Synthesis of N-alkoxy amines via catalytic oxidation of hydrocarbons

Article abstract of DOI:10.1002/adsc.200303194

Sterically hindered N-alkoxy amines 3 are synthesized in good yields by coupling nitroxides 2 with hydrocarbyl radicals generated in situ by t-BuOOH hydrogen abstraction from hydrocarbons. The reaction is catalyzed by copper halides as well as by onium iodides.

Full text of DOI:10.1002/adsc.200303194

FULL PAPERS
Synthesis of N-Alkoxy Amines via Catalytic Oxidation of
Hydrocarbons
a
a
a
a
Hans-J. Kirner, Franz Schwarzenbach, Paul A. van der Schaaf, Andreas Hafner,
a
a,
a
b
Val e¬ rie Rast, Markus Frey, * Peter Nesvadba, Guenther Rist
Ciba Specialty Chemicals Inc., Schwarzwaldallee 215, 4002 Basel, Switzerland
a
Fax: (þ41)-61-636-2332, e-mail: markus.frey@cibasc.com
b
Institut f¸r physikalische Chemie der Universit‰t Basel, Klingelbergstrasse 80, 4051 Basel, Switzerland
Fax: (þ41)-61-267-3855, e-mail: guenther.rist@unibas.ch
Received: October 30, 2003; Accepted: March 10, 2004
Abstract: Sterically hindered N-alkoxy amines 3 are Keywords: N-alkoxy amines; tert-butyl hydroperox-
synthesized in good yields by coupling nitroxides 2 ide; copper; hydrocarbons; nitroxides; oxidation;
with hydrocarbyl radicals generated in situ by t- phase-transfer catalysis; radical reactions; TEMPO
BuOOH hydrogen abstraction from hydrocarbons.
The reaction is catalyzed by copper halides as well
as by onium iodides.
Introduction
the accessibility of N-alkoxy amines via this process is
largely determined by their stability with respect to ther-
Sterically hindered amines (1, Scheme 1) are used as ad- mal cleavage (retro-reaction, Scheme 2) and in practice
1
2
ditives to prevent photo-oxidation of polyolefins. Their is limited to R , R ¼H, alkyl. Thus, a process that would
mode of action is believed to involve formation of the ni- allow preparation of N-alkoxy amines with lower CꢀO
1
2
troxide 2, a very efficient trap for carbon-centered radi- bond dissociation energy (e.g., R or R ¼phenyl, vinyl)
cals. The resulting N-alkoxy amine 3 in turn may scav- was highly desirable.
enge peroxy radicals to regenerate the nitroxide (Deni-
[1]
sov cycle).
In the past few years, N-alkoxy amines have attracted
interest as low-basicity analogues of 1 and as polymeri-
[2]
zation regulators. A possible way to prepare N-alkoxy Scheme 2. Formation of N-alkoxy amines (3) by trapping hy-
amines is to generate carbon-centered radicals in pres- drocarbyl radicals with nitroxides (2).
ence of nitroxides, e.g., via hydrogen atom abstraction
from a suitable substrate provided by an oxidant
Results and Discussion
(
Scheme 2). Such a process would be extremely attrac-
tive both from an economic and ecological point of
[4]
view. An early process is based on the MoO -catalyzed Copper Halides as Catalyst
3
decomposition of t-BuOOH, leading to the formation
.
[3]
of t-BuO radicals. The reaction requires relatively Using a model reaction based on TEMPO (2,2’,6,6’-tet-
high temperatures (up to 1008C) but allows direct cou- ramethylpiperidine N-oxide 2, R¼H), ethylbenzene
pling of nitroxides with simple hydrocarbons. However, and t-BuOOH, numerous transition metal salts and ox-
ides have been screened as catalyst. Most of them give
1
2
only modest yields of 3 (R¼H, R ¼Me, R ¼Ph), but
copper halides turned out to be highly active. Thus,
with as low as 0.05 mol % CuCl and under non-aqueous
2
conditions, high yields (>80%) of N-alkoxy amine 3
were obtained at 608C. Allylic and non-activated hydro-
carbons (e.g., cyclohexene, cyclohexane) were found to
workequally well, with 4-propoxy substituents being
Scheme 1. N-Oxides (2, nitroxides) and N-alkoxides (3, tolerated (Scheme 3). Running the model reaction at
NOR) of sterically hindered amines (1).
