Full text of DOI:10.1002/1521-3765(20010105)7:1<297::AID-CHEM297>3.0.CO;2-6

FULL PAPER
Mild and Efficient Flavin-Catalyzed H2O2 Oxidations
Alexander B. E. Minidis and Jan-E. Bäckvall*[a]
Abstract: Based on a previously discovered method for amine oxidations using
flavins as catalysts and hydrogen peroxide as oxidant, a comparative kinetic study
using NMR spectroscopy was undertaken with a series of flavins for amine and
thioether oxidations. Included in this series is the newly prepared 7,8-difluoro-1,3-
dimethyl-5-ethyl-5,10-dihydroalloxazine. This study shows that flavins, which bear
electron-donating groups on the aromatic ring and/or the N-10 position, are less
active and are deactivated during the course of the reaction. Moreover, flavins that
are alkylated at the N-1 position instead of the N-10 position and having either no
substituents or electron-withdrawing groups on the aromatic ring, remain the most
active and stable.
Keywords: flavins
catalysis ´ hydrogen peroxide ´ ox-
idation
´ homogeneous
Me
Me
N
Ph
N
Introduction
H
9
H
N
H
N
N
N
1
O
Me
Me
O
O
8
7
10
5
3
N
The oxidation of organic substrates as for example amines and
thioethers with stoichiometric flavin hydroperoxides has been
known for more than two decades.[1] A more economical and
environmentally friendly system with only catalytic amounts
of flavin has so far been limited to a few publications involving
N
N
4a
N
Me
N
Me
N
Me
6
Et
O
Et
O
Et
O
1
2
3
Me
N
Me
N
Me
N
H
N
H
N
H
N
O
Me
Me
O
F
F
O
the oxidation of secondary amines to nitrones,[2] sulfoxida-
N
N
N
[2, 3]
N
Me
N
Me
N
Me
tions,
and Baeyer± Villiger oxidation of selected sub-
Et
O
Et
O
Et
O
strates.[4] Only recently we discovered that flavin 1 could be
used as catalyst in a mild and efficient oxidation of tertiary
amines with hydrogen peroxide as terminal oxidant,[5] which
also leads to a synthetically more viable reaction.
4
6
5
Because of the difference in catalytic activity of earlier
employed flavins 2 ± 5 compared with 1,[1±5] we wanted to
investigate if there was a correlation between the catalytic
activity of these flavins and the substitution pattern and/or the
electronic properties due to the substituents. Thus, a series of
flavins (1 ± 6) was prepared and employed in the oxidation of
tertiary amines and thioethers. In addition to N oxidation,[5]
the sulfoxidation is of current interest, since most methods for
oxidation of thioethers to sulfoxides have several limita-
tions.[6, 7]
Results and Discussion
Flavins 1 ± 5 and the novel difluoro compound 6 were
synthesized according to literature procedures,[4, 5a, 8, 9] and
used as catalysts as described for the oxidation of tertiary
amines.[5a] To observe conversion over a given time some
reactions were followed by NMR spectroscopy, which also
allows for initial rate determination based upon a pseudo-first
order mechanism.[2]
Sulfoxidation with 1 as catalyst: We first studied the oxidation
of a set of different thioethers to sulfoxides by H2O2 catalyzed
by flavin 1 (Table 1). The results demonstrate the effective-
ness and mildness of the new oxidation system. Only a small
amount of catalyst 1 was necessary (1.8 mol%) and a minor
excess of the final oxidant H2O2 (1.75 equiv) was used.[10]
Methanol was chosen as a solvent and no inert reaction
conditions such as dry solvents or an oxygen-free atmosphere
were necessary.[1±3]
[a] Prof. Dr. J.-E. Bäckvall, Dr. A. B. E. Minidis
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University, 10691 Stockholm (Sweden)
Fax : (‡ 46)8-154908
E-mail: jeb@organ.su.se
Supporting information for this article is available on the WWW under
http://www.wiley-vch.de/home/chemistry or from the author. Kinetic
curves for the multiple addition experiments and a mass spectrum for 6
are available (four pages).
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0297 $ 17.50+.50/0
297

J.-E. Bäckvall and A. B. E. Minidis
FULL PAPER
Table 1. Flavin-catalyzed sulfide oxidation.[a]
of the thioether unit.[7e, 12] The aniline thioether 11
demonstrates, that deactivated amines are also compat-
ible with this oxidation system. Only a quinone
thioether could not be oxidized.[13] For dithiane 17
many procedures exist for its selective mono- and
dioxidation, however, all of them have drawbacks such
as the requirement of low temperature or strong
oxidants.[14] The flavin/H2O2 system yields only the
mono-oxide, even if two equivalents of hydrogen
peroxide are employed.
