New free-radical halogenations of alkanes, catalysed by N-hydroxyphthalimide. Polar and enthalpic effects on the chemo- and regioselectivity

Article abstract of DOI:10.1016/j.tetlet.2003.12.130

Reactions of alkanes with different halogenating systems are compared in order to explore the reactivity of phthalimido-N-oxyl radical in hydrogen abstraction; the importance of polar effects is emphasised.

Full text of DOI:10.1016/j.tetlet.2003.12.130

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 1607–1609
New free-radical halogenations of alkanes, catalysed
by N-hydroxyphthalimide. Polar and enthalpic effects
on the chemo- and regioselectivity
Francesco Minisci,a,* Ombretta Porta,a Francesco Recupero,a Cristian Gambarotti,a
Roberto Paganelli,a Gian Franco Pedullib and Francesca Fontanac
aDipartimento di Chimica, Materiali e Ingegneria Chimica G.Natta, Politecnico di Milano, via Mancinelli 7, 20131 Milano, Italy
bDipartimento di Chimica Organica A. Mangini, Universitꢀa degli Studi di Bologna, via S.Donato 15, 40127 Bologna, Italy
cDipartimento di Ingegneria Industriale, Universitꢀa di Bergamo, v.le Marconi 5, 24044 Dalmine (BG), Italy
Received 10 December 2003; revised 22 December 2003; accepted 23 December 2003
Abstract—Reactions of alkanes with different halogenating systems are compared in order to explore the reactivity of phthalimido-
N-oxyl radical in hydrogen abstraction; the importance of polar effects is emphasised.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Recently we have evaluated1 the bond dissociation
enthalpies (BDE) of the O–H bonds of a variety of
hydroxylamine derivatives, showing how the BDE val-
ues strongly increase from alkyl- or arylhydroxylamine
(ꢀ70 kcal molꢁ1) to mono- and di-acylhydroxylamines
(88.1 kcal molꢁ1 for N-hydroxyphthalimide, NHPI).
Moreover, we have determined1 the absolute rate con-
stant for hydrogen abstraction from a series of C–H
bonds by phthalimido-N-oxyl (PINO) radical, generated
in situ from NHPI (0.38, 2.4, 3.3, 28.3 and 0.047 Mꢁ1 sꢁ1
at 25 ꢁC, respectively, for Ph–CH3, Ph–CH2–Me, Ph–
CHMe2, Ph–CH2OH and cyclohexane) giving for the
first time a quantitative basis for the catalytic activity of
NHPI in aerobic oxidations.2;3 The hydrogen abstrac-
tions by t-BuOOÅ radical are considerably slower3
(0.036, 0.20, 0.22, 0.13, 0.0034 Mꢁ1 sꢁ1 at 25 ꢁC, respec-
tively, for Ph–CH3, Ph–CH2–Me, Ph–CHMe2, Ph–
CH2OH and cyclohexane) in spite of the fact that, BDE
values for O–H bonds in NHPI and ROO–H being
identical (88 kcal molꢁ1), the enthalpic effect in hydrogen
abstraction by PINO and t-BuOOÅ radicals should be
quite similar. The rate difference is particularly marked
with Ph–CH2OH, where hydrogen abstraction is 218
times faster by PINO than by t-BuOOÅ radical, and we3;4
have suggested that this different behaviour could be due
to a more pronounced electrophilic character of the
PINO radical (Eq. 1) compared to the peroxyl radical.
O
O
δ+
δ−
O
N
O
H
CHOH-Ph
N
O
O
ð1Þ
+ H-CHOH-Ph
O
.
N
OH
+ CHOH-Ph
O
Nitroxyl radicals are generally electrophilic, but the
polar character should be considerably enhanced by the
presence of two carbonyl groups in PINO (Eq. 2)
O
N+
O
N
O
O
O
O
ð2Þ
The effect should be similar to the one observed5 with
acylperoxyl, RC(@O)–OOÅ, compared to alkylperoxyl,
R–OOÅ, radicals: the former are more electrophilic than
the latter. This phenomenon could be more marked with
PINO than with acylperoxyl radicals, because nitrogen
can settle a positive charge, as in Eq. 2, better than
carbon or oxygen.
Keywords: N-Hydroxyphthalimide; Alkane halogenation; Free-radi-
cals.
