Reduction of 1-(2-Chloroethyl)-2-nitrobenzene and 1-(2-Bromoethyl)-2- nitrobenzene at carbon cathodes: Electrosynthetic routes to 1-nitro-2- vinylbenzene and 1H -indole

Article abstract of DOI:10.1149/1.3482019

Studies of the electrochemical reductions of 1-(2-chloroethyl)-2- nitrobenzene (1) and 1-(2-bromoethyl)-2-nitrobenzene (2) at carbon cathodes in dimethylformamide (DMF) containing tetramethylammonium tetrafluoroborate (TMABF₄) have been undertaken. Cyclic voltammograms for 1 and 2 exhibit three irreversible cathodic peaks, the first of which is attributed to one-electron reduction of the nitro group, and the resulting nitro radical-anion immediately reacts as a base with the adjacent alkyl halide moiety via an E2 elimination to produce 1-nitro-2-vinylbenzene. At a potential corresponding to the first cathodic peak, bulk electrolyses of 1 or 2 in the absence of a proton donor afford 1-nitro-2-vinylbenzene as principal product, along with 1H -indole and a dimeric species. However, when 1 or 2 is electrolyzed in the presence of either phenol or 2,4-pentanedione as a proton source, the only significant product is 1H -indole. A mechanistic picture for the reductions of 1 and 2 is proposed, and a separate examination of the electrochemical behavior of 1-nitro-2-vinylbenzene has been included as part of this work.

Full text of DOI:10.1149/1.3482019

Journal of The Electrochemical Society, 157 ͑11͒ F167-F172 ͑2010͒
F167
0
013-4651/2010/157͑11͒/F167/6/$28.00 © The Electrochemical Society
Reduction of 1-(2-Chloroethyl)-2-nitrobenzene and
-(2-Bromoethyl)-2-nitrobenzene at Carbon Cathodes:
Electrosynthetic Routes to 1-Nitro-2-vinylbenzene and 1H-Indole
1
,
z
Peng Du* and Dennis G. Peters**
Department of Chemistry, Indiana University, Bloomington, Indiana 47405, USA
Studies of the electrochemical reductions of 1-͑2-chloroethyl͒-2-nitrobenzene ͑1͒ and 1-͑2-bromoethyl͒-2-nitrobenzene ͑2͒ at
carbon cathodes in dimethylformamide ͑DMF͒ containing tetramethylammonium tetrafluoroborate ͑TMABF ͒ have been under-
4
taken. Cyclic voltammograms for 1 and 2 exhibit three irreversible cathodic peaks, the first of which is attributed to one-electron
reduction of the nitro group, and the resulting nitro radical–anion immediately reacts as a base with the adjacent alkyl halide
moiety via an E2 elimination to produce 1-nitro-2-vinylbenzene. At a potential corresponding to the first cathodic peak, bulk
electrolyses of 1 or 2 in the absence of a proton donor afford 1-nitro-2-vinylbenzene as principal product, along with 1H-indole
and a dimeric species. However, when 1 or 2 is electrolyzed in the presence of either phenol or 2,4-pentanedione as a proton
source, the only significant product is 1H-indole. A mechanistic picture for the reductions of 1 and 2 is proposed, and a separate
examination of the electrochemical behavior of 1-nitro-2-vinylbenzene has been included as part of this work.
©
2010 The Electrochemical Society. ͓DOI: 10.1149/1.3482019͔ All rights reserved.
Manuscript submitted June 28, 2010; revised manuscript received July 26, 2010. Published September 7, 2010. This was Paper
8
16 presented at the Vancouver, Canada, Meeting of the Society, April 25–30, 2010.
Investigations of the electrochemical reduction of aromatic nitro
compounds have a long history that dates back to the beginning of
the twentieth century. Among the numerous papers cited in
ployed chromatographic and spectroscopic techniques to separate,
identify, and quantitate the electrolysis products. In addition, be-
cause 1-nitro-2-vinylbenzene ͑3͒ is a major product of the electro-
chemical reductions of both 1 and 2 in the absence of an added
proton source, and because direct reduction of 3 in the presence of a
proton donor leads almost exclusively to 1H-indole, an investigation
of the behavior of 3 at carbon electrodes in the DMF–TMABF4
medium has been conducted. Conceivably, this research will provide
a strategy that can be used in the future for the electrosynthesis of
1
-4
reviews, most studies pertaining to the electroreduction of these
species ͑including nitrobenzene itself͒ have involved the use of vari-
ous protic media, whereas comparatively few studies have been car-
ried out in aprotic solvents. Some examples of the latter category of
5
work are the reduction of nitrobenzene in liquid ammonia, in di-
6
7
methylformamide ͑DMF͒ and acetonitrile, in tetrahydrofuran, and
in a room-temperature ionic liquid.
8
1
H-indole and its derivatives.
Derivatives of nitrobenzene are important chemicals which serve
as precursors for species that can be generated via reductive in-
tramolecular cyclization. Several publications have dealt with the
electrochemical reduction of o-substituted nitrobenzenes as an av-
enue for the synthesis of cyclic products. For example, Hazard and
9
,10
11
co-workers
and David and co-workers employed the electrore-
ductive cyclizations of 3-͑2-nitrophenyl͒-3-oxopropanoic acid to
form 2-͑benzo͓c͔isoxazol-3-yl͒acetic acid and of 5-isocyano-2,6-
dimethyl-4-͑2-nitrophenyl͒-1,4-dihydropyridine-3-carbonitrile to af-
ford
͔naphthyridine 6-oxide. Electrochemically promoted cyclization of
N-͑2-nitrophenyl͒acetamide to give 2-methyl-1H-benzo͓d͔-
5-amino-1-isocyano-2,4-dimethyl-3,10b-dihydrobenzo͓c͔͓2,
7
1
2
imidazole in quantitative yield was studied by Alberti et al. Each
of the aforementioned investigations involved the use of aqueous
media, but it is reasonable to expect similar results if such com-
pounds are electrolyzed in aprotic solvents ͑perhaps in the presence
of a suitable proton donor͒.
