Azobispyrazole Family as Photoswitches Combining (Near-) Quantitative Bidirectional Isomerization and Widely Tunable Thermal Half-Lives from Hours to Years**

Article abstract of DOI:10.1002/anie.202103705

Azobenzenes are classical molecular photoswitches that have been widely used. In recent endeavors of molecular design, replacing one or both phenyl rings with heteroaromatic rings has emerged as a strategy to expand molecular diversity and access improved photoswitching properties. Many mono-heteroaryl azo molecules with unique structures and/or properties have been developed, but the potential of bis-heteroaryl architectures is far from fully exploited. We report a family of azobispyrazoles, which combine (near-)quantitative bidirectional photoconversion and widely tunable Z-isomer thermal half-lives from hours to years. The two five-membered rings remarkably weaken the intramolecular steric hindrance, providing new possibilities for engineering the geometric and electronic structure of azo photoswitches. Azobispyrazoles generally exhibit twisted Z-isomers that facilitate complete Z→E photoisomerization, and their thermal stability can be broadly adjusted regardless of the twisted shape, overcoming the conflict between photoconversion (favored by the twisted shape) and Z-isomer stability (favored by the orthogonal shape) encountered by mono-heteroaryl azo switches.

Full text of DOI:10.1002/anie.202103705

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Azobispyrazole family as photoswitches combining (near-)
quantitative bidirectional isomerization and widely tunable thermal
half-lives from hours to years
Authors: Yixin He, Zhichun Shangguan, Zhao-Yang Zhang, Mingchen
Xie, Chunyang Yu, and Tao Li
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202103705
Link to VoR: https://doi.org/10.1002/anie.202103705

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
Azobispyrazole family as photoswitches combining (near-)
quantitative bidirectional isomerization and widely tunable
thermal half-lives from hours to years
Yixin He,[a] Zhichun Shangguan,[a] Zhao-Yang Zhang,*[a] Mingchen Xie,[a] Chunyang Yu,[a] and Tao Li*[a]
[a]
Y. He, Z. Shangguan, Dr. Z. Y. Zhang,* M. Xie, Prof. C. Yu, and Prof. T. Li*
School of Chemistry and Chemical Engineering, Frontiers Science Center for Transformative Molecules, Shanghai Key Laboratory of Electrical Insulation
and Thermal Aging
Shanghai Jiao Tong University
Shanghai 200240, China
E-mail: litao1983@sjtu.edu.cn; zy_zhang@sjtu.edu.cn
Supporting information for this article is given via a link at the end of the document.
Abstract: Azobenzenes are classical molecular photoswitches that
example, t1/2 ranging from seconds to hours can match the
have received widespread application. In recent endeavors of timescales of various biological events,[5] while ultralong t1/2 of
molecular design, replacing one or both phenyl rings by months to years are advantageous to information and energy
heteroaromatic ones is emerging as a strategy to expand the storage.[4a,6]
molecular diversity and access improved photoswitch properties.
Azobenzene (Ph-N=N-Ph) itself is
a highly adaptable
Many mono-heteroaryl azo molecules with unique molecular architecture because of its numerous structural
structure/properties have been developed, but the potential of bis- modification possibilities. Since the discovery of azobenzene
heteroaryl architecture is far from exploited. Here we report a family photoisomerization in 1937,[7] introducing substituent groups to
of azobispyrazoles, which combine (near-)quantitative bidirectional phenyl rings has produced a vast number of azobenzene
photoconversion and widely tunable Z-isomer thermal half-lives from derivatives.[8] Important progress includes ortho-substitutions
hours to years. The two five-membered rings remarkably weaken the based visible-light-switchable azobenzenes.[9] In recent
intramolecular steric hindrance, providing new possibilities for the
endeavors of molecular design, replacing one or both phenyl
engineering geometric and electronic structure of azo photoswitches. rings by heteroaromatic ones has attracted increasing
Azobispyrazoles generally exhibit twisted Z-isomers that facilitate attention.[10] The diversity of heterocycles and their distinct
complete Z→E photoisomerization, and their thermal stability can be nature from benzene provide tremendous opportunities to
adjusted independently of geometric shape, overcoming the conflict regulate the performance of azo switches. Six-membered rings
between effective photoconversion (favored by twisted shape) and such as pyridine[11] and pyrimidine[12] afforded heteroaryl azo
Z-isomer stability (favored by T shape) encountered by mono- switches in which coordination interactions and hydrogen
heteroaryl azo switches.
bonding could be utilized. Significantly, mono-heteroaryl azo
molecules (Ph-N=N-Het) based on five-membered rings
(imidazole,[13] pyrazole,[14] pyrrole,[14] thiophene,[15] isoxazole,[16]
etc.) showed some unusual features. Herges group[13b] and
Introduction
Fuchter
group[14a]
reported
that
Z-isomers
of
phenylazoimidazoles and phenylazopyrazoles could adopt a T-
shaped geometry (where the phenyl ring was orthogonal to the
azo-heterocycle moiety) not accessible by azobenzenes. Herges
et al[13b] disclosed that such T geometry could enlarge the
difference in conjugation degree between E and Z isomers and
thus their π-π* absorption bands were well separated, enabling
complete E→Z photoisomerization. A later systematic study by
Fuchter and Contreras-García et al[14b] revealed that the T
geometry was crucial for high Z-isomer stability (t1/2 up to 1000
days) but detrimental to Z→E photoconversion (due to the weak
n-π* absorbance of Z-isomers). Structural modifications that
Molecular photoswitches, which can be reversibly
interconverted between different states by light irradiations, have
inspired intensive efforts in developing light-responsive systems
for diverse applications.[1] Azobenzenes are the most widely
used photoswitches and play dominating roles in many areas,
such as photomechanical actuation of soft materials,[2]
photopharmacological modulation of physiological functions,[3]
and photochemical solar thermal energy storage.[4] For azo
photoswitches, probably the foremost factor determining their
quality is the E ⇆ Z photoisomerization yields in both directions.
