a in 56% yield. A desilylation was accomplished by
treatment with tetrabutylammonium fluoride (TBAF) in the
presence of AcOH (to avoid azulene formation[10]), affording al-
cohol 9a in 92% yield. By treatment with TsOH in refluxing
Scheme 3. Synthesis of aromatic VHF 3b. DCE=1,2-dichloroethane.
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
hibits some reluctance towards the oxidation. The structure of
3b was confirmed by X-ray crystallographic analysis (see
below). We then attempted to convert VHF 3b into DHA 3a.
This conversion was, however, not successful, even by heating
3b at 608C for 30 h in 1,2-dichloroethane (DCE). For compari-
son, the half-life of VHF 1b was merely 9.4 min under the
same conditions (see the Supporting Information).
Optical and switching properties
The optical and switching properties of DHA 2a were studied
by using UV/Vis absorption spectroscopy. In both polar and
nonpolar solvents, DHA 2a exhibited a characteristic absorp-
tion at lmax 342 nm in EtOH and cyclohexane. Unfortunately,
neither in cyclohexane at 258C nor in (deoxygenated) EtOH at
ꢀ1008C did 2a seem to undergo the desired photoisomeriza-
tion upon irradiation at 365 nm. Instead of the appearance of
the characteristic VHF absorption at 440–490 nm, the charac-
teristic DHA absorption simply disappeared and new higher
energy absorptions appeared. Thus, for 2a, other light-induced
reactions come into play, or, if in fact formed, an ultrafast con-
version of VHF into DHA may occur. We consider that the VHF
is simply too short-lived to be observed under the experimen-
tal conditions, even at low temperature (ꢀ60 to ꢀ1008C) at
which the quantum yield of ring-opening at the same time
should be decreased.[4] To identify the product(s) formed,
a sample of DHA 2a in CDCl3 was exposed to light at 365 nm
Characterization by NMR spectroscopic and X-ray crystallo-
graphic analyses
The aromatic characteristics of the benzenoid rings of 2a and
1
3b were probed experimentally by H NMR and 13C NMR spec-
troscopies as well as by X-ray crystal structure analyses. The
1H NMR chemical shifts of the two protons in the central ben-
zene ring in 2a were found in the aromatic region at d=7.94
and 7.55 ppm and 13C(H) resonances at d=133.22 and
119.19 ppm. Likewise, the protons of the fused benzene ring in
3b were found in the aromatic region (>7.25 ppm, overlap-
ping with the Ph resonances) and the 13C resonances at d>
120 ppm. The crystal structures (Figure 1) obtained from
single-crystal structure determination offered additional sup-
port for the benzenoid characteristics; all carbon atoms in the
naphthalene and benzene moieties of 2a and 3b were copla-
nar. The bond lengths of the naphthalene moiety of 2a were
similar to those found in naphthalene (see the Supporting In-
formation). For 3b, the bond lengths of the benzene ring were
in the range 1.36–1.42 ꢁ, hence close to the value of 1.39 ꢁ in
benzene.
1
1
for 60 h. By H/1H COSY and H/1H NOESY NMR spectroscopy,
six different DHA structures were observed (see the Supporting
Information), all products of sigmatropic rearrangements. Fur-
ther irradiation resulted in formation of the corresponding
1,3a-DHA 15 (Scheme 4) along with one other isomer as major
products.
Scheme 4. Light and Lewis acid (such as AlCl3 or BBr3) induced conversions
of DHA 2a.
Although formation of VHF could not be tracked by irradia-
tion with light, treatment with Lewis acids (LA) successfully
converted 2a into the corresponding VHF·LA complex (2b·LA)
(Scheme 4), the formation of which was monitored both by
UV/Vis and NMR spectroscopies. Thus, addition of BBr3 or AlCl3
to a solution of 2a in CH2Cl2 resulted in a redshift of the lon-
gest wavelength absorption maximum to 504 nm for VHF·BBr3
and to 503 nm for VHF·AlCl3 (Figure 2). Most importantly, the
1H and 13C NMR spectra of the VHF·AlCl3 complex (see the Sup-
porting Information) were very similar to those previously mea-
sured for the BBr3 and AlCl3 complex of 1b.[14] Upon addition
of water, the VHF·LA complex was broken up and then the
VHF instantaneously returned to the DHA. We note that, while
showing the characteristic DHA absorption peak, the initial ab-
Figure 1. Molecular structures of 2a (top; CCDC 1429767) and 3b (bottom;
CCDC 1421412) with displacement-ellipsoids of 50% for non-H atoms.
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
By using a bromination–elimination–cross-coupling proce-
dure,[10,15] we set out to prepare a DHA with a tolyl-substituent
at C-7 of the DHA. Firstly, treatment of DHA 2a with Br2 in
CH2Cl2 at ꢀ788C followed by LiHMDS in tetrahydrofuran (THF)
at 08C almost quantitatively yielded the bromo-substituted
DHA 16a (Scheme 5). Treatment with tolyl boronic acid under
our previously optimized Suzuki cross-coupling conditions
Figure 2. Top: UV/Vis absorption spectra in CH2Cl2 of DHA 2a (red), VHF·BBr3
complex (blue), VHF·AlCl3 complex (purple), DHA generated from addition of
water to BBr3 complex (black), and DHA generated from addition of water to
AlCl3 complex followed by a liquid–liquid extraction (green). Bottom: UV/Vis
absorption spectrum of VHF 3b in DCE.
sorption spectrum obtained upon addition of water is not
identical to that of DHA 2a. However, after an aqueous work-
up, the spectrum becomes identical to that of 2a. The conver-
sion between 2a and 2b is thus indeed reversible.