Comparing the results in Table 1 with the previously
reported results for N-oxidation,[5a] the rate enhance-
ments are in general equally high for the S- and N-
oxidation. Considering that the background reaction for
dialkyl sulfides is fast, the rate enhancement achieved
for such substrates is especially noteworthy. On the
other hand, even though the background reaction for
substrates such as diphenyl sulfide is very slow (11%
conversion in four days), the catalyzed reaction (11%
conversion in 160 min) results in a rate enhancement of
approximately the same range as most other substrates
(41:1 and 4280:1, respectively).
O
S
cat. 1
S
R1
R2
R1
R2
H2O2
CH3OH or CD3OD
Substrate
Product
mol% Time Conv./%[b] Rate
flavin [min] (yield/%)[c] enhance-
ment[d]
O
S
S
1.83
0.50
1.63
60
100
(quant)
74:1
1
2
(7100:1)[e]
8
7
7
8
360[f]
160
99[g]
(98)
n.d.
O
S
S
95.8
99.9
99.5
45:1
(4730:1)[e]
3
4
5
Br
Br
10
9
O
S
S
1.33
1.60
23
45
12:1
(1440:1)[e]
H2N
H2N
12
11
O
S
S
36:1
(3620:1)[e]
In addition, the reaction is not limited to the
arbitrarily chosen amounts of catalyst and hydrogen
peroxide. These amounts can be lowered to render an
equally effective, albeit slower, system, which never-
theless gives full conversion (Table 1, entry 1 vs 2 and
entry 6 vs 7).
MeO
MeO
14
2S
13
1.60
0.10
15
99
(quant)
12.5:1
(735:1)[e]
O
2S
6
7
16
15
15
16
175
98.5
2:1
(2040:1)[e]
As mentioned earlier, the solvent of choice was
methanol. However, other polar solvents such as
acetonitrile, acetone, tert-butanol, and DMSO work
almost equally well, although with lower rates (Table 2).
Methylene chloride works fine with urea hydrogen
peroxide, whereas with aqueous H2O2 the reaction stops
after about 25% conversion due to phase separation
and extraction of the flavin intermediate into the
aqueous phase. Sodium peroxide as terminal oxidant
gave almost no conversion.
99[h]
n.d.
n.d.
S
S
S
S
8
9
1.70
1.28
20
15
O
18
17
O
O
O
100
(quant)
S
S
HO
HO
5
5
20
19
O
O
O
S
S
10
1.08
30
99.0
n.d.
O
O
22
21
Catalytic amine- and sulfide-oxidations have been
described in the literature only for N,N-3,10-dimethy-
lated flavin derivatives such as 4 and 5.[1±3] Catalysts
based on such a substitution pattern result only in low
rate enhancements, whereas the N,N-1,3-dimethylated
flavin 1 is a highly active catalyst as shown above for
sulfoxidations and previously for amine oxidations.[5a]
These findings led us to investigate the possible reasons
for this difference in activity.
[a] Substrate (0.223 mmol) was dissolved in methanol (600 mL, [D4]MeOH for NMR
measurement). Catalyst (3 ± 4 mmol) and H2O2 (40 mL, 30% aq. solution,
1
0.392 mmol) were added at 258C (no inert conditions). For the NMR experiments
the tube was shaken well until a homogenous solution was obtained. [b] Determined
by 1H NMR spectra. [c] Isolated yield. [d] Describes the rate enhancement of
catalyzed versus noncatalyzed reaction (see Experimental Section for details). In
parenthesis the estimated ratio of reactivities of the catalytic flavin hydroperoxide
and H2O2 is given (see [e]). [e] Estimated ratio of reactivities of the catalytic flavin
hydroperoxide and H2O2, which is corrected for the amount of catalyst and amount
of H2O2 to obtain an appropriate comparison of the two peroxides ({rate
enhancement/amount catalyst)} ´ equiv. H2O2).[5] [f] Only 1.0 equiv H2O2 (30% in
water) was used. [g] TON ˆ 198. [h] With 2 equiv H2O2 no formation of 1,3-dioxide
was detectable after 1 h.
Relative rates of different flavins in catalytic N- and S-
oxidation: To gain further insight into the influence of
the substitution pattern of flavins on their catalytic
activity, flavins 1 ± 6 were prepared and compared in the H2O2
oxidation of tertiary amines (Figure 1) and sulfides (Figure 2).