In general, quantitative bidirectionality is highly desirable, since
incomplete photoconversion in any direction will compromise the
capability of photoswitching. Another essential property is the
thermal stability of metastable isomers (Z-forms, in general),
which spontaneously relax back to the thermodynamically stable
states. The suitable thermal back-isomerization kinetics,
represented by half-lives t1/2 (values are at 25 °C hereinafter, if
not specified), depends on the application of interest. For
twisted the conformation could improve the photoisomerization
[14b,17]
yields, which, however, shortened their Z-isomer t1/2
.
Phenylazothiophenes also allowed T-shaped Z-isomers; they
similarly displayed complete E→Z and unfavorable Z→E
photoconversion, while their t1/2 (20 °C) were at the timescale of
hours or less.[15] Azo switches bearing a benzoheterocycle such
as quinoline,[12b] indole,[18] indazole,[19] and purine[20] were also
1
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
reported. Their extended
π
system generally led to
bathochromic shifts of π-π* absorption bands and accelerated
back relaxation with t1/2 from nanoseconds to days. In spite of
the important progress on mono-heteroaryl azo switches, a great
challenge remains on how to achieve quantitative bidirectional
photoconversion and, at the same time, tune their t1/2 from short
to very long.
Bis-heteroaryl azo architecture (Het-N=N-Het) can further
expand the molecular diversity and offers a probably more
progressive way in tailoring the geometric and electronic
structure, which, however, has not received sufficient attention.
Fuchter group[21] designed and synthesized azobis(2-imidazole),
a visible-light-responsive azo switch with 90% and ~100%
conversions for E→Z and Z→E directions, respectively. Its Z-
isomer adopted a twisted conformation and t1/2 was only 16 s.
Very recently, Beves et al[22] studied some azobisbenzazoles
bearing benzimidazoles, benzoxazole or benzothiazoles, which
showed 73 – 84% E→Z conversions under visible light. Notably,
azobisbenzimidazole obtained a t1/2 of 520 s, which was among
the
longest
known
for
visible-light-switchable
Scheme 1. Synthesis routes of azobispyrazoles. Reagent and conditions. (a)
AcOH, concentrated HCl, aqueous NaNO2, 0 °C, then aqueous KOAc, MDA-
Na, 0 – 5 °C, 51 – 54%; (b) Methylhydrazine sulfate, KOAc, EtOH, refluxed, 65
azoheteroarenes.[22] Azobisthiophenes were shown to be
different from phenylazothiophenes in that their Z-isomers were
also twisted and t1/2 were up to 38 days (22 °C), yet their
photoisomerization yields remain to be determined.[23] By
–
80%; (c) MnO2, CHCl3, 80 °C, 25%; (d) CH3I, Cs2CO3, DMF, room
temperature, 83 – 85%.
combining
electron-rich
(thienyl)pyrrole/bithiophene
and
electron-deficient (benzo)thiazole, a series of push-pull bis-
heteroaryl azo switches were constructed, which were
responsive to visible light and their Z-isomer t1/2 ranged from
milliseconds to seconds.[24] In view of the very limited types of
bis-heteroaryl azo switches reported so far and their limitations
on bidirectional photoconversion and/or Z-isomer stability, the
potential of Het-N=N-Het architecture is far from exploited,
calling for the development of new molecular systems.
by systematic experimental and theoretical investigations,
azobispyrazoles show different structure-property relationships
from mono-heteroaryl azo switches. In particular, their twisted Z-
isomers can accommodate a wide range of t1/2 timescales
including hours, days, weeks, months, and years. Therefore, our
work demonstrates the great potential of Het-N=N-Het
architecture in molecular design, performance improvement, and
broad applications of azo photoswitches.
Here we present
a family of azobispyrazoles as new
photoswitches, as shown in Figure 1, all of which show excellent
bidirectional photoconversion, and their Z-isomer t1/2 can be
broadly tuned from hours to years by changing the linking
position between the azo group and pyrazole rings. As revealed
Results and Discussion
Molecular Design and Synthesis
By bridging two pyrazole rings with an azo group at C3, C4, or
C5 positions, a family of six azobispyrazoles can be formulated,
including three symmetrical members (3-3, 4-4, and 5-5) and
three asymmetrical members (3-4, 3-5, and 4-5), as shown in
Figure 1. Because of the distinct nature of the three kinds of
carbons on pyrazole, variations in linkage pattern can regulate
the molecular geometries and electronic structure of the whole
system, which is expected to have a significant effect on their
photoswitch properties.
Synthesis routes of azobispyrazole family were shown in
Scheme 1. In previous studies, phenylazopyrazoles containing
3,5-dimethylated 4-pyrazole were synthesized by diazo
couplings between phenylamine and acetylacetone with
subsequent ring formation.[14a,25] Inspired by this method, we first
attempted to prepare our azobispyrazoles containing methyl-free
4-pyrazole (3-4, 4-4, and 4-5) using pyrazolylamines and
malonaldehyde that was generated from in situ hydrolysis.
However, the yield of diazo coupling step was uncontrollable (5
– 50% in different trials) since the malonaldehyde was prone to
polymerize in solution during the hydrolytic process. Therefore,
we employed malonaldehyde sodium salt (MDA-Na), which can
Figure 1. Molecular design of azobispyrazole family (N-methylation is
necessary to prevent the azo-hydrazone or annular tautomerization).