Scheme 5. Installation of “spectator group” to investigate whether light-in-
duced ring-opening occurs.
VHF 3b was also studied by using UV/Vis absorption spec-
troscopy, showing a strong absorption at lmax 468 nm in DCE
(Figure 2). No sign of a thermally induced conversion into 3a
was observed, even by addition of Lewis acids, such as ZnCl2,
which has previously proven to catalyze the ring-closure reac-
tion.[14] Clearly, the VHF seems to be too stable to undergo the
ring-closure reaction as also suggested by calculations (see
below). Thus, the aromatic character gained by the VHF re-
versed the relative DHA–VHF stability.
(K3PO4, RuPhos, and Pd(OAc)2)[15] gave the desired DHA 17a in
59% yield (over three steps). With this tolyl-substituted DHA in
hand, it was exposed to white light for 10 h in CD3CN solution
to investigate whether an isomerization into isomer 18 oc-
1
curred. From H/1H COSY NMR spectroscopic analysis, it tran-
spires that whereas DHA 17a still undergoes the undesired sig-
matropic rearrangements, formation of an intermediate VHF
cannot be unambiguously established because an extremely
complex NMR spectrum resulted from irradiation. In principle,
a total of 12 DHA compounds can arise from a combination of
ring-opening and ring-closure in combination with sigmatropic
rearrangements. From the spectrum we were only able to un-
ambiguously confirm the presence of the undesired com-
pound 19 resulting from sigmatropic rearrangements (as well
as several other isomers). Although this observation is not en-
couraging, it does not rule out the formation of VHF. Thus, we
have previously established that the quantum yield of the
ring-opening does not necessarily affect the degree of isomeri-
zation from the 7- to the 6-substituted isomer. Instead, we be-
lieve that the polarization of the VHF transition state does.
Thus, a strongly electron-withdrawing substituent would place
a larger partial positive charge on C-4 compared with that on
C-8a (DHA core numbering) and could present a possible way
Further functionalization
Our next objective was to investigate the possibility of using
2a as a substrate for a subsequent functionalization of the
core, which is important for tuning of properties. In addition,
we hoped that attachment of a substituent group in the
seven-membered ring could shed light on whether a ring-
opening does occur upon irradiation or not. Given that we
have shown that pure 7-substituted DHAs partly isomerize into
their 6-substituted derivatives—via the VHF—upon irradiation
with light in polar solvents,[1e,10] we considered that it may be
possible to probe whether the DHA does indeed rapidly form
VHF by installation of a substituent at C-7 of the DHA core of
2a: formation of any 6-substituted DHA would be evidence of
a ring-opening/closure cycle.
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
to observe this isomerization. Thus, we set out to place a CN-
group on C-7 by using a three-step palladium-insertion–trans-
metalation–cyanation procedure. Gratifyingly, treatment of 16a
with Pd2dba3/tBu3P in CH3CN followed by addition of Bu3SnCN
in toluene gave the cyano-substituted DHA 20a in 32% yield
(over three steps). This compound was then exposed to light
at 366 nm in CD3CN and CD2Cl2 solutions; however, again we
could not unambiguously determine or rule out formation of
the desired isomerized compound 21. Instead, we again detect
products from undesired sigmatropic rearrangements and for-
mation of compound 22. In all, although the new derivatives
did not offer evidence for DHA to VHF photoactivity, we were
pleased to see that the new polycyclic system could be regio-
selectively functionalized, and, as calculations show below, in-
corporation of a cyano group strongly increases the energy
barrier for the VHF to DHA back-reaction, which is particularly
relevant in a broader context of solar thermal energy storage
systems, albeit the exact structure 20a is not suited for this
purpose.
Figure 3. Optimized structures (M06-2X/6-311+G(d)) for the couples 2a/2b
and 4a/4b and the transition state (TS) for each of the ring-closure reac-
tions.
Calculations
To understand the thermodynamics and kinetics of the systems
involving compounds 2–4, they were subjected to a computa-
tional study. Structures were optimized (geometries shown in
Figure 3) and energies calculated by using the M06-2X/6-311+
G(d) method in vacuum and in three different solvents, cyclo-
hexane (CH), dichloromethane (CH2Cl2), and acetonitrile
(MeCN; for details, see the Supporting Information).[17] Gibbs
free energies and enthalpies of the VHFs relative to those of
the DHAs are listed in Table 1. The calculations show a remarka-
bly high energy storage capacity of 2a/2b of 0.38 MJkgꢀ1
(Gibbs free energy) in vacuum (but decreasing slightly with sol-
vent polarity), which is almost four times that of the parent
system 1a/1b (0.11 MJkgꢀ1).[5] Calculations were also per-
formed for the cyano-substituted couple 20a/20b, which gave
Table 1. Calculated Gibbs free energy, enthalpy differences for DHA/VHF couples (M06-2X/6-311+G(d); 298.15 K and 1 atm) and Gibbs free activation en-
ergies for VHF to DHA ring-closure reactions (M06-2X/6-311+G(d); 298.15 K and 1 atm).
