1H NMR aided elucidation of products derived from photodegradation of ethyl 3-azido-4,6-difluorobenzoate in 2,2,2-trifluoroethanol

Article abstract of DOI:10.1016/j.tet.2009.03.020

The major products formed upon photolysis of ethyl 3-azido-4,6-difluorobenzoate in 2,2,2-trifluoroethanol-d₃ has been elucidated by ¹H NMR analysis of the product mixture. Among the products formed and structurally elucidated was a h

Full text of DOI:10.1016/j.tet.2009.03.020

Tetrahedron 65 (2009) 3863–3870
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
1H NMR aided elucidation of products derived from photodegradation
of ethyl 3-azido-4,6-difluorobenzoate in 2,2,2-trifluoroethanol
,
Magne O. Sydnes a, Masaki Kuse b, Issei Doi a, Minoru Isobe c
*
a Laboratory of Organic Chemistry, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa, Nagoya 464-8601, Japan
b Chemical Instrument Division, RCMS, Nagoya University, Furo-cho, Chikusa, Nagoya 464-8602, Japan
c Institute for Advanced Research, Nagoya University, Furo-cho, Chikusa, Nagoya 464-8601, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The major products formed upon photolysis of ethyl 3-azido-4,6-difluorobenzoate in 2,2,2-tri-
fluoroethanol-d3 has been elucidated by 1H NMR analysis of the product mixture. Among the products
formed and structurally elucidated was a hitherto unreported product formed during photolysis of aryl
azides, namely azoxybenzene 19. The structural assignments of the major components of the reaction
mixture were aided by comparison with 1H NMR data from synthetic reference materials and compound
isolation. MS, MS/MS, and HPLC analysis as well as UV spectroscopy was also employed in order to
confirm and aid the structural analysis.
Received 12 February 2009
Received in revised form 3 March 2009
Accepted 5 March 2009
Available online 11 March 2009
Ó 2009 Elsevier Ltd. All rights reserved.
1N
1. Introduction
*N3
N3
N
N
The photodegradation of aryl azides has been extensively ex-
amined over the last 80 years.1 It is commonly understood that
upon irradiation with UV light phenyl azide 1 is converted to singlet
phenylnitrene 3 with the loss of N2 via excited azide 2 (Scheme 1).2
At temperatures above 165 K phenylnitrene 3 rapidly ring-expands
via benzazirine 4 to ketenimine 5.3 Intermediate 5 can be trapped
by the addition of diethylamine, thus forming the corresponding
2-diethylamino-1H-azepine 6, which quickly rearranges to
2-diethylamino-3H-azepine 7.4 At moderate concentrations and in
1
2
3
4
5
Et2NH
NEt2
ISC
3N
NH2
NH
10
8
6
the absence of
a suitable trapping agent ketenimine gives
NEt2
N
predominantly polymeric tar.5 At temperatures below 165 K in-
tersystem crossing (ISC) dominates generating triplet phenyl-
nitrene 8 from singlet phenylnitrene 3, which can dimerize in order
to form the corresponding azobenzene 9 or the corresponding
aniline 10 via photoreduction.6
N
N
9
7
Scheme 1. Mechanism for the photodegradation of phenyl azide and some of the
products formed when conducted in the presence of diethylamine.
F
R1
H
N
N3
OH
CF3
Recently we reported that hemiaminal 11 (general structure
shown) was formed when a range of aryl azides, non-fluorinated or
fluorinated [including ethyl 3-azido-4,6-difluorobenzoate (12)],
was irradiated with a high-pressure mercury lamp (350 nm) at
ambient temperature in 2,2,2-trifluoroethanol (TFE) at a concen-
tration of 2.0 mM.7 The structural assignment in the previous work
was predominantly conducted by using nano-LC-ESI-Q-TOF-MS
and -MS/MS. In the work presented herein the photolysis of aryl
F
R
1 = H, F or OMe
2 = H or F
R2
CO2Et
12
R
11
azide 12 was performed at
a much higher concentration
(13.75 mM) enabling the reaction to be monitored by 1H NMR
spectroscopy.
* Corresponding author. Tel.: þ81 52 789 4109; fax: þ81 52 789 4111.
E-mail address: isobem@agr.nagoya-u.ac.jp (M. Isobe).
0040-4020/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2009.03.020

3864
M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
2. Results and discussion
2.2. Photolysis
Photolysis of a 13.75 mM degassed (degassed by sonication)
solution of ethyl 3-azido-4,6-difluorobenzoate (12) in 2,2,2-tri-
fluoroethanol-d3 (TFE-d3) contained in a Pyrex NMR tube (no ad-
ditional filter was used) with a low-pressure mercury lamp
(254 nm) at ambient temperature resulted in the formation of at
least five products. In order to aid assignment of signals in the 1H
NMR spectra of the reaction mixture as well as in the HPLC chro-
matograms we prepared synthetic samples of products known
from previously to be formed during photolysis of aryl azide 12.7
The compounds we decided to prepare were azobenzene 13 and
hydrazine 14. In addition we had a sample of the corresponding
aniline 15, viz. ethyl 3-amino-4,6-difluorobenzoate, from our pre-
vious work.
