Photoinduced rearrangement of aromatic N-chloroamides to chloroaromatic amides in the solid state: Inverted Πn-ΣN occupational stability of amidyl radicals

ARTICLE

pubs.acs.org/JPCA

Photoinduced Rearrangement of Aromatic N-Chloroamides

to Chloroaromatic Amides in the Solid State: Inverted Π_NꢀΣ_N

Occupational Stability of Amidyl Radicals

Panꢀce Naumov,*^,†Yildiray Topcu,^†,‡Mirjana Eckert-Maksiꢁc,*^,§Zoran Glasovac,^§Fabijan Pavoꢀseviꢁc,^§

Manoj Kochunnoonny,^†and Hideyuki Hara^||

^†Department of Material and Life Science, Graduate School of Engineering, Osaka University, 2ꢀ1 Yamada-oka, Suita,

Osaka 565ꢀ0871, Japan

^§Division of Organic Chemistry and Biochemistry, Rudjer Boꢀskoviꢁc Institute, Bijeniꢀcka c. 54, HRꢀ10000 Zagreb, Croatia

ESR Division, Bruker Biospin K.K., 3ꢀ9 Moriya-cho, Kanagawa-ku, Yokohama-shi, Kanagawa 221-0022, Japan

S Supporting Information

b

ABSTRACT: We report a solid-state photochemical rearrange-

ment reaction by which aromatic N-chloroamides exposed to UV

light or sunlight are rapidly and eﬃciently converted to chloroaro-

matic amides. The course, the intermediate (nascent chlorine vs

dichlorine) and the outcome of the reaction depend on the excita-

tion (exposure time, wavelength, and intensity) and on inherent

structural factors (the directing role of the substituents and, as

demonstrated by the diﬀerent reactivity of two polymorphs of N-chlorobenzanilide, the supramolecular structure). The photolysis of

the chloroamides provides facile photochemical access to arylamidyl radicals as intermediates, which in the absence of strong

hydrogen bond donors are stabilized in the reactant crystals by CꢀH/NꢀCl π interactions, thus, providing insight into their

3 3 3

structure and chemistry. Thorough theoretical modeling of the factors determinant to the stability and the nature of the spin-hosting

orbital evidenced that although the trans-Π_||state (Np spin) of the amidyls is normally preferred over the trans-Σ_^conﬁguration

(Nsp²spin), stabilization by aromatic conjugation, steric and geometry factors, as well as by electronic eﬀects from the substituents

can decrease the ΠꢀΣ gap in these intermediates signiﬁcantly, resulting in similar and, in the case of the orthogonal amide-phenyl

disposition, even reversed population of the unpaired electron in the two orbitals. Quantitative correlation established that the

inverted occupational spin stability and the Π_NꢀΣ_Ncrossover are collectively facilitated by the conformation, valence angle, and

disposition of the amide grouprelative to the aromatic system. The stabilization and detection of a trans-Σ_^radical was experimentally

accomplished by steric locking of the orthogonal trans-amide conformation with double ortho-tert-butyl substitution at the phenyl

ring. The eﬀects of the single para-phenyl substituents on the relative occupational stabilityof the arylamidyl radical states point out to

non-Hammett behavior. By including cumulative electronic eﬀects from multiple substitutions, four distinct families of the aromatic

amidyl radicals were identiﬁed. The Π state is the most stable structure of the N-phenylacetamidyl radical and of most of the

substituted arylamidyls, although the Σ_^and Π_^states can also be stabilized by introducing tert-butyl and nitro groups, respectively.

Recently, we investigated⁴the brief note⁵by Porter and Wilbur in

1927 that, in addition to the thermolysis of melted neat N-chloro-

acetanilide (N-chloro-N-acetylaminobenzene, 1a; Scheme 1), photo-

1. INTRODUCTION

Solid-state rearrangements,¹by means of which the molecular

topology is altered while the overall number of chemical bonds is

retained, are an indispensable “green”synthetic tool because they can

provide clean, solventless access to structures whose preparation by

conventional (wet) synthetic methods may be laborious and eco-

nomically unfeasible. Provided the reaction yields are suﬃciently

high, the regularly ordered and sterically restricted environment in

lysis of its crystals also aﬀords 1d. By UV-induced photolysis of the

NꢀCl bond, single crystals of 1a are quantitatively converted⁴to

para-chloroacetanilide (1d), which is chemically similar to one of the

most used analgesics and antipyretics, paracetamol (para-hydro-

xyacetanilide). Following the photolysis of the NꢀCl bond, a

chlorine atom from the N-chloroamide group of each molecule of

the solid state can signiﬁcantly enhance the reaction selectivity, thus,

1a is exchanged with a para-hydrogen atom from the aromatic ring of

bypassing the isolation/puriﬁcation steps and greatly simplifying the

an adjacent molecule within the slanted head-to-tail O HꢀC

3 3 3

synthetic procedure. Relative to the analogous rearrangements in

solution, however, the chemical transformations that are attainable

through solid-state rearrangements are markedly less common,²with

most of the examples being limited to the transfer of protons or small

chemical groups in organic materials, such as are the well-studied

methyl group rearrangements.³

hydrogen bond (HB) chains, whereby 1d is obtained in quantitative

Received: April 22, 2011

Revised:

May 25, 2011

Published: May 25, 2011

r

2011 American Chemical Society

7834

dx.doi.org/10.1021/jp203771c J. Phys. Chem. A 2011, 115, 7834–7848

|

The Journal of Physical Chemistry A

ARTICLE

yield. The preliminary results of this photoinduced rearrangement in

crystals of 1a, which proceeds through a photoinduced N-phenyla-

cetamidyl radical (Ph^•NAc) as intermediate, have indicated that the

selectivity remarkably supersedes that of the analogous light-induced,

heat-induced, or acid-catalyzed reactions in solution.⁶In fact, it has

been established that thermolysis or acid-catalyzed decomposition of

1a in solution involves molecular chlorine and aﬀords a very complex

mixture of products, including dimers and comparable amounts of all

three chloroacetanilide isomers 1bꢀ1d.⁶

well as of the succinimidyl radical (Figure 1a).¹¹Although the Π_N

state of the acyclic amidyl radicals is preferred over the Σ_Nstate, as

it has been demonstrated in the pioneering work of Lessard and

others,¹²steric constraints (e.g., tert-butyl substitution) as well as

other geometric factors can decrease the ΠꢀΣ gap signiﬁcantly,

leading to nearly equal probability for population of both available

orbitalswith the unpairedelectronortomixingof theoccupational

states. The probability of state quasidegeneracy increases with the

aromatic amidyls, where the electron/electron pair couple com-

petes for resonance with the aromatic π-system.¹³In such a case,

the diﬀerence in delocalization energy of the electron pair relative

to that of the unpaired electron could exceed the energy required

for promotion of one electron from a nonbonding Nsp²orbital to

the half-occupied Np orbital, ultimately resulting in delocalization

of the electron pair from the Np orbital. The situation can be

additionally complicated if the forms having the unpaired electron

at the oxygen atom (the O-centered states, Π_Oand Σ_O) are

explicitly accounted for, although the contributions of these states

are normally much smaller relative to Π_Nand Σ_N. Furthermore,

aromatic substituents of diﬀerent electronic aﬃnity can exert

either identical (“class S” amidyls, Hammett behavior) or opposite

(“class O” amidyls, non-Hammett behavior) eﬀect on the

stabilization.¹³Although some of the available experimentalresults

seem to be in favor of Π_Nas the ground state for several amidyl

radicals, the method-dependence of the theoretical calculations

and the occasional misassignments of ESR data have precluded a

consensus about the true nature of the hosting orbital.¹⁴

The amidyl radicals are important synthetic intermediates⁷

which have been pointed out recently for their key role in the

processes of human aging.⁸In situ produced amidyl radicals have

also been suggested as eﬃcient initiators for metal-catalyzed living

radical polymerization.⁹Nonetheless, most of the basic research

interest in these species originates from the long-standing con-

troversy related to the nature of the orbital that accommodates

the unpaired electron. A substantial amount of experimental

evidence exists showing that, similar to the related aminyl radicals,

the Π_Nradical state, which has the unshared electron pair in the

Nsp²hybrid orbital and the unpaired electron in the orthogonal

Np orbital, is the actual ground state of the acyclic amidyls¹⁰as

Scheme 1. Photoinduced Rearrangement of N-Chloroaceta-

nilide (1a) to para-Chloroacetanilide (1d) via Amidyl Radical

(Ph^•NAc) in the Solid State

Encouraged by the high reaction selectivity and yield observed

with 1a,⁴we hypothesized that this solid-state rearrangement is a

more general photochemical reaction of aryl-N-chloroamides. To

investigate that, we prepared a series of substituted N-chloroace-

tanilides (NCAs) and N-chlorobenzanilides (NCBAs; Scheme 2)

and analyzed their solid-state photochemistry together with that

of the parent molecule 1a, aiming to provide a more quantitative

mechanistic assessment. The results presented herein conﬁrm

that all studied compounds are photoreactive, endorsing this

reaction as a general solid-state photochemical rearrangement by

which NCAs and NCBAs are eﬃciently converted to the respec-

tive chloroaromatic amides in the solid state. Moreover, we found

that the reaction can be also induced and can proceed to

completion by using natural sunlight. By employing sunlight as

a green,¹⁵clean, traceless, and cost-free reagent,^16,17this solvent-

less procedure oﬀers great potential and provides access to

industrial amounts of a myriad of chloroaromatic ﬁne chemicals

and pharmaceuticals, herbicides, fungicides, bactericides, dyes, or

pigments.^18ꢀ21The green aspects of such a procedure pertain to

the photoinduced rearrangement step, although in principle the

amide reactants could be obtained by sonochemical coupling (as

we also demonstrate with two compounds in this work) and

possibly chlorinated with mechanochemical methods. We also

report evidence that the amidyl radical intermediates, which are

created by homolytic NꢀCl dissociation of the excited chloroamide

Figure 1. Simpliﬁed schematic representation of the orbital interaction

in Π_N(a, d, f) and Σ_N(b, c, e) states of simple aliphatic (a, b) and

aromatic (cꢀf) N-centered amidyl radicals (cis-conﬁguration of the

amidyl group). As exempliﬁed here with the aromatic radical created by

photolysis of N-chloroacetanilide (Ph^•NAc; cꢀf), the amidyl radical

center can assume with the π-electron system of the phenyl ring any

conformation between the limiting orthogonal (^; c, d) and parallel

( ; e, f) geometries. The O-centered states, which have minor contribu-

tion to the electronic structure, are not shown here.

