Effect of Extended Benzannelation Orientation on Bergman and Related Cyclizations of Isomeric Quinoxalenediynes

Article abstract of DOI:10.1021/acs.joc.7b02420

A combined computational and experimental study was conducted to examine the effect of extended benzannelation orientation on C¹-C⁵ and C¹-C⁶ cyclization of acyclic quinoxalenediynes. Calculations (mPW1PW91/cc-p

Full text of DOI:10.1021/acs.joc.7b02420

Subscriber access provided by University of Florida | Smathers Libraries
Article
Effect of Extended Benzannelation Orientation on Bergman
and Related Cyclizations of Isomeric Quinoxalenediynes
Stephanie A. Valenzuela, Alondra J. Cortes, Zakery J. E. Tippins, Morgan
H. Daly, Terell E. Keel, Benjamin F Gherman, and John D. Spence
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.7b02420 • Publication Date (Web): 09 Nov 2017
Downloaded from http://pubs.acs.org on November 10, 2017
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.
The Journal of Organic Chemistry 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 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Effect of Extended Benzannelation Orientation on Bergman and Related Cyclizations of
Isomeric Quinoxalenediynes
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
Stephanie A. Valenzuela, Alondra J. Cortés, Zakery J. E. Tippins, Morgan H. Daly, Terell E.
*
*
Keel, Benjamin F. Gherman and John D. Spence
Department of Chemistry, California State University, Sacramento, 6000 J Street, Sacramento,
California, 95819, United States
*Corresponding authors ghermanb@csus.edu; jspence@csus.edu
Table of Contents/Abstract Graphic
Abstract
A combined computational and experimental study was conducted to examine the effect of
1
5
1
6
extended benzannelation orientation on C –C and C –C cyclization of acyclic
quinoxalenediynes. Calculations (mPW1PW91/ccꢀpVTZ // mPW1PW91/6ꢀ31G(d,p)) on
terminal and phenylethynylꢀsubstituted 5,6ꢀdiethynylquinoxaline and 6,7ꢀdiethynylquinoxaline
1
6
showed C –C Bergman cyclization as the favored thermodynamic reaction pathway, with larger
1
6
C –C preference for the angular quinoxalenediynes due to gain of a new aromatic sextet.
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Kinetic studies, as a function of 1,4ꢀcyclohexadiene concentration, revealed retroꢀBergman ring
opening predominates over hydrogen atom abstraction (k > k ) for 6,7ꢀdiethynylquinoxaline
ꢀ1
2
while 5,6ꢀdiethynylquinoxaline undergoes irreversible Bergman cyclization indicative of a large
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
retroꢀBergman ring opening barrier (k > k ). The effect of extended linear versus angular
2
ꢀ1
benzannelation on reaction pathway shows in the contrasting photocyclizations of phenylethynyl
1
6
derivatives. While angular 5,6ꢀdiethynylquinoxalines gave exclusive C –C photocyclization,
1
5
linear 6,7ꢀdiethynylquinoxaline afforded C –C fulvene products. Computed singlet–triplet gaps
and biradical stabilization energies indicated weak interaction between the nitrogen lone pair and
proximal radical center in angular 5,6ꢀdiethynylquinoxalines. The overall data indicates
extended angular benzannelation effectively renders Bergman cyclization irreversible due to
favorable aromatic stabilization energy, while extended linear benzannelation results in increased
1
5
retroꢀBergman ring opening, allowing C –C cyclization to become a competitive reaction
channel.
2
ACS Paragon Plus Environment

Page 3 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Introduction
Intramolecular diradical cyclizations of conjugated polyunsaturated systems have become
1
powerful additions to the arsenal of ringꢀforming processes. The prototypical examples include
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
1
6
2
1
5
3
2
7
C –C Bergman and C –C SchreinerꢀPascal cyclizations of enediynes along with C –C
4
2
6
5
MyersꢀSaito and C –C Schmittel cyclizations of enyneꢀallenes (Figure 1). In particular,
Bergman cyclization has been extensively studied due to the ability of natural products such as
6
calicheamicin and dynemicin to cleave DNA upon cyclization of an enediyne unit. In addition
7
to the development of synthetic antitumor agents, diradical cyclizations have found applications
8
as a synthetic tool in the preparation of novel polycyclic aromatic systems, in the preparation of
9
10
polymeric materials, in the construction of molecular electronic nanodevices on surfaces, and
11
may serve as precursors to the formation of aromatic species in interstellar medium.
Furthermore, diradical cyclizations have provided a challenging platform towards the
development of computational models due to the multiconfigurational nature of diradical
1
2
species.
Figure 1. Diradical cyclization pathways of enediynes and enyneꢀallenes.
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
6
For the parent enediyne, (Z)ꢀhexꢀ3ꢀeneꢀ1,5ꢀdiyne, C –C Bergman cyclization is the
preferred thermal reaction pathway, where gain of an aromatic sextet places the pꢀbenzyne
1
3
diradical intermediate in a deep energy well. Incorporating the enediyne alkene within a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6
benzene ring, however, increases endothermicity of C –C cyclization as the aromaticity gain
1
4
upon converting 1,2ꢀdiethynylbenzene to naphthalene is attenuated. As a result, 1,2ꢀ
diethynylbenzene has a lower barrier towards retroꢀBergman ring opening, making cyclization to
1
5
the diradical reversible.
Experimentally, these energetic changes are manifested in a measured rate of
disappearance for 1,2ꢀdiethynylbenzene that becomes dependent upon hydrogen atom donor
1
6
concentration and provides the overall rate constant, k , under pseudoꢀfirst order conditions.
obs
In comparison, the rate of disappearance of simple nonꢀbenzannelated acyclic enediynes are
independent of hydrogen atom donor concentration and directly provide the rate of Bergman
17
cyclization, k1. This critical difference for benzannelated enediynes makes comparison of
literature data on Bergman cyclization kinetics complicated as the concentration of hydrogen
atom donor is highly variable across the available experimental data. This trend continues upon
extended benzannelation as Bergman cyclization of 2,3ꢀdiethynylnaphthalene also exhibits a rate
1
6
dependence on hydrogen atom donor concentration. In this case, however, the measured
disappearance rate is twoꢀfold slower compared to 1,2ꢀdiethynylbenzene with higher
concentrations of hydrogen atom donor required to detect formation of cyclized product.
Unfortunately, the increased endothermic cost upon benzannelation of an enediyne
1
6
cannot readily be offset through traditional substituent effects in the annealed ring. Upon C –C
cyclization, the developing diradicals are orthogonal to the ꢀsystem of the benzene ring
π
preventing direct conjugation between radical centers and substituents. This is evident by the
4
ACS Paragon Plus Environment

