Singlet vs Triplet Reactivity of Photogenerated α,n-Didehydrotoluenes

Article abstract of DOI:10.1021/acs.joc.7b00610

The reactivity of α,n-didehydrotoluenes (DHTs) in protic media (organic/aqueous mixtures) was explored by means of a combined computational and experimental approach. These intermediates were generated via a photoinduced double elimination process occurring in (chlorobenzyl)trimethylsilanes and led to the formation of a varied products distribution, depending on the isomer tested. Irradiation of ortho- and para-derivatives resulted, respectively, in the formation of triplet α,2- and α,4-DHTs, whose diradical reactivity led to both radical and polar products. On the other hand, irradiation of the meta-precursor led to the singlet α,3-DHT isomer. The latter showed a marked preference for the formation of polar products and this was rationalized, as supported by computational evidence, via the involvement of a zwitterionic species arising through interaction of the nucleophilic solvent with the benzylic position of the DHT.

Full text of DOI:10.1021/acs.joc.7b00610

Article
Singlet vs Triplet Reactivity of Photogenerated alpha,n-Didehydrotoluenes
Chiara Pedroli, Davide Ravelli, Stefano Protti, Angelo Albini, and Maurizio Fagnoni
J. Org. Chem., Just Accepted Manuscript • Publication Date (Web): 07 Jun 2017
Downloaded from http://pubs.acs.org on June 9, 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 38
The Journal of Organic Chemistry
1
2
3
4
Singlet vs Triplet Reactivity of Photogenerated α,nꢀDidehydrotoluenes
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
Chiara Pedroli, Davide Ravelli,* Stefano Protti, Angelo Albini, Maurizio Fagnoni*
PhotoGreen Lab, Department of Chemistry, Viale Taramelli 12, 27100 Pavia, Italy.
Fax: +39 0382 987323; Tel: +39 0382 987198.
Abstract: The reactivity of α,nꢀdidehydrotoluenes (DHTs) in protic media (organic/aqueous
mixtures) was explored by means of a combined computational and experimental approach. These
intermediates were generated via a photoinduced double elimination process occurring in
(chlorobenzyl)trimethylsilanes and led to the formation of a varied products distribution, depending
on the isomer tested. Irradiation of orthoꢀ and paraꢀ derivatives resulted, respectively, in the
formation of triplet α,2ꢀ and α,4ꢀDHTs, whose diradical reactivity led to both radical and polar
products. On the other hand, irradiation of the metaꢀprecursor led to the singlet α,3ꢀDHT isomer.
The latter showed a marked preference for the formation of polar products and this was rationalized,
as supported by computational evidences, via the involvement of a zwitterionic species arising
through interaction of the nucleophilic solvent with the benzylic position of the DHT.
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 38
1
2
3
4
5
6
Introduction
7
8
9
In organic molecules the covalent bond, that is the sharing of two (or more) electrons between two
atoms, is what typically makes molecules stable. Chemical reactions most often occur by cleaving a
bond and thus via high energy intermediates (a couple of radicals or a couple of ions),1 unless the
simultaneous formation of another bond leads back to stability (as typical of concerted processes).2
When the fragments are bonded, diradicals3 or zwitterions are formed (Scheme 1a).4
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
Scheme 1. a) Generation of a Biradical and a Zwitterion via Homolytic and Heterolytic Bond
Cleavage. b) Bergman Cycloaromatization of Enediynes for the Generation of 1,4ꢀBenzynes.
The competition between homoꢀ and heterolytic cleavage of a bond is thus a phenomenon of
general interest.1,5 The observed reactivity, viz. atom abstraction or cation (anion) transfer, is the
election tool to have insight on the lysis mechanism and enables a better characterization of the
intermediates. An intriguing case is that of cycloaromatizations, such as the Bergman reaction,6,7
where a new σ bond is formed at the expenses of two π bonds. As a result, a pꢀbenzyne
intermediate is generated, endowed with a peculiar dual reactivity, showing either a diradical1,5,8,9 or
a zwitterionic10 behavior (Scheme 1b). A related case is represented by didehydrotoluenes (DHTs,
II; Scheme 2), intermediates formally resulting from the elimination of two hydrogen atoms from a
toluene molecule, one from the benzylic position and one from the ring. Accordingly, these are
heterosymmetric species (σ1π1), where the two centers are not expected to interact due to geometric
2
ACS Paragon Plus Environment

Page 3 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
factors. Depending on the relative arrangement of the two radical centers, three isomeric DHTs,
namely α,4ꢀ (IIa), α,3ꢀ (IIb) or α,2ꢀDHT (IIc), are possible and all of them can exist in two
different spin states, either the singlet or the triplet. Indeed, theoretical investigations demonstrated
that the α,4ꢀ and α,2ꢀDHTs are ground state triplets (with a tripletꢀsinglet gap of 7.4 and 8.1 kcal
molꢀ1, respectively), while in the case of the α,3ꢀisomer the two states are closer in energy, with the
singlet slightly more stable than the triplet (3.0 kcal molꢀ1 gap).11,12
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
Scheme 2. Generation of α,nꢀDHTs via: a) MyersꢀSaito Cycloaromatization of EnyneꢀAllene I
and b) Photolysis of Benzyl Derivatives V.
Notably, intermediates containing the α,3ꢀDHT moiety are known to have a cytostatic activity, that
has been attributed to their capability of cleaving DNA.13 In fact, highly unsaturated hydrocarbons,
such as enyneꢀallenes (I, Scheme 2a) present in natural (as an example, the properties of antibiotic
Neocarzinostatin are due to its conversion to an activated derivative containing the enyneꢀallene
moiety)14,15 or artificial9,16,17 compounds, may cyclize via diradical intermediates. A typical case is
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 38
1
2
3
4
5
6
7
8
9
the MyersꢀSaito reaction, where α,3ꢀdidehydrotoluenes are exclusively formed.14,18 Indeed, model
studies have revealed the formation of products expected from both radical and ionic paths.1 As a
matter of fact, α,3ꢀdidehydrotoluenes can lead (among others) to both benzyl methyl ether (III) and
phenethyl alcohol (IV) in the reaction with MeOH in variable amounts.1,19 This calls for a
diradical/zwitterion dichotomy for α,3ꢀDHTs (see structures IIb / IIbꢀ/+) where, however, in both
cases either a hydrogen atom or a proton transfer from the medium to IIb or IIbꢀ/+ is the first step
involved.18d
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
While the formation of radical products can be easily explained by considering the diradical nature
of α,3ꢀDHT, more difficult is the rationalization of polar products formation. Originally, it was
proposed that OꢀH insertion products could be explained on the basis of a resonance between
diradical IIb and zwitterion IIbꢀ/+, sharing a negative charge at the dicoordinated carbon and a
positive one at the benzylic position.18d Such resonance, however, is symmetry forbidden and the
zwitterion is actually an excited state of IIb.20 Indeed, experimental studies ruled out the likelihood
of a single intermediate mechanism, where the diradical and the zwitterion are in resonance.20,21
Nevertheless, a recent report supported that a part of the MyersꢀSaito reaction follows a
nonadiabatic course. Thus, the zwitterion was predicted to form directly from the enyneꢀallene, in a
postꢀrateꢀdetermining bifurcation. A postꢀtransition state nonadiabatic transition from the groundꢀ
state singlet of enyneꢀallene to the excitedꢀstate singlet zwitterion was then invoked.22
Such scenario, however, applies only to the MyersꢀSaito reaction and to the reactivity of α,3ꢀDHT,
the only isomer accessible by this approach. In view of this, we reconsidered our recently developed
alternative access to all α,nꢀDHTs, including thermally inaccessible IIa and IIc isomers. This
involves a oneꢀphoton, double elimination of a nucleofugal group (chloride,19,23,24 sulfonate25 or
phosphate25 anions) and an electrofugal moiety (trimethylsilyl cation19,23 or carbon dioxide24) from
aromatic compounds (Vaꢀc), occurring efficiently at room temperature via triplet phenyl cations
3
(3VI+ or VII, respectively; Scheme 2b). Interestingly, the products distribution observed when
4
ACS Paragon Plus Environment

