5-azido-2-aminopyridine, a new nitrene/nitrenium ion photoaffinity labeling agent that exhibits reversible intersystem crossing between singlet and triplet nitrenes

Article abstract of DOI:10.1021/ja405637b

The photochemistry of a new photoaffinity labeling (PAL) agent, 5-azido-2-(N,N-diethylamino)pyridine, was studied in aprotic and protic solvents using femtosecond-to-microsecond transient absorption and product analysis, in conjunction with ab initio multiconfigurational and multireference quantum chemical calculations. The excited singlet S₁ state is spectroscopically dark, whereas photoexcitation to higher-lying singlet excited S₂ and S₃ states drives the photochemical reaction toward a barrierless ultrafast relaxation path via two conical intersections to S ₁, where N₂ elimination leads to the formation of the closed-shell singlet nitrene. The singlet nitrene undergoes intersystem crossing (ISC) to the triplet nitrene in aprotic and protic solvents as well as protonation to form the nitrenium ion. The ISC rate constants in aprotic solvents increase with solvent polarity, displaying a "direct" gap effect, whereas an "inverse" gap effect is observed in protic solvents. Transient absorption actinometry experiments suggest that a solvent-dependent fraction from 20% to 50% of nitrenium ions is generated on a time scale of a few tens of picoseconds. The closed-shell singlet and triplet nitrene are separated by a small energy gap in protic solvents. As a result, the unreactive triplet state nitrene undergoes delayed, thermally activated reverse ISC to reform the reactive closed-shell singlet nitrene, which subsequently protonates, forming the remaining fraction of nitrenium ions. The product studies demonstrate that the resulting nitrenium ion stabilized by the electron-donating 4-amino group yields the final cross-linked product with high, almost quantitative efficiency. The enhanced PAL function of this new azide with respect to the widely applied 4-amino-3-nitrophenyl azide is discussed.

Full text of DOI:10.1021/ja405637b

Subscriber access provided by Grand Valley State | University
Article
5-Azido-2-Aminopyridine, a New Nitrene/Nitrenium Ion
Photoaffinity Labeling Agent That Exhibits Reversible
Intersystem Crossing between Singlet and Triplet Nitrenes
Maxim S. Panov, Valentyna D. Voskresenska, Mikhail N.
Ryazantsev, Alexander N. Tarnovsky, and Robert Marshall Wilson
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 12 Nov 2013
Downloaded from http://pubs.acs.org on November 16, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
5
-Azido-2-Aminopyridine, a New Nitrene/Nitrenium Ion
Photoaffinity Labeling Agent That Exhibits Reversible
Intersystem Crossing between Singlet and Triplet Nitrenes
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
#
Maxim S. Panov, Valentyna D. Voskresenska, Mikhail N. Ryazantsev, Alexander N.
Tarnovsky,* R. Marshall Wilson*
Department of Chemistry and Center for Photochemical Sciences
Bowling Green State University
Bowling Green, OH 43403
Abstract
The photochemistry of a new photoaffinity labeling (PAL) agent, 5-azido-2-(N,N-
diethylamino)pyridine was studied in aprotic and protic solvents using femtosecond-to-
microsecond transient absorption, and product analysis, in conjunction with ab initio
multiconfigurational and multireference quantum chemical calculations. The excited
singlet S
excited S
relaxation path via two conical intersections to S
1
state is spectroscopically dark, whereas photoexcitation to higher-lying singlet
and S states drives the photochemical reaction towards a barrierless ultrafast
, where N elimination leads to the
2
3
1
2
formation of the closed-shell singlet nitrene. The singlet nitrene undergoes intersystem
crossing (ISC) to the triplet nitrene in aprotic and protic solvents as well as protonation to
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 58
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
form the nitrenium ion. The ISC rate constants in aprotic solvents increase with solvent
polarity displaying a “direct” gap effect, whereas an “inverse” gap effect is observed in
protic solvents. Transient absorption actinometry experiments suggest that a solvent-
dependent fraction from 20 to 50 % of nitrenium ions is generated on a time scale of a
few tens of picoseconds. The closed-shell singlet and triplet nitrene are separated by a
small energy gap in protic solvents. As a result, the unreactive triplet state nitrene
undergoes delayed, thermally-activated reverse ISC to reform the reactive closed-shell
singlet nitrene, which subsequently protonates, forming the remaining fraction of
nitrenium ions. The product studies demonstrate that the resulting nitrenium ion
stabilized by the electron-donating 4-amino group yields the final cross-linked product
with high, almost quantitative efficiency. The enhanced PAL function of this new azide
with respect to the widely applied 4-amino-3-nitrophenyl azide is discussed.
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
#
Present address: Department of Chemistry, Emory University, Atlanta, GA 30322,
United States.
*
Address the correspondence to these authors
2
ACS Paragon Plus Environment

Page 3 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
INTRODUCTION
Photoaffinity labeling and photocross-linking (PAL/PCL) continues to evolve as a major
1
-7
technique for studying molecular interactions in biological systems. In this method, the
PAL agent specifically and reversibly binds to the active site of the biomolecule where
irradiation generates an extremely reactive intermediates that is thought to rapidly and
irreversibly react with the target site before they can drift apart. Great progress in the
development of new PAL agents and understanding of the photoaffinity labeling
processes has been made over the past ten years. It is of critical important to determine
the photochemical mechanisms of these PAL agents in considerable details so as to aid in
the design of more effective agents and increase their labeling selectivity and efficiency.
The PAL agent, 4-fluoro-3-nitrophenyl azide, is a system that has been widely
applied in photoaffinity studies with different classes of biological molecules in recent
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
8
-13
years. Recently, we have studied the photochemical mechanism of its amine analog, 4-
amino-3-nitrophenyl azide (1), in the presence of nucleophiles using ultrafast transient
14
absoprtion and quantum-chemical computational techniques. This molecule is promoted
to its first excited singlet S -state (π, π*) by absorption of a 420-nm photon and
1
undergoes unproductive fast internal conversion to the vibrational manifold of the ground
electronic state on a ~350-fs time scale. On the other hand, 350-nm excitation of 1 to its
2
second excited singlet S -state (π, π*) leads to two reaction pathways: the recovery of the
electronic ground state of the parent molecule and N
2
elimination (Scheme 1). Excitation
at 305 nm favors this second nitrogen-elimination channel, but some internal conversion
back to the ground state still takes place. The presence of the nitro group as a strong
electron-withdrawing substituent on the benzene ring has a dramatic effect on the
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 58
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
reactivity of the diiminoquinone nitrenium ion intermediate 2 produced in this photolysis.
2
The generation of the nitrene 3 from 1 occurs on the S surface, and this reaction
intermediate is a powerful base that abstracts protons extremely rapidly from a variety of
sources to form diiminoquinone nitrenium ion 2. This nitrenium ion, augmented by the
nitro group, becomes a powerful electrophile, and reacts with the conjugate base of the
acid that led to its formation. The protonation of the nitrene by methanol observed in this
work ranks with the fastest intermolecular proton transfer processes known, which
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
include the protonation of singlet diphenylcarbene by methanol, τ = 9 ps, and the
16
protonation of water by several excited photoacids.
However, 1 has several disadvantages: the nitro group is bulky, so reactions with
some target molecules are sterically hindered, in addition the nitro group is photoreactive
itself. With these considerations in mind, we set out to design and synthesize a new azide
that might be used as a more effective photoaffinity-labeling agent than the nitro analog
1
. The 5-azido-2-aminopyridine (4) was selected as a possibly improved PAL reagent,
Fig. 1. Nitrogen-containing heterocyclic compounds with azido groups have attracted
17-20
considerable attention as photoaffinity labeling reagents.
Common photochemistry of
pyridyl azides includes photochemical ring expansion leading to seven-membered di-N-
heterocyclic ring via the singlet nitrenes, and generally is thought to be very similar to
1
7,18,21,22
photochemistry of other aryl azides.
However, just as with the azide 1, the
introduction of the amino substituent para to the azide in 4 suppresses the ring-expansion
mode and favors formation of nitrenium ion 5.
In this work, the photochemical mechanism of the azide 4 and possible
advantages of this modified system are investigated using product analysis, time-resolved
4
ACS Paragon Plus Environment