2
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
be isolated in pure solids and is stable over weeks in dry
atmosphere.[26] In this way, pyrazolyl hydrazone intermediates
(3-HM, 4-HM, and 3(5)-HM) were obtained in reproducible yields
of ~50%. Further reactions with methylhydrazine sulfate
constructed the 4-pyrazole rings, forming the other moieties of
azobispyrazoles. The other three members (3-3, 3-5, and 5-5)
were prepared in one pot through oxidation coupling of 1H-
pyrazol-3(5)-amine[27,28] and subsequent N-methylation at two
selectable positions, benefitting from the annular tautomerism of
1H-pyrazole. 3-3 and 5-5 were also obtained by electrochemical
synthesis in previous work.[29] The single-crystal structures of all
six members have been determined by X-ray crystallographic
analysis (see Section 4 in the Supporting Information, SI). We
note that photochromic properties of 3(5)-2H were recently
studied by König group.[28] It formed an equilibrium mixture of
different tautomers (due to annular or azo-hydrazone
tautomerism), and the preferred tautomeric form depended on
solvent environment, which remarkably affected molecular
property including photoisomerization yields and Z-state stability.
Photoisomerization and Electronic Transition Analysis
The photoisomerization properties of azobispyrazoles were
studied by UV-Vis absorption spectroscopy in acetonitrile
solution. As shown in Figure 2, the E-isomers exhibited strong
π-π* absorption bands in 300 – 400 nm region (εmax = 18 –
27×103 M-1 cm-1), and the peak positions varied with the linkage
pattern. As detailed in Table 1 and Figure 2, the π-π* λmax red-
shifted from 325 – 327 nm (3-3, 3-4, and 4-4) to 342 nm (3-5
Table 1. Spectroscopic data, E ⇆ Z photoisomerization and thermal Z→E isomerization properties of azobispyrazole family.
E-isomer[a] Z-isomer[a] Photoconversion[b] Quantum yield[c]
Thermal Z→E isomerization
[k]
[k]
[k]
[l]
π-π* λmax
(nm)
n-π* λmax
(nm)
π-π* λmax
(nm)
n-π* λmax
(nm)
E→Z
(%)
Z→E
(%)
ΔG‡
ΔH‡
ΔS‡
ΔG‡
calc
exp
exp
exp
E→Z
Z→E
t1/2 (d)
(kJ mol-1)
(kJ mol-1)
(J mol-1 K-1)
(kJ mol-1)
4-4
3-4
3-3
4-5
3-5
5-5
327
326
325
342
342
356
~395
~394
400
–
277
277
277
285
293
313
420
414
430
430
441
454
98[d]
97[d]
97[d]
97[e]
97[e]
94[f]
99 / ~100[g]
97 / 99[g]
0.43[d]
0.32[d]
0.34[d]
0.47[e]
0.42[e]
0.36[e]
0.76[h]
0.47[h]
0.62[h]
0.72[h]
0.56[h]
0.49[h]
681[i]
623[i]
118.3
118.0
112.7
110.1
106.5
99.1
105.6
104.2
96.1
95.4
91.1
86.2
-42.5
-46.4
-55.7
-49.4
-51.4
-43.3
109.2
108.0
103.9
100.9
99.3
98 / 99[g]
72[i]
25[i]
98 / 99[g]
~413
–
98 / 99[g]
6[i] (7[j])
0.3[i] (0.3[j])
99 / ~100[g]
91.4
[a] The λmax were taken from UV-Vis absorption spectra measured in MeCN (4×10-5 M). [b] The isomer composition were obtained through integrating the peak
areas of 1H NMR spectra measured in CD3CN solution (see Section 2.2 in SI). [c] The quantum yields were determined in MeCN, and the errors were within ±
0.02 (see Section 2.4 in SI). [d‒h] The samples were irradiated with lights of [d] 350 nm, [e] 365 nm, [f] 370 nm, [g] 523 / 530‒550 nm, and [h] 470 nm. [i]Values at
25 °C in DMSO, extrapolated from the experimental kinetics data at elevated temperatures using Arrhenius equation (see Section 3 in SI). [j] Values at 25 °C
measured in MeCN. [k] Values at 25 °C in DMSO, calculated through Eyring plot based on the experimental kinetics data (see Section 3 in SI). [l] Values at 25 °C
in DMSO, computed at B2PLYP-D3(BJ)/def2-TZVP//PBE0-D3(BJ)/6-31G** level of theory with solvation free energies at M05-2X/6-31G* level, considering the
weights of all possible pathways (Table S21).
Figure 2. UV-Vis absorption spectra with molar extinction coefficients of azobispyrazoles in MeCN (see the zoomed-in spectra of n-π* regions in Figure S2).
Insets show the E-isomer contents at 350 – 550 nm photostationary states (see tabulated results in Table S2 and corresponding spectra in Figure S9).
3
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
and 4-5) and further to 356 nm (5-5). In all cases, irradiating the
E-isomers at π-π* bands with appropriate light sources
(350/365/370 nm) resulted in almost complete conversions (94 –
98%) to Z-isomers, benefitting from the overwhelming
absorbance of E over Z isomers at the excited wavelengths.