Page 5 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
fact that paraꢀsubstituents on 1,2ꢀdiethynylbenzene have minor effects on Bergman cyclization
18
rates operating through relatively weak field effects. The challenge of tuning enediyne
cyclization is further complicated by the presence of two orthogonal ꢀsystems wherein changes
to the inꢀplane ꢀsystem (conversion to ꢀbond and two radical centers) control the reaction
barrier while changes to the outꢀofꢀplane ꢀsystem (gain of aromaticity) can influence overall
reaction energy. One approach to control the rate of Bergman cyclization of benzannelated
π
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
π
σ
π
1
9
2
0
enediynes is through reactant destabilization imparted by orthoꢀsubstituents. Here, steric
compression of the alkynes increases electron repulsion of the inꢀplane πꢀsystem in the reactant
that is alleviated in the transition state. A more extreme route to tune enediyne reactivity through
21
benzannelated substituents employs cyclization of enediyne radical anions. Under such
conditions, radical anionic cyclization of enediynes is accelerated upon benzannelation with
increased sensitivity to remote substituents due to crossing of outꢀofꢀplane and inꢀplane
molecular orbitals near the transition state. In comparison, phenyl groups located on the alkyne
termini are more sensitive to substituent effects as direct conjugation to the inꢀplane πꢀsystem is
available. For example, electron withdrawing groups present in 1,2ꢀbis(4ꢀ
1
6
nitrophenylethynyl)benzene were found to lower the C –C activation barrier compared to
2
2
electron rich paraꢀmethoxyphenyl substituents. Overall, however, substituent effects on
phenylethynyl groups have been much less studied as placement of phenyl groups on the alkyne
1
6
23
termini significantly raises the barrier towards C –C Bergman cyclization. Furthermore, bulky
2
4
1
5
phenylethynyl substituents can lead to phenylꢀshifted products and competing C –C
cyclization facilitated by gain of resonance stabilization to one of the radical centers from the
2
5
external phenyl group.
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In a previous study we reported computational data on diethynylquinoxaline isomers to
further probe the influence of extended benzannelation orientation on Bergman cyclization
2
5
energetics. Calculations revealed that endothermicity of Bergman cyclization to the
corresponding diradical intermediate can increase or decrease relative to 1,2ꢀdiethynylbenzene
(ꢀHrxn = 9.2 kcal/mol) based on the nature of the extended aromatic core. With extended linear
benzannelation, as in 6,7ꢀdiethynylquinoxaline 1 (Figure 2), Bergman cyclization becomes less
thermodynamically favorable (ꢀHrxn = 12.3 kcal/mol) as the aromatic stabilization gained upon
cyclization is further attenuated. In comparison, Bergman cyclization of 5,6ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
diethynylquinoxaline 2 is less endothermic (ꢀHrxn = 5.4 kcal/mol) and approaches the calculated
value for (Z)ꢀhexꢀ3ꢀeneꢀ1,5ꢀdiyne (ꢀHrxn = 0.88 kcal/mol) as a new aromatic sextet is produced,
restoring aromatic stabilization to the angularly fused diyl product. These results are consistent
with the earlier experimental observation that 2,3ꢀdiethynylnaphthalene undergoes a more
sluggish Bergman cyclization compared to 1,2ꢀdiethynylbenzene, likely due to rapid retroꢀ
1
6
Bergman cyclization. Unfortunately, no experimental data is available on 1,2ꢀ
diethynylnaphthalene to directly compare this effect.
Despite these changes to the reaction energy profile, aromatic or heteroaromatic cores are
2
6
commonly utilized when conducting photoꢀBergman cyclization to prevent cis–trans
2
7
isomerization of the enediyne alkene. For example, the seminal work of Turro et al. and Funk
2
8
et al. employed benzannelated enediynes in the photocyclization of acyclic and cyclic
2
9
30
enediynes, respectively. Subsequently, Jones et al. and Hirama et al. locked the central
double bond within an alicyclic ring to demonstrate that aromatic cores are not strictly required
for successful photoꢀBergman cyclization of acyclic enediynes. Interestingly, although direct
comparisons were not made, the isolated yield of photoadduct derived from 1,2ꢀ
6
ACS Paragon Plus Environment

Page 7 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
bis(phenylethynyl)cyclohexene (22%) reported by Jones is higher than values typically reported
2
4,31
for 1,2ꢀbis(phenylethynyl)benzene (1ꢀ10%).
Funk also noted increased conversion of an
angular naphthaleneꢀfused enediyne over the corresponding linear isomer, again suggesting
aromatic stabilization can influence photoꢀBergman cyclization yields. Based on our previous
calculations, along with available literature data, the nature of extended benzannelation can
provide unique opportunities to tune enediyne cyclization energetics.
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 the present paper we report the syntheses, computational studies, kinetic studies, and
photochemical reactivity of a series of diethynylquinoxalines to directly compare the effect of
linear versus angular extended benzannelation on enediyne reactivity (Figure 2). Calculations
were conducted on unsubstituted 6,7ꢀ and 5,6ꢀdiethynylquinoxalines 1 and 2 along with bisꢀ
phenylethynylquinoxaline isomers 3 and 4 to examine changes in reaction energetics along the
1
6
1
5
C –C and C –C reaction channels. Kinetic studies, as a function of 1,4ꢀcyclohexadiene
concentration, were then conducted on terminal quinoxalenediynes 1 and 2 to gain experimental
data to support our calculations. Finally, photochemical reactivity was examined on
phenylethynyl quinoxalenediynes 3–6 to determine if a change in extended benzannelation
orientation manifests in improved photoꢀBergman cyclization yields for quinoxalenediynes.
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
N
N
N
N
1
2
Ph
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Ph
Ph
N
N
N
N
3
5
4
Ph
Ph
Ph
Ph
N
N
N
N
6
Ph
Figure 2. Quinoxalenediynes 1–6.
Results and Discussion
1
5
Computational Analysis. A. C –C Cyclization Products. Prior to examining reaction
energetics, relative enthalpies (Table 1) and relative free energies (Table S3) were calculated for
1
5
the isomeric diyls derived from C –C cyclization of quinoxalenediynes 1–4 (Figure 3). For
1
5
linear derivatives 1 and 3, symmetry in the starting enediyne leads to one possible C –C
cyclization pathway, providing a mixture of E/Z diradical intermediates. However, angular
1
5
derivatives 2 and 4 lack symmetry and have two possible C –C cyclization pathways leading to
a total of four possible isomeric diradical intermediates. For terminal quinoxalenediynes 1 and 2,
1
5
the lowest energy isomers calculated for C –C cyclization were the corresponding Zꢀdiradical
3a
isomers 1ꢀZ and 2ꢀZ, respectively. For phenylꢀsubstituted quinoxalenediynes 3 and 4,
1
5
however, optimization of the E/Z C –C diradical isomers yielded the same structure in which
the exocyclic fulvene radical adopts a virtually 180° bond angle (177.4° in 3ꢀLin, 176.3° in 4ꢀ
Lin, and 174.3° in 4'ꢀLin) leading to essentially linear intermediates which lean 3–6° towards
8
ACS Paragon Plus Environment