DGDHA!VHF (top) and DHDHA!VHF (bottom) [kJmolꢀ1/MJkgꢀ1
]
Couple[a]
Vacuum
CH
CH2Cl2
MeCN
2a/2b
107/0.38
114/0.40
103/0.37
110/0.39
94.3/0.34
100.8/0.36
ꢀ64.8/ꢀ0.21
ꢀ55.1/ꢀ0.18
51.7/0.18
58.2/0.21
102/0.34
110/0.36
131/0.51
137/0.54
117/0.43
90.3/0.32
96.9/0.34
ꢀ66.1/ꢀ0.22
ꢀ56.3/ffi0.18
51.0/0.18
56.8/0.20
101/0.33
108/0.35
131/0.51
137/0.54
116/0.43
3a/3b[a]
4a/4b
ꢀ55.1/ꢀ0.18
ꢀ47.2/ꢀ0.15
58.5/0.21
64.7/0.23
110/0.36
ꢀ59.2/ꢀ0.19
ꢀ50.9/ꢀ0.17
56.0/0.20
62.1/0.22
106/0.35
114/0.37
131/0.51
137/0.54
117/0.43
20a/20b
23a/23b
24a/24b
NB/QC
116/0.38
131/0.51
137/0.54
117/0.43
125/0.46
65.5/0.71
64.2/0.70
124/0.46
66.0/0.72
64.7/0.70
124/0.46
66.7/0.72
65.4/0.71
123/0.46
66.9/0.73
65.7/0.71
DGVHF!TS [kJmolꢀ1
]
Couple
Vacuum
33.5
153
CH
CH2Cl2
17.4
141
74.8
33.6
68.3
74.1
MeCN
17.1
140
70.3
30.5
65.8
69.8
2a/2b
24.0
147
95.3
42.4
76.1
82.4
3a/3b[b]
4a/4b
101
20a/20b
23a/23b
24a/24b
48.1
81.5
88.0
[a] VHF 3b in its s-trans conformation. [b] VHF 3b in its s-cis conformation.
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
a similar value of 0.36 MJ kgꢀ1 in vacuum. The high storage ca-
pacity is, however, counterbalanced by a low activation barrier
of the VHF to DHA back-reaction. The ring-closure barrier
heights are also listed in Table 1. Thus, the Gibbs free activa-
tion energy is only 33.5 kJmolꢀ1 in vacuum for the conversion
of 2b into 2a, whereas it is 119.5 kJmolꢀ1 for the conversion
of 1b.[5] The low back-reaction barrier of 2b agrees with the ul-
trafast conversion into 2a observed experimentally when gen-
erating 2b by a Lewis acid followed by treatment with water.
Yet, the barrier is increased significantly for 20b (48.1 kJmolꢀ1),
albeit still much lower than for the unlocked VHF 1b. Thus, in-
corporation of the electron-withdrawing cyano group directly
onto the seven-membered ring has a beneficial influence in
regard to slowing down the back-reaction. Proceeding to 3a/
3b, we see that the situation is reversed. Now, the VHF is in
fact more stable than the DHA and the activation energy for
the back reaction is significantly higher than that of 1b. Calcu-
lations on the couple 4a/4b reveal that the locked structure
itself contributes to increasing the energy storage capacity,
which is almost doubled relative to that of the parent system
1a/1b.
Figure 4. DHA/VHF couples with one cyano group substituted for a hydro-
gen and methyl, respectively.
numbers relate to the diastereoisomers shown in Figure 4;
only small changes were obtained between the various iso-
mers (see the Supporting Information, page S97). Indeed, sub-
stituting a cyano group for a hydrogen atom increases the
energy storage capacity further (to +0.51 MJ kgꢀ1) and increas-
es the activation energy for the VHF into DHA conversion from
+33.5 to +81.5 kJmolꢀ1 (in vacuum). The heat release has
now reached 77% of that of the NB/QC couple. By using in-
stead a methyl substituent, the increase in the energy storage
capacity is not as significant (+0.43 MJ kgꢀ1), but the activa-
tion energy for the back-reaction is conveniently increased fur-
ther (+88.0 kJmolꢀ1). It should still occur faster than the back-
reaction of 4b (+100.8 kJmolꢀ1), for which we found previous-
ly an experimental half-life of only 13 ms in MeCN at room
temperature.[9] Thus, further modifications are clearly required
to slow down the ring-closure reaction further, but the calcula-
tions convincingly show how we can indeed fine-tune the
properties by minor structural changes—by either removing
a cyano group at C-1 or adding one at position C-7 of the DHA
core. The modifications may be combined with suitable donor/
acceptor functionalization in both the five- and seven-mem-
bered rings, which we have previously shown[16] is an impor-
tant tool for tuning the rate of the VHF to DHA ring-closure re-
action.
The loss/gain in aromaticity accompanying switching of the
two benzenoid DHA/VHF systems was further corroborated by
Nucleus Independent Chemical Shift (NICS) calculations.[18]
NICS(0) values were calculated at the center of the added rings
(colored rings in Scheme 1) using CAM-B3LYP/6-311+G(d).
DHA 2a has a negative NICS(0) value of ꢀ7.4 ppm, whereas
the value turns positive (+1.1 ppm) in VHF 2b. Oppositely,
a more negative NICS value of VHF 3b (ꢀ7.5 ppm) indicates
gain of aromaticity relative to DHA 3a (ꢀ2.5 ppm).