Scheme 2. Photoinduced Rearrangement of N-Chloroacetanilides (R₁= methyl) and N-Chlorobenzanilides (R₁= phenyl) in the

Solid State

7835

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Figure 2. IR spectral changes of NCAs and NCBAs after exposure to very weak UV light (,1 mW cm^ꢀ1). Spectra recorded at larger number of time

3

points are deposited as Figure S1 in the SI.

functionality and are normally very reactive species in solution,

can be notably stabilized in single crystals of the aromatic

N-chloroamides. Theoretical modeling showed that, instead of

the normally expected Π_Nstates, in some of these crystals, the

arylꢀamidyl intermediates can be exceptionally stabilized as Σ_N

spins. For the amidyl radicals, we also provide the most detailed

to date theoretical assessments of the eﬀect of structural factors,

including the conformation of the amide group, the delocaliza-

tion of the unpaired amidyl electron with the aromatic π-system,

as well as the steric attributes and substituent eﬀects, on the

relative orbital occupancy of the unpaired and paired electrons.

and characterization are deposited as Supporting Information,

SI). During the synthesis, we unintentionally obtained in pure

state an additional N-chlorinated product, N-Cl-3d, which has a

reactive NꢀCl bond. Except 7a,²²from light-protected, cooled

petroleum ether solutions, the products crystallized as colorless

or white crystalline materials. Upon exposure to continuous-

wave (cw) UV light with power densities of several mW cm^ꢀ2

,

3

the transformation became conspicuous in both IR and ¹H NMR

spectra even after a minute, and quantitative conversion of all

N-chloroamides to photostable products was achieved in several

minutes (typically in less than 10 min). Under identical condi-

tions, the chloroaromatic amides that do not contain NꢀCl bond

(e.g., 1bꢀ1d, 3d, and 5d) were not reactive; therefore, secondary

photochemical reactions of the product chloroaromatic amides

were excluded. As outlined in Scheme 1, the characteristic IR

spectral changes, shown for 1aꢀ5a and N-Cl-3d in Figure 2, are

consistent with the conversion of the N-chloroamides to chlor-

oaromatic amides: the ν(NCl) band disappeared, the amide

ν(NH) bands evolved, the strong ν(CO) singlet was replaced

2. RESULTS AND DISCUSSION

2.1. Preparation of the N-Chloroamides and Solid-State

Photochemistry. The reactants 1aꢀ7a (Scheme 2) were ob-

tained in high yields by N-chlorination of the respective anilides

with solid trichloroisocyanuric acid or hypochlorite solution (the

latter method was used to prepare 3a; the details of the synthesis

7836

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

with the amide I and II bands, and the phenyl ring vibrations

reflected the typical substitution-specific changes.

We were able to control the reaction rate simply by moderating

the excitation power. While the reaction was typically completed

in several minutes with strong UV light, by using very weak

incident light the reaction time could be extended up to several

hours, providing insight into the reaction mechanism and kinetics

(Figures S1 and S2 in the SI contain a large collection of IR and

NMR spectra). Spectral comparison of 1a, which was irradiated in

dispersed state (KBr) at high excitation power with pure o-, m-,

and p-chloroacetanilides (1b, 1c, and 1d), showed that 1d was

produced selectively in a few minutes. The purity of the products

was conveniently assessed from the IR region of the reporter

bands characteristic for the type of the phenyl disubstitution,

ortho, meta, and para (some typical diﬀerences between the

spectra of the isomeric 1b, 1c and 1d can be inspected from

Figure 1 in ref 4). With low excitation power and after much

longer irradiation times (several hours), however, the spectrum of

the product indicated a slightly diﬀerent outcome. Evidently, by

altering the reaction course, the excitation aﬀects the product

distribution, with purer products being obtained after shorter

reaction times (see the results of the product HPLC analysis of 1a

below). The product composition and the reaction rate were also

aﬀected by the excitation wavelength (Figure S3, SI). Selective

exposure to energy-ﬁltered excitation resulted in a diﬀerent

outcome in the ﬁnal IR spectrum: while red light (λ > 550 nm)

did not have any eﬀecton 1a even after prolonged irradiation, very

weak UV light at λ_max= 365 and 313 nm caused signiﬁcant

changes, which occurred even with excitation with λ =

260ꢀ313 nm and the reaction was typically completed in 2ꢀ3

h. Higher excitation energy by including the 254 nm line or

shorter wavelengths, however, produced a qualitatively diﬀerent

spectrum, indicating activation of additional photolytic pathways.

Intrigued by the observation of the strong sensitivity of NCAs

and NCBs to artiﬁcial UV light, we exposed crystals of these

compounds to direct natural sunlight (λ > 300 nm). The IR

spectra of solar-irradiated 1a and 5a, selected as typical examples

and deposited in the SI as Figures S4 and S5, conﬁrmed that with

direct sunlight all compounds underwent the solid-state rearran-

gement. The reaction times necessary for total conversion of the

chloroamides depended on the compound: for instance, while the

reaction of 5a (form A) was slower and took ∼4 h to completion,

the rearrangement of 1a was completed in less than 2 h. The

spectral analysis conﬁrmed that the products were identical to

those obtained by using artiﬁcial UV light. Larger-scale (grams)

experiments of thinly spread powder samples were also performed

and conﬁrmed that the procedure can be easily scaled up. Because

of its extreme simplicity and eﬃciency, this reaction of NCAs and

NCBAs could be considered as a convenient and eﬀective method

for preparation of large amounts of chloroaromatic amides by a

solar light-induced rearrangement, which oﬀers the potential for

application in the large-scale production of ﬁne chemicals.

Figure 3. (a) ESR spectra of 2aꢀ5a and N-Cl-3d irradiated in situ at

50 K with cw UV light (Hg lamp) in the SQUID cavity. (b) Experimental

ESR spectrum of UV-irradiated 6a at 50 and 200 K, plotted together with

the respective simulated ESR spectrum.

Figure 4. Left panel: Time-proﬁle of the relative concentrations of the

reactant and the products during the UV-photolysis of N-chloroaceta-

nilide (1a) monitored by HPLC. Right panel: HPLC traces of the

reaction mixture after 0, 3, and 9 h irradiation (a complete plot of all

HPLC traces was deposited in the SI), together with the trace of a

standard mixture of the three isomeric chloroacetanilides (bottom). The

peak from the (unsubstituted) acetanilide is denoted with an asterisk.

signals, the resonances were simple and structureless. The signal

from the UV-irradiated 3a decreased after the excitation was

terminated, while the signals of the other amides remained

unsaturated, and they continued to intensify even after several

minutesfrom the irradiationonset, showing that thepopulation of

the amidyl radicals increased with the irradiation time. The

extracted g values of 2aꢀ5a and N-Cl-3d were 2.044, 2.0037,

2.0047, 2.0047, and 2.0043 at 200 K, and 2.0041, 2.0041, 2.0044,

2.0067, and 2.0055 at 50 K. However, the solid-state broadening

of the ESR bands, which could neither be decreased at 50 K nor by

using single crystal specimen or glassy matrices, thwarted the

resolution of the hyperﬁne structure in these cases (Figure 3a). In

the case of the tri-tert-butyl precursor 6a, however, the simulated

spectrum at 50 K was consistent with a nitrogen-centered radical

(hfc = 21.2 G) in addition to two carbon-centered radicals

(Figure 3b, see the Discussion of the theoretical results below).

To scrutinize the time-course and the product distribution of

this reaction, as well as to assess their dependence on the power

of the incident UV light, we monitored (HPLC) the time-depen-

By employing ESR spectroscopy on photoirradiated crystalline

1aꢀ4a and N-Cl-3d at 200 and 50 K (Figure 3), as well as on

frozen glassy matrices of the reactants in cyclopropane, we

conﬁrmed the evolution of the amidyl radicals, along with

the reaction described in Scheme 1. All solids were ESR-silent

before irradiation. Upon exposure to UV light from a medium-

pressure Hg lamp in the resonance cavity, the samples developed

broad resonances from the photoinduced radicals (Figure 3). The

ESR spectrum of UV-irradiated 1a was identical to that reported

previously.⁴Except for N-Cl-3d, which exhibited more complex

dence of the reaction at very low excitation power (,1 mW cm^ꢀ2

Figure 4 and Figure S4, SI). To discriminate between the

;

7837

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Scheme 3. Reaction Proﬁle of the N-Chloroacetanilide Rearrangement Induced by Strong (Fast Reaction) and Weak UV Light

(Slow Reaction)^a

^aThe kinetics of the reaction aﬀects the diﬀusion and recombination of chlorine atoms, thus, determining the product composition.

conversion yields of diﬀerent molecules, all conversion yields

were uniformly lowered by intentionally decreasing the apparent

reaction rate with UV light of low intensity. Contrary to the high

selectivity (∼100% 1d) observed with high-intensity UV light,

the product analysis revealed that in addition to 1d, the long-term

exposure of pure solid 1a to low-intensity UV light aﬀorded two

additional products in smaller yield, which were identiﬁed with

the standard addition method as 1b and acetanilide (Figure 4).

After the reaction was completed (>99.6%) in about 9 h, the

main component (82.4%) was 1d, but the mixture contained

similar amounts of 1b and acetanilide (8.4 and 8.9%, re-

spectively), although 1c or other products could not be detected.

The result shows that about 8ꢀ9% of the chlorine in the long

term exposed sample of 1a evaporated from the crystal as gaseous

chlorine (the chlorine gas that emanated from these weakly

excited samples could be also detected organoleptically by its

characteristic strong odor), and the same amount of amidyl

radicals recombined later with hydrogen atoms from the solvent

to give acetanilide. Furthermore, under long-term irradiation, the

product analysis (HPLC) showed very clear dependence of the

reaction rate on the phenyl substitution (see Table S1 in the SI).

For instance, we found that while 100% of 1a, 2a, and N-Cl-3d

reacted in 3 h and 96% of 4a reacted after 6 h, only 13.9% of 3a

reacted even after 3 h irradiation. This result seems to be related

to the decrease of the ESR signal of 3a after the termination of

excitation. The most viable explanation for the higher inertness

of 3a is the deactivation eﬀect of the meta-methyl group on the

reaction and/or eﬃcient radical recombination.

chlorine. Under such circumstances, part of the halogenation

proceeds with chlorine molecules so that, although the para-

isomer is obtained as the main product (∼82.4%; 74.0% by

chlorination with Cl^•and 8.4% by chlorination with Cl₂), the

ortho-isomer 1b is also produced in signiﬁcant yield (∼8.4%),

without any detectable amounts of the meta-isomer. Being ortho,

para-activating, the amide group directs the secondary, Cl₂-

driven reaction to the para- and ortho-positions, which are

electronically preferred over the meta-position. The deﬁciency

of chlorine due to recombination and evaporation from the

crystal of gaseous chlorine (8ꢀ9% Cl^•, recombined as 4.4% Cl₂)

results in an identical amount of free amidyl radicals, which

subsequently abstract hydrogen atoms from the HPLC solvent to

yield acetanilide (∼8.9%). As a result, in addition to para-

chloroacetanilide, the ortho-chloro isomer 1b is obtained in

approximately equal yield to unsubstituted acetanilide. Notably,

this amount of free radicals (∼9%) is very close to the value 9.2%

determined in single crystal of 1a by X-ray photodiﬀraction,⁴

which appears to be in support of the current mechanistic

conclusions. This example nicely demonstrates that the reaction

rate has the critical role in determining the product distribution

in this radical-mediated, diﬀusion-controlled solid-state reaction.

2.2. Crystal Structures of the N-Chloroamide Precursors.

To further clarify the reasons behind different reactivities and to get

insight into the structure of the amidyl radical intermediates, we

determined the crystal structures of nine related compounds (Table

S2, SI), seven of which are N-chloroamide precursors of photo-

induced amidyl radicals (Figure 5).²²N-chlorobenzanilide (5a) was

obtained as two polymorphic modifications, orthorhombic (5a-A,

space group P2₁2₁2₁), and monoclinic (5a-B, space group P2₁/c).