The absolute rate constants in hydrogen abstraction
from substituted benzyl alcohols would support the
* Corresponding author. Tel.: +39-02-2399-3030; fax: +39-02-2399-
3080; e-mail: francesco.minisci@polimi.it
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.12.130

1608
F. Minisci et al. / Tetrahedron Letters 45 (2004) 1607–1609
importance of the electrophilic character of PINO rad-
ical, since a good Hammett correlation was found.6
Exceptions are observed in the cases of p-nitro- and
p-cyanobenzyl alcohols, which have the same rate con-
stants as unsubstituted benzyl alcohol. A possible
explanation of what appears to be a contradictory
behaviour could be related to a capto-dative effect,
which exceeds the sum of the individual substituent
(–OH and –NO2 or –CN) effects by a synergetic stabil-
isation. In this letter we report new free-radical halo-
genations of alkanes and substituted alkanes, catalysed
by NHPI, which further confirm the marked polar effect
in hydrogen abstraction from C–H bonds by PINO
radical.
polysubstituted derivatives, due to the high reactivity
and low selectivity of the chlorine atom. By using Br2
instead of CuCl2 under the same conditions, bromina-
tion of the alkane takes place; in this case the alkyl
bromides are formed through the fast reaction of alkyl
radical with Br2 (Eq. 9)
RÅ þ Br2 ! RBr þ BrÅ
ð9Þ
The bromine atom abstracts a hydrogen atom from
NHPI regenerating the PINO radical (Eq. 10); HBr is
oxidised to Br2 by HNO3 or NO2.
˜N–OH þ BrÅ ! ˜N–OÅ þ HBr
ð10Þ
We expect Eq. 10 to be much faster than hydrogen
abstraction from C–H bonds of alkanes by BrÅ for
enthalpic reasons (BDE values of the O–H bond in
ROO–H and of the H–Br bond are identical,
88 kcal molꢁ1, and we have evaluated1 that hydrogen
abstraction from NHPI by R–OOÅ radical is about seven
orders of magnitude faster, that is, 7.2 · 103 Mꢁ1 sꢁ1 at
30 ꢁC, than hydrogen abstraction from the C–H bonds
of alkanes, which is about 10ꢁ4 Mꢁ1 sꢁ1). This is clearly
When cyclohexane was oxidised by HNO3 under nitro-
gen or oxygen atmosphere in the presence of NHPI and
CuCl2, chlorocyclohexane was obtained without traces
of dichlorocyclohexanes at 35% conversion, based on
HNO3, indicating that the introduction of a Cl atom
deactivates the alkyl group towards further chlorination.
On the opposite, polychlorination easily occurs by Cl2
chlorination, even at lower conversions. No chlorination
by CuCl2 occurs under the same conditions in the
absence of NHPI, suggesting that hydrogen abstraction
from C–H bonds by PINO radical (Eq. 3), generated
from NHPI and HNO3 or NO2 (Eqs. 4 and 5) is a key
step of the process.
Table 1. Isomer distribution in chlorination, bromination and acet-
oxylation of substituted alkanes, catalysed by NHPIa
Method
Substrate (isomer distribution)
Cl—CH2—CH2—CH2—CH2—CH2—CH3
4.9 4.7 13.6 34.0 42.8 traces
˜N–OÅ þ R–H ! ˜N–OH þ RÅ
˜N–OH þ HNO3 ! ˜N–OÅ þ NO2 þ H2O
ð3Þ
ð4Þ
A
B
3.8
5.1
15.3 32.2 43.6 traces
C
D
6.1 5.5 13.6 30.9 43.9
---
--- traces 7.1 35.4 57.5
---
12
Cl2
Me2NCl12
1.8 9.4 21.3 23.9 26.5 17.2
--- --- 6.8 19.4 71.6 2.2
˜N–OH þ NO2 ! ˜N–OÅ þ HNO2 ! NO þ NO2 ð5Þ
The resulting cyclohexyl radical reacts very rapidly7 with
CuCl2, leading to chlorocyclohexane (Eq. 6).
MeOCO—CH2—CH2—CH2—CH2—CH3
6.1 12.1 28.3 53.5 ---
A
C
6.3 16.0 27.0 50.7 ---
k6
Me2NCl12;13
Me2NBr14
RÅ þ CuCl2 !R–Cl þ CuCl k6 > 109 Mꢁ1 sꢁ1
ð6Þ
---
---
0.7 6.3 87.3 6.7
0.4 5.9 87.0 6.8
Cu(II) is continuously regenerated from CuCl by oxi-
dation with HNO3, NO2 or O2. With stoichiometric
amount of HNO3 under N2 atmosphere the overall
stoichiometry is expressed by Eq. 7.