Page 5 of 38
The Journal of Organic Chemistry
1
2
3
generating IIb under thermal conditions from enyneꢀallene I and under irradiation conditions from
4
5
6
(3ꢀchlorobenzyl)trimethylsilane is roughly the same.19
7
8
9
It appeared worthwhile to unravel the nature of the involved intermediates by generating the three
isomeric α,nꢀDHTs and analyzing the distribution of the products obtained depending on the
reaction conditions. This investigation should help in the rationalization of the diradical / zwitterion
dichotomy observed in α,nꢀDHTs chemistry, being thus beneficial in the comprehension of the
mode of interaction of didehydrotoluenes with the surrounding medium in biological systems. We
thus investigated the photochemistry of the three isomeric (chlorobenzyl)trimethylsilanes 1aꢀc in
protic media and in the presence of suitable additives (Schemes 3ꢀ7). The rationalization of the
reactivity offered was supplemented also by computational analysis. The results of this study are
reported below.
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
As mentioned in the Introduction, the photochemical decomposition of 1aꢀc in protic solvents
involves a heterolytic dechlorination and ensuing desilylation from the benzylic position.19,23a The
study was carried out in organic/aqueous mixtures, maintaining the highest amount of water
possible (solvent/H2O ratio from 4/1 to 2/1) for dissolving 1aꢀc (0.025 M). This choice was based
on two grounds, viz. i) in view of the closeness to biologic conditions and ii) because previous
studies showed that dechlorination was efficient in the presence of water.19,23,24 The irradiation was
carried out for 15 h by using phosphorꢀcoated lamps centered at 310 nm. The products formed are
gathered in Table 1, where they have been divided in two classes, namely those assigned to the
phenyl cation reactivity (dechlorinated/silylated products) and those deriving from α,nꢀDHTs
(dechlorinated/desilylated products). The ratio between cation vs α,nꢀDHT deriving products and, in
the latter case, the ratio between radical vs polar products have been likewise reported (see further
the Discussion section and Table S1 in the Supporting Information).
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 38
1
2
3
4
5
6
7
8
9
Irradiation in aqueous alcohols (MeOH, EtOH, iPrOH, tBuOH) yielded both siliconꢀcontaining
products (benzylsilane 2, phenols 8aꢀc, as well as aromatic ethers 11aꢀc, 13aꢀc) and siliconꢀfree
compounds, viz. toluene 3, bibenzyl 4, αꢀ (or βꢀ in the case of tBuOH) benzylated alcohols 5, 9, 12
and 14, as well as benzyl alcohol 7 and benzyl ethers 6, 10 and 15 (see Table 1 and Schemes 3, 4).
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
Scheme 3. Photoreactivity of Chlorobenzylsilanes 1aꢀc in MeOH/H2O 4:1 Mixture.
Scheme 4. Photoreactivity of 1aꢀc in Aqueous Alcohols.
The quantum yield of decomposition (Φ−1) was particularly high in MeOH/H2O 4/1, but reasonable
values were found also in the other cases, the tBuOH/H2O 3/1 mixture being the medium where
compounds 1aꢀc were less photoreactive. As for products distribution, dechlorinated 2 was formed
in significant yields (20ꢀ49%) for 1b,c in all of the solvents tested. Radical products (3ꢀ5, 9, 12, 14)
were predominant in the case of paraꢀ and orthoꢀ chlorides 1a,c.
6
ACS Paragon Plus Environment

Page 7 of 38
The Journal of Organic Chemistry
1
2
3
4
5
Table 1. Irradiation of Compounds 1aꢀc in (Deuterated) Aqueous Alcohols.a
6
7
8
9
Products (yield %)c
Cation/DHTd
Radical/Polare
b
Halide
Φꢀ1
Cation products
α,nꢀDHT products
CH3OH/H2O 4/1
3, 1; 4, 23; 5, 56; 6, 5
5, 12; 6, 42; 7, 14
4, 5; 5, 23; 6, 3
EtOH/H2O 2/1
3, 4; 4, 29;
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
0.73f
0.82f
0.27f
2, 3
2, 25
2, 49
3/97
27/73
61/39
94/6
18/82
90/10
1a
1b
1c
0.55
2, 2
2/98
98/2
1a
9, 59; 10, 2
3, 1; 4, 5; 7, 18;
9, 32; 10, 14
0.28
0.41
2, 25; 8b, 4
2, 35; 8c, 1
29/71
52/48
54/46
100/0
1b
1c
3, 3; 9, 31
iPrOH/H2O 4/1
3, 12; 4, 7; 12, 56
3, 11; 12, 27
0.20
0.32
0.28
2, 4
5/95
53/47
63/37
100/0
100/0
100/0
1a
1b
1c
2, 37; 11b, 6
2, 45; 11c, 3
3, 6; 12, 22
tBuOH/H2O 3/1
4, 9; 7, 5;
14, 16; 15, 10
0.08
ꢀ
0/100
62/38
1a
2, 21; 8b, 10;
13b, 8
0.16
0.19
14, 16; 7, 18; 15, 13
45/55
50/50
34/66
53/47
1b
1c
2, 20; 8c, 3
7, 3; 14, 12; 15, 8
CH3OD/D2O 4/1
4, 23; 5ꢀd1, 67;
6ꢀd1, 3
2, 1
1/99
97/3
1a
4, 3; 5ꢀ d1, 12;
6ꢀd1, 22; 7ꢀd2, 7
4, 6; 5ꢀd1, 38; 6ꢀd1, 3
CD3OH/H2O 4/1
4ꢀd2, 14; 5ꢀd3, 22;
6ꢀd3, 11
2, 31
2, 40
41/59
46/54
34/66
94/6
1b
1c
ꢀ
0/100
32/68
41/59
77/23
0/100
76/24
1a
1b
1c
2ꢀd1, 23
2ꢀd1, 26
6ꢀd3, 39; 7, 9
4ꢀd2, 6; 5ꢀd3, 22;
6ꢀd3, 9
a
Conditions: 0.025 M solution of 1aꢀc in the chosen solvent irradiated at 310 nm. tirr = 15 h (see also Table
b
S1 in Supporting Information). Quantum yields of disappearance (Φꢀ1) of 1aꢀc (10−2 M in the chosen
c
d
media) were measured at 254 nm. Yields were based on consumed 1aꢀc and determined by GC analysis.
Ratio between phenyl cation vs DHT deriving products. e Ratio between radical vs polar α,nꢀDHTs deriving
products. f From ref. 19.
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 38
1
2
3
4
5
6
7
8
9
Table 2. Irradiation of 1aꢀc in Aqueous Ethers and Acetonitrile.a
Products (yield %)c
Halide
Conv. (%)b
Cation/DHTd
Radical/Polare
Cation
products
α,nꢀDHT
products
THF/H2O 2/1
4, 11; 16, 61
16, 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
73
55
43
2, 10
2, 43
2, 33
12/88
98/2
97/3
100/0
100/0
100/0
1a
1b
1c
16, 1
Dioxane/H2O 2/1
4, 10; 17, 66
4, 3; 17, 51
17, 32
100
100
100
2, 4
2, 41
2, 49
5/95
43/57
61/39
100/0
100/0
100/0
1a
1b
1c
CH3CN/H2O 4/1
4, 15; 18, 51
ꢀ
100
19
ꢀ
0/100
100/0
61/39
100/0
ꢀ
1a
1b
1c
2, 37
2, 23
19
4, 15
100/0
a See Table 1. b Conv. (%) = 1aꢀc consumption (%). cꢀe See Table 1.
By contrast, compounds arising from the polar pathway of DHTs were mainly formed in high yields
in the case of 1b. On the other hand, no polar product was obtained when using isopropanol/water
as medium, independently from the considered isomer, whereas a competition between the two
pathways took place in aqueous tertꢀbutanol.
8
ACS Paragon Plus Environment

