Ozonation of 1,1,2,2-tetraphenylethene revisited: Evidence for electron- transfer oxygenations

Article abstract of DOI:10.1002/(SICI)1522-2675(20000412)83:4<801::AID-HLCA801>3.0.CO;2-6

Ozonolyses of 1,1,2,2-tetraphenylethene (TPE, 1) have been described many times in the literature, but the reports are contradictory. This reaction is particularly important for understanding the mechanism of alkene ozonolysis, in view of possible stabilization of reactive intermediates by aryl groups. Thus, systematic investigations of ozonolysis in both aprotic solvents and in protic solvents are reported here. Attention is directed to the following details that have been underestimated in the past: i) the actual electronic structure of ground-state ozone (O₃), ii) differentiation between strained and unstrained alkenes, iii) the significance of both the O₃ concentration and the TPE concentration, iv) the influence of various solvents, including pyridine, v) the influence of the reaction temperature, vi) the role of electron-transfer catalysis (ETC) and, yii) the effect of structural modifications. Our results suggest that ozonolysis of TPE (1) does not include a 1,3-dipolar reaction step, but represents a particularly interesting example of electron-donor (TPE)/electron-acceptor (O₃) redox chemistry. The present investigations include several crucial results. First, pure 3,3,6,6-tetraphenyltetroxane (3, m.p. 221°(dec.)) and pure tetraphenylethylene ozonide (4, m.p. 153°(dec.)) are prepared for the first time, although 3 and 4 have long been known. Second, the singlet diradical character of O₃, lessened by means of hypervalent-electron interaction and predicted by different calculations, is evidenced via reaction with the spintrap galvinoxyl (2,6-bis(1-1-dimethylethyl)-4-{[3,5-bis(1,1- dimethylethyl)-4-oxocyclohexa-2,5-dien-1-ylidene]methyl}phenoxy; 8), and the zwitterionic reaction behavior of ground-state O₃ is ruled out. Third, the electronacceptor ability of O₃ is evidenced by reactions with suitable tetraaryl ethylenes: it is enhanced by addition of catalytic amounts of protons or Lewis acids. Fourth, the observed distribution of the O₃ O-atoms to the two different olefinic C-atoms of the unsymmetric alkene 27b is in full agreement with an initial single-electron transfer (SET) step, followed by a radical mono-oxygenation to cause the crucial C,C cleavage. Final dioxygenation should lead to the generally known products (ozonides, tetroxanes, hydroperoxides). The regioselectivity is found to be inconsistent with the expected decay of an intermediate primary ozonide. Finally, the treatment of 1,2-bis(4-methoxyphenyl)acenaphthylene (36) with O₃ (simultaneous transfer of three O-atoms) leads to the same experimental result as a stepwise transfer of one O-atom followed by a transfer of two O- atoms.

Full text of DOI:10.1002/(SICI)1522-2675(20000412)83:4<801::AID-HLCA801>3.0.CO;2-6

Helvetica Chimica Acta ± Vol. 83 (2000)
801
Ozonation of 1,1,2,2-Tetraphenylethene Revisited: Evidence for Electron-
Transfer Oxygenations1)
by Kurt Schank*, Horst Beck2), Michael Buschlinger2), Jörg Eder3), Thomas Heisel, Susanne Pistorius2),
and Christiane Wagner2)
Department of Organic Chemistry, University of Saarland, D-66041 Saarbrücken
(Fax: ‡49(681)302 4747; e-mail: kschank@rz.uni-sb.de)
Ozonolyses of 1,1,2,2-tetraphenylethene (TPE, 1) have been described many times in the literature, but the
reports are contradictory. This reaction is particularly important for understanding the mechanism of alkene
ozonolysis, in view of possible stabilization of reactive intermediates by aryl groups. Thus, systematic
investigations of ozonolysis in both aprotic solvents and in protic solvents are reported here. Attention is
directed to the following details that have been underestimated in the past: i) the actual electronic structure of
ground-state ozone (O3), ii) differentiation between strained and unstrained alkenes, iii) the significance of both
the O3 concentration and the TPE concentration, iv) the influence of various solvents, including pyridine, v) the
influence of the reaction temperature, vi) the role of electron-transfer catalysis (ETC) and, vii) the effect of
structural modifications. Our results suggest that ozonolysis of TPE (1) does not include a 1,3-dipolar reaction
step, but represents a particularly interesting example of electron-donor (TPE)/electron-acceptor (O3) redox
chemistry. The present investigations include several crucial results. First, pure 3,3,6,6-tetraphenyltetroxane (3,
m.p. 2218 (dec.)) and pure tetraphenylethylene ozonide (4, m.p. 1538 (dec.)) are prepared for the first time,
although 3 and 4 have long been known. Second, the singlet diradical character of O3, lessened by means of
hypervalent-electron interaction and predicted by different calculations, is evidenced via reaction with the spin-
trap galvinoxyl (2,6-bis(1,1-dimethylethyl)-4-{[3,5-bis(1,1-dimethylethyl)-4-oxocyclohexa-2,5-dien-1-ylidene]-
methyl}phenoxy; 8), and the zwitterionic reaction behavior of ground-state O3 is ruled out. Third, the electron-
acceptor ability of O3 is evidenced by reactions with suitable tetraaryl ethylenes: it is enhanced by addition of
catalytic amounts of protons or Lewis acids. Fourth, the observed distribution of the O3 O-atoms to the two
different olefinic C-atoms of the unsymmetric alkene 27b is in full agreement with an initial single-electron
transfer (SET) step, followed by a radical mono-oxygenation to cause the crucial C,C cleavage. Final
dioxygenation should lead to the generally known products (ozonides, tetroxanes, hydroperoxides). The
regioselectivity is found to be inconsistent with the expected decay of an intermediate primary ozonide. Finally,
the treatment of 1,2-bis(4-methoxyphenyl)acenaphthylene (36) with O3 (simultaneous transfer of three O-
atoms) leads to the same experimental result as a stepwise transfer of one O-atom followed by a transfer of two
O-atoms.
1. Introduction. ± To the best of our knowledge, the first ozonolysis of 1,1,2,2-
tetraphenylethene (TPE; 1) was described in 1925 by Fischer and Müller [1]. These
authors found a quantitative formation of benzophenone (2) upon ozonizing 1 in
CHCl3 at 788 (Scheme 1). Later experiments, carried out mostly in CCl4 at 08, led to
formation of the tetroxane 3 as the main product (52 ± 59% yield) [2]. But these yields
were calculated on the assumption that only one molecule of peroxide could be formed
from one molecule of olefin [2a]. Since the stoichiometry of tetroxane formation via
Presented in part at the 3rd International Conference on Heteroatom Chemistry, Riccione, Italy, 1992.
Part of Ph.D. theses of H.B. (1988), M.B. (1994), S.P. (1992), and C.W. (1998), University of Saarland.
Diploma thesis of J.E., University of Saarland, 1994.
1
)
)
)
2
3

802
Helvetica Chimica Acta ± Vol. 83 (2000)
ozonolysis needs two molecules of olefin, these yields should be halved, and it should
be recognized that benzophenone (2) was actually the main product. Furthermore, in
1968, Criegee reported that TPE (1) is an olefin that does not itself form an ozonide [3],
and it was only three years later that he accomplished the synthesis of TPE ozonide [4],
even via a route that had been described by others to be ineffective [2d,e]. His results
were later confirmed [5], and his approach is applied here. Though the wide range of
the melting points (with decomposition) described for ozonide 4 (155 ± 1698) does not
offer a criterion for the purity of samples, the difference from the reported average
melting and decomposition temperatures of tetroxane 3 (200 ± 2288) was too large to be
ignored [4].
Scheme 1. Literature Results for the Ozonolysis of TPE (1)
a) CHCl3, 788 [1]. b) CCl4, 08 [2]. c) CCl4, 08 or liquid 2, 558 [4]. d) MeOH, CCl4, 08 [6]. e) MeOH, CHCl3,
788 [7].
Similar inconsistencies were reported for ozonolysis of TPE in the presence of
4
Â
MeOH. Whereas Renard and Fliszar [6] ) reported a quantitative yield (104 ± 120% on
the basis of O3 consumed) of diphenyl methoxymethyl hydroperoxide (5), Robertson
and Verzino [7] described formation of diphenylmethyl dihydroperoxide (6) instead of
Â
5. Renard and Fliszar [6] did not report the isolation of hydroperoxide 5; 5 was later
prepared by another route, isolated, and characterized (m.p. 62 ± 648) [8], and it was
proposed that formation of the established bis(hydroperoxide) 6 occurs by a reaction
involving H2O2 [7]. Although H2O2 can be generated by ozonation of alcohols [9], it is
known that MeOH, in particular, is not attacked by O3 at temperatures below 208.
Even at higher temperatures, MeOH does not react with O3 in the presence of a
reactive olefin, since the olefin reacts quantitatively with O3 faster than the O3 can react
with MeOH to form H2O2, or the H2O2 can form 6, even in a large excess of the more
nucleophilic MeOH (see Scheme 1).
4
)
See also Table 2 in [2d].

Helvetica Chimica Acta ± Vol. 83 (2000)
803
A common proposal made by researchers who reported formation of peroxidic
products 3 ± 6 [2][4 ± 7] is that ozonolysis of 1 produces the carbonyl oxide of
benzophenone, 7, as the reactive precursor. However, the only established behavior of
7 is its decomposition in matrix [10] or in solution [11] to yield benzophenone and
oxygen; 3 or 4 should have been formed, at least in solution, whenever 7 or 2 were
present, but were not found [11] (Scheme 2).
Scheme 2. Generation and Decomposition of Benzophenone Oxide (7) (in solution, according to Griller and co-
workers [11])
From these contradictory literature reports, it seems unlikely that the ozonide 4
would form via the widely accepted Criegee mechanism [12]. The first step of the
Criegee mechanism involves a 1,3-dipolar cycloaddition of ground-state O3 (a known
electrophile having a very modest dipole moment of 0.53 D) to the strained and
distorted olefin TPE (1) [13], which seems highly improbable. Furthermore, 1 is known
to fail the usual olefin test with Br2 (AE mechanism).
In view of these manifold uncertainties, we began a systematic reinvestigation of
ozonations of TPE (1). We considered it essential to a) generate pure samples of 3 and
4 with reproducible melting and decomposition temperatures to use as standards in the
analysis of mixtures of ozonide 4 and tetroxane 3, b) determine the typical behavior of
ground-state O3 from experimental results, c) examine solvent, concentration, and
temperature effects in TPE ozonations, and d) explore substituent effects for
tetrasubstituted ethylenes compared to TPE ozonolyses. It seemed also prudent to
optimize conditions for single-electron transfer (SET).
2. Results. ± 2.1. Preparations of Pure Tetroxane 3 and Ozonide 4. Previously
reported preparations of 3 included reaction of 2 with H2O2 [2a], reaction of benzo-
phenone dichloride with 30% aqueous H2O2 [14], ozonolysis of 1 and of other 1,1-di-
phenylalk-1-enes [2], and photooxygenation of diphenyl diazomethane [15]. Melting points
(with decomposition) of 159 ± 1628 [8c], 1708 [16], 200 ± 2208 [17], 205 ± 2068 [2d][18],
206 ± 2088 [2b], 206.5 ± 2158 [14], 212.58 [2a], 213.5 ± 2148 [15], 214 ± 2158 [19], 216 ±
2178 [2c], and 225 ± 2288 [2f] have been reported5). Typical properties of 3 have been
described by Marvel and Nichols [14]. Because the reaction of benzophenone dichloride
with 30% aqueous H2O2 was not always reliable [2c], we prepared only small amounts
of 3 without application of O3. It was, however, difficult to obtain a pure substance:
first, minor impurities in 2 could not be removed by repeated recrystallization, and
5
)
The highest [2f] and lowest [8c] melting-point ranges reported are suspect because of inconsistencies within
these citations and with the literature cited therein.