MeOCO—CH2—CH2—CH2—CH3
5.4 30.1 44.5 18.7
15
Cl2
B
C
8.2 27.1 64.7 traces
9.4 26.6 64.0 traces
15
Br2
Me2NCl13
35.0 19.6 45.4
--- 13.3 79.9
---
6.7
3RH þ 2HNO3 þ 3HCl ! 3RCl þ 2NO þ 4H2O ð7Þ
In the presence of O2 with catalytic HNO3, NO is con-
tinuously oxidised to NO2 and the stoichiometry is given
by Eq. 8. Also in this case, chlorination is inhibited in
the absence of NHPI.
O2N—CH2—CH2—CH2—CH2—CH2—CH3
--- --- 9.2 32.2 58.6 ---
A
D
AcO—CH2—CH2—CH2—CH2—CH2—CH2—CH3
--- 3.5 5.4 20.5 28.3 42.3 ---
1
2
RH þ O2 þ HCl ! RCl þ H2O
ð8Þ
CH3—CH2—CH2—CH2—CH2—CH2—CH3
--- 45.9 36.9 17.2
In principle, the alkyl radical RÅ can also rapidly react
with O2, NO2 or NO (the rates of all these reactions are
often diffusion controlled), but the much higher con-
centration of CuCl2 in the reaction medium makes the
chlorination quite selective.
D
Me2 NCl16
1.1 55.6 29.0 14.3
a Typical procedures of halogenation and acetoxylation are: (A) a
solution of 2 mmol of substrate, 0.4 mmol of HNO3, 2 mmol of CuCl2
and 0.4 mmol of NHPI in 10 mL of AcOH were stirred at 100 ꢁC for
5 h under N2, the solution was diluted by water, extracted by CH2Cl2
and analysed by GLC, as previously9 reported: a 31% conversion
based on HNO3 was obtained; (B) as in (A) under O2, 47% conver-
sion; (C) as in (A) with 2 mmol of Br2 instead of CuCl2 and 0.02 mmol
of Cu(OAc)2 at 80 ꢁC, 35% conversion; (D) as in (A) with 2 mmol of
I2 instead of CuCl2, 37% conversion.
In order to evaluate the extent of polar and enthalpic
effects for hydrogen abstraction a series of substituted
alkanes has been investigated. Only monosubstitution
occurs and the results are reported in Table 1. The use of
Cl2 instead of CuCl2 leads to a complex mixture of

F. Minisci et al. / Tetrahedron Letters 45 (2004) 1607–1609
1609
supported by the observed selectivities, quite different
from those found in free-radical bromination by Br2 in
the absence of NHPI (Table 1).
hypothesis that the higher reactivity of PINO radical,
compared to peroxyl (ROOÅ), in the hydrogen abstrac-
tion from C–H bonds must be related to a more pro-
nounced electrophilic character, as shown by Eqs. 1 and
2. Unfortunately, a direct comparison between PINO
and peroxyl radicals in hydrogen abstraction from
substituted alkanes is prevented by the difficulty of
having clean reactions with peroxyl radicals.
In the presence of I2 the alkyl iodides, initially formed
from alkyl radicals, undergo solvolysis (the solvent is
acetic acid) under the reaction conditions, leading to
acetoxy derivatives. The reaction is catalytic in I2, since
it is regenerated from HI, formed during the solvolysis,
by oxidation with HNO3 or O2; the concentration of I2
in the solution, however, must be kept relatively high in
order to overcome the competition of the alkyl radical
reactions with NO2 or O2. No hydrogen abstraction can
take place by iodine atom from either NHPI or C–H
bonds, due to the low BDE value for H–I, but PINO
radical is generated according to Eqs. 4 and 5.
The effectiveness of the catalytic activity is not high with
alkanes, as we have evaluated1 the first order rate con-
stant for self-decay of PINO radical to be 0.1 sꢁ1 at
25 ꢁC, whereas the rate of hydrogen abstraction from
cyclohexane by PINO is rather low (0.047 Mꢁ1 sꢁ1 at
25 ꢁC); thus a considerable amount of NHPI is con-
sumed during the catalytic process. A higher effective-
ness has been observed with more reactive substrates:
benzyl alcohol4;6 is 600 times more reactive than cyclo-
hexane in hydrogen abstraction by PINO, while cum-
ene11 is 70 times more reactive than cyclohexane. This
gives particular interest to the functionalisation of these
substrates.4;6;11
In Table 1 some results obtained with a series of
substituted alkanes by NHPI catalysis are compared
with the analogous results obtained both in halogen-
ations with Cl2 and Br2, where ClÅ and BrÅ are the
hydrogen abstracting species, and in what Deno8 has
called the ÔMinisci halogenationÕ, where amino radical
cations, R2NHþÅ, from protonated haloamines are the
hydrogen abstracting species.9 All these results suggest
that:
References and notes
(i) In the reactions catalysed by NHPI, chemoselectivity
is much higher than in free-radical halogenations by Cl2
or Br2: the introduction of a halogen atom or an acetoxy
group determines a significant deactivation of the sub-
strate, allowing selective monosubstitution even at
considerable conversions. This behaviour is more similar
to halogenation by N-haloamines, which is particularly
sensitive to polar effects, than to halogenation by Cl2 or
Br2.
