Tuning azoheteroarene photoswitch performance through heteroaryl design

Article abstract of DOI:10.1021/jacs.6b11626

Photoswitchable compounds, which can be reversibly switched between two isomers by light, continue to attract significant attention for a wide array of applications. Azoheteroarenes represent a relatively new but understudied type of photoswitch, where one of the aryl rings from the conventional azobenzene class has been replaced with a fivemembered heteroaromatic ring. Initial studies have suggested the azoheteroarenes-the arylazopyrazoles in particular- to have excellent photoswitching properties (quantitative switching and long Z isomer half-life). Here we present a systematic computational and experimental study to elucidate the origin of the long thermal half-lives and excellent addressability of the arylazopyrazoles, and apply this understanding to determine important structure-property relationships for a wide array of comparable azoheteroaryl photoswitches. We identify compounds with Z isomer half-lives ranging from seconds to hours, to days and to years, and variable absorption characteristics, all through tuning of the heteraromatic ring. Conformation perhaps plays the largest role in determining such properties: the compounds with the longest isomerization half-lives adopt a T-shaped ground state Z isomer conformation and proceed through a T-shaped isomerization pathway, whereas the most complete photoswitching is achieved for compounds that have a twisted (rather than T-shaped) Z isomer conformation. By balancing these factors, we report a new azopyrazole 3pzH, which can be quantitatively switched to its Z isomer (<98%) with 355 nm irradiation, near-quantitatively (97%) switched back to the E isomer with 532 nm irradiation, and has a very long half-life for thermal isomerization (t_1/2 = 74 d at 25 °C). Given the large tunability of their properties, the predictive nature of their performance, and the other functional opportunities afforded by usage of a heteroaromatic system, we believe the azoheteroaryl photoswitches to have huge potential in a wide range of optically addressable applications.

Full text of DOI:10.1021/jacs.6b11626

Subscriber access provided by GAZI UNIV
Article
Tuning Azoheteroarene Photoswitch Performance through Heteroaryl Design.
Joaquín Calbo, Claire E. Weston, Andrew White, Henry Rzepa, Julia Contreras-García, and Matthew J Fuchter
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 23 Dec 2016
Downloaded from http://pubs.acs.org on December 24, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 29
Journal of the American Chemical Society
1
2
3
4
5
Tuning Azoheteroarene Photoswitch Performance through
Heteroaryl Design.
6
7
8
9
Joaquín Calbo†,a Claire E. Weston†,b Andrew J. P. White,b Henry S. Rzepa,b Julia
Contreras-García*,c,d Matthew J. Fuchter*,b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a Instituto de Ciencia Molecular, Universidad de Valencia, 46890 Paterna, Spain
b Department of Chemistry, Imperial College London, London SW7 2AZ, United Kingdom
c Sorbonne Universités, UPMC Univ Paris 06, UMR 7616, Laboratoire de Chimie Théorique, F-75005, Paris, France
d CNRS, UMR 7616, LCT, F-75005, Paris, France
† Both authors contributed equally to this work
* contrera@lct.jussieu.fr, m.fuchter@imperial.ac.uk
Abstract
Photoswitchable compounds, which can be reversibly switched between two isomers by light, continue to
attract significant attention for a wide array of applications. Azoheteroarenes represent a relatively new but
understudied type of photoswitch, where one of the aryl rings from the conventional azobenzene class has
been replaced with a five-membered heteroaromatic ring. Initial studies have suggested the azoheteroarenes –
the arylazopyrazoles in particular – to have excellent photoswitching properties (quantitative switching and
long Z isomer half-life). Here we present a systematic computational and experimental study to elucidate the
origin of the long thermal half-lives and excellent addressability of the arylazopyrazoles, and apply this
understanding to determine important structure-property relationships for a wide array of comparable
azoheteroaryl photoswitches. We identify compounds with Z isomer half-lives ranging from seconds to hours,
to days and to years, and variable absorption characteristics; all through tuning of the heteraromatic ring.
Conformation perhaps plays the largest role in determining such properties; where the compounds with the
longest isomerization half-lives adopt a T-shaped ground state Z isomer conformation and proceed through a T-
shaped isomerization pathway, whereas the most complete photoswitching is achieved for compounds that
have a twisted (rather than T-shaped) Z-isomer conformation. By balancing these factors, we report a new
azopyrazole 3pzH, which can be quantitatively switched to its Z-isomer (>98%) with 355 nm irradiation, near-
quantitatively (97%) switched back to the E isomer with 532 nm irradiation, and has a very long half-life for
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 29
1
2
3
4
5
6
7
8
thermal isomerisation (t1/2 = 74 d at 25 °C). Given the large tunability of their properties, the predictive nature
of their performance, and the other functional opportunities afforded by usage of a heteroaromatic system, we
believe the azoheteroaryl photoswitches to have huge potential in a wide range of optically addressable
applications.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Introduction
Photoswitchable compounds, which can be reversibly switched between two isomers by light, continue to
attract significant attention for a wide array of applications: from molecular motors, memory, and
manipulators to solar thermal storage.1 While there are several molecular properties that impact upon
photoswitch performance, two are of central importance: (1) the completeness of photoswitching at a given
wavelength of light, and (2) the thermal stability of each switched state. Among the classes of photochromic
compounds studied, the azobenzenes maintain a privileged position due to the ease of their synthesis, their
excellent photoswitching properties (such as high extinction co-efficient and quantum yields), and the fact that
their isomerization results in a significant change in molecular shape.2
In general, a significant amount is known on how the chemical structure of photochromic compounds impacts their
photochemical and thermal behavior.1e,2-3 In the case of azobenzenes, excellent progress has been made in tuning
the properties of azobenzenes via the substituents on the aromatic ring(s). For example, the past few years
have seen the successful design of azobenzene photoswitches that can be addressed with visible light,4 a
property of particular importance for the rapidly growing field of photopharmacology.5 Tetra-ortho-substituted
azobenzenes have shown high promise in this regard, such as the o-methoxy analogues reported by Woolley4a,6
and the o-fluoro compounds reported by Hecht.4b,7 Both can be switched with visible light due to separation of
the n-π* transitions of the E and Z isomers, which allows the E isomer to be switched at the long wavelength n-
π* absorbance, rather than via the more usual π-π* transition (which generally requires UVA irradiation). In the
case of the o-methoxyazobenzenes, separation of n-π* transitions is caused by a red-shift in the E isomer
transition due to unfavorable electronic interactions between the lone pairs of the methoxy and azo moieties,
which raises the n-orbital energy.2b Meanwhile for the fluoroazobenzenes, the separation of the n-π* transitions is
caused by a blue-shift in the Z isomer transition due to a lowering of the n-orbital energy by the electron
withdrawing fluorine substituents.4b,8
Increasing the thermal half-life of azobenzene Z isomers, while retaining good addressability9, has arguably
been the most challenging endeavor for this class of photoswitch. Apart from the cyclic azoaryl switches,4d,10
ACS Paragon Plus Environment

Page 3 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
the Z isomers of azobenzenes are generally metastable and thermally revert back to the E isomer over time.
Once again, the tetra-ortho-substituted azobenzenes have provided the optimum solution to this problem,
with the ortho-fluoroazobenzenes reported to have (in addition to excellent addressability) Z isomer thermal
half-lives of up to ca. 700 days (25 °C in DMSO).4b,7 This property was rationalized computationally as being
partially caused by the increased stability of the Z isomer, relative to the E isomer (vs azobenzene), and
partially due to a contribution of Z isomer versus TS dipole moments.4b,7
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We have recently reported an alternative approach to access azo photoswitches with excellent properties
(quantitative switching and long Z isomer half-life) through replacement of one of the azobenzene aryl rings
with a five-membered heteroaromatic ring.11 Such heteroaryl azo compounds are far less well studied12 than
their ubiquitous azobenzene counterparts. Specifically, we found arylazopyrazoles (4pzH and 4pzMe, Figure 1)
to have excellent potential against the state of the art, as they could be quantitatively switched in one or both
directions, depending on “ortho”-substitution,13 and had half-lives of 10 and 1000 days at 25 °C. Given this
performance, the arylazopyrazoles are beginning to be employed in applications including in supramolecular
complexes14 and as photopharmacological agents.15 The performance of the azoheteroaryl photoswitches is
very dependent on the nature of the heteroaromatic ring however, with the arylazopyrroles found to have
shorter thermal half-lives (21 s and 12 h16 for 2pyMe and 2pyH respectively, Figure 1) than the
arylazopyrazoles;11 some other heteroaryl azo compounds were found to have Z isomer half-lives of
microseconds to nanoseconds.12c Clearly, a better understanding of how the photoswitching properties and
performance can be tuned by a judicious choice of the type, positioning and substitution of the heteroaromatic
ring would enable the azoheteroaryl photoswitches to reach their full potential. We note that such design
‘rules’ should not only allow for optimization of intrinsic photoswitch performance, but also for the realization
of designs where the heteroaromatic ring provides other functional opportunities.17
There have been a large number of papers that use computational methods (see review by Burdette et al2a) to
study the photochemical isomerization mechanisms of azobenzenes, although relatively few of these concern
Z-E thermal isomerization.18 Recently, Wegner and co-workers carried out experimental and computational
studies into the thermal half-lives of sterically congested azobenzenes, and discovered that van der Waals
forces play an important role.18a Favorable London dispersion interactions between ortho-alkyl substituents in
the Z isomers are responsible for an increase in Z isomer stability upon increasing the bulkiness of the
substituents. Comparatively, the azoheteroaryl photoswitches are almost unstudied in this regard.11,19 Only
very recently, the Z-E photoisomerization mechanism of two arylazopyrazoles has been computationally
investigated.20 Notably, one of these compounds was found to exhibit near stereospecific excited-state
relaxation.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 29
1
2
3
4
5
6
7
8
9
Herein we report a combined computational and experimental study on novel azoheteroaryl photoswitches
derived from (nitrogen-based) five-membered heteroaromatic rings. We elucidate the origin of the long
thermal half-lives and excellent addressability of the arylazopyrazoles, and apply this understanding to a wide
array of comparable azoheteroaryl photoswitches. As such, we determine important structure-property
relationships as a function of the type of heteroaromatic ring, the connectivity of the ring (relative to the azo
group), and the ring substitution. Furthermore, we identify that calculation of the azo N=N Wiberg Index
provides a quick and simple means to approximate the thermal half-live of this class of photoswitch, which can
vary from seconds-days-months-years. This allows for rapid assessment of novel photoswitches based on
ground state properties, importantly not relying on time-consuming transition state or excited state
calculations. We believe these studies will greatly facilitate the further development of the azoheteroaryl
photoswitches and their application to a broad array of functional areas.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Results and Discussion
Thermal half-lives of the Z-azoheteroaryl photoswitches
Elucidation of the factors underpinning the Z-isomer thermal half-life for the pyrrole and pyrazole-based
photoswitches
We commenced our studies using the previously described11 pyrrole and pyrazole photoswitches 2pyH, 2pyMe,
4pzH and 4pzMe (see Figure 1, and Table 1 for half-lives) as a small set of representative nitrogen-based
azoheteroaryl photoswitches. Computational investigations were carried out in order to rationalize the structure-
property relationships that had been observed experimentally, and to enable the discovery of other azoheteroaryl
compounds with excellent photoswitching performance. Several conformations are available for such compounds,
through rotation of the aromatic rings relative to the azo bridge. All the possible conformers of the E and Z isomeric
forms were fully optimized using the hybrid exchange-correlation PBE0 functional21 including the Grimme’s
dispersion correction in its latest version (D3)22 and using the split-valence Pople’s 6-31G** basis set.23
ACS Paragon Plus Environment