804
Helvetica Chimica Acta ± Vol. 83 (2000)
second, chromatographic separation of 3 from 2 (silica gel/CH2Cl2; room temperature)
led inevitably to some catalytic decomposition of 3. Cooling the chromatographic
column to ca. 158 prevented this undesired decomposition, and in this way pure 3 was
accessible (m.p. 2218, with decomposition, determined independently with a copper
block and automatically recorded with a Fus-O-mat (Heraeus, D-Darmstadt)).
Surprisingly, samples of 3 generated by ozonolyses of TPE (1) and purified as
described above did not have melting points in excess of 215 ± 2168. Although the
1H-NMR spectrum of a sample having a melting point of 215 ± 2168 was identical to
that of one melting at 2218, the melting point of the former could not be increased by
further purification. This discrepancy is probably due to the presence of a persistent
trace impurity, most likely of the corresponding trimer, as indicated by mass
spectroscopy (CI-MS, EE 120 eV, CH4, 2318, 594 (0.08%).
Pure ozonide 4 was prepared according to the method of Criegee and Korber from 1
in liquid 2 at 558 [4]; pure 4 had a melting point of 1538 (dec.). Although the formulae
of 3 and 4 differ by only one O-atom, their 1H-NMR spectra are completely different
(Fig. 1). In particular, it should be emphasized that 3 displays a 4 :4 :2 :6 :4 pattern of
aromatic-H multiplets at 400 MHz and a 4 :6 :6 :4 pattern at 100 MHz [20], whereas 4
displays only two 8 :12 aromatic-H multiplets. The lower-field 8 arom. H multiplet
(7.37 ± 7.44 ppm) is not superimposable on the spectrum of the tetroxane H-atoms, and
can be used to calculate mutual impurities of 3 in 4, or of 4 in 3. Cross contamination of
3 and 4 is likely responsible for the different melting temperatures reported in
literature.
2.2. The Electronic Structure of Ground-State O3. The early paper by Trambarulo et
al. [21] succeeded in establishing the view that ground-state O3 is an allyl anion-like
zwitterion (A, Fig. 2). Contemporary calculations suggest that O3 is a singlet 1,3-
diradical (B, Fig. 2) [22] or a heteroallene-like species (C, Fig. 2) with a hypervalent
central O-atom [23]. Whichever the best description is, O3 displays the typical
properties of a diradical in its reactions with other radicals [24 ± 26]. In the course of
such a radical combination, one O3 O-atom is transferred to the other radical, resulting
in mono-oxygenation (Scheme 3). It has also been reported that carbon radicals react
analogously with O3 [27], but no details were given.
To probe the diradical nature of O3, we reacted the commercially available stable
organic standard radical and radical scavenger, galvinoxyl, with ozonized O2. It was
shown earlier that galvinoxyl (2,6-bis(1,1-dimethylethyl)-4-{[3,5-bis(1,1-dimethyleth-
yl)-4-oxocyclohexa-2,5-dien-1-ylidene]methyl}phenoxy; 8) is only relatively stable
towards O2 [28]; a solution of 8 (1.0 g, 2.37 mmol) in CCl4 (500 ml) was reported to
Scheme 3. Typical O-Transfer Radical Reactions of O3 or O2 (reactions with nitroso and nitrone spin traps [26])

Helvetica Chimica Acta ± Vol. 83 (2000)
805

806
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 2. Electronic configurations of O3 discussed in the literature (A [21], B [22], and C [23])
react completely with O2 in 24 h [28c] (Scheme 4). A corresponding conversion with
ozonized O2 (a stream of O2 at 1 l/min containing 1.8 mmol O3/min) was complete after
ca. 4 min. Both processes resulted in formation of aldehyde 10, which was rather stable
towards O2 alone, but could be easily converted to the substituted benzoic acid 10a by
ozonized O2 (confirmed by an independent experiment with commercially available
10). Polymeric impurities in both reaction mixtures were easily removed by adsorption
on alumina. This result supports the proposal that O3 is a diradical species.
To judge the contribution of the zwitterionic O3 formula A (Fig. 2), the ꢀpush-pullꢁ
olefin 2-(methoxymethylidene)propanedinitrile (11) was treated with ozonized O2 in
CH2Cl2 solution at temperatures between 788 and 08 [29] (Scheme 5). That no
reaction occurred, in spite of a tenfold excess of O3, suggests that formula A does not
significantly contribute to the structure of O3.
2.3. Results of the Ozonations of TPE (1) in Aprotic Medium. Ozonolysis of 1 in
CFCl3 at 1058 yielded mainly 2, and 3 in only trace amounts. Earlier ozonolyses of 1 in
CCl4 (cf. Scheme 1) were carried out in suspensions, since ca. 15 mm of 1 in CCl4 is the
solubility limit at 08 (cf. Entry 2 in Table 1). Therefore, the superior solubility of 1 in
CH2Cl2 was exploited for investigations below 08 for concentration studies, while CCl4
was used for corresponding investigations above 08. It was striking that the formation of
ozonide 4 was observed in considerable amounts at higher reaction temperatures, and
that trace amounts persisting at 788 caused depressions of the melting points of
Scheme 4. Oxidative Cleavage of Spin-Trap Galvinoxyl (8) by O2 [28] or O3 /O2
Scheme 5. Attempted Ozonation of Enol Ether 11 (ꢀpush-pullꢁ olefin)

Helvetica Chimica Acta ± Vol. 83 (2000)
807
Table 1. Products from Ozonolysis of 1 in Aprotic Medium under Various Conditions
En-
try
Solvent and other
conditions
Concen-
tration
of 1 [mm]
Reaction
temp. [8]
Yields
Ketone 2
[%]
Tetroxane 3
[%]
Ozonide 4
[%]
Ester 12
[%]
a
1
2
3
CFCl3
CCl4
CCl4
14
14
14
105
0
22.5
72
70
57
28
27.5
36
±
2.5
7
)
±
a
)
4
5
6
7
8
9
10
11
12
13
14
15
16
CCl4
14
14
14
6
55
0
0
78
78
78
78
78
71.5
80
77
81
77
75.5
88.5
86
quant.
98
14
17
16
14
20
23.5
10.5
0.5
4
2.5
6
5
3
ClCH2CH2Cl
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2/FeCl3
a
)
)
)
)
a
a
14
a
113
15
1
b
a
)
)
±
±
±
11.5
14
±
±
±
CH2Cl2 inversec)
CH2Cl2 ‡ 2d)
CH2Cl2 ‡ 2e)
AcOEt ‡ 2f)
2
! 0c)
±
±
2
±
±
±
20
0 or 22.5
a
14
14
100
14
0
0
55
78
)
quant.
±
14.6
±
g
)
±
±
CH2Cl2/pyridine 1 : 2h)
quant.
a) Trace amounts. b) FeCl3 added as 1 ml of a 2% soln. in abs. CH2Cl2. c) O3 at 19.5 mmol/l in CH2Cl2 at
788; 1 added dropwise at 3 mmol/l. d) 20 g of 2 in 100 ml. e) 30 g of 2 in 150 ml. f) 40 g of 2 in
h
150 ml; g) Not determined, 2 as cosolvent.
) Formation of pyridine N-oxide: ca. quant. yield by H-NMR
1
detection, or ca. 56% from conversion to the picrate (m.p. 179.5 ± 1808).
isolated tetroxane 3 samples. Phenyl benzoate (12) was an additional impurity, formed
during all reactions performed in CH2Cl2, and in those performed in CCl4 above 208. Its
removal, together with 2, from samples of 3 and 4 was easily accomplished by
chromatographic purification at 158. No discrepancies were observed for reaction
performed in 1,2-dichloroethane.
The addition of a catalytic amount of anhydrous FeCl3, a SET catalyst, to a solution
of 1 in CH2Cl2 (Entry 10 in Table 1) led to some surprising observations. The
.
‡
characteristic purple color of radical cation 1 [30] occurred only when the solvent had
not been degassed (with Ar in an ultrasound bath) to remove O2. This intense purple
color likewise occurred in the presence of air after addition of strong protic acids (the
.
‡
intensity of the color with CF3SO3H > CH3SO3H > H2SO4) and its source, namely 1 ,
could be characterized by ESR spectroscopy (Fig. 3)6). Surprisingly, when samples of 1
were freshly prepared by reductive methods (from benzophenone dichloride and
copper [31a] or from 2 by McMurry reduction [31b]), the purple color changed rapidly
to dark green (lmax 700 nm, log e 0.3 [32]), and the compound became ESR-inactive
[33]. A similar green color was reported during oxidation of 1 with aminium radical
cation salts [34]. Moreover, the purple or green color disappeared just before a molar
amount of O3 was introduced. The reactions followed the same course after
consumption of equivalent amounts of O3 (Entry 10 in Table 1). The product profile
from inverse ozonolysis of 1 in CH2Cl2 (Entry 11 in Table 1) was nearly the same (cf.
Entry 10 in Table 1), however, the deep blue color of the O3 solution obscured the color
6
)
Performed by L. Eberson, Lund, Sweden.

808
Helvetica Chimica Acta ± Vol. 83 (2000)
.
‡
‡
Fig. 3. ESR Spectrum of 1 from 1 in the presence of HFP/H /O2
development during formation of the charge-transfer complex (cf. Table 2). Never-
theless, charge transfer between 1 and O3 in a matrix experiment has been described [35].
Ozonolyses of 1 in CH2Cl2 or AcOEt solution in the presence of 2 led exclusively to
formation of 2 (Entries 12 ± 14 in Table 1), which establishes that ozonide formation
from a solution of 1 in liquid 2 at 558 (Entry 15 in Table 1) was not simply the result of
the usual cross-ozonolysis. The ozonide yields for Entries 4 and 15 in Table 1 indicate a
common temperature effect (558).
Only at best was a trace amount of ozonide 4 formed during ozonolysis at
temperatures ꢀ08 (Entries 1, 5 ± 14, and 16 in Table 1). The yield of 4 from ozonolysis
in CCl4 as a function of temperature (Entries 2 ± 4 in Table 1) and the similarity in the
yield of 4 under different solvent conditions at 558 (Entries 4 and 15 in Table 1) points
to a dependence on temperature in the formation of this product.
It is generally accepted that ozonolysis of simple olefins in the presence of pyridine
does not lead to formation of pyridine N-oxide [36], however ozonolysis of strained
olefin 1 in CH2Cl2/pyridine (Entry 16 in Table 1) proved to be an exception in that, in
addition to 2, pyridine N-oxide was formed quantitatively. The pyridine N-oxide
product was identified both by 1H-NMR spectroscopy and by precipitation of its picrate
(56% yield, m.p. 179.5 ± 1808 [37]7) after evaporation of excess pyridine.
In summary, ozonolysis of TPE (1) in an aprotic medium is apparently dependent
on solvent, concentration of 1, reaction temperature, and on the presence of a SET
catalyst.
2.4. Application of 1,1,1,3,3,3-Hexafluoropropan-2-ol (HFP) as a Radical-Cation
Stabilizer. Eberson et al. [39] provided evidence that radical cations from electron-rich
arenes can be stabilized by solvation with HFP. To determine whether a similar
.
‡
stabilization would be observed with radical cation 1 [30], several ozonations of 1
were carried out at 08 in CH2Cl2 solutions containing various amounts of HFP. Whereas
7
)
A melting point of 1678 recently reported [38] actually corresponds to the picrate of pyridine.