Page 9 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
Scheme 5. Photoreactivity of 1aꢀc in a) CH3OD/D2O 4:1 Mixture and b) CD3OH/H2O 4:1
Mixture.
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
CH2OCH3
CH2CH2OD
CH2OD
a)
h
15 h
2 +
4 +
1a-c
+
+
CH3OD/D2O 4:1
D
D
5-d1
6-d1
7-d2
D
TMS
+
b)
15 h
CD3OH/H2O 4:1
h
1a-c
4-d2
D
D
2-d1
CH2CD2OH
CH2OCD3
+
+
+ 7
H
6-d3
D
5-d3
The reaction was then extended to deuterated methanol/water mixtures. Comparison with the non
deuterated medium evidenced that shifting to CH3OD/D2O (Scheme 5a) halved the yield of
products 6 and 7 (in the case of 1a,b), but had only a small effect on the other products yield. On
the other hand, using the CD3OH/H2O mixture (Scheme 5b) diminished the yield of products 4 and
5 from 1a, and of products 5 and 7 from 1b (see Table 1, lower part).
In THF and 1,4ꢀdioxane (2:1 with water), bibenzyl 4 and benzylated ethers 16 and 17 were the main
products from the paraꢀ isomer (Table 2). Ether 17 was also formed in significant amounts from
both the orthoꢀ and metaꢀ derivatives (Scheme 6). Nitrile 18 was formed instead in aqueous
acetonitrile (4:1) only from 1a and no solvent incorporating products were detected in the case of
1b and 1c (Scheme 6).
The reaction was then studied in MeOH/water 4:1 in the presence of some additives (Scheme 7).
Thus, when 1,4ꢀcyclohexadiene (CHD, 0.2 M) was present, toluene 3, bibenzyl 4 (in the case of
1a), benzylcyclohexadiene 19, and benzyl methyl ether 6 (in the case of 1b), as well as benzylsilane
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 38
1
2
3
4
5
6
7
8
9
2, were formed in largely different yields from the three isomers (Table 3). CCl4 (0.5 M) caused a
lowering of the yield of 2, while the main products were (2,2,2ꢀtrichloroethyl)benzene 20 (that is
present only for isomers 1a,c) and the three isomeric 1ꢀchloroꢀnꢀ(2,2,2ꢀtrichloroethyl)benzenes 21aꢀ
c along with some ether 6 (only from the metaꢀ isomer).
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
Scheme 6. Photoreactivity of 1aꢀc in Aqueous Dioxane, Acetonitrile and THF.
Scheme 7. Photoreactivity of 1aꢀc in the Presence of 1,4ꢀCyclohexadiene and CCl4.
15 h
h
19
2 + 3 + 4 + 6 +
MeOH/H2O 4:1
(0.2 M)
1a-c
CH2CCl3 CH2CCl3
+
6 h
h
2 + 3 + 6 +
MeOH/H2O 4:1
CCl4 (0.5 M)
Cl
20
21a-c
10
ACS Paragon Plus Environment

Page 11 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Irradiations of 1aꢀc in MeOH/H2O 4/1 in the Presence of Additives.a
Products (yield %)c
Halide Conv. (%)b
Cation/DHTd
Radical/Polare
Cation
products
α,nꢀDHT
products
0.2 M 1,4ꢀCyclohexadiene
3, 40; 4, 21;
19, 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
99
2, 7
7/93
100/0
1a
3, 6; 6, 11;
53
61
2, 75
2, 53
77/23
89/11
50/50
100/0
1b
1c
19, 5
3, 3; 19, 3
0.5 M CCl4
56
8
ꢀ
20, 20; 21a, 33
6, 21; 21b, 72
20, 15; 21c, 22
0/100
5/95
100/0
77/23
100/0
1a
1b
1c
2, 5
32
2, 11
23/77
a See Table 1. b Conv. (%) = 1aꢀc consumption (%). cꢀe See Table 1.
Calculations
To rationalize the reactivity observed for the three isomeric α,nꢀDHTs in different media, we
undertook a computational investigation to document the details of the reaction path describing the
transfer of a hydrogen from the chosen solvent to DHTs. According to previous work by our
group,23 the theoretical method adopted here made use of a multiꢀconfigurational approach, viz. the
Complete Active Space SelfꢀConsistent Field (CASSCF) method, required to correctly handle the
peculiar electronic structure of DHTs, also including proper corrections to account for the solvent
effect (an implicit model has been used) and for the dynamic electron correlation term (see Section
2 in Supporting Information for details). Furthermore, we also evaluated the role of specific
interactions of DHTs with the medium by explicitly including solvent molecules into the simulation.
At the beginning, we focused our attention on the simplest system possible, viz. that involving the
DHT and one molecule of solvent (MeOH). In particular, we optimized the transition states (TSs)
describing the transfer of H from MeOH to the dicoordinated carbon of IIaꢀc, the most reactive site
of the intermediate.26 Two different kinds of transfer path have been considered, since the Hꢀatom
may be abstracted from either the CꢀH or the OꢀH positions. As apparent from the structure of 3IIa
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 38
1
2
3
4
5
6
7
8
9
(taken as a model) reported in Figure 1, the methanol molecule assumes two different arrangements
depending on the process considered, with the CMꢀO bond lying slightly tilted or perpendicularly
with respect to the ring of the DHT in the TS structures describing the CꢀH and the OꢀH transfer,
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
Figure 1. Transition state structures for the transfer of HꢀCH2OH (left part) and HꢀOCH3 (right
part) to 3IIa, taken as a model.
After optimization of the TSs, we investigated the corresponding intrinsic reaction coordinates
(IRCs) in both directions to have an idea of the energy changes involved in the process. The results
are gathered in Table 4, where both the overall free energy change (ꢀGCASMP2) and the free energy
barrier (ꢀG‡CASMP2) have been reported. The energy profiles along the whole IRC (100 steps in both
directions) are plotted in Figures 2 and 3 for Hꢀtransfer from HꢀCH2OH and HꢀOCH3, respectively,
in the case of 1,3IIa (part a) and 1,3IIb (part b). As for the CꢀH profile, the reaction is largely
exergonic and the different isomers show similar behaviors, with ꢀG‡ values between ca. 13 and 19
kcal molꢀ1. A similar scenario was found when considering the H transfer from the OꢀH position,
where the reaction is again exergonic, although with less favorable energy changes. However, the
observed ꢀG‡ values are very similar to the previous case (13 to 18 kcal molꢀ1 range). Apart from
MeOH, we also evaluated the reactivity of 1,3IIb at the HꢀC(CH3)2OH position of isopropanol, since
in the experiments we found a large amount of radicalꢀtype products when 1b was irradiated in
12
ACS Paragon Plus Environment