Page 9 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
the Zꢀisomer. As a result, optimization of isomers 1ꢀZ and 1ꢀE derived from 3 both lead to 3ꢀ
Lin. Similarly, optimization of 2ꢀZ/E and 2'ꢀZ/E derived from 4 provided 4ꢀLin and 4'ꢀLin,
respectively. A similar geometrical outcome was noted by Alabugin et al. in geometry
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
1
5
optimizations of the C –C cyclization products derived from diarylꢀsubstituted 1,2ꢀ
3
2
diethynylbenzene derivatives. The linearization of the radical center is attributed to
conjugation with the terminal aryl group and consequent rehybridization of the radical center.
1
5
The lowest energy isomers derived from C –C cyclization of 1–4 (1ꢀZ for 1, 3ꢀLin for 3, 2ꢀZ for
2, and 4ꢀLin for 4) were then used in the computation of the cyclization reaction energetics.
1
5
Figure 3. Isomers derived from C –C cyclization of (a) linear quinoxalenediynes 1 and 3 and
(
b) angular quinoxalenediynes 2 and 4.
1
5
Table 1. Relative Enthalpies (25 °C, gas phase, kcal/mol) for C –C Cyclization of 1–4 and
Corresponding Boltzmann Percentages (Cf. structures in Figure 3)
Substrate Products
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
linear
1ꢀZ
0
1ꢀE
1.90
3ꢀLin
n/a
1
(R=H)
(
96.11%) (3.89%)
3 (R=Ph)
n/a
n/a
0
(
100%)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
angular
2ꢀZ
ꢀ0.27
(53.31%) (1.30%)
n/a n/a
2ꢀE
1.93
4ꢀLin
n/a
2'ꢀZ
0
2'ꢀE
0.64
4'ꢀLin
n/a
2
(R=H)
(33.90%) (11.49%)
n/a n/a
4
(R=Ph)
ꢀ0.23
59.51%)
0
(
(40.49%)
B. Cyclization Reaction Energetics. To gain insight into the effect of extended benzannelation
1
6
1
5
orientation and alkyne substitution pattern on C –C and C –C cyclization energetics, we
examined activation enthalpy barriers and reaction enthalpies for quinoxalenediynes 1–4
computationally, with the results summarized in Table 2 (free energy barriers and reaction free
energies provided in Table S1). From the data, several interesting trends emerge.
Table 2. Computed Cyclization Thermodynamic Parameters (25 °C, gas phase, kcal/mol)
1
5
1
6
1
5
1
6
Substrate
linear
C –C
C –C
(C –C )ꢀ(C –C )
‡
‡
‡
ꢁ
H
ꢁHrxn
ꢁH
ꢁHrxn
ꢁꢁH
ꢁꢁHrxn
1
3
(R = H)
(R = Ph)
39.07
38.21
31.46
30.92
31.16
43.07
12.27
27.42
7.91
ꢀ4.86
19.19
3.50
1
0
ACS Paragon Plus Environment

Page 11 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
angular
2
4
(R = H)
(R = Ph)
40.26
40.16
32.61
32.68
30.32
42.36
5.50
21.11
9.94
ꢀ2.20
27.11
11.57
linearꢀangular
1
3
ꢀ 2
ꢀ 4
ꢀ1.19
ꢀ1.95
ꢀ1.15
ꢀ1.76
0.84
0.71
6.77
6.31
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
First, substitution of the terminal alkynes in 1 and 2 with phenyl groups has little effect
1
5
on reaction barrier or reaction energies along the C –C cyclization channel. Apparently, any
increased steric hindrance from phenyl substitution is compensated by the gain of resonance
3
b
1
6
stabilization to the developing exocyclic vinyl radical center. Upon C –C cyclization,
however, increased steric interactions from the addition of phenyl substituents cannot be offset
by resonance stabilization of the developing diradical. As a result, introduction of phenyl
1
6
substituents increases the C –C activation barrier by approximately 12 kcal/mol and reaction
energy by 15–16 kcal/mol for 3 and 4 compared to 1 and 2, respectively.
1
5
1
6
Second, when comparing C –C versus C –C reaction pathways for unsubstituted
1
6
derivatives 1 and 2, C –C Bergman cyclization is favored kinetically and thermodynamically.
1
5
However, upon phenyl substitution in 3 and 4, increased sterics leads to C –C becoming the
1
6
preferred reaction channel kinetically while aromatic stabilization in the C –C product results in
Bergman cyclization remaining the preferred thermodynamic reaction pathway. Overall, the
1
6
thermodynamic preference for C –C cyclization is more pronounced with the angular
2
5
derivatives 2 and 4 primarily due to formation of a new aromatic sextet in these cases.
Nucleusꢀindependent chemical shift calculations (NICS (Table S4) and NICSꢀ1 (Table
1
6
3)) confirm increased aromaticity in the newly formed 6ꢀmembered ring produced upon C –C
cyclization of the angular quinoxalenediynes compared to their linear counterparts. The NICSꢀ1
data reveal that the new aromatic sextet generated in 2 and 4 has aromaticity 7.9% greater and
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
6
1
1.1% greater than the 6ꢀmembered ring formed upon C –C cyclization of 1 and 3, respectively.
1
6
In addition, similar NICSꢀ1 values at the transition states upon C –C cyclization for the linear
‡
and angular analogs provide support for the small ꢁꢁH values calculated between 1 and 2 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6
between 3 and 4 (cf. Table 2). Furthermore, larger NICSꢀ1 values for the C –C cyclization
products derived from 2 and 4 support larger ꢀꢀHrxn values calculated between the linear and
angular quinoxalenediyne derivatives. In comparison, NICSꢀ1 calculations on the newly formed
1
5
5
ꢀmembered rings produced upon C –C cyclization of 1–4 confirm the negligible aromaticity
1
5
gain along the C –C reaction channel. Similar NICS and NICSꢀ1 calculations from Schreiner et
al. on (Z)ꢀhexꢀ3ꢀeneꢀ1,5ꢀdiyne and 1,2ꢀdiethynylbenzene likewise showed appreciable
1
3a
aromaticity in the Bergman cyclization transition states. Further, the decrease in the NICSꢀ1
value from the formation of the new aromatic sextet in pꢀbenzyne compared to the Bergman
product formed from the benzannelated reactant 1,2ꢀdiethynylbenzene is comparable to, though
somewhat smaller than, that calculated here between the Bergman cyclization of angular
quinoxalenediynes 2 and 4 and linear counterparts 1 and 3.
a
Table 3. Calculated NICS-1 Values for the 5- and 6-Membered Rings Formed in the
Cyclization Reactions of 1–4
1
5
1
6
Substrate
C –C
transition state
C –C
transition state
product
ꢀ2.2
ꢀ1.9
ꢀ1.7
ꢀ1.3
product
ꢀ11.9
ꢀ12.9
ꢀ10.2
ꢀ11.4
1
2
3
(R = H)
(R = H)
(R = Ph)
(R = Ph)
ꢀ1.5
ꢀ1.3
ꢀ1.2
ꢀ0.8
ꢀ9.6
ꢀ9.9
ꢀ7.6
ꢀ8.2
4
a
The NICSꢀ1 value for the pꢀbenzyne diradical singlet at the same level of theory is ꢀ12.6.
Finally, when examining the effect of linear versus angular extended benzannelation, the
data indicates steric hindrance combined with aromatic stabilization energy can influence
12
ACS Paragon Plus Environment