Helvetica Chimica Acta ± Vol. 83 (2000)
809
ozonation in the absence of HFP (Table 1) gave initially colorless to yellow solutions
that ultimately decolorized, the presence of HFP in the ozonolysis solvent led to
permanently yellow (HFP/1 4 :1) to reddish brown (HFP/1 > 20 :1) solutions. This
visible change was dependent on a different composition of products. Inclusion of HFP
led to formation of the Wagner-Meerwein rearrangement product 14, for the first time
during ozonolyses of TPE. The amount of 14 formed was directly proportional to the
HFP content, while the corresponding yields of 2 and of phenyl benzoate (12) were
inversely proportional (Figs. 4 and 5).
Fig. 4. Ozonolysis of TPE (1) at 08 in CH2Cl2 (20 ml) as a function of HFP concentration
2.5. Oxirane Formation during Ozonolysis of 16 as a Substitute for 1. Unexpectedly,
tetraphenyloxirane (15), corresponding to strained olefin 1, was not found in product
mixtures resulting from ozonations carried out with a mixture of O3/O2, although the
related electron-transfer photo-oxygenation of 1 [40] generated a small amount (1.5%)
of 15, and a 3% yield of 15 from 1 was reported for an oxygenation carried out in a

810
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 5. Hitherto undescribed products from ozonolyses of TPE (1)
mixture of NO2/O2 [30b]. Hence, strained and distorted 1 was replaced by undistorted8)
(E)-1,2-di(tert-butyl)-1,2-diphenylethylene (16) [41], and its ozonation products were
studied for comparison [42] (Scheme 6). Unlike 1, 16 was rather unreactive towards O3,
and the C,C-cleavage product, (tert-butyl)phenyl ketone (17), was the main product.
Cyclic peroxides corresponding to 3 or 4 were not found, but the very unstable
polymeric peroxides 20, which had variable O contents (0.83 to 1.77 O per molecule 16
consumed), could be isolated from the solutions after ozonolysis. Each reaction
furnished oxirane 18 and tert-butyl benzoate (19), no traces of the isomeric phenyl
pivalate could be detected.
Scheme 6. Products from Ozonation of 16 under Various Conditions
2.6. Ozonolysis of TPE (1) in the Presence of MeOH. Repetition of ozonolyses of 1
Â
in the presence of absolute MeOH, according to Renard and Fliszar [6] and to
Robertson and Verzino [7], led to formation of a rather unstable hydroperoxide, which
decomposed to yield 2. Since this hydroperoxide did not display a MeO signal in its
1H-NMR spectrum after careful removal of MeOH at low temperature and low
pressure, the methoxy hydroperoxide 5 [6] must be ruled out. Furthermore, no
dismutation of 5 to yield 6 and benzophenone dimethyl acetal had occurred. Produc-
tion of benzophenone dimethyl acetal was not observed during ozonolysis or workup.
Ozonolysis of 1,1-diphenylethylene in a mixture of t-BuOH and 80% H2SO4, which
has been described as a countercurrent method for the efficient ozonization of olefins,
gave (tert-butoxy)diphenylmethyl hydroperoxide, related to the corresponding (me-
thoxy)diphenylmethyl hydroperoxide (5) [2b]. We were, however, despite many
attempts, unable to confirm this result.
2.7. Ozonolyses of Different Tetraaryl Ethylenes under Various Conditions. 2.7.1.
Tetrakis(4-methoxyphenyl)ethene (TME; 21). The results of ozonolyses of tetra 4-
methoxy-substituted 1, i.e., 21, either in CH2Cl2 or in CCl4 under various conditions
(Scheme 7), may be found in Table 2. Whereas a red-violet charge-transfer complex
gradually appeared under aprotic conditions (Entries 1 ± 5 in Table 2), the expected
.
‡
blue color of radical cation 21 [43] instantaneously appeared in the presence of either
an acid catalyst or a one-electron acceptor, when O3 was introduced (Entries 6 ± 11 in
Table 2). It is worth mentioning that no traces of ozonide 22 were found even under the
8
)
With respect to the CˆC bond only.

Helvetica Chimica Acta ± Vol. 83 (2000)
811
Scheme 7. Ozonolysis of TME (21) under Various Conditions (see also Table 2)
same conditions that worked for a TPE ozonolysis (cf. Entry 4 in Table 1 and Entry 12
in Table 2). In contrast to 1, 21 suffered ion-radical-induced polymerization.
2.7.2. 9-(Diphenylmethylidene)-9H-fluorene (27a) and 9-[Bis(4-methoxyphenyl)-
methylidene]-9H-fluorene (27b). These unsymmetric, undistorted8) olefins [44] were
ozonized in aprotic solvents in order to examine possible ozonide formation, or at least
regioselective peroxide formation. As in the case of distorted TME, 27a,b did not form
ozonides, tetroxane 3, or tetroxane 24. Only tetroxane 29, previously described as a
product of the ozonolysis of 9,9'-bifluorenylidene (27c) [45], was obtained in both
cases. Since ozonolysis of 27c was reported to give dispiro[9H-fluorene-9,3'-[1,2]diox-
etane-4',9''-9H-fluorene] [46] in pinacolone as solvent, we have examined that reaction
under various conditions. In aprotic solvents, 29 was isolated [45]. In contrast to the
report by Yang and Carr [46a], the highest yield of tetraoxane 29 was obtained via
ozonolysis of 27c in pinacolone (Scheme 8).
Table 2. Results of Ozonolyses (with a dry stream of O2/O3 at 1 l/min (ca. 1.8 mmol O3/min) until the reaction
was complete, followed by dry O2 for 15 min, and dry N2 for 1 h at r.t.) of TAE (21) under Various Conditions
Entry Solvent Cata- Concen-
Reaction Products and Yieldsa)
Color
Time for
appearance
of colorb)
[min]
lyst
tration
temp. [8]
Ketone Tetroxane Ester Polymer
23 [%] 24 [%] 25 [%] 26 [%]
of 21 [mm]
1
2
3
4
5
6
7
8
9
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CCl4
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CCl4
±
±
±
±
±
4.4
14
112.9
14
78
78
78
0
69c)
69
66
67
55
9
10
11
7
23
±
2
±
3.5
±
±
9
8
7.5
6
2
13
13
15.5
20
20
12
15
11
red-violet 2.5
red-violet
red-violet 15
3
red-violet
red-violet
blue
blue
blue
4
5
14
0
d
e
e
e
e
e
e
)
4.4
4.4
4.4
4.4
4.4
4.4
14
78
78
78
78
78
78
55
77
11
)
)
)
)
)
)
f
)
)
70.5
77.5
68.5
75
75
56
12.5
11.5
12.5
11.5
11
g
h
)
15.5
13.5
14
blue
blue
i
10
11
12
)
)
j
blue
k
±
10
5.5
28.5
)
±
a) Determined by 1H-NMR. b) After introduction of O3/O2. c) Yield determined as 2,4-dinitrophenyl hy-
drazone 65%. d) 5Â 10 3 mmol CF3SO3H. e) Color change instantaneous. f) 5Â 10 6 mmol CF3SO3H. g) 5 Â
10 3 mmol BF3 in Et2O. h) 5 Â 10 6 mmol BF3 in Et2O. i) 0.24 mol (4-BrC6H4)3NSbCl6. j) Gaseous
HCl. k) Pale yellow solution with formation of a colorless powder giving positive peroxide test; products
became resinous during attempts to isolate; insoluble in CDCl3.

812
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 8. Ozonation of 9-Alkylidene-9H-fluorenes 27 under Various Conditions
a) 27a in CH2Cl2, 608. b) 27b in CCl4, 08. c) 27b in CH2Cl2, 788. d) 27c in pinacolone, 508. e) 27c in CH2Cl2,
788. f) 27c in CCl4, 08. g) Inverse ozonolysis, addition of 27c in CH2Cl2, 788, to a solution of O3 in CH2Cl2,
788.
A marked difference was found when solutions of undistorted8) 27b and of
distorted8) 27c were ozonized in MeOH/CH2Cl2 1:1 at 08: 27b gave rise to bis(4-
methoxyphenyl) ketone (23) and the methoxy hydroperoxide 31 [8b] in nearly
quantitative yields, whereas 27c led to a mixture of ketones 28 and 33, with 31 present in
only trace amounts (Scheme 9).
An independent preparation of 31 by photo-oxygenation of 9-diazofluorene in the
presence of MeOH by the method of Sawaki et al. [8b] was possible in principle, but
irradiation caused 31 to decompose nearly as quickly as it was formed. Therefore, for
comparison, 31 was obtained by simple dioxygenation of 9-methoxy-9H-fluorene (32)
by air in the presence of Triton B (cf. Scheme 9); the known spiro ketone 33 [34] was
also independently generated.
2.7.3. 1,2-Bis(4-methoxyphenyl)acenaphthylene (36). Because it represents an
undistorted (but Baeyer-strained) tetraaryl olefin possessing a rigid C-skeleton, 36
was synthesized for ozonolyses. Its ozonolysis (as a simultaneous transfer of three
Scheme 9. Ozonation of 9-Alkylidene-9H-fluorenes 27b, c in the Presence of MeOH
‡
a) MeOH/CH2Cl2 1:1, 08; 99%. b) Triton B, pyridine, 108, followed by H ; 42%. c) MeOH/CH2Cl2 1 :1, 08; 28/
33 ca. 7.5 :1.

Helvetica Chimica Acta ± Vol. 83 (2000)
813
Scheme 10. Synthetic Access to Rigid 1,2-Bis(4-methoxyphenyl)acenaphthylene (36)
O-atoms) was compared with its epoxidation, followed by C,C cleavage and final
dioxygenation of the intermediate red 1,3-diyl or carbonylylide 44 (a stepwise transfer
of three O-atoms; cf. Scheme 12), as had previously been described for the
corresponding 1,2-diphenylacenaphthylene [47]. Synthesis of 36 was carried out in
two ways, starting either from 1,8-naphthoic anhydride (34) or from acenaphthene-1,2-
quinone (37) [48] (Scheme 10, Paths a and b, resp.).
Whereas ozonide 4 of TPE (1; an olefin with torsional strain) was formed in
acceptable yields only by heating (i.e., under abnormal ozonation conditions), ozonide
37 from strained 36 was formed as expected for olefins that are able to form ozonides at
low temperatures (i.e., under normal conditions). Moreover, ozonolysis of 36 under
inverse-ozonation conditions led to essentially the same result as a normal ozonolysis
(in contrast to ozonolysis of 1, cf. Table 1). Whereas 3 was a rather stable ozonide, 39
was very sensitive and was decomposed by heating, by irradiation with visible light, or
by acid catalysis to yield keto ester 40. The molecular structure of 40 was confirmed by
its hydrolysis to phenol 41 and 4-methoxybenzoic acid (42) (Scheme 11).
Stepwise O-transfer to 36 suffered the disadvantage that the ozonide 39 was
photolabile, but the oxirane 43 must be irradiated to cause C,C cleavage to generate the
intermediate 46, which is necessary for further photodioxygenation (Scheme 12).
Scheme 11. Formation and Reactions of 1,2-Bis(4-methoxyphenyl)acenaphthylene Ozonide (39)
a) CH2Cl2, 788. b) CCl4, 08. c) MeOH/CH2Cl2 2 :1, 608. d) CH2Cl2, inverse, 788. e) CCl4, 808, 48 h; 12%
(1H-NMR). f) Xylene, 1358, 6 d; 85%. g) MeOH, reflux (808), 24 h; 40%. h) CS2, r.t., 6 d; 23%. i) Toluene, 08,
hn (500-W daylight lamp), 7 h; 81%. j) CH2Cl2, 08, cat. amount of HSO3Cl, 1 h; 100%. k) Toluene, Ph3P, 808,
6 h; 72%. l) CH2Cl2, TFA, m-CPBA, r.t., 72 h; 74%. m) KOH/EtOH, reflux, 2 h, followed by 1m H2SO4.

814
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 12. Formation and Reactions of 1,2-Dihydro-1,2-bis(4-methoxyphenyl)acenaphthylene Epoxide (43)
a) KHSO5, acetone/CH2Cl2, buffer, 5 ± 108, 4 h; 82%. b) MeOH, reflux, 3 d. c) m-CPBA, NaHCO3, CH2Cl2, r.t.,
25 h. d) hn, l ˆ 254 nm.
Oxirane 43 suffered exclusively heterolytic C,O cleavage in boiling MeOH (Path b).
Olefin 36 could not be epoxidized with m-chloroperbenzoic acid (m-CPBA), because
the intermediate oxirane 43 was immediately cleaved at the C O and C C bonds to
yield a mixture of the Wagner-Meerwein-like rearrangement product 45 and diketone
35 (Path c). A selective C C cleavage was obtained only by irradiation (Path d).
3. Discussion. ± Pure samples of tetroxane 3 and ozonide 4 that show reproducible
melting points have been generated for the first time. Hitherto reported controversial
values likely reflect errors or contaminated samples. In addition, it has been established
that ozonolysis of 1 in the presence of MeOH does not yield a methoxyalkyl
hydroperoxide.
Experimental evidence indicates that ground-state O3 is an open-shell singlet
diradical instead of the generally accepted closed-shell zwitterion. Independent
calculations by Harcourt [22a] led to a similar conclusion, and pointed out short-
comings in earlier calculations, specifically that the contribution of a spin-paired
diradical Lewis structure had been largely ignored: ꢀUntil recently, any contribution to
resonance of the ꢀlong-bondꢁ (or spin-paired diradical) Lewis structure A has been
largely ignored. However, it has now been calculated to be a very important structureꢁ
and ꢀConsiderable attention is given to the use of Pauling 3-electron bonds and increased-
valence structures for providing qualitative VB descriptions of electronic structure. The
increased-valence structures for electron-rich molecules ± for example B and C ± are
equivalent to resonance between standard and long-bond Lewis structures, and usually
involve Pauling 3-electron bonds as diatomic componentꢁ.