To benchmark our energy storage capacities against that of
the norbornadiene/quadricyclane couple (NB/QC; Scheme 6),
this isomer pair was also subjected to a computational study
using the same level as for the DHA/VHF couples. The energies
Conclusions
Scheme 6. Isomerization between norbornadiene (NB) and quadricyclane
Efficient synthetic protocols for preparing two DHA-VHF sys-
tems were developed based on either loss or gain of aromatici-
ty upon photoisomerization. Future work will seek to modify
further the aromatic DHA structure to avoid light-induced sig-
matropic rearrangements and to allow storage of the energy
in the metastable VHF form for a sufficiently long time. Indeed,
our computational study clearly predicts that benzannulation
in combination with locking the VHF structure is a strong tool
for increasing the energy storage capacity of the DHA/VHF
system, providing a value of ca. 0.4 MJ kgꢀ1. Our calculations
also show that by removing one cyano group at C-1, we will
move further in the right direction, both increasing the energy
storage capacity and slowing down the back reaction. Howev-
er, the back reaction still needs to be retarded by other struc-
tural modifications because the calculated energy of activation
remains too small. It is in this regard noteworthy that the new
(QC).
obtained in vacuum, CH, CH2Cl2, and MeCN are listed in
Table 1. In vacuum, the Gibbs free storage energy is 0.71 MJ
kgꢀ1, and the heat release for the back reaction is 0.70 MJ
kgꢀ1. Accordingly, the couple 2a/2b has a heat release that is
about 57% of that of NB/QC (in vacuum). The latter couple
likely represents the upper energy storage limit [MJ kgꢀ1] for
photochromic molecules, and the obtained value for 2a/2b is
high considering the large molecular weight of this system.
We have previously seen that removing one cyano group in
the parent system 1a/1b can lead to both increased energy
storage capacity and slower back-reaction.[5b] We therefore
subjected compounds 23a/23b and 24a/24b (Figure 4) to
a computational study; the results are included in Table 1. The
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
polycyclic DHA system can be regioselectively functionalized in
the seven-membered ring, offering an important tool for
tuning the molecular properties. In addition, although not solv-
ing the unwanted light-induced rearrangements, our trapping
of the VHF species by complexation with a Lewis acid presents
an alternative solution to preventing the fast back reaction. Ul-
timately, we hope that our iterative approach, combining syn-
thesis and quantum-chemical calculations, will allow us to
reach closed-cycle molecular energy storage systems that satis-
fy not only large storage energies but also controllable energy
releases. The challenge is to make all the small structural varia-
tions work in concert.
X-ray crystal structure analysis
CCDC 1429767 (2a) and 1421412 (3b), contain the supplementary
crystallographic data for this paper. These data are provided free
of charge by The Cambridge Crystallographic Data Centre.
Naphtho[2,1-a]azulene-12,12(11aH)-dicarbonitrile (2a)
Method 1: To a solution of 8a (98.9 mg, 240 mmol) in benzene
(25 mL), TsOH·H2O (4.7 mg, 27 mmol, 0.11 equiv) and water (ca.
0.01 mL, 0.05 mmol, 2 equiv) were added, and the mixture was
heated to reflux for 2 h (Note: as the desilylation commences, the
yellow color strengthens, but after the elimination of water, a decol-
oration is evident). The resulting slightly yellow solution was dilut-
ed with Et2O (150 mL) and poured into saturated aqueous NaHCO3
(50 mL). The organic phase was separated, dried with MgSO4, and
concentrated in vacuum. Purification by flash column chromatogra-
phy (10% EtOAc/heptanes) gave 2a (47.7 mg, 170 mmol, 71%) as
a slightly yellow solid.
Experimental Section
Method 2: To a solution of 9a (95.4 mg, 0.320 mmol) in benzene
(25 mL), TsOH·H2O (6.1 mg, 32 mmol) was added and the reaction
mixture was heated to reflux for 2 h. The resulting slightly yellow
solution was diluted with Et2O (150 mL) and poured into saturated
aqueous NaHCO3 (50 mL). The organic phase was separated, dried
with MgSO4, and concentrated in vacuo. Purification by flash
column chromatography (10% EtOAc/heptanes) gave 2a (60.1 mg,
0.214 mmol, 67%) as a slightly yellow solid. TLC (10% EtOAc/hep-
tanes): Rf =0.25 (becomes brown upon treatment with vanillin and
General procedures
All reactions were performed under an inert atmosphere by using
either nitrogen with a gas-bubbler or argon with a balloon. All
chemicals and solvents were used as received unless otherwise
stated. p-Toluenesulfonic acid was dried by azeotropic distillation
of water using benzene. Benzene and THF were distilled from
a Na/benzophenone couple. Acetonitrile and CH2Cl2 were purified
and dried using activated Al2O3 (drying-tower). CDCl3 was purified
by passing through activated Al2O3. Thin-layer chromatography
(TLC) was carried out on commercially available pre-coated plates
(silica 60) with fluorescence indicator. Chromatographic purifica-
tions were performed on silica (SiO2) with a pore size of 60 ꢁ and
a particle size of 15–40 mm using the dry column vacuum chroma-
tography method, as described previously[15,19] or 40–63 mm using
flash column chromatography. Mass spectra were recorded with an
ESP-MALDI-FT-ICR spectrometer equipped with a 7 T magnet (prior
to the experiments, the instrument was calibrated with NaTFA clus-
ter ions) or a MicrOTOF-QII spectrometer using ESP (calibrated with
formic acid). All 1H NMR and 13C APT (attached proton test) NMR
spectra were recorded with a 500 MHz instrument equipped with
a (non-inverse) cryoprobe (500.1300/125.7578 MHz), with a Varian
300-MHz (300.0787 Hz) instrument with a penta-probe, or Bruker
500-MHz (499.9731/125.7183 MHz) instrument with a broad band-
1
light heating). H NMR (500 MHz, CDCl3): d=8.30 (ddd, J=8.4, 1.2,
0.9 Hz, 1H), 7.94 (d, J=8.6 Hz, 1H), 7.93 (ddd, J=8.2, 1.2, 0.9 Hz,
1H), 7.76 (ddd, J=8.4, 8.1, 1.2 Hz, 1H), 7.62 (ddd, J=8.1, 8.2,
1.2 Hz, 1H), 7.55 (d, J=8.6 Hz, 1H), 6.79–6.71 (m, 2H), 6.67–6.59
(m, 1H), 6.37 (ddd, J=10.0, 6.0, 1.9 Hz, 1H), 5.89 (dd, J=10.0,
4.3 Hz, 1H), 3.82 ppm (dt, J=4.3, 1.9 Hz, 1H) (not all spin-systems
paired); 13C NMR (126 MHz, CDCl3): d=137.59, 135.03, 134.66,
133.22, 131.64, 131.55, 131.28, 129.60, 129.10, 128.57, 127.64,
127.61, 122.83, 119.45, 119.19, 117.55, 115.20, 113.24, 51.66,
42.10 ppm; HRMS (ESP+ FT-ICR, added formic acid): m/z calcd for
[C20H12N2Na]+ 303.08927 [MNa]+; found 303.08990; C20H12N2
(280.32): calcd C 85.69, H 4.31, N 9.99; found C 85.73, H 4.02, N
9.74.