Page 5 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
New photoswitches
Literature photoswitches
Fuchter et al 20
3-pyrazoles
N
4-pyrazoles
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
3pzH
3pzMe
4pzH
4pzMe
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
2-pyrroles
5-pyrazoles
N
N
N
N
N
N
N
N
N
N
N
N
N
5pzH
5pzMe
2pyH
2pyMe
3-pyrroles
23
imidazoles
Herges et al
N
N
N
N
N
N
N
N
N
N
N
N
N
N
3pyH
3pyMe
4im
5im
Otsuki et al 22
triazole and tetrazole
N
N
N
N
N
N
= "ortho"-CH3
N
N
N
N
N
N
N
N
N
2im
tet
3tri
Figure 1. Azoheteroaryl photoswitches considered in this study.
Consistent with our previous study (Figure 1),11 compounds such as 2pyH and 4pzH, which contain heteroaromatic
rings that are not substituted in both the ortho positions, exhibit a Z isomer conformation with a perpendicular T
shape (Figure 2a). Such a T-shaped conformation is only accessible for azo compounds with five-membered
heteroaromatic rings, and not for conventional azobenzenes.2a,12b The introduction of bulky groups in both ortho-
positions of the heteroaromatic ring (such as in 2pyMe and 4pzMe) disfavors the T-shaped conformation and
instead a twisted conformation is predicted, which has been crystallographically characterized for 4pzMe.11
Compounds exhibiting twisted Z isomer conformations are observed experimentally to have shorter Z isomer half-
lives relative to their less-hindered T-shaped analogues (Table 1).11 By introducing only one methyl group into the
ortho position, such as is the case for 2pyH, the two possible conformers can coexist (Figure 2a), and so both T-
shaped and twisted conformers need to be taken into account to provide accurate theoretical half-life times (vide
infra).
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 29
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. a) The two possible Z conformations when there is both a methyl group (bonded in this case to a nitrogen atom) and a
hydrogen atom in the ortho positions, such as in 2pyH. b) The corresponding NCI surfaces: green areas indicate weak non-covalent
interactions whereas narrow red regions designate steric ring clashes.
From analysis of the relative enthalpies (ΔHZ-E) between the E and Z minimum-energy conformations, the Z isomer
was found to be more than 10 kcal/mol less stable than the corresponding E isomer for all the compounds examined
(see Table S1 in the Supplementary Information – SI). This difference increases upon introduction of two ortho
methyls onto the heteroarene to ca. 13 kcal/mol, which provides a thermodynamic driving force for accelerated
isomerization. Given this, it was postulated that the increase in Z isomer energy for 2pyMe and 4pzMe was due to a
destabilization of the Z isomer ground state by the (sterically induced) twisted conformation. However, a closer
inspection of the computed conformers reveals a subtle interplay of interactions, as can be visualized using a non-
covalent index (NCI) analysis.24 Z isomer stabilization of derivatives containing at least one ortho-hydrogen on the
heteroarene (2pyH and 4pzH) can occur through interactions present in the T-shaped conformation, which are
absent in the bis-o-methylated compounds (2pyMe and 4pzMe). A localized surface arises between the ortho-
hydrogen and the benzene ring (Figure 2b), which indicates the presence of a favorable C-H···π interaction. As a
reference, the interaction energy for a T-shaped benzene dimer at the PBE0-D3/6-31G** level is calculated to be
2.90 kcal/mol. However, the twisted conformations of the bis-o-methylated compounds (2pyMe and 4pzMe) exhibit
a greater attractive NCI surface as a result of a larger dispersive contribution, despite the increased steric repulsion
experienced (Figure 2b). When computed through Grimme’s D3 approach,22 the dispersive contribution to the
stability of the T-shaped Z isomer compounds 2pyH and 4pzH is ca. 8 kcal/mol, whereas it is ca. 11 kcal/mol in the
case of the twisted methylated compounds 2pyMe and 4pzMe (see Table S1). For highly substituted azobenzenes,18a
the dispersive contribution has been reported to overcome the steric repulsion and stabilize the Z isomer, increasing
the isomerization half-life. Similarly, in the azoheteroaryl compounds, the computational investigation of
heteroarene substituents more bulky than a methyl group was found to result in stabilizing dispersive interactions in
ACS Paragon Plus Environment

Page 7 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
the Z isomer that surpass any steric congestion (see Table S2 and Figure S24 in the SI). Thus there appears to be a
compensation principle in effect where the introduction of substituents destabilizes the structure through steric
clashes, but the substituted ring adopts a conformation that maximizes mutual stabilizing dispersive interactions.
Despite this, the destabilizing steric effect appears to dominate for the systems under primary study here (ortho H vs
Me), with the (twisted) substituted heteroarene Z isomers (2pyMe and 4pzMe) being higher in energy than
comparable compounds with T-shaped Z isomer conformers (2pyH and 4pzH).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The computed conformation-induced differences in the relative energies between E and Z isomers do not however
correlate with the half-live trends observed experimentally (Table 1). For example, calculations of the free energy for
Z-E isomerization indicate that the Z isomer of 4pzMe is 15.8 kcal/mol less stable than the E isomer, whereas in 2pyH
the difference is slightly smaller (11.7 kcal/mol); however, the half-lives of the Z isomers of these compounds follow
the opposite trend: 10 days and 12 hours, respectively. It would therefore appear that there is a significant kinetic
component to the thermal isomerization rates observed.
Table 1. Theoretical thermodynamic and kinetic parameters including the free energy for the Z-E isomerization reaction (ΔGZ-E, kcal/mol)
and the free energy barrier (ΔG‡, kcal/mol) for the lowest energy conformer of each Z-isomer, and the weighted average half-life times
(WA t1/2) for all four possible TSs compared with the reported experimental values.
ΔGZ-E
-11.7
-14.8
-11.5
-15.8
ΔG‡ a
WA t1/2
1.2 h
120 s
108 days
1 day
Experimental t1/2
12 h c
2pyH
2pyMe
4pzH
22.2 (II)
19.8 (II)
26.4 (I)
23.7 (I)
21 s b
1000 days b
10 days b
4pzMe
The type of the minimum-energy TS is indicated within parenthesis. Data taken from Fuchter et al.11 c Lit. value of 3 h reported by
a
b
Fuchter et al11 was measured at ~10-2 M (see further discussion below).
Given this, transition states (TS) and activation barriers (ΔG‡) were computed for the different Z-E isomerization
processes at the DFT PBE0-D3/6-31G** level of theory. Two principle mechanistic pathways for azobenzene
isomerization have been previously reported – rotation about the N=N bridge and inversion at one of the azo
nitrogen atoms – however, variations on these have also been suggested.2a In the case of azoheteroaryl
photoswitches, the inversion mechanism was found to be the lowest energy pathway for the thermal isomerisation
of all the compounds analyzed (see Figures S25 and S26 for further details). Inversion could occur either at the azo
nitrogen bearing the phenyl ring (type-I, Figure 3a) or at the azo nitrogen bearing the heteroaryl ring (type-II).
Moreover, and in accordance with the conformational possibilities afforded by these novel photoswitches, the type-I
TS is reminiscent of the Z isomer ground state conformation, where the heteroaryl ring is (almost) perpendicular to
the benzene ring, whereas the type-II TS is closer to the E isomer conformation, where the two aromatic rings are
(almost) coplanar. Since there are two possible conformations for the inversion transition state when the system is
non symmetric on either side of the azo group (Figure 3a), there are four possible transition states to be considered.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 29
1
2
3
4
5
6
7
8
9
Analysis of all these TSs yielded, in general, similar energetic barriers for the four possible pathways. Changing the
computational method, the basis set, or moving from the gas phase to a continuum solvent model for acetonitrile all
had negligible effect on the computed relative energy barriers (Table S3). All the pathways were therefore weight-
averaged to obtain mean half-lives according to a Boltzmann distribution of the TS energy barriers, assuming an
equally-balanced ratio of the two Z conformers.25 Theoretical evaluation of the free energy was carried out in gas
phase, including thermal corrections to the total partition function at 298.15 K. Using transition state theory,26 the
rate constants for isomerization were calculated, assuming a first-order reaction, and from this we computed the
final weighted-average half-lives (WA t1/2 in Table 1). Despite the reasonably small energy differences (in the range
of few kcal/mol) calculated for the transition state barriers of the different compounds, large differences in the half-
lives for the photoswitches are predicted, which show good agreement with the available experimental data (Table
1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. a) The two lowest energy Z-E thermal isomerization pathways of the representative photoswitch 2pyH; b) NCI surfaces
calculated from the PBE0-D3/6-31G** electron density for the TS of compounds 2pyH, 2pyMe, 4pzH and 4pzMe that lead to the lowest
free energy barriers. The TS type (either I or II) is indicated within parenthesis.
ACS Paragon Plus Environment