similarly low catalyst loadings (0.05 mol %), but with
554
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/adsc.200303194
Adv. Synth. Catal. 2004, 346, 554±560

N-Alkoxy Amines via Catalytic Oxidation of Hydrocarbons
FULL PAPERS
Table 1. CuBr -catalyzed reaction of TEMPO (2, R¼H) with ethylbenzene (10 equivs.) at 608C using aqueous 70% t-
2
[a]
BuOOH (3 equivs.): effect of added phase-transfer catalyst (PTC).
[
b]
[b]
Entry
CuBr [equivs.]
PTC [equivs.]
Yield [%]
Conversion [%]
2
1
2
3
4
0.0005
0.0005
0.0005
0.01
None
35
73
78
83
48
92
100
100
Bu NBr [0.01]
4
Oct NMeBr [0.01]
3
None
[
[
a]
b]
Results based on 3.2 mmol scale.
Yield of N-alkoxy amine 3 (R¼H, R ¼Me, R ¼Ph) and conversion of TEMPO after 1 h/608C, respectively.
1
2
>
BrꢁCl. Fluoride did not catalyze the reaction. Iodo
compounds other than Bu NI gave significantly lower
4
yields (ca. 10% using Bu NI , NaI, LiI, I or Bu NOH/
4
3
2
4
I ) or no product at all [iodobenzene, Bu N (o-iodoben-
2
4
zoate), Bu NIO or Bu NIO ]. The more hydrophobic
4
3
4
4
Oct NMeI exhibited similar activity as compared to
3
Bu NI. The catalyst could also conveniently be prepared
4
from, e.g., sodium iodide and a suitable onium salt (e.g.,
chloride, hydrogen sulfate, etc.) via in situ anion ex-
change. Immobilized onium iodides (e.g., tributylme-
thylammonium iodide bound to polystyrene) worked
Scheme 3. Synthesis of N-alkoxy amines 4 to 6 from nitroxide
2
(R¼O-propyl), hydrocarbons and t-BuOOH/(cat.) CuCl .
2
Conditions (4 mmol scale): nitroxide (1 equiv.), hydrocarbon
(
11.7 equivs.), t-BuOOH (53.4% in decane, 6.9 equivs.), 608C, also and could be filtered off after the reaction, but the
CuCl (1 equiv.), time (4: 0.0005 equiv., 60 min; 5: 0.0005 turnover frequency (TOF) was lower due to the hetero-
2
equiv., 12 min; 6: 0.0019 equiv., 200 min); yields (crude) given geneous nature of the reaction. Non-activated sub-
in parenthesis.