2-(4-((tert-Butyldimethylsilyl)oxy)-3,4-dihydronaphthalen-
1(2H)-ylidene)malononitrile (6)
1
probe. All H and 13C chemical shifts are referenced to the residual
solvent peak (CDCl3: dH =7.26 ppm, dC =77.16 ppm; [D6]DMSO:
dH =2.50 ppm, dC =39.52 ppm; CD3CN dH =1.94 ppm, dC =
1.32 ppm). Elemental analyses were performed at the Department
of Chemistry, University of Copenhagen or at London Metropolitan
University. All spectroscopic measurements (including photolysis)
were performed in a 1-cm path length quartz cuvette. UV/Vis ab-
sorption spectra were obtained by scanning the wavelength from
800 to 200 nm. Photoswitching experiments were performed with
a 150 W xenon arc lamp equipped with a monochromator; the
DHA absorption maximum (lowest energy absorption) for each in-
dividual species was chosen as the wavelength of irradiation (line
width Æ2.5 nm). The thermal back-reaction was performed by
heating the sample (cuvette) with a Peltier unit in the UV/Vis spec-
trophotometer (temperature kept at 608C Æ0.18C) or, at low tem-
perature, in a cryostat (ꢀ1008C to ꢀ608C Æ0.18C). Photoswitching
experiments of 1a/1b and 3b were performed in non-deoxygenat-
ed solvents as it has been shown that this does not significantly
improve stability; experiment with 2a/2b were performed in de-
oxygenated solvents to rule out oxidative degradation.
To a mixture of 5 (6.30 g, 22.8 mmol) and malononitrile (3.91 g,
59.3 mmol, 2.6 equiv), AcOH (10.1 mL, 251 mmol, 11 equiv) and
HMDS (5.3 mL, 25 mmol, 1.1 equiv) were added and the resulting
dark-red to black reaction mixture was stirred and heated at reflux
for 10 h. The reaction mixture was diluted with CH2Cl2 (400 mL),
washed with water (3ꢂ100 mL), dried with MgSO4, filtered, and
concentrated in vacuum. The crude residue was passed through
a plug of silica (50% CH2Cl2/ petroleum spirit), which gave 6
(7.18 g, 22.1 mmol, 97%) as a light-yellow oil. 1H NMR (500 MHz,
CDCl3): d=8.16 (ddd, J=7.9, 1.2, 0.6 Hz, 1H), 7.59 (ddd, J=7.5, 7.6,
1.2 Hz, 1H), 7.51 (m, 1H), 7.44 (ddd, J=7.9, 7.5, 1.3 Hz, 1H), 4.78
(dd, J=7.4, 3.6 Hz, 1H), 3.26 (dt, J=18.4, 6.3 Hz, 1H), 2.95 (ddd, J=
18.4, 7.3, 6.1 Hz, 1H), 2.18–1.86 (m, 2H), 0.91 (s, 9H), 0.15 (s, 3H),
0.06 ppm (s, 3H) (not all spin systems could be paired); 13C NMR
(126 MHz, CDCl3): d=171.87, 144.44, 133.95, 129.15, 128.34, 127.97,
127.04, 113.81, 113.25, 80.69, 68.66, 30.35, 28.88, 25.85, 18.24,
ꢀ4.39, ꢀ4.57 ppm; HRMS (MALDI+ FT-ICR, dithranol): m/z calcd for
[C19H24N2OSiNa]+ 347.15501 [M+Na+]; found 347.15526.
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
or low-melting solid. TLC (10% EtOAc/heptanes): Rf =0.38 (two
closely running spots, becomes brown upon treatment with vanil-
lin and light heating). TLC (20% EtOAc/heptanes): Rf =0.53
(yellow); m.p. 111–1168C; 1H NMR (500 MHz, CDCl3): d=7.72–7.66
(m, 1H [major/minor]), 7.53–7.48 (m, 1H [major/minor]), 7.45–7.37
(m, 2H [major/minor]), 6.64 (dd, J=11.3, 6.3 Hz, 1H [minor]), 6.60
(dd, J=11.3, 6.3 Hz, 1H [major]), 6.52 (dd, J=11.3, 6.2 Hz, 1H
[minor]), 6.49 (dd, J=11.3, 6.2 Hz, 1H [major]), 6.34–6.29 (m, 1H
[major/minor]), 6.28 (br d, J=6.2 Hz, 1H), 5.86 (dd, J=10.2, 3.9 Hz,
1H [major]), 5.84 (dd, J=10.2, 3.9 Hz, 1H [minor]), 5.03–4.97 (m,
1H), 3.82 (dt, J=3.9, 2.0 Hz, 1H [major]), 3.76 (dt, J=3.9, 2.0 Hz, 1H
[minor]), 2.76–2.59 (m, 2H [major/minor]), 0.94 (s, 9H [minor]), 0.91
(s, 9H [major]), 0.14 (s, 3H [minor]), 0.13 (s, 3H [major]), 0.08 (s, 3H
[minor]), 0.04 ppm (s, 3H [major]); 13C NMR (126 MHz, CDCl3): d=
141.71, 141.22, 139.45, 139.39, 138.87, 138.46, 133.68, 133.17,
131.19, 131.13, 131.09, 130.64, 129.76, 129.70, 127.88, 127.55,
127.55, 127.41, 126.54, 126.31, 123.37, 123.36, 119.92, 119.89,
118.50, 117.96, 115.07, 115.06, 112.86, 112.72, 68.66, 68.63, 51.41,
51.32, 42.84, 42.47, 31.66, 31.43, 25.95, 25.90, 18.25, 18.22, ꢀ4.32,
ꢀ4.35, ꢀ4.61, ꢀ4.65 ppm; HRMS (ESP+ FT-ICR, added formic acid):
m/z calcd for [C26H28N2OSiNa]+ 435.18631 [MNa]+; found
435.18712.