With the reference data in hand work could shift toward the
photolysis experiments. A 13.75 mM degassed solution (degassed
by sonication) of ethyl 3-azido-4,6-difluorobenzoate (12) in TFE-d3
was irradiated with a low-pressure mercury lamp in a Pyrex NMR
tube (no additional filter was used during these experiments). The
reaction was monitored by conducting 1H NMR experiments after 2,
4, 6, 8, 10, 14, 18, 22, and 26 min. Simultaneously samples for ana-
lytical HPLC analysis were collected and frozen on dry ice. These
samples were analyzed shortly thereafter. From the 1H NMR spectra
and HPLC chromatograms we could see that the concentration of
the products slowly increased until ca. 18 min and then leveled off.
The 1H NMR spectrum and HPLC chromatogram after 26 min
irradiation showed that there was significant amount of unreacted
starting material left in the reaction mixture. Further irradiation of
the solution did not increase the conversion of the starting mate-
rial. Most likely due to the dark color of the reaction mixture
functioning as a filter, thus blocking out most of the light. Based on
the data for the reference material of azobenzenes 13a and 13b we
could easily determine that both E- and Z-azobenzene were formed
during the photolysis (see Figs. 1 and 2). We could also conclude
that hydrazine 14 was not formed under these conditions. This
conclusion was based on the fact that we could not see a triplet in
the 1H NMR spectrum at 7.49 ppm and no signal in the HPLC
chromatogram at tR 9.62 min [HPLC conditions: Develosil ODS-UG-
5 (4.6ꢀ250 mm i.d.), CH3CN/H2O 4:1, 0.5 mL/min]. The fact that
compound 14 was not formed in this work, but was formed in our
previous work7 could be due to a concentration effect and/or an
effect due to the change of light source (350 nm/254 nm). In our
previous work the photolysis was conducted predominantly with
a high-pressure mercury lamp and with a sample concentration of
2.0 mM.
2.1. Synthesis of reference material
We found azobenzene 13 to be a rather difficult compound to
prepare due to the fact that several of the potential starting ma-
terials are activated toward SNAr reactions. However, after some
experimentation we were able to obtain the desired substrate,
although in poor yield, by using a method reported by Leyva and
co-workers8 To this end, aniline 16 was treated with potassium
ferricyanide [K3Fe(CN)6] and potassium hydroxide in ethanol/water
(1:1), thus forming azobenzene carboxylic acid, which was directly
converted to the corresponding ethyl ester prior to column chro-
matography. Concentration of the relevant fractions gave a 3:1
mixture of the E and Z isomer of azobenzene, viz. compounds 13a
and 13b, respectively, in a combined yield of 7% (yield not opti-
mized) (Scheme 2). The major isomer was assigned to be the
E-isomer based on literature precedent stating that generally the
E-isomer is the most stable of the two possible isomers.9 Part of
the aforementioned mixture was then converted to the corre-
sponding hydrazine 14 upon exposure to sodium dithionite in
water/methanol/dichloromethane 6.4:6.4:1 at reflux by using
a method reported recently by Sydnes and co-workers.10 By such
means compound 14 could be obtained in 50% yield at 50% con-
version (yield not optimized). Although the yield of 13a and 13b
was low it still provided us with sufficient amount of material for
our purpose.
The product in the reaction mixture with tR 9.51 min was
assigned to be hemiaminal 17 based on its UV spectrum obtained
by HPLC analysis. The UV spectrum of compound 17 (Fig. 3A) is very
similar to the UV spectrum we obtained for hemiaminal 18 (Fig. 3B)
in our previous work.7 In our previous work the structure of
compound 18 was securely proved by ESI-Q-TOF-MS and -MS/MS
including proton/deuterium (H/D) exchange experiments followed
by MS and MS/MS analysis.7 We therefore assign the structure as
CO2Et
1) K3Fe(CN)6, KOH,
EtOH/H2O 1:1,
reflux, 6 h
2) H2SO4 (cat.),
EtOH, reflux, 3 h
F
F
F
F
F
NH2
N
N
N
+
N
F
F
F
F
F
CO2H
CO2Et
CO2Et
CO2Et
16
13a
13b
obtained as a 3:1 mixture of 13a and 13b
combined yield of 7%
Na2S2O4,
MeOH/H2O/CH2Cl2
6.7:6.7:1, reflux, 1.3 h
CO2Et
F
F
F
NH2
H
N
N
F
H
F
F
CO2Et
15
CO2Et
14 50% (at 50% coversion)
Scheme 2. Synthesis of azobenzene 13 and hydrazine 14 and the structure of aniline 15.

M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
3865
Figure 1. 1H NMR spectrum (9.0–5.8 ppm) obtained after 26 min irradiation of a 13.75 mM solution of ethyl 3-azido-4,6-difluorobenzoate (12) in TFE-d3. The signals derived from
the various compounds are assigned at the top of the spectrum.
depicted for compound 17 based on analogy with hemiaminal 18. In
order to further prove the structural assignment of the compound
we calculated the UV spectrum for compound 17 using SPARTAN.
As we can see from the calculated UV spectrum depicted in
Figure 3C it matches well with the UV spectrum obtained for
hemiaminal 17 (Fig. 3A). It was not possible to discern the signals
derived from compound 17 in the 1H NMR spectrum most likely
due to overlapping signals. Due to the unstable nature of hemi-
aminals in general we did not attempt to isolate this compound.