These Z-isomers displayed more evident n-π* absorption bands
in visible range than E-isomers, and exciting the n-π* transition
at band tails using green lights (523 nm or longer wavelengths)
could induce quantitative backward isomerization for all
azobispyrazole members. In contrast, the mono-pyrazolyl azo
corresponded to HOMO-1→LUMO transition (Figure 3a). It
appeared that n-π* transitions require less excitation energies
than π-π* but with larger gaps (H-1–L gaps vs. H–L gaps). This
discrepancy,
which
was
also
found
in
other
azoheteroarenes,[13a,14b,16,17c,18a] can be ascribed to the difference
in exciton binding energies.[30]
For Z-isomers, the frontier molecular orbital (FMO) theory
could not well describe the nature of electronic excitations, since
the frontier occupied molecular orbitals mixed
n and π
composition and each transition was contributed by multiple sets
switches (phenylazopyrazoles), regardless of the linkage pattern, of orbitals. Therefore, the natural transition orbital (NTO)
suffered from heavy overlapping of the n-π* bands between two
isomers, which restricted the yields/efficiency of their Z→E
photoisomerization,[14] and thus structural modification by
substitution at the phenyl/pyrazole ring was necessary to
improve the performance.[14a,17] Moreover, 5-5 could be used as
a visible-light-induced photoswitch, which provided 85% E→Z
and ~100% Z→E conversions by 400 and 530 – 550 nm light
irradiations, respectively (Figure 2, inset and Table S2).
Azobispyrazoles could be soluble in various solvents even in
water (slightly soluble), therefore we examined the solvent effect
on UV-Vis absorption spectra using five solvents (n-hexane,
EtOH, MeCN, DMSO, water). As shown in Table S5, the π-π*
bands of E-isomers generally red-shifted with the increase of
solvent polarity, while water induced blue shifts for the 5-
pyrazole series (3-5, 4-5, and 5-5). The n-π* λmax of Z-isomers
did not show obvious variation with solvent polarity except that
water induced blue shifts again. On the whole, solvents had a
limited effect on spectral characteristics and did not apparently
affect E ⇆ Z photoisomerization yields (see the UV-Vis spectra
in different solvents in Figures S12 – S15).
analysis[31,32] was performed, and we found that the S0→S1 and
S0→S2 excitations could be almost identified as n-π* and π-π*
transitions, respectively (Figure 3b).
All E-isomers of azobispyrazoles showed a planar geometry
and their n-π* transitions were symmetry-forbidden with
negligible oscillator strength (f) of 0.0000 – 0.0001 (Tables S22
– S24). By contrast, the Z-isomers generally adopt a twisted
geometry where two rings and azo group formed three dihedral
angles φ1–3 (acute angles), as illustrated in Figure 4. This broke
the geometric symmetry and made n-π* transitions more
allowed with remarkably increased f of 0.01 – 0.09 (Tables S22
– S24). As a result, Z-isomers showed relatively strong n-π*
absorbance against E-isomers, responsible for quantitative Z→E
photoisomerization.
We then determined the photoisomerization quantum yields
of azobispyrazoles in acetonitrile solution (see results in Table 1
and S4). For E→Z (π-π* excitation) and Z→E (n-π* excitation)
isomerizations, the quantum yields were 0.32 – 0.47 and 0.47 –
0.76, respectively, which were generally higher than those of
azobenzene (0.10 – 0.15 and 0.41 – 0.58).[8a] Quantum yield
results of heteroaryl azo were very limited in literature, but the
advantage of Ph-N=N-Het molecules over azobenzene was also
shown by azoimidazoles,[13a] azopyrrole,[14a] azopyrazoles,[14a,28]
and azothiophene.[15a] More experimental and theoretical efforts
are needed to gain a clear understanding that how heteroaryl
substitution makes the photochemical processes more
productive.
In order to reveal the relationship between molecular
structures and electronic transition characteristics, density
functional theory (DFT) and time-dependent DFT studies were
carried out for all possible conformers of azobispyrazoles (see
Section 5 in SI). For the E-isomers, the π-π* absorption band
(S0→S2 excitation) corresponded to HOMO→LUMO transition
(Figure 3a). The calculated λmax agreed well with the
experimental values, which red-shifted with the reduction of H–L
gaps. This reflected the dependence of π-conjugation degree on
the linkage pattern. According to the electron-donating
conjugative effect from pyrrole-type N lone pair, the 5-pyrazole
forms a “complete” conjugation pathway with the azo group
while those from 4- and 3-pyrazole are only “partial”.[14b] As a
result, the molecular architecture bearing more 5-pyrazole units
showed larger π-conjugation extension in pyrazole-azo-pyrazole
system, making the π-π* λmax follow the order of 5-5 > 3(4)-5 >
3(4)-3(4). The n-π* band (S0→S1 excitation) of E-isomers
Figure 3. (a) Frontier molecular orbitals and (b) natural transition orbitals for
the predominant conformers of 4-4 (isosurface value of 0.05). Results of other
molecules were illustrated in Figures S35 – S40.
4
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
Z-3-3
Z-3-4
(3) (4)
Z-3-5
(4) (3)
Z-4-4
(2) (1)
Z-4-5
(3) (1)
Z-5-5
(1) (3)
Z-Azobenzene
(Z-Azo)
Conformation
Population (%)
φ1 (°)
(2)
(1)
60.4 23.5
53.9 39.2
9.8 12.0
18.6 39.2
36.3 39.2
41.6 30.0
0.0 44.5
75.0 13.8
25.6 0.0
11.2 0.0
34.7 0.0
30.2 0.0
64.2 30.7
36.5 32.7
50.7 24.9
70.9 39.3
4.8 10.3
7.1 19.3
39.0 29.3
87.8 12.2
13.6 28.8
7.2 11.0
63.6 28.8
38.6 28.8
100
51.6
9.1
φ2 (°)
0.0
9.8
9.4
9.9
φ3 (°)
0.0 20.3
0.0 32.4
28.6 32.7
32.6 32.7
51.6
51.6
Avg.(φ1,φ3) (°)
Figure 4. The geometric structure and dihedral angles of Z-isomers (optimized at PBE0-D3(BJ)/6-31G** level of theory in MeCN).
