Unusually Facile Thermal Homodienyl-[1,5]-Hydrogen Shift Reactions in Photochemically Generated Vinyl Aziridines

Article abstract of DOI:10.1002/chem.201600479

A range of photochemically generated tri- and tetracyclic vinyl aziridines have been found to undergo a general and surprisingly low temperature ring opening through a [1,5]-hydrogen shift reaction. The rate of the process was found to be highly dependent on the structure and substitution around the azirdine ring and the alkene terminus, with some substrates being observed to undergo ring opening at temperatures as low as 25 °C. The rigid nature of these polycyclic systems precludes a conformational explanation of these rate differences, and an Eyring study confirmed a negligible entropic barrier to the reaction. However, the Eyring plots for two different aziridines systems showed a significant difference in their enthalpies of activation. It is therefore believed that the levels of aziridine ring strain, as well as electronic effects, are the dominant factors in this sequence.

Full text of DOI:10.1002/chem.201600479

DOI: 10.1002/chem.201600479
Full Paper
&
Synthetic Methods |Hot Paper|
Unusually Facile Thermal Homodienyl-[1,5]-Hydrogen Shift
Reactions in Photochemically Generated Vinyl Aziridines
Jonathan P. Knowles* and Kevin I. Booker-Milburn*[a]
Abstract: A range of photochemically generated tri- and tet-
racyclic vinyl aziridines have been found to undergo a gener-
al and surprisingly low temperature ring opening through
a [1,5]-hydrogen shift reaction. The rate of the process was
found to be highly dependent on the structure and substitu-
tion around the azirdine ring and the alkene terminus, with
some substrates being observed to undergo ring opening at
temperatures as low as 258C. The rigid nature of these poly-
cyclic systems precludes a conformational explanation of
these rate differences, and an Eyring study confirmed a negli-
gible entropic barrier to the reaction. However, the Eyring
plots for two different aziridines systems showed a significant
difference in their enthalpies of activation. It is therefore be-
lieved that the levels of aziridine ring strain, as well as elec-
tronic effects, are the dominant factors in this sequence.
Introduction
Aziridines have long proved to be useful synthetic intermedi-
ates due to their ability to undergo nucleophilic ring openings
and thermal [2+3] cycloaddition reactions.[1] Vinyl aziridines
possess a further manifold of reactivity[2] as the p-bond ena-
bles the possibility of SN2’ ring processes,[3] additional cycload-
dition chemistry under Pd-catalysis[4] and Tsuji–Trost[5]-type pro-
cesses. We recently reported the novel, photochemical synthe-
sis of a range of tricyclic vinyl azridines 3 (via cyclobutanes 2)
starting from N-butenyl pyrrole precursors 1 as shown in
Scheme 1, and demonstrated that scale-up of these reactions
could be achieved in a simple manner through the use of a flu-
orinated ethylenepropylene (FEP) flow reactor.[6] Such products
possess a wealth of stereochemistry and functionality for fur-
ther derivatisation, perhaps most interestingly their potential
as dipolarophiles and the possibility of performing nucleophilic
attack at various sites.
Scheme 1. Thermal rearrangement of tricyclic aziridines.
it was quickly revealed that an alternative non-catalysed pro-
cess was operating at temperatures greater than 508C. This
was identified as a thermal homodienyl-[1,5]-hydrogen shift
(retro-ene reaction), leading to imines 4 as shown in Scheme 1.
Whereas this type of rearrangement has been extensively stud-
ied in vinylcyclopropanes,[8] such reactions have less frequently
been reported to occur in vinylaziridines. Somfai[9] has studied
the generality of this type of process with conformationally
flexible vinylaziridines, and there have been other infrequent
reports of this rearrangement, often as a competitive side reac-
tion.[10]
This potential to generate a high degree of structural com-
plexity in only two steps, combined with ease of scale-up,
prompted us to investigate a number of reaction possibilities.