The new 7,8-difluoro derivative 6 was synthesized according to
Scheme 1. Attempts to synthesize the 7,8-dichloro derivative
of 1 failed under the reductive alkylation conditions em-
ployed for preparing such reduced flavin structures.[5a] Flavins
bearing other electron-withdrawing substituents such as a
nitrile would most likely fail under these conditions as well.[15]
As expected, aliphatic sulfides and electron-rich aryl alkyl
sulfides are oxidized faster than electron-poor aryl alkyl
sulfides (Table 1).[6, 7] It is noteworthy that no overoxidation
was observed in any of the cases studied. Esters (e.g. 21), acids
(e.g. 19), and alkenes (e.g. 21) are also compatible with this
oxidation system.[11] Other mild sulfoxidation methods such as
catalytic MTO (methyltrioxorhenium) with H2O2 oxidize
ꢀ
both the carbon carbon double bond, as well as the sulfur
298
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0298 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

Flavin-Catalyzed Oxidations
297±302
Table 2. Solvents and oxidants in oxidation of 7.
result of a disadvantageous substituent combination: elec-
tron-donating groups on the aromatic ring and trialkylation at
nitrogens in positions 3, 5, and 10. Flavin 4, which differs from
5 only in that the 7,8-dimethyl groups have been removed, was
a better catalyst than 5.
The mechanism of the flavin-catalyzed N- and S-oxidations
is given in Scheme 2. Initial preactivation by molecular
oxygen produces the flavin hydroperoxide I.[16] Transfer of
the electrophilic oxygen to the substrate via II gives the 4a-
hydroxy intermediate III. Elimination of the hydroxide ion
from III gives IV, which can react with hydrogen peroxide to
regenerate flavin hydroperoxide I. Taking into consideration
the reaction mechanism in Scheme 2 and the fact that the
oxygen transfer (I ! II ! III) to the substrate is most likely
the rate-limiting step,[5a] electron-donating groups on the
aromatic ring seem to stabilize the peroxo intermediate I, thus
making it less reactive.[17] Consequently, a flavin with two
Oxidant
Solvent
mol%
flavin 1 [min]
Time Conv.[%][a]
30% H2O2
30% H2O2
30% H2O2
30% H2O2
30% H2O2
30% H2O2
30% H2O2
H2O2 ´ urea
H2O2 ´ urea
H2O2 ´ urea
H2O2 ´ urea
CD3OD
CD3OD
CD3CN
[D6]acetone
tBuOH
[D6]DMSO
CD2Cl2
CD2Cl2
±
60
55
60
60
60
60
20[b]
60
60
60
60
60
7.5
100
76
23
60
14
26
100
91
100
100
13
1.6
1.7
1.7
1.5
2.9
1.7
1.7
1.8
MeOH
MeOH/H2O (20:1) 1.7
abs. MeOH[c]
1.6
1.8
Na2CO3 ´ 1.5H2O2 MeOH
[a] Conversion to sulfoxide 8. [b] Reaction stopped after ca. 25%
conversion due to phase separation and extraction of the flavin inter-
mediate into the aqueous phase. [c] Methanol was dried over activated
molecular sieves with a stream of argon passing through the solvent for 1 h.
A general trend (Figures 1
and 2) is that N,N,N-1,3,5-tri-
substituted flavins (1, 2, and 6)
show significantly higher activ-
ity than the corresponding
N,N,N-3,5,10-trisubstituted an-
alogues (3, 4, and 5). As can
also be seen from the Figures,
flavin 1 is an efficient catalyst
with high initial rates resulting
in full substrate conversions.
Catalyst 6 gave full conversion
within about 1 h, although the
initial rate was slightly lower
than that of 1. Flavin 2 showed
the same high initial rate as
catalyst 1 in the N-oxidation,
but the activity is lost during the
course of the reaction. A sim-
Figure 1. Comparison of flavins in the N-oxidation of N,N-dimethyl benzylamine.
ilar deactivation of 2 was also
observed in the S-oxidation. A
likely explanation for the deac-
tivation of 2 is that the electron-
donating groups on the aromat-
ic ring (7,8-dimethyl) make it
electron rich enough to become
sensitive towards oxidative deg-
radation. Nevertheless, flavin 2
results in a much higher sub-
strate conversion than catalysts
3 ± 5. Compounds 3 ± 5 differ
from 1 in that they have elec-
tron-donating substituents on
the nitrogen in position 10
(i.e., they are N,N,N-3,5,10-tri-
substituted) and when em-
ployed as catalysts they result
in only
a slightly enhanced
activity over the noncatalyzed
reaction. The very low activity
of 5 actually seems to be the
Figure 2. Comparison of flavins in the S-oxidation of 7.