Page 13 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
iPrOH/H2O. Indeed, the model highlighted lower ꢀG‡ and more negative ꢀG values with respect to
the case of MeOH. Finally, we considered the case of the HꢀCH2CN position in acetonitrile, for
which we investigated the paraꢀ and metaꢀ isomers only, since they show a ground state with
different spin multiplicity (the triplet and the singlet, respectively). A more varied situation has been
observed in this case, since, as far as the ground state is concerned, the paraꢀ isomer showed less
favorable parameters with respect to the CꢀH of MeOH, while the opposite was true for the metaꢀ
one (Table 4).
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. Calculated Parameters for the TSs Describing the Transfer of H from Solvent to IIaꢀ
c.a
ꢀG‡
ꢀGCASMP2
CASMP2
DHT
[kcal·molꢀ1]
[kcal·molꢀ1]
Reactivity at HꢀCH2OH
1IIa
3IIa
1IIb
3IIb
1IIc
3IIc
19.14
16.22
ꢀ23.06
ꢀ13.50
ꢀ13.60
ꢀ18.72
ꢀ24.00
ꢀ14.28
16.12
12.79
16.06
15.05
Reactivity at HꢀOCH3
14.05
1IIa
3IIa
1IIb
3IIb
1IIc
3IIc
ꢀ14.40
ꢀ4.77
ꢀ7.07
ꢀ8.69
ꢀ14.22
ꢀ5.08
15.77
17.65
14.75
13.32
16.34
Reactivity at HꢀC(CH3)2OH
1IIb
3IIb
9.59
ꢀ16.82
8.49
ꢀ20.21
Reactivity at HꢀCH2CN
1IIa
3IIa
1IIb
3IIb
14.12
17.72
13.77
13.83
ꢀ21.91
ꢀ9.92
ꢀ13.80
ꢀ16.04
a
The data referred to the most stable spin isomer have been reported in bold.
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 38
1
2
3
4
5
6
7
8
9
An interesting remark is related to the charge (qESP data have been reported; see also Computational
Details) associated with the Hꢀatom during the transfer step (purple values in Figures 2 and 3). In
the case of HꢀCH2OH the charge always fell in the ꢀ0.1 to +0.1 range along the entire IRC (Figure
2). Some differences have been observed in the case of iPrOH and MeCN, where the charge of the
Hꢀatom was slightly negative (around ꢀ0.15 at the TS; Figure S6) in the former case, while it
assumed positive values (up to ca. +0.2, particularly in the first part of the IRC; Figure S8) in the
latter. Different was the case of the HꢀOCH3 profiles, where a positive charge was associated with
the Hꢀatom and it assumed much larger values (up to +0.5; Figure 3).
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. IRC plots of the TSs describing the reactivity of: a) IIa and b) IIb with methanol at the
HꢀCH2OH position (E0(CASSCF,MeOH) data have been reported, as detailed in Tables S5ꢀS8; see
Section 2.1). ESP charges (qESP) for HꢀCH2OH and the dicoordinated carbon on relevant points
(first, TS and last point) along the IRC are shown in purple and grey, respectively, as derived from
the results of calculations at the CPCMꢀCASSCF/6ꢀ31G(d,p) level of theory in MeOH.
14
ACS Paragon Plus Environment

Page 15 of 38
The Journal of Organic Chemistry
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 3. IRC plots of the TSs describing the reactivity of: a) IIa and b) IIb with methanol at the
HꢀOCH3 position (E0(CASSCF,MeOH) data have been reported, as detailed in Tables S11ꢀS14; see
Section 2.2). ESP charges (qESP) for HꢀOCH3 and the dicoordinated carbon on relevant points (first,
TS and last points) along the IRC are shown in purple and grey, respectively, as derived from the
results of calculations at the CPCMꢀCASSCF/6ꢀ31G(d,p) level of theory in MeOH.
This charge localization is highly dependent on the OꢀH distance and decreases along the IRC,
finally reaching the same values observed above for CꢀH around the endꢀpoints of the reaction path.
Interestingly, we also verified that almost no charge localization occurred at the dicoordinated
carbon atom, independently from the process and the isomer considered, with charge values ranging
from ꢀ0.1 to +0.1 (grey values in Figures 2 and 3). The only exception to this trend is represented by
the process describing the transfer of HꢀCH2OH to IIc, where negative values (between ꢀ0.15 and ꢀ
0.2) have been observed in the last part of the IRC (see Figure S2 for details).
Another issue is related to the starting point of the reported IRCs. Actually, we tried to optimize the
corresponding structures to check for the formation, if any, of complexes between the DHTs and the
molecule of solvent. Any attempt, however, failed and we never observed any coordination of
methanol (either through the CꢀH or the OꢀH group) to the dicoordinated carbon of the DHT,
independently from the isomer and spin considered.
As mentioned in the Introduction, the involvement of a zwitterionic intermediate to account for the
reactivity observed in the case of IIb has been long debated in the literature.22 Accordingly, we
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 38
1
2
3
4
5
6
7
8
wanted to address this point in our simulation, with particular regards to the involvement of such
electronic state during the Hꢀtransfer step from HꢀOCH3. Thus, we took the geometries of the points
1
along the IRC (from the starting point to the TS) reported in Figure 3b for IIb and calculated the
9
energy of the first excited state for each of them (see Section 2.5 in Supporting Information for
details) to obtain a reaction profile describing the Hꢀatom transfer step along the excited state
surface. The results are gathered in Figure 4, where it can be appreciated that the excited state
profile lies ca. 50 kcal molꢀ1 above the IRC starting point of the ground state surface. Another
interesting point is related to the charge associated to the Hꢀatom. It is noteworthy that less positive
charge values are consistently observed along the excited state surface with respect to the ground
state (purple values). This can be rationalized on the basis of a strong negative charge localization at
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
the C3 atom (the dicoordinated carbon) of IIb*, counterbalancing the electronegativity of the
oxygen atom.
Figure 4. Reaction profile for the transfer of the Hꢀatom from HꢀOCH3 to the dicoordinated carbon
1
of IIb on the ground state (solid symbols) and the excited state (open symbols) surfaces
(E0(CASSCF,MeOH) data have been reported, as detailed in Table S23). ESP charges (qESP) for HꢀOCH3
and the dicoordinated carbon (C3) on relevant points (first and last points) along the reaction path
are shown in purple and grey, respectively, as derived from the results of calculations at the CPCMꢀ
CASSCF/6ꢀ31G(d,p) level of theory in MeOH. The green circle indicates the position of the
16
ACS Paragon Plus Environment