Geometric Insight
Interestingly, a couple of planar conformers were found in Z-3-
4 and Z-3-5 (φ1 = φ2 = φ3 = 0°), which were stabilized by an
intramolecular C–H···N hydrogen bond between the pyridine-
type N atom on 3-pyrazole ring and a C–H group on the other
ring (Figure 4 and S31). These planar conformers showed small
f of 0.0018 – 0.0026 (Tables S22 – S24). Z-3-4 had two planar
conformers (conformer (1) and (3), see Figure 4 and S31) and
their total population accounted for 55% (Boltzmann distribution,
see Table S19), explaining its relatively low n-π* band observed
in Figure 2, while the absorbance of Z-3-5 was of “normal”
intensity because its sole planar conformer (3) was of small
distribution (only 14%). Planar Z-isomers were never observed
in Ph-N=N-Ph or Ph-N=N-Het type molecules but were also
found in other five-membered ring based azobisheteroarenes
involving C–H···N hydrogen bond, including azobis(2-
imidazole),[21] its protonated product,[21] and protonated
azobis(benzothiazole).[22] This unique phenomenon shows again
that two five-membered rings result in weakened steric
hindrance, and it can be overcome by a weak hydrogen bond.
For the E-isomers, two five-membered rings provide a more
spacious molecular environment to accommodate bulky
substituents. This was directly demonstrated using ortho methyl
as the probing group. As shown in Figure 5b, the introduction of
four ortho methyl groups to azobenzene forced both phenyl rings
to rotate away from coplanarity with the azo group, resulting in a
highly twisted geometry. When one of the phenyl rings was
replaced by a pyrazole ring, half of the crowdedness was
eliminated. By contrast, the azobispyrazole structure was free of
any steric perturbation and kept the planar geometry. In this
case, the electronic excitation characteristics of the parent
molecule were not substantially affected by the bulky methyls
(Table S26). It is known that the twisted geometry inevitably
interrupts the conjugation of a π-system, which is undesirable in
many conditions. Introducing fluorines is the only reported way
to maintain the tetra-ortho-substituted azobenzene planar,
benefiting from the small atomic size.[9c] Azobispyrazole and
other Het-N=N-Het molecules may act as a more versatile
platform where electronic effect of bulky substituents can also be
utilized without affecting the effective conjugation.
For the Z-isomer geometry, azobispyrazoles are actually in a
similar situation with azobenzenes (also twisted) where the two
rings and azo group avoid coplanarity due to steric reasons.
However, Z-azobispyrazoles are less twisted from planar
geometry than Z-azobenzene (Z-Azo): as detailed in Figure 4,
Z-Azo is characterized by large φ1 and φ3 (both over 50°)[33]
while for Z-azobispyrazoles at least one of the twist angles is
smaller than 30° (averaged values of φ1 and φ3: 28 – 40°). This
difference indicates that the intramolecular steric hindrance
associated with six-membered ring is remarkably weakened by
two five-membered rings. This viewpoint can also be understood
from an energetic perspective. We calculated the Gibbs free
energy difference between the twisted geometry (energy
minimum conformation) and a near-planar geometry (φ1 =φ2 =φ3
= 10°). As shown in Figure 5a, the energy increase was 190
kJ/mol for Z-Azo to reach the near-planar geometry, but much
lower energy increases of 49 – 101 kJ/mol were found for the
series of Z-azobispyrazoles.
Figure 5. (a) Gibbs free energy difference between normal twisted and
restricted near-planar geometries (computed at B2PLYP-D3(BJ)/def2-
TZVP//PBE0-D3(BJ)/6-31G** level of theory in MeCN with solvation free
energies at M05-2X/6-31G* level). (b) The geometric structure and dihedral
angles of ortho-tetramethyl azo molecules (optimized at PBE0-D3(BJ)/6-31G**
level of theory in MeCN).
5
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
Thermal Z→E Isomerization
processes via rotation and inversion by relaxed potential energy
scans, as depicted in Figure 6a. Our results suggested that the
lowest energy pathway followed the rotational mechanism,
instead of the inversion mechanism found for mono-pyrazolyl
azo switches.[14b] Based on the computed Gibbs free energies of
Z-isomers and transition states (GZ and GTS), theoretical free
energy barriers (ΔG‡) were obtained (Table 1 and S21), which
were well correlated with the experimental values, as illustrated
in Figure 6b.
Since GTS and GZ were influenced by intramolecular
stabilizing/destabilizing interactions, the noncovalent interaction
(NCI) analysis[32,35] of transition states (TSs) and Z-isomers was
performed. The results of their predominant conformers were
shown in Figure 6c. For the TSs of the molecules with 5-
pyrazole (4-5, 3-5, and 5-5), the bulky N-methyl on 5-pyrazole
were adjacent to the azo group, enabling them to form
dispersive interactions, while such stabilizing interactions were
absent for the other three molecules (their N-methyls were away
from azo groups). Therefore, it was easier for the 5-pyrazole
series to overcome the energy barriers. For the Z-isomers of this
series, the N-methyl group on 5-pyrazole of 4-5 orientated
towards the side of the remaining azo-pyrazolyl part; it gave rise
to the largest dispersive surface that stabilized Z-4-5 the most,
The thermal Z→E isomerization of azobispyrazoles followed
first-order kinetics (see Section 3 in SI), and their Z-isomer t1/2
and thermodynamic activation parameters at 25 °C were listed in
Table 1. Very long t1/2 of nearly 2 years were achieved by 4-4
and 3-4, making them rare azo switches possessing highly
stable Z-isomers.[9c,14a,34] The t1/2 could be further regulated to
months, weeks, days, and hours by adjusting the linkage pattern.