Recent work in our laboratories[7] has demonstrated that these
complex vinyl aziridines indeed undergo a variety of efficient
Pd-catalysed ring-opening processes under unusually mild pro-
cesses for non-activated (i.e., at N) aziridines. During this study,
[a] Dr. J. P. Knowles, Prof. Dr. K. I. Booker-Milburn
School of Chemistry
In our case, it was significant to note that these reactions
would proceed smoothly at relatively low temperatures;
indeed, further studies showed it to operate as low as ambient
temperature with some substrates. We therefore elected to
study the scope and mechanism of this unusually facile reac-
tion and the results are reported here.
University of Bristol
Cantock’s Close, Bristol, BS8 1TS (UK)
E-mail: k.booker-milburn@bristol.ac.uk
jon.knowles@bristol.ac.uk
Supporting information for this article including supplementary figures and
tables, full experimental procedures, analytical data, kinetics data, as well
1
as copies of H and 13C NMR data is available under http://dx.doi.org/
10.1002/chem.201600479.
Chem. Eur. J. 2016, 22, 11429 – 11434
11429
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
to confirm that a standard, concerted mechanism was still op-
erating and also to gain a better understanding of the reason
for this large variation in reaction rate.
Results and Discussion
To carry out a broad study of this reaction, we needed to syn-
thesise a diverse range of substituted pyrroles and investigate
their subsequent photochemistry. Substrates were selected
based on the availability of the starting pyrroles as well as the
success of the photochemical step. The substituted pyrroles
were synthesised using standard heterocyclic techniques and
were then irradiated at 254 nm. We also took the opportunity
to explore, for the first time, the photochemistry of more com-
plex tetrahydroindolone-derived starting materials,[11] which
gave access to the more elaborate tetracyclic aziridines 5 and
7 upon irradiation at 254 nm (see Supporting information for
full details).
To this end, the reactions in Table 1 reactions were per-
formed in [D8]toluene at 1008C and the progress monitored
using a 500 MHz NMR with 1,3,5-trimethoxybenzene as an in-
ternal standard. An initial screen of concentrations for sub-
strates 15 and 19 (entries 6 and 8) showed the reactions to be
first order. All subsequent data was found to fit first-order ki-
netics, and plots of ln[substrate] versus t gave the rate con-
stants shown in Table 1.[13] For substrate 9, the reaction was so
rapid at 1008C, it was repeated at 808C to generate a second,
more accurate rate constant. Substrates 5 and 7 were found to
be too reactive to be monitored at 1008C, and were followed
at the temperatures specified.
With this diverse range of aziridines in hand, we then stud-
ied their thermal rearrangement (Table 1). As can be seen, the
[1,5]-hydrogen shift reaction occurs in high yield for all sub-
strates, tolerating a range of functional groups (ester, amide,
nitrile and ketone). The yield remains high even upon substitu-
tion of the double bond (entries 7–13), with only two examples
(entries 9 and 10) displaying minor products from side reac-
tions.[12] In every case, a single diastereomer was produced. Re-
arrangement of compound 27 (entry 12) was also successful,
indicating that deuterium could be transferred as well as hy-
drogen, albeit at a reduced rate compared to undeuterated 19
(entry 8, vide infra). Compounds 21, 23, 25, 27, 29 and 31
gave complete stereospecificity at the newly formed tertiary
centre, and the relative stereochemistry of these products was
The results show some clear trends, which deserve further
comment. The most reactive substrates proved to be the tetra-
cyclic species 5 and 7, with the next most reactive being 9,
which possessed electron-withdrawing groups (EWG) on both
aziridine ring carbons. Although these high reaction rates were
in some ways unsurprising as all three of these substrates
were seen to rearrange upon standing at room temperature,
this degree of reactivity appears not to have been previously
documented in any other class of vinyl aziridine.[14] Substrates
with an EWG conjugated to the double bond of the starting
material (entries 4 and 5) were seen to be second fastest, with
the extra EWG increasing the rate by a factor of 5 (i.e., entry 5
versus entry 8). Aziridines with a single EWG (entries 6–8) fol-
lowed this pattern and reacted more slowly again. Slower still
were those substrates possessing a methyl group at the migra-
tion terminus position (entries 9–13), all of which proceeded
extremely slowly. It can also be seen that transfer of deuterium
rather than hydrogen (entries 8 and 12) slows the reaction sub-
stantially (vide infra).