The dependence of the product composition on the reaction

rate can be explained by considering the structure and reaction

mechanism presented in Scheme 3. The structure of 1a before

the excitation is composed of head-to-tail hydrogen bonded

1

The H NMR spectra in DMSO (Figure S2 in the SI) as a

solvent of medium polarity showed that all compounds exist as

dynamic mixtures of the cis- and trans-conformations of the

N-chloroamide group. Surprisingly, in contrast to the generally

higher stability and thermodynamic preference of the trans- over

the cis-conformation of the amide group, but in accordance with

the structure of 1a,⁴except the tert-butyl derivative 6a (see the

discussion of the steric eﬀects below), the amide group in the

crystals of all N-chlorinated compounds assumes the cis-con-

formation. This is clearly a consequence of the chlorine substitu-

tion. Indeed, as exempliﬁed by the structures of benzanilide,^23a

5d (this work and ref 23b) and the three crystallographically

independent molecules in 3d (this work), an unsubstituted

amide group usually adopts the more stable, trans-conformation.

C(para)ꢀH Cl molecules, each of which is 85.6:14.4-disor-

3 3 3

dered over a pseudomirror plane.⁴The plane of the chloroamido

group is quasiperpendicular to the phenyl ring. In the case of a

fast reaction (high incident power), following the homolysis of

the NꢀCl bond, the reaction is triggered by chlorine atoms,

which substitute the closest (para) hydrogen atoms. The rapid

release of chlorine atoms within a short time interval prevents

their diﬀusion and recombination, thus, favoring the Cl^•-driven

para-hydrogen substitution as a dominant process and aﬀords 1d

as the sole product. Conversely, prolonged exposure of the

crystals to weak UV light facilitates diﬀusion-controlled pro-

cesses and recombination of the chlorine atoms to molecular

7838

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Figure 5. ORTEP diagrams (30% probability level) of the molecular structures of six N-chloroamide amidyl radical precursors (from the two

polymorphs of 5a, only the molecule in the more stable polymorph 5a-A is shown). In the crystal, 1a exists as the cis-isomer which is disordered

(85.6:14.4) over a pseudomirror plane.⁴The stick-type plot on the far right represents overlapped representation of the structures, which are

superimposed at the phenyl ring. Color codes: 1a, black; 2a, red; 3a, blue; N-Cl-3d, ochre; 4a, green; 5a, magenta; 6a, violet.

Table 1. Selected Intramolecular Parameters and Percent Conversions of Eight N-Chloroamides

parameter^a

1a

2a

3a

N-Cl-3d

4a

5a-A

5a-B

6a

amide group conformation

cis

trans

—(amideꢀphenyl)/ꢀ

—(O9ꢀC8ꢀN7ꢀCl)/ꢀ

—(C1ꢀN7ꢀC8)/ꢀ

—(N7ꢀC8ꢀC10)/ꢀ

d(C8dO9)/Å

87.2

74.4

87.8

81.7

90

72.3

71.0

89.4

ꢀ0.6

126.9(3)

116.8(4)

1.204(5)

ꢀ3.7

127.9(2)

115.5(2)

1.218(3)

ꢀ0.3

127.8(4)

115.7(5)

1.235(7)

4.1

0

2.3

ꢀ0.2

126.4(2)

117.6(2)

1.215(3)

ꢀ179.2

123.4(2)

117.5(3)

1.197(3)

127.3(5)

116.3(4)

1.222(6)

127.6(2)

114.9(2)

1.205(3)

125.9(3)

116.5(3)

1.247(4)

intramol H-bonds

CꢀH Cl

CꢀH Clp

CꢀH Cl

3 3 3

conversion (strong UV)^b

conversion (weak UV)^b/%

QC^c

21.4

QC^c

13.9

QC^c

35.2

QC^c

95.9

QC^c

NM^d

62ꢀ65

48ꢀ49

^aThe atom labels refer to Scheme 1. ^bThe photoconversions (expressed as % of the N-chloroamide that reacted) were determined by IR spectroscopy or

HPLC. “Strong UV” here generally refers to incident UV powers typically on the order of 1 mW cm^ꢀ1, while “weak UV” refers to powers ,1

3

mW cm^ꢀ1. ^cQuantitative conversion. ^dNot measured.

3

On the contrary, in the crystals of 6a, which were obtained

following identical synthetic sequence (chlorination of the

amide, puriﬁcation, and crystallization), the N-chloroamide

group is locked by the steric bulk of the two ortho-tert-butyl

groups and exists in the crystal as pure trans-conformation (we

could not observe any products other than the trans-isomer).

The trans-conformation in 6a is additionally stabilized by four

intramolecular hydrogen bonds (Table S4, SI) that bridge

oxygen/chlorine atoms to the methyl hydrogen atoms. The

structure of the amidyl radical in the crystal of UV-irradiated

1a,⁴which exists in identical (cis) conformation and similar

geometry with the parent molecule before the photolysis,

indicates that the amidyl radicals obtained from the other NCAs

and NCBAs are probably also created and exist in the solid state

with molecular conformations that resemble those of the respec-

tive N-chloro precursors.

A selection of relevant parameters pertaining to the molecular

geometries and intermolecular contacts in these crystals is listed

in Tables 1 and S4 (SI). In the structures of all compounds, the

planes of the chloroamide group and the phenyl ring are nearly

orthogonal, but they are exactly perpendicular only in 4a, where

the molecule sits on a special position, and in 6a, where the amide

functionality is constrained to orthogonality by the symmetric

steric crowding from the two tert-butyl groups (see the over-

lapped representation in Figure 5 and the values listed in

Table 1). The largest deviation from orthogonality was observed

in the structures of the two polymorphs of 5a, 5a-A (72.3ꢀ), and

5a-B (71.0ꢀ), where the twisting is enhanced by molecular strain

caused by the 1,2-diphenyl substitution. The molecular twisting

in 5a-A and the trans-conformation in 6a are also reﬂected in the

longest (1.247(4) Å) and the shortest (1.197(3) Å) carbonyl

bond within the set of studied N-chloroamides.

In the absence of strong hydrogen-bond donors, the molecular

packing in the structures of the sterically noncongested 2aꢀ4a

and N-Cl-3d is accomplished by dimerization through CꢀH

3 3 3

O bonds and weak interactions involving the chlorine atoms

(Figure 6, Table S4). In the crystal of 2a, by utilizing C9ꢀH9B

3

O1 bonds between the carbonyl oxygen and the ortho-methyl

3 3

proton the molecules are arranged as centrosymmetric dimers

(Figure 6a). The dimeric units are translated and connected by

halogenꢀhalogen contacts Cl1 Cl1⁰(3.229(4) Å, sum of the

3 3 3

van der Waals radii 3.50 Å), forming chains along the ab-diagonal

(Figure 6b). The chains, which are related to each other by a

2-fold screw axis, are further connected by weak interactions

C4ꢀH4 O1. Similar to 2a, being associated through hydrogen

3 3 3

bonds C2ꢀH2 O1, the molecules in the structure of 3a form

3 3 3

centrosymmetric dimers, but they utilize the carbonyl oxygen and

ortho-aromatic proton instead (Figure 6c). The dimers are

translated and connected by very weak C8ꢀH8B Cl1 contacts

3 3 3

between the ortho-methyl group and the chlorine atom, forming

chains (Figure 6d). The chains, which are related by a 2-fold screw

axis, expand the packing further through hydrogen bonds

C4ꢀH4 O1. A similar centrosymmetric dimer, associated by

3 3 3

C2ꢀH2 O1 bonds between the carbonyl oxygen O1 and the

aromatic proton H2, was also observed in the structure of 4a

(Figure 6e). The dimers are translated along the b-axis, forming

chains. The chains contact through very weak N1ꢀCl1

π

3 3 3

interactions forming a zigzag pattern (Figure 6f). The Cl1-to-

centroid distance is 3.634(2) Å, which falls within the 0.2 Å limit

from the van der Waals cutoﬀ for Cl-to-C interactions (3.45 Å). In

7839

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Figure 6. Stick-model representation of the hydrogen bonds and crystal packing in the structures of 2a (a, b), 3a (c, d), 4a (e, f), and N-Cl-3d (g, h).

the crystal of N-Cl-3d, the molecules are also associated as

centrosymmetric dimers that involve an ortho-aromatic proton

(Figure 6g), similar to what is observed with 3a. These dimers are

translated along the b-axis, forming strings via CꢀH O bonds,

3 3 3

which are further organized by very weak CꢀH Cl interac-

3 3 3

tions between the meta-methyl hydrogen andthe para-substituted

chlorine atom into layers along the c-axis (Figure 6h). The layers

are connected by very weak interactions N1ꢀCl1 π along the

3 3 3

c-axis. The Cl C distance is 3.635 Å and, thus, within 0.2 Å of

3 3 3

the van der Waals Cl-to-C limit of 3.45 Å. Here, as well as in the

case of 4a, the presence of weak Cl π interactions is supported

3 3 3

by the absence of other, much stronger interactions.

Contrary to the N-chloroamides with small (methyl) substit-

uents, where the supramolecular structure is determined by

dimerization through the (CꢀH O)₂synthon, in the struc-

3 3 3

tures of the sterically congested 5a and 6a, where such close

contacts are not feasible, the molecular assembly is directed by

halogen bonds N1ꢀCl O (Figure 7).²⁴In the structure of 5a

3 3 3

(form A, see below), the molecules are associated by halogen

bonds N1ꢀCl1 O1 into strings (Figure 7a). The distance

3 3 3

Cl1 O1 (3.025(3) Å) is shorter than the sum of the van der

3 3 3

Waals radii of Cl and O (3.27 Å). The molecules, translated along

the a-axis, form chains by weak interactions C6ꢀH6 Cl1,

3 3 3

which further expand by weak π π interactions into layers

3 3 3

along the c-axis (Figure 7b). Utilizing CꢀH π, πꢀπ, and

3 3 3

Figure 7. Stick-model representation of the hydrogen bonds and crystal

NꢀCl O interactions, the molecules form a helix along the

3 3 3

packing of 5a (a, b) and 6a (c, d).

b-axis. In the crystal of 6a, where the dimerization is impeded by

the tert-butyl groups, the molecules form strings through halogen

bonds N1ꢀCl1 O1, with d(Cl1 O1) = 3.312(4) Å

in the course of several days as bunches of long colorless prisms

(the conditions which lead to preferential crystallization of either

of these polymorphs were not optimized). Form 5a-A was

obtained much more frequently than 5a-B and it appears as a

thermodynamically more stable form at room temperature. Each

of the two structures contains a single type of 5a molecules with a

nearly planar amide group (a dihedral angle of 2.27(60)ꢀ in 5a-A

and 0.18(44)ꢀ in 5a-B). The relative disposition of the two

phenyl rings in the two structures is practically identical

(69.83(11)ꢀ and 69.23(9)ꢀ), but the molecules differ significantly

3 3 3

(Figure 7c), a distance that is shorter than the sum of the

respective van der Waals radii (3.27 Å). The adjacent strings

are connected through CꢀH O interactions in a helical

3 3 3

fashion (Figure 7d).