Page 13 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
cyclization energetics and reaction pathway. For terminal quinoxalenediynes, derivatives 1 and
1
6
1
5
2
both favor C –C over C –C cyclization as the lower barrier pathway (by 7.9–9.9 kcal/mol) as
well as the more energetically favorable reaction channel (by 19.2–27.1 kcal/mol). For each
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
1
6
parameter, angular quinoxalenediyne 2 has a stronger preference for C –C cyclization which
1
5
produces a new aromatic sextet. Furthermore, for 1 and 2 the retro C –C barrier (~7.6 kcal/mol)
1
6
is significantly lower than the corresponding retro C –C barrier (18.9–24.8 kcal/mol), with the
1
6
C –C retroꢀBergman barrier of angular derivative 2 the highest retroꢀbarrier calculated in this
study, again highlighting the importance of the new aromatic sextet in the angularly fused
1
6
product. Supporting this data, terminal quinoxalenediynes 1 and 2 exclusively undergo C –C
Bergman cyclization under thermal conditions (vide infra). When examining the effect of
extended linear versus angular benzannelation on phenylethynyl derivatives 3 and 4, however,
the situation is not as clear. For the linear derivative 3, addition of phenyl substituents leads to
1
5
1
6
C –C becoming the preferred kinetic pathway (favored by 4.9 kcal/mol) and, while C –C
remains the preferred thermodynamic channel (favored by 3.5 kcal/mol), attenuated aromatic
stabilization gain and increased steric hindrance leads to this isomer having the smallest reaction
1
6
1
5
enthalpy favoring C –C . In comparison, angular phenylꢀsubstituted 4 has a small C –C kinetic
1
6
preference (favored by 2.2 kcal/mol), while the gain of a new aromatic sextet leads to C –C
Bergman cyclization being the preferred thermodynamic reaction channel (favored by 11.6
1
6
kcal/mol). Overall, from this data one may anticipate aromatic stabilization along the C –C
channel for angular quinoxalenediyne 4 to be the dominant energetic factor leading to a
successful photoꢀBergman cyclization. However, photolysis of linear quinoxalenediyne 3, with
the smallest retroꢀBergman ring opening barrier, is less predictable and may lead to a mixture of
cyclization products.
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
C. Cyclization Reaction Electronics. In addition to aromatic stabilization energy, electronic
interactions between the proximal nitrogen atom and the enediyne πꢀorbitals or developing diyl
intermediate can also influence reaction energetics for angular quinoxalenediynes.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Upon examining the reactant angular quinoxalenediynes, Mulliken charge populations
reveal a small repulsion between the nitrogen atom (ꢀ0.493) and the proximal alkyne πꢀbond (ꢀ
0.181) in angular derivative 2. In comparison, for 4, the charge on the nitrogen atom is nearly
equal (ꢀ0.499) to that in 2, while the charge on the proximal alkyne πꢀbond in 4 is significantly
smaller (ꢀ0.039), leading to minimal repulsion. For both 2 and 4, however, this electronic
repulsion has little impact on the geometry of the enediyne functional group (Table 4). For
1
6
example, the C –C distances and enediyne bond angles (as measured from the midꢀpoint of
alkyne to midꢀpoint of alkene to midꢀpoint of alkyne) are nearly equal for quinoxalenediynes 1–
4
3
. In addition, the alkyne bond lengths and bond orders are equal between 1 and 2 and between
and 4.
Table 4. Geometric Parameters for Quinoxalenediynes 1–4
1
6
Substrate
C –C
alkyne bond
enediyne
Wiberg
a
b
a
distances length (Å) bond angle
bond order
(Å)
1
2
3
(R = H)
(R = H)
(R = Ph)
(R = Ph)
4.126
4.136
4.086
4.145
1.208
1.208, 1.208
1.214
89.90°
89.31°
90.50°
89.11°
2.822
2.810, 2.817
2.654
4
1.214, 1.215
2.638, 2.646
a
Alkyne bond lengths and bond orders are identical due to symmetry in 1 and 3. For 2 and 4, the
alkyne proximal to the N atom is listed first.
b
Measured in the enediyne functional unit from midꢀpoint of alkyne to midꢀpoint of alkene to
midꢀpoint of alkyne.
14
ACS Paragon Plus Environment