1
proven through the H NMR NOE studies. This stereospecificity
is indicative of a concerted and facially selective thermal [1,5]-
hydrogen shift, and this view was supported by subsequent la-
belling studies.
Whereas this screen showed the reaction to be highly gener-
al, we were interested to observe cases (entries 13 and 14)
where more forcing conditions than toluene at reflux were re-
quired. This reinforced previous observations in which at-
tempts had been made to perform chemistry on such aziri-
dines; the conversion of aziridine 15 to imine 16 was first
noted during an attempted nucleophilic ring opening at 508C,
whereas aziridine 29 had been found to be unreactive under
identical conditions. Remarkably, substrates 5, 7 and 9 were
seen to undergo conversion to the imine product at room
temperature. This difference in reaction rate was intriguing,
both from a theoretical point of view and because a better un-
derstanding of the factors accelerating this reaction might
help minimise its competition as a side reaction in other trans-
formations.[7] Additionally, previous studies, most notably those
of Somfai,[9a,b] had generally shown this type of [1,5]-hydrogen
shift to be facile only when an activating group (e.g., ester,
phenyl) was present at the migration origin. Subsequent stud-
ies by Somfai of the [1,5]-hydrogen shift reactions of various
substituted, conformationally flexible vinyl aziridines showed
relatively small differences in reaction rate unless the alkene
geometry was varied.[9c] These variations in rate were largely
ascribed to steric effects on conformation, which we felt were
unlikely to have a large impact in rigid systems such as ours.
Given this, we elected to perform a more detailed study, both
The effect of solvent polarity on the rate of reaction was of
interest to us, and substrate 15 was selected to perform a brief
solvent screen (Table 2). It can be seen that increasing solvent
polarity leads to a small but significant increase in reaction
rate, likely consistent with a small degree of charge formation
in the transition state.
We then proceeded to investigate the incorporation of deu-
terium into the reacting aziridines. Deuterated substrates 33a,
33b and 34 were synthesised through alkylation of the pyrrole
with the appropriate tether using either phase transfer or Mit-
sunobu reaction conditions. The tethers themselves were syn-
thesised by reduction of the appropriate ester/aldehyde using
LiAlD4 (see Supporting Information for full details).
Thermolysis of bisdeuterated amide 34 showed that the
transfer of deuterium rather than hydrogen slowed the reac-
tion significantly, with a kH/kD of 4.0 (Table 3). However, al-
though a primary kinetic isotope effect (KIE) is clearly present
here, the presence of two deuterium atoms makes the ob-
served effect a combination of both this and a secondary ki-
netic isotope effect. Additional studies on a diasteroemeric
mixture of monodeuterated 33 allowed us to remove this com-
plication by following the reaction of each diastereomer sepa-
rately. The chemical shifts for the exo/endo-hydrogen atoms of
Chem. Eur. J. 2016, 22, 11429 – 11434
www.chemeurj.org
11430
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Yields and rate constants for the [1,5]-hydrogen shift reactions.
[b]
Entry
Substrate
Product
Yield [%][a]
T [8C]
Concentration [m]
Rate, k [sÀ1
]
Relative rate
1
80
80[c]
0.144
1.8Æ0.1”10À2
2520
1860
2
89
80[c]
0.235
1.33Æ0.02”10À2
100
80[d]
0.121
0.121
2.25Æ0.03 ”10À3
7.59Æ0.05 ”10À4
315
106
3
4
67
85
100
100
0.136
0.196
4.32Æ0.06”10À4
4.28Æ0.02”10À4
60.4
5
83
59.9
25.0
100
100
100
0.084
0.14
0.28
1.76Æ0.01”10À4
1.79Æ0.02”10À4
1.80Æ0.02”10À4
6
7
8
100
90
100
0.093
1.31Æ0.01”10À4
18.3
11.3
100
100
0.25
0.13
8.05Æ0.02”10À5
8.00Æ0.05”10À5
98
9
71[e]
100
0.10
7.26Æ0.09”10À5
10.2
10[f]
71[e]
100
100
0.10
<7”10À5
<10
11
70
0.099
6.4Æ0.1”10À5
9.0
2.7
[g]
12
–
100
0.094
1.93Æ0.02”10À5
13
14
74
90
100
100
0.24
0.20
1.63Æ0.01”10À5
7.15Æ0.08”10À6
2.3
1
[a] Isolated yield after chromatography. [b] See Supporting Information for concentration profiles and log plots. Errors calculated using a least squares anal-
ysis. [c] Reaction proved too fast to follow at 1008C. [d] Reaction also performed at 808C to obtain a more accurate rate constant. [e] Yield reduced by
competing elimination reaction.[12] [f] The rate constant for this substrate represents the sum of both the desired rate constant and that of the competing
elimination reaction.[13] [g] Product not isolated.