0947-6539/01/0701-0299 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
299

J.-E. Bäckvall and A. B. E. Minidis
FULL PAPER
between I and III so that the con-
centration of I becomes very low.[19]
Consequently, even though step I !
II may have become faster for 6, a
lower concentration of I will result
in an overall reaction rate which is
less than one would expect by just
considering the oxidation potential
of I.
Me
N
H
N
H
N
H
N
NH2
NH2
O
O
O
F
F
F
F
N
N
O
F
F
O
a
c
+
NH
* H2O
NH
N
N
Me
O
O
Et
O
23 R = H
24 R = CH3
b
6
Scheme 1. Synthesis of 6. a) H3BO3, HOAc; b) K2CO3, MeI, DMF, 508C; c) H2, Pd/C, CH3CHO, HCl,
EtOH/H2O.
The flavin hydroperoxide derived from O2 and 5, analogous
to I, has been previously studied by Bruice[1c±e] and shown to
be between 5 Â 103 and 104 times as active as hydrogen
peroxide in stoichiometric N oxidations. This rate enhance-
ment is of the same order of magnitude as that estimated for
the hydroperoxide of 1.[5a] An interesting question therefore is
why 5 does not work well in the catalytic reaction. To explain
this one would have to assume that the reduced form IIIb
(analogous to III) is not efficiently recycled to flavin hydro-
peroxide because of degradation and/or slow kinetics. It is
likely that the slow step in this recycling is the dissociation of
the hydroxy group to give the charged species IVb. This is
supported by results by Murahashi,[2] who found that the rate-
limiting step in S- and N oxidation with 7,8-dimethyl-N,N-
3,10-dimethyl-flavin corresponding to 5[20] is the elimination
of the hydroxy ion from the 4a-hydroxy intermediate IIIb to
give a cationic isoalloxazine (IVb). In fact, for the oxidation
of a secondary amine the rate constant for step IIIb ! IVb
was 3600 times smaller than the rate constant for the O
transfer from flavin hydroperoxide to nitrogen.[20] For the
N,N-1,3-dimethylated flavins (1, 2, and 6) the step III ! IV
(Scheme 2) should be faster due to formation of a fused
aromatic system (alloxazine).[21] Thus, the higher catalytic
activity of 1, 2, and 6 would be due to a more efficient
recycling of III to I (Scheme 2).
Scheme 2. Proposed catalytic cycle.
electron-withdrawing substituents in the 7- and 8-positions
and methylated nitrogens in positions 1 and 3 is expected to
result in increased reaction rates. When the 7,8-difluoro flavin
6 was employed as catalyst, this flavin did indeed show
different properties compared with all other catalysts. There is
a significant induction period, presumably due to the fact that
the flavin precursor is less reactive towards activation by
molecular oxygen.[18] After this initial phase though, the
reaction proceeds quite rapidly.
If more substrate and oxidant are added at the end of the
reaction, the subsequent reaction rates are as fast as in the first
run in both the N- (Figure 1) and S oxidation (Figure 2). This
was demonstrated for catalyst 6 in the S-oxidation of
4-methylphenyl methylsulfide (7) and N oxidation of N,N-
dimethylbenzylamine and for catalyst 1 in S-oxidation of
4-methylphenyl methylsulfide (7) (see Supporting Informa-
tion). The fact that 6 is obviously as stable as 1 in contrast to
all other flavin catalysts supports the hypothesis of electron-
donating groups having a negative effect on the catalytic
properties. The reason why it does not show an overall higher
reaction rate could be due to a change in the rate-limiting step.
With 6, the reoxidation of III ! I, that is, IV! I, (Scheme 2)
should be slower compared with that of 1. If I turns into a
more powerful oxidant (higher oxidation potential) because
of the electron-withdrawing groups, it should be more difficult
to reoxidize the corresponding reduced form IV back to I.
Thus, the reoxidation may become the rate-determining step
and there would not be full equilibrium between IV and I.