Page 15 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
To probe the effect of the nitrogen atoms on the diyl intermediates, we examined
biradical stabilization energies (BSE) and singlet–triplet (S–T) splittings for quinoxalenediynes
1
5
1
6
1
–4 along the C –C and C –C reaction channels (Figure 4). Close spatial proximity between
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 nitrogen atom and one of the radical centers in the angular diyl intermediates can affect the
3
3
34
BSE and S–T gap, each of which provide estimates of through bond coupling in the diyl. As
a result, any electronic effect from the proximal nitrogen atom in the angular quinoxalenediynes
may be indirectly monitored by relative changes in the BSE and S–T gap compared to the linear
1
6
derivative. First, the C –C cyclization product derived from 1 has a S–T gap and BSE
35,36
35,36
33
comparable to pꢀbenzyne,
1,4ꢀnaphthalyne,
and 5,8ꢀquinolyne, with BSE values
3
6
1
6
underestimating the S–T gap as previously observed. However, for the C –C cyclization
product derived from angular quinoxalenediyne 2, both the S–T gap and BSE are lower, which
suggests through space interaction between the nitrogen lone pair and the proximal radical that
weakens the through bond orbital interaction between the two radical centers. The phenylꢀ
substituted derivatives 3 and 4 similarly show a decrease in the S–T gap and BSE in going from
1
6
the linear to the angular C –C product, although the values for the S–T gap and BSE are overall
1
5
smaller compared to 1 and 2. Second, the C –C product 1ꢀZ has a larger S–T gap and BSE
1
6
compared to the C –C product from 1 as expected due to the antiꢀperiplanar geometry of the
3
6
1
5
radical centers in 1ꢀZ. In addition, the C –C cyclization product 2ꢀZ also has a larger S–T gap
1
6
and BSE versus the C –C cyclization product from 2. However, there is not as significant of a
1
5
difference for the S–T gap and BSE between 1 and 2 along the C –C channel compared to the
1
6
C –C channel as the distance from the proximal radical carbon to the nitrogen atom in the
1
5
1
6
diradical products from 2 is 0.21 Å greater in the C –C than in the C –C pathway (Table 5)
1
6
1
5
which minimizes any interactions. As with the C –C reaction channel, the C –C products 3ꢀ
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Lin and 4ꢀLin from the phenylꢀsubstituted derivatives also show smaller S–T gaps and BSEs
1
than 1ꢀZ and 2ꢀZ. However, the decrease in the S–T gap is greater than that seen along the C –
6
1
5
C channel upon phenyl substitution, as along the C –C channel the exocyclic radical centers in
ꢀLin and 4ꢀLin rehybridize and gain resonance stabilization with the phenyl group, thereby
reducing the through bond coupling.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
16
ACS Paragon Plus Environment

Page 17 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
HST
BSE H
1
6
C -C Cyclization
N
R
R
N
N
R
R
N
N
R
R
1
3
-2.85
-1.64
1
3
1.39
0.65
+
2
N
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
R
R
R
R
R
R
R
R
N
N
N
N
N
N
N
N
2
4
-2.07
-1.04
2 0.83
4 0.26
+
+
1
5
C -C Cyclization
N
N
N
N
N
N
1-Z -5.14
2-Z -4.95
+
+
3.03
2.92
N
N
N
N
N
N
N
N
N
+
+
N
R
R
R
N
N
R
2
.80
.37
R
R
N
N
N
N
N
3-Lin -3.16
R
+
+
R
R
R
R
R
N
N
N
N
N
2
R
R
R
R
N
N
N
R
R
2.08
N
N
N
R
N
N
R
4
-Lin -2.82
+
R
+
R
N
N
N
N
2.53
R
R
Figure 4. Singlet–triplet splitting (kcal/mol) and biradical stabilization energies (kcal/mol) of
cyclization products from quinoxalenediynes 1–4.
1
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 5. Distance Between Proximal Nitrogen Atom of the Quinoxaline Group and Carbon
Radical (Å) During Cyclization Reactions of 2 and 4
2
(R = H) 4 (R = Ph)
1
5
C –C TS
3.028
3.086
2.856
2.878
3.019
3.080
2.860
2.888
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
5
C –C PRD
1
6
C –C TS
1
6
C –C PRD
The interaction between the radical sites in the diyl intermediates and the proximal
nitrogen atom influences, to a small yet discernable degree, the reactivity of the angular
1
6
quinoxalenediynes. For the C –C pathway transition states and products, the Mulliken charge
populations (Table 6) show that the carbon radical proximal to the nitrogen atom bears a more
positive charge (difference ranging from +0.037 to +0.064) in the angularly fused cases
compared to the linear analogs, while the carbon radical distal to the nitrogen atom bears a more
negative charge (difference ranging from ꢀ0.015 to ꢀ0.025). This represents a small step towards
1
6
37
zwitterionic character for the C –C diradical in the angularly fused cases, and stabilization for
the slightly greater development of the cationic center is provided by the spatially aligned lone
pair of the proximal nitrogen atom in the Bergman cyclization product of 2 and 4. For 4, this
‡
effect leads to a lower activation enthalpy (∆∆H = ꢀ0.71 kcal/mol vs. linear) and contributes to
the lower reaction enthalpy (∆∆Hrxn = ꢀ6.31 vs linear). In 2 though, this effect combines with the
‡
destabilization of reactant 2 to give larger differences versus the linear analog (∆∆H = ꢀ0.84
kcal/mol and ∆∆Hrxn = ꢀ6.77 vs. linear). The dominant energetic contribution to the ∆∆Hrxn
1
6
1
5
values, though, remains aromaticity gain upon C –C cyclization of 2 and 4. For the C –C
pathway transition states and products, the Mulliken charge populations (Table 6) show that the
carbon radical proximal to the nitrogen atom bears only a slightly more positive charge
(difference ranging from +0.029 to +0.038) in the angularly fused cases compared to the linear
18
ACS Paragon Plus Environment