2.3. Structures and Reactivity of the Two Polymorphs of

5a. By slow evaporation of cooled (∼277 K) petroleum ether

solutions of 5a, from individual batches, we obtained in the pure

state two nonsolvated polymorphs: form A (5a-A) appeared after

one day as large colorless blocks, while form B (5a-B) crystallized

7840

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Figure 8. Overlapped representation of the molecules (a), intermolecular bonds (b, c), and molecular packing (d, e) in the polymorphs 5a-A (b, d) and 5a-B(c,e).

in respect to the phenyl ring twists from the amide plane:

both phenyl rings of 5a-A, where the interplanar angles between

the amide group and the rings are 73.99(11) and 52.84(13)ꢀ, are

twisted relative to the rings of 5a-B, 69.52(8) and 35.89(13)ꢀ

(Figure 8a). In the absence of strong hydrogen bond donors,

the supramolecular structures of the two polymorphs are domi-

nated by weak interactions (halogen bonds, hydrogen bonds,

and πꢀπ interactions). By limiting the availability of the chloro-

amide group for participation in intermolecular interactions, the

phenyl twist around the amide bridge indirectly determines the

primary supramolecular structure. Whereas in form 5a-A the

Consequently, all IR measurements were performed by pelletiz-

ing previously UV-irradiated samples, this approach aﬀording

qualitatively identical products.

1

According to the H NMR analysis (see Figures S7, S8, and

Table S3 in the SI), under identical conditions, 51ꢀ52% of the

product from 5a-B and 35ꢀ38% of the product from 5a-A were

obtained after 3 h of UV irradiation. The diﬀerent reactivity

correlates well with the diﬀerent structures: whereas in 5a-A

the reactive chlorine atom is halogen-bonded to oxygen atom

O1 at d(O1 Cl1) = 3.025(3) Å, in 5a-B the chlorine

3 3 3

atom appears unbound, with closest distance d(Cl1 Cl1) =

3 3 3

molecules are connected through halogen bonds N1ꢀCl1

3.614(3) Å. Thus, the higher reactivity of 5a-B can be attributed

to the presence of non-halogen-bonded chlorine, which can react

readily after dissociation of the NꢀCl bond. Relative to 5a-B, the

halogen bond in the crystals of form 5a-A decreases the reactivity,

because after their creation, a larger amount of chlorine atoms

remains within the radical cage and in the vicinity of the nitrogen

center of the geminate amidyl radicals, with which they can

recombine to regenerate the reactant.

3 3 3

O1 in chains (Figure 8b,d), in form 5a-B they are associated

through hydrogen bonds C3ꢀH3 O1 (Figure 8c,e). In the

3 3 3

structure of 5a-A, the strings are further connected by weak

C6ꢀH6 Cl1 interactions along the a-axis and by weak πꢀπ

3 3 3

interactions along the c-axis (Figure 8d). In 5a-B, the molecular

chains are linked by weak C12ꢀH12 π and πꢀπ interactions

3 3 3

(Figure 8e).

2.4. Model Testing and Electronic Structure of the N-

Phenylacetamidyl Radical. The production of amidyl radicals

during the photoinduced homolysis of the N-chloroamides

described above provides an opportunity to study their structure

and chemistry in the solid state. A substantial amount of existing

experimental and computational evidence has indicated that the

Π_Nradical state, having an unshared electron pair in the Nsp²

hybrid orbital and an unpaired electron in the orthogonal Np

orbital (Figure 1a), is the actual ground state of the acyclic amidyl

radicals¹⁰as well as of the cyclic succinimidyl radical.¹¹Prediction

of the electronic structure of the phenyl-substituted amidyl

intermediates created during the photolysis of the NCAs and

NCBAs, exemplified by the simplest radical, N-phenylacetamidyl

(Ph^•NAc), is far more complex because the degree of delocaliza-

tion of the spin between the limiting orthogonal and parallel

geometries relative to the phenyl π-electron system is expected

to affect the occupational preference of the unpaired spin

The polymorphism of 5a, where in the two polymorphs

the same molecule exists with diﬀerent amide-phenyl angles

(73.99(11)ꢀ in 5a-A, as compared to 69.52(8) in 5a-B), and it

also utilizes diﬀerent intermolecular interactions (halogen vs

hydrogen bonding), provides a basis to establish a more quanti-

tative correlation between the crystal structure and the solid-state

reactivity. Notably, the photoreactivity of the two forms is

quantitatively diﬀerent, but it also depends on the treatment of

the sample. During the course of the IR spectroscopic product

analysis, we found that the pressure exerted during the pellet

preparation had a very diﬀerent eﬀect on the reaction of the two

polymorphs. Regardless of whether it was ﬁrst pelleted with KBr

and irradiated in the pellet, or it was irradiated in pure state and

pelletized afterward, form 5a-B always aﬀorded the same pro-

duct. On the other hand, as shown in Figure S9 in the SI, when

irradiated in a pellet, 5a-A exhibited a profoundly diﬀerent IR

spectrum relative to the sample irradiated in a pure state.

7841

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

(Figure 1cꢀf).¹³Aimed at identification of the electronic struc-

ture of the intermediates created during the rearrangement of the

N-chloroacetanilide derivatives, and considering conditions that

could facilitate alteration of the orbital occupancy, we embarked

on the most comprehensive to date theoretical modeling of the

effects of various geometries and substituents on the relative

stability of the Π_Nand Σ_Nstates of Ph^•NAc.

Our calculations at the UB3LYP/6-31G(d) level excellently

reproduced the results obtained by Muchall et al.^12aon alkyla-

midyl radicals within 1 kJ mol^ꢀ1, indicating that a similar method

could be applied to the arylamidyl radicals. In the absence of

experimental gas-phase data on the geometry of Ph^•NAc, and to

independently test the reliability of the UB3LYP-calculated

geometry, we also optimized the geometry of Ph^•NAc using

the CASSCF method. We utilized the state-averaged approach

with two roots, corresponding to the ground state and the ﬁrst

excited state of equal weights. The structure was optimized for

the ground state (the ﬁrst root in the CASSCF(13,11)/6-31G(d)

calculation). Analysis of the wavefunction showed predominantly

single conﬁgurational character of the ground state, with 79%

weight of the leading conﬁguration. Inspection of the natural

occupation numbers revealed that the unpaired electron popu-

lates the orbital 36a (Figure 9), having the highest electron

density in the N 2p_yorbital and being signiﬁcantly delocalized

over the aromatic ring.

Notably, no signiﬁcant conjugation of the free electron with

the carbonyl π orbital could be observed. The calculated bond

distances closely reproduce those optimized at the UB3LYP level

(Table S5, SI), as expected from the dominant single conﬁgura-

tional character of the ground-state structure; the CꢀC distances

in the two structures are practically identical, and the CꢀN bond

is by 0.007 Å longer relative to the UB3LYP value. The only

exception is the CdO distance, which is shorter by 0.022 Å in

the CASSCF model. The largest diﬀerence in the angles was

observed for Θ(C6C1N7C8), which is smaller by 6.1ꢀ in the

CASSCF relative to the UB3LYP-optimized structure (the

atom labels refer to Scheme 1). In the CASSCF model the amide

group is more twisted, as evidenced by the larger (Δ = 18.7ꢀ)

Θ(C1N7C8O9) angle. In view of the concurrence of the

UB3LYP and CASSCF results, the former method was selected

for an in-depth analysis as an economically more advantageous

approach.

Prior to an in-depth analysis of the geometry and substituent

eﬀects on the radical stability, it is convenient to determine the

energy gaps between the limiting orthogonal and parallel geo-

metries in the two major orbital states of the parent radical.

Under the assumption of a C_ssymmetry, the Π states are of ²A⁰⁰

2

symmetry, and the ∑ states adopt A⁰symmetry. For all sym-

metry-constrained models, we calculated (UB3LYP/6-31G(d))

the relative stabilities of the cis-amide and trans-amide conforma-

tions (with respect to the relative disposition of the phenyl ring

and the methyl group), apart from the Σ_^state, for which we

could not locate the cis-isomer. The calculated energies are

displayed in Tables 2 and 3 together with that of the fully

optimized (unconstrained) structure, and the respective spin

Table 3. Electronic Energies E_el(hartrees) and Energy Dif-

ference between the Π and Σ_^States, ΔE_el(ΣꢀΠ)(kJ mol^ꢀ1

of the trans-Conformers of some para-Substituted Ph^•NAc

Radicals Constrained to C_sSymmetry Calculated with Dif-

ferent Computational Methods

)

radical method

Ph^•NAc

Π (2-A⁰⁰)

Σ_^(2-A⁰)

ΔE_el(ΣꢀΠ)

UB3LYP/6-31G(d)

G3(MP2)-RAD

CBS-RAD

ꢀ439.60588

ꢀ439.03849

ꢀ438.98483

ꢀ438.70146

ꢀ439.59224

ꢀ439.02029

ꢀ438.96585

ꢀ438.67956

35.8

47.8

49.8

57.5

ROMP2

p-CN-Ph^•NAc

UB3LYP/6-31G(d)

G3(MP2)-RAD

p-NMe₂-Ph^•NAc

UB3LYP/6-31G(d)

G3(MP2)-RAD

ꢀ531.84713

ꢀ531.16739

ꢀ531.83335

ꢀ531.14847

36.2

49.7

ꢀ573.58494

ꢀ572.854

ꢀ573.56910

ꢀ572.83403

41.6

52.4

p-F-Ph^•NAc

UB3LYP/6-31G(d)

G3(MP2)-RAD

ꢀ538.84075

ꢀ538.19413

ꢀ538.82619

ꢀ538.17486

38.2

50.6

p-NO₂-Ph^•NAc

UB3LYP/6-31G(d)

G3(MP2)-RAD

Figure 9. Semioccupied molecular orbital of the trans-N-phenylaceta-

midyl radical (Ph^•NAc) calculated at the CASSCF(13,11)/6-31G(d)

level of theory.

ꢀ644.10340

ꢀ643.33232

ꢀ644.09092

ꢀ643.31431

32.8

47.3

Table 2. Electronic Energies, E_el(hartrees), and Energy Diﬀerence between the Π_Nand Σ_NStates, ΔE_el(ΣꢀΠ)(kJ mol^ꢀ1

Calculated with the B3LYP/6-31G(d) Method for the N-Phenylacetamidyl Radical (Ph^•NAc) with a C_sSymmetry^a

)

conformation

state

E_el(UB3LYP/6-31G(d))

ΔE_el(ΣꢀΠ)

ÆS²æ

NImag^b

parallel, trans-

Π (²A⁰⁰)

Σ (²A⁰)

Π (²A⁰⁰)

Σ (²A⁰)

Π (²A⁰⁰)

Σ (²A⁰)

Π (²A⁰⁰)

(²A)

ꢀ439.60588

0.0

0.7805

0.7554

0.7858

0.7546

0.7667

0.7762

0.7706

0.7835

1

2

0

ꢀ439.58328

59.4

0.0

parallel, cis-

ꢀ439.59188

ꢀ439.56854

61.3

0.0

orthogonal, trans-

ꢀ439.58270

ꢀ439.59224

ꢀ28.1

orthogonal, cis-

ꢀ439.58008

optimized structure

ꢀ439.60863 (ꢀ439.46709)

ꢀ7.2^c

^aThe total energy of the fully optimized structure is given in parentheses. ^bNumber of imaginary frequencies. ^cRelative to the parallel trans-conformer.