Chem. Eur. J. 2016, 22, 11429 – 11434
www.chemeurj.org
11431
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
why some of these species were sufficiently reactive to under-
go the rearrangement at room temperature. We hypothesised
that a combination of ring strain and a highly rigid conforma-
tion was responsible. Eyring plots have previously been em-
ployed to probe similar systems,[15] and we applied this ap-
proach to substrates 5 and 13. These were chosen as they re-
flect the most reactive substrate (5) and a substantially less re-
active substrate (13) that still underwent the rearrangement at
Table 2. Solvent screen for the [1,5]-hydrogen shift reaction of amide 15.
[a]
Entry
Solvent
k [sÀ1
]
Relative rate
1
2
3
[D8]toluene
[D5]nitrobenzene
[D6]DMSO
1.79Æ0.01”10À4
2.79Æ0.06”10À4
3.66Æ0.05”10À4
1
1.58
2.07
a
range of temperatures below the boiling point of
[D8]toluene. This generated the rate constants in Table 4. It can
be seen that in the case of 5, the reaction rate is appreciable
even at room temperature, faster in fact than for some sub-
strates at 1008C (i.e., Table 1, entries 12 and 13). From the
plots of ln(k/T) versus 1/T, the enthalpies and entropies of acti-
vation were calculated (see Supporting Information) and are
presented in (Figure 1).
[a] See Supporting Information for concentration profiles and log plots.
Errors were calculated using a least squares analysis.
Table 3. Kinetic isotope effect studies.
[a]
Entry
Substrate
k [sÀ1
]
kH/kD
Table 4. Variation of rate constant with temperature for substrates 5 and
13.
1
1.79Æ0.2”10À4
2.07Æ0.04”10À4
0.41Æ0.2”10À4
0.447Æ0.03”10À4
0.86 (15/33a)
[a]
k [sÀ1
]
Entry
Substrate
T [8C]
80
60
40
25
1.8Æ0.1”10À2
2.45Æ0.03”10À3
2.37Æ0.03”10À4
3.40Æ0.01”10À5
1
2
3
4
4.6 (33a/34)
4.4 (15/33b)
0.92 (33b/34)
100
90
80
4.28Æ0.02”10À4
1.651Æ0.003”10À4
5.496Æ0.004”10À5
1.563Æ0.001”10À5
2
70
[a] See Supporting Information for concentration profiles and log plots.
Errors were calculated using a least squares analysis.
[a] See Supporting Information for concentration profiles and log plots.
Errors were calculated using a least squares analysis.
the CH2 of interest were assigned using 1H NMR TOSCY and
NOESY studies (see Supporting Information), allowing the iden-
tification of the two isomers, Dexo (33a) and Dendo (33b). Thus,
following the reaction of this diastereomeric mixture by
1H NMR spectroscopy allowed the rates of the two reactions to
be determined. The two pairs of results give an average pri-
mary KIE of 4.5 (kH/kD =4.4 for 15 versus 33b, and 4.6 for 33a
versus 34), and an average secondary KIE of 0.89 (kH/kD =0.86
for 15 versus 33a and 0.92 for 33b versus 34).
Similarly, studies involving ketone 27 were seen to give es-
sentially the same results (Table 1, entries 8 and 12), with
a combined primary and secondary kinetic isotope effect of
4.2. In addition to showing the reaction to be stereospecific
through a hydrogen-atom transfer, these results are also con-
sistent with the reaction proceeding though a concerted, al-
though somewhat asynchronous mechanism.
Figure 1. Eyring plots and thermodynamic parameters for the reactions of
substrates 5 and 13.