Page 9 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
NCI analysis was employed to disentangle the effects underpinning this energetic ordering. Pyrroles 2pyH and
2pyMe exhibit the lowest TS energy barriers and a TS similar to the E isomer conformation, where the two aromatic
rings attempt to adopt a coplanar relationship (type-II). The presence of the ortho methyl group in 2pyMe results in
more significant twisting away from such co-planarity relative to 2pyH. The NCI surfaces of the pyrrole TS (Figure 3b)
show a number of stabilizing dispersive interactions. For example, when quantified through Grimme’s approach, we
find the stabilizing dispersive contribution to compound 2pyMe to be 10.2 kcal/mol, compared with a smaller value
of 7.1 kcal/mol for 2pyH. The lowest-energy TS for pyrazoles 4pzH and 4pzMe has a T-shaped conformation, similar
to the Z isomer ground state conformation, where the heteroaryl ring is perpendicular to the phenyl ring (type-I).
Interestingly, NCI analysis indicates dispersion is almost absent in the (lowest-energy) TS of 4pzH (Figure 3b); the
dispersive contribution is calculated to be only 6.2 kcal/mol (see Table S1). As in the case of the ground state
conformations, methyl substitution at the ortho position leads to greater dispersive effects. These results suggest
that the TSs are stabilized by non-covalent interactions, mainly dispersion, and such interactions determine the TS
energetics: compound 2pyMe has the highest stabilization through dispersion and therefore the lowest energy
barrier, whereas compound 4pzH has the lowest dispersion stabilization and therefore the highest TS energy.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In summary, the isomerization rate (and therefore the half-lives) of photoswitches 2pyH, 2pyMe, 4pzH and 4pzMe
are controlled by both the Z isomer ground state stabilization and the TS energy for a given isomerization process. In
general, the pyrazoles appear to have much longer isomerization half-lives than the pyrroles due to their preference
to adopt T-shaped (type-I) TSs, which lack significant dispersive stabilization and hence have a higher TS energy.
Methyl substitution in the ortho position twists the ground state Z isomer conformations of both the pyrroles and
pyrazoles away from the T-shape which, through steric effects, destabilizes the ground state Z isomer of 2pyMe and
4pzMe relative to 2pyH and 4pzH, respectively. The compound with the longest thermal Z isomer half-life, 4pzH
(1000 days), exhibits both a T-shaped ground state Z isomer conformation and proceeds through a T-shaped
isomerization pathway. The role of the T-shape conformation in these processes is therefore crucial.
Synthesis and characterization of new azoheteroaryl photoswitches
In order to elucidate structure-property relationships for a wider array of azoheteroaryl photoswitch systems and
provide predictive trends to enable future azoheteroaryl photoswitch design, it was necessary to synthesize and
characterize a larger number of analogues (Figure 1). These were considered in comparison with 2pyH, 2pyMe, 4pzH
and 4pzMe, and with the phenylazoimidazole photoswitches previously characterized in the literature.12a,12b The
principle variables considered when selecting novel candidate compounds for synthesis were: i) the number and
positioning of the nitrogen atoms (and therefore formally the heteroaryl ring present); and ii) the number and
positioning of heteroaryl methyl substituents. The new analogues (Figure 1) were synthesized using similar routes to
those previously described (see SI).11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 29
1
2
3
4
Table 2. Experimental Z-E thermal isomerisation kinetics at 25 °C, including the re-measured values for literature compounds
2pyH, 4im and 5im.
5
6
7
Position of azo
on heteroarene
Rate (k) / s-1
Half-life (t1/2)
8
9
tetrazole
triazole
4
tet a
2.1 × 10-7
3.8 × 10-7
39 d
21 d
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
3
3tri a
3pzH a
1.1 × 10-7
1.5 × 10-6
74 d
3
3pzMe b
5.8 d
4
4
4pzH a,e
7.6 × 10-9
7.7 × 10-7
1000 d
10 d
pyrazoles
4pzMe b,e
5
5
2
2
5pzH b
7.4 × 10-7
1.6× 10-5
6.5 × 10-5
3.3 × 10-2
11 d
12 h
12 h
21 s
5pzMe c
2pyH c,f
2pyMe c,d,e
pyrroles
3
3
3pyH c,d
2.2 × 10-6
7.6 × 10-4
3.7 d
3pyMe c,d
0.25 h
2
4
5
2im g
2.2 x 10-5
3.3 × 10-6
1.2 × 10-6
9 h
4im c,h
5im c,i
2.4 d
6.5 d
imidazoles
a
b
c
Measured using NMR at a range of elevated temperatures, ~10-2 M in d6-DMSO; NMR, ~10-2 M in d3-MeCN; UV-vis, 10-5 M in
d
e
f
anhydrous MeCN. Rates obtained were dependent on the water content of solvent. Reported by Fuchter et al11. Lit. value of 3 h
reported by Fuchter et al11 was measured at ~10-2 M. Reported by Otsuki et al12a
.
A thermal half-life of 5 hh and 22 di (solvent or
g
h,i
concentration not known) was reported by Herges et al12b
.
Thermal isomerization rates and half-lives were measured for the Z isomers of the newly synthesized compounds
(Table 2, which includes the literature compounds for comparison). Following E-Z switching at 355 nm (vide infra),
ACS Paragon Plus Environment

Page 11 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
rates at 25 °C were obtained using either UV/vis or NMR spectroscopy (Figures S15-18): The kinetics of analogues
with half-lives of less than 1 day (or those showing a concentration-dependence – vide infra) were measured using
UV/vis spectroscopy at a concentration of 1 x 10-5 M in acetonitrile. Otherwise, kinetics were more conveniently
measured by NMR spectroscopy at ~10-2 M; either at 25 °C in d3-acetonitrile or, for analogues with half-lives of
longer than 10 days at 25 °C, (for example tet, 3tri, 3pzH), at a range of higher temperatures (70 – 110 °C) in DMSO-
d6. For the latter cases, ΔH‡ and ΔS‡ values were extracted using Eyring plots, and from these, the isomerization
rates and half-lives at 25 °C were calculated (see SI).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Clear experimental trends are apparent from the data summarized in Table 2. When comparing compounds with the
same azo linkage position on a given heteroarene series, the bis-ortho-methylated photoswitches had faster thermal
isomerisation rates than their singly or non-methylated analogues, consistent with the analysis above. The position
of the ring nitrogen atom(s) also had a significant effect on the rate. By separating the photoswitches into those with
≤ 1 ortho-hydrogens and those with two ortho-hydrogens, the 2-azopyrroles had shorter half-lives than the
corresponding 3-azopyrroles (t1/2: 2pyMe < 3pyMe, 2pyH < 3pyH), and for the azopyrazoles, the 5-analogues had
the shortest half-lives and the 4-analogues had the longest half-lives (t1/2: 5pzH < 3pzH < 4pzH, 5pzMe < 3pzMe <
4pzMe). It should be noted that although 3pzMe has been included in the pzMe series, it differs from the other
compounds (5pzMe and 4pzMe) in that it has only one ortho-methyl and a basic nitrogen in the other ortho-position
(as opposed to two ortho-methyls). The finding that this compound fits into the pzMe series suggests that ortho
basic nitrogens have a similar effect on the thermal isomerisation rate as o-methylation. Consistent with this, the
presence of a nitrogen atom in the ortho position was computationally found to disfavor the T-shaped
conformation, and results in a twisted conformation of the Z isomer (Figure 4). For the imidazole compounds, 4im
and 5im both have one ortho-hydrogen so can be considered as part of the same series (t1/2: 5im > 4im) due to the
formation of a T-shaped conformation (as discussed by Herges and co-workers).12b 2im however has one ortho-
methyl and one ortho basic nitrogen so adopts a twisted conformation.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 29
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Characteristic dihedral angles of the lowest energy conformations of representative Z isomer compounds with basic nitrogen
atoms in the ortho position of the heteroarene (bottom) compared with the long-living T-shaped Z isomer of photoswitch 4pzH (top).
More extensive structure property relationships: Computational vs experimental trends
An analogous computational treatment to that described for compounds 2pyH, 2pyMe, 4pzH and 4pzMe was
performed on all the newly synthesized compounds, as well as the other previously reported azoheteroaryl
photoswitches in the literature. Comparison of the Z-E isomerization half-lives obtained experimentally to those
computed (taking into account both thermodynamic and kinetic factors) is shown in Figure 5. In general, the
calculated half-lives were good predictions of the experimental values (i.e. within the same order of magnitude), and
the majority of half-lives were predicted in the correct order, shortest to longest (Figure 5). The comparison of the
theoretical vs experimental half-lives for this larger array of compounds confirms the physical effects (outlined
above) that underpin the thermal half-life of these azoheteroaryl photoswitches: compounds that adopt a more T-
shaped conformation, such as 3pyH and 5pzH, had slower thermal isomerisation rates than their twisted analogues
(3pyMe and 5pzMe). In addition to this, the effect seen upon changing the position of the azo linkage on the
heteroarene was correctly predicted for the pyrroles and pyrazoles, with the sole exception of 3pzMe.
Experimentally, the order of increasing half-life for this series was 5pzMe < 3pzMe < 4pzMe, whereas the calculated
half-life for 3pzMe was longer than 4pzMe. As discussed above, this compound differs from the other azopyrazoles
studied as it has one ortho-methyl and one ortho basic nitrogen, rather than two o-methyls or one o-methyl and one
ortho CH group. The azo-positional effect can be rationalized by the extent of conjugation between the heteroarene
and the phenylazo moiety (vide infra).
ACS Paragon Plus Environment