Even if the reoxidation does not become the rate-limiting step
the increased oxidation potential may shift the equilibrium
Me
N
Me
N
Me
Me
N
O
O
Me
Me
N
O
O
N
N
N
Me
N
Me
O
HO
H
IVb
IIIb
Conclusion
In summary, we have extended the use of our simple flavin/
hydrogen peroxide system to the oxidation of thioethers to
sulfoxides, which are accessible in short reaction times in
quantitative yields. The comparative kinetic study of different
flavin catalysts has shown that those based on the nitrogen
substitution of the natural riboflavin with substituents on N-3
and N-10 are less active catalysts than their corresponding
N,N-1,3-dialkylated counterparts. Furthermore, it was shown
that flavins, which are alkylated at positions N-1 and N-3,
having either no substituents such as 1 or electron-with-
drawing groups such as 6 on the aromatic ring, result in highly
efficient catalysts. The higher activity of N,N,N-1,3,5-trisub-
stituted flavins (alloxazine system, 1, 2, and 6) over their
N,N,N-3,5,10-trisubstituted counterparts (isoalloxazine sys-
300
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0300 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

Flavin-Catalyzed Oxidations
297±302
excess of solid sodium dithionite instead of a dithionite solution was used
during the work-up. Notably the product is highly air sensitive in solution
tem, 3 ± 5) is thought to be due to the higher rate of
dissociation of the hydroxy group from the 4a-hydroxy
intermediate of the former flavins to give an alloxazine
(III). Flavins 1 and 6 are each accessible in only three steps
with their precursors being much easier to handle than those
bearing substituents at the N-10 instead of the N-1 position.[22]
These two flavins are therefore the preferred choice of
catalyst for the described mild and efficient N- and S
oxidation. Finally, these results are also important in view of
developing more efficient chiral flavin catalysts,[3] where a
suppression of the relatively fast background reaction is
essential.
and
a rapid work-up is necessary. This demands that all necessary
equipment must be assembled and accessible in advance. Reaction with
7,8-difluoro-1,3-dimethylalloxazine (600 mg, 2.15 mmol) yielded a yellow-
green solid (65 mg, 10%). Due to its oxygen sensitivity, the NMR sample of
6 was prepared under argon in degassed CDCl3 having a layer of aqueous
(D2O) Na2S2O4 on top; a MS spectra could also be obtained[26] using this
solution directly after preparation. 1H NMR (CDCl3, 300 MHz): d ˆ 6.71
(dd, J ˆ 7.5, 11.4 Hz, 1H; arom H), 6.38 (dd, J ˆ 7.2, 10.2 Hz, 1H; arom H),
5.42 (br, 1H; NH), 3.45 (s, 3H; N-CH3), ꢁ3.44 (overlap; q, J ˆ 7.2 Hz, 2H;
CH2), 1.16 (t, J ˆ 6.9 Hz, 3H; CH3); 19F NMR (CDCl3, 282.2 MHz): d ˆ
ꢀ142.2 (m, 1F), ꢀ143.7 (m, 1F); MS (EI): m/z (%): 309 (19), 308 (77)
‡
[M] , 307 (91), 279 (100).
General procedure for the kinetic study: The conversion rates of the
reactions performed in the presence or absence of flavin catalyst were
determined by integration of the corresponding 1H NMR signals for the
ꢀ
sulfide and product sulfoxide. For most compounds the R SMe signals
were integrated; however, in a few cases CHSMe or CH2SMe signals were
integrated due to better separation from other peaks. The rate enhance-
ment was calculated by division of the rate at low conversion (ꢂ10%), for
the catalyzed reactions usually after one minute reaction time.
Experimental Section
General: 1H, 13C, and 19F NMR spectra were recorded on a Varian
300 MHz and 400 MHz Gemini instruments. Chemical shifts are reported
in ppm using residual solvent as internal standard.[23] IR spectra were
measured on a Perkin ± Elmer Spectrum One FTIR. MS spectra were
measured on a Thermo QuestGCQ plus via direct inlet. Elemental
analyses were performed by Analytische Laboratorien, Lindlar (Germa-
ny). Reagents were purchased from Lancaster or Aldrich, except 1,2-
diamino-4,5-difluorobenzene (Apollo). Deuterated solvents were pur-
chased from Cambridge Isotope Laboratories, except [D6]acetone (Sigma)
and CD3CN (Dr. Blaser AG, Basel). 6-(Methylthio)hexanoic acid (19) was
prepared according to Vederas and Liu.[24] Flavins 2 ± 5 have all been
previously described and were synthesized accordingly.[4, 8, 9] Flavin 1 was
prepared according to Bergstad and Bäckvall.[5a] For the final work-up of
1 ± 5, the method described below for compound 6 was used.