It can be seen that in both cases the entropy of activation is
extremely low, suggesting that minimal reorganisation is
needed to reach the reactive conformation. This is consistent
with our hypothesis that the constrained and inflexible nature
of these aziridines is a key factor in their high reactivity.
Indeed, similar Eyring analysis of the Somfai data gives a DS°
of À61 JKÀ1 molÀ1, revealing a considerably larger entropic bar-
rier.[16] This, however, does not explain the large difference in
Although these studies confirmed that these tricyclic sub-
strates reacted through the same mechanism determined by
Somfai for less constrained examples,[15] they shed no light on
why there was such a wide range of reaction rates, or indeed
Chem. Eur. J. 2016, 22, 11429 – 11434
www.chemeurj.org
11432
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
rate constants observed across the series. This owes to the
considerable variation in the enthalpy of activation, which is
nearly 20 kJmolÀ1 higher for 5 than that of 13.
It seems probable that at least part of the explanation for
this difference is the level of strain present in the aziridine ring.
For comparison, the DH° of the acyclic system investigated by
Somfai[15] is 88.9 kJmolÀ1, making this barrier somewhat less
than that of even our most reactive substrate. Clearly, the
value can reflect both the degree of ring strain and other fac-
tors, making it possible that this lower barrier is owed to the
presence of an activating group (i.e., CO2tBu) at the migration
origin.
These results also permit the calculation of rate constants
for tetracyclic substrate 5 at other temperatures. At 1008C,
a temperature at which the high reaction rate precluded mea-
surement by 1H NMR spectroscopy, this calculated rate is
0.11 sÀ1, more than 15000 times faster than that of the least re-
active substrate (Table 1, entry 14) at the same temperature.
This large variation in rate explains some of the behaviour ob-
served when attempting catalytic reactions with these species,
and clearly underlines the importance of choosing appropriate
substrates for use in the palladium-catalysed processes we re-
cently reported.
Figure 2. X-ray crystal structure of compound 17.[17]
duced reaction rates, although it remains unclear whether this
is an entropic or enthalpic effect.
Conclusion
Although this additional mode of reactivity has the potential
to cause complications when attempting other transformations
on these aziridines, we have shown that it is a general and syn-
thetically useful process in its own right. In addition, the kinetic
data shows that certain structural features accelerate this ther-
mal process, enabling better design of photochemical prod-
ucts for further catalytic transformation.
In conclusion, we have shown that the [1,5]-hydrogen shift re-
actions of photochemically generated tricyclic aziridines is
a general process that frequently occurs at unusually low tem-
perature. Such reactions provide easy and stereospecific access
to a wide range of highly functionalised bicyclic and tricyclic
imines. There is considerable variation in reaction rate, and
several reactions are unusually rapid, especially in view of the
lack of an activating group normally required to accelerate
such processes. Given the constrained nature of these systems, Acknowledgements
and negligible entropy of activation, it is unlikely this variation
is due to steric effects on the conformation of the reacting
molecules as has been proposed for other systems. More plau-
sibly, our suggestion of this acceleration instead of what would
be expected is in large part due to the compounds already ex-
isting in a suitable conformation to undergo reaction. For ex-
ample, Figure 2 shows the crystal structure of compound 17,[17]
demonstrating that the migrating endo-hydrogen is orientated
above the p-system of the reacting alkene, and there is little
distance (3 Š) between it and the migration terminus.
We would like to thank EPSRC for funding (EP/K006053/1),
Mairi Haddow for X-ray crystallography, and Dr. Craig Butts and
Mr. Paul Lawrence for their invaluable assistance with the NMR
studies.
Keywords: aziridine · photochemistry · reaction kinetics ·
rearrangement · sigmatropic
[1] a) L. Degennaro, P. Trinchera, R. Luisi, Chem. Rev. 2014, 114, 7881–7929;
b) X. E. Hu, Tetrahedron 2004, 60, 2701–2743; c) P. Dauban, G. Malik,
Angew. Chem. Int. Ed. 2009, 48, 9026–9029; Angew. Chem. 2009, 121,
9188–9191; d) B. Kang, A. W. Miller, S. Goyal, S. T. Nguyen, Chem.