Page 17 of 38
The Journal of Organic Chemistry
1
2
3
optimized structure (shown in the inset along with some relevant parameters) of the complex
4
5
between 1IIb* and MeOH.
6
7
8
9
However, literature data indicates that zwitterionic 1IIb* should lie ca. 30ꢀ40 kcal molꢀ1 (depending
on the adopted theoretical approach) above the ground state,22 much lower in energy than what
found here. This could be due to our theoretical approach, based on single point calculations, thus
providing vertical excitation energies and not a realistic scenario of the potential energy surface.
Accordingly, we tried to optimize the structure of the excited DHT (1IIb*) in the presence of a
MeOH molecule and, to our delight, we found a complex based on an interaction between the OꢀH
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
moiety and the dicoordinated carbon of IIb*. The energy of this structure has been reported in
Figure 4 (see the green circle) and was found to be much lower in energy, around 36 kcal molꢀ1
above the ground state, in accordance with the literature.22 The structure of this complex (see
geometry in Figure 4) also deserves some comments. First of all, the HꢀOCH3 and C3 carbon
distance is around 2.0 Å, clearly highlighting an interaction between the two sites. Furthermore, the
DHT skeleton is perfectly planar, but the hexagon ring is distorted, with a C2ꢀC3ꢀC4 angle around
112°, and protrudes towards the MeOH molecule. Interestingly, this optimization also highlighted
the formation of a partial negative charge (ꢀ0.48) at the dicoordinated carbon, confirming the
different electronic distribution observed with respect to the ground state.
The above data do not account for the varied reactivity observed, since almost all of the isomeric
DHTs offer very similar behaviors in the simulation system adopted. At the same time, the
involvement of the zwitterionic excited state, despite being low in energy, seems not feasible.
However, there are sparse reports in the literature highlighting the role of the solvent in tuning the
reactivity of some intermediates, in particular of carbenes.27 We mentioned above that the
dicoordinated carbon (phenyl radical site) of ground state DHTs does not interact with the solvent.
We surmised, however, that the situation could be different in the case of the benzylic position.
Accordingly, we exposed the Cα atom of 1,3IIb to the hydroxyl group of MeOH and tracked the
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 38
1
2
3
4
5
6
7
8
9
effect of modifying the CαꢀO distance on the resulting complex (see Section 2.6 in Supporting
Information). The results are gathered in Figure 5 and it is important to highlight a marked
difference between the singlet and the triplet state. In the latter case, we observed a monotonic
behavior, with the energy increasing with the decrease of the CαꢀO distance. Furthermore, the
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
electronic structure of IIb was not affected when approaching the hydroxyl group to the benzylic
1
carbon. By contrast, in the ground state IIb, both the geometry and the electronic structure were
heavily affected. Three different stages can be observed. The first one concerns with a CαꢀO
distance ≥ 2.0 Å, where an energy variation around + 30 kcal molꢀ1 was observed and the DHT
structure remained planar with a peculiar diradical character (σ1π1 configuration, with two unpaired
electrons and no charge localization). In the second region (1.7 Å < CαꢀO distance < 2.0 Å), the
geometry started to deviate from planarity, with the C2ꢀC3ꢀC4 moiety acquiring a sort of cumulenic
character, and a partial negative charge developed at C3. In the inset of Figure 5, the structure of
point A has been reported, showing the highly distorted geometry observed in this region and
highlighting the (almost) full negative charge located at C3. In the third region (CαꢀO distance ≤ 1.6
Å), a sharp discontinuity was observed (see point B) and the DHT structure assumed again a planar
1
geometry, strictly resembling that reported above for excited state IIb* (also showing a marked
negative charge at the dicoordinated carbon).
18
ACS Paragon Plus Environment

Page 19 of 38
The Journal of Organic Chemistry
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 5. Energetic profile for the interaction of HꢀOꢀCH3 with the Cα atom of 1,3IIb
(E0(CASSCF,MeOH) data have been reported, as detailed in Table S25). ESP charges (qESP) for the
dicoordinated carbon (C3) on relevant points are shown in grey, as derived from the results of
calculations at the CPCMꢀCASSCF/6ꢀ31G(d,p) level of theory in MeOH. The structure shown
corresponds to that of point A.
Figure 6. Energetic profile for the TSs describing the transfer of HꢀOCH3 to the C3 atom of 1IIb as
a function of the distance of the Cα atom to the hydroxyl group of a second MeꢀOꢀH molecule
(E0(CASSCF,MeOH) data have been reported, as detailed in Table S26). ESP charges (qESP) for the
exchanged HꢀOCH3 and the C3 atom on relevant points are shown in violet and grey, respectively,
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 38
1
2
3
as derived from the results of calculations at the CPCMꢀCASSCF/6ꢀ31G(d,p) level of theory in
4
5
MeOH. The structure shown corresponds to that of point A.
6
7
8
9
A final simulation was undertaken to see the effect of the interaction of MeOH at the benzylic
position on the TS describing the Hꢀatom transfer from HꢀOCH3 to the dicoordinated carbon. The
energetic profile is reported in Figure 6, where the energy of the TSs have been reported.
Interestingly, the profile was similar to that observed in Figure 5 and different regions can be
extrapolated. In the first part (CαꢀO distance ≥ 2.0 Å), the energy of the TSs increased with the
decrease of the CαꢀO distance, while small variations were observed both in terms of geometry (the
DHT structure remains planar) and of charges associated to both the exchanged Hꢀatom and C3 (a
small positive value has been found in the latter case). By contrast, for CαꢀO distance < 2.0 Å, the
energy of the TSs fell down and this discontinuity was accompanied by a marked change in terms of
charges, particularly for the case of the C3 atom, that acquired a partial negative charge (up to ꢀ
0.39). As a result, the energy of the TS when the CαꢀO distance was 1.7 Å, was comparable to that
observed when the MeOH at the benzylic position was not interacting with the DHT (< 0.5 kcal
molꢀ1 difference; see the structure of point A reported in the inset of Figure 6). It is worth noting,
however, that the course of the Hꢀatom transfer step was completely different, as testified by the
different charges of C3 involved.
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
Discussion
The photoinduced generation of cations/diradicals from the examined chlorobenzylsilanes 1aꢀc was
carried out at room temperature and the mass balance of the products formed was decent in most
cases. The products structure and distribution leave no doubt about the nature of the intermediates
involved. As an example, the yields of products obtained in the photoinduced reactions of 1b are
comparable to those reported for reactions occurring via α,3ꢀDHT under thermal conditions.13 As
mentioned in the Introduction, the generation of DHTs reported here is a single photon, two steps
20
ACS Paragon Plus Environment