strates like cyclohexane required higher temperatures
(
808C) for the reaction to work. TOF and yield were
generally lower as compared to CuBr , the reaction
2
aqueous 70% t-BuOOH instead, caused the yield to not going to completion even after further addition of
drop to 35% (Table 1, Entry 1) unless some phase-trans- t-BuOOH (Figure 1). Thus, in the reaction of 4-benzoyl-
fer catalyst was added (Table 1, Entry 2). Alternatively, oxy-TEMPO (2, R¼OCOPh) to the N-alkoxy amine 3
1
2
yields could be improved by increasing the copper con- (R¼OCOPh, R ±R ¼pentamethylene), Ph PI gave
4
centration (Table 1, Entry 4). Thus, 0.05 mol % CuBr2/ the highest yield (74%; CuBr : 82%) and conversion
2
1
mol % Bu NBr was found to be about equally effec- (88%; CuBr : 95%) but at low TOF (16/h; CuBr : 60/
4
2
2
tive as 1 mol % CuBr . The more hydrophobic Oct
h). Use of polar, inert cosolvents (chlorobenzene and
2
3
NMeBr did not give much better results compared to the like) thought to increase the solubility of Ph PI did
4
Bu NBr (Table 1, Entry 3 versus 2). Cupric and cuprous not remarkably improve TOF. Replacing a phenyl group
4
ions were found to be similarly effective, but for the re- by i-propyl, ethyl, hexyl or hexadecyl resulted in lower
action to worksome halide counter ions (chloride, bro-
yield but higher TOF, the effect being partially paral-
mide) from either the copper or the tetraalkylammoni- leled by an increase of the n-octanol/water partition co-
um cation need to be present. Thus, CuSO did not efficients (log K ). Thus, maximum TOF and yield did
4
ow
workunless some Bu NBr was added. Bu NHSO on not exceed 36/h and 61%, respectively. Similar values
4
4
4
the other hand was not effective as phase-transfer cata- were found for Oct PI (37/h, 58%), shortening of the al-
4
lyst, the activity of CuBr being largely the same whether kyl chains resulting in lower TOF (Bu PI: 29/h, 61%) and
2
4
or not Bu NHSO was added.
yield (Et PI: 15/h, 46%). A dramatic drop in yield was
observed if more than two phenyl groups are replaced
4
4
4
by methyl (PhPMe I: 16/h, 39%) or if the methyl group
3
[5]
Onium Iodides as Catalyst
is activated (Ph PCH CO MeI: 16/h, 36%). Maximum
3
2
2
TOF and yield of tetraalkylammonium iodides (Fig-
Bu NBr by itself was found to be active as catalyst. Thus, ure 2) was found to be very similar to the corresponding
4
in the absence of any copper salt and under the condi- tetraalkylphosphonium iodides, the values depending
tions shown in Table 1, Bu NBr gave about 10% prod- on log K rather than the accessibility (Q) of the ammo-
4
ow
uct. Running the reaction without any additive gave nium cation or the reactivity of the alkyl substituents
[
6]
no product at all. Surprisingly, however, Bu NI exhibit- (benzyl vs. alkyl: OctBzlNMe ꢁOct NMe). Below
4
2
3
ed similar catalytic activity as compared to CuBr (Ta- some threshold (log K <1.7, Bu NI: 29/h, 63%) TOF
2
ow
4
ble 2, Entry 4), the order of decreasing activity being I>
and yield started to decrease (Et NI: 14/h, 38%). Trial-
4
Adv. Synth. Catal. 2004, 346, 554±560
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
555

FULL PAPERS
Hans-J. Kirner et al.
Table 2. Bu NHal (0.01 equiv.) catalyzed reaction of TEM-
4
PO (2, R¼H) with ethylbenzene (10 equivs.) at 608C using
aqueous 70% t-BuOOH (3 equivs): effect of halide (absence
[a]
of copper).
Entry
[
b]
[b]
Hal
Yield [%]
Conversion [%]
1
2
3
4
F
±
9
10
80
±
2.5
8
Cl
Br
I
94
[
[
a]
b]
Results based on 3.2 mmol scale.
1
2
Yield of N-alkoxy amine 3 (R¼H, R ¼Me, R ¼Ph) and
conversion of TEMPO after 1 h/608C, respectively.
Figure 2. R NI (0.01 equiv.) catalyzed reaction of 4-benzoyl-
4
oxy-TEMPO (2, R¼OCOPh) with cyclohexane (10 equivs.)
at 808C using aqueous 70% t-BuOOH (3 equivs): effect of R.
Figure 3. Overlaid plots of pH, N-alkoxy amine (NOR) for-
mation, acetophenone formation and nitroxide conversion
versus time for the reaction of TEMPO (2, R¼H, 1 equiv.)
and ethylbenzene (10 equivs.) with t-BuOOH aq. 70%
Figure 1. R PI (0.01 equiv.) catalyzed reaction of 4-benzoyl-
4
oxy-TEMPO (2, R¼OCOPh) with cyclohexane (10 equivs.)
at 808C using aqueous 70% t-BuOOH (3 equivs.): effect of
R. Results based on 3 mmol scale; log Kow ¼n-octanol/water
(3 equivs.)/Bu NI (0.01 equivs.) at 608C (absence of copper).