Page 5 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(
from 100 femtoseconds to several microseconds) transient absorption, and ab initio
quantum chemical calculations. The photocross-linking chemistry of 4 is shown to
proceed through a previously unrecognized mechanism in which the highly reactive
closed-shell singlet nitrene 6 and the unreactive triplet nitrene 7 are in equilibrium with
each other and proceeds with nearly quantitative efficiency cross-linked products. Thus,
reaction of the reactive nitrene 6 is bimodal. This species is formed rapidly in the
femtosecond time domain and either forms the nitrenium ion 5 or undergoes ISC to 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
triplet nitrene 7 in the picosecond time domain. This ISC process is both environment-
23-27
dependent following the direct “gap” effect
in aprotic solvents and the “inverse gap”
2
3-27
effect
in protic solvents, and reversible. The reversibility of the ISC greatly extends
the effective lifetime of the closed-shell singlet nitrene, and in many ways is analogous to
28
the mechanism of thermally activated delayed fluorescence.
METHODOLOGY
Synthetic procedures and sample characterization. All experiments were performed on
o
the azide 4 at 22 C. Solvents used (methanol, ethanol, 2-propanol, n-butanol, t-butanol,
toluene, methyl acetate, acetonitrile and cyclohexane) were purchased from Aldrich or
EMD Chemicals. Imidazole hydrochloride was purchased from Aldrich. Methyl acetate,
toluene, and acetonitrile were used in anhydrous form.
14
Ultrafast transient absorption. The set-up described previously is based on an
amplified Ti:sapphire laser system (90-fs, 800 nm, and 1 kHz). Transient absorption (ΔA)
spectra of 4 were recorded at excitation wavelengths of 360, 310, and 255 nm produced
by a TOPAS-C optical parametric amplifier. Femtosecond pulses tunable from 280 to
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 58
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
3
90 nm from another TOPAS-C or a white-light continuum from 340 to 720 nm were
used as probe light. The instrument response function (150 fs, fwhm) was measured using
a Gaussian-like ΔA signal at time zero due to stimulated Raman scattering from neat
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
29
solvent. All ΔA spectra of 4 were measured at magic angle polarization conditions and
corrected for the group velocity dispersion of the probe light with an accuracy of 30 fs by
using cross-phase modulation or two-photon (pump plus probe) absorption signals from
3
0,31
neat solvent.
Multiexponential fits of ∆A kinetic traces and decay-associated global
3
2
fits of ΔA spectra were employed to characterize the temporal evolution of the reactive
30,31
species. The solvent signals
persist for 100-200 fs after excitation, so this region (300
fs upon 255-nm excitation) was excluded from fitting. The ΔA kinetic traces as a function
−t /τi
of delay time (t) were fitted to a sum of exponential functions: ∆A = ∆A + Aie
,
0
∑i
where τ
i i 0
are the time constants, A are the amplitudes, ΔA is the permanent spectrum at
infinitely long times (>> 1 ns). The ΔA spectra were globally fitted to a sum of
exponentials
∑
ε
j
(λ) exp(-t/τj,g), where τj,g are the time constants, and ε (λ) are the decay-
j
j
associated spectra characteristic of the τj,g time components and re-constructed from these
values based on the assumption of a consecutive reaction mechanism.
Microsecond transient absorption. The ΔA spectra were collected using a Proteus
3
3
spectrometer equipped with a 150 W Xe arc lamp, a monochromator and photodiode
detectors. Excitation 355-nm pulses of ~10-ns width with energy of 3-5 mJ from a
Nd:YAG laser/OPO system (10 Hz) were directed to the sample with an absorbance of
0
.6-0.8 at 355 nm.
6
ACS Paragon Plus Environment