A) With flavin catalyst
4-Methylphenyl methyl sulfoxide: Flavin 1 (1.06 mg, 3.89 mmol) was added
to a solution of 4-methylphenyl methylsulfide (30 mL, 0.223 mmol) in
[D4]MeOH (600 mL) in an NMR tube. The tube was shaken well, followed
by the addition of H2O2 (40 mL, 30% aq. solution, 0.392 mmol) to the
solution and the time measurement was started. If shaken well again, all of
1 dissolved immediately to give a yellow solution. The reaction was
followed by 1H NMR spectroscopy (usually at 1 ± 2 min intervals for the
first 10 min, then 5 min intervals for another 20 min and 10 min intervals
until completion).[27] The NMR data for 8 was in accordance to the
literature.[28]
Oxidation of other thioethers: The reactions were followed by 1H NMR
spectra for 20 ± 60 min (1 ± 2 min intervals for the first 5 ± 10 min, thereafter
5 min intervals).[29] The NMR data of the products were in accordance to
literature: 1-methanesulfinyl-4-methoxy-benzene (14),[28] 1-bromo-4-meth-
anesulfinyl-benzene (10),[30] 4-methanesulfinyl-aniline (12),[31] 1,3-dithiane-
1-oxide (18),[32] dibutylsulfoxide (16),[28] diphenylsulfoxide,[28] 6-(methyl-
sulfinyl)hexanoic acid (20).[24]
7,8-Difluoroalloxazine (23): The procedure described by Bergstad and
Bäckvall was followed[5a] using diamino-difluorobenzene (500 mg,
3.47 mmol) which gave
a pale yellow powder (868 mg, 86%). M.p.
>2808C;[25] 1H NMR ([D6]DMSO, 300 MHz): d ˆ 11.99 (br, 1H; NH),
11.77 (br, 1H; NH), 8.27 (dd, J ˆ 8.4, 10.8 Hz, 1H; arom H), 7.97 (dd, J ˆ
8.0, 11.6 Hz, 1H; arom H); 13C NMR ([D6]DMSO, 75.4 MHz): d ˆ 188.7,
160.6, 150.5, 147.8, 139.0 (dd, J ˆ 444.8, 12.2 Hz), 134.7 (dd, J ˆ 434.0,
12.0 Hz), 116.6 (d, J ˆ 16.8 Hz), 113.7 (d, J ˆ 18.3 Hz); 19F NMR
([D6]DMSO, 376.3 MHz): d ˆ ꢀ126.8 (ddd, J ˆ 20, 11, 9 Hz, 1F), ꢀ133.9
B) Without flavin catalyst
4-Methylphenyl methyl sulfoxide: H2O2 (40 mL, 30% aq. solution,
0.392 mmol) was added to
a solution of 7 (30 mL, 0.223 mmol) in
‡
[D4]MeOH (600 mL) in an NMR tube. The reaction was monitored
immediately by 1H NMR for 24 h (77% conversion). The NMR data of
product 8 was in accordance to literature data.[28]
(ddd, J ˆ 20, 11, 9 Hz, 1F); MS (EI): m/z (%): 251 (15), 250 (100) [M] ; IR
(KBr): nÄ ˆ 3564(m), 3478(m), 3073(s), 2932(m), 2840(s), 1747(vs),
1724(vs), 1685(vs), 1635(m), 1589(s), 1578(s), 1513(vs), 1497(s),
1446(m), 1404(m), 1354(s), 1287(s), 1245(s), 1228(s), 1159(m),
1041(w), 892(s), 865(m), 844(m), 811(m), 802(m), 748(m), 690(w),
658(m), 638(w), 592(m), 528(s), 497(s), 455 cmꢀ1 (m); elemental analysis
calcd (%) for C10H4F2N4O2 (250.2): C 48.01, H 1.61; found C 47.01,
H 1.71.
Oxidation of other thioethers: The reactions were followed for 60 ± 180 min
for comparison and initial rate determination in connection with the
catalyzed reactions. For diphenylsulfide the reaction was followed for 5 d
(14% conversion).
General procedure for preparation of sulfoxides (4-methylphenyl methyl
sulfoxide): Flavin 1 (1.06 mg, 3.89 mmol) was added to a solution of
4-methylphenyl methylsulfide (30 mL, 0.223 mmol) in MeOH (600 mL),
followed by the addition of H2O2 (40 mL, 30% aq. solution, 0.392 mmol)
and stirred for one hour. Alternatively the reaction mixture from the NMR
experiments can be used and submitted to work-up as follows: A small
amount of dithionite (ca. 10 mg) was added and the mixture diluted with
diethyl ether (20 mL). Washing with water and drying over sodium sulfate
gave a white solid (31.5 mg, >99%). The NMR data of product 8 was in
accordance to the literature.[25]
7,8-Difluoro-1,3-dimethylalloxazine (24): The procedure described by
Bergstad and Bäckvall was followed,[5a] using 7,8-difluoroalloxazine
(717 mg, 2.86 mmol) and yielded a yellow solid (785 mg, 98.5%). M.p.