4
partition coefficient (calculated); the reaction with Ph PI as
Results based on 6.4-mmol scale.
4
catalyst was repeated on a 35 mmol scale to afford after chro-
1
2
matography 63% of N-alkoxy amine 3 (R¼O CPh, R ±R ¼
2
pentamethylene).
monitored and plotted versus time. Results are shown
in Figure 3. t-BuOOH feed is paralleled by a sharp in-
kylsulfonium iodides showed the lowest activity (Oct2 crease in pH from 7.5 (0% fed) to 8.6 (50% fed), followed
SMeI: 4/h, 12%). Running the reaction under non-aque- by a decrease down again to 7.5 (100% fed). This effect is
ous conditions (t-BuOOH 70% in cyclohexane, mo- specific for the iodide/t-BuOOH interaction, the pH of
lecular sieve) was found to increase TOF, but the yield 70% aqueous t-BuOOH itself being slightly acidic. As
was not affected as demonstrated with Ph PI (18/h, the reaction progresses the pH slowly increases from
4
7
6%; aqueous: 16/h, 74%) and Bu NI (52/h, 62%; aque- 7.5 to finally 8, presumably due to the slightly basic prod-
4
ous: 29/h, 63%). Higher catalyst loadings increased TOF uct. There is clear evidence of an induction period, no
but at the expense of yield.
product being formed until about 50% of t-BuOOH has
been introduced. A plot of % product formation versus
time is linear between 10% and 80% and the rate was cal-
culated assuming zero-order kinetics to give k ¼1.70%/
min (TOF¼102/h, TON¼100). No gap between product
Comparison of pH Profiles
The redox chemistry of halides being pH dependent it formation and TEMPO conversion is seen. The reaction
was speculated that the iodide-catalyzed reaction might is accompanied by the formation of acetophenone (k¼
exhibitadistinctivepHoptimum,thusallowingforthere- 0.15%/min) as major side product, the total amount at
action to be further optimized. First, the pH and product 100% TEMPO conversion being 9% (relative to 100%
formation of the Bu NI-catalyzed model reaction was product). A similar increase and decrease in pH during
4
556
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 554±560

N-Alkoxy Amines via Catalytic Oxidation of Hydrocarbons
FULL PAPERS
Figure 4. Plot of N-alkoxy amine (NOR) and acetophenone
formation rates and their ratio as a function of pH for the re-
action of TEMPO (2, R¼H, 1 equiv.) and ethylbenzene
Figure 6. Plot of N-alkoxy amine (NOR) and acetophenone
formation rates and their ratio as a function of pH for the re-
action of TEMPO (2, R¼H, 1 equiv.) and ethylbenzene
(10 equivs.) with t-BuOOH aq. 70% (3 equivs.)/CuCl2
(0.1 molar) at 608C: effect of added phase-transfer catalyst
(
(
10 equivs.) with t-BuOOH aq. 70% (3 equivs.)/Bu NI
4
0.01 equiv.) at 608C (absence of copper).
(
Oct NMeCl, 0.01 equiv).
3
Stability of N-Alkoxy Amines Towards Oxidation
The CꢀH bonds alpha to the oxygen of the N-alkoxy
amine group are activated with respect to hydrogen
atom abstraction, thus rendering the product more
prone towards oxidation compared to the substrate.
As long as the substrate is used in large excess, over-ox-
idation is not an issue. However, as the excess is lowered
below some threshold (about 3 equivalents), product
oxidation increasingly occurs. This was demonstrated
1
2
Figure 5. Plot of N-alkoxy amine (NOR) and acetophenone
formation rates and their ratio as a function of pH for the re-
action of TEMPO (2, R¼H, 1 equiv.) and ethylbenzene
with N-alkoxy amine 3 (R¼H, R ¼Me, R ¼Ph), aceto-
phenone and the amine 1 being formed as main prod-
ucts.