7842

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Figure 10. Total spin densities in the Π and Σ states for parallel and

orthogonal conformations of the cis- and trans-isomers, based on the

UB3LYP/6-31G(d) calculations. Optimization of the cis-∑_^structure

led to the corresponding trans-isomer as the minimum.

densities are plotted in Figure 10. It should be noted that under

the C_ssymmetry constraint, the calculated structures correspond

to saddle points on the potential energy surface and, therefore,

only their electronic energies are compared.

Along with what is expected from the conformational land-

scape of the amide functional group, the results in Table 2 show

that in all cis/trans pairs, the trans-conformers are consistently

more stable than their respective cis counterparts. Also consistent

with the expectations based on the earlier results on acyclic

amidyls is the signiﬁcantly higher stability of the Π states relative

to the respective Σ states of the planar radical, the diﬀerence

being ∼60 kJ mol^ꢀ1in both (cis and trans) conformations. To

our surprise, however, we found that in the case of the trans-

isomer of the orthogonal Π/Σ pair, the occupational stability of

the two states is inverted, that is, the Σ_^state is ∼28 kJ mol^ꢀ1

more stable than the Π_^state (Table 2). This result indicates

that, for the orthogonal disposition of the amide group and the

phenyl ring, the phenylamidyl radicals can indeed exist as Σ

radicals because of the stronger stabilization by conjugation of

the π-electron sextet with the unpaired electron (relative to the

unshared electron pair) when it occupies the Nsp²orbital. This

result has important implications for the targeted design of

precursors of Σ amidyl radicals: in the absence of other electronic

eﬀects, such precursors should have the amidyl group in the

trans-conformation orthogonal to the phenyl ring. To verify

this hypothesis, we prepared molecules 6a and 7a where the

N-chloroamidyl group is locked by the double ortho-tert-butyl

substitution in its trans-conformation, as conﬁrmed by the XRD

structure of 6a (note that, although the para-methyl derivative 4a

is also constrained by symmetry to the orthogonal conformation,

due to the N-chloro-substitution of the precursor, in the absence

of transꢀcis isomerization following the photolysis, the radical is

expected to exist in the crystal in the cis-conformation). As

Figure 11. (a) Potential energy surface of the trans-N-phenylacetamidyl

radical (Ph^•NAc) scanned as a rigid species with the UB3LYP/6-31G(d)

method. (b, c) Two views on the potential energy surfaces of the ground

state and the ﬁrst excited state calculated for Ph^•NAc at the CASSCF-

(13,11)/6-31G(d) level as a relaxed potential energy surface scan over

the valence angle θ(C8N7C1) and the dihedral angle θ(C8N7C1C2) as

internal coordinates.

described above, upon photolysis 6a aﬀorded a nitrogen-centered

radical (Figure 3b), which, considering the inability of the amide

group to isomerize to its cis-conformation due to the steric bulk of

the tert-butyl groups, can now be identiﬁed as a Σ_^radical state.

The stabilizing eﬀect appears as a pronounced shortening of

the bond between the aromatic ring and the radical center

(C1ꢀN7 in Scheme 1) relative to structures where such inter-

action is not possible (see Tables S5 and S6 in the SI). Inspection

of the geometries of the calculated C_sstructures reveals that on

going from planar to the orthogonal structures, that is, by

increasing the dihedral angle between the phenyl ring and the

acetamidyl group (θ(C6C1N7C8)), the amide valence angle at

the nitrogen atom (θ(C1N7C8)) also opens up. This trend is

illustrated in Figure 11a, which shows the three-dimensional

7843

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

Table 4. Selected Geometry Parameters of Substituted trans-N-Phenylacetamidyl Radicals^a

substituents^b

R¹

R²

R³

type

d(C1ꢀN7)

d(N7ꢀC8)

d(C8ꢀO9)

θ(C6C1N7C8)

θ(C1N7C8O9)

θ(C1N7C8)

H

Π

1.344

1.339

1.341

1.344

1.338

1.342

1.341

1.332

1.320

1.342

1.340

1.339

1.322

1.323

1.314

1.374

1.365

1.394

1.398

1.388

1.394

1.398

1.391

1.399

1.372

1.384

1.375

1.371

1.381

1.336

1.334

1.342

1.431

1.437

1.225

1.223

1.230

1.226

1.223

1.227

1.222

1.231

1.222

1.229

1.233

1.227

1.238

1.241

1.234

1.215

1.213

9.7

46.0

47.9

31.7

41.9

49.3

38.0

49.9

43.5

55.8

38.9

26.5

47.1

23.1

23.7

0.1

122.7

122.6

122.4

122.6

122.8

122.5

122.7

129.8

131.2

129.8

126.1

127.7

152.2

151.2

148.2

153.7

119.1

120.3

H

CN

Π

8.6

H

NMe₂

F

Π

10.9

10.0

8.4

H

Π

H

CHO

OCH₃

NO₂

NMe₂

H

Π

H

Π

10.8

8.2

H

Π

Me

CF₃

TMS

Ph

Me

CF₃

TMS

Ph

Π

27.4

23.5

26.8

29.3

35.0

70.1

69.6

93.1

52.6

93.1

93.0

Π

NMe₂

H

Π

t-Bu

NO₂

t-Bu

NO₂

H

Σ_^

Π_^

t-Bu

NMe₂

NO₂

H

39.3

0.0

NO₂

Π_^

0.0

ꢁ

^aThe bond distances (d) are given in angstr€oms, and the dihedral and valence angles (θ) in degrees (the atom labels refer to Scheme 1). ^bR¹and R²

denote the substituents at the C2 and C6 atoms, whereas the R³denotes the substituent attached to the C4 atom in Scheme 1.

potential energy surface of trans-Ph^•NAc constructed at the

UB3LYP/6-31G(d) level by twisting the (otherwise rigid)

molecule from its planar to the orthogonal conformation (by

changing the amideꢀphenyl dihedral angle from 0 to 90ꢀ) while

expanding the amide valence angle (from 105 to 175ꢀ). At this

level of theory, the predicted minimum-energy structure has a

phenylꢀacetamidyl dihedral angle of ∼10ꢀ and an amide valence

angle of ∼125ꢀ. These values are supported by full optimization

of the parent radical, which ends up in a minimum having

θ(C6C1N7C8) and θ(C1N7C8) of 9.7 and 122.7ꢀ, respectively.

It is important to note that this structure, with essentialy pseudo-

2a⁰⁰conﬁguration, is only 7.4 kJ mol^ꢀ1more stable than the

structure of the symmetry-constrained trans-Π radical.

The potential energy surfaces of the ground and excited states

were obtained by partial optimization of each structure, followed

by state-average single point calculations over the two (Π and Σ)

states. Inspection of the conﬁgurations at each surface shows that

at lower dihedral angles (θ(C8N7C1C2) < 60ꢀ) the ground-

state conﬁguration is best described as a Π radical, irrespective of

the value of the valence angle. By increasing the dihedral angle

above 60ꢀ, for a valence angle larger than 155ꢀ the ground state

assumes Σ character. However, no characteristic Σ minimum was

found on the surface of the ground state. This is in accordance

with the Π character of the parent phenyl amidyl radical

observed experimentally.⁴The results also suggest that formation

of the Σ-radical would require additional stabilization of the

perpendicular conformation through electronic and/or steric

eﬀects. In that regard, substitution of the phenyl ring by sterically

demanding groups in the ortho position relative to the site of the

amidyl group attachment is particularly promising.

and Σ_^states the free electron appears signiﬁcantly delocalized

over the aromatic ring, even more than over the carbonyl group,

as expected based on the energy diﬀerences between the

corresponding fragment orbitals.

2.5. Substituent Effects on the Separation of the Σ and Π

States. As concluded from the energies in Table 2, the most stable

geometry of the Ph^•NAc radical in Figure 10 is the trans-isomer of

the Π state (trans-Π ), showing that the stabilization by inter-

action between the Np orbital carrying the unpaired electron and

the π-electron sextet in the planar structure is the most effective.

The second most stable state, separated from the former by a gap

of 35.8 kJ mol^ꢀ1, is trans-Σ_^. This result is not surprising because

in both structures the interaction between the free electron on the

radical center and the aromatic π system is symmetry-allowed.

Having identiﬁed the most stable conformers and the structure

of the ground state of the unsubstituted Ph^•NAc, we now turn to

its para-substituted derivatives, aiming at evaluation of the eﬀects

of aromatic substitution on the Σ_^ꢀΠ energy gap. For that

purpose, the structures of several para-substituted radicals con-

strained to C_ssymmetry were initially optimized at the same level

of theory (UB3LYP/6-31G(d)) as the parent radical Ph^•NAc.

Subsequently, the calculated energies of all radicals were reﬁned

with more advanced methods, CBS-RAD, ROMP2/6-311þG-

(3df,2p)//UB3LYP/6-31G(d) (ROMP2), and G3(MP2)-RAD,

which have been purposefully designed to reproduce thermo-

chemical data of radicals with high accuracy.²⁵It should be

emphasized that comparison of the results obtained by using the

UB3LYP/6-31G(d), ROMP2, and G3(MP2)-RAD methods is

particularly straightforward because they are based on the same

level of geometry optimization. This fact is of importance for the

amidyl radicals, for which it has been demonstrated that they are

particularly demanding in that respect.^26,27

Analysis of the calculated spin densities in Figure 10 is also

revealing: it shows a fraction of the spin density on the oxygen

atom of the carbonyl goup in all structures, indicating smaller

contribution from the O-centered states. In addition, in the Π

The energies and selected geometrical details of the substi-

tuted amidyls are listed in Tables 3 and 4 respectively, and the

7844

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

energies for all other considered conformers are summarized in

Tables S7 and S8 in the SI. The Σ_^ꢀΠ energy gaps of Ph^•NA

calculated at higher levels of theory are larger than the respective

UB3LYP values, with the largest discrepancy observed with the

ROMP2 model. Notably, G3(MP2) and CBS-RAD methods

aﬀorded similar results, but since spin contamination in all

structures calculated with the latter method was larger than

1.00, these results were excluded from the calculations. Albeit

limited to only four computational models, these results show

that all employed methods predict the same stability order of the

Σ_^and Π states. The results also suggest that the calculations

performed with the UB3LYP/6-31G(d) method aﬀord reliable

estimates of the diﬀerence in energies and can be safely used to

predict the ground-state structures of substituted Ph^•NAc radi-

cals and the thermochemical properties of their higher energy

states. As a second and more important conclusion from these

results, we found that with one exception (the strongly electron-

donating NMe₂group) the eﬀect of para-substitution of the

aromatic ring on the Σ_^ꢀΠ energy gap is very small (Table 3).

Therefore, we conclude that, similar to what has been observed

with the phenylaminyl radicals,¹³the phenylamidyl radicals (or,

at least, their trans-conformers) will exhibit a non-Hammet

behavior (“class S”) with both push and pull substituents having

similar eﬀects on the spin density at the nitrogen.