Page 7 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Computational Details. All calculations were performed for the azide 4, where Et was
replaced by Me to minimize the computational cost, and its products (closed-shell singlet
nitrene 6, open-shell singlet nitrene 8, triplet nitrene 7, nitrenium ion 5, adduct/product
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
34
35
9
/10) using Gaussian 03 and MOLCAS 7.4. The 4 and 6-10 geometries were
3
6
optimized using MP2/cc-pVDZ, where solvent effects were accounted by the
3
7
conductor-like polarizable continuum model (CPCM). Also, the complete active space
self-consistent field (CASSCF) methodology was used to optimize the 4-10 geometries at
3
8
CASSCF/6-31G* CPCM. The conical intersection seems (CIs) were optimized at
39
CASSCF(12,10)/6-31G*, whereas the Moplot graphical program was employed to
visualize the lower cone of optimized CIs. Minimum energy paths (MEPs) were
4
0
computed using the intrinsic reaction coordinate method (IRC) at CASSCF(14,12)/6-
41
3
1G*. Using the complete active space second order perturbation theory (CASPT2) ,
vertical excitation transition (VET) and oscillator strength (f) calculations were
performed at the optimized geometries utilizing the CASPT2//CASSCF protocol and the
42
scalar relativistic atomic natural orbital basis set (ANO-RCC).
SYNTHESIS AND PRODUCT ANALYSIS
The route used to synthesize 4 is shown in Scheme 2. Irradiation of the solutions of 4
under a nitrogen atmosphere was conducted in a Pyrex tube at 350 nm in a Rayonet
o
Photochemical reactor at 20-25 C. The photolysis of 4, performed in the presence of
alcohols (HOR), leads to the formation of addition products 10 (Nuc = OR) in excellent
yields, Table 1 and Scheme 3. The irradiation of 4 in 2-PrOH leads to the formation of
only one addition product 11 (the 2-substituted adduct), the 4 or 5-substituted products
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 58
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
were not observed. The addition process with 4 proceeds much more effectively, in a
range from 80% to 100%, in comparison to the nitro analog 1. Moreover, in 4 the
conversion of starting material to product was more rapid and proceeds to completion in
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
0
.5 to 4 hours. Experiments in saturated solutions of phenol and dimethylammonium salt
in acetonitrile provided complete conversion to ratios of 10/12 = 100/0 and 98/2,
respectively, Table 1. Furthermore, irradiation of 4 in tert-butanol leads to complete
conversion of starting material to the addition product 10, Nuc= -tOBu. Moreover,
photolysis with imidazole hydrochloride in acetonitrile led exclusively to the
corresponding adduct 13, Scheme 3. It is noteworthy that no substitution reaction
occurred in the presence of imidazole in acetonitrile. Photolysis of 4 in cyclohexane was
completed within 3 hours and afforded a single product: the dimer 14, which suggests the
intermediacy of triplet nitrene in this non-protonating solvent. The aniline reduction
product 12 and the dimer 14 in ratio 3:1 were obtained upon photolysis of 4 in toluene.
COMPUTATIONAL RESULTS
Photochemical reaction paths. The CASPT2//CASSCF(16,13)/ANO-RCC VETs of 4
from the ground state (S
show that the lowest energy transition S
In contrast, the S → S transition at λmax = 314 nm (f = 0.0274) and the S
transition at λmax = 260 nm (f = 0.1999) are both intense. The S , S and S states are all
π, π*).
0
) to the three low-lying excited singlet states (S
→ S has a small oscillator strength, f = 0.0002.
→ S
1 2 3
, S and S )
0
1
0
2
0
3
1
2
3
(
The CASSCF photochemical reaction paths following excitation of 4 into S
1
2
, S ,
and S
3
were obtained to determine which one is responsible for N
8
2
extrusion. Following
ACS Paragon Plus Environment

Page 9 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
vertical excitation of 4 into the S
1
state, the MEP starting from the Franck-Condon (FC)
point reveals an S /S conical intersection (CI), Fig. 2. The MEP re-started from the
1
0
vicinity of the CI leads to the formation of ground-state 4 or photoproducts (nitrene and
molecular nitrogen). This nitrene has two singly occupied molecular orbitals: the 2p-π (5)
and in-plane 2p orbital (12) on N, and therefore, is identified as the open-shell singlet (8)
species. Following vertical excitation of 4 into S , the S and S potential energy surfaces
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
2
2
1
approach each other along the MEP; however, the intersection point is not reached within
the scan. The MEP starting from the S FC region locates the S /S CI, and then proceeds
3
3
2
to the CI(S /S
2 1
). From the S
2
/S
1
CI, the relaxation process on the S
1
surface leads to the
formation of the nitrene and N
2
, Fig. 3. This nitrene exhibits a closed-shell electronic
configuration with the empty in-plane p-orbital on a nitrogen atom and the doubly
occupied the 2p-π orbital, corresponding to the unrelaxed closed-shell singlet nitrene
geometry, which after optimization yields the relaxed geometry of this species where the
orbital occupancies are reversed (nitrene 6), Fig. 4. The active space orbital occupancies
39
and the Moplot topologies are given in Supporting Information.
Product geometries and energetics. Fig. 4 shows the CASSCF molecular orbitals
included in the active space of the triplet nitrene 7, open-shell singlet nitrene 8, and
closed-shell singlet nitrene 6. Fig. 5 shows the optimized geometries of these species. The
pyridine moieties in the optimized geometries of 7 and 6 are very similar. The
-1
CASSCF(10,8)/6-31G* level of theory predicts that 8 is 21 kcal mol and 6 is 33 kcal
-1
mol higher in energy than 7 in vacuum. However, the energy gap between 6 and 7
-1
becomes much smaller in polar solvents such as methanol (5.5 kcal mol ). In contrast,
for the azide 1 the CASSCF(10,8)/6-31G* calculations predict much larger energy
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 58
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
difference between the triplet and the closed-shell singlet nitrenes in the gas phase and
-1
methanol, 40 and 36 kcal mol , respectively, Supporting Information. The above trend is
supported by electric dipole moments of the involved species. Electric dipole moments of
the triplet and the closed-shell singlet nitrenes derived from 1 and 4 are calculated to be
6.1 and 6.9 D, and 4.1 and 7.8 D, respectively. A larger dipole moment is typically
correlated with a stronger dependence of the corresponding electronic state energy on
medium polarity. Indeed, 4 is substituted with a powerful electron-donating amino group,
which should stabilize the closed-shell singlet nitrene with respect to other product states
due to π-electron donation to the empty p-orbital on N. Consequently, 6 is more polar
than 7, and their energy gap is expected to decrease with increasing solvent polarity, in
the agreement with the calculations. The CASPT2 VETs for 6-8 and several other
reactive species derived from irradiation of 4 are shown in Table 2.
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
SPECTROSCOPIC RESULTS
Steady-state absorption. The UV-vis absorption spectrum of 4 in 2-PrOH consists of two
-1
-1
broad bands, a weak band is located at 340 nm (ε = 5969 M cm ) and stronger one is
-1
-1
centered at about 280 nm (ε = 18690 M cm ), Fig. 1. The UV-vis spectra of 4 exhibit no
solvatochromic shift. It should be noted that solutions of 4 does not exhibit any
noticeable fluorescence upon irradiation in the 280-360 nm range.
Ultrafast transient absorption. In neat cyclohexane upon 360-nm excitation, any
significant ΔA signals occur at delay times between -50 and 50 fs. Following 360-nm
excitation of 4 in cyclohexane, 100-fs ΔA spectrum consists of a 330-nm UV band and a
broad visible band with a maximum at ~485 nm (Fig. 6A). Between 1 and 20 ps, transient
1
0
ACS Paragon Plus Environment