Commun. 2009, 3928–3930; e) S. Gandhi, A. Bisai, B. A. B. Prasad, V. K.
Singh, J. Org. Chem. 2007, 72, 2133–2142; f) R. Maeda, R. Ishibashi, R.
Kamaishi, K. Hirotaki, H. Furuno, T. Hanamoto, Org. Lett. 2011, 13, 6240–
6243; g) H.-T. Yang, M.-L. Xing, Y.-F. Zhu, X.-Q. Sun, J. Cheng, C.-B. Miao,
F.-B. Li, J. Org. Chem. 2014, 79, 1487–1492.
We speculate that in the anomalously fast (i.e., RT) cases,
this low entropic barrier combines with an increased level of
ring strain. In compounds 5 and 7, this derives from the addi-
tion of an extra ring, and in compound 9, it comes from the
addition of a bulky substituent on the aziridine ring. In the
other substrates, less strain is present and the reactions occur
at lower rates; however, other effects can still be observed in
these systems. For instance, the presence of an EWG conjugat-
ed to the non-reacting end of the p-system tends to lead to
faster reactions, presumably by reducing the enthalpy of acti-
vation, and substitution of the migration terminus leads to re-
[2] H. Ohno, Chem. Rev. 2014, 114, 7784–7814.
[3] a) P. Lu, Tetrahedron 2010, 66, 2549; b) S. H. Krake, S. C. Bergmeier, Tetra-
hedron 2010, 66, 7337; c) S. Stankovic, M. D’hooghe, S. Catak, H. Eum,
M. Waroquier, V. Van Speybroeck, N. De Kimpe, H.-J. Ha, Chem. Soc. Rev.
2012, 41, 643.
Chem. Eur. J. 2016, 22, 11429 – 11434
www.chemeurj.org
11433
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[4] a) D. C. D. Butler, G. A. Inman, H. Alper, J. Org. Chem. 2000, 65, 5887–
5890; b) K. Aoyagi, H. Nakamura, Y. Yamamoto, J. Org. Chem. 2002, 67,
5977–5980; c) M. A. Lowe, M. Ostovar, S. Ferrini, C. C. Chen, P. G. Law-
rence, F. Fontana, A. A. Calabrese, V. K. Aggarwal, Angew. Chem. Int. Ed.
2011, 50, 6370–6374; Angew. Chem. 2011, 123, 6494–6498; d) K.
Fugami, Y. Morizawa, K. Ishima, H. Nozaki, Tetrahedron Lett. 1985, 26,
857–860.
tures with the corresponding imine. Yields of the remaining photo-
chemical reactions are tabulated in the SI.
[5] a) B. M. Trost, M. Osipov, G. Dong, J. Am. Chem. Soc. 2010, 132, 15800–
15807; b) A. A. Cantrill, A. N. Jarvis, H. M. I. Osborn, A. Ouadi, J. B. Swee-
ney, Synlett 1996, 847–848; c) T. Kawamura, N. Matsuo, D. Yamauchi, Y.
Tanabe, H. Nemoto, Tetrahedron 2013, 69, 5331–5341.
[6] a) K. G. Maskill, J. P. Knowles, L. D. Elliott, R. W. Alder, K. I. Booker-Mil-
burn, Angew. Chem. Int. Ed. 2013, 52, 1499–1502; Angew. Chem. 2013,
125, 1539–1542; b) L. D. Elliott, J. P. Knowles, P. J. Koovits, K. G. Maskill,
M. J. Ralph, G. Lejeune, L. J. Edwards, R. I. Robinson, I. R. Clemens, B.
Cox, D. D. Pascoe, G. Koch, M. Eberle, M. B. Berry, K. I. Booker-Milburn,
Chem. Eur. J. 2014, 20, 15226–15232.
[12] In the cases of 21 and 23, elimination products of type 35 were also
formed.
[7] E. E. Blackham, J. P. Knowles, J. Burgess, K. I. Booker-Milburn, Chem. Sci.
2016, 7, 2302–2307.