Page 13 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5. Comparison between the theoretical vs experimental half-life times for the list of photoswitches under study.
In our initial comparison of the experimental and computed half-lives, there were two compounds (4im and 5im) –
previously reported by Herges and co-workers12b – for which the predicted values were observed to be notably
more divergent than the experimental values obtained from the literature. Since there was no apparent reason why
these compounds in particular should be outliers, combined with the fact that the concentrations or solvents in
which half-lives were measured were not reported, we decided to synthesize 4im and 5im and re-measure their
thermal half-lives. The compounds were synthesized following the literature method reported by Herges.12b
Confirmation of the structures of isomeric compounds 4im and 5im, including unambiguous assignment of 4im by X-
ray crystallography (see Figure S2 in the SI), was consistent with the previous structural assignment.12b An interesting
effect was observed however when attempts were made to determine the thermal isomerization rates of 4im and
5im: the measured rate of 4im was concentration-dependent, with faster rates observed at high concentrations
(>>10-4 M) (Table 2). Although the origin of this effect is unclear, it seems that there may be bimolecular pathways at
higher concentration that facilitate isomerization for this compound. At concentrations below 10-3 M, no
concentration dependence on the half-life was observed. 5im did not show a concentration dependence; however,
under the conditions of this study (10-5 M concentration in acetonitrile, 25 °C) a shorter half-live was obtained than
that previously reported.12b The half-lives for 4im and 5im, when measured at the same concentration of the
majority of the other photoswitches included in this study (10-5 M, see Table 2) were 2.4 d and 6.5 d, respectively
(c.f. the reported values of 5 h and 22 d, which were presumably carried out under different conditions). The
inclusion of these revised values into our comparative analysis gave an improved fit (Figure 5).
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In light of the discovery of a concentration-dependence on the thermal half-life, we examined whether other
compounds under study exhibited a comparable effect. To complete the azoimidazole series, 2im was also
synthesized and found to have the same half-life when measured at 10-5 M (by UV/vis) and at 10-2 M (by NMR),
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 29
1
2
3
4
5
6
7
8
which matched the value reported by Otsuki and co-workers (t1/2 of 9 h).12a For the novel photoswitches synthesized
in this study, a representative set27 were examined and it was found that the azopyrroles show a concentration
dependence, whereas the azopyrazoles do not (see Figures 19-23 in the SI for more details). Specifically, the
published half-life of 2pyH was 3 hours when recorded at ~10-2 M and 12 hours when measured at 10-5 M, whereas
it shows negligible concentration dependence at concentrations below ~10-3 M. This observation is crucial when
designing new photoswitches and envisaging their technological use: concentration may be a useful parameter for
tuning switch properties.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
It was also found that the azopyrroles are sensitive to water content in the solvent (as reported previously for
2pyMe28), whereas the other heteroaryl systems were not; where increased water content leads to accelerated
thermal isomerisation. Although the origin of this effect has not been further studied, it is plausible that polar
isomerization transition states are more stabilized in the presence of water, thus leading to faster isomerization. The
+
fact that pyrroles are good electron donors (σAr = -1.96 for N-methyl-2-pyrrolyl)29 into the azophenyl moiety may
lead to the formation of a more polar TS for heteroaryl photoswitches. Theoretical calculations confirmed that the
pyrrole-based transition states have larger electron-donor character than the pyrazole analogues, and that this
difference is more accentuated upon inclusion of water solvent effects (Figure S27).
Rapid approximation of Z-isomer thermal half-life
While in-depth computational study of the thermodynamics and kinetics of the isomerization process provides
insight into the structural-property effects of the azoheteroaryl photoswitches, and allows for assessment of the
magnitude of the Z isomer thermal half-life, such an analysis clearly requires TS structures to be obtained; which is
not practical for large screening of diverse sets of prospective new heteroaromatic photoswitch designs. We
therefore sought to define a predictive metric that would be simpler/quicker to calculate and inform on likely
photoswitch performance. We selected calculation of the N=N bond order (the Wiberg index30 or WI) of the
optimized Z conformers as a potentially useful metric. Bond orders were calculated as the sum of the off-diagonal
square of the density matrix obtained at the PBE0-D3/6-31G** level (see the SI for more details), and averaged for
the two possible Z isomer conformations of each photoswitch. As expected, the WIs obtained show a linear
correlation with respect to the azo N=N bond length (Figure S28a). In all systems studied, the N(azo)-C(benzene)
bond length is computed to be larger than the N(azo)-C(heteroarene) bond length. The difference between these
two magnitudes (NC asymmetry)18f is also predicted to correlate with the N=N bond length (Figure S28b). Since the
N=N bond length highly depends on the extent of π-conjugation in the photoswitch, WI calculations will give
valuable information on the strength of the N=N double bond and, presumably, on the relative half-life times.
Compounds that exhibit a twisted Z isomer conformation, due to ortho-substitution, were found to result in
elongation of both the N-C hindered bond, and the N=N azo bond, with a concomitant decrease in the bond order
ACS Paragon Plus Environment

Page 15 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(smaller Wiberg index). This was particularly noticeable for the bis-o-methylated 2pyMe and 4pzMe when compared
to their less hindered analogues 2pyH and 4pzH (Figure 6). The nature and connectivity of the heteroaromatic ring
was found to impact more noticeably on the Wiberg index calculated. For example, the 2-pyrroles have lower
Wiberg indices (e.g. 2pyH, and 2pyMe) than the 3-pyrroles and all of the pyrazoles. The effect upon changing the
connectivity of the heteroaryl ring (or conversely viewed as changing the azo positioning) can be qualitatively
explained in terms of conjugative electron donation from the N-methyl lone pair to the azophenyl system (Figure 6).
For the 2-azopyrroles and the 5-azopyrazoles, the position of the N-methyl relative to the azo group allows more
extensive conjugation of the lone pair into the azophenyl relative to the other regioisomers. This improved electron
donation from the heteroaryl ring likely leads to a decrease in the bond order through polarization of the N=N bond
(push-pull type effect). Similar electronic effects are apparent upon changing the type of heteroaryl ring: for
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
example, pyrroles (σAr = -1.96, N-methyl-2-pyrrolyl)29 are better electron donors into the azo group than pyrazoles
+
(σAr+ = -0.41, -0.99, -0.29 for N-methyl-3-pyrazolyl, N-methyl-4-pyrazolyl, and N-methyl-5-pyrazolyl respectively),31
and hence have lower Wiberg indices (bond orders) relative to the pyrazoles.
'COMPLETE' HETERARENE CONJUGATION
'PARTIAL' HETEROARENE CONJUGATION
R
R
Ph
Ph
R
N
N
N
N
N
N
R=H 3pyH
WI 1.828
R=H 2pyH
WI 1.791
R=Me 3pyMe WI 1.822
R=Me 2pyMe WI 1.754
2-azopyrrole
3-azopyrrole
R
R
N
N
Ph
N
R
Ph
N
Ph
N
N
N
N
N
N
N
N
R
R=H 4pzH
R=Me 4pzMe WI 1.837
WI 1.849
R=H 5pzH
R=Me 5pzMe WI 1.825
WI 1.841
R=H 3pzH
R=Me 3pzMe WI 1.874
WI 1.877
3-azopyrazole
4-azopyrazole
5-azopyrazole
Figure 6. Conjugation of the heteroarene into the azophenyl moiety.
Comparison of the experimental t1/2 values with the Wiberg indices for all the azoheteroaryl photoswitches studied is
shown in Figure 7. Despite the simplified nature of the Wiberg index to approximate the structural and
conformational effects of this diverse range of compounds, a clear trend appears, although residuals are significant
(R2 = 0.54). When compared with the theoretical half-life times, the correlation is more significant (R2 = 0.82, see
Table S4 and Figure S29). Photoswitches with ‘partial’ heteroarene conjugation (Figure 6), such as 3pzH, 3tri, tet,
etc., are predicted to provide the largest Wiberg indices, correlating well with long experimental half-lives. On the
other hand, ‘complete’ heteroarene conjugated systems, such as 2pyH, 2pyMe, 5pzMe, are calculated with the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 29
1
2
3
4
5
6
7
8
9
lowest Wiberg indices out of the compounds studied, which correlates well with their short half-life values. The
largest outliers in this trend are those most impacted by ortho substitution or lack thereof. For example,
photoswitch 4pzH has a calculated half-life two orders of magnitude larger than expected from its Wiberg index. As
discussed above, its lack of ortho substitution allows for a T-shaped Z isomer ground state and T-shaped TS barrier.
In contrast, photoswitch 3pyMe exhibits a half-life nearly two orders of magnitude shorter than what is expected
from its Wiberg index. In this case, the bis-ortho-substitution more significantly destabilizes the Z isomer with
respect to the transition state, permitting an unexpectedly fast Z-E isomerization (see also compound 2pyMe). Other
key outliers include those with concentration dependent half-lives, such as 2pyH and 5im. We note that when 4pzH,
2pyH, and 5im are removed from the comparison, the trend with respect to experimental t1/2 values vs the Wiberg
index is much improved (R2 = 0.76).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Correlation between the experimental half-lives and the Wiberg index colored by families: bis-ortho-methylation (red), mono-
ortho-methylation (blue), and no ortho-methylation (green). The last family is subdivided in the number of sp2 nitrogen atoms in ortho-
position (light – one atom, medium – zero atoms, and dark – two atoms).
Z-E photoisomerization
The UV/vis absorbances of azoarene E and Z isomers determines their practical utility: their addressability.9
Overlapping absorbances in the characteristic electronic transitions between such isomers leads to incomplete
photoswitching. Therefore, suitable separation of the λmax values for transitions associated with each isomer is highly
desirable for good/quantitative two-way photoswitching at defined wavelengths. The low-energy n-π* band is often
exploited for Z-E photoisomerisation of azobenzenes; however, there is frequently overlap of the absorbances
between the two isomers in this spectral window. Several modified azobenzenes, such as Woolley’s o-methoxy4a,6a
ACS Paragon Plus Environment