Page 11 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
absorption narrows, blue-shifts and decays, Fig. 6B. From 50 ps to several hundreds of
picoseconds, the 313- and 456-nm product absorption bands grow and the 350-nm
ground-state bleach signal becomes pronounced, indicating that the overlapping 330-nm
band decays (Fig. 6C). The 1000-ps ΔA spectrum displays the 313-nm band as well as the
visible 456-nm band with two shoulders at 436 nm and ~510-515 nm. The 330- and 485-
nm transients agree with the CASPT2 VETs calculated for the singlet closed-shell nitrene
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
6
. A 313-, 436-, and 456-nm ΔA bands at long times are consistent the CASPT2 VETs
predicted for the triplet nitrene 7 (Fig. 6).
The 100- and 200-fs ΔA spectra of 4 in 2-PrOH following 360-nm excitation
display an intense ~322 nm band and a broad ~475-nm band, Fig. 7. From 1 to 50 ps, the
3
22-nm band and the red ΔA absorption decay concurrently with a rise of the 314-nm
band of 7 and the visible 475-nm band, which (albeit with a minor decay) remain to 1000
ps (Fig. 7C). The ∆A spectral evolution of 4 upon 310-nm excitation (Supporting
Information) and 255-nm excitation (Fig. 7D-F) is rather similar. However, the 300-fs
UV ΔA band is much broader upon 255-nm excitation compared to that at 360-nm,
suggesting that the species responsible for the short-time UV band is produced with an
excess of vibrational energy. This nascent species is assigned to 6, consistent with the
CASPT2 VET calculations.
The ∆A spectra were also measured following 360-nm excitation of 4 in methanol
(
Fig. 8) and several other solvents, Supporting Information. The ∆A spectra were also
measured for 4 in a mixture of acetonitrile and imidazole hydrochloride. Imidazole
mimics the amino acid histidine and, therefore, is a good model for studies of 4 as a
potential PCL/PAL system. Spectral profiles of the visible product ∆A bands in several
1
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 58
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
selected solvents are compared in Fig. 9. For 4 in protic solvents, the spectral evolution of
transient absorption resembles that observed in 2-PrOH. The 470-475 nm band forms on
3
a time scale of ten to several tens of picoseconds, being the fastest in CH OH and slowest
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
in butanols. With decreasing solvent polarity, this visible band exhibits a minor
hypsochromic shift. The 470-475 nm band decays faster in the alcohols with shorter
carbon chain lengths. In methanol solutions investigated to a 2.5-ns delay time, the decay
of this band is accompanied by a rise of transient absorption in the 530-580 nm range,
Fig. 8B, inset. In 2-PrOH, the aforementioned rise occurs to a much lesser degree. In the
aprotic solvents and acetonitrile/imidazole hydrochloride mixture, the ∆A spectral
evolution resembles that observed in cyclohexane, albeit occurring much faster. The
visible absorption band exhibits no noticeable decay within 1 ns, except for the
aforementioned mixture where some decay is observed similar to t-BuOH.
Fitting analysis. The best-fit parameters of multiexponential fits of ∆A kinetic traces
performed at wavelengths characteristic of the temporal evolution of the transient species
are summarized in Supporting Information. Two components (time constants, τ and τ )
1
2
are dominant in all solvents investigated, describing the short-time spectral reshaping
), and the formation of the 470-475 bands as well as the decay of the UV and red
absorption (τ ). In 2-PrOH, τ = 16 ps, independent of the excitation wavelengths, Table
3 and Supporting Information. In protic solvents, τ values range from 10.5 ps in CH OH
(τ
1
2
2
2
3
to 19.5 ps in t-BuOH, generally becoming larger with increasing carbon chain length of
the solvent molecule. Under the assumption that the visible bands single-exponentially
relax to zero, the following decay time constants can be deduced (in ns): 4.2 (MeOH), 6.5
1
2
ACS Paragon Plus Environment

Page 13 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(
2
EtOH), 18 (2-PrOH), 12.5 (n-BuOH), and 13.7 (t-BuOH). In aprotic solvents, τ changes
from 8.1 ps in acetonitrile to 38 ps in less polar toluene. Cyclohexane is unique in the
sense that τ and τ are significantly larger compared to the other solvents investigated: τ
= 22 ps and τ = 280 ps, Table 3.
A satisfactorily global fit to the ~350-620 nm ΔA spectra in all solvents requires a
sum of two decay-associated spectra (ε ) obeying exponential time dependences (time
1
2
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
2
i
constants τ1,g and τ2,g,) and a permanent spectrum, see Fig. 10 for 4 in 2-PrOH and
cyclohexane. The τ1,g represents spectral narrowing and the τ2,g describes the ~450 nm
product absorption rise, Table 3. The τ1,g and τ2,g values agree with the τ
1
2
and τ ,
indicating the consistency between these two fitting procedures.
Transient absorption actinometry. The formation of the nitrenium ion 5 and triplet
nitrene 7 products was quantified in MeOH, 2-PrOH, n-ButOH, CH CN, and C
measure of strength of interaction of these species with the solvent, we used E
3
6
H12. As a
T
(30)
4
3
polarity parameter. We assume that the visible ΔA amplitude quantifies the amount of 7
formed. This is justified as 5 has the red-shifted transitions of low oscillator strengths,
Table 4. The formation of 7 in C
yield (φTn,c = 1). Triplet-triplet absorption in C
quantum yield of the formation of the triplet species (φ ) in other solvents using eq. (1):
6
H12, where no protonation is possible, has unit quantum
6
H12 is used for the determination of the
Tn
αTn
εTn,c
×
×
φTn,c
=
φTn
(1)
αTn,c
εTn
The αTn and αTn,c slopes of the dependence of the ∆A signal at the visible band
maximum versus excitation energy were measured under the same experimental
1
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 58
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
conditions at t = 100 ps (1000 ps in C
6
H
12) when the band formation is complete. The
maximum extinction coefficients of the visible band of 5 in C
6
H12 (εTn,c ) and other
solvent (ε ) were evaluated based on the conservation rule of the transition dipole
Tn
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
44
moment strength (µ), eq. (2):
−
1
2
|
µ| = const nν εTn
STn (v)dν
(2)
∫
band
In eq. (2), n is the solvent refractive index, S(v) is the spectral profile of the visible band
normalized to unity at the maximum and plotted against wavenumber ( ). The ε values
and the α-ratio in cyclohexane and acetonitrile were similar, which suggests φ = 1 in
ν
Tn
CH
factor of 4.4 in MeOH) is a main reason, despite an offset by ε values smaller by a 1.33-
factor, as compared to C 12. We note that the above treatment assumes that (i) 6
3
CN. In the protic solvents, φTn << 1 (Table 4), for which smaller αTn slopes (e.g., a
6
H
converts only to 7, which is justified as it is the only relaxation path readily available to
6, see Discussion for more details; (ii) there is only one radiationless decay channel in
electronically excited 4 leading to 6 in unit quantum yield. The barrierless nature of the
S
2
3 6 3
, S states of 4 and the similarity in quantum yields for C H12 and CH CN solutions of
drastically different polarity support the second assumption. We determined ε (475 nm)
Tn
-1
-1
=
5415 M cm of 7 in CH CN using φ = 1 and the benzophenone actinometer in
3
Tn
45-48
-1
-1 46
benzene,
where the triplet state (εTb,max (532 nm) = 7220 M cm ) is formed in unit
45
quantum yield (φTb =1 ). Then, quantum yields upon 255-nm excitation of 2 were
4
5
-1
-1 46
measured using benzophenone in CH CN (φ =1, εTb,max (525 nm) = 6250 M cm ).
3
Tb
1
4
ACS Paragon Plus Environment