Helvetica Chimica Acta ± Vol. 83 (2000)
815
The prevalence of the zwitterionic description for O3 is perplexing given that it was
excluded already 35 years ago by Gould and Linnett [49a], who concluded that the O3
molecule is best described in terms of two three-electron bonds, rather than as a
resonance hybrid of two valence bond structures each involving a single and double
bond. More recent calculations also describe O3 as a diradical species [49b][50]; O-
atoms with two cumulative multiple bonds have been called ꢀhypervalentꢁ [23].
Harding and Goddard [51] concluded 20 years ago that experiments describing the
electrophilic nature of O3 were misinterpreted on the basis of the assumed non-
electrophilic behavior of 1,3-biradicals. However, the electrophilic nature of O-
centered radical species is well-established and is consistent with the experimentally
established electrophilic character of ground-state O3; the presence of a positively
charged O-atom is not required. There can be no electron delocalization in the 4p
system of O3, which possesses three parallel p-orbital axes and 1,3-diradical character
with a hypervalent central O-atom having two approximately orthogonal p-systems,
because these features require different molecular geometries. Important experimen-
tally established properties of ground state O3 include: 1) O3 is a peroxy radical; 2) O3
as a strong oxidant is an efficient one-electron acceptor; 3) aside from O2, O3 is the
most effective mono-oxygenating agent known; and 4) although organic hydrotrioxides
or trioxides (including the so-called ꢀprimaryꢁ ozonides) have been described, they
cannot result from concerted reactions of 1,3-diradical ozone. Ozonation reactions must
involve at least two-steps, to permit energy evolution from nascent O H or O
C
bonds to the solvent and to avoid simultaneous cleavage of week O O bonds. Since
formation of trioxides was not observed in the present investigation, this property will
not be further discussed (see the mechanistic investigations of Pryor et al. [52]).
The reaction mechanism for alkene ozonolysis proposed by Criegee is generally
accepted, but many exceptions to the accepted mechanism of this fundamental reaction
have been described. Recently, 1,3-dipolar cycloaddition of O3 to alkenes (the first step
of Criegeeꢁs mechanism) has been rejected, on the basis of experimental data, in favor
of an initial single-electron-transfer (SET) step to generate alkene-cation-radical/
ozone-anion-radical pairs [53], in agreement with our observations. But the further
conclusions of these authors follow classical ideas and are inconsistent with our
observations. 1) O3 is a metastable compound, and its decay requires an activation
energy of ca. 105 kJ/mol (25 kcal/mol) to yield atomic and molecular oxygen. However,
its anion undergoes very rapid cleavage [54], thus successful alkene ozonolyses should
9
be initiated by alkene-catalyzed decomposition of O3 ). The progress of the reaction
depends on the structure of the alkene and on the reaction conditions (cf. Table 1). 2)
Most alkenes should undergo a reversible SET to O3 as a first weakly endothermic step.
However, ozonate is a reactive oxygen radical (anion) and should oxygenate alkene
cation radicals very quickly and exothermally via mono-oxygen exchange (Scheme 13).
Depending on the character of the alkene cation radicals, quasi-epoxidations or open-
chain oxygenations can occur. However, isomeric reactive intermediates can undergo
cleavage of the s-bond via either exergonic isomerization [55] or b-cleavage of the
9
)
Cf. a) Y. A. Maletin, R. C. Cannon, Theor. Exper. Chem. 1998, 34, 57; b) P. Maslak, Top. Curr. Chem. 1999,
168, 1; c) P. Maslak, J. N. Narvaez, Angew. Chem. 1990, 102, 302; Angew. Chem., Int. Ed. 1990, 29, 283.
(added in proof).

816
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 13. Fate of the p-Bond during Alkene Ozonolysis
corresponding oxyl radical to generate a carbonyl compound and a carbene cation
radical. The former situation could have occurred with the rigid cyclic alkene 36, and
represents an oxygenative ring enlargement of a strained five-membered ring to a less
strained six-membered ring. Our earlier communication on cyclopentene ring
ozonolyses [56] already challenged the Criegee mechanism. In the case of electron
exchange between unsaturated hydrocarbons and O3, it is not necessary that the initial
SET be exergonic; exergonic SET is required only when stationary concentrations of
free cation radicals are needed for spectral observations. If alkene cation radicals are
needed only to lead to subsequent exergonic oxygenations (cf. formation of 1,2-
dioxetanes via Nelsenꢁs CRCC oxygenation of tetraalkyl ethylene cation radicals by O2
[57]), then they usually cannot be established for oxygenations. Nevertheless, they are
required for oxygenations. Related cases have recently been discussed in the
comprehensive review by Schmittel and Burghart [58a].
The second type of intermediate should occur with distorted alkenes like TPE (1)
and its derivatives, where cation radicals must be a good deal more distorted than the
.
‡
neutral radicals. Mono-oxygenation of 1 should lead to cation radical 47, which has
been proposed as an intermediate during electron-transfer-catalyzed oxirane C O ring
cleavage of 15 induced by one-electron acceptors [59a]. Strong solvation of 47 during
ozonolysis of 1 with HFP [39] favors an intramolecular Wagner-Meerwein-like
rearrangement to yield benzpinacone (14; cf. Fig. 4). The corresponding rearrange-
ment has been observed in the mass spectrum of 15 [59b], whereas thermal cleavage at
558 (cf. Table 1) should lead to 2 and the diphenyl carbene cation radical 48. Such
reactive intermediates have been studied in the gas phase [55l] as well as in solution
[60] (Scheme 14), but these would be too reactive to be observed during ozonation.
Bifunctional intermediates like 48 have been shown in kinetic studies to react as well
with O2 as with acetone (and other nucleophiles) at comparable rates [60]. These data
exclude a direct proof of 48 in the present case. Therefore, unprecedented ozonide
formation by thermal activation, as well as the formation of geminal bis(hydroper-
oxide) 6 during ozonolysis of 1 with an O3/O2 mixture in the presence of MeOH (cf.
Scheme 1), follow from the assumed intermediacy of 48 via 51. Auto-reaction of 48
after neutralization in the cold should react via 50 to give the dimer 3, with the trimer as
a trace impurity (Scheme 15).
It has been found that ozonolysis of 9-[bis(4-methoxyphenyl)]methylidene-9H-
fluorene (27b) in the presence of MeOH leads regioselectively to the methoxy
hydroperoxide 31, contradictory to expectation based on the Criegee mechanism.
Ozonolysis favors hydroperoxide formation on the olefinic C-atom of the CˆC double
bond (i.e., the carbonyl oxide C-atom), offering better stabilization of an intermediate

Helvetica Chimica Acta ± Vol. 83 (2000)
817
Scheme 14. C O (Path a) and C C (Path b) Cleavage of Tetraphenyloxirane Radical-Cation Intermediates
a) O3, CH2Cl2/HFP (cf. Fig. 4). b) O2, CH2Cl2, catalysis by Ph3CSbCl6 or by (4-Br C6H4)3NSbCl6 [59a]. c) O3,
CCl4, 558 (cf. Table 1). d) O2, MeCN, hn, anthracene-9,10-dicarbonitrile, 1,1'-biphenyl, MeCN [5b].
Scheme 15. Possible Mechanism of Atypical Ozonolyses of 1 Involving Carbene Radical-Cation Intermediate
48. Path a: neutralization via nucleophile-electrophile interaction; Path b: dioxygenation via C O radical
combination.
positive charge. On the other hand, ozonation of bisfluorenylidene (27c) in the
presence of MeOH does not form the methoxy hydroperoxide 31 (as expected), but
rather 28 and the Wagner-Meerwein-like rearrangement product 33 (cf. Scheme 9).
Formation of the corresponding rearrangement product 14 (via ozonation of 1)
required the presence of HFP (cf. Fig. 4). The divergent behavior of 27b and 27c, each
containing 9H-fluoren-9-ylidene moieties, is ascribed to the planarity of the double-
bond skeleton [44] of 27b in contrast to the strongly distorted double bond (cf. its dark
red color) of 27c [13]. Consequently, only 27b is expected to follow the proposed
epoxidation path shown in Scheme 13, and only 27b should form the methoxy
hydroperoxide 31 during ozonolysis in the presence of MeOH through the inter-
mediates 52 ± 57 (Scheme 16); 27c should follow the proposed open-chain O-exchange
path leading ultimately to the Wagner-Meerwein-like rearrangement product 33
(similar to Scheme 14).
Nucleophilic substitutions at carboxonium ions (see the last step of Scheme 16) are
well-known [61]. In the course of these reactions, carbonyl compounds (e.g., 23) serve
as leaving groups.

818
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 16. Proposed Mechanism for Formation of 31 by Ozonolysis of 27b in the Presence of MeOH
The mechanism of oxirane formation during alkene ozonations (cf. partial-cleavage
products in [12b]) is uncertain. Ozone itself is proposed as the epoxidizing agent, or
these are O3-alkene reaction products, particularly dioxiranes; during ozonolysis of 2,3-
dimethylbut-2-ene, formation of dimethyl dioxirane (DMDO) has been proposed as an
intermediate leading to epoxide formation [62]. In the present case, an attempt to
epoxidize distorted 1 with DMDO failed to yield tetraphenyloxirane (15). Similar
behavior for diphenyldioxirane (13) could explain the failure to form 15 during
ozonation of 1, although the presence of the isomer of 13, phenyl benzoate (12) in some
experiments argues for formation of 13 as an intermediate [10]. On the other hand, in
the case of ꢀundistortedꢁ8) alkene 16, formation of epoxide 18 (cf. Scheme 6) furnished
experimental proof for epoxide formation during alkene ozonation, which failed in the
case of 1.
The formation of pyridine N-oxide during ozonation of 1 in the presence of pyridine
(cf. Entry 16 in Table 1) conflicts with the usual results of the so-called ꢀpyridine effectꢁ
during alkene ozonolyses [63]. Whereas it was confirmed in independent experiments
that ozonolysis of several other alkenes leads neither to formation of pyridine N-oxide
from added pyridine nor to deoxygenation of added pyridine N-oxide during
ozonations in CH2Cl2, distorted 1 behaved differently. In accord with the reactivity
pattern of radical cations [58a], the mechanism proposed in Scheme 17 involves initial
nucleophilic attack of pyridine on the radical cation 47 during ozonolysis of 1.
Scheme 17. Proposed Formation of Pyridine N-Oxide via Addition of Pyridine to 47 (cf. Scheme 14) during
Ozonolysis of 1 in the Presence of Pyridine