The compound with tR 13.1 min in the HPLC chromatogram was
isolated by utilizing a reversed phase column (see Experimental
part for details) in order to help structural assignment. 1H NMR
analysis of the product in TFE-d3 showed four sets of down field
triplets (8.87, 8.51, 7.04, and 6.97 ppm integrating for one proton
each) in addition to two quartets (4.33 and 4.31 ppm integrating for
two protons each) and two triplets (1.29(8) and 1.29(7) ppm in-
tegrating for three protons each). 19F NMR analysis in CDCl3
revealed that the compound contained four fluorine atoms and low
resolution MS (EI) showed a molecular ion at m/z 414. Subjection of
this molecular ion to high resolution MS gave the molecular for-
mula C18H14F4N2O5. The molecular mass was also confirmed by ESI-
Q-TOF-MS, and a proton/deuterium exchange experiment followed
by MS analysis, using our well established method.7,11 The proton/
deuterium exchange experiment revealed a molecular mass of m/z
416 (MþD)þ confirming that the product had no exchangeable
proton. Unfortunately due to the small amount of available com-
pound it was not possible to obtain a 13C NMR spectrum.
Nevertheless, based on the information at hand we concluded that
the structure of the compound is as depicted for azoxybenzene 19.
Comparison of the UV spectrum of compound 19 (lmaxz310 nm)
(Fig. 4) with similar compounds from the literature shows a good
match.12
It is plausible that substrate 19 was formed from azobenzene
13a and/or 13b by oxidation,13 however, a different mode of for-
mation is also plausible. It is possible that compound 19 was
formed upon reaction between N-oxide 20 and triplet phenyl-
nitrene 21. In an attempt to verify any of the intermediates en route
to compound 19 we subjected the photolysis mixture to nano-LC-
ESI-Q-TOF-MS. And indeed we were able to find a compound with
mass m/z 216 (MþH)þ, which could stem from the N-oxide 20. The
compound giving rise to the triplet at 8.67 ppm in the 1H NMR
spectrum (see Fig. 1) (the remaining signals could not be discerned
due to overlapping) could possibly be derived from N-oxide 20.
O
F
F
3N
N
F
F
CO2Et
20
CO2Et
21
Regardless of if compound 19 was formed by oxidation of azo-
benzene or via a reaction between compounds 20 and 21 we still
could not unambiguously point out the oxygen source for this

3866
M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
Figure 3. (A) UV spectrum of hemiaminal 17; (B) UV spectrum of hemiaminal 18 (data
taken from Ref. 7); (C) calculated UV spectrum for compound 17 using SPARTAN.
trace of water in the solvent; (2) the oxygen originated from the
solvent; (3) molecular oxygen that still remained dissolved in the
solvent due to poor degassing prior to photolysis. First the pho-
tolysis was conducted in a mixture of ethanol/H128O water 95:5
followed by isolation of azoxybenzene 19 by preparative HPLC.
Submission of the isolated product to MS (EI) analysis failed to
show any enhancement of 18O in the compound, thus ruling out
water as the oxygen source. Next we tested if the source of oxygen
was the solvent. Indeed, when the reaction was conducted in other
solvents, 1,1,1,3,3,3-hexafluoropropan-2-ol, 2,2,2-trichloroethanol,
and ethanol, we did see the formation of compound 19. However,
when the reaction was performed in non-degassed acetonitrile still
some azoxybenzene was formed. These results together with the
former points toward oxygen in the air as the source of oxygen in
this reaction. Finally when acetonitrile was degassed very carefully
the formation of compound 19 dropped to almost zero proving that
the fifth oxygen in azoxybenzene 19 stems from oxygen in the air.
A final experiment was conducted where oxygen was bubbled
Figure 2. HPLC chromatogram [HPLC conditions: Develosil ODS-UG-5 (4.6ꢀ250 mm
i.d.), CH3CN/H2O 4:1, 0.5 mL/min] of the reaction mixture obtained after 26 min irra-
diation of a 13.75 mM solution of ethyl 3-azido-4,6-difluorobenzoate (12) in TFE-d3:
(A) analyzed at 295 nm; (B) analyzed at 254 nm. The signals derived from the various
compounds are assigned on the top of the chromatograms. The tR for the respective
compounds are as follows: compound 22 7.73 min, hemiaminal 17 9.51 min, SM (12)
10.30 min, azo 11.32 min, dimer (azoxybenzene) 19 13.08 min, and azo 17.35 min.
reaction. A few experiments were therefore conducted in order to
try to shade some light on this question. Initially we envisaged
three potential sources for the oxygen: (1) the oxygen came from

M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
3867
Figure 5. 1H NMR spectra of photolysis mixture (400 MHz, TFE-d3, room temperature)
(9.0–5.5 ppm shown). (A) Aryl azide 12 prior to photoirradiation; (B) after 2 min ir-
radiation (the sample was degassed before photoirradiation) (12/22 1:0.04); (C) after
2 min irradiation with oxygen bubbling through the sample (12/22 5:1).
coupling constant (J) 3.2 Hz and 13.6 Hz, respectively. Furthermore,
the spectrum also shows, as expected, two signals resulting from
the ethyl ester [4.20 ppm (q, J¼7.2 Hz, 2H) and 1.19 ppm (t,
J¼7.2 Hz, 3H)]. The rather small coupling (J) constant (3.2 Hz) for
the doublet at 6.88 ppm and the fact that both the down field
protons appeared as doublets was intriguing. Further analysis by
13C NMR only enabled us to discern the signals for six out of the
nine carbons embodied in compound 22 (see Experimental for
details). Syringe injection ESI-Q-TOF-MS revealed a molecular ion
with mass m/z 200 for compound 22. Subjection of this molecular
ion to MS/MS gave rise to the following fragments, m/z 173,171, and
170. The fragment at m/z 173 probably arises from loss of HCN,
while m/z 171 and 170 are derived from loss of ethyl radical and
ethane, respectively. We then submitted the sample to our standard
H/D exchange conditions,7,11 which confirmed that the product had
no exchangeable proton. Furthermore, the fragments at m/z 171
and m/z 170 both had a mass increase of 1 Da, thus further con-
firming that the product had no exchangeable proton. The UV
spectrum of the product is depicted in Figure 6 and shows a red
shift relative to the starting material (lmaxz290 nm).