Experimental
Reagents.— Each of the following compounds ͑with a purity or
concentration indicated in parentheses͒ was purchased from com-
mercial sources and was used without further purification:
For a number of years, our laboratory has been interested in the
reduction of halogenated organic substances to afford a variety of
compounds, and some of our work has centered on electroreductive
2
-nitrophenylethyl alcohol ͑97%͒, trichloroacetonitrile ͑98%͒,
13-20
intramolecular cyclizations.
recently became interested in the electrochemical reduction of
-͑2-chloroethyl͒-2-nitrobenzene ͑1͒ and 1-͑2-bromoethyl͒-2-
In view of this earlier research, we
triphenylphosphine ͑98.5%͒, n-dodecane ͑99%͒, p-toluenesulfonyl
chloride ͑99%͒, hexane ͑HPLC grade͒, ethyl acetate ͑HPLC grade͒,
phenol ͑99%͒, 2,4-pentanedione ͑99%͒, 1H-indole ͑99%͒, palladium
on charcoal ͑10%͒, tert-butyllithium ͑1.7 M in pentane͒, nitroben-
zene ͑reagent grade, ACS͒, ͑2-chloroethyl͒benzene ͑99%͒, styrene
1
nitrobenzene ͑2͒, species that have not been previously investigated.
After synthesizing these compounds, we have used cyclic voltam-
metry and controlled-potential ͑bulk͒ electrolysis to probe and char-
acterize the direct reduction of these compounds at carbon cathodes
in dimethylformamide ͑DMF͒ containing tetramethylammonium tet-
͑
99%͒, and azobenzene ͑99%͒. Solutions for all electrochemical ex-
periments were deaerated with zero-grade argon. Dimethylforma-
mide ͑DMF, 99.9%͒ was employed without further purification as
the solvent for cyclic voltammetry and controlled-potential elec-
trolysis, and tetramethylammonium tetrafluoroborate ͑TMABF4͒
served as the supporting electrolyte.
rafluoroborate ͑TMABF ͒ as a supporting electrolyte in both the
4
presence and absence of added proton sources, and we have em-
Synthesis of 1-(2-chloroethyl)-2-nitrobenzene (1).— A modifica-
tion of a published procedure was used for the preparation of 1. A
mixture of trichloroacetonitrile ͑37.5 mmol͒ and triphenylphosphine
*
Electrochemical Society Student Member.
2
1
*
* Electrochemical Society Fellow.
z
E-mail: peters@indiana.edu
Downloaded on 2015-01-17 to IP 128.193.164.203 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

F168
Journal of The Electrochemical Society, 157 ͑11͒ F167-F172 ͑2010͒
͑
1
37.5 mmol͒ was dissolved in 50 mL of dry dichloromethane in a
00 mL round-bottom flask, and 2-nitrophenylethyl alcohol ͑25
mmol͒ was added dropwise into the stirred solution at room tem-
perature under a nitrogen atmosphere. After 30 min, the solution was
concentrated by means of rotary evaporation to obtain the crude
product. A silica-gel column was employed to separate and purify
the desired compound; the eluent consisted of a 10:1 mixture of
1
hexane and ethyl acetate. Spectroscopic data for 1 are as follows: H
NMR ͑300 MHz, CDCl ͒ ␦ 3.34 ͑m, 2H͒, 3.82 ͑m, 2H͒, 7.39–7.98
3
+
͑
−
͓
m, 4H͒; MS ͑70 eV͒ m/z 187, M ͑2͒; 185, M+ ͑6͒; 168, ͓M
+
+
+
17͔ ͑100͒; 150, ͓M − Cl͔ ͑trace͒; 103, ͓M − Cl − 47͔ ͑52͒; 77,
+
M − Cl − 73͔ ͑100͒.
Synthesis of 2-(o-nitrophenyl)ethyl tosylate.— This compound
2
2
was prepared according to a procedure published by Tipson.
A
solution of 2-nitrophenylethyl alcohol ͑25 mmol͒ in 100 mL of pyr-
idine was treated with p-toluenesulfonyl chloride ͑30 mmol͒ in an
ice bath, and the solution was kept at 0°C for 2 h; then the mixture
was allowed to come to room temperature for a 30 min period.
Absolute methanol was used as solvent to recrystallize the desired
1
product, which was characterized spectroscopically: H NMR ͑300
MHz, CDCl ͒ ␦ 2.41 ͑s, 3H͒, 3.24 ͑t, 2H͒, 4.33 ͑t, 2H͒, 7.26–7.95
3
+
+
͑
͓
m, 8H͒; MS ͑70 eV͒ m/z 321, M ͑trace͒; 275, ͓M − 46͔ ͑8͒; 155,
M − 166͔ ͑71͒; 136, ͓M − 185͔ ͑73͒; 91, ͓M − 230͔ ͑100͒.
Figure 1. Cyclic voltammograms recorded with a glassy carbon electrode
+
+
+
2
−1
͑area = 0.077 cm ͒ at 100 mV s in oxygen-free DMF containing 0.050
M TMABF : ͑A and B͒ 2.0 mM nitrobenzene, ͑C͒ 2.0 mM 1, and ͑D͒ 2.0
4
Synthesis of 1-(2-bromoethyl)-2-nitrobenzene (2).— Compound 2
mM 2. For curve A, the scan goes from 0 to Ϫ0.8 to 0 V; for curves B–D, the
scan goes from 0 to Ϫ2.0 to 0 V.
2
3
was prepared according to a published procedure. Anhydrous
lithium bromide ͑20 mmol͒ was added to a stirred, refluxing solution
of 2-͑o-nitrophenyl͒ethyl tosylate ͑10 mmol͒ in 50 mL of acetone,
and the mixture was heated overnight. After removal of acetone, the
mixture was treated with water, and the crude product was extracted
into diethyl ether–pentane. Then the solvent was removed, and the
desired product was recrystallized from ethanol and characterized by
Separation, identification, and quantitation of electrolysis prod-
ucts.— At the end of each controlled-potential ͑bulk͒ electrolysis,
the solvent–supporting electrolyte solution containing the products
was partitioned between diethyl ether and water, after which the
ether phase was dried over anhydrous sodium sulfate, concentrated
by means of rotary evaporation, and analyzed with the aid of gas
chromatography. Peak areas for the various products were deter-
mined with respect to that of an internal standard ͑n-dodecane͒
added in known amount to the solution prior to the start of each
electrolysis. Details concerning the procedures employed for the
means of spectroscopy: 1H NMR ͑300 MHz, CDCl ͒ ␦ 3.45 ͑m,
2
3
+
H͒, 3.67 ͑m, 2H͒, 7.39–8.01 ͑m, 4H͒; MS ͑70 eV͒ m/z 231, M
+
+
͑
−
trace͒; 229, M ͑trace͒; 150, ͓M − Br͔ ͑100͒; 103, ͓M − Br
+
+
47͔ ͑38͒; 77, ͓M − Br − 73͔ ͑71͒.