2-(4-((tert-Butyldimethylsilyl)oxy)-2-(cyclohepta-2,4,6-trien-1-
yl)-3,4-dihydronaphthalen-1(2H)-ylidene)malononitrile (7)
To a stirred suspension of 6 (1.14 g, 3.53 mmol) and mortared tro-
pylium tetrafluoroborate (659 mg, 3.70 mmol, 1.05 equiv) in CH2Cl2
(250 mL) at ꢀ788C, Et3N (0.52 mL, 3.7 mmol, 1.05 equiv) was added
dropwise and the reaction mixture was stirred for 16 h while the
temperature was allowed to slowly reach rt. The reaction mixture
was poured into saturated aqueous NH4Cl (100 mL) and the phases
were separated. The organic phase was washed with saturated
aqueous NH4Cl (4ꢂ100 mL) and the combined aqueous phases
were extracted with CH2Cl2 (100 mL). The organic extract was
washed with saturated aqueous NH4Cl (100 mL) and combined
with the other organic phase. The combined organic phases were
dried with MgSO4, filtered and concentrated in vacuo, which gave
a ca. 1:3 diastereomeric mixture of 7 (1.44 g, 3.47 mmol, 98%) as
a light-yellow foam. A sample was purified for characterization by
flash column chromatography (15% EtOAc/heptanes), which gave
7 as a colorless oil. TLC (20% EtOAc/heptanes): Rf =0.51 and 0.56
(UV254 nm); 1H NMR (500 MHz, CDCl3): d=7.88 (br d, J=8.3 Hz, 1H
[minor]), 7.73 (br d, J=7.8 Hz, 1H [major]), 7.63–7.52 (m, 2H
[major/minor]), 7.45–7.39 (m, 1H [major/minor]), 6.70–6.65 (m, 1H
[major/minor]), 6.64–6.57 (m, 1H [major/minor]), 6.29–6.22 (m, 2H
[major/minor]), 5.24 (dd, J=9.5, 6.6 Hz, 1H [minor]), 5.20 (dd, J=
9.5, 6.7 Hz, 1H [major]), 5.14–5.08 (m, 1H [major/minor]), 4.89 (dd,
J=6.7, 4.4 Hz, 1H [minor]), 4.66–4.61 (m, 1H [major]), 3.70–3.64 (m,
1H [minor]), 3.47 (ddd, J=9.3, 7.3, 5.4 Hz, 2H), 2.38–2.27 (m, 1H
[major/minor]), 2.18 (dt, J=13.6, 4.4 Hz, 1H [minor]), 2.11 (dt, J=
9.3, 6.7 Hz, 1H [major]), 1.87 (dt, J=9.9, 6.6 Hz, 1H [minor]), 1.79
(ddd, J=13.6, 8.3, 5.4 Hz, 1H [major]), 0.98 (s, 9H [major]), 0.88 (s,
9H [minor]), 0.21 (s, 3H [major]), 0.19 (s, 3H [major]), 0.13 (s, 3H
[minor]), ꢀ0.02 (s, 3H [minor]) ppm (not all spin-systems paired);
13C NMR (126 MHz, CDCl3): d=176.73 [minor], 175.66 [major],
143.05 [major], 142.91 [minor], 133.38 [major], 133.25 [minor],
131.62 [minor], 131.37 [major], 130.78, 130.07, 129.12, 128.63,
127.89 [minor], 127.84 [major], 127.53, 127.02, 126.99, 126.91,
126.86, 126.85, 125.73, 122.11 [minor], 122.11 [major], 122.00,
121.12, 113.47 [minor], 113.20 [major], 113.18 [minor], 112.81
[major], 83.53 [major], 82.12 [minor], 67.43 [minor], 66.39 [major],
44.28 [major], 42.60 [minor], 41.87 [minor], 39.80 [major], 35.56
[minor], 32.77 [major], 26.02 [major], 25.76 [minor], 18.33 [major],
18.06 [minor], ꢀ4.13 [major], ꢀ4.36 [minor], ꢀ4.59 [minor],
ꢀ4.77 ppm [major] (one signal missing due to overlap, five minor
aliphatic peaks dismissed as grease); HRMS (MALDI+ FT-ICR, dithra-
nol): m/z calcd for [C26H31N2OSi]+ 415.22002 [M+H]+; found
415.22023.