Page 15 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The values obtained were similar to those upon 350-nm excitation within the
experimental uncertainty of 20 %.
Microsecond transient absorption. Fig. 11A shows the ∆A spectra of 4 in n-BuOH after
excitation at 355 nm. A broad absorption band at 530-550 nm decays concurrently with a
rise of the 380-410 nm absorption (time constant, 670-ns) . In 2-PrOH and t-BuOH, 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
∆A spectra of 4 behave similarly, displaying a 2.5 μs decay time constant of the 530-550
nm band. Importantly, 4 in toluene (Fig. 11B) gives rise to a much larger ∆A signal
compared to n-BuOH under the same excitation conditions. In toluene, the 460-nm ∆A
band (134 ns) greatly resembles the 1-ns ∆A spectrum in our ultrafast experiments. From
1
34 to 1864 ns, another absorption band at 440 nm develops.
DISCUSSION
0 2
Photochemical reaction paths. The 340 and 280 nm bands are assigned as S → S and
S
0
→ S
predictions from the CASPT2 calculations. The S
small oscillator strength and, consequently, is absent in the spectrum. We conclude that
the CASPT2 calculations accurately reproduce the transitions from the ground S state to
the first three excited singlet (S , S , and S ) states. The 255-nm excitation used in this
study promotes 4 to S . On the other hand, 360-nm excitation is at the low-energy wing
of the S → S transition, and therefore, should promote 4 to S without vibrational
energy excess. Excitation at 310 nm can be assigned to the S → S transition populating
vibrational excited S , with some contribution from the overlapping S → S transition.
3
, based on the agreement between their positions and relative intensities with the
0
→ S transition carries a negligibly
1
0
1
2
3
3
0
2
2
0
2
2
0
3
1
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 58
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 2
According to the MEP calculations, excitation of 4 to S results in N elimination
with the formation of the open-shell singlet nitrene, 8, Fig. 2. However, photoexcitation
to S seems not to be possible because of low oscillator strength of the S → S electronic
1
0
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
transition. Following excitation to the S and/or S states, 4 relaxes barrierlessly through
2
3
two CIs (S
3
/S
2
2 1 0
and S /S ) to form the closed-shell singlet nitrene 6, Fig. 3. Because the S
→
S
2
and S
0
→ S electronic transitions carry all the oscillator strength and lead to the
3
same product (6), the photochemistry of 4 is not excitation-wavelength selective. The
presence of the S /S CI easily accessible from the Franck-Condon regions on the S and
3
2
3
S
2
potential energy surfaces explains the lack of any appreciable fluorescence of 4.
Moreover, stimulated emission of 4 was not observed upon either 255- and 360-nm
excitation, consistent with the excited molecule reaching the S /S CI very rapidly (<<
rate constants in molecules containing
There is no evidence that ISC in 4 takes place on a 100-fs time scale, as
3
2
11
-1
1
50 fs). ISC may occur with 10
s
2
7,49
heteroatoms.
5
0
1
was previously reported in the S state of 1-nitronaphthalene. Therefore, we conclude
that the radiationless relaxation of 4 proceeds through the manifold of singlet excited
states.
The radiationless decay pathways of the closed-shell singlet nitrene 6 include
internal conversion to the open-shell singlet nitrene 8 and ISC to the triplet nitrene 7,
51,52
Scheme 4. The former path is highly forbidden
because the change in orbital angular
momentum is not balanced by a corresponding change in electron spin angular
momentum, but ISC is allowed from 6 to 7 because the angular momentum associated
with the electron spin flip is coupled with the angular momentum associated with the
53
redistribution of electron density between orthogonal molecular orbitals. Indeed, the
1
6
ACS Paragon Plus Environment

Page 17 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
most intense vertical electronic transitions of 7 and 6 correspond to π → π* and n → π*,
respectively (Fig. 4. In 7: orbital 5 (HOMO)→ orbital 7; in 6: orbital 1→ orbital 8).
3
1 2
Consequently, the electronic configurations of 7 and 6 correspond to (π, π*) and 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
respectively. Therefore, the only spin-orbit allowed processes are ISC back and forth
5
3,54
between triplet and closed-shell singlet states.
For aromatic systems such as
5
3,55
derivatives of aza-benzenes, -naphthalenes, and –phenanthrenes,
it is known that the
fusion of N-heteroatoms into an aromatic ring significantly enhances ISC because of
more efficient spin−orbit coupling between electronic configurations involving the
nitrogen lone pair and π-orbitals. One would then expect fast ISC in 6 derived from 4.
Detailed studies of the photochemistry of the azide 1 reported the formation of a very
basic closed-shell singlet nitrene that subsequently undergoes protonation in alcohol
solvents on a several picosecond time scale to yield a metastable nitrenium ion, in
addition to undergoing less efficient ISC. The azide 4 investigated in this work is
structurally related to 1, and one might speculate that 6 may act as a proton acceptor in
the manner analogous to that of the closed-shell singlet nitrene derived from 1 and that
neither the open-shell singlet nitrene 8 nor the triplet nitrene 7 would be protonated since
these attachments would lead to nitrenium ion excited states.
Ultrafast dynamics. The S
very rapidly (<< 150 fs) on account of the barrierless reaction paths and lead to
elimination of N with formation of the closed-shell nitrene 6. Indeed, 6 (330-350 nm)
2 3
and S photochemical reactions of 4 are expected to proceed
2
was detected within the first 150-fs following excitation Fig. 6A. Numerous examples of
intense UV (330-350 nm) absorption of singlet nitrenes derived from aryl azides are well
1
4,56-58
known.
1
Narrowing of the absorption of 6 and its time scale (τ ), ranging from about
1
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 58
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
2
ps in polar solvents to 18 ps in cyclohexane, are typical of neutral polyatomic
5
9-
molecules as they dissipate their excess vibrationally energy to the surrounding solvent.
6
3
In cyclohexane, toluene, and methyl acetate, where protonation is not possible, 6 is
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
expected to convert solely to 7 via ISC (time constant, τ ).
2
The visible transitions of 5 have low oscillator strengths in comparison to those of
, so that the absorption in the vicinity of the 450-nm band maxima is dominated by 7.
7
However, the visible ∆A absorption in protic and aprotic solvents of similar polarity (to
minimize polarity effects on the band shape) is somewhat broader in the former solvents,
Fig. 9. This difference is observed as early as 20 ps after excitation, Supporting
Information. This would be consistent with the formation of some nitrenium ions on a
several picosecond time scale. Furthermore, if 6 relaxes solely to 7, the quantum yield for
the formation of 7 should be unity in all solvents. However, the trend observed is that the
τ
2
becomes shorter but the triplet quantum yield decreases with the protic solvent
polarity, suggesting the presence of competing reaction pathways, i.e. protonation and
ISC. The τ component is then due to both ISC and protonation in solvents where
protonation is possible. The τ values in 2-PrOH are independent of the photon energy
2
2
(16.5 ps for 255-, 310- and 360-nm excitation, Table 4), suggesting that the ISC and
protonation pathways originate from vibrationally relaxed 6. The fact that 4 undergoes
the consecutive photochemical reaction path with branching (ISC and protonation)
occurring only during the last relaxation step allows one to interpret the transient cascade
global fit as follows: ε
the cold 6 produced by relaxation of ε
decays. For 6 that decays via two competing processes, ISC and protonation, its lifetime
A
is the hot 6 “instantaneously” formed upon excitation of 4, ε
B
is
A
, and ε is the long-lived 7 or 7/5 generated as ε
C
B
1
8
ACS Paragon Plus Environment