At this stage we suspected that this compound might be the
corresponding ketenimine 23 (Fig. 7), however, its long lifetime
(slowly decomposing over two weeks when stored at room tem-
perature) in TFE-d3 suggested otherwise. Furthermore, the 1H NMR
spectrum shows two down field doublets, which do not, at least at
first glance, match the expected 1H NMR spectrum of the keteni-
mine. It was at least expected that the proton between the two
fluorine atoms should give rise to a doublet of doublets or a triplet
due to coupling with the two fluorines. However, some solvated
form of ketenimine, such as indicated in structure 24 (Fig. 7),
should not be totally ruled-out.
Figure 4. UV spectrum for compound 19.
through the sample during photolysis. Under these conditions we
did not observe an increase in the formation of azoxybenzene 19,
rather the contrary, as evident from HPLC analysis. In fact, bubbling
oxygen through the sample during the photolysis gave only trace
amount of compound 19. However, under these conditions we did
see a dramatic increase in the amount of the compound with tR
7.73 min in the HPLC chromatogram. The fact that azoxybenzene 19
was not formed in a significant amount when oxygen was bubbled
through the reaction mixture during photolysis serves as evidence
that the compound most likely is formed via a triplet process. The
fact that oxygen is a triplet quencher14 results in this process almost
totally stopping up when oxygen is bubbled through the sample
solution during photolysis. While conducting the experiments in
different solvents we found that the highest yield of compound 19
(15%) was obtained when ethanol was used as the solvent. Using
ethanol as the solvent also resulted in an increased formation of
aniline 15, which was only formed in trace amount at this con-
centration when TFE was used as solvent, as evident from HPLC
analysis.
Interestingly the yield of azoxybenzene 19 was reduced to half
when the reaction concentration was doubled, viz. 27.5 mM. De-
spite the fact that generally a concentration increase would result
in an increased rate of dimerization. One probable explanation for
this could be that the solution turns very dark colored during the
photolysis, thus forming a filter blocking out the light and by such
means hampering further reaction. From our previous work7 we
know that compound 19 is not formed under rather dilute condi-
tions, namely 2.0 mM, so a higher concentration is essential in
order for this compound to be formed.
There is previously only one report of low temperature (ꢁ86 ꢂC)
13C NMR analysis of 13C enriched ketenimine 5 incorporated into
hemicarcerand.15
The remarkable long lifetime for compound 22 at room tem-
perature in TFE could be explained by the unique physical proper-
ties of this solvent, such as low nucleophilicity.16 It has been
reported that another fluorinated alcohol, namely 1,1,1,3,3,3-hexa-
fluoropropan-2-ol, increases the half-lives of radical cations
dramatically.17 It is likely that TFE can have a similar effect on
intermediates formed during photolysis given that the physical
properties of the two alcohols are similar.16
The compound giving rise to the two doublets at 6.88 and
5.96 ppm and with a tR 7.73 in the HPLC chromatogram was formed
in an 18% yield after 26 min irradiation together with a range of
other products (Figs. 1 and 2). Fortunately this compound (referred
to as compound 22 from this point onwards) was the predominant
product after 2 min irradiation, however, the yield was very low
(12/22 1:0.04). As alluded to previously the yield of this compound
could be improved (12/22 5:1) when photolysis was conducted for
2 min with oxygen bubbling through the sample solution. By such
means it was possible to obtain good 1H NMR data for compound
22 (Fig. 5).
Ketenimines are known to be readily trapped by diethylamine,
thus forming the corresponding 2-diethylamino-3H-azepine.4
A sample mixture containing compound 22 generated by irradia-
tion of azide 12 in TFE with oxygen bubbling through the sample
The 1H NMR spectrum of compound 22 shows two down field
doublets at 6.88 and 5.96 ppm integrating for one proton each with

3868
M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
analysis is therefore required in order to draw a conclusion as to the
actual structure of compound 22.
F
NEt2
N
F
EtO2C
25
3. Conclusion
The work reported herein represents the first thorough analysis
by 1H NMR of products derived from photodegradation of aryl
azide. We have shown that 1H NMR is an excellent analytical
method to investigate the products formed upon photolysis of ethyl
3-azido-4,6-difluorobenzoate. The strategy employed in this work
has broad generality for the analysis of products derived from
photolysis of aryl azides. 1H NMR supported by MS, MS/MS, and UV
analysis enabled firm confirmation of four out of the five major
products formed. In order to unambiguously prove the structure of
compound 20 further work is required.
4. Experimental section
4.1. General experimental
Reactions were monitored by thin-layer chromatography (TLC)
carried out on 0.25 mm silica gel coated glass plates 60F254 using
UV light as visualizing agent and basic KMnO4 solution followed by
heating as developing agent. Silica gel 60 (particle size 0.063–
0.2 mm ASTM) was used for flash chromatography. Ethyl 3-azido-
4,6-difluorobenzoate (13) was prepared according to our previously
reported method.8 Solvents and reagents for LC/MS were pur-
chased and prepared as follows: acetonitrile (hypergrade for LC/MS,
Merck, Germany), CF3CO2D (D¼99.8%) (Merck, Germany), tri-
fluoroacetic acid (Nacalai tesque, Japan), deuterium oxide
(D2O¼99.9%, Cambridge Isotope Laboratory, USA) and water were
of MilliQ grade. Deuterated solvents were freshly opened prior to
use and was used for 1 day when capped under an atmosphere of
argon.