Such a widely tunable t1/2 could meet the kinetic requirements of
a broad spectrum of photo-responsive systems. We could find
that the Z-isomer stability did not rely on its geometric shape,
since the twisted geometry allowed widely tunable t1/2. This was
in stark contrast with the structure-property relationship
established for mono-heteroaryl molecular system based on
imidazole, pyrazole, pyrrole, triazole, and tetrazole, for which the
T and twisted shape dictates long and short t1/2, respectively.[14b]
A clear trend between Z-isomer t1/2 and bridging position was
found: 4-pyrazole unit favored long t1/2 the most, followed by the
3-pyrazole and finally the 5-pyrazole, and t1/2 took orders from
the combination of the two units, i.e., 4-4 > 3-4 > 3-3 > 4-5 > 3-
5 > 5-5. In order to understand the thermal Z→E isomerization
behavior of azobispyrazole family, we studied their isomerization
Figure 6. (a) Typical relaxed potential energy scans of the thermal Z→E isomerization of azobispyrazoles in DMSO (taking 3-3 as an example) at PBE0-D3(BJ)/6-
31G** level of theory (using broken symmetry (BS) approach for rotational pathway). (b) Linear correlation between experimental and computed free energy
barriers for azobispyrazoles in DMSO. (c) NCI surfaces of the predominant Z-isomers and transition states for azobispyrazoles in DMSO (isosurface value of
0.75). The surfaces were colored on a blue-green-red scale according to the values of sign(λ2)·ρ, ranging from -0.04 to 0.02 au.[35]
6
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
explaining its longest t1/2 in this series. The higher stability of Z-3-
5 than Z-5-5 was probably attributed to the relatively stronger
stabilizing effect of intramolecular C–H···N hydrogen bond in
planar Z-3-5. For the other three molecules (4-4, 3-4, and 3-3),
their Z-isomers displayed similar NCI surfaces, and their TSs
faced the same situation, which made it difficult to qualify the
difference in their energy barriers through NCI pictures. Indeed,
ΔG‡ of their considered conformers were only slightly different
(ΔG‡ – ΔG‡ = 1.0 kJ/mol, ΔG‡ – ΔG‡ = 2.6 kJ/mol).
Interestingly, for the TSs, a dispersive interaction between the
N-methyl and adjacent C–H group on 3-pyrazole was observed,
whereas this interaction on 4-pyrazole was negligible. This might
contribute to the lower ΔG‡ of the molecules with more 3-
pyrazole units.
Conflict of interest
The authors declare no competing financial interest.
Keywords: azobenzene • heterocycle • molecular photoswitch •
photoisomerization • pyrazole
[1]
a) W. Szymański, J. M. Beierle, H. A. V. Kistemaker, W. A. Velema, B.
L. Feringa, Chem. Rev. 2013, 113, 6114-6178; b) K. Glusac, Nat.
Chem. 2016, 8, 734-735; c) M. Baroncini, S. Silvi, A. Credi, Chem. Rev.
2020, 120, 200-268; d) M. Kathan, S. Hecht, Chem. Soc. Rev. 2017, 46,
5536-5550; e) J. Zhang, Q. Zou, H. Tian, Adv. Mater. 2013, 25, 378-
399; f) A. Goulet-Hanssens, F. Eisenreich, S. Hecht, Adv. Mater. 2020,
32, 1905966; g) L. Wang, Q. Li, Chem. Soc. Rev. 2018, 47, 1044-1097;
h) X. Huang, T. Li, J. Mater. Chem. C 2020, 8, 821-848; i) C. Sun, C.
Wang, R. Boulatov, ChemPhotoChem 2019, 3, 268-283; j) W. Xu, S.
Sun, S. Wu, Angew. Chem. Int. Ed. 2019, 58, 9712-9740.
4-4
3-4
3-4
3-3
Conclusion
[2]
[3]
a) A. H. Gelebart, D. Jan Mulder, M. Varga, A. Konya, G. Vantomme, E.
W. Meijer, R. L. B. Selinger, D. J. Broer, Nature 2017, 546, 632-636; b)
T. Ube, T. Ikeda, Angew. Chem. Int. Ed. 2014, 53, 10290-10299.
a) J. Broichhagen, J. A. Frank, D. Trauner, Accounts Chem. Res. 2015,
48, 1947-1960; b) F. Riefolo, C. Matera, A. Garrido-Charles, A. M. J.
Gomila, R. Sortino, L. Agnetta, E. Claro, R. Masgrau, U. Holzgrabe, M.
Batlle, M. Decker, E. Guasch, P. Gorostiza, J. Am. Chem. Soc. 2019,
141, 7628-7636.
In conclusion, we have designed and investigated a family of
azobispyrazoles as new bis-heteroaryl azo photoswitches. All six
members show excellent bidirectional E ⇆ Z photoconversion,
and their Z-isomer t1/2 cover a wide range of timescales including
years, months, weeks, days, and hours. Azobispyrazole system
bearing 5-pyrazole moiety shows higher π-conjugation, so the π-
π* λmax of E-isomer follows the order of 5-5 > 3(4)-5 > 3(4)-3(4).
5-5 can be used as a bidirectionally visible-light-activated
photoswitch. The linkage pattern also directs the thermal back-
isomerization kinetics, and t1/2 follow the order of 4-4 > 3-4 > 3-3
> 4-5 > 3-5 > 5-5. This can be explained by the different
intramolecular stabilizing interactions brought by 4-, 3-, 5-
pyrazole units in TSs and Z-isomers.