Page 19 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
is: τ
constant. The quantum yields for ISC (φTn ) and protonation (φp ) are defined as: φTn
and φp = κ , with φp = 1 - φ , so that κISC and κ can be determined. As shown
in Table 4, the κ increases with an increase of protic solvent polarity. These κ values
2
= 1/(κ
P
+ κISC), where κ
P
is the protonation rate constant and κISC is the ISC rate
=
κ
ISC
τ
2
P
τ
2
P
Tn
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
p
p
are about three times smaller than those for the protonation reaction involving the closed-
14
shell singlet nitrene species derived from the azide 1 in the same solvents.
Nitrene intersystem crossing. Increasing polarity of the aprotic solvent leads to a
dramatic acceleration of ISC in 6: the κISC increases by ~34 times upon going from C H
6
12
to CH
polarity slows ISC with the κISC in MeOH smaller than that in 2-PrOH by a factor of
1.8. Seeking an explanation, we note that 6 has both non-bonding electrons localized on
3
CN, Fig. 12 and Table 4. In the protic solvents, the trend is opposite: increasing
~
the same orbital of the N atom (the doubly occupied in-plane 2p-orbital (6), Fig. 4), and
as a result, its dipole moment is larger than that of the biradical triplet species. As the
solvent polarity increases, 6 is expected to be more strongly stabilized than 7, which
decreases the singlet-triplet energy gap between these two species. This conjecture is
supported by the CASSCF(10,8)/6-31G* calculations of the energy gap between 6 and 7
-1
-1
(5.5 kcal mol in MeOH vs. 33 kcal mol in vacuum).
With increasing singlet-triplet gap, ISC between the involved states generally
2
3-25
decelerates. This is conventionally known as the “direct” energy gap law,
explained by poorer Franck-Condon factors with increasing singlet-triplet energy gap.
The propensity rule indicates that the absorption of one quantum has the highest
and is
2
3,24
2
6
probability in the transition, and as a result, the ISC transition probability is favored by
1
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 58
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
25
high frequency vibrational modes (e.g., the C–H stretching vibrations). Yet, ISC rate
constants sometime increase with increasing energy gap. The so-called “inverse” gap
2
2
26
effect has been observed in aryl carbenes. Usually, this is attributed to the lack of the
effective coupling between the singlet and triplet states separated by a small energy gap
and having sparse manifolds of vibrational states. As the energy gap increases, the
acceptor vibrational levels in the vicinity of the donor state become increasingly
congested, leading to faster ISC. In words of Englman and Jortner, ISC from the closed-
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
2
5
shell singlet 6 to triplet 7 nitrene is in the “weak coupling limit”. This case is observed
in large molecules where the vibronic coupling is not very strong, giving rise to only
slight displacements of the initial and final electronic potential energy surfaces. The 6 and
7
species indeed shows very similar geometries with the modest maximum deviation in
the bond lengths of ~0.17 Å (Fig. 5). The weak coupling limit predicts a large increase of
ISC rate constants (roughly exponential for large gaps) with decreasing energy gap,
which explains the large κISC variation observed in the aprotic solvents. Polarity of protic
solvents decreases the energy gap between 6 and 7, and these species enter the small
energy gap domain where the singlet-triplet gap becomes comparable to the frequency of
the promoting high-frequency vibrational mode. With further increase of solvent polarity,
the gap decreases further, the promoter state density becomes sparse, weakening the
coupling of vibronic states of 7 and 6, and causing slower ISC. This “inverse” gap law is
observed from 2-PrOH to MeOH, Fig. 12. In an alternative interpretation, the high-
frequency mode may no longer be the coupling mode for small gaps, and lower-
2
2
frequency promoting modes with weaker couplings may become more important.
2
0
ACS Paragon Plus Environment