Page 21 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
process. The general features of the photoinduced heterolysis of chlorobenzenes via the triplet state
are well known,28,29 as is the fact that a nucleophilic solvent (such as water) stabilizes the
intermediate triplet phenyl cations,19,23,24,30 thus favoring the occurrence of its peculiar reactions
(e.g. hydrogen abstraction from the medium). However, cleavage of the CꢀSi bond may take place
competitively, leading to the corresponding DHTs. Two intermediates were then mainly responsible
of the variety of arylated derivatives formed, viz. the triplet cation (3VI+) and the DHT (II), as
apparent from Tables 1ꢀ3.
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
From triplet phenyl cations to
α,n-DHTs. Competing processes. The consumption of 1aꢀc was
significant (apart selected cases) in all of the protic media tested herein (Tables 1, 2), also when
additives such as tetrachloromethane or 1,4ꢀcyclohexadiene were present (Table 3). The desired
desilylation to α,nꢀDHTs suffered, however, of competing chemistry28,29,31 from the short lived (a
lifetime of 10 ns could be estimated)32 triplet phenyl cations (3VI+aꢀc, Scheme 8).
Accordingly, hydrogen abstraction from the solvent (Scheme 8, path b) took place significantly with
the orthoꢀ and metaꢀ isomers. In these cases, the yield of benzylsilane 2 was in the 20 to 50% range,
roughly following the BDE of the CꢀH bonds of the medium. Shifting from CH3OH to CD3OH led
to the formation of 2ꢀd1 supporting the involvement of a H(D) transfer from the CH3/CD3 group. In
the case of the orthoꢀ and metaꢀ isomers 1b,c, a minor process occurred since ISC from the firstꢀ
3
formed triplets VI+b,c to the more stable singlet cations (1VI+b,c)23a allowed the isolation of
1
solvolysis products 8b,c, 11b,c and 13b (paths c, d). The generation of VI+b, however, opened an
1
interesting route to singlet diradical IIb (path e'). On the contrary, with the paraꢀ isomer 1a,
processes competitive with desilylation (path e) were always negligible, and reduction occurred
only as a minor path (0ꢀ10%) even in the presence of excellent Hꢀdonors, such as CHD.
DHT chemistry. According to what observed in our previous computational and experimental
3
1
investigations,23a desilylation of triplet VI+aꢀc and singlet (only for VI+b) phenyl cations to the
3
1
corresponding IIaꢀc or IIb, respectively, (Scheme 8, paths e, e') is a feasible process in protic
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 38
1
2
3
4
5
6
7
8
9
media. Based on the present and previous computational evidences,23a DHTs are heterotopic
diradical species (σ1,π1), with the two unpaired electrons not interacting, independently from the
isomer and the spin state considered. However, they exhibit a different reactivity as demonstrated
by the formation of polar products (such as benzyl alcohols and ethers) from α,3ꢀDHT, initially
assigned to a ionic mechanism rather than to a radical path.18,20,22 Data in our hands suggest that all
bare (i.e. not influenced by the medium where they are generated) DHTs are "purely diradical"
species behaving in a similar way, and reaction occurs from the lowestꢀlying spin state for each
isomer considered (see the bold data in Table 4). The more reactive site of the diradicals is the
σ center, since the π center is stabilized by resonance.26 Thus, hydrogen transfer (HT, Scheme 9,
path a) operated by the phenyl radical site26 from the solvent (ethers, acetonitrile and alcohols), as
well as from CHD, was the first step and occurred preferentially at the CꢀH site.
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
Scheme 8. Reactivity of Triplet Phenyl Cations 3VI+aꢀc Photogenerated from 1aꢀc.
22
ACS Paragon Plus Environment

Page 23 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
The capability of aryl radicals in HT from the surrounding environment has been deeply
investigated in the literature.33 In alcohols, the rate constant for CꢀH cleavage occurring in primary
CꢀH (BDE = 100.0±2.0 kcal molꢀ1)34 of tertꢀbutanol was measured to be up to 30 times slower than
from labile αꢀoxy CꢀH bond (91±1.0 kcal molꢀ1)34 in isopropanol35 for which CꢀH cleavage
occurred exclusively. On the other hand, in methanol both the CꢀH (96.0±0.1 kcal molꢀ1)34 and OꢀH
(104.2±0.9 kcal molꢀ1)34 bonds were cleaved with a ratio of ca. 9:1.36 Analogously, in ethanol and
tBuOH, a competitive hydrogen transfer from OꢀH and CꢀH bonds occurred, depending on both
statistics and thermodynamic effects. Moreover, hydrogen abstraction by phenyl radical was
described as slightly exothermic (4.2 kcal molꢀ1) also from water and observed in a few experiments
at lowꢀtemperature in argon matrices.37
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 computational data reported in Table 4 highlight that DHTs can abstract efficiently both
electrophilic (from CH3CN, BDE CꢀH = 96 kcal molꢀ1)34 and nucleophilic (from the CꢀH positions
of alcohols) hydrogens thanks to the reactivity imparted by the phenyl radical site. Experimental
data in our hands pointed out that αꢀhydrogens in ethers (e.g. THF, BDE = 92.1±1.6 kcal molꢀ1)34
were selectively extracted. Turning to the CꢀH vs OꢀH competition for the case of CH3OH, very
similar activation parameters have been found in the present simulation, making a generalization
unfeasible. This topic has been already debated in the literature, since theoretical simulations,
showing a preference for the OꢀH abstraction, were apparently in disagreement with experimental
data.22 The reason for this failure was however found in an inaccurate handling of hydrogen
bonding by the simulation adopted and ruled out the possibility that the α,3ꢀDHT diradical could
react at the OꢀH site.22 We propose, however, that all DHTs can promote both reactivities through a
homolytic mechanism, with a preference towards the CꢀH vs the OꢀH site (see data gathered in
Table 4). This is evident by analyzing the data for 1a, where we can surmise that the only
responsible for the observed DHTꢀderiving products is 3IIa. Actually, we observed a decrease of the
radical/polar products ratio when the BDE of the CꢀH bond of the solvent increases. This is
particularly clear if we consider the results from the experiments carried out in tBuOH and in
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 38
1
2
3
4
5
6
7
8
9
CD3OH. The experiments in deuterated solvents further support the fact that HT occurs from the Oꢀ
H or the CꢀH, respectively, in the formation of polar and radical products, according to the
deuteration pattern observed on the aromatic ring. As depicted in Scheme 9, this HT event leads to
benzyl radicals (path a), that upon coupling with the radicals arising from the solvent, are
responsible for the formation of alcohols 5, 7, 9, 12, 14, benzyl ethers 6, 10, 15 and products 16ꢀ20
(path b). Benzyl radicals can also dimerize to bibenzyl 4 (path c), as well as abstract a hydrogen
(path d) to yield toluene 3. HT also occurs from 1,4ꢀCHD, and coupling of the generated radicals
affords alkyl benzene 19.
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
Scheme 9. Reactivity of Photogenerated α,nꢀDHTs.
H
O
R
H
O R
+
f
H
1IIbC
-
RO-H
6-7, 10, 15
H
e
only for
R-O-H 1IIb
R
H
H
C-H or O-H
H abstraction
H
b
CH2R
H
a
1,3IIa-c
H
H
R'-H
a'
5-7, 9-10,
12, 14-15,
16-20
CCl4
c
d
3
4
H
H
b'
21a-c
CCl3
Cl
Chlorine abstraction from CCl4 followed by coupling of the resulting benzyl/Cl3C• radical pair
(paths a', b') affords tetrachlorinated aromatics 21aꢀc18a (notice that radical products 4 and 5 are not
formed in the presence of such additive). However, though both polar and radical products can be
observed via a radical pathway, the latter will be predominant, since homolytic cleavage of a CꢀH
bond always requires a lower energy than to cleave OꢀH bond.
24
ACS Paragon Plus Environment