Synthesis of 1-nitro-2-vinylbenzene (o-nitrostyrene) (3).— We
prepared this compound from 2-͑o-nitrophenyl͒ethyl tosylate ac-
cording to a published procedure. Spectroscopic data for 3 are as
follows: H NMR ͑300 MHz, CDCl ͒ ␦ 5.47 ͑d, 1H͒, 5.73 ͑d, 1H͒,
7
1
36
quantitation of products are described in an earlier publication;
24
product yields reported in this paper correspond to the absolute per-
centage of starting material incorporated into a particular species.
Gas chromatography–mass spectrometry ͑GC-MS͒ and gas
chromatography–high-resolution mass spectrometry ͑GC-HRMS͒
were used to confirm the identities of all products.
1
3
+
.17 ͑dd, 1H͒, 7.35–7.95 ͑m, 4H͒; MS ͑70 eV͒ m/z 149, M ͑trace͒;
+
+
+
32, ͓M − 17͔ ͑72͒; 120, ͓M − 29͔ ͑78͒; 91, ͓M − 58͔ ͑100͒; 77,
+
+
͓
M − 72͔ ͑100͒; 51, ͓M − 98͔ ͑60͒.
Synthesis of 1H,1ЈH-2,2Ј-biindole, 1H,1ЈH-3,3Ј-biindole, and
Results and Discussion
1
1
H,1ЈH-2,3Ј-biindole.— We
prepared
1H,1ЈH-2,2Ј-biindole,
Cyclic voltammetric behavior of nitrobenzene, 1-(2-chloroethyl)-
2-nitrobenzene (1), and 1-(2-bromoethyl)-2-nitrobenzene (2).— Cy-
clic voltammograms recorded at 100 mV s− for the direct reduc-
tion of separate 2.0 mM solutions of nitrobenzene, 1-͑2-
chloroethyl͒-2-nitrobenzene ͑1͒, and 1-͑2-bromoethyl͒-2-nitro-
benzene ͑2͒ at a freshly polished glassy carbon electrode in DMF
H,1ЈH-3,3Ј-biindole, and 1H,1ЈH-2,3Ј-biindole according to pro-
25-27
cedures in the literature.
Mass spectrometric data for these com-
1
pounds are as follows: ͑a͒ for 1H,1ЈH-2,2Ј-biindole, m/z ͑70 eV͒
32, M ͑100͒; 204, ͓M − 28͔ ͑18͒; 176, ͓M − 56͔+ ͑4͒; 116,
+
+
2
͓
+
+
M − 116͔ ͑11͒; 89, ͓M − 143͔ ͑7͒; ͑b͒ for 1H,1ЈH-3,3Ј-biindole,
+
+
+
m/z ͑70 eV͒ 232, M ͑100͒; 204, ͓M − 28͔ ͑26͒; 176, ͓M − 56͔
containing 0.050 M TMABF under an argon atmosphere are shown
4
+
͑
1
−
−
12͒; 115, ͓M − 117͔+ ͑18͒; 88, ͓M − 144͔ ͑8͒; ͑c͒ for
in Fig. 1.
+
H,1ЈH-2,3Ј-biindole, m/z ͑70 eV͒ 232, M ͑100͒; 204, ͓M
Curve A of Fig. 1 is a cyclic voltammogram with cathodic and
anodic peak potentials of Ϫ0.44 and Ϫ0.27 V, respectively, that is
attributable to the reversible one-electron reduction of nitrobenzene
to its radical–anion and to the subsequent reoxidation of the radical–
28͔ ͑22͒; 176, ͓M − 56͔+ ͑7͒; 116, ͓M − 116͔+ ͑18͒; 102, ͓M
+
+
130͔ ͑12͒. These results are in agreement with those reported in
earlier publications.
25,28,29
anion back to nitrobenzene; this process has been well documented
Electrodes, cells, and instrumentation.— Details and pertinent
literature concerning the electrodes, cells, and instrumentation for
cyclic voltammetry and controlled-potential ͑bulk͒ electrolysis can
5,37-39
in previous work.
When more negative potentials are accessed
during the reduction of nitrobenzene ͑curve B, Fig. 1͒, a second
cathodic peak is observed at Ϫ1.25 V, which has an accompanying
30-32
be found in previous papers.
All potentials in the present work
5,39
anodic peak at Ϫ0.94 V; in accord with earlier studies,
we be-
are given with respect to a reference electrode consisting of a
cadmium-saturated mercury amalgam in contact with DMF satu-
rated with both cadmium chloride and sodium chloride; this refer-
ence electrode has a potential of Ϫ0.76 V vs the aqueous saturated
lieve that this second cathodic peak arises from one-electron reduc-
tion of the nitrobenzene radical–anion to the nitrobenzene dianion,
which is protonated by residual water in the solvent–supporting
electrolyte, whereupon further reduction takes place probably to
form nitrosobenzene and then phenylhydroxylamine. Each of the
33-35
calomel electrode ͑SCE͒ at 25°C.
Downloaded on 2015-01-17 to IP 128.193.164.203 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑11͒ F167-F172 ͑2010͒
anodic peaks for curve B is noticeably smaller in height than its
F169
associated cathodic peak; under the conditions of our experiments,
we conclude that this diminution of the anodic peaks is due to com-
petition between protonation by adventitious water ͑which initiates
follow-up irreversible electrochemical and chemical processes͒ and
oxidation of the nitrobenzene radical–anion or dianion.