Page 21 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Protonation mechanisms. Based on Fig. 11B, the 460-nm ∆A band (134 ns) in toluene is
assigned to 7. This conclusion implies that in aprotic solvents, 7 is long-lived under our
excitation conditions. In protic solvents, a much weaker ~530 nm ∆A band (200 ns in n-
BuOH, Fig. 11A) may be explained by a relatively small extinction coefficient in the
responsible species, such as the nitrenium ion 5, Table 2. The 530-nm band is the
precursor for the 410-nm absorption band, which based on its spectral position (Table 2)
can be attributed to the adduct 9 formed from the nitrenium ion 5. The similar ∆A band
following the decay of 7 is observed in methanol, Fig. 8B, insert.
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 photolysis of 4 in protic solvents leads to the formation of the addition
products 10 in a very high yield, Table 1. It is clear that the precursor of the addition
products 10 is the nitrenium ion 5. In principle, this species can be formed by the
6
4-66
protonation of either the closed-shell singlet or triplet nitrene.
The ∆A spectra of 4
reveal that a fraction of 6 undergoes protonation on a time scale of several tens of
picoseconds, whereas the remaining fraction undergoes ISC to form 7. However, since
direct protonation of the triplet nitrene is less likely, since protonation of 7 to 5 requires a
67
spin inversion. Therefore, the most reasonable mechanism for the nitrenium ion
formation is the protonation of the closed-shell singlet nitrene 6. The next question we
address is how to reconcile the observations of very high yields of both the addition
products 10 (Table 1) and the triplet nitrene 7 (Table 4) in protic solvents.
A likely scenario is thermal activation of the triplet nitrene 7 coupled with reverse
68-70
ISC into the closed-shell singlet nitrene 6.
Such a mechanism has been invoked
previously for protonation of the nitrene photochemically derived from 5-azido-8-
7
0
methoxypsoralen. Consequently, a closed-shell singlet nitrene reformed may reacts with
2
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 58
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
a source of protons forming a nitrenium ion. Similar observations have been made
7
0
68,69
previously by McClelland and Bettinetti
and co-workers. Equilibrium was observed
7
0
between singlet and triplet nitrenes derived from 5-azido-8-methoxypsoralen. However,
in this work the spin-orbital correlations between these species were not fully
appreciated, Scheme 4, and the lower energy open-shell singlet nitrene was coupled with
the triplet nitrene. In the present system, it has been shown that the closed-shell singlet
nitrene, is involved in this chemistry. With this adjustment, it now becomes clear how the
nitrene might be stored in an unreactive triplet state and returned over relatively long
periods of time via ISC to the reactive closed-shell singlet state, which is rapidly
protonated to form the reactive nitrenium ion. However, such a situation is possible only
if (i) in 6, rate constants of protonation to 5 and direct ISC to 7 are comparable, Table 4,
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
(
ii) 6 and 7 are separated by a relatively small energy gap making the reverse ISC of 7
8
-1
thermodynamically feasible at ambient temperatures (a 5×10 s thermal activation rate
-1
71
constant for a 5.5 kcal mol energy gap predicted by the Arrhenius expression with a
12 -1
5
×10 s pre-exponential factor is consistent with a decay of 7 and a rise of 5 in
methanol on a 2.5-ns time scale, Fig. 8) and (iii) dimerization of 7 is much slower than
reverse ISC of 7 and protonation of the reformed 6 (~3% excited fraction of 4 (1.1 mM)
1
0
-1 -1
upon our conditions suggests a 1 µs reaction time for dimerization of 6 with 10 M s
7
2
second-order diffusion rate constants typical for the solvents investigated. ). Therefore,
the emerging picture is that a significant fraction of nascent 6 forms 7. However, this
triplet state is populated temporally, and 7 re-populates 6 on a time scale of several to
several tens of nanoseconds determined by the solvent-dependent triplet-singlet energy
gap. The reformed 6 subsequently undergoes rapid protonation. In a sense, this situation
2
2
ACS Paragon Plus Environment

Page 23 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
resembles the well-known, thermally activated delayed fluorescence phenomenon where
the triplet excited state behaves as a “parking lot” for the fluorescing singlet excited
2
7
state. Based on the κ and κISC rate constants (Table 4), we estimate that a fraction of 5
P
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
is formed directly on ultrafast time scales (e.g. about one fifth in 2-PrOH), whereas the
remaining fraction of 5 is formed more slowly as a result of thermally-activated reverse
ISC from 7 to 6 followed by protonation of 6. As a result, the cross-linking product yield
is 80-100% (Table 5).
Scheme 5 summarizes the reaction steps following photolysis of 4 in cyclohexane
and 2-PrOH. In cyclohexane, a 287-ps lifetime of 6 is due to ISC to 7, and 7 eventually
forms the dimer 14 detected in product analysis, Scheme 3. The formation of the dimer
1
4 or aniline reduction product 12 (possibly present in a μs time domain as a 440-nm
absorption band, Fig. 11B), is perfectly consistent with the formation of triplet nitrene in
aprotic solvents. A 16.5-ps lifetime of 6 in 2-PrOH is due to ISC and protonation. A
diiminoquinone-like species, the nitrenium ion 5, stabilized by the electron-donating 4-
amino group slowly collapses forming the adduct 9/cross-linked product 10. We note that
the protonation step should lead to the formation of a contact ion pair (CIP), which
equilibrate to the solvent-separated ion pair (SSIP) via proton relay (Grotthuss
7
3-76
mechanism
). The CIP to SSIP relaxation observed for 1 via some red shift in the CIP
1
4
absorption band is not detected in the present study for 4, likely due to overlapping
absorption of 7 and kinetic mechanisms involved in the formation of 5.
CONCLUSIONS
2
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 58
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
The photochemical mechanism of 5-azido-2-(N,N-diethylamino)pyridine (4) in various
solvents have been proposed on the basis of ultrafast and nanosecond transient absorption
and product analysis, supported by multiconfigurational (CASSCF) and multireference
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
(
CASPT2) calculations. When 4 is excited to S and S states, it follows the ultrafast
2 3
barrierless reaction path via two conical intersections towards the lowest singlet excited
state, where N elimination takes place. The nascent closed-shell singlet nitrene 6 is
S
1
2
formed within the first 150 fs after excitation and undergoes vibrational relaxation with
time constants ranging from 2 ps (polar solvents) to 18 ps (cyclohexane). In aprotic
solvents, vibrationally relaxed 6 undergoes ISC on a solvent-dependent time scale of
several to several hundred picoseconds, giving rise to the triplet nitrene 7. The
corresponding kISC rate constants increase and the energy gap between 6 and 7 decreases
with increasing solvent polarity, and therefore, these kISC values obey the “direct” gap
law. In the protic solvents studied, vibrationally relaxed 6 undergoes both ISC to 7 and
protonation to form the nitrenium ion 5. An increase of the protic solvent polarity makes
ISC slower, so these kISC values follow the “inverse” gap law. The protonation rate
constants increase linearly with the increase in an empirical parameter of solvent polarity,
T
E (30). Experimental results suggest that ambient thermally-activated reverse ISC of 7 to
6
in polar protic solvents is thermodynamically possible, which is supported by the
CASSCF calculations of the energy gap between the closed shell singlet and triplet
-1
nitrenes in methanol (5.5 kcal mol ). As a result, 6 reformed from 7 tends to be
protonated forming 5 within a time domain limited by the lifetime of 7. In 2-PrOH, about
four fifth of the nitrenium ion 5 is produced via the aforementioned mechanism on a 10-
2
4
ACS Paragon Plus Environment