Page 17 of 29
Journal of the American Chemical Society
1
2
3
4
5
and Hecht’s o-fluoroazobenzenes,4b,7 have been reported in which the n-π* band of the two isomers are sufficiently
well separated to allow photoswitching for both E-Z and Z-E isomerization using this transition. In particular, several
of the o-fluoroazobenzene analogues developed by Hecht have improved (near quantitative) photoconversion
(≤95% Z and ≤97% E) to those usually seen for azobenzenes.7
6
7
8
9
In our initial study,11 we showed that the nature (pyrrole versus pyrazole) and ring (methyl) substitution of four
azoheteroaryl compounds significantly impacts their absorption spectra and, in turn, their photoswitching behavior.
Most notably, we found that excitation of the π-π* transition of E-4pzMe at 355 nm resulted in complete
photoswitching (>98% Z isomer), whereas irradiation of the tail of the n-π* absorbance of Z-4pzMe switched it back
to >98% E isomer. Thus, 4pzMe could be quantitatively switched in both directions, a result which surpassed all of
the state-of-the-art azo switches.11 While 4pzMe additionally exhibited an impressive thermal half-life for the Z
isomer (10 days), compound 4pzH was found to be the best performing photoswitch of the initial study in terms of Z
isomer half-life (~1000 days, vide supra). Conversely, less efficient (Z to E) switching was possible for 4pzH, due to an
overlap of the Z and E isomer n-π* absorbances. Given the increased understanding developed above for the
features of the azoheteroaryl switches that impact thermal half-life, we sought to develop comparable
understanding of changes in isomer specific transitions, and thus obtain a clear design rationale for how to obtain
heteroazoaryl switches with both excellent two-way photoswitching and Z-isomer thermal half-life.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Frontier molecular orbital analysis
First, we analyzed the energy of the frontier molecular orbitals responsible of the lowest-lying singlet excited states
for the Z and E isomers of the heteroazoaryl photoswitches. Given that two different conformations are possible for
each isomer, the frontier molecular orbitals (FMO) of each isomer were calculated at the PBE0-D3/6-31G** level of
theory for both conformers, in all cases. Figure 8 shows representative HOMO, HOMO-1 and LUMO frontier MOs for
2pyH. Both E isomer conformations are planar, and differ only in the rotational orientation of the heteroaromatic
ring, relative to the azophenyl group. These two conformations have very similar HOMO and LUMO levels and
topologies (see Figure 8), indicating that the relative orientation of the heteroarene, and thus the E isomer
conformation, does not have a dramatic effect on the frontier MOs (Figure S30). The HOMO and LUMO of the E-
isomer are clearly of a π-nature whereas the HOMO-1 is an n-type orbital.32 The conformation of the Z isomer had a
much more dramatic effect on the frontier MOs. While the symmetry of the HOMO and HOMO-1 mostly shows
mixed n/π character for the Z isomer conformers, compounds able to form an orthogonal T-shaped conformation
(Figure 8) result in a more clearly defined n-type HOMO and a π-type HOMO-1 orbital due to symmetry constraints.
The LUMO in all the Z isomers is of a π nature.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 29
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Frontier molecular orbitals of the representative photoswitch 2pyH in the minimum-energy orientations for Z (top) and E
isomers (bottom).
ACS Paragon Plus Environment

Page 19 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Figure 9. Frontier molecular orbital (FMO) diagram for the Z isomers. FMO energies are plotted as green (HOMO) and red (LUMO) lines,
with the computed energy given by the left-hand side axis. HOMO-LUMO energy gaps are displayed as bars, with the computed energy
given by the right-hand side axis. Each compound is associated with two sets of data which correspond to the two possible conformers.
Twisted and T-shaped arrangements are labelled as W and T, respectively.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The significant variation in Z isomer frontier MOs energy levels, as a function of conformation, can clearly be
observed across the series in Figure 9, in contrast to that found for the E isomer (Figure S30). For the Z isomers, the
photoswitches that experience the largest differences in the FMO energy between the two conformers are Z-2pyH,
4im, 5im, 3pyH and 5pzH. These compounds contain one substituted ortho position (either N-Me, C-Me or a basic N
lone pair) and one unsubstituted position (C-H). As such, they can exhibit either a twisted or a T-shaped
conformation (see Figure 8 for 2pyH). In the T-shaped conformation, the n-orbital is stabilized while the π*-orbital
becomes higher in energy, resulting in a considerable increase of the HOMO-LUMO energy gap.
Following this analysis, we sought to correlate the frontier molecular orbitals with their associated electronic
transition energy. The 20 lowest-lying singlet excited states were computed for each conformer under the time-
dependent density functional theory (TDDFT) framework33 at the PBE0/6-31G** level. The reliability of the PBE0
functional was assessed by comparison with other popular hybrid functionals (see Tables S5-S8 in the SI).34 The
frontier molecular orbital gaps were compared to the associated electronic transition energies in order to analyze
the amount of monodeterminantal character of the most important excited states in Z and E isomers (see SI). For
the E isomers, the azoheteroarene photoswitches follow a linear correlation between the S1 energy and the H-1–L
gap (Figure S31a), suggesting that this transition is best described by the one-electron promotion for the H-1 to the
LUMO. The H-1 is the typical n-type lone pair orbital for the azo group (Figure 8). We note that while the ground
state is usually well represented by a single Slater-determinant, the excited state often consists of several
determinants.35 Thus, depending on the mixture of configurations, the size of the two electron-terms and the orbital
energy differences, the H → L transition is not necessarily the lowest in energy.36 The S2 transition for the E isomer,
in turn, corresponds to the H → L electronic transiꢀon, with a linear correlaꢀon between the S2 energy and the H–L
gap (Figure S31b). For the Z-isomers, the lowest-lying S1 excited correlates excellently with the H–L gap in all
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 29
1
2
3
4
5
6
7
compounds (Figure S32a); the correlation being indicative of the monodeterminantal character of this electronic
transition. Conversely, the S2 excited state is calculated as being multiconfigurational, showing a poor correlation
with the H-1 → L gap (Figure S32b). This analysis shows that care should be taken in the analysis of optical
transitions relative to HOMO-LUMO energies.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Experimental photoswitching
In order to provide a larger experimental dataset to compare with this computational analysis, UV/vis spectra and
photoswitching experiments were recorded for each of the new photoswitches in acetonitrile. Since the best
photoconversions were generally achieved with 355 nm (E-Z) and 532 nm (Z-E) irradiation, these wavelengths were
principally used in the switching experiments (see Table 3). Azopyrazole E-5pzH could be quantitatively switched to
its Z isomer using 355 nm irradiation, while the majority of other compounds were switched to photostationary
states (PSSs) containing a mixture of isomers. Z-3pyMe could be near-quantitatively (98%) switched back to its E
isomer using 532 nm, while the majority of other compounds gave PSSs containing a mixture of isomers, due to
overlapping absorbances between E and Z isomers at this wavelength. Notably however, one of the azopyrazoles
newly synthesized for this study - E-3pzH - could be quantitatively switched to its Z-isomer (>98%) with 355 nm
irradiation and near-quantitatively (97%) switched back to the E isomer with 532 nm irradiation. Theoretical
calculations for Z-3pzH predict conformers slightly distorted from a perfectly T-shaped disposition (Figure S33),
resulting in an increased oscillator strength for the n-π* transition compared to, e.g., the analogous 4pzH system
(Table 3), thus allowing improved switching at 532m for Z-3pzH. This, combined with its long half-life for thermal
isomerisation (t1/2 = 74 d at 25 °C), makes it an extremely desirable photoswitch (i.e. comparable photoconversions
to 4pzMe,11 but with increased Z isomer thermal stability). PSS compositions and extracted Z isomer spectra (Figures
S11 and S12), were determined as described in the SI (and as demonstrated in Figures S4 to S10). From this,
absorption maxima (λmax) and molar extinction coefficients (ε) were obtained for the π-π* and n-π* absorbances of
both isomers of each compound (Table 3).
Table 3. Photostationary states and spectral data for the azoheteroarenes.
Photostationary states
E isomer π-π*
E isomer n-π*
Z isomer π-π*
Z isomer n-π*
ACS Paragon Plus Environment