Curve C in Fig. 1, which depicts a cyclic voltammogram for the
reduction of 1 at a glassy carbon electrode in DMF containing 0.050
M TMABF , exhibits three cathodic peaks at Ϫ0.44, Ϫ1.04, and
4
Ϫ1.93 V. It seems evident that the first cathodic peak for 1 corre-
sponds to the first cathodic peak for nitrobenzene; however, the peak
current for the first step of reduction of 1 is noticeably larger than
that of the first step of reduction of nitrobenzene. Another important
feature of the first cathodic peak for 1 is that cycling the potential
between 0 and Ϫ0.60 V shows no anodic peak for scan rates up to
−
1
4
0 V s , quite unlike the reversibility seen for nitrobenzene. Thus,
we propose that the initially formed nitro radical–anion of 1 under-
goes a rapid follow-up intramolecular reaction in which the nitro
radical–anion ͑acting as a base͒ abstracts an ␣ proton from the
o-chloroethyl moiety, leading to formation of 1-nitro-2-
vinylbenzene ͑which is shown later in this report to be electroactive
in the same region of potentials as both nitrobenzene and 1͒. Notice
that the peak potential of the second cathodic peak for 1 is shifted
positively with respect to the peak potential of the second cathodic
peak for nitrobenzene, an observation that seems to be correlated
with a complex mixture of products obtained when 1 is electrolyzed
at potentials corresponding to its second cathodic peak. Moreover,
the second cathodic peak for 1 is smaller in magnitude than the first
cathodic peak for 1 ͓a feature also seen for the reduction of 2 ͑curve
D, Fig. 1͔͒, but this phenomenon cannot yet be rationalized. Finally,
we assign the third irreversible cathodic peak of 1 to reductive
cleavage of the carbon–chlorine bond.
As shown by curve D in Fig. 1, a cyclic voltammogram for the
reduction of 2 shows three cathodic peaks at Ϫ0.34, Ϫ0.99, and
Ϫ1.89 V that closely resemble those seen for 1 ͑curve C͒. Indeed,
the only major difference between curves C and D is that the first
cathodic peak potential for 2 is Ϫ0.34 V, which is 100 mV more
positive than the first cathodic peak potential for 1. We believe this
difference is due to the fact that the intramolecular reaction between
the electrogenerated nitro radical–anion and the o-bromoethyl group
of 2 is more facile because the carbon–bromine bond of 2 is lower in
energy than the carbon–chlorine bond of 1.
Figure 2. Consecutive cyclic voltammograms recorded with a glassy carbon
2
−1
electrode ͑area = 0.077 cm ͒ at 100 mV s in oxygen-free DMF contain-
ing 0.050 M TMABF4 and 2.0 mM 2; curve A is the first scan, followed
without interruption by curves B and C. For all curves, the scan goes from 0
to Ϫ0.6 to 0 V.
tuted nitrosobenzene in DMF–0.050 M TMABF , and we observed
4
four cathodic peaks ͑with the first three showing reversibility͒; how-
ever, the most pertinent finding was that the reversible process as-
sociated with electrogeneration and subsequent oxidation of the
radical–anion of nitrosobenzene exhibits cathodic and anodic peak
potentials of Ϫ0.23 and Ϫ0.066 V, respectively, at a scan rate of
−
1
1
00 mV s , so it is unlikely that either of the cathodic peaks seen
in curves B and C of Fig. 2 can be attributed to reduction of a
compound such as 1-nitroso-2-vinylbenzene.
Cyclic voltammetric behavior of 1-(2-chloroethyl)-2-nitro-
benzene (1) and 1-(2-bromoethyl)-2-nitrobenzene (2) in the presence
of a proton donor.— Cyclic voltammograms of 2.0 mM 1 or 2 in
the presence of 10.0 mM phenol or 2,4-pentanedione as proton
Displayed in Fig. 2 is a set of cyclic voltammograms for the first
three consecutive scans from 0 to Ϫ0.6 to 0 V for a 2.0 mM solution
sources were recorded with a glassy carbon electrode at
100 mV s− in DMF containing 0.050 M TMABF4. A comparison
of cyclic voltammograms for 2.0 mM 1 by itself ͑curve A, which is
a repetition of curve C from Fig. 1͒ and 2.0 mM 1 in the presence of
10.0 mM phenol ͑curve B͒ is displayed in Fig. 3. For curve B, only
one large, broad irreversible cathodic peak having a peak potential
of Ϫ0.46 V is observed; the second cathodic peak seen for curve A
does not appear for curve B. Similar cyclic voltammograms were
obtained for both 1 and 2 in the presence of either phenol or 2,4-
pentanedione. These results imply that the presence of a proton do-
nor favors further reduction of the nitro radical–anion of 1 or 2, and
that there is formation of products which ͑i͒ are not reducible at a
more negative potential and ͑ii͒ differ in their distribution from
those derived from the reduction of 1 or 2 in the absence of a proton
source.
1
of 2 in DMF–0.050 M TMABF , where the focus is exclusively on
4
the time-dependent changes for the first cathodic peak associated
with reduction of 2. Curve A of Fig. 2 is the response that has
already been described for the first peak in curve D of Fig. 1; once
again, the cathodic peak potential is confirmed to be Ϫ0.34 V, and
there is almost no anodic process on the reverse scan. However,
curves B and C of Fig. 2 ͑corresponding to the second and third
potential sweeps͒ are superimposable and reveal two overlapping
cathodic peaks at Ϫ0.33 and Ϫ0.40 V. We interpret curve A as
showing the reduction of 2 to its nitro radical–anion form, but no
significant anodic peak is seen because oxidation of the nitro
radical–anion is largely prevented by its rapid intramolecular inter-
action with the neighboring o-bromoethyl group. In a parallel ex-
periment, when we recorded three consecutive cyclic voltammo-
grams for a 2.0 mM solution of unsubstituted nitrobenzene in DMF–
Controlled-potential (bulk) electrolyses of 1-(2-chloroethyl)-
2-nitrobenzene (1) and 1-(2-bromoethyl)-2-nitrobenzene (2).— Com-
piled in Table I are coulometric data and product distributions for
controlled-potential ͑bulk͒ electrolyses of 5–20 mM solutions of 1
or 2 at reticulated vitreous carbon cathodes in DMF containing
0
.050 M TMABF , a classic one-electron reversible process was
4
observed and all three curves were superimposable. Curves B and C
of Fig. 2 appear to provide additional evidence for the reaction
between the nitro radical–anion and the o-bromoethyl moiety of 2;
the first cathodic peak ͑attributed to reduction of the nitro group of
0.050 M TMABF and in the absence of a deliberately added proton
4
2
͒ is attenuated as a consequence of this reaction, and the second
donor; each entry represents the average of at least two separate
experiments. Coulometric n values indicate the average number of
electrons involved in the reduction of each molecule of 1 or 2.