Helvetica Chimica Acta ± Vol. 83 (2000)
819
Subsequent reaction steps could include radical b-cleavage of 58, formation of 13, and
finally formation of pyridine N-oxide [64].
Experimental Part
1. General. O3 was generated from a dry stream of O2 at a flow rate of 1 l/min in a Sander 301.7 laboratory
ozonizer at 0.6 A, which affords O3 at a rate of ca. 1.8 mmol/min as determined by iodometric titrations, For
ꢀnormalꢁ ozonolyses, the O3/O2 stream was bubbled through the alkene soln.; for ꢀinverseꢁ ozonolyses, abs.
CH2Cl2 (100 ml) was first saturated with O3 at 788 by introduction of the O3/O2 mixture for 10 min before
addition of the alkene. The O3 content (ca. 5 mmol O3) of the sat. soln. was determined by titration with 0.1m
Ph3P in abs. CH2Cl2 at 788, until the dark blue coloration disappeared. TLC Foils Alugram SILG/UV254
(Macherey-Nagel, Germany) were used to continuously monitor the progress of the conversions. Column
chromatography (CC) separations were carried out on silica gel (70 ± 325 mesh ASTM) or on neutral alumina
from Macherey-Nagel with a variety of columns and solvents. M.p.: a calibrated Kofler hot stage for
instantaneous measurements that minimize depression of the m.p. caused by formation of decomposition
products during the measurement (uncorrected); a Fus-O-mat (Heraeus, Germany) apparatus for fast
measurements, and a classical copper block for slow measurements. UV/VIS Spectra: Kontron Uvikon 860
spectrometer; lmax in nm. IR Spectra: Beckman IR-33 spectrometer; in cm 1. NMR Spectra: Bruker AM 400
1
spectrometer (in CDCl3; H: 400 MHz and 13C: 100 MHz); d in ppm rel. to internal Me4Si (unless otherwise
specified), J in Hz. CI-MS: CH4, 120 eV, 2208; in m/z (rel %).
Note: All manipulations of O3 and peroxide reaction products were carried out in an open-air laboratory
behind a safety shield, and excess O3 was destroyed with KI in dil. aq. AcOH, to minimize risk of violent
decompositions or explosions. For reactions with less reactive olefins, which leave larger amounts of unreacted
O3, we recommend using solns. of sodium thiosulfate, a catalytic amount of KI, and a few drops of AcOH in H2O
to reduce residual O3.
2. Preparation of Tetrasubstituted Ethenes. 1,1,2,2-Tetraphenylethene (1) was obtained from Merck and was
used without further purification. Alternatively, 1 could be generated and purified starting from benzophenone
dichloride and Cu powder by the reductive procedure of Buckles and Matlack [31a], or from benzophenone (2)
by the McMurry procedure according to Mukaiyama et al. [31b]. Samples of 1 prepared by these methods were
spectroscopically identical to commercially available 1, and were suitable for ozonolytic cleavage. However,
.
.
‡
‡
they were unsuitable for ESR spectroscopy of 1 . The purple radical cation 1 could be easily prepared for
ESR spectroscopy by air oxidation of commercially available 1 in the presence of FeCl3 or strong protic acids
(e.g., conc. H2SO4, MsOH, or, preferably, triflic acid; cf. Fig. 3). Traces of transition-metal ions that survived
simple recrystallizations caused a change in soln. color from ESR-active purple to ESR-inactive green (lmax
700 nm, log e ˆ 0.3). The rate of this color change varied.
ˆ
(E)-1,2-Di(tert-butyl)-1,2-diphenylethene (16) was prepared according to Leimner and Weyerstahl
[41a][42].
1,1,2,2-Tetrakis(4-methoxyphenyl)ethene (TME; 21) was prepared according to the method of Buckles and
Womer [65]. However, 21 readily formed stable solvates with various solvents. Thus, recrystallization from dry
toluene furnished a 1:1 solvate which melted and resolidified at 1748, but after further heating displayed the lit.
m.p. of 184 ± 1858.
9-(Diphenylmethylidene)-9H-fluorene (27a) [66] and 9-[bis(4-methoxyphenyl)methylidene]-9H-fluorene
(27b) [44] were prepared as described.
9,9'-Bifluorenylidene (27c). 9-Bromo-9H-fluorene [67] (4.9 g, 20 mmol) was dissolved in t-BuOH (80 ml) at
808, t-BuOK (6.7 g, 60 mmol) was added in portions with stirring, and the mixture was refluxed for 0.5 h. After
cooling, the mixture was added to stirred cold 2n H2SO4 (150 ml). The red precipitate was filtered by suction,
washed with H2O (50 ml), and dried (P2O5). Recrystallization from EtOH/CCl4 1:10 furnished 27c (4.8 g,
75%). Red crystals. M.p. 1868 ([68]: 185 ± 1878).
1,2-Bis(4-methoxyphenyl)acenaphthylene (36). Path a: Mg (1.2 g, 50 mmol) was covered with dry Et2O
(10 ml), and 1-bromo-4-methoxybenzene (9.4 g, 50 mmol) in dry Et2O (50 ml) was added dropwise. The
mixture was heated at reflux for 1 h until the metal was dissolved (after 50 min, a small amount of 1-bromo-4-
methoxybenzene was added to complete the reaction). Powdered 1,8-naphthoic anhydride (34; 2 g, 10 mmol)
was gradually added to the cooled soln. (internal temp. 20 ± 258), dry benzene (20 ml) was added for better
dissolution, the mixture was stirred for 0.5 h at r.t., then was heated at reflux for 12 h. After cooling, the mixture
was added to a sat. aq. NH4Cl soln.; the product was extracted with CH2Cl2 (50 ml), dried (Na2SO4), the solvent

820
Helvetica Chimica Acta ± Vol. 83 (2000)
was removed in vacuo, and the oily residue partially crystallized at 158. Recrystallization from acetone
furnished 1,8-bis(4-methoxybenzoyl)naphthalene (35) (0.5 g, 13%). Colorless crystals. M.p. 2188 (215 ± 2168
[48]).
Path b: as in Path a, but with acenaphthylene-1,2-quinone (37) to furnish a mixture of stereoisomers of 1,2-
bis(4-methoxyphenyl)-1,2-dihydroacenaphthylene-1,2-diol (38) in 34% yield [48]. A McMurry reduction
mixture was prepared by cautious dropwise addition of abs. THF (40 ml) to TiCl4 (3.0 g, 15.8 mmol) in a
carefully dried 3-neck flask. Dry Zn dust was added in small portions and the black mixture was heated at reflux
for 1 h. A soln. of either 35 (2.0 g, 5 mmol) or 38 (2.0 g, 5 mmol) in warm abs. THF (120 ml or 40 ml, resp.) was
added dropwise with stirring, and the mixture was refluxed for 14 ± 18 h (with monitoring by TLC (silica gel;
CH2Cl2/petroleum ether 1:1)) until 35 or 38 was consumed. The cooled mixture was added to Et2O (200 ml),
followed by cautious addition of ice water (100 ml). After the usual workup of the Et2O layer and
recrystallization from EtOH/acetone, 36 (78 ± 82%) was obtained. Red crystals. M.p. 1228. IR (KBr): 1610
(CˆC). 1H-NMR (CDCl3) 7.83 (d, J ˆ 8.1, 2 H); 7.7 (d, J ˆ 6.9, 2 H); 7.57 (t, J ˆ 7, 2 arom. H); 7.38
(m, 4 arom. H); 6.91 (m, 4 arom. H); 3.84 (s, 2 MeO). 13C-NMR: 158.8 (arom. C OMe); 140.4; 137; 131.2;
128.4; 127.9; 127.8; 127; 123.5; 113.9 (arom. C); 55.2 (MeO). Anal. calc. for C26H20O2 (364.4): C 85.69, H 5.53;
found: C 86.03, H 5.48.
Note: This procedure was generally carried out with amounts of reactant ranging from 1.5 ± 200 mmol,
which afforded the indicated yields, providing that larger batches (>200 ml solvent) were concentrated to 1ꢂ4 of
the original volume before workup.
3. Preparation of Pure 3,3,6,6-Tetraphenyl-1,2,4,5-tetroxane (3) and Pure 3,3,5,5-Tetraphenyl-1,2,4-
trioxolane (4). Crude samples of 3, obtained via the ozonolytic [2d] or non-ozonolytic [14] procedure, and 4
[4] were prepared and recrystallized as described. Since trace impurities of ketone 2 and occasionally of ester 12
could not be quantitatively removed by recrystallization, solns. of 3 or 4 were chromatographed on a water-
cooled (<158) silica gel column. In either case, 3 or 4 was the first analyte to elute. Pure samples of 3 obtained
non-ozonolytically: m.p. 2218 (dec.), 3 obtained ozonolytically: m.p. 215 ± 2168 (dec.). Pure samples of 4: m.p.
1538 (dec.). Mixtures of 3 and 4 obtained by ozonolysis of 1 in CH2Cl2 or in CCl4 could not be separated
chromatographically; such samples of tetroxane 3: m.p. 205 ± 2068 (dec.).
4. Experiments to Differentiate between Singlet Diradical Character or Zwitterionic Character for Ground-
State O3. 4.1. Oxygenation of Galvinoxyl by O2. According to the procedure of Greene and Adam [28c], pure
galvinoxyl (ˆ2,6-bis(1,1-dimethylethyl)-4-{[3,5-bis(1,1-dimethylethyl)-4-oxocyclohexa-2,5-dien-1-ylidene]me-
thyl}phenoxy; 8; ACROS Chimica; 1.0 g, 2.37 mmol) was dissolved in dry CCl4 (500 ml) and bubbled with a
stream of dry O2 at r.t. Decolorization was observed after 16 h, or after 6.5 h with half as much 8 dissolved in dry
1,2-dichloroethane, a slightly better solvent for dioxygen. The solvent was removed by rotatory evaporation at
r.t., and a sample of the resulting viscous orange residue was dissolved in CDCl3 for 1H-NMR analysis: no
peroxidic products were found.
4.2. Oxygenation of Galvinoxyl by O3 /O2. A corresponding soln. of galvinoxyl (1.0 g, 2.37 mmol) was
prepared in dry CCl4 (500 ml) and bubbled with a stream of dry O3/O2 at r.t.; the O3 content was ca. 1.8 mmol/
min at a gas flow rate of 1 l/min. The soln. was decolorized within 4 min; galvinoxyl consumed ca. 3 equiv. of O3.
1H-NMR Monitoring showed formation of the corresponding principal products 2,6-di(tert-butyl)cyclohexa-2,5-
diene-1,4-dione (9), 3,5-di(tert-butyl)-4-hydroxybenzaldehyde (10), 3,5-di(tert-butyl)-4-hydroxybenzoic acid
(10a), and a resinous polymer. The products 9 [69], 10 [70], and 10a [71] were identified by comparison with
samples synthesized independently by published methods (9 and 10a) or obtained commercially (10).
4.3. 2-(Methoxymethylidene)propanedinitrile (11), prepared according to the procedure of Jones [72] and
dissolved in dry CH2Cl2, did not react with even a 10-fold excess of O3 at temps. ranging from 788 to 08 [29].
5. Preparation of Putative Ozonolysis Products by Other Procedures. Phenyl benzoate (12) [73a] and tert-
butyl benzoate (19) [73b,c] were prepared by a general procedure from a soln. of PhOH or t-BuOH, resp., in
pyridine to which was added an equimolar amount of PhCOCl: 12, (m.p. 708); 19, (b.p.10 94 ± 958)10).
Pyridine N-oxide picrate [37], benzpinacone (ˆ phenyl triphenylmethyl ketone, 14) [74], tetraphenyloxirane
(15) [75], 1,1-dimethyl-3-phenylpropan-2-one (17) [76], 3,3,6,6-tetrakis(4-methoxyphenyl)tetraoxane (24) [77],
4-methoxyphenyl 4-methoxybenzoate (anisyl anisoate, 25) [78], 9,10-dihydrospiro[9H-fluorene-9,9'-phenan-
threne]-10-one (33) [79], 1,8-Bis(4-methoxybenzoyl)naphthalene (35), and 1,2-Bis(4-methoxyphenyl)-1,2-
dihydroacenaphthylene-1,2-diol (38) [48] were prepared by published methods.
10
)
A b.p.2 of 968 reported in [73b] was found to be erroneous, and the b.p.10 found here is consistent with that
obtained via a different synthetic procedure.