1. Amorati, R.; Lucarini, M.; Mugnaini, V.; Pedulli, G. F.;
Minisci, F.; Fontana, F.; Recupero, F.; Astolfi, P.; Greci,
L. J. Org. Chem. 2003, 68, 1747.
2. Review: Ishii, Y.; Sakaguchi, S.; Iwahama, T. Adv. Synth.
Catal. 2001, 343, 393; recent papers: Hara, T.; Iwahama,
T.; Sakaguchi, S.; Ishii, Y. J. Org. Chem. 2001, 66, 6425;
Tsujimoto, S.; Sakaguchi, S.; Ishii, Y. Tetrahedron Lett.
2003, 44, 5601; Nobuyoshi, K.; Basudeb, S.; Espenson, J.
H. J. Org. Chem. 2003, 68, 9364; Arnaud, R.; Milet, A.;
Adamo, C.; Einhorn, C.; Einhorn, J. J. Chem. Soc., Perkin
Trans. 2 2002, 1967.
(ii) The regioselectivity for chlorination and bromina-
tion in the presence of NHPI is quite similar, strongly
suggesting the involvement of the same hydrogen
abstracting species (PINO radical, Eq. 3); this regio-
selectivity is quite different from the one observed with
Cl2 or Br2 in the absence of NHPI, which suggests a
large polar effect in hydrogen abstraction by PINO
radical, even though this effect is lower compared to
R2NHþÅ radicals. Differences in acetoxylation selectivity
are likely related to the solvolysis of the initially formed
alkyl iodides.10
3. Review: Minisci, F.; Recupero, F.; Pedulli, G. F.; Luca-
rini, M. J. Mol. Catal. A 2003, 204–205, 63.
4. Minisci, F.; Punta, C.; Recupero, F.; Fontana, F.; Pedulli,
G. F. Chem. Commun. 2002, 688.
5. Bravo, A.; Bjørsvik, H.-R.; Fontana, F.; Minisci, F.; Serri,
A. J. Org. Chem. 1996, 61, 9409.
6. Minisci, F.; Recupero, F.; Punta, C.; Gambarotti, C.;
Paganelli, R.; Pedulli, G. F. Eur. J. Org. Chem. 2004, 109.
7. Kochi, J. K. Acc. Chem. Res. 1974, 7, 351.
8. Deno, N. C. In Methods in Free-Radical Chemistry;
Huyser, E. S., Ed.; Dekker: New York, 1972; p 147.
9. Review: Minisci, F. In Substituent Effects in Free-Radical
Chemistry; Viehe, H. G., Ed.; Reidel: 1986. p 391.
10. Minisci, F.; Recupero, F.; Punta, C.; Gambarotti, C.;
Paganelli, R. Tetrahedron Lett. 2003, 44, 6919.
11. Minisci, F.; Recupero, F.; Gambarotti, C.; Cecchetto, A.;
Pedulli, G. F.; Fontana F. Org. Proc. Res. Dev., in press.
12. Minisci, F.; Galli, R.; Bernardi, R. Chim. Ind. (Milan)
1967, 49, 594; Minisci, F.; Gardini, G.; Bertini, F. Can. J.
Chem. 1970, 48, 544.
(iii) The enthalpic effects also considerably affect selec-
tivity: the methyl group, even if it is the least deactivated
by the polar substituents, reacts only in traces, due to
the higher BDE values of the C–H bonds compared to
those in the –CH2– group. Also the similar selectivity in
positions 1 and 2 of 1-chlorohexane can be explained in
terms of higher polar deactivation of position 1, bal-
anced by the more favourable enthalpic effect.
13. Minisci, F.; Galli, R.; Galli, A.; Bernardi, R. Tetrahedron
Lett. 1967, 2207.
14. Minisci, F.; Galli, R.; Bernardi, R. J. Chem. Soc., Chem.
Commun. 1967, 903.
15. Singh, H.; Tedder, J. M. J. Chem. Soc. B 1964, 4737.
16. Bernardi, R.; Galli, R.; Minisci, F. J. Chem. Soc. B 1968,
324.
(iv) In the acetoxylation of n-heptane position 2 is
somewhat more reactive than positions 3 and 4, which
show the same reactivity, as already9 observed with N-
chloroamines. All these results strongly support the