1978C; 1H NMR (CDCl3, 300 MHz): d ˆ 8.08 (dd, J ˆ 8.0, 9.6 Hz, 1H;
arom H), 7.77 (dd, J ˆ 7.6, 10.4 Hz, 1H; arom H), 3.80 (s, 3H; N-CH3), 3.60
(s, 3H; N-CH3); 13C NMR (CDCl3, 75.4 MHz): d ˆ 159.1, 155.0 (dd, J ˆ
263.2, 16.1 Hz), 151.7 (dd, J ˆ 257.2, 16.0 Hz), 150.2, 145.4, 141.1 (d, J ˆ
12.1 Hz), 136.8 (d, J ˆ 10.7 Hz), 129.5, 115.9 (dd, J ˆ 17.5, 2.3 Hz), 113.3 (d,
J ˆ 18.3 Hz), 29.7, 29.3; 19F NMR (CDCl3, 376.3 MHz): d ˆ ꢀ122.6 (ddd,
J ˆ 18.3, 10.7, 7.6 Hz, 1F), ꢀ129.6 (ddd, J ˆ 18.3, 10.7, 7.6 Hz, 1F); MS (EI):
‡
m/z (%): 280 (7), 279 (39), 278 (76) [M] , 167 (100); IR (KBr): nÄ ˆ
3550(m), 3476(m), 3413(s), 3044(m), 2962(w), 2924(w), 1727(s),
1680(vs), 1637(m), 1564(s), 1513(m), 1487(s), 1432(m), 1414(s), 1379(s),
1360(m), 1299(s), 1248(s), 1198(m), 1174(m), 1101(w), 1066(w), 982(w),
913(m), 858(m), 832(w), 810(w), 764(vw), 745(w), 731(w), 645(w),
615(m), 568(w), 489(w), 478(m), 459 cmꢀ1 (m); elemental analysis calcd
(%) for C12H8F2N4O2 (278.2): C 51.81, H 2.90; found C 52.05, H 3.01.
Acknowledgement
Financial support from the Swedish Research Council for Chemical
Engineering and the Swedish Foundation for Strategic Research is
Ê
gratefully acknowledged. We would like to thank Dr. Katarina Färnega
Â
Â
(nee Bergstad), Dr. Anna L. E. Minidis (nee Larsson) and Dr. Benjamin
Pelcman, Karobio AB, for helpful suggestions and Joakim Löfstedt for the
help with mass spectroscopy.
7,8-Difluoro-1,3-dimethyl-5-ethyl-5,10-dihydroalloxazine (6): A modified
procedure from Bergstad and Bäckvall was followed,[5a] where a large
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0301 $ 17.50+.50/0
301

J.-E. Bäckvall and A. B. E. Minidis
FULL PAPER
[1] a) A. E. Miller, J. J. Bischoff, C. Bizub, P. Luminoso, S. Smiley, J. Am.
Chem. Soc. 1986, 108, 7773; b) S. Oae, K. Asada, T. Yoshimura,
Tetrahedron Lett. 1983, 24, 1265; c) S. Ball, T. C. Bruice, J. Am. Chem.
Soc. 1980, 102, 6498; d) S. Ball, T. C. Bruice, J. Am. Chem. Soc. 1979,
101, 4017; e) T. C. Bruice, Acc. Chem. Res. 1980, 13, 256.
[12] H. Adolfsson, A. Converso, K. B. Sharpless, Tetrahedron Lett. 1999,
40, 3991. An exception is the MTO/urea hydrogen peroxide system,
which is compatible with alkenes, see ref. [7d].
[13] Only partial decomposition of starting material (2-(phenylthio)-1,4-
benzoquinone) could be detected by NMR.
[2] S.-I. Murahashi, T. Oda, Y. Masui, J. Am. Chem. Soc. 1989, 111,
5002.
[14] See for esample V. K. Aggarwal, I. W. Davies, R. Franklin, J.
Maddock, M. F. Mahon, K. C. Molloy, J. Chem. Soc. Perkin Trans. 1
1994, 2363.
[3] a) S. Shinkai, T. Yamaguchi, O. Manabe, F. Toda, J. Chem. Soc. Chem.
Commun. 1988, 1399; b) S. Shinkai, T. Yamaguchi, A. Kawase, A.
Kitamura, O. Manabe, J. Chem. Soc. Chem. Commun. 1987, 1506.
[4] C. Mazzini, J. Lebreton, R. Furstoss, J. Org. Chem. 1996, 61, 8.
[5] a) K. Bergstad, J.-E. Bäckvall, J. Org. Chem. 1998, 63, 6650; b) K.
Bergstad, S. Y. Jonsson, J.-E. Bäckvall, J. Am. Chem. Soc. 1999, 121,
10424.