Figure 12. Schematic presentation of the total spin densities of the

N-phenylacetamidyl radical (a) and its 2,6-dimethyl-4-NMe₂(b), 2,6-di-

t-butyl-phenyl (c), and 2,4,6-tri-NO₂(d) derivatives, calculated by the

B3LYP/6-31G(d) method (the contours are plotted as 0.07 e).

Despite the fact that para-phenyl substitution alone is ex-

pected to have a small eﬀect on the Σ_^ꢀΠ energy gap of

Ph^•NAc, we anticipated that the relative stability could still be

signiﬁcantly aﬀected by cumulative substituent eﬀects, inclusive

of eventual steric restrictions. To test such a hypothesis, we

optimized the structures of a series of Ph^•NAc radicals substi-

tuted with typical σ-donors (Me, t-butyl, SiH₃, TMS), σ-acceptor

(CF₃), π-donors (OMe, NMe₂), π-acceptors (CN, CHO, and

NO₂), π-donor/σ-acceptor (F), and the phenyl group. The

substituents were selected so as to test the premise across a wide

range of interactions with the aromatic ring. The optimizations

were carried out at the UB3LYP/6-31G(d) level, and the

geometry minima were conﬁrmed in all cases.

as in Ph^•NAc and its para-substituted derivatives. The spin

distribution, as exemplified by 2,6-dimethyl-4-NMe₂-Ph^•NAc

(Figure 12b) also ressembles that of the Π radicals. Notably,

the radicals with bulky TMS groups and phenyl rings at ortho-

positions also belong to this group.

Group 3. This group encompasses radicals with tert-butyl

groups in the ortho-position, for which it was predicted that they

exist in the Σ ground state. This is indicated by the dihedral angle

(C6C1N7C8) and the valence angle (C1N7C8) at the radical

center in 2,4,6-tri-tert-butyl- and 2,4,6-tri-tert-butyl derivatives,

which for both radicals assume values close to 70 and 150ꢀ,

respectively. Noteworthy, substitution of the para-position in the

2,6-di-tert-butyl derivative with the ꢀNMe₂group leads to

practically orthogonal alignment of the amidyl group relative to

the aromatic ring, thus, enabling more efficient delocalization of

the free electron over the aromatic ring. Conversely, nitro

substitution at the same position induces decrease of the dihedral

angle to 52.6ꢀ and simultaneous twist around CNꢀCO within

the acetamidyl fragment. Prompted by this result, we approached

modeling of the structure of the N-(2,6-di-tert-butyl-phe-

nyl)benzamidyl radical, where the CH₃group in the aceta-

midyl fragment is replaced by the more sterically demanding

phenyl group. The resulting structure was very similar to that of

the N-(2,6-di-tert-butyl-phenyl)acetamidyl radical, with the

C6C1N7C8 and C1N7C8 angles at 70.3 and 150.8ꢀ, respec-

tively. All these radicals exhibit spin density distribution char-

acteristic for Σ_^radicals, as illustrated with the N-(2,6-di-tert-

butyl-phenyl)acetamidyl radical in Figure 12c. In turn, this result

strongly supports the above assumptions that the presence of

bulky tert-butyl groups in the ortho-positions is a critical pre-

requisite for stabilization of the Σ state.

The optimized geometries are listed in Table 4 (the total and

ZPVE-corrected energies, and coordinates of the optimized

structures are deposited as Table S9 in the SI). Perusal of the

results shows that the substituted phenylamidyl radicals can be

classiﬁed into four groups:

Group 1. This group includes the parent radical Ph^•NAc and its

para-substituted analogues. These radicals are characterized by a

value of the dihedral angle θ(C6C1N7C8) <10ꢀ and angle at the

radical center ∼122ꢀ. These values are in excellent agreement

with those predicted by the search of energy surface of Ph^•NAc

shown in Figure 11a. Furthermore, these radicals exhibit pro-

nounced twisting around the CNꢀCO bond (31.7ꢀ49.9ꢀ). In

turn, this result shows that the effects of both electron pull and

electron push groups in the para-position on the geometry of

Ph^•NAc are only marginal. This also holds true for the spin

density distribution, which is characteristic for a Π radical, with

the largest fraction residing at the nitrogen atom (Figure 12a).

Group 2. This group includes radicals having Me, TMS, or CF₃

at the ortho-positions and a hydrogen atom or NMe₂group at the

para-position of the phenyl ring with respect to the site of

substitution of the amidyl group. The only major difference

to the radicals in Group 1 is the larger (∼10ꢀ20ꢀ) dihedral

angle C6C1N7C8, which is presumably due to the presence

of substituents in the ortho-aromatic positions. The twisting

around NCꢀCO, however, is of the same order of magnitude

Group 4. In radicals with two and three nitro groups, the

amidyl group assumes practically orthogonal orientation to the

aromatic ring, with the angle at the nitrogen atom being close to

that in the Π radicals. Contrary to the other radicals, while most

of the spin density is localized at the radical center, partially it is

7845

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

delocalized over the ortho-nitro groups. This strongly indicates

that the free electron in these radicals resides in the plane of the

aromatic ring, thus, preventing interaction with the π-system of

the phenyl group, as expected for the Π_^radicals. To the best of

our knowledge, neither theoretical nor experimental data have

been reported yet on stable Π_^amidyl radicals.

4. EXPERIMENTAL SECTION

4.1. Synthesis, Purification, Characterization, and Crystal-

lization. All synthetic and recrystallization procedures (except

for the synthesis of the amide precursors) were performed in

the dark, under reduced light or under red safelight (>660 nm).

The precursors of 6a and 7a, N-(2,4,6-tri-tert-butylphenyl)-

acetamide and N-(2,4,6-tri-tert-butylphenyl)benzamide, were

prepared by sonochemical synthesis from 2,4,6-tri-tert-butylani-

line and acetyl chloride or benzoyl chloride, respectively. The

NCAs and NCBAs were prepared by chlorination of the respec-

tive anilides and benzanilides with trichloroisocyanuric acid

(TCA),²⁸following a procedure described for chlorination of

amides, lactams, and carbamates,²⁹as exemplified with acetani-

lide below (detailed synthetic procedures for the other chlori-

nated compounds and additional analytical data are deposited as

SI). In the case of the meta-methyl acetanilide, the procedure

always resulted in the doubly substituted (N,p-) product; there-

fore, chlorination with NaOCl was employed. The radicals

created by light or additional heating render the procedures for

preparation of these compounds very sensitive to the reaction

conditions. Especially, the solvent composition, exposure to

room light and the storage temperature are critical to obtain

and preserve the products in pure state. The products were

stored at ꢀ20 ꢀC and were generally stable in the solid state

(after standing at room temperature, 2a turns into a semisolid

mass), but on prolonged storage (months), some of them

showed very slow isomerization, occasionally accompanied by

off-white coloration. The crystallization was performed from

light-protected solutions in petroleum ether below room tem-

perature, eventually with the addition of a very small amount of

acetone. The repeated attempts to crystallize multiply purified/

dried 7a were not successful and the compound was always

obtained as a transparent, glassy solid.

4.2. X-ray Diffraction. The XRD data were collected (ω-

scans) with APEX2 diffractometer (Bruker AXS),³⁰using Mo KR

X-rays monochromated by a confocal multilayer mirror and CCD

detector. The integrated and scaled data were empirically cor-

rected for absorption effects with SADABS.³¹Structural solution

and FMLS refinements based on F²were performed with

SHELXS-97 and SHELXL-97.³²For the structures before irradia-

tion, all non-H atoms were assigned anisotropic ADPs, and the

hydrogen atoms were either included as riding bodies or refined

from the electron density (Table S1). The data quality of 3a could

not be improved further due to instability of the crystal and the

sensitivity of the sample to X-rays. Although in this particular case

the data quality was less than optimal, the accuracy was sufficient

for the discussion of the general geometrical features and com-

parison with the other amides studied in this work.

3. CONCLUSIONS

With a series of N-chlorinated acetanilides and benzanilides,

here we demonstrated a new photoinduced rearrangement in the

solid state, whereby aromatic N-chloroamides are eﬃciently

converted to chloroaromatic amides. As evidenced by the crystal

structures of the N-chloroamide precursors, one of which was

obtained as two polymorphs of diﬀerent reactivity, the product

composition depends on the substituent eﬀects, molecular pack-

ing, excitation wavelength, and irradiation time. The reaction

outcome is also determined by the diﬀusion of the reactive species

created by NꢀCl photolysis, which can be in the form of nascent

chlorine or (after recombination) dichlorine. With strong UV

excitation, the transformation proceeds mainly with nascent

chlorine, and as a result, the reaction is selective and proceeds

rapidly, being completed typically within several minutes. Slow-

ing of the kinetics by weak and prolonged UV excitation activates

side reactions as a result of recombination of a small portion of the

chlorine atoms to molecular chlorine. The molecular chlorine

reacts partially, and the remaining escapes from the crystal.

Importantly, the reaction can proceed to completion in the solid

state even by using direct natural sunlight, thus, providing a

solventless route to a plethora of compounds based on the

physiologically relevant para-chlorophenylamide moiety.

The computational results presented in this work provide

additional evidence for the conclusions reached in a recent

experimental study⁴that the ground-state structure of the

Ph^•NAc radical corresponds to a twisted Π radical. The amidyl

group deviates (∼10ꢀ) from the plane of the phenyl ring, and the

calculated spin density distribution resembles that of an entirely

planar Π-type structure. The acetamidyl group is twisted around

the CNꢀCO bond by ∼40ꢀ. The Π state is also energetically

favored in radicals with CN, F, OMe, NMe₂, or CHO group in

para-disposition relative to the phenyl ring, as well as in radicals

with Me, CF₃, TMS, and phenyl groups adjacent to the acet-

amidyl group. On the other hand, the phenyl ring and the amidyl

group in the molecular structure of the radical with tert-butyl

groups at the ortho-positions are juxtaposed at 70ꢀ, and the

structure resembles the Σ_^state. This conclusion is supported by

the calculated spin density distribution. Introducing a dimethy-

lamino group at the para-position of the aromatic ring in such

radical results in orthogonal alignment of the two subunits.

Finally, for radicals with two or three nitro groups at the phenyl

ring, two of which occupy the ortho-positions, the Π_^state

appears to be the most stable structure. Apart from the Π and Σ_^

radicals, where the spin density is also distributed over the

aromatic ring, in all radicals the main fraction of the spin density

resides at the nitrogen radical center and, to a smaller extent, at

the carbonyl oxygen atom.

4.3. Computational Details. Geometry optimization was

performed using the density functional theory (DFT) with the

three-parameter hybrid functional UB3LYP³³and the 6-31G(d)

basis set and further checked by frequency analysis at the same

level of theory and the same basis set. The 6-31G(d) basis set was

chosenasa compromisebetweenaccuracyand computational cost

for the size of moderately large organic molecules which includes

polarization functions on non-hydrogen atoms that are required

for an accurate description of radical electronic structure.³⁴

Geometries for the parent radical obtained by this approach were

found to be practically identical to those calculated previously with

the 6-31G(d,p) basis set.⁴The nature of the optimized structures

was checked by frequency analysis. The geometry of the parent

The results presented here aid our understanding of the

substituent and geometry factors that aﬀect the electronic

structure of the arylamidyl radicals. The inferences have strong

predictive value in terms of the design of precursors of the target

radicals in a particular electronic state, based on a judicious

choice of substituents at the aromatic ring.