Page 19 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
analogs. Again, this can be explained by the greater separation between the nitrogen atom and
1
5
the proximal radical center in the C –C cyclization pathway, minimizing any stabilization of
zwitterionic resonance structures by the nitrogen lone pair.
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 6. Mulliken Charge Populations for Carbon Radicals During Cyclization Reactions
a
for 1–4
R=H (2)
R=H (1)
R=Ph (4)
R=Ph (3)
C
C
C
C
C
C
C
C
Radical
Radical
Radical
Radical
Radical
Radical
Radical
Radical
Distal
to N
Proximal Distal to Proximal Distal to Proximal Distal to Proximal
to N
N
to N
N
to N
N
to N
1
5
C –C TS
0.006
0.016
0.077
0.025
ꢀ0.309
ꢀ0.221
ꢀ0.012
ꢀ0.036
ꢀ0.032
ꢀ0.013
0.013
ꢀ0.012
ꢀ0.302
ꢀ0.225
0.013
ꢀ0.012
ꢀ0.091
ꢀ0.059
0.001
ꢀ0.007
0.000
0.028
ꢀ0.081
ꢀ0.066
ꢀ0.126
ꢀ0.088
ꢀ0.063
ꢀ0.052
0.015
0.033
ꢀ0.063
ꢀ0.052
1
5
C –C PRD
1
6
C –C TS
1
6
C –C PRD
a
Data for the nitrogen proximal to the carbon radical are provided in Table S5.
Finally, the Mulliken spin populations at the transition states (Table 7) for the radical
1
5
carbons in the angularlyꢀfused C –C reaction are larger (more so for 4 vs. 3 compared to 2 vs. 1)
than for the linear analogs, indicating more radical delocalization in the linear cases. The
1
5
1
5
essentially identical C –C distances in the transition states between 1 and 2 (∆C ꢀC = 0.004 Å)
1
5
and between 3 and 4 (∆C ꢀC = 0.001 Å) suggests the spin population differences are not due to
differences in timing of the transition states along the reaction coordinate. The greater spin
1
5
delocalization for the linear derivatives contributes to the C –C cyclization becoming less
‡
‡
kinetically favorable for 2 vs. 1 (∆∆H = +1.19 kcal/mol) and 4 vs. 3 (∆∆H = +1.95 kcal/mol),
where the latter value is larger due to the greater difference in radical delocalization in 3 vs. 4
compared to 1 vs. 2.
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 7. Mulliken Spin Populations for the Carbon Radicals During the Cyclization
a
Reactions for 1–4
R=H (2)
R=H (1)
R=Ph (4)
R=Ph (3)
C
C
C
C
C
C
C
C
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Radical
Proximal
to N
Radical
Distal
to N
Radical
Proximal
to N
Radical
Distal
to N
Radical
Proximal
to N
Radical Radical
Distal Proximal
Radical
Distal
to N
to N
0.476
0.583
to N
0.572
ꢀ0.600
1
5
C –C TS
ꢀ0.620
0.629
ꢀ0.286
0.598
0.319
ꢀ0.620
ꢀ0.319
ꢀ0.617
ꢀ0.589
ꢀ0.455
0.600
1
6
C –C TS
0.289
a
Data for the nitrogen proximal to the carbon radical are provided in Table S5.
Overall, the data suggests weak interaction between the nitrogen atom and the proximal
radical center in the arynes derived from angular quinoxalenediynes 2 and 4 that is more
1
6
1
5
measurable along the C –C versus C –C channel that leads to a small increase in zwitterionic
character. These effects, however, impart relatively small changes to the reaction energetics
1
6
compared to aromatic stabilization energy along the C –C channel for the angular derivatives.
Future studies will more systematically examine aromatic stabilization energies and biradical
stabilization energies to further refine the role of aromaticity and electronic effect imparted by
nitrogen atoms on cyclization of quinoxalenediynes 1–4 that is beyond scope of the current
paper.
1
6
Syntheses. We previously reported the synthesis and thermal C –C cyclization of 6,7ꢀ
2
5
diethynylquinoxaline 1 starting from 1,2ꢀdiiodoꢀ4,5ꢀdinitrobenzene. In the present study two
alternative synthetic routes were employed to prepare quinoxalenediynes 2–6 to examine kinetics
and photochemical reactivity. Linear 6,7ꢀbis(phenylethynyl)quinoxaline derivatives 3 and 5
were prepared in three steps from commercially available 5,6ꢀdibromoꢀ2,1,3ꢀbenzothiadiazole 7
38
as outlined in Scheme 1. Sonogashira coupling of 7 with phenylacetylene readily afforded
benzothiadiazole 8 in 78% yield. Sulfur extrusion with LiAlH produced diamine 9 and
4
20
ACS Paragon Plus Environment

Page 21 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
immediate condensation with glyoxal or 3,4ꢀhexanedione in ethanol afforded linear
quinoxalenediynes 3 and 5 in twoꢀstep yields of 72% and 50%, respectively. An analogous route
was employed to generate angular derivatives 5,6ꢀdiethynylquinoxaline 2 and 5,6ꢀ
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
39
bis(phenylethynyl)quinoxaline 4 from 4,5ꢀdibromoꢀ2,1,3ꢀbenzothiadiazole 10. Sonogashira
coupling of 10 with trimethylsilylacetylene gave benzothiadiazoleꢀbased enediyne 11 in 65%
yield. Subsequent treatment with LiAlH followed by condensation with glyoxal and
4
desilylation produced 5,6ꢀdiethynylquinoxaline 2 in a three step 45% yield. Similarly,
Sonogashira coupling between phenylacetylene and 10 gave benzothiadiazoleꢀbased enediyne 13
in 46%. Finally, treatment of 13 with LiAlH and condensation with glyoxal produced angular
4
quinoxalenediyne 4 in a twoꢀstep 75% yield.
Scheme 1. Syntheses of Quinoxalenediynes from Dibromobenzothiadiazoles
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Syntheses of quinoxalenediynes from dibromoꢀ2,1,3ꢀbenzothiadiazole isomers is our
preferred method to introduce alterations to the quinoxaline core as condensation with
appropriate 1,2ꢀdicarbonyls is conducted as the final step. In addition, we sought development
of an alternative synthetic route that would employ late stage insertion of the enediyne motif
more suited to modify alkyne substitution pattern in the future. This was accomplished for
preparation of the angular 5,6ꢀdiethynylquinoxaline system starting from 4ꢀaminoꢀ3ꢀnitrophenol
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
15 and used to prepare quinoxalenediyne 6 as outlined in Scheme 2. Hydrogenation of 15
followed by condensation with 3,4ꢀhexanedione afforded 2,3ꢀdiethylꢀ6ꢀhydroxyquinoxaline 16 in
a twoꢀstep 99% yield. Selective bromination of 16 with NBS then afforded 5ꢀbromoꢀ6ꢀ
hydroxyquinoxaline 17 in 86% yield and treatment with Nꢀphenyltriflimide gave triflate 18 in
92% yield. Finally, Sonogashira coupling of 18 with phenylacetylene provided 2,3ꢀdiethylꢀ5,6ꢀ
bis(phenylethynyl)quinoxaline 6 in 42% yield.
Scheme 2. Alternative Synthesis of Angular Quinoxalenediyne 6
O N
OH
1. H2, Pd/C, MeOH
N
N
OH
2
H N
O
2
2.
15
16
O
MeOH
NBS
CH CN, reflux
(
99%)
3
(86%)
Br
Br
PhN(Tf) , Et N
2
3
N
N
OTf
CH Cl
N
N
OH
2
2
(92%)
18
17
Ph
Ph
Pd(PPh ) , CuI
3
4
o
N
N
Et N, 50
C
3
H
Ph
(42%)
6
22
ACS Paragon Plus Environment