5-Hydroxy-5,6-dihydronaphtho[2,1-a]azulene-12,12(11aH)-
dicarbonitrile (9a)
To a solution of 9a (340 mg, 824 mmol) in THF (50 mL), AcOH
(0.47 mL, 8.24 mmol, 10 equiv) and
a 1m solution of TBAF
(1.65 mL, 1.65 mmol, 2 equiv) in THF were added successively and
the reaction mixture was heated to reflux for 16 h. The reaction
mixture was diluted with Et2O (100 mL), washed with water (3ꢂ
50 mL) and saturated aqueous NaHCO3 (50 mL), dried with Na2SO4,
filtered, and concentrated in vacuum. Purification by flash column
chromatography (CH2Cl2) gave 9a (226 mg, 0.758 mmol, 92%) as
1
a yellow glass. M.p. 164–1658C; H NMR (500 MHz, CDCl3): d=7.75
(t, J=7.7 Hz, 1H [major/minor]), 7.52–7.48 (m, 2H [major/minor]),
7.44–7.39 (m, 1H [major/minor]), 6.64–6.59 (m, 1H [major/minor]),
6.54–6.48 (m, 1H [major/minor]), 6.35–6.28 (m, 2H [major/minor]),
5.88 (dd, J=10.2, 3.8 Hz, 1H [minor]), 5.83 (dd, J=10.1, 3.9 Hz, 1H
[major]), 4.98–4.94 (m, 1H [major/minor]), 3.82 (dt, J=3.8, 1.9 Hz,
1H [minor]), 3.78 (dt, J=3.9, 1.8 Hz, 1H [major]), 2.93–2.68 (m, 2H
[major/minor]), 1.84 ppm (br s, 1H); 13C NMR (126 MHz, CDCl3): d=
141.12 [minor], 140.93 [major], 138.77 [minor], 138.45 [major],
137.78 [minor], 137.55 [major], 132.63 [major], 132.43 [minor],
131.40 [major], 131.25 [minor], 130.94 [major], 130.78 [minor],
129.94 [minor], 129.90 [major], 129.59 [major], 129.53 [minor],
128.33 [major], 128.08 [minor], 127.90 [minor], 127.66 [major],
127.27 [major], 127.26 [minor], 123.79 [minor], 123.76 [major],
120.00 [minor], 119.67 [major], 118.93 [minor], 118.51 [major],
115.10 [major], 114.87 [minor], 112.78 [minor], 112.74 [major],
67.64 [minor], 67.45 [major], 51.26 [major], 51.19 [minor], 42.82
[minor], 42.58 [major], 30.58 [major], 30.43 [minor] ppm; HRMS
(ESP+ FT-ICR, added formic acid): m/z calcd for [C20H14N2ONa]+
321.09983 [MNa]+; found 321.10041.
5-((tert-Butyldimethylsilyl)oxy)-5,6-dihydronaphtho[2,1-
a]azulene-12,12(11aH)-dicarbonitrile (8a)
To a stirred solution of 7 (1.23 g, 2.96 mmol) in anhydrous MeCN
(150 mL) at ꢀ408C, NOBF4 (719 mg, 6.16 mmol, 2.08 equiv) was
added, and the reaction mixture was stirred for 2 h (full conversion
according to TLC analysis). The resulting yellow reaction mixture
was diluted with cold CH2Cl2 (150 mL) and at ꢀ408C, a solution of
pyridine (0.50 mL, 6.2 mmol, 2.1 equiv) in CH2Cl2 (50 mL) was
added. The temperature was allowed to slowly reach ꢀ208C over
2 h and the reaction mixture was poured into water (100 mL) and
extracted with CH2Cl2 (3ꢂ100 mL). The combined extracts were
dried with Na2SO4, filtered, and concentrated in vacuum. Purifica-
tion by dry column vacuum chromatography (0–100 CHCl3/hep-
tanes, 10% increments, 40 mL fractions) followed by flash column
chromatography (10% EtOAc/heptanes) gave a ca. 1:3 diastereo-
meric mixture of 8a (679 mg, 1.65 mmol, 56%) as a dark-yellow oil
Acknowledgements
University of Copenhagen and the Carlsberg Foundation are
acknowledged for supporting this work.
Keywords: electrocyclic reactions
· energy conversion ·
molecular devices · photochromism · renewable resources
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
[1] 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, 63–106; c) M. B. Nielsen, S. L. Broman, M. ꢁ. Petersen,
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, ARKI-
VOC 2014, 249–263; e) S. L. Broman, M. B. Nielsen, Phys. Chem. Chem.
Phys. 2014, 16, 21172–21182.
[2] a) S. Lara-Avila, A. V. Danilov, S. E. Kubatkin, S. L. Broman, C. R. Parker,
M. B. Nielsen, J. Phys. Chem. C 2011, 115, 18372–18377; b) S. L. Broman,
S. Lara-Avila, C. L. Thisted, A. D. Bond, S. Kubatkin, A. Danilov, M. B. Niel-
sen, Adv. Funct. Mater. 2012, 22, 4249–4258; c) S. T. Olsen, T. Hansen,
K. V. Mikkelsen, J. Phys. Chem. C 2015, 119, 14829–14833.
Chem. A 2015, 119, 896–904; b) M. Cacciarini, A. B. Skov, M. Jevric, A. S.
Hansen, J. Elm, H. G. Kjaergaard, K. V. Mikkelsen, M. B. Nielsen, Chem.
Eur. J. 2015, 21, 7454–7461.
[6] Z.-I. Yoshida, J. Photochem. 1985, 29, 27–40.
[7] R. H. Mitchell, Eur. J. Org. Chem. 1999, 2695–2703.
[8] Y. Yang, Y. Xie, Q. Zhang, K. Nakatani, H. Tian, W. Zhu, Chem. Eur. J.