Figure 6. UV spectrum of compound 22.
(26 min irradiation) was treated with excess diethylamine. By HPLC
analysis we could determine that the intermediate 22 was fully
converted to other products after 1.5 h at room temperature. The
reaction mixture, which contained several new compounds as ev-
ident from HPLC analysis, was then subjected to ESI-Q-TOF-MS
analysis. These analysis revealed the formation of a compound with
m/z 273 (MþH)þ corresponding to the mass of the expected
diethylamine insertion product 25 derived from ketenimine 23.
Subjecting this molecular ion to MS/MS analysis displayed the
structurally important fragment at m/z 200, which most likely
corresponds to ketenimine 23. 1H NMR analysis of the reaction
mixture was not attempted due to the complexity of the reaction
mixture and the high concentration of diethylamine in the solution.
However, the MS/MS data obtained for the compound strongly
suggest that the substrate giving rise to the mass m/z 273 (MþH)þ is
azepine 25.
4.2. Instrumentation
NMR chemical shifts are reported as d values in parts per million
(ppm) using tetramethylsilane
(
d
¼0.00 ppm) or 2,2,2-tri-
fluoroethanol-d3 (
d
¼5.02 ppm) as internal standard for proton (1H)
NMR and residual chloroform (
d
¼77.0 ppm) or residual 2,2,2-tri-
fluoroethanol (
d
¼126.3 ppm) as internal standard for carbon (13C)
NMR. Fluorine (19F) NMR spectra were referenced externally to
1,1,1-trifluorotoluene at
d¼0.00 ppm. HPLC analysis of reaction
mixtures derived from photolysis experiments was analyzed by
HPLC; Develosil C30-UG-5 (4.6ꢀ250 mm i.d.), CH3CN/H2O 4:1,
1 mL/min and the effluent were monitored at both 295 and 254 nm.
MS analysis was conducted utilizing a Q-TOF mass spectrometer
(Micromass, Manchester, UK) equipped with a Z-spray type ESI
source. Data were acquired and processed using MassLynx version
3.4. All samples were desalted and separated by non-split type pre-
packed gradient (PPG) system and appropriately assembled nano-
HPLC system (JASCO, Tokyo, Japan) using a Develosil ODS-HG-5
column (Nomura, 150 mmꢀ0.3 mm i.d.) before on-line ESI-MS and
MS/MS analysis. For H-MS and H-MS/MS analysis the column was
Even though some of the data reported herein points toward
ketenimine 23 as the most likely candidate for compound 20, there
are also, on the other hand, facts that suggests otherwise. Further
CH2CF3
O
F
F
H
N
N
F
F
equilibrated with 260
acid at a flow rate of 10
gradient from 0 to 100% of acetonitrile containing 0.025% tri-
fluoroacetic acid for 40 min at a flow rate of 5 L/min. For the H/D
mL water containing 0.025% trifluoroacetic
EtO2C
23
EtO2C
mL/min and then developed using a linear
24
Figure 7. Possible structures for compound 22.
m

M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
3869
exchange experiments (D-MS and D-MS/MS) the column was
equilibrated with 260 L D2O containing 0.025% D-trifluoroacetic
acid at a flow rate of 10 L/min and then developed using a linear
gradient from 0 to 100% acetonitrile containing 0.025% D-tri-
fluoroacetic acid at a flow rate of 5 L/min. The column effluent was
monitored at 210 nm and then introduced into the electrospray
nebulizer without splitting. For H-MS and H-MS/MS the column
was equilibrated with isocratic 40% acetonitrile/water containing
(E-isomer), 14.2 (E-isomer), 14.1 (Z-isomer), five signals from the
aromatic rings associated with the Z-isomer and the signal from the
carbonyl carbon associated with both isomers could not be dis-
m
m
cerned; 19F NMR (CDCl3, 376 MHz)
d
ꢁ36.3 (E-isomer), ꢁ40.1
m
(Z-isomer), ꢁ49.2 (Z-isomer), ꢁ46.5 (E-isomer); MS (EIþ) m/z 398
ꢃ
(Mþ , 68%), 353 (22), 213 (86), 185 (46), 157 (98), 140 (100), 112 (97),
ꢃ
101 (77), 51 (37); HRMS (EIþ) found: Mþ , 398.0872. C18H14F4N2O4
ꢃ
requires Mþ , 398.0890; tR in HPLC 17.75 min (E-isomer) and
0.025% trifluoroacetic acid at a flow rate of 5
mL/min. For the proton
11.48 min (Z-isomer).
exchange experiments (D-MS and D-MS/MS) the column was
equilibrated with isocratic 40% acetonitrile/D2O containing 0.025%
4.6. Hydrazine 14
D-trifluoroacetic acid at a flow rate of 5 mL/min. Selected MS and
MS/MS spectra were also measured utilizing syringe injection ion
trap HCT Plus mass spectrometer (Bruker Daltonics, Bremen, Ger-
many) equipped with an orthogonal ESI source. Data were acquired
and processed using Compass version 1.2 (esquireControlÔ and
DataAnalysisÔ version 3.2) (Bruker Daltonics, Bremen, Germany),
respectively. All MS experiments were preformed in the positive
ion mode.