Page 23 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Quinoxalenediynes 1–6 were isolated as stable solids and purified by column
chromatography or recrystallization prior to examining reactivity. Terminal quinoxalenediynes 1
and 2 were prepared to conduct kinetic studies while phenyl substituents in derivatives 3–6 were
added to promote photochemical reactivity. Ethyl groups present in 5 and 6 were added to
facilitate solubility and to prevent the potential for radical reactivity alpha to the ring nitrogen
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
4
0
atoms (Minisci reaction ).
Electronic absorbance spectra for quinoxalenediynes 3–6 in CH Cl were characterized
2
2
by a strong absorption band between 280–300 nm with one or two lower energy absorbance
bands from 360–380 nm. Finally, emission spectra of phenylethynyl quinoxalenediynes 3–6
gave a single broad emission band ranging from 399–449 nm, with larger Stokes shifts observed
for angular isomers 4 and 6.
Prior to examining reaction kinetics, bulk thermal cyclizations of 1 and 2 were conducted
in chlorobenzene at 180 °C containing excess 1,4ꢀcyclohexadiene (Scheme 3). For both terminal
diethynylquinoxaline isomers, Bergman cyclization was the only observed reaction channel,
1
6
supporting our calculations indicating C –C cyclization is the preferred kinetic and
thermodynamic pathway. For linear 6,7ꢀdiethynylquinoxaline 1, however, formation of
benzo[g]quinoxaline 19 was accompanied by slow formation of hydrogenated 5,10ꢀ
dihydrobenzo[g]quinoxaline 20 (combined 75% yield, with pure 19 isolated in 30% yield upon
25
DDQ oxidation of the product mixture). In comparison, thermal cyclization of 5,6ꢀ
diethynylquinoxaline 2 afforded benzo[f]quinoxaline 21 in 84% yield with no evidence of the
corresponding hydrogenated product forming.
Scheme 3. Thermal Cyclization of Quinoxalenediynes 1 and 2
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Kinetic Studies. Kinetics for 1 and 2 were followed under conditions originally described by
4
1
42
Grissom et al. and more recently modified by Alabugin et al. (Figure 5). Observed rate
constants for the disappearance of quinoxalenediynes and formation of benzoquinoxaline
products were monitored as a function of 1,4ꢀcyclohexadiene concentration. As described by
Alabugin, monitoring the effect of 1,4ꢀcyclohexadiene concentration on the overall reaction
kinetics allows determination of k , the true first order rate constant for cyclization, as well as the
1
k /k ratio graphically from equation 2 (assuming k is slow in the presence of excess 1,4ꢀ
ꢀ1
2
3
4
2
cyclohexadiene and formation of hydrogenated adduct 20 from 19 is minimal). Deconvolution
of rate constants from the measured kobs is ideally suited to examine the effect of extended
benzannelation orientation on Bergman cyclization of quinoxalenediynes 1 and 2. In the case of
linear 1, diluted aromaticity gain predicted computationally should lead to a fast retroꢀBergman
cyclization, with overall kinetics highly dependent on 1,4ꢀcyclohexadiene concentration and a
relatively large k /k ratio. In comparison, for angular 2 the gain of a new aromatic sextet will
ꢀ1
2
either render Bergman cyclization irreversible, and thus show no dependence on 1,4ꢀ
cyclohexadiene concentration, or show a weaker dependence on 1,4ꢀcyclohexadiene
concentration with a smaller k /k ratio (assuming comparable k values for 1 and 2). As a
ꢀ1
2
2
result, monitoring the effect of 1,4ꢀcyclohexadiene concentration on reaction kinetics for 1 and 2
24
ACS Paragon Plus Environment

Page 25 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
can support our computational data and provide experimental insight into the effect of extended
benzannelation orientation on Bergman cyclization energetics.
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 5. Kinetic model for cyclization of quinoxalenediynes (reference 42).
Kinetic plots monitoring the disappearance of quinoxalenediynes 1 and 2 in the presence
of excess 1,4ꢀcyclohexadiene gave good linearity monitored to 2–3 halfꢀlives indicating first
order or pseudoꢀfirst order decay. Similarly, product formation from 1 showed excellent
linearity through the first 1–2 reaction halfꢀlives. However, at longer reaction time, product
analysis was complicated by secondary formation of 20 and the kinetic significance of k . As a
4
result, plots of product formation from 1 were only plotted through the first 1–2 halfꢀlives and
the small contribution of product 20 was included in the determination of total percent of product
43
at time t. Product analysis for 2, which showed no indication of secondary products via
transfer hydrogenation from 1,4ꢀcyclohexadiene, gave good correlation out to 2–3 halfꢀlives.
Observed rate constants for the disappearance of quinoxalenediynes 1 and 2 and
formation of benzoquinoxaline products as a function of 1,4ꢀcyclohexadiene concentration at 180
°
C are given in Table 8. For linear isomer 1, formation of benzoquinoxaline products gave a
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
slower kobs value compared to disappearance of starting quinoxalenediyne at low 1,4ꢀ
cyclohexadiene concentration (50 equivalents or less). This can be explained by significant
polymerization of the diradical intermediate with k becoming a kinetically significant step. At
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
higher 1,4ꢀcyclohexadiene concentrations (100–250 equivalents) both the disappearance of
quinoxalenediyne 1 and formation of cyclized products 19 and 20 showed a strong dependence
on 1,4ꢀcyclohexadiene concentration with comparable observed rate constants through the first
1
–2 reaction halfꢀlives. At longer reaction times, however, k becomes significant leading to
4
more scatter in the plots of product formation from 1. Despite this complexity, it is clear that
,7ꢀdiethynylquinoxaline 1 shows a strong dependence on 1,4ꢀcyclohexadiene concentration,
6
indicative of a reversible Bergman cyclization, as the rate of product formation increases 6ꢀfold
across the range of 1,4ꢀcyclohexadiene concentrations examined. Finally, a plot of 1/kobs versus
44
2
1
/[CHD] for both quinoxalenediyne disappearance (R = 0.972) and benzoquinoxaline
2
ꢀ4 ꢀ1
formation (R = 0.975) provided an averaged k value of 7.72 ± 1.4 × 10 (s ) and a k /k ratio
1
ꢀ1
2
of 4.30 ± 0.85 for the cyclization of 6,7ꢀdiethynylquinoxaline 1 at 180 °C.
Table 8. Observed Rate Constants for the Disappearance of Quinoxalenediynes and
a
Formation of Benzoquinoxalines at 180 °C
4
4
4
DEQ [1,4ꢀCHD] M
kobs sm × 10
kobs pdt × 10
k1× 10
7.72 ± 1.4
k /k
-1 2
4.3 ± 0.9
b
1
0.258
0.280
0.505 ± 0.03
0.528 ± 0.03
0.878 ± 0.04
1.15 ± 0.03
1.34 ± 0.08
1.56 ± 0.07
1.71 ± 0.05
1.80 ± 0.08
33.1 ± 1.5
0.314 ± 0.01
0.378 ± 0.02
0.860 ± 0.05
1.23 ± 0.07
1.39 ± 0.08
1.60 ± 0.11
1.74 ± 0.10
1.86 ± 0.15
29.8 ± 2.3
0
0
0
.553
.824
.835
1
.10
1.29
.37
0.280
1
c
2
29.0 ± 1.2
0
0
.410
.600
29.0 ± 1.1
28.0 ± 2.0
30.2 ± 2.2
27.8 ± 1.7
26
ACS Paragon Plus Environment