Page 25 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Very similar conclusions can be drawn for 1c, while the product distribution is completely different
for 1b, despite the similar computational results obtained for IIb. In our opinion, this different
behavior can be ascribed to the fact that IIb is the only isomer showing a singlet ground state.
Indeed, the peculiar behavior of singlet DHTs has been previously shown to have some generality,
as demonstrated in the case of methoxyꢀsubstituted α,nꢀDHTs.23b Turning to the computational
results, it is worth noting that the pathways previously described apply for bare DHTs only (Scheme
9, path a), while the situation may be different if a suitable model able to appreciate specific
interactions with the nucleophilic solvent is considered. Thus, forcing the MeOH molecule to
coordinate at the benzylic site causes a charge separation in 1IIb, tagged as complexed 1IIb (1IIbC in
Scheme 9) with the development of a (partial) negative charge at the dicoordinated carbon, and
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
renders it similar to the first excited state IIb*. It is no surprise that both Figures 5 and 6 show a
sharp discontinuity, because in both cases the approach of the lone pair of oxygen of MeOH to the
benzylic position causes an electronic reorganization, ultimately leading to the population of the
1
excited state of IIb (and in turn causing the observed discontinuity). In our opinion, however, this
is not the most important part of Figures 5 and 6, while it should be highlighted the behavior
observed in the points with a CαꢀO distance in the 1.8ꢀ2.0 Å range, where it is demonstrated that a
1
charge separation in IIb is actually feasible. The present model is obviously approximated and
shows that such coordination occurs with an energy increase. However, we believe that several
MeOH molecules could concur to this interaction (in turn, facilitating it) and that a continuum of
situations may arise (without arriving at the extreme case involving the formation of the excited
state), where the development of the negative charge is strictly related with the extent of the
interaction of the solvent at the benzylic position. Such chargeꢀseparated species, as previously
suggested,22 can form polar products (such benzyl ethers 6, 10, 15 and benzyl alcohol 7) through an
ionic addition mechanism. Furthermore, it is interesting to highlight that such polarization can
operate on singlet species only, in strict accordance with experimental results. A zwitterion
reactivity is thus effectively present in the case of chloride 1b. Thus, in the presence of H donors
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 38
1
2
3
4
5
6
7
8
such as isopropanol or CHD, a radical behavior arising from bare 1IIb is the main or even the only
observed, whereas, when using tBuOH/water mixture as the medium as well as in the presence of
1
deuterated CD3OH, we have a chance to observe the reactivity of IIbC, which afforded polar
9
products in amounts larger than those observed for other isomers. Similarly, Carpenter et al.
proposed, in the case of thermally generated α,3ꢀDHT, a nonadiabatic surface crossing from the
ground state to the first excitedꢀstate zwitterionic singlet surface.22
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 similar behavior was described by Sander when investigating the ground state of
diphenylcarbene.27a However, in that case a hydrogen bonding with the solvent caused an inversion
of the ground state of the intermediate, whereas in the present investigation the nucleophilic site of
the medium caused the formation of a chargeꢀseparated species, strictly resembling the effect of the
1
excitation from S0 to S1 in IIb. The nucleophilic addition to a radical center in other aromatic
diradicals has been also investigated (by having recourse to both computational and experimental
analyses) as in the case of paraꢀbenzynes, in order to explain the biosynthetic pathway to
cyanosporasides,10 where the formal addition of HCl onto a generated 1,4ꢀdidehydrobenzene has
been observed. In that case, the reaction took place with an energy barrier largely due to the
solvation of the nucleophile, and nucleophilic addition was followed by protonation.
Conclusions
Since the discovery of the MyersꢀSaito cycloaromatization of enyneꢀallenes, the competing
diradical/zwitterion reactivity of didehydrotoluenes has been the object of detailed investigations,
most of them limited by the exclusive accessibility of the α,3ꢀisomer under thermal conditions. The
photochemical approach adopted in the present study allowed the generation of all of the DHT
isomers and highlighted the dependence of the reactivity of diradicals on the actual spin state
involved. Thus, bare α,nꢀDHT intermediates exhibit a unique diradical reactivity responsible for the
formation of radical and polar products (the latter in a minor amounts), whereas interaction of
26
ACS Paragon Plus Environment