We reveal that the two five-membered rings on Het-N=N-Het
architecture can remarkably weaken the intramolecular steric
hindrance compared with phenyl ring(s) on Ph-N=N-Ph and Ph-
N=N-Het. This brings new possibilities for engineering the
geometric and electronic structure of azo photoswitches. Z-
azobispyrazoles generally adopt a twisted geometry, which
favors n-π* absorption strength and facilitates complete Z→E
photoisomerization. Moreover, the twisted Z-isomers provide
widely tunable thermal t1/2 without compromising the
photoisomerization yields. This overcomes the inherent conflict
between effective Z→E photoisomerization (favored by twisted
shape) and Z-isomer stability (favored by T shape) that exists in
the Ph-N=N-Het class based on five-membered N-heterocycles
(imidazole, pyrazole, pyrrole, triazole, and tetrazole). Therefore,
azobispyrazoles show great potential in developing high-
performance photocontrollable systems and can inspire the
rational design of new photoswitches making use of Het-N=N-
Het architecture.
[4]
a) L. Dong, Y. Feng, L. Wang, W. Feng, Chem. Soc. Rev. 2018, 47,
7339-7368; b) S. Wu, H. J. Butt, Macromol. Rapid Comm. 2020, 41,
1900413; c) Z. Zhang, Y. He, Z. Wang, J. Xu, M. Xie, P. Tao, D. Ji, K.
Moth-Poulsen, T. Li, J. Am. Chem. Soc. 2020, 142, 12256-12264; d) Z.
Wang, R. Losantos, D. Sampedro, M. Morikawa, K. Börjesson, N.
Kimizuka, K. Moth-Poulsen, J. Mater. Chem. A 2019, 7, 15042-15047.
a) W. Lin, M. Tsai, R. Rajappa, R. H. Kramer, J. Am. Chem. Soc. 2018,
140, 7445-7448; b) R. J. Mart, R. K. Allemann, Chem. Commun. 2016,
52, 12262-12277.
[5]
[6]
S. Hvilsted, C. Sánchez, R. Alcalá, J. Mater. Chem. 2009, 19, 6641-
6648.
[7]
[8]
G. S. Hartley, Nature 1937, 140, 281.
a) H. M. D. Bandara, S. C. Burdette, Chem. Soc. Rev. 2012, 41, 1809-
1825; b) E. Merino, Chem. Soc. Rev. 2011, 40, 3835-3853.
[9]
a) O. Sadovski, A. A. Beharry, F. Zhang, G. A. Woolley, Angew. Chem.
Int. Ed. 2009, 48, 1484-1486; b) A. A. Beharry, O. Sadovski, G. A.
Woolley, J. Am. Chem. Soc. 2011, 133, 19684-19687; c) D. Bléger, J.
Schwarz, A. M. Brouwer, S. Hecht, J. Am. Chem. Soc. 2012, 134,
20597-20600; d) H. A. Wegner, Angew. Chem. Int. Ed. 2012, 51, 4787-
4788; e) S. Samanta, T. M. McCormick, S. K. Schmidt, D. S. Seferos, G.
A. Woolley, Chem. Commun. 2013, 49, 10314-10316; f) M. Dong, A.
Babalhavaeji, S. Samanta, A. A. Beharry, G. A. Woolley, Accounts
Chem. Res. 2015, 48, 2662-2670; g) M. J. Hansen, M. M. Lerch, W.
Szymanski, B. L. Feringa, Angew. Chem. Int. Ed. 2016, 55, 13514-
13518; h) L. N. Lameijer, S. Budzak, N. A. Simeth, M. J. Hansen, B. L.
Feringa, D. Jacquemin, W. Szymanski, Angew. Chem. Int. Ed. 2020, 59,
21663-21670.
[10] S. Crespi, N. A. Simeth, B. König, Nature Reviews Chemistry 2019, 3,
133-146.
[11] a) Y. Wang, X. Ge, G. Schull, R. Berndt, H. Tang, C. Bornholdt, F.
Koehler, R. Herges, J. Am. Chem. Soc. 2010, 132, 1196-1197; b) S.
Venkataramani, U. Jana, M. Dommaschk, F. D. Sönnichsen, F. Tuczek,
R. Herges, Science 2011, 331, 445.
Acknowledgements
[12] a) E. Procházková, L. Čechová, J. Kind, Z. Janeba, C. M. Thiele, M.
Dračínský, Chem. Eur. J. 2018, 24, 492-498; b) Y. Xu, C. Gao, J.
Andréasson, M. Grøtli, Org. Lett. 2018, 20, 4875-4879.
This work was supported by the National Key Research and
Development Program of China (2017YFA0207500), National
Natural Science Foundation of China (22022507, 51973111),
[13] a) J. Otsuki, K. Suwa, K. K. Sarker, C. Sinha, J. Phys. Chem. A 2007,
111, 1403-1409; b) T. Wendler, C. Schütt, C. Näther, R. Herges, J. Org.
Chem. 2012, 77, 3284-3287.
Beijing
National
Laboratory
for
Molecular
Sciences
(BNLMS202004) and China Postdoctoral Science Foundation
(2020M681279). We acknowledge the Instrumental Analysis
Center of Shanghai Jiao Tong University, especially the help of
Dr. Lingling Li for single-crystal XRD analysis.
[14] a) C. E. Weston, R. D. Richardson, P. R. Haycock, A. J. P. White, M. J.
Fuchter, J. Am. Chem. Soc. 2014, 136, 11878-11881; b) J. Calbo, C. E.