Product yields listed in the table denote the absolute amount of
starting material incorporated into each species and are accurate to
peak is almost certainly due to reduction of the product of this
reaction ͑namely, intermediate 6, the electrochemical behavior of
which is discussed later in this paper͒. In another separate experi-
ment, we examined the cyclic voltammetric behavior of unsubsti-
Downloaded on 2015-01-17 to IP 128.193.164.203 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

F170
Journal of The Electrochemical Society, 157 ͑11͒ F167-F172 ͑2010͒
the adjacent o-chloroethyl moiety. Second, a solution containing a
mixture of 10.0 mM ͑2-chloroethyl͒benzene and 10.0 mM nitroben-
zene in DMFϪ0.050 M TMABF was electrolyzed at a reticulated
4
vitreous carbon cathode held at Ϫ0.50 V, and the product distribu-
tion ͑determined with the aid of GC-MS͒ consisted of unreacted
starting materials ͑Ͼ80%͒, styrene ͑ϳ10%͒, and a small amount
͑
Ͻ2%͒ of azobenzene, the last compound arising from reductive
40,41
coupling of nitrobenzene.
Results of the second experiment sup-
port those of the first experiment and, in addition, rule out the like-
lihood that the radical–anion of nitrobenzene can act catalytically to
promote reductive cleavage of the carbon–chlorine bond of ͑2-
chloroethyl͒benzene, since radical-derived products ͑ethylbenzene
and 1,4-diphenylbutane͒ expected from such a bimolecular redox
process were not detected; indeed, the potentials associated with
one-electron reduction of the nitro group and direct reductive cleav-
age of the chloroethyl substituent are too disparate to make this
redox process happen during an electrolysis of 1 by itself.
When 2 was subjected to direct bulk electrolysis at a reticulated
vitreous carbon cathode held at a potential of Ϫ0.45 V in DMF–
0
.050 M TMABF and in the absence of a proton donor, the coulo-
4
metric n value was found to be 1.8–1.9 and, as shown in entries 4–6
of Table I, the product distribution consisted of 1-nitro-2-
vinylbenzene ͑71–73%͒, 1H-indole ͑12–14%͒, and a dimeric species
͑5–7%͒. These results are very similar to those for the reduction of
Figure 3. Cyclic voltammograms recorded with a glassy carbon electrode
2
−1
͑
area = 0.077 cm ͒ at 100 mV s in oxygen-free DMF containing 0.050
M TMABF : ͑A͒ 2.0 mM 1 and ͑B͒ 2.0 mM 1 in the presence of 10 mM
4
phenol. For curve A, the scan goes from 0 to Ϫ2.0 to 0 V; for curve B, the
scan goes from 0 to Ϫ1.8 to 0 V.
1
, although the n value for reduction of 2 is larger, 3 is formed in
slightly lower yield, and both cyclic products ͑4 and 5͒ are obtained
in somewhat higher yields.
At this point, it is worthwhile to comment on the cyclic voltam-
metric peak currents observed in Fig. 1 and the coulometric n values
listed in Table I. First, if the cathodic peak currents measured at
Ϫ0.44 V for curves A and B of Fig. 1 are attributable to the revers-
ible one-electron reduction of nitrobenzene, it appears that the cur-
rents obtained for the first cathodic peaks of curves C and D in Fig.
Ϯ3% absolute; no starting material was detected at the end of any
electrolysis.
Entries 1–3 in Table I reveal that 1 undergoes reduction with an
n value of 1.7–1.8 when electrolyzed in the absence of any added
proton donor at a potential ͑Ϫ0.50 V͒ slightly more negative than its
first cathodic peak. Three electrolysis products were obtained: ͑i͒
1
are indicative of processes corresponding to n values of 1.3–1.4.
1-nitro-2-vinylbenzene, 3 ͑76–79% yield͒; ͑ii͒ 1H-indole, 4 ͑9–11%
We ascribe this extra consumption of electricity to follow-up elec-
trochemistry of an intermediate ͑proposed later in the mechanistic
portion of this paper͒ that arises from reduction of 1 or 2 on the time
scale of cyclic voltammetry. Second, a complete ͑bulk͒ electrolysis
of 1 or 2 typically requires on the order of 60 min ͑compared with
the 40 s needed for a full cathodic–anodic potential sweep in cyclic
voltammetry͒. Therefore, it is not surprising that the coulometric n
values in Table I range from 1.7 to 1.9, because there is more than
adequate time for electrogenerated intermediates to undergo addi-
tional chemical and electrochemical reactions, as evidenced by the
fact that 1H-indole and a dimer ͑both arising from overall four-
electron processes͒ are minor products obtained from bulk electroly-
ses of 1 and 2. When coulometric n values calculated theoretically
on the basis of the observed product distributions in Table I are
compared with the experimentally measured n values, we believe
there is reasonable agreement between these results, in particular,
due to the facts that our observed total product recoveries are only
91 Ϯ 2% and that any atmospheric oxygen entering the sealed elec-
trolysis cell will undoubtedly interact with the initially formed
radical–anion intermediate ͑causing an overconsumption of electric-
ity͒.
yield͒; and ͑iii͒ a dimer, 5, believed to be 1,1Ј-biindole ͑3–4%
yield͒. To obtain insight as to how the major product ͑3͒ might be
formed, two simple experiments were carried out. First, a solution
containing 10.0 mM 1 and 10.0 mM tert-butyllithium ͑serving as a
base͒ in DMF containing 0.050 M TMABF was stirred for 10 min,
4
after which an analysis of the solution by means of GC-MS showed
that 1 was converted quantitatively to 3. This result suggests that
during an electrolysis, the nitro radical–anion generated from 1
might act as a base to initiate an E2 elimination of chloride ion from
Table I. Coulometric data and product distributions for direct
electrolytic reduction of 1-(2-chloroethyl)-2-nitrobenzene (1) or
1
-(2-bromoethyl)-2-nitrobenzene (2) at reticulated vitreous car-
bon cathodes in DMF containing 0.050 M TMABF4.
Product distribution
a
͑
%͒
Starting material
E
͑V͒
b
c
d
Entry
͑mM͒
n
3
4
5
Total
1
2
3
4
5
6
1 ͑5͒
1 ͑10͒
1 ͑20͒
2 ͑5͒
2 ͑10͒
2 ͑20͒
Ϫ0.50 1.72
Ϫ0.50 1.78
Ϫ0.50 1.75
Ϫ0.45 1.81
Ϫ0.45 1.85
Ϫ0.45 1.86
79
77
76
73
71
73
9
4
3
3
6
5
7
92
91
89
91
90
93
To probe how the presence of a proton donor affects the bulk
reductions of 1 and 2, we repeated the experiments corresponding to
entries 2 and 5 of Table I, except that a fivefold molar excess of
either of two different proton donors ͑phenol or 2,4-pentanedione͒
was deliberately added to the system prior to an electrolysis. For
these electrolyses, regardless of which proton donor was present, the
average coulometric n value was 4 and 1H-indole was obtained in
11
10
12
14
13
a
Yield expressed as the percentage ͑determined by means of gas chro-
matography͒ of starting material incorporated into each product: 3
9
8–99% yield. Thus, in the presence of a suitable proton donor,
reductive cyclization of 1 and 2 occurs exclusively to afford
H-indole, which is not electroactive under the experimental condi-
=
1-nitro-2-vinylbenzene; 4 = 1H-indole; 5 = dimer.
b
1
Each number in parentheses is the millimolar concentration of start-
ing material.
Average number of electrons per molecule of starting material.
Addition of a proton donor ͑phenol or 2,4-pentanedione͒ leads to
tions. If one merely compares the cathodic peak currents at Ϫ0.44 V
for curves A and B in Fig. 3, the conclusion might be reached that
approximately twice as many electrons are transferred to 1 in the
presence than in the absence of a proton donor. However, the volt-
c
d
1
H-indole as the only significant product ͑see text for discussion͒.
Downloaded on 2015-01-17 to IP 128.193.164.203 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑11͒ F167-F172 ͑2010͒
ammetric peak ͑curve B͒ obtained with the added proton donor is
F171
very broad and of an integrated area that is definitely compatible
with the observed coulometric n value of 4 for a bulk electrolysis of
1
or 2.
Several bulk electrolyses of 1 and 2 were carried out at a poten-
tial ͑Ϫ1.20 V͒ corresponding to the second cathodic peak for
each compound. A complex mixture of products was always ob-
tained, including 1H-indole ͑Ͼ40%͒, the previously mentioned
dimer ͑ϳ15%͒, 2-vinylaniline ͑3%͒, 2-ethylaniline ͑trace͒, species
of higher molecular mass such as trimers or tetramers ͑Ͼ10%͒, and
as many as five more unidentified materials formed in small quan-
tities. These electrolyses lack synthetic value because the product
distributions are highly irreproducible.
Efforts to identify the dimeric product.— By employing both
GC-MS and GC-HRMS, we have concluded that a dimeric product
͑
1,1Ј-biindole, 5͒ is formed upon reduction of 1 or 2, and we pro-
pose that this species arises from nitrogen–nitrogen coupling of the
radical precursor of 1H-indole. Mass spectral data for the dimer are
+
+
as follows: MS ͑70 eV͒ m/z 232, M ͑100%͒; 204, ͓M − 28͔
+
+
͑
16%͒; 116, ͓M − 116͔ ͑91%͒; 106, ͓M − 126͔ ͑10%͒; high res-
olution chemical ionization mass spectrometry ͑HRMS͒ exact mass
calculated for C H N ͓M ͔, 232.1000; found, 232.0971. In prin-
+
16
12
2
ciple, coupling of two radical precursors of 1H-indole can occur in
four ways, three of which involve carbon–carbon bond formation
to give 1H,1ЈH-2,2Ј-biindole, 1H,1ЈH-3,3Ј-biindole, or
Figure 4. Cyclic voltammograms recorded with a glassy carbon electrode
2
−1
͑
area = 0.077 cm ͒ at 100 mV s in oxygen-free DMF containing 0.050
M TMABF and 2.0 mM 3. For curve A, the scan goes from 0 to Ϫ0.8 to 0
4
V; for curve B, the scan goes from 0 to Ϫ2.0 to 0 V.
1
H,1ЈH-2,3Ј-biindole and one of which proceeds via nitrogen–
nitrogen bond formation to afford 1,1Ј-biindole. Since a carbon–
carbon single bond is much more stable than a nitrogen–nitrogen
single bond, we readily synthesized 1H,1ЈH-2,2Ј-biindole,
and Ϫ0.27 V, respectively, that is associated with the nitro group.
When the cyclic voltammetric scan for 3 is extended to more nega-
tive values ͑curve B, Fig. 4͒, a second cathodic peak is observed at
Ϫ1.17 V, along with an anodic peak at Ϫ0.84 V.
Because 3 is the principal product of the reduction of both 1 and
2 in the absence of a proton donor ͑Table I͒, it is useful to know
whether 3 can be further reduced to some new species at its first
cathodic peak potential. For this purpose, we conducted bulk elec-
trolyses of 5.0 mM solutions of 3 at reticulated vitreous carbon
cathodes held at Ϫ0.50 V in DMF containing 0.050 M TMABF4 in
the absence and presence of a fivefold molar excess of a proton
donor ͑phenol͒. For experiments done without added phenol, the
products were unreacted 3 ͑Ͼ60%͒, 1H-indole ͑ϳ20%͒, 1-amino-
1
H,1ЈH-3,3Ј-biindole, and 1H,1ЈH-2,3Ј-biindole, as described ear-
lier in this report. When the gas chromatographic retention time for
each of these three dimers was measured independently under iden-
tical experimental conditions, the difference among them was no
more than 5–10 s, but the retention time for the dimer obtained from
reduction of 1 or 2 was approximately 5 min longer, signifying that
the dimeric species we obtained is none of the three biindoles with a
carbon–carbon bond linkage. Therefore, we infer that the nitrogen–
nitrogen bonded 1,1Ј-biindole is the electrolytically formed dimer,
even though the nitrogen–nitrogen single bond enthalpy
−
1
͑
160 kJ mol ͒ is less than one-half of the carbon–carbon single
−
1
bond enthalpy ͑356 kJ mol ͒. Indirect evidence for the existence
of 1,1Ј-biindole has appeared in a recent paper by Masson and
2-vinylbenzene ͑ϳ6%͒, and a dimeric species ͑ϳ4%͒. These find-
4
2
43,44
Gagné, who investigated the reaction between 1H-indole and poly-
phosphoric acid at 150°C. Using Fourier transform infrared ͑FTIR͒
spectroscopy, these workers observed that the intensity of the amine
ings are supported by two earlier publications,
which indicate
that 3 can be catalytically reduced to 1H-indole and 1-amino-2-
vinylbenzene by the use of carbon monoxide at high pressure; the
relative yields of the two products can be altered if reaction condi-
tions are changed. When we electrolyzed a 5.0 mM solution of 3 in
the presence of 25.0 mM phenol, 1H-indole was obtained as the
major product along with tiny amounts of both 1-amino-2-
vinylbenzene and the dimer, and none of the starting material re-
mained unreduced. All of these results for the reduction of 3 imply
that the two minor products ͑1H-indole and the dimer͒ derived from
the electrolysis of 1 or 2 in the absence of a proton donor emanate
from reduction of a precursor of 3.
͑
N–H͒ absorbance of 1H-indole decreased substantially as a conse-
quence of the reaction, a result attributed to thermally induced head-
to-head coupling of 1H-indole to afford 1,1Ј-biindole. Moreover, if
one envisages that a dimer might arise via attack of the radical
precursor ͑C H N•͒ of 1H-indole on one of the carbon–carbon
8
6
double bonds of an already formed molecule of 1H-indole, the re-
sulting species would ͑after necessary follow-up abstraction of two
hydrogen atoms from the medium͒ have a chemical formula of
C H N and an exact mass of 234.1157, which differ from the
1
6
14
2
observed results cited earlier in this paragraph. We are attempting to
devise other experiments that can be used to identify the dimer
unambiguously.
Mechanistic features of the reduction of 1-(2-chloroethyl)-2-
nitrobenzene (1) and 1-(2-bromoethyl)-2-nitrobenzene (2).— A plau-
sible set of mechanistic steps that lead to formation of the various
products is proposed in Scheme 1. Earlier, in discussing the cyclic
voltammetric characteristics of nitrobenzene, we indicated that the
radical–anion can be obtained through a one-electron reduction.
Therefore, as illustrated in Reaction 1, the reversible one-electron
reduction of the nitro group of 1 or 2 to form the radical–anion is
proposed as the initial step in the electrochemical process; after this
initial electron transfer, we envisage an intramolecular proton trans-
fer which eliminates a halide ion to form intermediate ͑6͒ via an E2
reaction. To account for the appearance of 1-nitro-2-vinylbenzene
͑3͒ as the major product of the reduction of 1 or 2 in the absence of
Electrochemical behavior of 1-nitro-2-vinylbenzene (3).— To add
to the body of information about the reduction of 1 and 2, an ex-
amination of the electrochemical behavior of 1-nitro-2-vinylbenzene
͑
3͒ was carried out by means of cyclic voltammetry and controlled-
potential ͑bulk͒ electrolysis. As displayed in Fig. 4, cyclic voltam-
mograms for the reduction of 2.0 mM solutions of 3 at a glassy
carbon electrode in DMF containing 0.050 M TMABF under an
4
argon atmosphere are nearly identical to those for nitrobenzene seen
in Fig. 1. Curve A of Fig. 4 reveals the reversible, one-electron
redox couple, with cathodic and anodic peak potentials of Ϫ0.47
Downloaded on 2015-01-17 to IP 128.193.164.203 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

F172
Journal of The Electrochemical Society, 157 ͑11͒ F167-F172 ͑2010͒
H
H
Indiana University assisted in meeting the publication costs of this ar-
ticle.
X
X
H
H
H
H
H
H
_
e
E2
X
H
H
H
References
(1)
_
_
O
O
OH
N
O
.N
.
N
1. W. Kemula and T. M. Krygowski, in Encyclopedia of Electrochemistry of the
Elements, Vol. 13, A. J. Bard and H. Lund, Editors, p. 77, Marcel Dekker, New
York ͑1979͒.
2. A. J. Fry, in The Chemistry of Amino, Nitroso, and Nitro Compounds and Their
Derivatives, S. Patai, Editor, p. 319, John Wiley & Sons, New York ͑1982͒.
O
O
..
..
1
2
: X = Cl
: X = Br
6
3
. H. Lund, in Organic Electrochemistry, 4th ed., H. Lund and O. Hammerich, Edi-
tors, p. 379, Marcel Dekker, New York ͑2001͒.
. S. Torii, Electro-organic Reduction Synthesis, p. 190, Kodansha, Tokyo/Wiley-
VCH, Weinheim ͑2006͒.
oxidation by O2
during workup
(
2)
3)
6
6
4
NO2
3
5. W. H. Smith and A. J. Bard, J. Am. Chem. Soc., 97, 5203 ͑1975͒.
6
. M. F. Bento, M. J. Medeiros, M. I. Montenegro, C. Beriot, and D. Pletcher, J.
Electroanal. Chem., 345, 273 ͑1993͒.
. J.-C. Gard, J. Lessard, and Y. Mugnier, Electrochim. Acta, 38, 677 ͑1993͒.
_
_
OH
_
e
(
7
N
O
8. D. S. Silvester, A. J. Wain, L. Aldous, C. Hardacre, and R. G. Compton, J. Elec-
troanal. Chem., 596, 131 ͑2006͒.
9
. R. Hazard, M. Jubault, C. Mouats, and A. Tallec, Electrochim. Acta, 31, 489
1986͒.
7
͑
1
0. R. Hazard, M. Jubault, C. Mouats, and A. Tallec, Electrochim. Acta, 33, 1335
͑1988͒.
_
_
_
OH
H+
_
e
+
e
(
4)
5)
7
9
1
1
1. J. Y. David, J. P. Hurvois, A. Tallec, and L. Toupet, Tetrahedron, 51, 3181 ͑1995͒.
2. A. Alberti, P. Carloni, L. Greci, P. Stipa, R. Andruzzi, G. Marrosu, and A. Trazza,
J. Chem. Soc., Perkin Trans. 2, 1019 ͑1991͒.
3. W. M. Moore and D. G. Peters, Tetrahedron Lett., 13, 453 ͑1972͒.
4. W. M. Moore, A. Salajegheh, and D. G. Peters, J. Am. Chem. Soc., 97, 4954
N
N
OH
8
9
1
1
(
+
dimer
͑
1975͒.
15. B. C. Willett, W. M. Moore, A. Salajegheh, and D. G. Peters, J. Am. Chem. Soc.,
01, 1162 ͑1979͒.
N
H
5
1
4
1
6. R. Shao, J. A. Cleary, D. M. La Perriere, and D. G. Peters, J. Org. Chem., 48, 3289
͑1983͒.
_
_
_
OH
, H+, e
_
e
,
(
6)
3
7
17. R. Shao and D. G. Peters, J. Org. Chem., 52, 652 ͑1987͒.
18. M. S. Mubarak, D. D. Nguyen, and D. G. Peters, J. Org. Chem., 55, 2648 ͑1990͒.
19. W. A. Pritts and D. G. Peters, J. Electroanal. Chem., 380, 147 ͑1995͒.
20. A. P. Esteves, D. M. Goken, L. J. Klein, M. J. Medeiros, and D. G. Peters, J.
Electroanal. Chem., 560, 161 ͑2003͒.
Scheme 1.
2
1. W. Pluempanupat, O. Chantarasriwong, P. Taboonpong, D. O. Jang, and W. Chava-
siri, Tetrahedron Lett., 48, 223 ͑2007͒.
2. R. S. Tipson, J. Org. Chem., 9, 235 ͑1944͒.
3. K. B. Wiberg and B. R. Lowry, J. Am. Chem. Soc., 85, 3188 ͑1963͒.
4. S. Alunni, R. Ruzziconi, and O. Teofrasti, Res. Chem. Intermed., 25, 483 ͑1999͒.
25. C. Koradin, W. Dohle, A. L. Rodriguez, B. Schmid, and P. Knochel, Tetrahedron,
an added proton donor, we show in Reaction 2 that 6 undergoes
oxidation by dioxygen when the electrolyzed solution is exposed to
the atmosphere during workup of the catholyte; similar behavior is
well known for the electrogenerated radical–anion of ordinary ni-
trobenzene. In an actual bulk electrolysis of 1 or 2 in an oxygen-free
environment, intermediate 6 can be expected to accept another elec-
tron, followed by expulsion of hydroxide ion to afford 1-nitroso-2-
vinylbenzene ͑7͒, as shown by Reaction 3. As quickly as 7 is pro-
duced, it can be reduced, protonated, and reduced again, as seen in
Reaction 4, to form an intermediate ͑8͒ that loses hydroxide ion to
afford a nitrene ͑9͒. Nitrenes have been proposed as highly reactive
intermediates that arise from the reductive deoxygenation of deriva-
tives of 1-nitro-2-vinylbenzene in systems such as refluxing triethyl
2
2
2
59, 1571 ͑2003͒.
2
2
6. J. Bergman, Acta Chem. Scand., 25, 1277 ͑1971͒.
7. N. Wahlström, J. Slätt, B. Stensland, A. Ertan, J. Bergman, and T. Janosik, J. Org.
Chem., 72, 5886 ͑2007͒.
8. U. Berens, J. M. Brown, J. Long, and R. Selke, Tetrahedron: Asymmetry, 7, 285
͑1996͒.
9. V. Bocchi and G. Palla, Tetrahedron, 40, 3251 ͑1984͒.
0. K. L. Vieira and D. G. Peters, J. Electroanal. Chem., 196, 93 ͑1985͒.
1. J. A. Cleary, M. S. Mubarak, K. L. Vieira, M. R. Anderson, and D. G. Peters, J.
Electroanal. Chem., 198, 107 ͑1986͒.
32. M. P. Foley, P. Du, K. J. Griffith, J. A. Karty, M. S. Mubarak, K. Raghavachari, and
D. G. Peters, J. Electroanal. Chem., 647, 194 ͑2010͒.
3. L. W. Marple, Anal. Chem., 39, 844 ͑1967͒.
4. C. W. Manning and W. C. Purdy, Anal. Chim. Acta, 51, 124 ͑1970͒.
5. J. L. Hall and P. W. Jennings, Anal. Chem., 48, 2026 ͑1976͒.
6. W. A. Pritts, K. L. Vieira, and D. G. Peters, Anal. Chem., 65, 2145 ͑1993͒.
2
2
3
3
4
5
phosphite or a palladium͑II͒ trifluoroacetate–phenanthroline cata-
46
3
3
3
3
lyst in DMF under carbon monoxide. If a nitrene intermediate
does form, it should enter into the pathways for production of both
1
H-indole ͑4͒ and the dimer ͑5͒, as suggested by Reaction 5. Finally,
for our separate examination of the electroreduction of 3 by itself, it
37. W. Kemula and R. Sioda, J. Electroanal. Chem., 7, 233 ͑1964͒.
3
3
4
4
4
4
8. B. S. Jensen and V. D. Parker, J. Chem. Soc., Chem. Commun., 367 ͑1974͒.
9. P. Zuman and Z. Fijalek, J. Electroanal. Chem., 296, 583 ͑1990͒.
0. A. Cyr, P. Huot, G. Belot, and J. Lessard, Electrochim. Acta, 35, 147 ͑1990͒.
1. E. Steudel, J. Posdorfer, and R. N. Schindler, Electrochim. Acta, 40, 1587 ͑1995͒.
2. J.-F. Masson and M. Gagné, Energy Fuels, 22, 3402 ͑2008͒.
3. C. Crotti, S. Cenini, B. Rindone, S. Tollari, and F. Demartin, J. Chem. Soc., Chem.
Commun., 784 ͑1986͒.
is likely that 7 arises according to Reaction 6, after which Reactions
4
and 5 ensue. Further research is planned in an effort to provide
more details about these intriguing processes.
Acknowledgment
High-resolution mass analyses were performed in the Indiana
University Mass Spectrometry Facility; the HRMS system was pur-
chased with funds provided by the National Institutes of Health
grant no. 1S10RR016657-01.
4
4
4
4. S. Tollari, A. Penoni, and S. Cenini, J. Mol. Catal. A: Chem., 152, 47 ͑2000͒.
5. B. C. G. Söderberg, Curr. Org. Chem., 4, 727 ͑2000͒.
6. I. W. Davies, J. H. Smitrovich, R. Sidler, C. Qu, V. Gresham, and C. Bazaral,
Tetrahedron, 61, 6425 ͑2005͒.
Downloaded on 2015-01-17 to IP 128.193.164.203 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).