Sodium dithionite (13.6 mg, 0.0782 mmol) was added to a solu-
tion of azobenzenes 13a and 13b (3.0 mg, 0.00753 mmol) stirred at
reflux in a mixture of MeOH/H2O/CH2Cl2 (1.0 mL/1.0 mL/0.15 mL).
The resulting reaction mixture was heated at reflux for 1.3 h before
being cooled to room temperature and poured into ice-water
(10 mL). The aqueous phase was extracted with Et2O (2ꢀ20 mL) and
the combined organic fractions were dried over Na2SO4. Filtration
and concentration gave a dark yellow oil, which was subjected to
preparative TLC (hexane/EtOAc 4:1) to afford fractions A and B.
Concentration of fraction A (Rf 0.46) gave 1.1 mg (37%, recovery)
of compound 13a as a yellow solid, which was identical, in all re-
spects, with the starting material.
4.3. Ethyl 3-azido-4,6-difluorobenzoate (12)
1H NMR (CF3CD2OD, 400 MHz)
d
7.57 (dd, J¼7.0 and 9.0 Hz, 1H),
6.84 (t, J¼10.6 Hz, 1H), 4.28 (q, J¼7.2 Hz, 2H), 1.27 (t, J¼7.2 Hz, 3H).
For the remaining spectroscopic data see Ref. 8.
Concentration of fraction B (Rf 0.38) gave 1.9 mg of a 3.8:1
mixture of hydrazine 14 and starting material 13a as a yellow solid.
This implements that 1.5 mg (50%, at 50% conversion) of the desired
4.4. Ethyl 3-amino-4,6-difluorobenzoate (15)
hydrazine 14 was formed. 1H NMR (CDCl3, 400 MHz)
d 7.57 (dd,
1H NMR (CF3CD2OD, 400 MHz)
d
7.40 (dd, J¼7.2 and 9.6 Hz, 1H),
J¼6.8 and 9.6 Hz, 2H), 6.91 (dd, J¼10.0 and 10.8 Hz, 2H), 4.34 (q,
6.77 (t, J¼10.6 Hz, 1H), 4.25 (q, J¼7.2 Hz, 2H), 1.26 (t, J¼7.2 Hz, 3H).
J¼7.2 Hz, 4H), 1.35 (t, J¼7.2 Hz, 6H); 1H NMR (CF3CD2OD, 400 MHz)
For the remaining spectroscopic data see Ref. 8.
d
7.49 (t, J¼8.2 Hz, 2H), 6.81 (t, J¼10.6 Hz, 2H), 4.20 (q, J¼7.2 Hz, 4H),
1.20 (t, J¼7.2 Hz, 6H); 19F NMR (CDCl3, 376 MHz)
d
ꢁ53.5, ꢁ61.3; MS
ꢃ
4.5. Azobenzenes 13a and 13b
(EIþ) m/z 400 (Mþ , 20%), 200 (28), 155 (100), 127 (82), 101 (32), 100
ꢃ
ꢃ
(22); HRMS (EIþ) found: Mþ , 400.1003. C18H16F4N2O4 requires Mþ
,
To a stirred solution of amine 16 (62.3 mg, 0.309 mmol) was
added a solution of KOH (47.4 mg, 0.845 mmol in 1.5 mL EtOH/
water 1:1) followed by slow addition of potassium ferricyanide
(408.3 mg, 1.24 mmol). The resulting reaction mixture was heated
at reflux for 7 h before being cooled to room temperature, filtered,
and diluted with CH2Cl2 (10 mL). The aqueous phase was acidified
with 1 N HCl and extracted with CH2Cl2 (3ꢀ10 mL). The combined
organic fractions were dried over Na2SO4. Filtration and concen-
tration under reduced pressure gave a dark orange oil, which was
diluted with EtOH (20 mL) and H2SO4 (10 drops) was added. The
resulting reaction mixture was heated at reflux for 3 h before being
cooled to room temperature and diluted with EtOAc (40 mL). The
organic phase was washed with brine/water (1ꢀ10 mL of a 1:1
mixture) and brine (1ꢀ10 mL) before being dried over Na2SO4.
Filtration and concentration gave a dark yellow oil, which was
subjected to flash chromatography (silica, hexane/EtOAc/Et3N
90:9.95:0.05/85:14.95:0.05). Concentration of the relevant fractions
(Rf 0.4 in hexane/EtOAc 85:15) gave 4.1 mg (7%) of the desired
product as a yellow solid and as a 3:1 mixture of E (13a) and Z (13b)
400.1046; tR in HPLC 9.62 min.
4.7. Isolation of azoxybenzene 19
The photolysis mixture was purified by HPLC (Cosmosil 5C18-AR
(10ꢀ250 nm), CH3CN/H2O 4:1, 2.0 mL/min, 295 nm). Concentration
of the relevant fraction under reduced pressure gave a light-yellow
oil. 1H NMR (CF3CD2OD, 400 MHz)
d
8.87 (t, J¼8.0 Hz, 1H), 8.51 (t,
J¼7.6 Hz, 1H), 7.04 (t, J¼10.2 Hz, 1H), 6.97 (t, J¼10.4 Hz, 1H), 4.33 (q,
J¼7.2 Hz, 2H), 4.31 (q, J¼7.2 Hz, 2H), 1.29(8) (t, J¼7.2 Hz, 3H), 1.29(7)
(t, J¼7.2 Hz, 3H); 1H NMR (CDCl3, 400 MHz)
9.04 (t, J¼8.0 Hz, 1H),
d
8.64 (t, J¼7.8 Hz, 1H), 7.11 (t, J¼9.8 Hz, 1H), 7.04 (t, J¼10.2 Hz, 1H),
4.44 (q, J¼7.2 Hz, 2H), 4.42 (q, J¼7.2 Hz, 2H), 1.42 (t, J¼7.2 Hz, 3H),
1.41 (t, J¼7.2 Hz, 3H); 19F NMR (CDCl3, 376 MHz)
d
ꢁ35.5, ꢁ37.7,
ꢃ
ꢁ41.1, ꢁ44.6; MS (EIþ) m/z 414 (Mþ , 100%), 394 (85), 369 (63), 213
(37), 199 (88), 157 (79), 140 (46), 115 (43), 101 (37); MS (ESI) m/z 415
ꢃ
([MþH]þ, 100%), 395 (4); HRMS (EIþ) found: Mþ , 414.0828.