[6] For accounts on sulfoxide synthesis and application see: a) K.-D.
Gundermann, K. Hümke, Methoden Org. Chem. (Houben-Weyl)
Organische Schwefel-Verbindungen, Vol. E11/Teil 1 (Ed.: D. Kla-
mann), Thieme, Stuttgart, New York, 1985; b) T. Durst in Compre-
hensive Organic Chemistry, Vol. 5 (Ed.: D. N. Jones), Pergamon,
Oxford, 1979, Sections 11.6 and 11.7; c) G. Solladie in Comprehensive
Organic Synthesis, Vol. 6 (Ed.: B. M. Trost), Pergamon, Oxford, 1991,
Section 1.5; d) S. Uemura in Comprehensive Organic Synthesis, Vol. 6
(Ed.: B. M. Trost), Pergamon, Oxford, 1991, Section 6.2.
[15] K. Kindler, J. Lieb. Ann. Chem. 1931, 485.
[16] This is in analogy to what has been reported for 1,5-dihydroflavins
3
with O2, see ref. [1e] and C. Kemal, T. W.Chan, T. C. Bruice, J. Am.
Chem. Soc. 1977, 99, 7272.
[17] The oxidation of thioethers can be looked upon as a borderline case,
since for example the equilibrium between I, IV, and III can play a
greater role depending on experimental conditions.
[18] An alternative activation mechanism, even though less likely, where
hydrogen peroxide electrophilically (!) attacks the flavin, can not be
ruled out.
[19] Murahashi and co-workers have previously discussed such possible
equilibria, see ref. [2].
[20] Murahashi used the perchlorate analogue of IVb as catalyst precursor.
[21] a) H. I. X. Mager, W. Berends, Tetrahedron 1976, 32, 2303; b) S.
Ghisla, P. Hemmerich, J. Chem. Soc. Perkin Trans. 1 1972, 1564.
[22] Fro example, the solubility and yields are rather low for most
precursors substituted at positions 7, 8, and/or 10.
[7] For some examples where extreme or less economical conditions seem
Â
necessary see: a) S. Vayssie, H. Elias, Angew. Chem. 1998, 110, 2246;
Angew. Chem. Int. Ed. 1998, 37, 2088 (use of in situ nitrous acid);
b) P. E. Correa, D. P. Riley, J. Org. Chem. 1985, 50, 1787 (high pressure
oxygen); c) C. Duboc-Toia, S. Menage, C. Lambeaux, M. Fontecave,
Tetrahedron Lett. 1987, 38, 3727 (m-oxo diferric complexes);
d) H. Q. N. Gunaratne, M. A. McKervey, S. Feutren, J. Finlay, J.
Boyd, Tetrahedron Lett. 1998, 39, 5655 (methyltrioxorhenium (MTO)
plus urea hydrogen peroxide); e) J. H. Espenson, Chem. Commun.
1999, 479 (review on MTO‡H2O2).
[23] H. E. Gottlieb, V. Kotlyar, A. Nudelman, J. Org. Chem. 1997, 62, 7512.
[24] Y. Liu, J. C. Vederas, J. Org. Chem. 1996, 61, 7856.
[25] The exact value was not determinable, since the melting point
exceeded the max. temperature of the apparatus.
[26] Nevertheless, due to the low concentration and its instability, it was
not possible to obtain a 13C NMR spectrum.
[27] When using less catalyst/peroxide the reactions were followed up to
360 min with 30 ± 60 min intervals after the first hour.
[28] M. H. Ali, W. C. Stevens, Synthesis 1997, 764.
[29] When using less catalyst/peroxide the reactions were followed up to
360 min with 30 ± 60 min intervals after the first hour.
[8] G. Eberlein, T. C. Bruice, J. Am. Chem. Soc. 1983, 105, 6685.
[9] S. Ghisla, U. Hartmann, P. Hemmerich, F. Müller, Liebigs Ann. Chem.
1973, 1388.
[10] The excess was mainly used to obtain better accuracy in the
determination of the initial rate of the noncatalyzed reaction.
[11] Furthermore, simple alkenes such as styrene did not oxidize under the
reaction conditions.
[30] R. S. Cooke, G. S. Hammond, J. Am. Chem. Soc. 1970, 92, 2739.
[31] J. B. Hyne, J. W. Greidanus, Can. J. Chem. 1969, 47, 803.
Â
[32] E. Juaristi, J. Guzman, V. V. Kane, R. S. Glass, Tetrahedon 1984, 40,
1477.
Received: September 18, 2000 [F2734]
302
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0302 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1