7846

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

radical was also calculated using UQCISD method within the

CBS-RAD(UQCISD,B3LYP)³⁵calculation protocol and with the

CASSCF(13,11)/6-31G(d) method.³⁶The active space in the

CASSCF calculations comprised 13 electrons in 11 orbitals

(3 highest occupied and 3 lowest unoccupied benzene π-orbitals,

π and π* (CdO) orbitals, Nn_πand Nn_σorbital, and O_norbital).

For refinement of the energies calculated for the UB3LYP/6-

31G(d) optimized structures ROMP2/6-311þG(3df,2p)//

UB3LYP/6-31G^37,38(ROMP2) and G3(MP2)-RAD³⁹calcula-

tion models were used. The average spin expectation values ÆS²æ of

thecalculated structuresrangefrom 0.76 to0.78, which suggestsno

significant error due to spin contamination. Electron spin densities

were obtained from Mulliken population analysis. The calculations

were performed using Gaussian 03,⁴⁰Gaussian 09,⁴¹or Molcas⁴²

quantum chemistry programs. For visualization of the geometries

and electron and spin densities Molden⁴³and GaussView 4.1

programs were used. The geometries of all stationary points

described in the paper are provided as Supporting Information.

COE program, “the Global Education and Research Center for

Bio-Environmental Chemistry” from the Ministry of Education,

Culture, Sports, Science and Technology, Japan. We acknowl-

edge the generous allocation of the computing time at cluster

Isabella in the Computing Center of the University in Zagreb. We

thank the anonymous reviewer who provided very detailed

comments that helped to improve greatly the quality of this paper.

’ REFERENCES

(1) (a) Tanaka, K. Solvent-Free Organic Synthesis; Wiley-VCH: Wein-

heim, 2003. (b) Ramamurthy, V. Tetrahedron 1986, 42, 5753–

5839. (c) Ramamurthy, V.; Eaton, D. F. Chem. Mater. 1994, 6,

1128–1136. (d) Ramamurthy, V.; Schanze, K. S. Solid State and Surface

Photochemistry; Marcel Dekker: New York, 2000. (e) Turro, N. J. Modern

Molecular Photochemistry; University Science Books: Sausalito, 1991.

(f) Turro, N. J., Ramamurthy, V., Scaiano, J. C. Modern Molecular Photo-

chemistry of Organic Molecules; University Science Books: Sausalito, 2010.

(2) Reviews: (a) Ramamurthy, V.; Venkatesan, K. Chem. Rev. 1987,

87, 433–481. (b) Garcia-Garibay, M. A. Acc. Chem. Res. 2003,

36, 491–498. (c) Ohashi, Y. Acc. Chem. Res. 1988, 21, 268–274.

(d) Drickamer, H. G.; Bray, K. L. Int. Rev. Phys. Chem. 1989,

8, 41–64. Selected examples: (e) Natarajan, A.; Tsai, C. K.; Khan,

S. I.; McCarren, P.; Houk, K. N.; Garcia-Garibay, M. A. J. Am. Chem. Soc.

2007, 129, 9846–9847. (f) Boldyreva, E. V. Russ. J. Coord. Chem. 2001,

27, 1–28. (g) Boldyreva, E. V.; Podberezskaya, N. V.; Virovets, A. V.;

Burleva, L. P.; Dulepov, V. E. J. Struct. Chem. 1993, 34, 128–138.

(h) Masciocchi, N.; Kolyshev, A. N.; Dulepov, V. E.; Boldyreva, E. V.;

Sironi, A. Inorg. Chem. 1994, 33, 2579–2585. (i) Vithana, C.; Uekusa, H.;

Sekine, A.; Ohashi, Y. Cryst. Growth Des. 2005, 1755–1760. (j) Kyono, A.

J. Photochem. Photobiol., A 2007, 189, 15–22. (k) Naumov, P.; Makreski,

P.; Petruꢀsevski, G.; Runꢀcevski, T.; Jovanovski, G. J. Am. Chem. Soc. 2010,

132, 11398–11401.

(3) (a) Kaftory, M. In Organic Solid State Reaction; Toda, F., Ed.;

Kluwer Academic Publishers: United Kingdom, 2002; pp 47ꢀ68.

(b) Handelsman-Benory, E.; Botoshansky, M.; Greenberg, M.; Shteiman,

V.; Kaftory, M. Tetrahedron 2000, 56, 6887–6897. (c) Sukenik, C. N.;

Bonapace, J. A. P.; Mandel, N. S.; Bergman, R. G.; Lau, P.-Y.; Wood, G. J.

Am. Chem. Soc. 1975, 97, 5290–5291. (d) Sukenik, C. N.; Bonapace,

J. A. P.; Mandel, N. S.; Bergman, R. G.; Lau, P.-Y.; Wood, G. J. Am. Chem.

Soc. 1977, 99, 851–858. (e) McNeill, K.; Andersen, R. A.; Bergman, R. G.

J. Am. Chem. Soc. 1995, 117, 3625–3626. (f) Greenberg, M.; Shteiman,

V.; Kaftory, M. Acta Crystallogr. 2001, B57, 428–434. (g) Venugopalan,

P.; Venkatesan, K.; Klausen, J.; Novotny-Bregger, E.; Leumann, C.;

Eschenmoser, A.; D€unitz, J. D. Helv. Chim. Acta 2004, 74, 662–669.

(4) Naumov, P.; Sakurai, K.; Tanaka, M.; Hara, H. J. Phys. Chem. B

2007, 111, 10373–10378. Based on the position of the unpaired electron

on the nitrogen atom, in the original report we named the species that is

obtained by homolysis of the NꢀCl bond of 1a “aminyl radical”. From

the viewpoint of the adjacency of a carbonyl group, however, which

aﬀects signiﬁcantly the electronic properties of the radical, a more

suitable term to describe this and similar species would be “acylaminyl”

or “amidyl radical”. The latter term, which is also consistent with the

generally accepted terminology in more recent organic chemistry

reports, was adopted in this article.

’ ASSOCIATED CONTENT

S

Supporting Information. Experimental details of the

b

synthetic and characterization procedures and results, HPLC

product analysis (Table S1), complete crystallographic data in

tabular (Table S2) and CIF format, complete refs 40 and 41,

product analysis of 5a (NMR, Table S3), intramolecular hydro-

gen bonds (Table S4), optimized geometry of the N-phenylace-

tamidyl radical (Table S5), theoretical geometry of the N-

methylacetamidyl radical (Table S6), electronic and total en-

ergies of the Π and Σ states at the UB3LYP/6-31G(d) level

(Table S7), electronic and total energies of the Π and Σ states at

the G3(MP2)-RAD level (Table S8), Cartesian coordinates of

the optimized structures (Table S9), IR spectra of irradiated

compounds (Figure S1), NMR spectra of irradiated compounds

(Figure S2), eﬀects of the excitation on the products (Figure S3),

results from the HPLC product analysis (Figure S4), eﬀects of

sunlight on the IR spectra of 1a (Figure S5), eﬀects of sunlight on

the IR spectra of 5a (Figure S6), NMR product analysis of form A

of 5a (Figure S7), NMR product analysis of form B of 5a (Figure

S8), eﬀects of the preparation on the IR spectra of 5a (Figure S9),

and results from the theoretical calculations (Figures S10ꢀS67).

This material is available free of charge via the Internet at http://

pubs.acs.org.

’ AUTHOR INFORMATION

Corresponding Author

*E-mail: pancenaumov@gmail.com (P.N.); mmaksic@emma.

irb.hr (M.E.-M.).

Notes

^‡On leave from the Chemistry Department, Arts and Sciences

Faculty, Nevsehir University, Turkey.

(5) Porter, C. W.; Wilbur, P. J. Am. Chem. Soc. 1927, 49, 2145–2149.

(6) (a) Bender, G. Ber. 1886, 18, 293–295. (b) Chattaway, F. D.;

Orton, K. J. P. J. Chem. Soc. 1899, 75, 1046–1048. (c) Hodges, F. W.

J. Chem. Soc. 1933, 240–246. (d) Ayad, K. N.; Beard, C.; Garwood, R. F.;

Hickinbottom, W. J. J. Chem. Soc. 1957, 2981–2986. (e) Olson, A. R.;

Halford, R. S.; Hornel, J. C. J. Am. Chem. Soc. 1937, 59, 1613–1615.

(f) Badr, M. Z. A.; Aly, M. M.; Salem, S. S. Tetrahedron 1978,

34, 123–125. (g) Fontein, C. J. Rec. Trav. Chim. Pays-Bas 1928,

47, 635–667. (h) Olson, A. R.; Porter, C. W.; Long, F. A.; Halford,

R. S. J. Am. Chem. Soc. 1936, 58, 2467–2472. (i) Mathews, J. H.;

Williamson, R. V. J. Am. Chem. Soc. 1923, 45, 2574–2584. (j) Blanksma,

J. J. Rec. Trav. Chim. Pays-Bas 1902, 21, 366–373.

’ ACKNOWLEDGMENT

This work is dedicated to Academician Gligor Jovanovski, the

leader of the structural chemistry in Macedonia, on the occasion of

his 65th birthday. The work was ﬁnancially supported by the

Ministry of Science, Education and Sport of Croatia (Project No.

098-0982933-2920), the Japanese Science and Technology

Agency, JST (FRBforGYRprogram), and partially also by a Global

7847

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

The Journal of Physical Chemistry A

ARTICLE

(7) (a) Mackiewicz, P.; Furstoss, R. Tetrahedron 1978,

34, 3241–3260. (b) Konig, T.; Hoobler, J. A.; Klopfenstein, C. E.;

Hedden, G.; Sunderman, F.; Russell, B. R. J. Am. Chem. Soc. 1974,

96, 4573–4577. (c) Baird, N. C.; Kathpal, H. B. J. Am. Chem. Soc. 1976,

98, 7532–7535. (d) Dewar, M. J. S.; Pakiari, A. H.; Pierini, B. J. Am.

Chem. Soc. 1982, 104, 3243–3244. (e) Forrester, A. R.; Johansson, E. M.;

Thomson, R. H. J. Chem. Soc., Perkin Trans. 1 1979, 1112–1119.

(8) (a) Rees, M. D.; Hawkins, C. L.; Davies, M. J. J. Am. Chem. Soc.

2003, 125, 13719–13733. (b) Rees, M. D.; Davies, M. J. J. Am. Chem. Soc.

2006, 128, 3085–3097. (c) Robaszkiewicz, A.; Bartosz, G.; Soszynski, M.

Mech. Ageing Dev. 2008, 129, 572–579. (d) Robaszkiewicz, A.; Bartosz,

G.; Soszynski, M. Free Radical Res. 2008, 42, 30–39.

(9) Percec, V.; Grigorias, C. J. Polym. Sci. 2005, 43, 5283–5299.

(10) (a) Lessard, J.; Griller, D.; Ingold, K. U. J. Am. Chem. Soc. 1980,

102, 3262–3264. (b) Sutcliﬀe, R.; Ingold, K. U.; Lessard, J. J. Am. Chem.