Helvetica Chimica Acta ± Vol. 83 (2000)
821
(E)-2,3-Di(tert-butyl)-2,3-diphenyloxirane (18) [42]. A soln. of commercially available m-CPBA (2.75 g,
11 mmol, peroxide content of 70 ± 75%) in dry CHCl3 (50 ml) was added dropwise to a soln. of 16 [41a] (2.92 g,
10 mmol) in dry CHCl3 at r.t. over 5 min. The mixture was stirred (20 h), an additional portion of m-CPBA
(0.8 g, 3.3 mmol) added, and the mixture was stirred for another 3 h until 16 was consumed (TLC monitoring
(silica gel; CH2Cl2/petroleum ether 1:1)). The mixture was diluted with Et2O (100 ml), extracted 3 Â with 10%
aq. Na2CO3 (200 ml portions), dried (MgSO4), the solvent removed, and the resulting residue recrystallized
from MeOH to afford 18 (1.74 g, 56%). White crystals. M.p. 156 ± 1578. IR (KBr): 3090, 3065, 2980, 2940, 2910,
2880, 1955, 1605, 1495, 1485, 1450, 1395, 1365, 1265, 1245, 1220, 1200, 1080, 1030, 950, 905, 885, 780, 730, 710.
1H-NMR (CDCl3): 7.45 (m, 4 arom. H); 7.29 (m, 6 arom. H); 0.57 (s, 2 t-Bu). 13C-NMR (CDCl3): 139.8; 130.9;
128.6; 127.1; 126.8; 126.0 (arom. C); 75.4 (oxiranyl C); 36.8 (Me3C); 28.6 (Me3C). Anal. calc. for C22H28O
(308.5): C 85.66, H 9.15, found: C 85.51, H 9.12.
Dispiro[9H-fluorene-9,3'-(1,2,4,5-tetroxane)-6',9''-[9H]fluorene] (29). Exhaustive ozonolyses of 27c,
followed by recrystallization from dioxane, furnished 29. M.p. (dec.) 2148 [45]: (201 ± 2038). We also prepared
29 by ozonolysis of 9-(methoxymethylidene)-9H-fluorene with recrystallization from dioxane: m.p. (dec.) 2168
‡
‡
‡
‡
[29]. CI-MS: 393 (18.7, [M ‡ 1] ), 392 (26, M ), 377 (68, [M ‡ 1 O] ), 361 (87, [M ‡ 1 O2] ), 196 (7.6, [M/
‡
‡
‡
2] ), 180 (100, [C13H8O] ), 152 (83, [C12H8] ).
Benzo[c]benzopyran-2-one (30). Since the literature procedures [80] gave mixtures in our hands, we
prepared 30 from fluorenone (28) by Baeyer-Villiger oxidation. A soln. of (CF3CO)2O (2.1 ml, 3.15 mmol) in
dry CH2Cl2 (5 ml) was added dropwise to a suspension of 28 (1.8 g, 10 mmol) and H2O2/urea 1:1 adduct (1.4 g,
15 mmol) in dry CH2Cl2 (50 ml). The mixture was stirred at r.t. for 48 h, a second portion of H2O2/urea adduct
was added, and stirring was continued at r.t. for a further 72 h. The mixture was then filtered, the org. phase
extracted with H2O (50 ml), and dried (MgSO4). After removal of solvent, the resulting yellow oil, which
crystallized at r.t., was heated with aq. 2n NaOH (10 ml) at 808 for 10 min, filtered, and the cooled filtrate
extracted with Et2O (20 ml) to remove unreacted yellow fluorenone 28. The aq. phase was acidified with 2n HCl
and extracted 2 Â with Et2O. The usual workup furnished beige-colored crude 30 (1.5 g), which was
recrystallized from petroleum ether/dry EtOH (1.1 g, 56%). Colorless crystals. M.p. 928 ([80]: 94.58, [80b]: 938).
9-Methoxy-9H-fluoren-9-yl Hydroperoxide (31). A 40% soln. of Triton B [81] in pyridine (0.05 ml) was
diluted with pyridine (1.25 ml) and saturated with O2 at 108, and a soln. of 9-methoxy-9H-fluorene (32)11
)
(1.0 g, 5.1 mmol) in dry pyridine (20 ml) was added dropwise over 95 min (with continuous bubbling of O2
through the mixture). After warming to r.t., the mixture was buffered with one drop of glacial AcOH, diluted
with ice water (150 ml), and the mixture repeatedly extracted with Et2O. The combined org. layer was washed
with 5% aq. HCl, H2O, aq. sat. NaHCO3, again with H2O, and dried (Na2SO4). The ether was removed in vacuo
without heating to yield 31. Yellow oil; the titrated peroxide content was 27%; 31 was contaminated with
unreacted 32 (43%) and fluorenone 28 (29%, as determined by 1H-NMR analysis). All attempts to purify 32 led
to formation of 28.
8-(4-Methoxybenzoyl)naphthylen-1-yl 4-Methoxybenzoate (40). A soln. of m-CPBA (70 ± 75%; 3.7 g,
corresponding to 2.6g (15 mmol peracid)) in CH2Cl2 (30 ml) was dried (MgSO4) and added to a soln. of 35
(2.3 g, 5.8 mmol) in dry CH2Cl2 (30 ml). The mixture was cooled to 08 and protected from light with Al foil. A
soln. of TFA (0.44 ml, 5.8 mmol) in dry CH2Cl2 (2 ml) was added dropwise, the mixture was stirred for 48 h, a
second portion of m-CPBA (3.7 g) in dry CH2Cl2 (30 ml) was added, and stirring was continued for another
48 h. This procedure was repeated for another 48 h, until 35 was completely consumed (TLC monitoring, silica
gel; CH2Cl2). Subsequently, the soln. was filtered and the liquid phase washed with 10% aq. Na2SO3, sat. aq.
NaHCO3, 2 Â with H2O, and dried (MgSO4). After removal of the solvent in vacuo, the resulting brownish oil
crystallized on stirring with a mixture of Et2O (3 ml) and acetone (10 ml). Recrystallization (AcOEt), and
additional chromatographic purification of the mother liquor afforded 40 (1.76 g, 74%). M.p. 134 ± 1368. IR
(KBr): 1735 (OCˆO), 1670 (CˆO). 1H-NMR (CDCl3): 8.0 (d, J ˆ 8.3, 1 H); 7.85 (d, J ˆ 8, 1 H); 7.72 (m, 2 H);
7.55 (m, 2 H); 7.39 (m, 3 H); 7.29 (m, 1 H); 6.83 (m, 2 H); 6.6 (d, JAB ˆ 9.1, 2 H); 3.89 (s, MeO); 3.77 (s, MeO).
13C-NMR (CDCl3): 196.3 (CˆO); 164.3 (OCˆO); 163.9; 163.6 (arom. CO); 146.0; 135.7; 135.3; 132.5; 132.2;
130.1; 129.4; 126.2; 125.7; 125.4; 124.7; 121.5; 120.6; 113.4; 113.1 (arom. C); 55.5; 55.4 (MeO). Anal. calc. for
C26H20O5 (412.5): C 75.72, H4.89; found: C 75.70, H 4.74.
8-(4-Methoxybenzoyl)naphthalen-1-ol (41). A mixture of 40 (0.7 g, 1.7 mmol) and KOH (0.2 g, 3.4 mmol)
in EtOH (25 ml) was heated under reflux for 2 h. The volume was reduced by evaporation to ca. 10 ml. The
11
)
Attempts to synthesize 31 via the Li salt were unsuccessful and led to oxidative dimerization; 9-methoxy-
9H-fluorene is known [82a] and was prepared according to [82b].

822
Helvetica Chimica Acta ± Vol. 83 (2000)
mixture was added to ice-water (80 ml) and acidified with 1m H2SO4. The aq. phase was extracted 3 Â with
AcOEt (30 ml), the resulting org. phase was washed with H2O, dried (MgSO4), and the solvent was evaporated
in vacuo. The resulting yellowish crystalline mixture of 41 and 4-methoxybenzoic acid (42) was separated by CC
(silica gel, AcOEt). The first fraction yielded 41 (0.4 g, 89%), which was recrystallized from AcOEt. Colorless
crystals. M.p. 148 ± 1508. IR (KBr): 3150 (OH), 1655 (CˆO). 1H-NMR (CDCl3): 8.05 (dd, 1J ˆ 8, 2J ˆ 1.3, 1 H);
7.85 (d, J ˆ 8.3, 2 H); 7.53 (m, 2 H); 7.44 (m, 2 H); 7.05 (d, J ˆ 8.4, 1 H); 6.96 (d, JAB ˆ 8.3, 2 H); 3.88 (s, MeO);
1.6 (br. s, OH). 13C-NMR (CDCl3): 200.7 (CˆO); 163.8 (arom. C OMe); 152.5 (arom. C OH); 135.7; 134.9;
133.1; 131.9; 131.4; 129.1; 127.3; 124.2; 122.2; 121.3; 113.6 (arom. C); 55.6 (MeO). Anal. calc. for C18H14O3
(278.3): C 77.68, H 5.07; found: C 77.80, H 5.19.
4-Methoxybenzoic acid (42) (0.12 g, 46%), obtained from the second fraction, was identical to an authentic
sample.
1,2-Bis(4-methoxyphenyl)-1,2-epoxy-1,2-dihydroacenaphthylene (43). A mixture of KHSO5 (caroate; 10 g,
corresponding to 30 mmol KHSO5) and NaHCO3 (13 g, 154 mmol) in distilled H2O (100 ml) and 0.002m aq.
Na2(EDTA) (20 ml) was vigorously stirred at 58 while a soln. of 36 (4.2 g, 13.8 mmol) and 18-crown-6 (0.2 g,
0.07 mmol) in CH2Cl2 (80 ml) and acetone (15 ml) was added dropwise. After stirring for 4 h at 5 ± 108, 36 was
consumed (TLC monitoring) and the mixture was poured into ice-cold H2O. Separation of the org. phase and
usual workup afforded a yellow oil that crystallized from Et2O/petroleum ether on cooling: 43 (1.56 g, 82%).
Colorless crystals. M.p. 138 ± 1398. UV (CH2Cl2): 285 (10230). 1H-NMR (CDCl3): 7.85 (d, J ˆ 8.2, 2 H); 7.53
(t, J ˆ 8.1, 2 H); 7.46 (d, J ˆ 7.2, 2 H); 7.39 (d, JAB ˆ 8.2, 4 H); 6.85 (d, JAB ˆ 8.6, 4 H); 3.77 (s, 2 MeO). 13C-NMR
(CDCl3): 159.3; 141.7; 138.1; 132.5; 129.2; 127.3; 126.2; 124.7; 123.0; 113.5 (arom. C); 76.0 (C
(MeO). Anal. calc. for C26H20O3 (380.5): C 82.08, H 5.3; found: C 82.02, H 5.43.
O C); 55.2
1,2-Bis(4-methoxyphenyl)-2-methoxy-1,2-dihydroacenaphthylen-1-ol (44). Oxirane 43 (0.2 g, 0.52 mmol)
was refluxed in abs. MeOH (20 ml) until complete methanolysis was observed by TLC (ca. 3 d). The solvent was
removed in vacuo, and the resulting viscous yellow-brown oil was chromatographed (silica gel; CH2Cl2). The
Wagner-Meerwein rearrangement product 2,2-bis(4-methoxyphenyl)-1,2-dihydroacenaphthylen-1-one (45) [48]
was collected in the first fraction (40 mg, 20%). The second fraction contained 44, which was crystallized from
AcOEt (110 mg, 56%). Colorless crystals. M.p. 205 ± 2068. IR (KBr): 3470 (OH). 1H-NMR (CDCl3): 7.89 (d, J ˆ
8.3, 1 H); 7.83 (d, J ˆ 8.2, 1 H); 7.61 (m, 2 H); 7.44 (d, J ˆ 6.8, 1 H); 7.34 (d, J ˆ 7, 1 H); 7.05 (d, J ˆ 8.8, 2 H); 6.64
(d, J ˆ 8.8, 2 H); 6.33 (s, 4 H); 4.93 (s, OH); 3.72 (s, MeO); 3.57 (s, MeO); 3.07 (s, MeO). 13C-NMR (CDCl3):
158.8; 157.9 (arom. C OMe); 146.9; 139.3; 138.0; 135.9; 131.0; 129.7; 129.4; 129.0; 128.1; 127.9; 126.9; 125.7;
124.4; 123.5; 121.3; 113.7; 112.9; 112.2 (arom. C); 93.1 (C OMe); 90.6 (C OH); 55.1; 55.0; 52.1 (MeO). Anal.
calc. for C27H24O4 (412.5): C 78.62, H 5.86; found: C 78.38, H 5.76.
6. Photooxygenation of 43. Oxirane 43 (0.38 g, 1 mmol) was dissolved in dry toluene (100 ml), degassed
with ultrasound under N2, and irradiated (500-W Hg high-pressure lamp) under N2 at r.t. After 10 min, a few ml
of the intensely red soln. was transferred to a dry Ar-filled UV cell, and an absorption with lmax 532.5 nm was
recorded. Dry O2 was bubbled through the remainder of the irradiated soln., which became instantaneously
decolorized. Complete conversion of 43 after 10 h was indicated by TLC (alumina, toluene/AcOEt 10 :1). The
solvent was removed in vacuo, and the remaining yellow-brown oil was crystallized from Et2O/MeOH, affording
40 (0.24 g, 58%).
7. Oxidation of Alkene 36 by m-CPBA. A mixture of 36 (0.6 g, 1.64 mmol), CH2Cl2 (50 ml), sat. aq.
NaHCO3 (50 ml), and m-CPBA (70 ± 75%, 0.3 g, ca. 1.8 mmol) was stirred for 24 h at r.t. Another portion of m-
CPBA (0.3 g) was added and stirring continued for 1 h until 36 was completely consumed (TLC monitoring),
and the mixture was worked up. Instead of epoxide 43, rearranged ketone 45 (0.13 g, 21%, m.p. 1518 (acetone)
[48]), and the C C cleavage product 35 (0.43 g, 67%, m.p. 2188 (acetone) [48]), were isolated by CC (silica gel,
CH2Cl2) in the first and the second fractions, resp.
8. General Ozonation Procedures. 8.1. Tetraarylethylenes (TPE, 1: Table 1 and Fig. 4; TAE, 21: Table 2; 9-
alkylidene-9H-fluorenes, 27a,b,c: Scheme 8; and 36: Scheme 11). Routine ozonolyses were carried out under
normal conditions, i.e., introduction of a dry stream of O2/O3 (ca. 3%) at a flow rate of 1 l/min (corresponding to
ca. 1.8 mmol O3/min) into a soln. of the alkene. When the reaction was complete, dry O2 was bubbled through
the mixture for further 15 min at the reaction temp., and dry N2 was bubbled through the mixture for 1 h at r.t.
before the mixture was worked up (see Table 2 for details regarding catalysts, quantities, concentrations,
solvents, and reaction temp.). The following procedures from Schemes 8 and 11 were selected as typical
examples of ozonolyses of 27c and of 36. Olefin 27c (1.0 g, 3 mmol) in dry 3,3-dimethylbutan-2-one
(pinacolone, 50 ml) was ozonized at 508 until the red color disappeared. The solvent was removed in vacuo,
and the remaining yellow-white residue was stirred with MeOH (5 ml), stored for 12 h at 08, and filtered by
suction. The resulting crystals were suspended in MeOH (3 ml), heated to reflux, the hot suspension was