ꢃ
C18H14O5N2F4 requires Mþ , 414.0839.
isomer. 1H NMR (CDCl3, 400 MHz)
d
8.41 (t, J¼8.2 Hz, 2H, E-isomer),
4.8. Compound 22
7.67 (t, J¼7.8 Hz, 2H, Z-isomer), 7.11 (t, J¼10.2 Hz, 2H, E-isomer), 6.81
(t, J¼9.6 Hz, 2H, Z-isomer), 4.43 (q, J¼7.2 Hz, 4H, E-isomer), 4.37 (q,
J¼7.2 Hz, 4H, Z-isomer), 1.42 (t, J¼7.2 Hz, 6H, E-isomer), 1.37 (t,
1H NMR (CF3CD2OD, 400 MHz)
d
6.88 (d, J¼3.2 Hz, 1H), 5.96 (d,
J¼13.6 Hz, 1H), 4.20 (q, J¼7.2 Hz, 2H), 1.19 (t, J¼7.2 Hz, 3H); 13C NMR
J¼7.2 Hz, 6H, Z-isomer); 1H NMR (CF3CD2OD, 400 MHz)
d
8.32
d
164.3 (d, JC–F¼2 Hz), 139.3 (d, JC–F¼20 Hz), 119.9 (d, JC–F¼4 Hz),
(t, J¼8.0 Hz, 2H, E-isomer), 7.55 (t, J¼7.6 Hz, 2H, Z-isomer), 7.02
(t, J¼10.4 Hz, 4H, E-isomer), 6.79 (t, J¼10.0 Hz, 2H, Z-isomer), 4.31
(q, J¼7.2 Hz, 4H, E-isomer), 4.23 (q, J¼7.2 Hz, 4H, Z-isomer), 1.29
(t, J¼7.2 Hz, 6H, E-isomer), 1.23 (t, J¼7.2 Hz, 6H, Z-isomer); 13C NMR
104.2 (dd, JC–F¼33 and 79 Hz), 65.5, 14.1, three signals were over-
lapping or obscured; MS (ESI): m/z 200 ((MþH)þ, 100%), 173 (8), 171
(42), 170 (34).
(CDCl3, 101 MHz)
d
164.1 (dd, JC–F¼12 and 270 Hz, E-isomer), 162.7
4.9. Compound 25
(dd, JC–F¼12 and 269 Hz, E-isomer), 163.0 (dd, JC–F¼4 Hz, E-isomer),
162.3 (dd, JC–F¼5 Hz, E-isomer), 125.9 (Z-isomer), 122.3 (E-isomer),
106.7 (dd, JC–F¼24 and 27 Hz, E-isomer), 61.8(9) (Z-isomer), 61.8(6)
MS (ESI): m/z 273 ((MþH)þ, 100%), 247 (10), 200 (5), 86 (90),
74 (25).

3870
M.O. Sydnes et al. / Tetrahedron 65 (2009) 3863–3870
4.10. General procedure for photolysis experiments (no
oxygen bubbling through the solution during irradiation)
M.O.S. and M.K. are grateful for the provision of a JSPS Postdoctoral
Fellowship for foreign researchers and a Grant-in-Aid for Encour-
agement of Young Scientists (B) (19780087), respectively. I.D.
would like to thank the Global COE program and Ono Pharma-
ceuticals for financial support.
A degassed solution of aryl azide 12 (13.75 mM) in 2,2,2-tri-
fluoroethanol-d3 (for 1H NMR experiments) or 2,2,2-tri-
fluoroethanol (for other experiments) was irradiated at ambient
temperature for the required time in a Pyrex NMR tube by using
a low-pressure mercury lamp. The samples were then subjected to
1H NMR and/or MS and MS/MS analysis.
References and notes
1. For an overview of the subject, see for example: (a) Abramovitch, R. A.; Davis,
B. A. Chem. Rev. 1964, 64, 149–185; (b) Iddon, B.; Meth-Cohn, O.; Scriven, E. F. V.;
Suschitzky, H.; Gallagher, P. T. Angew. Chem., Int. Ed. Engl. 1979, 18, 900–917; (c)
Wentrup, C. Tetrahedron 1974, 30, 1301–1311; (d) Platz, M. A. Acc. Chem. Res.
1995, 28, 487–492; (e) Borden, W. T.; Gritsan, N. P.; Hadad, C. M.; Karney, W. L.;
Kemnitz, C. R.; Platz, M. S. Acc. Chem. Res. 2000, 33, 765–771; (f) Gritsan, N. P.;
Platz, M. S. Chem. Rev. 2006, 106, 3844–3867; (g) Platz, M. S. In Reactive
Intermediates Chemistry; Moss, R. A., Platz, M. S., Jones, M., Jr., Eds.; Wiley:
New York, NY, 2004; pp 510–559.