[8] a) E. Piers, I. Nagakura, J. E. Shaw, J. Org. Chem. 1978, 43, 3431–3432;
b) B. M. Trost, P. H. Scudder, J. Org. Chem. 1981, 46, 506–509; c) B. M.
Trost, P. L. Ornstein, J. Org. Chem. 1982, 47, 748–751; d) E. Piers, A. R.
Maxwell, Can. J. Chem. 1984, 62, 2392–2395; e) E. Piers, A. R. Maxwell,
N. Moss, Can. J. Chem. 1985, 63, 555–557; f) R. J. Loncharich, K. N.
Houk, J. Am. Chem. Soc. 1988, 110, 2089–2092; g) P. A. Parziale, J. A.
Berson, J. Am. Chem. Soc. 1990, 112, 1650–1652; h) E. Nakamura, K.
Kubota, M. Isaka, J. Org. Chem. 1992, 57, 5809–5810; i) Y. L. Lin, E. Turos,
J. Am. Chem. Soc. 1999, 121, 856–857; j) Y. L. Lin, E. Turos, J. Org. Chem.
2001, 66, 8751–8759.
[9] a) J. Šhman, P. Somfai, D. Tanner, J. Chem. Soc., Chem. Commun. 1994,
2785–2786; b) P. Somfai, J. Šhman, Tetrahedron Lett. 1995, 36, 1953–
1956; c) J. Šhman, P. Somfai, Tetrahedron 1999, 55, 11595–11600.
[10] a) K. Jayaprakash, C. S. Venkatachalam, K. K. Balasubramania, Tetrahedron
Lett. 1999, 40, 6493–6496; b) D. Borel, Y. Gelas-Mialhe, R. Vessi›re, Can.
J. Chem. 1976, 54, 1590–1598; c) A. Hassner, R. D’Costa, A. T. McPhail,
W. Butler, Tetrahedron Lett. 1981, 22, 3691–3694; d) A. Hassner, W.
Chau, Tetrahedron Lett. 1982, 23, 1989–1992; e) W. H. Pearson, Tetrahe-
dron Lett. 1985, 26, 3527–3530; f) T. Hudlicky, J. O. Frazier, G. Seoane,
M. Tiedje, A. Seoane, L. D. Kwart, C. Beal, J. Am. Chem. Soc. 1986, 108,
3755–3762; g) J. Šhman, T. Jårevang, P. Somfai, J. Org. Chem. 1996, 61,
8148–8159; h) K. Ishii, T. Sone, T. Shigeyama, M. Noji, S. Sugiyama, Tetra-
hedron 2006, 62, 10865–10878.
[13] In the case of 23, the competing elimination reaction was appreciable
(final ratio 3:2); in this case, the rate constant is therefore a sum of the
first-order rate constants for these two processes (see Supporting Infor-
mation).
[14] The only rate constants available for such systems are those of Somfai
(see ref. [15]). Our most reactive system (5) rearranges 58 times faster
(k=1.8”10À2 sÀ1 versus 3.104”10À4 sÀ1) under similar conditions (808C,
in toluene and para-xylene, respectively).
[15] A. HussØnius, P. Somfai, D. Tanner, J. Phys. Org. Chem. 1996, 9, 623–625.
[16] Comparison of our data with that of ref. [15] was complicated by their
use of an Arrhenius rather than an Eyring plot to calculate entropy.
Consequently this value represents a recalculated DS° using the raw
data given in ref. [15] (p-xylene) using an Eyring approach. For compari-
son, the original, Arrhenius-derived value by Somfai was DS° =
À53 JKÀ1 molÀ1
.
[17] CCDC 1488690 (17) contains the supplementary crystallographic data
for this paper. These data are provided free of charge by The Cam-
bridge Crystallographic Data Centre.
[11] Irradiation of tetrahydroindolones 35a and 35b yielded tetracyclic aziri-
dines 5 and 7, respectively, in moderate yield. These yields are lower
than would be anticipated due to the [1,5]-hydrogen shift reactions of
these species proceeding efficiently at room temperature, leading to
difficulties in their purification and the products being isolated as mix-
Received: February 1, 2016
Published online on July 6, 2016
Chem. Eur. J. 2016, 22, 11429 – 11434
www.chemeurj.org
11434
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim