The first ESR observation of radical species in subcritical water

Article abstract of DOI:10.1016/j.tet.2008.11.074

Triphenylmethanol was treated in subcritical and supercritical water. A radical species, triphenylmethyl radical, was directly generated from triphenylmethanol in subcritical and supercritical water without using any radical initiator. The radical formati

Full text of DOI:10.1016/j.tet.2008.11.074

Tetrahedron 65 (2009) 807–810
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
The first ESR observation of radical species in subcritical water
Kazuya Kobiro a , Masahiko Matsura , Hiroyuki Kojima , Koichi Nakahara
,
a
a
b
*
a
Department of Environmental Systems Engineering, Kochi University of Technology, 185 Miyanokuchi, Tosayamada, Kami, Kochi 782-8502, Japan
Institute for Fundamental Research, Suntory Ltd., 1-1-1 Wakayamadai, Shimamoto-cho, Mishima-gun, Osaka 618-8503, Japan
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 August 2008
Received in revised form 14 November 2008
Accepted 18 November 2008
Available online 27 November 2008
Triphenylmethanol was treated in subcritical and supercritical water. A radical species, triphenylmethyl
radical, was directly generated from triphenylmethanol in subcritical and supercritical water without
using any radical initiator. The radical formation was confirmed by direct electron spin resonance (ESR)
measurement in high-temperature and high-pressure subcritical water and by capturing the radical
intermediate using hydrogen donors in supercritical water.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
generation of radical species from triphenylmethanol (1), direct
observation of radical species in sub-CW using ESR, and the re-
Subcritical water (sub-CW) and supercritical water (SCW) have
attracted much attention in recent years not only from the view
point of green chemistry but also their unique properties, such as
low viscosity, high density, low polarity, and high water solubility
action behavior of the generated radical species in SCW.
2. Results and discussion
1
of organic compounds. Recently, sub-CW and SCW have been used
The triphenylmethyl radical is considered one of the most stable
as reaction media and reaction catalysts for organic reactions, such
2
radicals. We selected this radical to investigate the reaction be-
havior of radical(s) generated in sub-CW and SCW, because of the
following reasons: (i) the radical has no hetero bond that can be
easily cleaved under the drastic reaction conditions of sub-CW and
SCW; (ii) the benzene moieties of the radical are known to exhibit
an attractive interaction with water molecules of sub-CW and
as H–D exchange reaction in supercritical deuterium oxide, the
3
Beckmann rearrangement of cyclohexanone oxime, pinacol rear-
3
rangement of 2,3-dimethyl-2,3-butanediol, the Cannizzaro re-
4
5
6
action of formaldehyde and benzaldehyde, the Heck reaction,
7
8
C–Si bond cleavage, benzil–benzilic acid rearrangement, and the
Claisen–Schmidt reaction. More recently, unique reactions such as
9
19
SCW; and (iii) the radical has a highly symmetrical structure,
which simplifies the analysis of reaction products. Although there
are several precursors from which the triphenylmethyl radical can
be obtained, 1 seems to be one of the most suitable substrates, since
the cleaved OH species and water molecules has very similar
structural analogy and the cleaved OH would adapt to the sur-
rounding water molecules.
the oxidation of alcohols to obtain the corresponding aldehydes or
ketones without the use of any oxidizing reagent or catalyst have
been reported.1
0,11
All the above-mentioned examples are con-
þ
ꢀ
cerned with ionic reactions that are catalyzed by H and/or OH
ions generated from water in situ
2
–9
or by hydrogen bonding
between the substrates and two water molecules to achieve an
10–13
eight-membered ring transition state of the reaction.
On the other hand, radical reactions are another common fea-
Two possibilities exist when the carbon–OH bond of 1 is cleaved
(Scheme 1). One is the ionic (heterolytic) fission of the bond to
generate a triphenylmethyl cation and a hydroxyl anion. The other is
the radical (homolytic) fission of the bond to generate triphenyl-
methyl and hydroxyl radicals. Another radical fission to generate
diphenylhydroxymethyl and phenyl radicals is also possible. Semi-
14
ture of the reactions that occur particularly in less dense SCW as
exemplified in the waste treatment of polymers and hazardous
15
compounds in the presence of a large amount of oxygen. Addition
1
6
17
of hydrogen peroxide, pulse radiolysis of water, and acoustic
cavitation18 are also used to generate radical species in SCW. Al-
2
0
empirical molecular orbital calculations (AM1) were performed to
estimate thebond dissociation energies of theC–OH and C–Ph bonds
of 1 in vacuum and in water (Scheme 1). The hydration effect was
though radical generation in water using these approaches is rel-
atively easy, little is known about the direct observation of radical
species generated in sub-CW and SCW. In this study, we report the
21
also considered using COSMO method in the calculations. The
bond dissociation energy for the ionic fission in water to generate
triphenylmethyl and hydroxyl ions was calculated to be 71 kcal/mol,
while those for the radical fissions to generate diphenylhydroxy-
methyl and phenyl radicals, and triphenylmethyl and hydroxyl
*
Corresponding author.
E-mail address: kobiro.kazuya@kochi-tech.ac.jp (K. Kobiro).
0
040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.11.074

808
K. Kobiro et al. / Tetrahedron 65 (2009) 807–810
bond dissociation energy
OH
kcal/mol
Ph
in vacuum in water
Ph
Ph
Ph
+
OH
ionic fission
200
48
71
46
45
Ph
Ph
1
Ph
H
Ph OH
H
OH
OH
Ph
+
+
+
+
Ph
Ph
Ph
Ph radical fission
radical fission
Ph
Ph
Ph
Ph
Ph
Ph
2
3
4
1
Scheme 2. Reaction of 1 in supercritical water.
OH
46
Scheme 1. Ionic and radical fissions of C–OH and C–Ph bonds of 1 and their bond
dissociation energies in vacuum and in water calculated by MOPAC (AM1/COSMO
method).
Table 1
Reaction of 1 in supercritical water with and without hydrogen donora
Entry
Hydrogen donor
Time (min)
Recovery (%)
Yield (%)
radicals were 46 and 45 kcal/mol, respectively. The corresponding
dissociation energies in vacuum were calculated to be 200, 48, and
1
2
3
4
1
2
3
no
30
20
30
9
1
0
26
>99
96
42
0
3
9
0
0
b
4
6 kcal/mol, respectively. Based on these calculations, the radical
5
b
fissions would be more plausible than the ionic fission.
6
ESR spectroscopy is well known to be one of the most powerful
methods to detect radical species directly. However, the glass
sample tubings as well as the ESR instrument itself must tolerate
the drastic conditions of such high-temperature and high-pressure.
A commercial ESR instrument was modified to be equipped with
the temperature control unit, which maintains the sample tem-
a
ꢁ
Reaction conditions: 0.192 mmol of 1, 380 C, 0.35 g/mL water density.
b
An equimolar amount of hydrogen donor (0.192 mmol) was used.
equimolar amount of hydrogen donor 5 or 6, (almost) qualitative
amount of reduction product 2 was obtained in each case (entries 2
and 3 in Table 1). These results clearly indicate that the reaction
proceeds via the triphenylmethyl radical intermediate, and that the
radical easily abstracts a hydrogen atom of the hydrogen donors
coexisting in the reaction system to give the reduction product 2, as
shown in Scheme 3. If no other hydrogen source was available,
however, the radical would abstract a hydrogen atom from the
other molecules instead to afford the same reduction product 2 or
the radical would be submitted to further reactions to yield 3 and 4.
perature being high by blowing the temperature controlled N
2
gas.
ꢁ
We succeeded in observing the ESR signals of 1 in sub-CW (300 C)
using a quartz sample tubing instead of glass ones. As seen in
Figure 1, characteristic ESR signals were observed, which clearly
indicate the formation of radical species in the sub-CW. Due to the
uncertainty of the hyperfine structures of the observed ESR signals,
it is difficult to put forward a conclusive discussion of the radical
species. However, the observed signals could be derived from tri-
phenylmethyl radicals generated in sub-CW, since the signal shape
and coupling constant (a¼0.22 mT) are similar to the reported
2
5
H
OH
Ph
Ph
H
OH
Ph
H
22
Ph
Ph
shape and value in toluene (a
p
¼0.286 mT). To our knowledge,
Ph
Ph
Ph
2
Ph
this is the first successful example of the direct ESR measurement of
radical species in sub-CW.
1
Scheme 3. Plausible reaction pathway of 1 to give 2 in supercritical water.
When alcohol 1 was treated with SCW in an SUS316 batch type
2
3
tubular reactor in a preparative scale, the reaction proceeded
smoothly to provide 26% of reduction product triphenylmethane
To clarify the overall reaction behaviorof compound 1 in SCW, the
reaction temperature and water density were varied, as shown in
Table 2. In the presence of water, the recovery of substrate 1 de-
creased with increasing temperature and the (almost) complete
2
4
(
2), 42% of dehydrated cyclization product 9-phenylfluorene (3),
and 9% of 9-phenyl-9-fluorenol (4) with 9% recovery of substrate 1
Scheme 2 and entry 1 in Table 1). If these products are derived
(
ꢁ
consumption of 1 was achieved at above 400 C (entries 3–6). Re-
from the radical intermediate(s), as suggested by the ESR mea-
surement, it is reasonable to use a hydrogen donor such as 9,10-
dihydroanthracene (5) or 2,5-di(tert-butyl)-1,4-hydroquinone (6)
to trap the radical intermediate(s). When alcohol 1 was treated
duction product 2 and dehydrated cyclization product 3 were
obtained at all test temperatures. Yield of 2 increased gradually with
increasing temperature, although the yield of 3 almost saturated at
around 400 C (entries 3–6). Cyclization product 4 having an OH
group was produced only at low temperatures (<380 C) with a low
ꢁ
ꢁ
with SCW (380 C, 0.35 g/mL water density) in the presence of an
ꢁ
2
6
yield (9%)(entries1 and2). Formationof2 wasalmostindependent
Table 2
a
Reaction of 1 in subcritical and supercritical water
Entry Temp Reaction
Water
Recovery Yield
ꢁ
(
C)
time (min) density (g/mL) (%)
(%)
1
2
3
4
7
1
2
3
4
5
6
7
330
380
400
420
440
460
420
30
30
30
30
30
30
30
0.35
0.35
0.35
0.35
0.35
0.35
0
36(9)
9(5)
0
2(1)
1(2)
0
16(2) 26(10) 9(4)
26(3) 42(6) 9(3)
0
0
0
0
0
0
27(1) 56(5)
31(3) 47(4)
37(5) 44(19)
38(1) 47(1)
0
0
0
0
0
16(3)
36(5)
0
11(2)
a
Figure 1. ESR spectrum of 1 in subcritical water (2.3
mmol of 1/46
mL of water at
Recovery and yields are mean values of at least three runs. Standard deviation is
given in parentheses.
ꢁ
300 C).

K. Kobiro et al. / Tetrahedron 65 (2009) 807–810
809
of the water density (entries 4 and 7). Interestingly, benzophenone
7) was obtained instead of 3 and 4 in the absence of water.
OH
Ph
(
Ph
Ph
1
2
.1. Reaction pathway
Considering the experimental data, we propose the following
Ph
Ph
OH
Ph
reaction pathway. In the case of simple pyrolysis (no water), ho-
molyticfissionwouldoccurtogive radicalspecies viaboth patha and
path b (Scheme 4). Path a should lead to the formation of diphe-
nylhydroxymethyl and phenyl radicals. Benzophenone (7) should be
formed if a hydrogen atom of the diphenylhydroxymethyl radical is
abstracted.27 On the other hand, path b should lead to the formation
of triphenylmethyl and hydroxyl radicals. Then, reduction product
OH
Ph
Ph
2
should be produced if the triphenylmethyl radical abstracts a
Ph
Ph
4
Ph
hydrogen atom fromthe other molecules that might be derived from
the residues such as decomposed products.
H
5
H
1
4
3
3
5
1
[
1,5]
2
2
H
H H
9
8
PhH
OH
O
Ph
[1,5]
path a
path b
Ph
OH
Ph
Ph 7
Ph
Ph
b
Ph
H
a
Ph
Ph
Ph
1
H
Ph
OH
3
Ph
Ph
Ph
Ph 2
Scheme 6. Reaction pathway of triphenylmethyl radical giving 3.
Scheme 4. Plausible reaction pathways of 1 in the absence of water.
The presence of water, however, changes the reaction behavior
Scheme 5). When an organic compound is immersed in SCW, water
3. Conclusion
(
molecules surround and strongly solvate the organic molecule to
We successfully observed the ESR signals of a radical species of
2
8
form a water ‘cage’ around it. The cage in which the radical species
would be generated should lengthen the lifetime of the radical
species bysuppressing the diffusionprocess of the organic molecule.
This should promote the recombination of the radical species and
triphenylmethyl radical in sub-CW generated from triphenylme-
thanol. The water density of the reaction media played a crucial
role in the generation of radical species. In the absence of water,
benzophenone was obtained along with a large amount of tri-
phenylmethane, while dehydrated cyclization product 3 was
obtained as the main product along with 2 in the presence of
water.
2
9
prevent the reaction from proceeding through path a to lead to 7.
"
cage" by water
molecules
OH
Ph
PhH
4
. Experimental section
O
Ph
Ph
Ph
Ph
Ph 7
path a
path b
4.1. Materials and methods
OH
b
a
"
cage" by water
molecules
Ph
Ph
Ph
GC analyses were done on a Shimadzu GC-17A equipped with
a CBP-5 glass capillary column. GC–MS analyses were performed on
a Shimadzu QP5050. H NMR spectra were recorded on a Varian
H
Ph
Ph
1
Ph
1
OH
Ph
Ph 2
Ph
Unity Inova 400 spectrometer. ESR spectrum was recorded on
a JEOL JES-RE-2X-ESR spectrometer. Triphenylmethanol (1), tri-
phenylmethane (2), 9,10-dihydroanthracene (5), n-heptadecane,
and n-heneicosane were purchased from Nacalai Tesque, Inc.
9-Phenyl-9-fluorenol (4) was purchased from Aldrich. 2,5-Di(tert-
butyl)-1,4-hydroquinone (6) was purchased from Wako Pure
Chemicals Industries, Ltd. Reverse osmotic water was used as a re-
action medium in all reactions.
H O
2
3
Scheme 5. Plausible reaction pathways of 1 in the presence of water.
Moreover, the cage should also prevent the radical in-
termediates, triphenylmethyl and hydroxyl radicals, from diffusing
in path b. Now, if a hydrogen atom is supplied from outside, the
triphenylmethyl radical would abstract the hydrogen atom to af-
ford reduction product 2, particularly in the presence of hydrogen
sources such as the hydrogen donors. Instead, if the fairly reactive
hydroxyl radical abstracts a hydrogen atom of the triphenylmethyl
radical, dehydrated cyclization product 3 should be obtained
4.2. Reaction of triphenylmethanol (1) in water
Alcohol 1 (50 mg, 0.192 mmol) and water (N2 bubbled) were
introduced into an SUS316 tubular reactor of 10.0 mL inner volume.
The reactor was purged by N and sealed with a screw cap, which
was equipped with a thermocouple for the measurement of the
2
(
Scheme 5). The reaction pathway to give 3 from the triphenyl-
reactor temperature. The reactor was then put in a melting-salt
bath, which was kept at an appropriate temperature, and heated for
an appropriate time. It took about 20–30 s for the inner reactor
temperature to be raised up to 380–460 C. The reaction was
quenched by placing the reactor into an ice-water bath. After the
methyl radical is shown in Scheme 6. The triphenylmethyl radical
would rearrange and cyclize successively to give intermediate 8,
which would undergo a [1,5] hydrogen shift twice to give 3, as
ꢁ
3
0
31
demonstrated by Miyagawa et al. and Shi et al.

810
K. Kobiro et al. / Tetrahedron 65 (2009) 807–810
reactor was completely cooled, the screw cap was opened and the
reaction mixture was extracted with ether. The organic phase was
separated and the solvent was evaporated in vacuo to give the
crude products. The products were purified further by GPC (JAI gel
11. Wang, P.; Kojima, H.; Kobiro, K.; Nakahara, K.; Arita, T.; Kajimoto, O. Bull. Chem.
Soc. Jpn. 2007, 80, 1828–1832.
1
2. Takahashi, H.; Hisaoka, S.; Nitta, T. Chem. Phys. Lett. 2002, 363, 80–86.
3. Takahashi, H.; Hashimoto, H.; Nitta, T. J. Chem. Phys. 2003, 119, 7964–7971.
14. B u¨ hler, W.; Dinjus, E.; Ederer, H. J.; Kruse, A.; Mas, C. J. Supercrit. Fluids 2002, 22,
7–53.
5. Rice, S. F.; Steeper, R. R. J. Hazard. Mater. 1998, 59, 261–278.
6. Lee, G.; Nunoura, T.; Matsumura, Y.; Yamamoto, K. Ind. Eng. Chem. Res. 2002, 41,
5427–5431.
7. Feng, J.; Aki, S. N. V. K.; Chateauneuf, J. E.; Brennecke, J. F. J. Am. Chem. Soc. 2002,
24, 6304–6311.
8. Yim, B.; Maeda, Y. Chem. Lett. 2002, 31, 452–453.
19. Schneider, G. M. Ber. Bunsen-Ges. Phys. Chem. 1972, 76, 325–331.
0. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985,
07, 3902–3909.
21. Klamt, A.; Sch u¨ u¨ rmann, G. Z. J. Chem. Soc., Perkin Trans. 2 1993, 799–805.
1
3
1
Hþ2H, chloroform), if necessary, and were identified by compar-
1
1
1
ing their H NMR and GC–MS spectra with those of authentic
samples. Recovery of 1 and yields of the products were determined
by internal standard method on GC and/or integration ratio of
signals on 1H NMR spectra. n-Heptadecane and n-heneicosane
were used as internal standards for GC measurements.
1
1
1
2
1
4
.3. Measurement of ESR spectra
22. Maki, A. H.; Allendoerfer, R. D.; Danner, J. C.; Keys, R. T. J. Am. Chem. Soc. 1968,
90, 4225–4231.
Alcohol 1 (0.60 mg, 2.3ꢂ10ꢀ mol) and 0.046 mL of water (N
6
2
2
3. Same products, 2, 3, and 4, were obtained in both sub-CW and SCW as shown
in Table 2. Taking into account that physical properties of sub-CW and SCW,
such as density, polarity, and ion product, change continuously between sub-
CW and SCW in accordance with the temperature, we think it would be pos-
sible to discuss the reaction mechanism in sub-CW and SCW conditions in the
same manner.
bubbled) was placed in a quartz tubing (o.d.: 2.7 mm, i.d.: 1.5 mm).
The tubing was sealed (ca. 8 cm height) after the air was replaced
by argon. Pressure resisting test of the sample quartz tubings was
performed by heating at 300 C for 5 min before use. The tubing
was allowed to measure the ESR spectra at 300 C. The sample
ꢁ
ꢁ
24. Asher, S. E.; Browne, S. E.; Cornwall, E. H.; Frisoli, J. K.; Salot, E. A.; Sauter, E. A.;
Trecoske, M. A.; Veale, P. S., Jr. J. Am. Chem. Soc. 1984, 106, 1432–1440.
tubing was heated by blowing the temperature controlled N
2
gas in
2
5. Two electron transfer followed by proton shift from hydrogen donor, 5 or 6, to
could be possible to afford the reduction product, however the calculated
a double quartz tubing of high-temperature unit.
1
energies (AM1) necessary for the first electron transfer processes from 5 and 6
to 1 are quite larger (183 and 165 kcal/mol, respectively) than the bond dis-
sociation energy for the homolytic fission of 1 (46 kcal/mol, Scheme 1).
References and notes
ꢁ
2
6. When alcohol 4 was treated in SCW (420 C, 30 min, 0.35 g/mL water density),
1
. For review, see: Akiya, N.; Savage, P. E. Chem. Rev. 2002, 102, 2725–2750 and
references cited therein.
41% of 3 along with 48% recovery of 4 were obtained. Although the reaction
pathway to give 4 from 1 is not clear yet, 4 could be derived from 3 by attack of
the generated hydroxyl radical two times.
2
3
. Yao, J.; Evilia, F. J. Am. Chem. Soc. 1994, 116, 11229–11233.
. Ikushima, Y.; Hatakeda, K.; Sato, O.; Yokoyama, T.; Arai, M. J. Am. Chem. Soc.
27. Formation of benzopinacol could be possible, if the reaction proceeds via the
ꢁ
2
000, 122, 1908–1918.
. Tsujino, Y.; Wakai, C.; Matsubayashi, N.; Nakahara, M. Chem. Lett. 1999, 28, 287–
88.
. Ikushima, Y.; Hatakeda, K.; Sato, O.; Yokoyama, T.; Arai, M. Angew. Chem., Int. Ed.
001, 40, 210–213.
. Zhang, R.; Zhao, F.; Sato, M.; Ikushima, Y. Chem. Commun. 2003, 1548–1549.
radical species. When benzopinacol was exposed to high-temperature (250 C,
4
5
6
GC measurement), disproportionation occurred to give almost equal amounts
of benzophenone (7) and benzhydrol. Judging from this result, benzopinacol
cannot be stable under the reaction conditions even if it is formed by coupling
of two diphenylhydroxymethyl radicals in the reaction.
2
2
28. Tucker, S. C.; Maddox, M. W. J. Phys. Chem. B 1998, 102, 2437–2453.
29. Coupling products of radicals such as biphenyl and tetraphenylmethane were
not detected. The water cage could prevent the two radical species from
meeting together to give the coupling products.
7. Itami, K.; Terakawa, K.; Yoshida, J.-i.; Kajimoto, O. J. Am. Chem. Soc. 2003, 125,
6058–6059.
8. Comisar, C. M.; Savage, P. E. Green Chem. 2005, 7, 800–806.
9
1
. Nolen, S. A.; Liotta, C. L.; Eckert, C. A.; Gl a¨ ser, R. Green Chem. 2003, 5, 663–669.
0. Arita, T.; Nakahara, K.; Nagami, K.; Kajimoto, O. Tetrahedron Lett. 2003, 44,
30. Miyagawa, N.; Karatsu, T.; Futami, Y.; Kunihiro, T.; Kiyota, A.; Kitamura, A. Bull.
Chem. Soc. Jpn. 1996, 69, 3325–3329.
1083–1086.
31. Shi, M.; Okamoto, Y.; Takamuku, S. J. Chem. Res., Synop. 1990, 346–347.