2. (a) Lwowski, W. In Nitrenes; Wiley: New York, NY, 1970; (b) Scriven, E. F. V.
Azides and Nitrenes; Academic: New York, NY, 1984.
3. (a) Schrock, A. K.; Schuster, G. B. J. Am. Chem. Soc. 1984, 106, 5228–5234; (b)
Gritsan, N. P.; Yuzawa, T.; Platz, M. S. J. Am. Chem. Soc. 1997, 119, 5059–5060; (c)
Gritsan, N. P.; Zhu, S.; Hadad, C. M.; Platz, M. S. J. Am. Chem. Soc. 1999, 121,
1202–1207.
4. (a) Doering, W. v. E.; Odum, R. A. Tetrahedron 1966, 22, 81–93; (b) Sundberg,
R. J.; Brenner, M.; Suter, S. R.; Das, B. P. Tetrahedron Lett. 1970, 11, 2715–2718; (c)
DeGraff, B. A.; Gillespie, D. W.; Sundberg, R. J. J. Am. Chem. Soc. 1974, 96,
7491–7496.
5. Meijer, E. W.; Nijhuis, S.; Von Vroonhoven, F. C. B. M. J. Am. Chem. Soc. 1988, 110,
7209–7210.
6. Leyva, E.; Platz, M. S.; Persy, G.; Wirz, J. J. Am. Chem. Soc. 1986, 108, 3783–3790.
7. Sydnes, M. O.; Doi, I.; Ohishi, A.; Kuse, M.; Isobe, M. Chem. Asian J 2008, 3,
102–112.
8. Leyva, E.; Medina, C.; Moctezuma, E.; Leyva, S. Can. J. Chem. 2004, 82,
1712–1715.
4.11. General procedure for photolysis experiments with
oxygen bubbling through the solution during irradiation
A solution of aryl azide 12 (13.75 mM) in 2,2,2-trifluoroethanol-
d3 (for 1H NMR experiments) or 2,2,2-trifluoroethanol (for other
experiments) was irradiated at ambient temperature for the
required time in a Pyrex NMR tube while bubbling oxygen through
the sample by using a low-pressure mercury lamp. The samples
were then subjected to 1H NMR and/or MS and MS/MS analysis.
4.12. Procedure for treating compound 22 with diethylamine
A solution of aryl azide 12 (13.75 mM) in 2,2,2-trifluoroethanol
was irradiated for 26 min at ambient temperature in a Pyrex NMR
tube while bubbling oxygen through the sample by using a low-
pressure mercury lamp. Diethylamine (1.0 mL, 9.67 mmol) was
then added to the reaction mixture and the resulting reaction
mixture was thoroughly shaken. The NMR tube was then left in the
dark at room temperature for 1.5 h when all of the intermediate
had reacted as judged by HPLC analysis. A sample of the reaction
mixture was then subjected to MS and MS/MS analysis.
9. Kucharski, S.; Janik, R.; Motschmann, H.; Radu¨ge, C. New J. Chem. 1999, 23,
765–771.
10. Sydnes, L. K.; Elmi, S.; Heggen, P.; Holmelid, B.; Malthe-Sørensen, D. Synlett
2007, 1695–1698.
11. (a) Kuse, M.; Kanakubo, A.; Suwan, S.; Koga, K.; Isobe, M.; Shimomura, O. Bioorg.
Med. Chem. Lett. 2001, 11, 1037–1040; (b) Kanakubo, A.; Isobe, M. J. Mass Spec-
trom. 2004, 39, 1260–1267.
4.13. General procedure for the preparation of samples for MS
and MS/MS analysis
12. Furin, G. G.; Fedotov, M. A.; Yakobson, G. G.; Zibarev, A. V. J. Fluorine Chem. 1985,
28, 273–290.
All samples for H-MS and H–MS/MS were further diluted with
13. When azoxybenzene is prepared synthetically it is generated by oxidation of
the corresponding azobenzene, see for example: Takahashi, H.; Ishioka, T.;
Koiso, Y.; Sodeoka, M.; Hashimoto, Y. Biol. Pharm. Bull. 2000, 23, 1387–1390.
14. (a) Sydnes, L. K.; Nilssen, A. V.; Jørgensen, E.; Pettersen, A. Acta Chem. Scand.
1992, 46, 661–668; (b) Barltrop, J. A.; Doyle, J. D. Excited States in Organic
Chemistry; John Wiley & Sons: New York, NY, 1975.
methanol/acetonitrile (1:1) to 500 pmol
D-MS and D–MS/MS were further diluted with methanol-d4/
m
Lꢁ1 and all samples for
acetonitrile to 500 pmol m .
Lꢁ1
15. Warmuth, R.; Makowiec, S. J. Am. Chem. Soc. 2005, 127, 1084–1085.
16. (a) Be´gue´, J.-P.; Bonnet-Delpon, D.; Crousse, B. Synlett 2004, 18–29; (b) Shuklov,
I. A.; Dubrovina, N. A.; Bo¨rner, A. Synthesis 2007, 2925–2943.
17. Eberson, L.; Hartshorn, M. P.; Persson, O. J. Chem. Soc., Perkin Trans. 2 1995,
1735–1744.
Acknowledgements
Financial support from a Grant-in-Aid for Specially Promoted
Research (16002007) from MEXT is gratefully acknowledged.