Soc. 1981, 103, 7685–7686. (c) Sutcliﬀe, R.; Griller, D.; Lessard, J.;

Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 624–628. (d) Baird, N. C.;

Taylor, K. F. Can. J. Chem. 1980, 58, 733–738. (e) Danen, W. C.; Gellert,

R. W. J. Am. Chem. Soc. 1972, 94, 6853–6854. (f) Brown, C.; Lawson,

A. J. Tetrahedron Lett. 1975, 3, 191–194.

(11) (a) Lund, A.; Samskog, P. O.; Eberson, L.; Lunell, S. J. Phys.

Chem. 1982, 86, 2458–2462. (b) Kasai, P. H. J. Am. Chem. Soc. 1992,

114, 2875–2880. (c) Lind, J.; Merenyi, G. In N-Centered Radicals;

Alfassi, Z. B., Ed.; John Wiley: Chichester, 1998; pp 563ꢀ575.

(d) Apeloig, Y.; Schreiber, R. J. Am. Chem. Soc. 1980, 102, 6146–6147.

(12) (a) Muchall, H. M.; Werstiuk, N. H.; Lessard, J. J. Mol. Struct.

(THEOCHEM) 1999, 469, 135–142. (b) Glover, S. A.; Goosen, A.;

McCleland, C. W.; Schoonraad, J. L. J. Chem. Soc., Perkin Trans. 2

1986, 645–652. (c) Teeninga, H.; Nieuwpoort, W. C.; Engberts, J. B. F.

N. Z. Naturforsch. B 1981, 36, 279–281.

(20) (a) Waisser, K.; Bures, O.; Holy, P.; Kunes, J.; Oswald, R.;

Jiraskova, L.; Pour, M.; Klimesova, V.; Kubicova, L.; Kaustova, J. Arch.

Pharm. 2003, 336, 53–53. (b) Kubota, K.; Funatsu, M.; Kanzawa, K.;

Ishikawa, A.; Takeuchi, M. Jpn. Kokai Tokkyo Koho JP 11209328 A

19990803, 1999.

(21) (a) Batch, J. J.; Parry, K. P.; Rowe, C. F.; Lawrence, D. K.;

Brown, M. J. Ger. Oﬀen. DE 2819878 19781116, 1978. (b) Chakraborty,

K.; Devakumar, C. Plant Breed. 2006, 125, 441–447. (c) Chakraborty,

K.; Devakumar, C.; Tomar, S. M. S.; Kumar, R. J. Agric. Food Chem. 2003,

51, 992–998. (d) Richardson, A. G.; Pierce, J. S.; Reid, E. E. J. Am. Chem.

Soc. 1952, 74, 4011–4012.

(22) Presumably as a result of absence of even weak intermolecular

interactions, 7a was always obtained as a colorless glassy matrix.

(23) (a) Bowes, K. F.; Glidewell, C.; Low, J. N.; Skakle, J. M. S.;

Wardell, J. L. Acta Crystallogr. 2003, C59, o1–o3. (b) Gowda, B. T.;

Tokarꢀcík, M.; Koꢀzíꢀsek, J.; Sowmya, B. P.; Fuess, H. Acta Crystallogr.

2008, E64, o769–o769.

(24) (a) Metrangolo, P.; Neukirch, H.; Pilati, T.; Resnati, G. Acc.

Chem. Res. 2005, 38, 386–395. (b) Metrangolo, P.; Carcenac, Y.;

Lahtinen, M.; Pilati, T.; Rissanen, K.; Vij, A.; Resnati, G. Science 2009,

323, 1461–1464.

(25) Coote, L. M.; Lin, C. Y.; Zipse, H. In Carbon-Centered Free

Radicals and Radical Cations; Forbes, M. D. E., Ed.; John Wiley and Sons:

New Jersey, 2010; pp 83ꢀ102.

(26) Zipse, H. Top. Curr. Chem. 2006, 263, 163–189.

(27) (a) Wood, G. P. F.; Moran, D; Jacob, R; Radom, L. J. Phys.

Chem. A 2005, 109, 6318–6325. (b) Wood, G. P. F.; Henry, D. J.;

Radom, L. J. Phys. Chem. A 2003, 107, 7985–7990.

(28) De Luca, L.; Giacomelli, G.; Nieddu, G. Synlett 2005, 2,

223–226.

(13) (a) Walter, R. I. J. Am. Chem. Soc. 1966, 88, 1923–1930.

(b) Nelsen, S. F.; Landis, R. T.; Kiehle, L. H.; Leung, T. H. J. Am. Chem.

Soc. 1972, 94, 1610–1614. (c) Danen, W. C.; West, C. T.; Kensler, T. T.

J. Am. Chem. Soc. 1973, 95, 5716–5724.

(29) Hiegel, G. A.; Hogenauer, T. J.; Lewis, J. C. Synth. Commun.

2005, 35, 2099–2105.

(30) APEX2, ver. 2.1ꢀ4, and SAINT, ver. 7.34A; Bruker AXS Inc.:

Madison, Wisconsin, 2007.

(14) Danen, W. C.; Neugebauer, F. A. Angew. Chem., Int. Ed. 1975,

14, 783–789.

(31) Sheldrick, G. M. SADABS, version 2.05; University of

G€ottingen: G€ottingen, Germany.

(15) (a) Roth, H. D. Angew. Chem., Int. Ed. Engl. 1989,

28, 1193–1207. (b) Albini, A.; Fagnoni, M. Green Chem. 2004, 6, 1–6.

(c) Albini, A.; Fagnoni, M. ChemSusChem 2008, 1, 63–66. (d) Protti, S.;

Fagnoni, M. Photochem. Photobiol. Sci. 2009, 8, 1499–1516.

(16) (a) Lewis, N. S. Science 2007, 28, 798–801. (b) Yoon, T. P.;

Ischay, M. A.; Du, J. Nature Chem. 2010, 2, 527–532. (c) Lewis, N. S.;

Nocera, D. G. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 15729–

15735. (d) Balzani, V.; Credi, A.; Venturi, M. ChemSusChem 2008,

1, 26–58.

(32) Sheldrick, G. M. SHELXS-97, Program for X-ray Crystal

Structure Determination; University of G€ottingen: Germany, 1997.

Sheldrick, G. M. SHELXL-97, Program for X-ray Crystal Structure

Reﬁnement; University of G€ottingen: Germany, 1997.

(33) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (b) Lee,

C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789. (c) Miehlich,

B.; Savin, A. K.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989,

157, 200–206. (d) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.;

Frisch, M. J. J. Phys. Chem. 1994, 98, 11623–11627.

(17) (a) Oelgem€oller, M.; Jung, C.; Mattay, J. Pure Appl. Chem.

2007, 79, 1939–1947. (b) Esser, P.; Pohlmann, B.; Scharf, H.-D. Angew.

Chem., Int. Ed. 1994, 33, 2009–2023.

(18) Randall Clark, C.; McMillian, L. J. Pharm. Sci. 1990, 71,

220–222.

(34) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. In Ab

Initio Molecular Theory; Wiley: New York, 1986.

(35) Mayer, P. M.; Parkinson, C. J.; Smith, D. M.; Radom, L. J. Chem.

Phys. 1998, 108, 604–615.

(36) (a) Ruedenberg, K.; Cheung, L. M.; Elbert, S. T. Int. J. Quantum

Chem. 1979, 16, 1069–1101. (b) Roos, B. O.; Siegbahn, P. E. M. Int. J.

Quantum Chem. 1980, 17, 485–500. (c) Roos, B. O.; Taylor, P. R.;

Siegbahn, P. E. M. Chem. Phys. 1980, 48, 157–173.

(37) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502–16513.

(38) Coote, M. L. J. Phys. Chem. A 2004, 108, 3865–3872.

(39) (a) Henry, D. J.; Parkinson, C. J.; Radom, L. J. Phys. Chem. A

2002, 118, 7927–7936. (b) Henry, D. J.; Sullivan, M. B.; Radom, L.

J. Chem. Phys. 2003, 118, 4849–4860.

(40) Frisch, M. J. et al. Gaussian 03, Revision E.01; Gaussian, Inc.:

Wallingford, CT, 2004.

(41) M. Frisch, M. J. et al. Gaussian 09, Revision A.02; Gaussian,

Inc.: Wallingford, CT, 2009.

(42) G. Karlstr€om, Lindh, R.; Malmqvist, P.-Å.; Roos, B. O.; Ryde, U.;

Veryazov, V.;Widmark, P.-O.;Cossi, M.; Schimmelpfennig, B.; Neogrady,

P.; Seijo, L.; , Comput. Mater. Sci. 2003, 28, 222ꢀ239, MOLCAS 7.4.

(43) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000,

14, 123ꢀ134, MOLDEN 3.7.

(19) (a) Eble, A.; Reckmann, U.; Baur, P.; Reizlein, K. PCT Int.

Appl. WO 2005044221 A2 20050519, 2005. (b) Zakarya, D.; Larfaoui,

E. M.; Boulaamail, A.; Lakhliﬁ, T. Roum. Chem. Quart. Rev. 2000,

7, 127–137. (c) Laudien, R.; Mitzner, R. J. Chem. Soc., Perkin Trans. 2

2001, 11, 2226–2229. (d) Laudien, R.; Mitzner, R. J. Chem. Soc., Perkin

Trans. 2 2001, 11, 2230–2232. (e) Mizuno, T.; Kino, T.; Ito, T.; Miyata,

T. Synth. Commun. 2000, 30, 1675–1688. (f) Henriet, F.; Marechal, P.-Y.

Commun. Agric. Appl. Biol. Sci. 2009, 74, 471–478. (g) Klementova, S.;

Zemanova, M. Res. J. Chem. Environ. 2008, 12, 5–11. (h) Kieloch, R.;

Rola, H.; Sumislawska, J. Prog. Plant Protect. 2007, 47, 138–146.

(i) Idziak, R.; Woznica, Z.; Pelczynski, W. Prog. Plant Protect. 2007,

47, 121–124. (j) Gao, M.; Song, Z.; Song, W.; Dai, S. Res. J. BioTechnol.

2008, 3, 5–10. (k) Amorisco, A.; Losito, I.; Carbonara, T.; Palmisano, F.;

Zambonin, P. G. Rapid Commun. Mass Spectrom. 2006, 20, 1569–1576.

(l) Tixier, C.; Meunier, L.; Bonnemoy, F.; Boule, P. Int. J. Photoenergy

2000, 2, 1–8. (m) Nelieu, S.; Perreau, F.; Bonnemoy, F.; Ollitrault, M.;

Azam, D.; Lagadic, L.; Bohatier, J.; Einhorn, J. Environ. Sci. Technol.

2009, 43, 3148–54.

7848

dx.doi.org/10.1021/jp203771c |J. Phys. Chem. A 2011, 115, 7834–7848

Article Doi

Photoinduced rearrangement of aromatic N-chloroamides to chloroaromatic amides in the solid state: Inverted Πn-ΣN occupational stability of amidyl radicals

DOI: 10.1021/jp203771c

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jp203771c

Full text of DOI:10.1021/jp203771c

Products guided by the article

R&D Labs maybe for 99848-91-6

Relevant to this article

Hot Product