Page 21 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
E-Z (% Z)
Z-E (% E)
λmax
/ nm / M-1 cm-1
10-2 × ε
λmax
10-2 × ε
λmax
10-2 × ε
λmax
/ nm / M-1 cm-1
10-2 × ε
/ nm / M-1 cm-1
/ nm / M-1 cm-1
~290 24.3 ± 2.2 b
272 48.9 ± 5.7 b
272 90.8 ± 11
289 69.4 ± 8.4
286 51.5 ± 6.0 b
296 55.3 ± 6.2 b
tet
76 ± 2
90 ± 2
> 98
72 ± 3
84 ± 3
97 ± 3
79 ± 3
87 ± 3
86 ± 3
310
310
320
341
325
340
167 ± 15
187 ± 22
146 ± 17
162 ± 20
174 ± 20
186 ± 21
442 3.73 ± 0.34
428
10.1 ± 0.92 b
12.7 ± 1.5 b
10.8 ± 1.3
3tri
438
425
5.05 ± 0.59
7.78 ± 0.91
6.76 ± 0.8
434
422
430
438
451
3pzH
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5pzH
> 98
425
8.73 ± 1.1
3pzMe
5pzMe
96 ± 2
94 ± 2
428
6.12 ± 0.72
8.54 ± 0.95
13.6 ± 1.6 b
22.1 ± 2.5 b
435
~410 a
87.5 ± 7.9 c
311
405
462
479
423
14.2 ± 1.3 c
33.2 ± 2.9 c
60.8 ± 4.5 c
15.7 ± 1.3 b
a
3pyH
3pyMe
2pyMed
2pyHd
77 ± 3
85 ± 5
363
345
394
385
193 ± 18
166 ± 15
282 ± 18
221 ± 18
-
~410 a
~430 a
~413 a
52.0 ± 4.6 c
300
a
85 ± 3
98 ± 2
>98
-
85 ± 3 g
70.4 ± 5.7 c
346
a
-
84 ± 2 g
82 ± 3
162 ± 13 b
333
a
-
4pzMed
4pzHd
2ime
>98
>98
>98
335
328
363
336
362
227 ± 18
166 ± 13
170
425
417
~450
-
9.63 ± 0.76
296 57.7 ± 4.6
275 80.6 ± 6.4
441
23.0 ± 1.8
70 ± 3
6.40 ± 0.51
403
6.66 ± 0.53
>98 h
>95 i
98 i
-
-
12.6
-
-
-
-
-
-
-
-
-
-
-
-
4imf
148
-
-
5imf
45 j
258
-
The position of the band (λmax in nm) and the corresponding extinction coefficient (ε in M-1 cm-1) are indicated. See SI for
a
b,c
UV/vis spectra. n-π* absorbance appears as a shoulder on the π-π* absorbance, hence ε cannot be obtained.
These values are for the extracted (b) or estimated (c) pure Z isomer. d Literature values reported by Fuchter et al.11 e
Literature values reported by Otsuki et al.12a Literature values reported by Herges et al.12b All E-Z PSSs were
f
measured at 355 nm irradiation except for the following which were measured at: g 415 nm; h 363 nm; i 365 nm. All
j
Z-E PSS were measured at 532 nm irradiation except for the following which was measured at: 455 nm irradiation.
n−π* electronic transition
TDDFT calculations were used to further rationalize the experimental results obtained, and in light of the analysis
above. Each conformer was treated independently, and the effect of the conformation (T-shaped, twisted, etc.) was
analyzed throughout. In the planar azoheteroarene E isomers, a mirror plane exists through the plane generated by
the azo function (for both conformers), and therefore the n−π* transition becomes symmetry-forbidden at the
equilibrium geometry. Indeed, negligible intensity is computed for the n-π* (H-1 → L) transition of the E isomer
(Table S5). Persico and co-workers37 reported that the intensity of the n-π* absorption band for E-azobenzene
depends on the effect that vibrational motions have on the electronic structure. Therefore, the anharmonicity, the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 29
1
2
3
4
5
6
7
thermal distributions, and the solvent effects all have to be taken into account. Vibrational and rotational motion
leads to a population of molecules away from the equilibrium geometry, which provide weakly allowed n-π*
transitions and correspondingly a small absorbance. This rationale is clearly also relevant for azoheteroarenes, with
a weak n-π* transition observed experimentally for the E isomers of all the compounds studied (Table 3).
8
9
The Z isomer n-π* (H → L) transition, on the other hand, is computed from weak to moderately intense, depending
on the photoswitch and conformation under analysis (Table S6). The most important feature affecting the
magnitude of the n-π* band intensity in the Z isomer is the potential to form a T-shaped conformation. In the
perfect T-shaped conformer, as in the planar E isomer, a mirror plane exists through the plane of the azo function
and, therefore, the transition becomes symmetry-forbidden at the equilibrium geometry. As a result, all the
compounds with the azoheteroaryl group orthogonal to the benzene ring (CCNN dihedral ≈ 90°, T-shaped) are
computed to have a singlet S1 excitation with negligible oscillator strength (Figure 10). Conversely, the
photoswitches that cannot adopt a T-shaped conformation exhibit an S1 state with oscillator strengths up to f =
0.090. The correlation of the CCNN dihedral angle to the computed intensity of the n-π* transition, for all the
photoswitches studied, can be seen in Figure 10. Such conformation-induced tuning of the magnitude of the n-π*
transition intensity is clearly observable from the experimental data. For example, the Z isomers of photoswitches
with two ortho-methyls on the heteroarene (5pzMe, 3pyMe) cannot adopt a T-shaped conformation (vide supra),
and therefore had a considerably larger n-π* absorbance than their non- or singly o-methylated analogues (Table 3).
Likewise, the presence of two heteroaryl basic nitrogens adjacent to the azo group (tet, 3tri), or one basic nitrogen
and one ortho-methyl (3pzMe), increased the intensity of the Z isomer n-π* band relative to that of the E isomer
(Table 3), due to an increasingly twisted conformation. Since the T-shaped conformation is additionally responsible
for an increased H-L band gap (vide supra), compounds exhibiting such a conformation in their Z isomer should
result in a shift of the n-π* band to higher energies. Experimentally this is indeed the case – for example, the λmax for
the n-π* transitions of Z-3pyH and Z-5pzH occur at shorter wavelengths than those of Z-3pyMe and Z-5pzMe,
respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 23 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Figure 10. Correlation between the intensity of the lowest-lying S1 excited state of n-π* nature with respect to the dihedral between the
benzene and the azo group in the Z isomer. The two possible conformers for each isomer are included.
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
An increase in the number of N atoms is predicted to result in a systematic stabilization of the HOMO and LUMO for
the Z isomers, and a global decrease in the HOMO-LUMO gap (Figure 9). When comparing the Z-isomers of 3pzH,
4pzH, 4im, 3tri and tet, there is a general decrease in energy of the n-π* electronic transition upon increasing the
number of N atoms as calculated by TDDFT (Table S6). For example, the n-π* electronic excitation is predicted to
change from 409 nm in 4pzH to 455 nm in 3tri, although little difference is found moving to tet (453 nm). This trend
is supported by the experimental data: λmax = 403, 434, and 428 nm for 4pzH, 3tri and tet, respectively. In terms of
conjugation effects resulting from heteroaryl connectivity, a consistent shift of the n-π* (H → L) transition to longer
wavelengths was predicted for compounds going from ‘partial’ to ‘complete’ heteroarene conjugation (Table S6 and
Figure 6). For instance, the n-π* excitation of ‘partial’-conjugated 3-azopyrroles 3pyH and 3pyMe is calculated at 438
and 473 nm, respectively, which moves to 458 and 490 nm for the ‘complete’-conjugated 2-azopyrroles 2pyH and
2pyMe. This also holds for the ‘partial’-conjugated 3- (3pzH and 3pzMe) and 4-azopyrazoles (4pzH and 4pzMe)
compared to the ‘complete’-conjugated 5-pyrazole compounds (5pzH and 5pzMe). These trends are in excellent
agreement with the experimental position of the n-π* band in the Z isomers gathered in Table 3.
π-π* electronic transition
For the majority of the compounds studied, the π-π* electronic transition of the E isomer was computed to exhibit a
moderately intense absorbance band with oscillator strength f = 0.5 – 0.9. Since the π-π* electronic transition of the
E isomers is assigned to a HOMO to LUMO transition (H → L), the computed H-L band gap trends can be compared
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 29
1
2
3
4
5
6
7
8
9
to the λmax position of the E isomer π-π* transition. When comparing compounds where the N-Me group is located
in the same position relative to the azo linkage (3pzH vs 3tri vs tet), a systematic increase in the energy of the π-π*
transition is computed as the number of nitrogen atoms increases: from 3.99 eV (311 nm) in 3pzH to 4.09 eV (303
nm) in tet. When considering compounds with two nitrogen atoms in the heterocyclic ring, but positioned
differently (4pzH vs 4im vs 3pzH), the π-π* bands of 4pzH and 4im, were predicted to have comparable π-π*
transition energies of 315 and 319 nm (3.94 and 3.89 eV) respectively, both lower than 3pzH, which has a calculated
transition energy of 311 nm (3.99 eV). These trends are supported by the experimental data, although, as for the n-
π* band, little/no difference is observed for 3tri versus tet: λmax = 336, 328, 320, 310, and 310 nm for E-4im, 4pzH,
3pzH, 3tri and tet respectively. The π-π* band of 3pyMe and 3pyH was less red-shifted than for the 2-azopyrrole
analogues (2pyMe and 2pyH); a result of ‘partial’ versus ‘complete’ conjugation of the azo group with the
heteroarene (see Figure 6). For similar reasons, the 5-azopyrazole analogue 5pzH had red-shifted π-π* absorbances
relative to the 3-azopyrazole and 4-azopyrazole analogues (3pzH and 4pzH), while a small red-shift was seen for
bis(o-methylated) 5-azopyrazole E-5pzMe relative to 4-azopyrazole E-4pzMe. The calculated data are in excellent
agreement with the experimental trends for the 3-, 4- and 5-azopyrazoles, whereas the significant conformational-
dependence of the π-π* S2 band position (Table S5) hinders a full comparison between 3- and 2-azopyrroles.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Band separation
Given that the extent of photoswitching in the PSS is primarily a function of the overlap of bands in the absorption
spectrum, and the corresponding intensity of the relevant transitions, it is important to fully understand the
structure-property relationships that underpin band separation. While the discussion above individually compares
the position and intensity of the main n-π* and π-π* transitions as a function of both the azoheteroaryl photoswitch
and isomer dependent conformations, a comparison of the computed versus experimental band separation is
warranted. Such a comparison is shown in Table 4. In general, the computed band separations qualitatively correlate
with those observed experimentally. By comparing compounds with and without ortho-methyl substitution, a
significant enhancement especially of the Z-E photoconversion is achieved when substituting the ortho H by a
methyl group. For instance, % Z-E photoconversion increases from 84, 85, 70 and 79% in 2pyH, 3pyH, 4pzH and
5pzH, respectively, to >98, 98, >98 and 86% for 2pyMe, 3pyMe, 4pzMe and 5pzMe (Table 4). This improvement in
the photoconversion efficiency is a result of n-π* band separation between Z and E isomers, which increases upon
moving from H-substituted to methyl-substituted photoswitches (Table 4), a result of the effect of the increasingly
twisted conformation of the Z-isomer (vide supra). In the case of 4pzH/4pzMe, inversion of the n-π* band positions
between E and Z is found both experimentally and theoretically. Pyrrole heteroaromatic switches show in general
poorer E-Z photoconversion efficiencies with respect to the pyrazole analogues, a result of the difference between
the n-π* band of the Z isomer and the π-π* of E isomer (Table 4). For example, the experimental n-π*(Z) minus π-π*
ACS Paragon Plus Environment