Weston, A. J. P. White, H. S. Rzepa, J. Contreras-García, M. J. Fuchter,
J. Am. Chem. Soc. 2017, 139, 1261-1274.
7
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
[15] a) C. Slavov, C. Yang, A. H. Heindl, H. A. Wegner, A. Dreuw, J.
Wachtveitl, Angew. Chem. Int. Ed. 2020, 59, 380-387; b) A. H. Heindl,
H. A. Wegner, Chem. Eur. J. 2020, 26, 13730-13737.
[16] P. Kumar, A. Srivastava, C. Sah, S. Devi, S. Venkataramani, Chem.
Eur. J. 2019, 25, 11924-11932.
[17] a) S. Devi, M. Saraswat, S. Grewal, S. Venkataramani, J. Org. Chem.
2018, 83, 4307-4322; b) L. Stricker, M. Böckmann, T. M. Kirse, N. L.
Doltsinis, B. J. Ravoo, Chem. Eur. J. 2018, 24, 8639-8647; c) Z. Y.
Zhang, Y. He, Y. Zhou, C. Yu, L. Han, T. Li, Chem. Eur. J. 2019, 25,
13402-13410.
[18] a) N. A. Simeth, S. Crespi, M. Fagnoni, B. König, J. Am. Chem. Soc.
2018, 140, 2940-2946; b) N. A. Simeth, A. Bellisario, S. Crespi, M.
Fagnoni, B. König, J. Org. Chem. 2019, 84, 6565-6575.
[19] R. Travieso-Puente, S. Budzak, J. Chen, P. Stacko, J. T. B. H.
Jastrzebski, D. Jacquemin, E. Otten, J. Am. Chem. Soc. 2017, 139,
3328-3331.
[20] D. Kolarski, W. Szymanski, B. L. Feringa, Org. Lett. 2017, 19, 5090-
5093.
[21] C. E. Weston, R. D. Richardson, M. J. Fuchter, Chem. Commun. 2016,
52, 4521-4524.
[22] A. D. W. Kennedy, I. Sandler, J. Andréasson, J. Ho, J. E. Beves, Chem.
Eur. J. 2020, 26, 1103-1110.
[23] P. R. Huddleston, V. V. Volkov, C. C. Perry, Phys. Chem. Chem. Phys.
2019, 21, 1344-1353.
[24] a) M. M. M. Raposo, A. M. C. Fonseca, M. C. R. Castro, M. Belsley, M.
F. S. Cardoso, L. M. Carvalho, P. J. Coelho, Dyes Pigm. 2011, 91, 62-
73; b) P. J. Coelho, M. C. R. Castro, A. M. C. Fonseca, M. M. M.
Raposo, Dyes Pigm. 2012, 92, 745-748; c) P. J. Coelho, M. C. R.
Castro, S. S. M. Fernandes, A. M. C. Fonseca, M. M. M. Raposo,
Tetrahedron Lett. 2012, 53, 4502-4506; d) J. Garcia-Amorós, M. C. R.
Castro, P. Coelho, M. M. M. Raposo, D. Velasco, Chem. Commun.
2013, 49, 11427-11429.
[25] H. V. Patel, K. A. Vyas, S. P. Pandey, P. S. Fernandes, Synthetic
Commun. 1992, 22, 3081-3087.
[26] A. Trivella, S. Coussan, T. Chiavassa, Synthetic Commun. 2008, 38,
3285-3290.
[27] P. López-Tarifa, G. Sánchez-Sanz, I. Alkorta, J. Elguero, D. Sanz, A.
Perona, R. M. Claramunt, J. Mol. Struct. 2012, 1015, 138-146.
[28] K. Rustler, P. Nitschke, S. Zahnbrecher, J. Zach, S. Crespi, B. König, J.
Org. Chem. 2020, 85, 4079-4088.
[29] a) B. V. Lyalin, V. L. Sigacheva, V. A. Kokorekin, T. Y. Dutova, G. M.
Rodionov, V. A. Petrosyan, Russ. Chem. B.+ 2018, 67, 510-516; b) B.
V. Lyalin, V. L. Sigacheva, V. A. Kokorekin, V. A. Petrosyan,
Tetrahedron Lett. 2018, 59, 2741-2744.
[30] J. Bredas, Mater. Horizons 2014, 1, 17-19.
[31] R. L. Martin, J. Chem. Phys. 2003, 118, 4775-4777.
[32] T. Lu, F. Chen, J. Comput. Chem. 2012, 33, 580-592.
[33] D. B. Konrad, G. Savasci, L. Allmendinger, D. Trauner, C. Ochsenfeld,
A. M. Ali, J. Am. Chem. Soc. 2020, 142, 6538-6547.
[34] J. Calbo, A. R. Thawani, R. S. L. Gibson, A. J. P. White, M. J. Fuchter,
Beilstein J. Org. Chem. 2019, 15, 2753-2764.
[35] E. R. Johnson, S. Keinan, P. Mori-Sánchez, J. Contreras-García, A. J.
Cohen, W. Yang, J. Am. Chem. Soc. 2010, 132, 6498-6506.
8
This article is protected by copyright. All rights reserved.

10.1002/anie.202103705
Angewandte Chemie International Edition
RESEARCH ARTICLE
Entry for the Table of Contents
Insert graphic for Table of Contents here.
Azobispyrazoles are a family of bis-heteroaryl azo photoswitches that combine excellent E ⇆ Z photoisomerization yields and widely
tunable Z-isomer thermal half-lives from hours to years, showing the great potential of Het-N=N-Het architecture in developing high-
performance molecular photoswitches.
9
This article is protected by copyright. All rights reserved.