Page 27 of 38
The Journal of Organic Chemistry
1
2
3
4
5
6
singlet α,3ꢀisomer with the solvent resulted in the formation of a polarized species (termed
“complexed DHT”), which allowed for the formation of polar compounds.
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
Experimental Section
General Considerations. 1H NMR spectra were recorded with a 300 MHz spectrometer, while 13C
NMR spectra were recorded with a 75 MHz spectrometer. The attributions were made on the basis
1
of H and 13C NMR, as well as DEPTꢀ135 experiments; chemical shifts are reported in ppm
downfield from TMS. GCꢀMS analyses were carried out using a Thermo Scientific DSQII single
quadrupole GC/MS system. A Restek Rtxꢀ5MS (30 m × 0.25 mm × 0.25 ꢁm) capillary column was
used for analytes separation with helium as carrier gas at 1 mL minꢀ1. The injection in the GC
system was performed in split mode and the injector temperature was 250 °C. The GC oven
temperature was held at 60 °C for 5 min, increased to 250 °C by a temperature ramp of 10 °C minꢀ1
and held for ten min. The transfer line temperature was 270 °C and the ion source temperature 250
°C. Mass spectral analyses were carried out in full scan mode.
The employed solvents (acetonitrile, water, methanol, ethanol and isopropanol) were of HPLC
purity. tertꢀButanol, tetrahydrofuran, 1,4ꢀdioxane, CCl4 and 1,4ꢀcyclohexadiene (reagent grade)
were freshly distilled before use. Compounds 1aꢀc19 have been synthesized by following a known
procedure. The amounts of 2ꢀ7, 8b,c25 9, 10,38 12, 14, 15,39,40 18 and 2041 have been determined on
the basis of calibration curves by comparison with either commercial standards or synthesized
compounds. Compounds 13b, 19,18a 21a and 21c have been identified and quantified on the basis of
GCꢀMS analyses (see Table S1 for further details).
(3-Isopropyloxybenzyl)trimethylsilane (11b). The title compound was prepared by adapting a
procedure described for the synthesis of 5ꢀmethoxyꢀ1,3ꢀbenzodioxole.42 A mixture of 8b25 (1.0 g,
5.5 mmol) and K2CO3 (0.40 g, 2.9 mmol) in acetone (34 mL) was stirred for 30 min, then 2ꢀ
bromopropane (42 mmol) was added dropwise and the resulting mixture was refluxed for 48h. After
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 38
1
2
3
4
5
6
7
8
9
cooling to room temperature, water (20 mL) was added and the mixture extracted with diethyl ether
(3×20 mL), the organic phases reunited, dried over MgSO4 and then concentrated. Purification of
the crude product by column chromatography (eluant: neat cyclohexane) afforded 0.13 g of 11b
(oil, 11% yield). 11b: 1HꢀNMR (CDCl3, 300 MHz) δ 0.05 (s, 9H), 1.35ꢀ1.40 (d, 6H, J = 6 Hz), 2.08
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
(s, 2H), 4.50ꢀ4.55 (m, 1H), 6.55ꢀ6.60 (m, 3H), 7.10ꢀ7.15 (m, 1H). 13CꢀNMR (CDCl3, 75 MHz) δ ꢀ
1.9 (CH3), 22.0 (CH3), 27.1 (CH2), 69.6 (CH), 111.4 (CH), 115.7 (CH), 120.4 (CH), 128.9 (CH),
142.0, 157.7. IR (ν, cmꢀ1): 3028, 2956, 1597, 1485, 1258, 1157, 851. MS (m/z): 222 (M+, 5), 189
(10), 132 (50), 117 (55), 104 (38), 91 (26), 73 (100).
1-Chloro-3-(2,2,2-trichloroethyl)benzene (21b). Compound 21b was obtained by adapting a
procedure for the synthesis of 20.41 CHCl3 (1.9 mL, 24 mmol) was added dropwise to a cooled (0
°C) solution of NaH (60% dispersion in mineral oil, 0.5 g, 12 mmol) in dry DMF (30 mL). 3ꢀ
Chlorobenzylbromide (0.5 mL, 3.8 mmol) was thus added dropwise and the solution stirred for 15
h. The reaction mixture was poured over ice (50 g) and extracted with hexane (3×50 mL). The
organic phases were reunited, dried over MgSO4 and concentrated. Purification of the crude product
by column chromatography (eluant: neat hexane) afforded 0.12 g of 21b43 (oil, 13% yield). 21b: 1Hꢀ
NMR (CDCl3, 300 MHz) δ 3.93 (s, 2H), 7.30ꢀ7.40 (m, 4H), 7.45ꢀ7.50 (m, 1H). 13CꢀNMR (CDCl3,
75 MHz) δ 59.1 (CH2), 98.1, 128.5 (CH), 129.3 (CH), 129.7 (CH), 131.5 (CH), 133.9, 134.9. IR (ν,
cmꢀ1): 3067, 2926, 1575, 1082, 955, 805. MS (m/z): 244 (M+, 12), 207 (10), 136 (9), 127 (32), 125
(100), 89 (8), 75 (7).
Photochemical experiments. The photochemical reactions were performed by using nitrogenꢀ
purged solutions of 1aꢀc (0.025 M) in the chosen medium in quartz tubes. Irradiations were
performed in a multiꢀlamp reactor fitted with ten 15W phosphorꢀcoated lamps (maximum of
emission at 310 nm). The reaction course was followed by means of GC and GCꢀMS analyses.
Workup of the irradiated mixtures involved concentration in vacuo and chromatographic separation
28
ACS Paragon Plus Environment

Page 29 of 38
The Journal of Organic Chemistry
1
2
3
4
5
by using silica gel as stationary phase. Quantum yields were measured at 254 nm (1 Hg lamp,
15W).
6
7
8
9
Irradiation of 1a in a tetrahydrofuran/water 2:1 mixture. A degassed solution of 1a (149 mg, 0.75
mmol, 0.025 M) in a THF/H2O 2:1 mixture (30 mL) was irradiated for 15h at 310 nm. The
photolyzed solution was then extracted with diethyl ether (3×20 mL), the organic phases reunited,
dried over MgSO4 and concentrated. Purification of the residue by column chromatography (eluant:
neat hexane) afforded 74 mg of 2ꢀbenzyltetrahydrofuran (16, 61% yield). GC analysis of the
irradiated solution revealed also the presence of benzyltrimethylsilane (2, 10% yield) and 1,2ꢀ
diphenylethane (4, 11% yield). Spectroscopic data of 16 were in accordance with the literature.44
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
Irradiation of 1a in a 1,4-dioxane/water 2:1 mixture. A degassed solution of 1a (149 mg, 0.75
mmol, 0.025 M) in a 1,4ꢀdioxane/H2O 2:1 mixture (30 mL) was irradiated for 15h at 310 nm. The
photolyzed solution was then extracted with diethyl ether (3×20 mL), dried over MgSO4 and
concentrated. Purification of the residue by column chromatography (eluant: neat hexane) afforded
88 mg of 2ꢀbenzylꢀ1,4ꢀdioxane (17, 66% yield).45 GC analysis of the solution revealed also the
1
presence of benzyltrimethylsilane (2, 4% yield) and 1,2ꢀdiphenylethane (4, 10% yield). 17: Hꢀ
NMR (CDCl3, 300 MHz) δ 2.60ꢀ2.80 (AB part of an ABX system, 2H), 3.35ꢀ3.40 (X' part of an
A'B'X' system, 1H), 3.65ꢀ3.85 (m, 6H), 7.20ꢀ7.35 (m, 5H). 13CꢀNMR (CDCl3, 75 MHz) δ 38.3
(CH2), 66.3 (CH2), 66.8 (CH2), 70.8 (CH2), 76.0 (CH), 126.4 (CH), 128.3 (CH), 129.1 (CH), 137.3.
IR (ν, cmꢀ1): 3028, 2596, 1122, 700. MS (m/z): 178 (M+, 3), 117 (6), 92 (10), 91 (55), 87 (100), 86
(68), 65 (15), 59 (7), 43 (16).
Computational Methods. All the calculations were carried out using the Gaussian 09 program
package.46 In our investigation, the optimization of all of the stationary points was carried out with
the CASSCF (Complete Active Space SelfꢀConsistent Field) method by using the standard 6ꢀ
31G(d,p) basis set. No symmetry constraint was applied to the investigated structures. The
occupancies of the orbitals included in the active space were carefully checked, and the values
29
ACS Paragon Plus Environment