Helvetica Chimica Acta ± Vol. 83 (2000)
823
filtered, and the product recrystallized from dioxane to yield 29 (0.25 g, 21%). Colorless crystals. M.p. 2148 (2038
[45][80b]). Workup of the mother liquor of the hot suspension in MeOH yielded additional 29 (60 mg, for a
total yield of 0.31 g, 26%) together with fluorenone 28 (0.72 g, 67%, m.p. 838 (EtOH)). Refluxing 29 in PhCl did
not lead to chemiluminescence as reported by Yang and Carr [46a]. Because of the >108 deviation from
literature values [45][80b] for the m.p. of 29, this product was further characterized by spectroscopic methods.
IR (KBr): 3020, 1615, 1590, 1455, 1315, 1225, 1210, 1160, 1030, 1000, 760. 1H-NMR ((D6)DMSO): 7.98 (d, J ˆ 7.7,
4 H); 7.81 (d, J ˆ 7.7, 2 H); 7.52 (m, 8 H); 7.39 (t, J ˆ 7.4, 2 H); 6.87 (m, 2 H). 13C-NMR ((D6)DMSO): 135.5;
‡
‡
129.4; 123.6; 121.1 (arom. C); 95.4 (O
1
C
O). CI-MS: 393.3 (18.7 [M ‡ 1] ), 392.3 (26, M ), 377.3 (68.3, [M ‡
‡ ‡
‡
‡
O] ), 361.3 (87.4 [M ‡ 1 O2] ), 180.2 (100, [C13H8O] ), 152.2 (83.1 [C12H8] ).
8.2. 1,2-Bis(4-methoxyphenyl)acenaphthylene Ozonide (39).Olefin 36 (2.0 g, 5.5 mmol) in dry CH2Cl2
(150 ml) was ozonized at 758 until the red-orange color had disappeared, followed by workup according to the
general procedure to yield 39 (1.95 g, 86%), which could be recrystallized from warm (but not refluxing)
MeOH. Nearly white crystals. M.p. 1618. IR (KBr): 3060, 3020, 2980, 2940, 2905, 2850, 1620, 1585, 1520, 1500,
1470, 1440, 1355, 1305, 1280, 1255, 1180, 1090, 1075, 1050, 1030, 1005, 830, 790. 1H-NMR (CDCl3): 7.83 (m, 2 H);
7.65 (m, 4 H); 7.37 (m, 2 H); 6.99 (m, 6 H); 3.85 (s, 2 MeO). 13C-NMR (CDCl3): 160.5 (arom. C OMe); 135.8,
132.9; 128.4; 128.3; 125.7; 125.5; 123.4; 113.7 (arom. C); 110.0 (ozonide C); 55.4 (MeO). Anal. calc. for C26H20O5
(412.5): C 75.72, H 4.89; found: C 75.73, H 5.01.
9. Generation of 8-(4-Methoxybenzoyl)naphthalen-1-yl 4-Methoxybenzoate (40) from Ozonide 39. 9.1.
Photo-catalyzed Rearrangement. Ozonide 39 (0.5 g, 1.2 mmol) in abs. toluene (100 ml) was irradiated, under N2
at 08, with a 500-W halogen day-light lamp (7 h). After storing the mixture for several days at 208, crystalline
40 was obtained in nearly quant. yield. Recrystallization from AcOEt yielded colorless product (0.4 g, 81%),
which was identical to the product from peracid oxidation of diketone 35.
9.2. Acid-Catalyzed Rearrangement. To a stirred soln. of 39 (0.2 g, 0.5 mmol) in abs. CH2Cl2 (20 ml) under
N2, one drop of HSO3Cl was added at 08. The colorless soln. immediately became reddish brown. TLC analysis
(silica gel; CH2Cl2/petroleum ether 1:1) revealed quant. isomerization of 39 to keto ester 40. After 1 h, the soln.
was extracted with sat. aq. NaHCO3, followed by H2O, then dried (MgSO4), and the solvent was removed in
vacuo, yielding homogeneous 40 in quant. yield (TLC).
We cordially thank L. Eberson (Lund, Sweden) and J. Fabian (Dresden, Germany) for helpful discussions
and for providing valuable references, R.F. Langler (Sackville, Canada) for helpful discussions and proof
reading, and Mrs. B. Boeffel for the organization of this manuscript.
REFERENCES
[1] H. Fischer, R. Müller, Hoppe-Seylerꢁs Z. Physiol. Chem. 1925, 148, 157.
[2] a) C. S. Marvel, V. Nichols, J. Org. Chem. 1941, 6, 296; b) N. A. Milas, P. Davis, J. T. Nolan Jr., J. Am. Chem.
Â
Soc. 1955, 77, 2536; c) J. I. G. Cadogan, D. H. Hey, W. A. Sanderson, J. Chem. Soc. 1960, 4897; d) S. Fliszar,
Â
D. Gravel, E. Cavalieri, Can. J. Chem. 1966, 44, 67; e) S. Fliszar, J. B. Chylinska, Can. J. Chem. 1968, 46, 783;
f) K. J. McCullough, A. R. Morgan, D. C. Nonhebel, P. L. Pauson, G. J. White, J. Chem. Res. (Miniprint)
1980, 0601.
[3] R. Criegee, Chimia 1968, 22, 392.
[4] R. Criegee, H. Korber, Chem. Ber. 1971, 104, 1812.
[5] a) M. Miura, M. Nojima, J. Am Chem. Soc. 1980, 102, 288; b) A. P. Schaap, L. Lopez, S. D. Gagnon, J. Am.
Chem. Soc. 1983, 105, 663.
Â
[6] J. Renard, S. Fliszar, Can. J. Chem. 1969, 47, 3333.
[7] J. C. Robertson, W. J. Verzino Jr., J. Org. Chem. 1970, 35, 545.
[8] a) Y. Ogata, Y. Sawaki, M. Shiroyama, J. Org. Chem. 1977, 42, 4061; b) Y. Sawaki, H. Kato, Y. Ogata, J. Am.
Chem. Soc. 1981, 103, 3832; c) K. Griesbaum, W.-S. Kim, N. Nakamura, M. Mori, M. Nojima, S.
Kusabayashi, J. Org. Chem. 1990, 55, 6153.
[9] S. Rakovski, D. Cherneva, Int. J. Chem. Kinetics 1990, 22, 321.
[10] W. W. Sander, J. Org. Chem. 1989, 54, 333.
[11] a) F. W. Hardstock, J. M. Kanabus-Kaminska, D. Griller, Int. J. Chem. Kinetics 1989, 21, 157, b) D. D. M.
Wayner, D. Griller, in ꢀAdvances in Free Radical Chemistryꢁ Vol. 1: Ed. D. D. Tanner, JAI Press,
Greenwich, 1990, pp. 181 ± 182.

824
Helvetica Chimica Acta ± Vol. 83 (2000)
[12] a) R. Criegee, Angew. Chem. 1975, 87, 765; R. Criegee, Angew. Chem., Int. Ed. 1975, 14, 745; b) P. S. Bailey,
ꢀOzonation in Organic Chemistry; Part 1: Olefinic Compoundsꢁ, Academic Press, New York, 1978; c) G.
Zvilichovsky, B. Zvilichovsky, in ꢀThe Chemistry of Functional Groups: Supplement E 2, The Chemistry of
Hydroxyl, Ether, and Peroxide Groupsꢁ, Ed. S. Patai, Wiley-Interscience, Chichester, 1993, pp. 687 ± 784.
[13] a) W. Luef, R. Keese, in ꢀTopics in Stereochemistryꢁ Vol. 20, Eds. E. L. Eliel, S. H. Wilen, John Wiley &
Sons, New York, 1991, pp. 231 ± 318; b) J. Sandström, in ꢀThe Chemistry of Functional Groups, Supplement
A 3: The Chemistry of Double-Bonded Functional Groupsꢁ, Ed. S. Patai, John Wiley & Sons Ltd.,
Chichester, 1997, pp. 1253 ± 1280.
[14] C. S. Marvel, V. E. Nichols, J. Am. Chem. Soc. 1938, 60, 1455.
[15] P. D. Bartlett, T. G. Traylor, J. Am. Chem. Soc. 1962, 84, 3408.
[16] S. W. Lebedew, J. Andreewsky, A. Matyuchkina, Ber. Dtsch. Chem. Ges. 1923, 56, 2349.
[17] E. P. Kohler, N. K. Richtmyer, J. Am. Chem. Soc. 1930, 52, 2038.
[18] H. Beck, Ph.D. Thesis, University of Saarland, 1988.
[19] O. L. Mageli, C. S. Sheppard, in ꢀOrganic Peroxidesꢁ Vol. I, Ed. D. Swern, John Wiley & Sons, New York,
1970, p. 32.
[20] P. Bladon, K. J. McCullough, A. R. Morgan, D. C. Nonhebel, P. L. Pauson, G. J. White, J. Chem. Res.
(Miniprint) 1980, 3701.
[21] R. Trambarulo, S. N. Ghosh, C. A. Burrus Jr., W. Gordy, J. Chem. Phys. 1953, 21, 851.
[22] a) R. D. Harcourt, in ꢀLecture Notes in Chemistryꢁ, Vol. 30, Springer-Verlag, Berlin, 1982; b) P. C. Hiberty,
Isr. J. Chem. 1983, 23, 10; c) S. D. Kahn, W. J. Hehre, J. A. Pople, J. Am. Chem. Soc. 1987, 109, 1871; d) K.
Andersson, P. Borowski, P. W. Fowler, P.-Š. Malmquist, B. O. Ross, A. J. Sadley, Chem. Phys. Lett. 1992,
190, 367.
[23] D. L. Cooper, J. Gerratt, M. Raimondi, J. Chem. Soc., Perkin Trans. 2 1989, 1187.
[24] a) D. Bahnemann, E. J. Hart, J. Chem. Phys. 1982, 86, 252; b) P. Zurer, Chem. Eng. News 1995, 73, 4; c) J. S.
Francisco, M. M. Maricq, Acc. Chem. Res. 1996, 29, 391.
[25] a) H. S. Johnston, D. M. Yost, J. Chem. Phys. 1949, 17, 386; b) D. D. Davis, J. Prusazcyk, M. Dwyer, P. Kim,
J. Phys. Chem. 1974, 78, 1775; c) H. Suzuki, T. Takeuchi, T. Mori, J. Org. Chem. 1996, 61, 5944 and related
communications.
[26] a) P. S. Bailey, J. E. Keller, J. Org. Chem. 1968, 33, 2680; b) W. Rundel, in ꢀMethoden der Organischen
Chemieꢁ 4. Aufl., Georg Thieme Verlag, Stuttgart, 1968, p. 440; c) H. G. Padeken, O. von Schickh, A.
Segnitz, in ꢀMethoden der Organischen Chemieꢁ 4. Aufl., Georg Thieme Verlag, Stuttgart, 1971, p. 106;
d) R. E. Erickson, T. M. Myszkiewicz, J. Org. Chem. 1965, 30, 4326; e) J. L. Gagnon, T. R. Walters, W. W.
Zajac Jr., J. H. Buzby, J. Org. Chem. 1993, 58, 6712.
[27] P. S. Bailey, ꢀOzonation in Organic Chemistry, Part 2: Nonolefinic Compoundsꢁ, Academic Press, New
York, 1982, p. 266.
[28] a) M. S. Kharasch, B. S. Joshi, J. Org. Chem. 1957, 22, 143; b) P. D. Bartlett, T. Funahashi, J. Am. Chem. Soc.
1962, 84, 2596; c) F. D. Greene, W. Adam, J. Org. Chem. 1963, 28, 3550.
[29] S. Pistorius, Ph.D. Thesis, 1992, University of Saarland.
[30] a) I. C. Lewis, L. S. Singer, J. Chem. Phys. 1965, 43, 2712; b) E. Bosch, J. K. Kochi, J. Am. Chem. Soc. 1996,
118, 1319.
[31] a) R. E. Buckles, G. M. Matlack, Org. Synth., Coll. Vol. IV 1963, 914; b) T. Mukaiyama, T. Sato, J. Hanna,
Chem. Lett. 1973, 1041.
[32] H.-W. Adolph, M. Zeppepauer, Departments of Biochemistry, and Pharmacy and Environmental
Technology, University of Saarland, personal communication.
[33] H.-H. Ruf, K. Gersonde, IBMT Fraunhofer-Institut für Biomedizinische Technik, St. Ingbert, Germany,
personal communication.
[34] F. Ciminale, L. Lopez, G. Mele, Tetrahedron 1994, 50, 12685.
[35] J.-P. Desvergne, H. Bouas-Laurent, J. Catalysis 1978, 51, 126.
[36] P. S. Bailey, ꢀOzonation in Organic Chemistry, Part 1: Olefinic Compoundsꢁ, Academic Press, New York,
1978, p. 133; P. S. Bailey, ꢀOzonation in Organic Chemistry, Part 2: Nonolefinic Compoundsꢁ Academic
Press, New York, 1982, p. 144.
[37] J. Meisenheimer, Ber. Dtsch. Chem. Ges. 1926, 59, 1848.
[38] H. T. Openshaw, ꢀA Laboratory Manual of Qualitative Organic Analysisꢁ, 3rd edn., Cambridge University
Press, Cambridge, 1965, p. 80; R. Kluge, M. Schulz, S. Liebsch, Tetrahedron 1996, 52, 5773.
[39] a) L. Eberson, M. P. Hartshorn, O. Persson, J. Chem. Soc. Perkin Trans. 2 1995, 1131; L. Eberson, M. P.
Hartshorn, O. Persson, J. Chem. Soc., Perkin Trans. 2 1995, 1735; b) L. Eberson, O. Persson, F. Radner,