Page 25 of 29
Journal of the American Chemical Society
1
2
(E) difference is recorded to increase from 42 nm in 3pyH to 102 nm in 3pzH, thus explaining the E-Z
3
4
photoconversion improvement from 77 to >98%.
5
6
7
8
9
While intrinsically good band separation enables improved photoswitching, judicious choice of the irradiation
wavelength, particularly to spectral regions significantly outside of the absorption maxima, can allow for further
improvement in the completeness of photoswitching (albeit with a decrease in quantum yield). A key case in point in
this study is photoswitch 3pzH. Using light at 532 nm (the n-π* λmax of Z-3pzH is 422 nm), excellent Z-E
photoisomerisation was achieved for 3pzH (97% E), although a relatively long irradiation time (~15 minutes) was
required to reach the PSS. We tentatively assign this to a difference in oscillator strengths between E and Z isomers
for 3pzH at 532 nm: the Z-3pzH conformation (which is slightly distorted from a perfectly T-shaped disposition -
Figure S33) results in an increased oscillator strength for the n-π* transition at 532 nm compared to the planar E-
3pzH. Thus, once again, the chemical structure and conformation of the potential photoswitch can help us
comprehend and design compounds that can achieve better photoconversions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. Experimental and theoretical separation (values in nm are given as Z minus E) of the n-π* and π-π* bands
for the Z and E isomers.
n-π* separation (Z – E)
experimental theoreticala experimental theoreticala experimental theoreticala
π-π* separation (Z – E)
n-π* (Z) – π-π* (E)
Compound
2pyH
2pyMe
4pzH
4pzMe
4im
10
49
-14
16
--
16
50
-38
15
-15
8
-52
-48
-53
-39
--
-38
-29
-36
-30
-28
-29
-11
-32
-19
-23
-23
-16
-12
-9
38
85
75
106
--
113
135
94
141
127
119
145
90
--
5im
--
--
--
2im
--
22
-9
--
3pyH
3pzH
5pzH
3tri
-5
-52
-48
-52
-38
-20
-44
-46
-39
42
-3
-13
6
102
89
143
135
135
152
150
148
117
5
-4
-25
-28
28
22
-2
124
118
111
115
113
tet
-14
16
51
10
5pzMe
3pyMe
3pzMe
-6
a
Theoretical band separation for the isomers was calculated by taking the simple average band position between
the two conformers for a given isomer.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 29
1
2
3
Conclusions
4
5
6
7
8
9
We present a systematic computational and experimental study on the azoheteroarene photoswitches:
photochromic azoaryl molecules consisting of one five-membered heteroaromatic ring and one benzene ring.
Modification of the type of the heteroaromatic ring, as well as its positioning (relative to the azo group) and
substitution has been found to have a huge effect on the performance of this class of switch. This is most notable in
the Z isomer thermal half-life data, where we identify compounds with half-lives ranging from seconds to hours, to
days and to years. In depth computational study has revealed the thermal Z-E isomerization rate to be determined
by both thermodynamic (Z-E isomer relative stability) and kinetic (activation barrier) factors, which are, in turn,
influenced by the structure and conformation of the photoswitch in question. Conformation perhaps plays the
largest role, where the compounds with the longest isomerization half-lives adopt a T-shaped ground state Z isomer
conformation and proceed through a T-shaped isomerization pathway. Non-covalent interaction analysis can readily
explain the stabilization or destabilization of such conformations.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conversely, the most complete photoswitching is achieved for compounds that have a twisted (rather than T-
shaped) Z-isomer conformation; where such a conformation is achieved through the use of substituents or basic
nitrogen atoms in the ortho positions. Such twisting particularly results in an increase in oscillator strength for the n-
π* band of the Z-isomer and therefore allows for the selection of wavelengths that can selectively address each
isomer. Ultimately, to achieve complete bidirectional photoswitching and a long Z isomer half-life, a balance must be
found between these two conformational states. In this study, such a balance is exemplified by the newly discovered
azopyazole 3pzH, which can be quantitatively switched to its Z isomer (>98%) with 355 nm irradiation, near-
quantitatively (97%) switched back to the E isomer with 532 nm irradiation, and has a very long half-life for thermal
isomerisation (t1/2 = 74 d at 25 °C). Z-3pzH has a conformation only slightly distorted from a perfectly T-shaped
disposition (Figure S33), which retains the slow thermal isomerization of the Z isomer, but results in an increased
oscillator strength for the n-π* transition compared to, e.g., the analogous 4pzH system.11
We also show the Wiberg index to be suitably useful as a predictive measure for estimating thermal half-lives,
allowing for large numbers of compounds to be screened without the need to obtain free energies for the E and Z
isomer conformations and TSs, which have high computational cost. Although outliers are present in our analysis,
which limit quantitative comparison, our data (see Figure 8) clearly shows that assessment of the Wiberg index gives
a useful ‘first approximation’ as to whether a given switch is likely to have a short, medium or long half-life. As such,
this should allow identification of photoswitches that would be suited to particular applications, where either short
or long thermal half-lives may be required.
Given the large tunability of their properties, the predictive nature of their performance, and the other functional
opportunities15,17a afforded by usage of a heteroaromatic system, we believe the azoheteroaryl photoswitches to
ACS Paragon Plus Environment