Page 27 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
a
ꢀ1
ꢀ1
b
Pseudoꢀfirst order rate constants (s ) for 1. First order rate constants (s ) for 2. Initial
ꢀ3
c
ꢀ3
concentration of 1 = 5.3 × 10 M. Initial concentration of 2 = 5.5 × 10 M.
Kinetic data for angular quinoxalenediyne 2 again showed a faster rate of starting
material disappearance compared to the measured rate of product formation at the lowest 1,4ꢀ
cyclohexadiene concentration examined (50 equivalents), indicative of polymerization. At
higher 1,4ꢀcyclohexadiene concentrations (75–100 equivalents), measured rates for
disappearance of quinoxalenediyne 2 and formation of benzoquinoxaline 21 were within
experimental error. More importantly, disappearance of quinoxalenediyne 2 and formation of
benzoquinoxaline 21 showed no dependence on hydrogen atom donor concentration as the
measured rates were in excellent agreement. These results are indicative of an irreversible
Bergman cyclization with the measured rate directly providing k . Overall, the averaged k rate
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
1
1
ꢀ3
constant measured for Bergman cyclization of 5,6ꢀdiethynylquinoxaline 2 was 2.90 ± 0.12 × 10
ꢀ1
(s ).
The above experimental data supports our calculations and is consistent with previous
literature data in a number of ways. First, the observed change in rate determining step from
hydrogen atom abstraction for 1 to Bergman cyclization for 2 is consistent with the larger
calculated retroꢀBergman barrier for 2 (assuming hydrogen atom abstraction for 1 and 2 are
comparable). Second, the faster k rate constant measured for 2 compared to 1 supports the
1
lower Bergman cyclization barrier calculated for 2. Experimentally, at 180 °C k is 3.8 times
1
‡
faster for angular derivative 2 compared to the linear isomer 1 (or ꢁꢁH = 1.20 kcal/mol). Based
‡
on our calculations (Tables S6 and S7) at 180 °C, this ratio is predicted to be 2.32:1 (or ꢁꢁH =
0
.76 kcal/mol) in favor of 2, in excellent agreement with the measured values from the kinetic
studies. Third, the observed rate of disappearance of 1 at 180 °C in the presence of 100ꢀfold
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 54
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ꢀ5
excess 1,4ꢀcyclohexadiene (8.78 × 10 ) is significantly slower than the rate of disappearance of
ꢀ4 42
1
,2ꢀdiethynylbenzene reported by Alabugin et al. under identical conditions (7.83 × 10 ).
Assuming comparable k rates, this result is indicative of a larger k rate constant upon linear
2
ꢀ1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
extended benzannelation present in 1. Furthermore, the measured k /k ratio for
ꢀ1
2
quinoxalenediyne 1 (4.3) is significantly larger with k becoming the dominant term compared to
-1
k /k values measured for 3ꢀnitroꢀ1,2ꢀdiethynylbenzene (0.16 at 164 °C) and 4ꢀnitroꢀ1,2ꢀ
ꢀ1
2
4
2
diethynylbenzene (0.10 at 170 °C) by Alabugin. Fourth, though only limited data is available
to plot equation 2, the rate of thermal Bergman cyclization of 2,3ꢀdiethynylnaphthalene shows a
strong 1,4ꢀcyclohexadiene dependence similar to quinoxalenediyne 1 with a large k /k ratio (5.1
ꢀ1
2
1
6
at 152 °C). Finally, gain of a new aromatic sextet may explain earlier Bergman cyclization
kinetic data reported in the literature including the increased reactivity of cinnolinoenediynes
45
compared to analogous benzo–fused enediynes and the absence of 1,4ꢀcyclohexadiene
4
6
concentration dependence on the rate of cyclization of imidazole–fused enediynes.
The combined computational and kinetic data indicates that the retroꢀBergman ring
opening barrier for 6,7ꢀdiethynylquinoxaline 1 is extremely low and leads to k becoming faster
ꢀ1
than the rate of hydrogen atom abstraction k . However, retroꢀBergman ring opening is
2
effectively shut down for the isomeric 5,6ꢀdiethynylquinoxaline 2 as the gain of a new aromatic
sextet decreases Bergman cyclization endothermicity. Though the contribution from the nitrogen
lone pair remains to be fully dissected, it is evident that extended benzannelation orientation in
quinoxalenediynes 1 and 2 has a significant influence on Bergman cyclization energetics.
Photochemical Reactivity. With terminal quinoxalenediynes 1 and 2 readily undergoing
1
6
thermal C –C Bergman cyclization, we turned our attention towards photocyclization of
phenylethynyl derivatives. Photochemical reactivity of quinoxalenediynes 3–6 was examined in
28
ACS Paragon Plus Environment

Page 29 of 54
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
a Rayonet RPR 100 reactor equipped with sixteen 3000 Å lamps. Initial studies were conducted
on angular 5,6ꢀbis(phenylethynyl)quinoxaline 4 in isopropanol, the most common solvent used
in photoꢀBergman cyclizations. Under dilute conditions, complete consumption of
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
quinoxalenediyne 4 was observed in four hours, providing a low and variable yield (1–5%) of
1
6
C –C cyclization product 22 (Scheme 4). Addition of 1,4ꢀcyclohexadiene led to more
reproducible yields of benzoquinoxaline 22, though the highest yield obtained was only
marginally improved (7%). Optimal yields were obtained upon increasing reaction concentration
(3 mL isopropanol per mg 4, 200 equivalents 1,4ꢀcyclohexadiene, 4 h), providing 22 in 13%
yield (based on 32% recovered quinoxalenediyne 4). An increase in reaction time gave higher
conversion of 4, however, significant polymerization led to lower isolated yields of 22 upon
extended irradiation. Similarly, attempts to further increase reaction concentration led to poor
solubility of quinoxalenediyne 4 in isopropanol, while addition of coꢀsolvents (acetonitrile,
toluene) reduced isolated yields of 22. Finally, with a significant absorbance band near 360 nm,
photolysis of quinoxalenediyne 4 under our optimized conditions with 3500 Å lamps provided a
comparable 10% yield of benzoquinoxaline 22 (based on 21% recovered starting material).
Rapid consumption of quinoxalenediyne 4 under dilute conditions, along with evidence
of higher molecular weight adducts observed by GCMS analysis, led us to speculate that radical
4
0
additions alpha to the quinoxaline nitrogen atoms (Minisci reaction ) may contribute to rapid
degradation of starting material. To limit this potential side reaction, ethyl substituents were
introduced at positions 2 and 3 of the quinoxaline core in 6. Irradiation of quinoxalenediyne 6
under our optimized conditions employed for 4 (3 mL isopropanol per mg 6, 200 equivalents
1,4ꢀcyclohexadiene, 300 nm, 4 hours) gave 49% conversion of 6 with 17% yield of
benzoquinoxaline 23, a 30% increase in yield compared to 4. It should be noted, however, that
29
ACS Paragon Plus Environment