Helvetica Chimica Acta ± Vol. 83 (2000)
825
M. P. Hartshorn, Res. Chem. Intermed. 1996, 22, 799; c) L. Eberson, M. P. Hartshorn, F. Radner, O. Persson,
Chem. Commun. 1996, 215; d) L. Eberson, O. Persson, Acta Chem. Scand. 1996, 50, 395; e) L. Eberson,
M. P. Hartshorn, O. Persson, F. Radner, Chem. Commun. 1996, 2105; f) L. Eberson, M. P. Hartshorn, F.
Radner, O. Persson, J. Chem. Soc., Perkin Trans. 2 1998, 59.
[40] J. Eriksen, C. S. Foote, J. Am. Chem. Soc. 1977, 99, 6455; J. Eriksen, C. S. Foote, J. Am. Chem. Soc. 1980, 102,
6083.
[41] a) J. Leimner, P. Weyerstahl, Chem. Ber. 1982, 115, 3697; b) J. E. Gano, C. Kluwe, K. Kirschbaum, A. A.
Pinkerton, P. Sekher, E. Shrzypczak-Jankun, G. Subramaniam, D. Lenoir, Acta Crystallogr. Sect. C: Cryst.
Struct. Commun. 1997, 53, 1723.
[42] C. Breyer, Diplomarbeit, University of Saarland, 1998.
[43] V. Kozhevnikov, L. Eberson, New J. Chem. 1992, 16, 351.
[44] M. L. T. M. B. Franco, B. J. Herold, J. C. Evans, C. C. J. Rowlands, J. Chem. Soc., Perkin Trans 2 1988, 443.
[45] R. Criegee, G. Lohaus, Liebigs Ann. Chem. 1953, 583, 6.
[46] a) N. C. Yang, R. V. Carr, Tetrahedron Lett. 1972, 5143; b) N. C. Yang, J. Liebman, J. Org. Chem. 1974, 39,
1782.
[47] K. Schank, M. Buschlinger, J. Prakt. Chem. 1995, 337, 196.
[48] W. E. Bachmann, E.-J. W. Chu, J. Am. Chem. Soc., 1936, 58, 1118.
[49] a) R. D. Gould, J. W. Linnett, J. Chem. Soc., Faraday Trans 1963, 59, 1001; b) K. Yamaguchi, K. Takada, Y.
Otsuji, K. Mizuno, in ꢀOrganic Peroxidesꢁ, Ed. W. Ando, John Wiley & Sons Ltd., Chichester, 1992, pp. 1 ±
100.
[50] W. A. Goddard III, T. H. Dunning Jr., W. J. Hunt, P. J. Hay, Acc. Chem. Res. 1973, 6, 368.
[51] L. B. Harding, W. A. Goddard III, J. Am. Chem. Soc. 1978, 100, 7180.
[52] a) W. A. Pryor, D. G. Prier, D. F. Church, J. Am. Chem. Soc. 1983, 105, 2883; b) W. A. Pryor, N. Ohto, D. F.
Church, J. Am. Chem. Soc. 1983, 105, 3614; c) D. H. Giamalva, D. F. Church, W. A. Pryor, J. Am. Chem.
Soc. 1986, 108, 7678.
[53] X.-M. Zhang, Q. Zhu, J. Org. Chem. 1997, 62, 5934.
[54] L. E. Bennett, P. Warlop, Inorg. Chem. 1990, 29, 1975; b) L. Forni, D. Bahnemann, E. J. Hart, J. Phys.
Chem. 1982, 86, 255; c) A. R. Forrester, V. Purushotham, J. Chem. Soc., Faraday Trans. 1 1987, 83, 211.
[55] a) W. J. Bouma, J. K. MacLeod, L. Radom, J. Am. Chem. Soc. 1979, 101, 5540; b) K. N. Houk, N. G.
Rondan, C. Santiago, C. J. Gallo, R. Wells-Gandour, G. W. Griffin, J. Am. Chem. Soc. 1980, 102, 1504;
c) B. C. Baumann, J. K. MacLeod, J. Am. Chem. Soc. 1981, 103, 6223; d) W. J. Bouma, D. Poppinger, S.
Saebo, J. K. MacLeod, L. Radom, Chem. Phys. Lett. 1984, 104, 198: e) T. Bally, S. Nitsche, E. Haselbach,
Helv. Chim. Acta 1984, 67, 86; f) T. Clark, Chem. Commun. 1984, 666; g) L. D. Snow, F. Williams, Chem.
Phys. Lett. 1988, 143, 521; h) M. C. R. Symons, J. L. Wyatt, Chem. Phys. Lett. 1988, 146, 473; i) C.
Wesdemiotis, B. Leyh, A. Fura, F. W. McLafferty, J. Am. Chem. Soc. 1990, 112, 8655; j) H. D. Roth, Top.
Curr. Chem. 1992, 163, 133; k) T. Clark, Top. Curr. Chem. 1996, 177, 1; l) M. N. Eberlin, A. E. P. M. Sorrilha,
F. C. Gozzo, R. S. Pimpin, J. Am. Chem. Soc. 1997, 119, 3550.
[56] K. Schank, S. Pistorius, M. Weiter, F. Werner, J. Prakt. Chem. 1995, 337, 409.
[57] a) S. F. Nelsen, Acc. Chem. Res. 1987, 20, 269; b) S. F. Nelsen, in ꢀAdvances in Oxygenated Processesꢁ, Ed.
A. L. Baumstark, JAI Press Inc., Greenwich, Vol. 2, 1990, pp. 159 ± 179; c) M. Buschlinger, Ph.D. Thesis,
University of Saarbrücken, 1994.
[58] a) M. Schmittel, A. Burghart, Angew. Chem. 1997, 109, 2658; M. Schmittel, A. Burghart, Angew. Chem.,
Int. Ed. 1997, 36, 2550; b) M. Patz and S. Fukuzumi, J. Phys. Org. Chem. 1997, 10, 129.
Â
[59] a) L. Lopez, L. Troisi, Tetrahedron Lett. 1989, 30, 3097; b) H. E. Audier, J. F. Dupin, M. Fetizon, Y.
Hoppilliard, Tetrahedron Lett. 1966, 2077.
[60] D. G. Stoub, J. L. Goodman, J. Am. Chem. Soc. 1997, 119, 11110.
[61] a) H. Perst, ꢀOxonium Ions in Organic Chemistryꢁ, Verlag Chemie, Weinheim, 1971, p. 81; b) S. Hünig,
Angew. Chem. 1964, 76, 400; S. Hünig, Angew. Chem. Int. Ed. 1964, 3, 548.
[62] R. W. Murray, W. Kong, S. N. Rajadhyaksha, J. Org. Chem. 1993, 58, 315.
[63] a) K. Griesbaum, Brennst. Chem. 1969, 50, 212; b) E. H. Pryde, J. C. Cowan, in ꢀTopics in Lipid Chemistryꢁ,
Ed. F. D. Gunstone, Logos Press, London 1971, Vol. 2, pp. 45 ± 48; c) P. S. Bailey, ꢀOzonation in Organic
Chemistry. Part 1: Olefinic Compoundsꢁ, Academic Press, New York, 1978, p. 133; P. S. Bailey, ꢀOzonation
in Organic Chemistry. Part 2: Nonolefinic Compoundsꢁ, Academic Press, New York, 1982, pp. 144 ± 146.
[64] M. Ferrer, F. Sanchez-Baeza, M. Messeguer, Tetrahedron 1997, 53, 15877.
[65] R. E. Buckles, W. D. Womer, J. Am. Chem. Soc. 1958, 80, 5055.

826
Helvetica Chimica Acta ± Vol. 83 (2000)
[66] a) J. W. Hill, J. A. Jenson, J. G. Yaritz, J. Chem. Educ. 1986, 63, 916; b) V. Kaufmann, Ber. Dtsch. Chem.
Ges. 1896, 29, 73.
[67] J. R. Sampey, E. E. Reid, J. Am. Chem. Soc. 1947, 69, 234.
[68] P. M. G. Bavin, Can. J. Chem. 1960, 38, 882.
[69] K. Omura, Synthesis 1998, 1145.
[70] G. M. Coppinger, T. W. Campbell, J. Am. Chem. Soc. 1953, 75, 734.
[71] E. Müller, A. Rieker, R. Mayer, K. Scheffer, Liebigs Ann. Chem. 1961, 645, 36.
[72] R. G. Jones, J. Am. Chem. Soc. 1952, 74, 4889.
[73] a) L. Gattermann, H. Wieland, W. Sucrow, ꢀDie Praxis des organischen Chemikersꢁ, 43rd edn., de Gruyter,
Berlin, New York, 1982, p. 704; b) J. F. Norris, G. W. Rigby, J. Am. Chem. Soc. 1932, 54, 2088; c) M. Pfanne,
Monatsh. Chem. 1911, 32, 509.
[74] W. E. Bachmann, Org. Synth., Coll., Vol. II, 1943, 73.
[75] W. A. Mosher, F. W. Steffgen, P. T. Lansbury, J. Org. Chem. 1961, 26, 670.
[76] J. H. Ford, C. D. Thompson, C. S. Marvel, J. Am. Chem. Soc. 1935, 57, 2619.
[77] S. K. Agarwal, R. W. Murray, Photochem. Photobiol. 1982, 31.
[78] G. Cilento, J. Am. Chem. Soc. 1953, 75, 3748.
[79] a) M. Gomberg, W. E. Bachmann, J. Am. Chem. Soc. 1927, 49, 236; b) T. Minami, N. Matsuzaki, Y. Ohshiro,
T. Agawa, J. Chem. Soc., Perkins 1, 1980, 1731; c) L. Lopez, G. Mele, C. Mazzeo, J. Chem. Soc., Perkin
Trans. 1 1994, 779.
[80] a) G. Wittig, G. Pieper, Ber. Dtsch. Chem. Ges. 1940, 73, 295; b) R. Criegee, W. Schnorrenberg, J. Becke,
Liebigs Ann. Chem. 1949, 7.
[81] a) Y. Sprinzak, J. Am. Chem. Soc. 1958, 80, 5449; b) L. F. Fieser, M. Fieser, ꢀReagents for Organic
Synthesisꢁ, Vol. 1, Wiley and Sons, Sydney, 1967, 1252.
[82] a) A. Kliegl, Ber. Dtsch. Chem. Ges. 1929, 62, 1327; b) P. Wan, E. Krogh, Chem. Commun. 1985, 1207.
Received February 2, 2000