Page 27 of 29
Journal of the American Chemical Society
1
2
3
4
have huge potential in a wide range of optically addressable applications. The approaches hereby presented provide
a means to rationally and rapidly design novel photoswitches to achieve a given performance requirement.
5
6
7
8
9
Supporting Information Available. Text, tables, figures, and CIF files giving synthetic methods, NMR and
crystallographic data, photochemical methods including additional spectra and kinetics, and computational details.
This material is available free of charge via the Internet at http://pubs.acs.org. Raw data files are openly available at
DOI: 10.6084/m9.figshare.4287302.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Acknowledgements. MJF and CEW thank the Engineering and Physical Sciences Research Council for funding. JC
acknowledges the Spanish Ministry of Education, Culture, and Sport (MECD) for an FPU grant. We also thank Dr
Kuimova (Imperial College London) for equipment access.
References
(1)
a) Gostl, R.; Senf, A.; Hecht, S., Chem. Soc. Rev. 2014, 43, 1982-1996. b) Bléger, D.; Hecht, S., Angew.
Chem. Int. Ed. 2015, 54, 11338-11349.
L., Chem. Rev. 2013, 113, 6114-6178.
c) Szymański, W.; Beierle, J. M.; Kistemaker, H. A. V.; Velema, W. A.; Feringa, B.
d) Feringa, B. L.; Browne, W. R. Molecular Switches; VCH: Weinheim, 2011.
e) Russew, M.-M.; Hecht, S., Adv. Mater. 2010, 22, 3348-3360.
Zheng, J. O.; Nocera, D. G.; Grossman, J. C., Nat Chem 2014, 6, 441-447.
f) Kucharski, T. J.; Ferralis, N.; Kolpak, A. M.;
g) Zhitomirsky, D.; Cho, E.; Grossman, J. C.,
Adv. Energy Mater. 2016, 6, 1502006.
2013, 3, 3260. i) Saydjari, A. K.; Weis, P.; Wu, S., Adv. Energy Mater. 2016, 1601622.
(2) a) Bandara, H. M. D.; Burdette, S. C., Chem. Soc. Rev. 2012, 41, 1809-1825.
G. A., Chem. Soc. Rev. 2011, 40, 4422-4437.
(3) a) Bastianelli, C.; Caia, V.; Cum, G.; Gallo, R.; Mancini, V., J. Chem. Soc., Perkin Trans. 2 1991, 679-683.
b) Irie, M.; Fukaminato, T.; Matsuda, K.; Kobatake, S., Chem. Rev. 2014, 114, 12174-12277. c) Feng, W.; Luo,
d) Luo, W.; Feng, Y.; Cao, C.; Li, M.; Liu, E.; Li, S.; Qin, C.; Hu, W.; Feng,
h) Feng, Y.; Liu, H.; Luo, W.; Liu, E.; Zhao, N.; Yoshino, K.; Feng, W., Sci. Rep.
b) Beharry, A. A.; Woolley,
W.; Feng, Y., Nanoscale 2012, 4, 6118-6134.
W., J. Mater. Chem. A 2015, 3, 11787-11795.
(4)
a) Beharry, A. A.; Sadovski, O.; Woolley, G. A., J. Am. Chem. Soc. 2011, 133, 19684-19687.
b)
Bléger, D.; Schwarz, J.; Brouwer, A. M.; Hecht, S., J. Am. Chem. Soc. 2012, 134, 20597-20600.
c) Yang, Y.; Hughes, R. P.;
Aprahamian, I., J. Am. Chem. Soc. 2012, 134, 15221-15224. d) Siewertsen, R.; Neumann, H.; Buchheim-Stehn, B.; Herges,
R.; Näther, C.; Renth, F.; Temps, F., J. Am. Chem. Soc. 2009, 131, 15594-15595.
(5)
Broichhagen, J.; Frank, J. A.; Trauner, D., Acc. Chem. Res. 2015, 48, 1947-1960.
(6) a) Samanta, S.; Beharry, A. A.; Sadovski, O.; McCormick, T. M.; Babalhavaeji, A.; Tropepe, V.; Woolley, G.
A., J. Am. Chem. Soc. 2013, 135, 9777-9784. b) Samanta, S.; McCormick, T. M.; Schmidt, S. K.; Seferos, D. S.;
Woolley, G. A., Chem. Commun. 2013, 49, 10314-10316.
a) Velema, W. A.; Szymanski, W.; Feringa, B. L., J. Am. Chem. Soc. 2014, 136, 2178-2191.
b)
(7)
Knie, C.; Utecht, M.; Zhao, F.; Kulla, H.; Kovalenko, S.; Brouwer, A. M.; Saalfrank, P.; Hecht, S.; Bléger, D.,
Chem. Eur. J. 2014, 20, 16492-16501.
(8)
Other successful attempts to red shift azo photoswitch absorbance include through BF2-coordination.
See e.g. Refs. 4c and Yang, Y.; Hughes, R. P.; Aprahamian, I., J. Am. Chem. Soc. 2014, 136, 13190-13193.
(9)
(10)
van Dijken, D. J.; Kovaříček, P.; Ihrig, S. P.; Hecht, S., J. Am. Chem. Soc. 2015, 137, 14982-14991.
Hammerich, M.; Schütt, C.; Stähler, C.; Lentes, P.; Röhricht, F.; Höppner, R.; Herges, R., J. Am. Chem. Soc.
2016, 138, 13111-13114.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 29
1
2
(11)
Weston, C. E.; Richardson, R. D.; Haycock, P. R.; White, A. J. P.; Fuchter, M. J., J. Am. Chem. Soc. 2014,
3
136, 11878-11881.
4
(12)
a) Otsuki, J.; Suwa, K.; Narutaki, K.; Sinha, C.; Yoshikawa, I.; Araki, K., J. Phys. Chem. A 2005, 109, 8064-
5
6
7
8
8069. b) Wendler, T.; Schütt, C.; Näther, C.; Herges, R., J. Org. Chem. 2012, 77, 3284-3287.
Lobo, M.; Nonell, S.; Velasco, D., Angew. Chem. Int. Ed. 2012, 51, 12820-12823.
c) Garcia-Amorós, J.; Díaz-
(13)
Substitution at one or both of the positions of the 5-member heteroarene adjacent to the azo group are
referred to as “ortho” throughout. Although this is not formally the correct usage of the term, it allows for simple
comparison with the ortho-substituted azobenzenes.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(14)
Stricker, L.; Fritz, E.-C.; Peterlechner, M.; Doltsinis, N. L.; Ravoo, B. J., J. Am. Chem. Soc. 2016, 138, 4547-
4554.
(15)
Weston, C. E.; Kraemer, A.; Colin, F.; Yildiz, Ö.; Baud, M. G. J.; Meyer-Almes, F.-J.; Fuchter, M. J., ACS
Infect. Dis. 2016, DOI: 10.1021/acsinfecdis.1026b00148.
(16)
(17)
Value reported in this work.
a) Weston, C. E.; Richardson, R. D.; Fuchter, M. J., Chem. Commun. 2016, 52, 4521-4524.
b)
Schütt, C.; Heitmann, G.; Wendler, T.; Krahwinkel, B.; Herges, R., J. Org. Chem. 2016, 81, 1206-1215.
(18) a) Schweighauser, L.; Strauss, M. A.; Bellotto, S.; Wegner, H. A., Angew. Chem. Int. Ed. 2015, 54, 13436-
13439. b) Wang, Y.-P.; Zhang, Z.-X.; Xie, M.; Bai, F.-Q.; Wang, P.-X.; Zhang, H.-X., Dyes Pigments 2016, 129, 100-108.
c) Ikegami, T.; Kurita, N.; Sekino, H.; Ishikawa, Y., J. Phys. Chem. A 2003, 107, 4555-4562. d) Wazzan, N. A.;
e) Dokić, J.; Gothe, M.; Wirth, J.; Peters, M.
Richardson, P. R.; Jones, A. C., Photochem. Photobiol. Sci. 2010, 9, 968-974.
V.; Schwarz, J.; Hecht, S.; Saalfrank, P., J. Phys. Chem. A 2009, 113, 6763-6773. f) Cembran, A.; Bernardi, F.; Garavelli, M.;
Gagliardi, L.; Orlandi, G., J. Am. Chem. Soc. 2004, 126, 3234-3243.
(19)
(20)
Nabati, M., J. Phys. Theor. Chem. IAU Iran, 2016, 12, 325–338.
Wang, Y.-T.; Liu, X.-Y.; Cui, G.; Fang, W.-H.; Thiel, W., Angew. Chem. Int. Ed. 2016, DOI:
10.1002/anie.201607373.
(21)
(22)
(23)
Adamo, C.; Barone, V., J. Chem. Phys. 1999, 110, 6158-6170.
Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H., J. Chem. Phys. 2010, 132, 154104.
Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S.; Defrees, D. J.; Pople, J. A., J. Chem.
Phys. 1982, 77, 3654-3665.
(24)
a) Johnson, E. R.; Keinan, S.; Mori-Sánchez, P.; Contreras-García, J.; Cohen, A. J.; Yang, W., J. Am. Chem.
Soc. 2010, 132, 6498-6506.
D. N.; Yang, W., J. Chem. Theory Comput. 2011, 7, 625-632.
b) Contreras-García, J.; Johnson, E. R.; Keinan, S.; Chaudret, R.; Piquemal, J.-P.; Beratan,
(25)
Assuming a 50% population of each conformer was found to give a comparable correlation between
theoretical and experimental half-life times to a Boltzmann distribution of the two possible Z conformers.
(26)
(27)
Eyring, H., J. Chem. Phys. 1935, 3, 107-115.
Only compounds with half-lives of less than 1 day at rt (2pyH, 3pyH, 5pzMe) were studied in order to
access accurate comparisons in a timely fashion. 2pyMe was not studied due to its very short half-life.
(28)
(29)
(30)
(31)
(32)
See the Supporting Information of Ref. 11.
Hill, E. A.; Gross, M. L.; Stasiewicz, M.; Manion, M., J. Am. Chem. Soc. 1969, 91, 7381-7392.
Wiberg, K. B., Tetrahedron 1968, 24, 1083-1096.
Noyce, D. S.; Sandel, B. B., J. Org. Chem. 1976, 41, 3640-3642.
We note that this analysis contrasts with the n-π* nature of the lowest-lying (HOMO → LUMO) electronic
transition commonly assigned to E isomer azo photoswitches; the π-π* transition being assigned to higher energies.2a The
lowest-lying excitation is not always described by a pure HOMO → LUMO monoexcitaꢀon. We used TDDFT equaꢀons to
find the best linear combination of monoexcitations and the corresponding energies. See Ref. 36.
(33)
a) Jamorski, C.; Casida, M. E.; Salahub, D. R., J. Chem. Phys. 1996, 104, 5134-5147.
b) Casida, M. E.;
Jamorski, C.; Casida, K. C.; Salahub, D. R., J. Chem. Phys. 1998, 108, 4439-4449.
Gross, E. K. U., Phys. Rev. Lett. 1996, 76, 1212-1215.
c) Petersilka, M.; Gossmann, U. J.;
(34)
(35)
Plater, M. J.; Harrison, W. T. A.; Rzepa, H. S., J. Chem. Res. 2015, 39, 711-718.
Szabo, A.; Ostlund, N. S. Modern Quantum Chemistry: Introduction to Advanced Electronic Structure
Theory; Dover Publications, 1989.
(36)
Casida, M. E.; Ipatov, A.; Cordova, F. In Time-Dependent Density Functional Theory; Marques, M. A. L.,
Ullrich, C. A., Nogueira, F., Rubio, A., Burke, K., Gross, E. U. K., Eds.; Springer Berlin Heidelberg: Berlin, Heidelberg, 2006, p
243-257.
(37)
Cusati, T.; Granucci, G.; Persico, M.; Spighi, G., J. Chem. Phys. 2008, 128, 194312.
ACS Paragon Plus Environment

Page 29 of 29
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table of Contents Graphic
ACS Paragon Plus Environment