Page 25 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
00 ns time scale whereas the remaining fraction is formed directly on an ultrafast time
scale. In MeOH, this ratio is ~50:50.
Product studies suggest that 4 is more efficient in the formation of cross-linked
products compared to 1. The first reason for the enhanced cross-linking efficiency is that
the S state of 1, reached by excitation at 420 nm, is not reactive towards N elimination:
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
2
this S
1
1
excited state population returns back to the ground state. In 4, the S state is
spectroscopically dark and optical excitation can promote this system only to the higher-
lying S and S excited states. The molecule then follows the barrierless ultrafast
2
3
relaxation path to form the closed-shell singlet nitrene 6. This species is a precursor to the
nitrenium ion 5 and cross-linking product 10. Thus, 1 can only harvest a part of the UV-
vis spectrum for its cross-linking function, whereas the entire absorption spectrum of 4 is
useful. The second reason is the relative energies of the closed-shell singlet and triplet
nitrene species. In 4, these species are separated by the relatively small energy gap which
is surmountable by thermal activation of the triplet molecules. As a result, the unreactive
triplet state behaves as a “parking lot” for the highly reactive closed-shell singlet nitrene,
and the nitrenium ion. This process resembles the well-known phenomenon of thermally-
activated delayed fluorescence. In 1, the energy gap between the triplet and the closed-
shell singlet nitrenes is much larger, and not surmountable via ambient thermal
activation.
Consequently, the azide 4 offers substantial advantages over the azide 1 as a
photoaffinity agent. Comparison of the photochemical characteristics of 1 with 4 is
shown in Table 5. One of the main disadvantages of many aryl azides as PAL agents is
their photoaffinity labeling at wavelengths shorter than 280 nm, the wavelengths of light
2
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 58
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
leading to photochemical damage to the biological samples. The azide 4 is highly
photoactive at wavelengths longer than 290 nm. Moreover, excitation of 4 leads to more
efficient formation of cross-linked substitution products. It is noteworthy, that even
though 1 absorbs in the visible region, 420-nm excitation of 1 does not yield a nitrene.
The coupling reactions of 4 proceed more efficiently than those of 1 with full conversion
of starting material, and the photolysis of 4 in the presence of different nucleophiles form
mainly addition products with fewer side products than 1, which is probably due to the
lack of unproductive nitro group photochemistry that occurs with 1. Finally, the PAL
conventional wisdom is that cross-linking should proceed rapidly following excitation of
the PAL agent. While this happens to some degree with 4, since a significant fraction of
cross-linking does occur within a few picoseconds of excitation, a large portion the
initially formed singlet nitrene is converted into an unreactive triplet form which lives
long enough to probe a variety of active sites or active site conformations. Furthermore,
some of these sites will be more effective at initiating reverse ISC by virtue of their
polarity or lack thereof. It shall be interesting to see how this type of site selectivity can
affect PAL processes. Finally, the mechanism described here, equilibrium between
closed-shell singlet and triplet nitrenes, finally accounts for the singlet-like reactivity that
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
6
8-70
has been observed for triplet nitrenes in other systems.
Supporting Information
Synthesis of starting material, product characterization, description of the time-resolved
spectroscopic techniques used, CASCCF active-space orbitals of the azide 4 and nitrenes
derived from 1 and 4, minimum-energy CI Moplot topologies, optimized geometries and
2
6
ACS Paragon Plus Environment

Page 27 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
relative energies of the nitrenes, and transient absorption spectra of 4 in 2-PrOH upon
3
10-nm excitation and other solvents upon 360-nm excitation with full sets of the best-fit
parameters. This material is available free of charge via the Internet at http://pubs.acs.org.
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
Acknowledgments
This work was supported by the NSF CAREER award (Grant CHE-0847707, ANT) and
the R. Marshall and Antonia G. Wilson Fund. An allocation of computer time from the
Ohio Supercomputer Center is gratefully acknowledged.
Figure Captions
Figure 1. UV-vis absorption spectra of 4 in several solvents normalized to unity at the
absorption maximum. The arrows indicate the excitation wavelengths used in the current
study.
Figure 2. CASSCF MEPs starting from the S Franck-Condon point (FC) of 4 to the
1
1 0
S /S conical intersection (CI) and from the CI towards the formation of the
photoproducts.
3
Figure 3. CASSCF MEPs from the S FC point of 4 towards the formation of the
3 2 2 1
photoproducts via the S /S and S /S CIs.
Figure 4. The occupancy of the molecular orbitals of the nitrenes derived from 4 upon
completing their geometry optimization at the CASSCF(10,8)/6-31G* level of theory.
Figure 5. The optimized geometries of the triplet 7, the open-shell 8 and closed-shell
singlet 6 nitrenes computed using CASSCF(10,8)/6-31G*/CPCM in methanol.
2
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 58
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
Figure 6. ΔA spectra of a 0.25-mm flowing jet of 4 (1.3 mM) in cyclohexane excited at
3
60 nm (pulse energy, 4.1 μJ). Delay times between the probe and pump pulses (in
picoseconds) and shown in the legends. Top panels show the allowed CASPT2 vertical
excitation energies of 6, 7, and 8 where bar thickness is proportional to the calculated
oscillator strengths.
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
Figure 7. ΔA spectra of 4 (1.1 mM, 0.2-mm flow cell) in 2-PrOH for various delay times
(
in picoseconds, shown in the legends). Solutions were excited either at 360 nm (A-C) or
-1
255 nm (D-F) with an energy 6.1 μJ pulse . The ΔA signals for λ > 525 nm (0.1 ps) are
due to solvent cross-phase modulation and impulsive stimulated Raman scattering, and
from 395 to 405 nm (≤ 0.2 ps) are due to solvent stimulated Raman scattering.
Figure 8. ΔA spectra of 4 (1.1 mM, 0.2-mm flow cell) in MeOH excited at 360 nm (0.8
-1
μJ pulse ). The inset shows the 100- and 2500-ps ΔA spectra normalized at the long-
wavelength wing.
Figure 9. The visible product bands (normalized to unity) from the 1-ns ΔA spectra of 4
measured in protic and aprotic solvents of similar polarity. A: 2-PrOH and CH
(30) = 48.4 and 45.6, respectively. B: t-BuOH and C , E (30) = 43.3 and 38.9,
respectively. Non-polar C (30) = 34.3) is added for comparison.
Figure 10. The decay-associated spectra (ε ) and time constants (τ1,g and τ2,g) extracted
3
CN,
E
T
3
H
6
O
2
T
6 T
H12 (E
i
6
from the global fit of the ΔA spectra of 4 in 2-PrOH (A-C) and C H12 (D-F) upon 360-nm
excitation.
Figure 11. ∆A spectra of 4 in n-BuOH (A) and toluene (B) at several delay times (shown
in the legends) after 355-nm nanosecond excitation. In B, the ∆A 1-ns spectrum measured
in ultrafast experiments in toluene solutions is scaled to the 134-ns ∆A spectrum.
2
8
ACS Paragon Plus Environment

Page 29 of 58
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Figure 12. The intersystem crossing rate constants (κISC) from the closed-shell singlet 6
T
nitrene to the triplet nitrene 7 as a function of a solvent polarity parameter E (30). For
-1
cyclohexane, toluene, and methyl acetate, κISC = (τ
2
) . The κISC entries for the remaining
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
solvents are from Table 4.
2
9
ACS Paragon Plus Environment