2012, 18, 11685–11694.
[9] S. L. Broman, O. Kushnir, M. Rosenberg, A. Kadziola, J. Daub, M. B. Niel-
sen, Eur. J. Org. Chem. 2015, 4119–4130.
[10] S. L. Broman, M. ꢁ. Petersen, C. G. Tortzen, K. Kilsꢃ, A. Kadziola, M. B.
Nielsen, J. Am. Chem. Soc. 2010, 132, 9165–9174.
[11] A. M. Khan, G. R. Proctor, L. Rees, J. Chem. Soc. C 1996, 990–994.
[12] K. C. Srivastava, S. Dev, Tetrahedron 1972, 28, 1083–1091.
[13] S. L. Broman, S. L. Brand, C. R. Parker, M. ꢁ. Petersen, C. G. Tortzen, A.
Kadziola, K. Kilsꢃ, M. B. Nielsen, ARKIVOC 2011, 51–67.
[14] C. R. Parker, C. G. Tortzen, S. L. Broman, M. Schau-Magnussen, K. Kilsꢃ,
M. B. Nielsen, Chem. Commun. 2011, 47, 6102–6104.
[3] a) R. Boese, J. K. Cammack, A. J. Matzger, P. Pflug, W. B. Tolman, K. P. C.
Vollhardt, T. W. Weidman, J. Am. Chem. Soc. 1997, 119, 6757–6773; b) Y.
Kanai, V. Srinivasan, S. K. Meier, K. P. C. Vollhardt, J. C. Grossman, Angew.
Chem. Int. Ed. 2010, 49, 8926–8929; Angew. Chem. 2010, 122, 9110–
9113; c) T. J. Kucharski, Y. Tian, S. Akbulatov, R. Boulatov, Energy Environ.
[15] S. L. Broman, M. Jevric, A. D. Bond, M. B. Nielsen, J. Org. Chem. 2014, 79,
41–64.
´
Sci. 2011, 4, 4449–4472; d) K. Moth-Poulsen, D. Coso, K. Bçrjesson, N.
Vinokurov, S. Meier, A. Majumdar, K. P. C. Vollhardt, R. A. Segalman,
Energy Environ. Sci. 2012, 5, 8534–8537; e) E. Durgun, J. C. Grossman, J.
Phys. Chem. Lett. 2013, 4, 854–860; f) K. Moth-Poulsen, in Organic Syn-
thesis and Molecular Engineering (Ed.: M. B. Nielsen), Wiley, Hoboken,
2014, pp. 179–196; g) T. J. Kucharski, N. Ferralis, A. M. Kolpak, J. O.
Zheng, D. G. Nocera, J. C. Grossman, Nat. Chem. 2014, 6, 441–447; h) K.
Bçrjesson, D. Coso, V. Gray, J. C. Grossman, J. Q. Guan, C. B. Harris, N.
Hertkorn, Z. R. Hou, Y. Kanai, D. Lee, J. P. Lomont, A. Majumdar, S. K.
Meier, K. Moth-Poulsen, R. L. Myrabo, S. C. Nguyen, R. A. Segalman, V.
Srinivasan, W. B. Tolman, N. Vinokurov, K. P. C. Vollhardt, T. W. Weidman,
Chem. Eur. J. 2014, 20, 15587–15604; i) A. Lennartson, A. Roffey, K.
Moth-Poulsen, Tetrahedron Lett. 2015, 56, 1457–1465; j) D. Zhitomirsky,
E. Cho, J. C. Grossman, Adv. Energy Mater. 2015, 1502006.
[16] S. L. Broman, M. Jevric, M. B. Nielsen, Chem. Eur. J. 2013, 19, 9542–9548.
[17] In the calculations, we do not take the excited state dynamics into ac-
count but assume that the bond breaking will lead to the lowest free
energy VHF conformation.
[18] a) P. von Raguꢄ Schleyer, C. Maerker, A. Dransfeld, H. Jiao, N. J. R. van Ei-
kema Hommes, J. Am. Chem. Soc. 1996, 118, 6317–6318; b) Z. Chen,
C. S. Wannere, C. Corminboeuf, R. Puchta, P. von Raguꢄ Schleyer, Chem.
Rev. 2005, 105, 3842–3888.
[19] a) D. S. Pedersen, C. Rosenbohm, Synthesis 2001, 16, 2431–2434; b) S. L.
Broman, A. U. Petersen, C. G. Tortzen, J. Vibenholt, A. D. Bond, M. B. Niel-
sen, Org. Lett. 2012, 14, 318–321.
[4] H. Gçrner, C. Fischer, S. Gierisch, J. Daub, J. Phys. Chem. 1993, 97, 4110–
4117.
[5] a) 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.
Received: March 13, 2016
Revised: July 12, 2016
Published online on && &&, 0000
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
Full of energy: An increase in the solar
energy storage capacity of the dihy-
droazulene (DHA)–vinylheptafulvene
(VHF) photo/thermoswitch is achieved
by fusing a benzene ring to the DHA
isomer (see figure). In contrast, the rela-
tive stability is interchanged by benzan-
nulation of the VHF isomer.
&
Energy Storage and Conversion
A. B. Skov, S. L. Broman,* A. S. Gertsen,
J. Elm,* M. Jevric, M. Cacciarini,
A. Kadziola, K. V. Mikkelsen,*
M. B. Nielsen*
&& – &&
Aromaticity-Controlled Energy Storage
Capacity of the Dihydroazulene-
Vinylheptafulvene Photochromic
System
&
&
Chem. Eur. J. 2016, 22, 1 – 10
www.chemeurj.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!