(
10 equivs.) with t-BuOOH aq. 70% (3 equivs.)/CuCl
0.1 molar) at 608C.
2
(
Hydrogenation of O-Cycloalkenylhydroxylamines
t-BuOOH feed is seen for the reaction catalyzed by
CuCl2.
Taking into account the somewhat lower reactivity of
To monitor the pH profile, the model reaction was run onium iodides compared to copper halides in the cou-
at different pH values kept constant during addition of t- pling of nitroxides with non-activated hydrocarbons,
BuOOH and throughout the reaction. Rates were calcu- the latter coupling products can also be synthesized via
lated as above assuming zero-order kinetics and then hydrogenation of initially formed reaction products
plotted versus pH. Results are shown in Figure 4. The with ethylene-type unsaturated hydrocarbons such as
pH profile exhibits a flat maximum at pH 9, kbeing
O-cycloalkenyl hydroxylamines, accessible in good
1
0% higher only as compared to the reaction lacking yields via onium iodide catalysis. Thus, O-cyclohexenyl
pH control. As the pH gets more acidic, kdecreases by
about 20% (pH 7) and 70% (pH 4). The ratio of product
over acetophenone formation rate is very similar be-
tween pH 10 and 7 but gets higher at lower pH.
hydroxylamine 7, synthesized in good yield using
The CuCl catalyzed model reaction exhibits a com-
2
plementary pH profile (Figure 5). There is no distinctive
maximum, kbeing constant at pH 3 and 4. If the pH is
raised to 5, precipitation of the catalyst starts and k
drops by 60%. This drop is not seen in presence of
phase-transfer catalyst (Figure 6). At pH 4, kis about
four times higher compared to the reaction lacking any
phase-transfer catalyst. However, at the same time the
acetophenone formation rate is about twenty times
Scheme 4. Synthesis of N-alkoxy amines 7 and 8 from nitro-
xide 2 (R¼OCOEt), cyclohexene and t-BuOOH/(cat.)
Bu NI followed by hydrogenation. Conditions: i) cyclohexene
4
(
10 equivs.), t-BuOOH aq. 70% (1.5 equivs.), Bu NI
4
higher, resulting in up to 27% acetophenone at the (0.01 equivs.), 608C, 88% (crude); ii) hexane, 10% Pd/C
end of the reaction.
(0.05 equivs.), 258C, H (4 bar), 91% (crude).
2
Adv. Synth. Catal. 2004, 346, 554±560
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
557

FULL PAPERS
Hans-J. Kirner et al.
Bu NI as catalyst, was hydrogenated to afford N-alkoxy
4
amine 8 (Scheme 4).
Discussion
The t-BuOOH/copper halide/Bu NBr system has been
4
investigated in detail by Sasson who reported the oxida-
tion of p-activated methylenes such as tetraline (1,2,3,4-
tetrahydronaphthalene) affording 1-t-butylperoxytetra-
[
7]
lin as main product. Although there is no mechanistic
evidence presented in the article, it is likely that hydro-
Figure 7. CuBr -catalyzed reaction of TEMPO (2, R¼H)
2
.
gen abstraction by t-BuO radicals is the rate-determin- with ethylbenzene (10 equivs.) at 608C using aqueous 70%
.
.
ing step, the formation of t-BuO and t-BuOO radicals t-BuOOH (3 equivs.): effect of added phase-transfer catalyst
[
8]
from t-BuOOH/Cu being well documented. The effi- on the N-alkoxy amine formation rate (Table 1, Entries 1 and
2
).
cient formation of N-alkoxy amines can be explained
by the high trapping rate and concentration of nitroxyl
.
relative to t-BuOO radicals. Indeed, only minor
amounts of the corresponding t-butyl-(1-methyl-1-ben- transport of Cu catalyst into the organic phase becomes
zyl) peroxide could be detected, acetophenone being rate determining (Figure 7). It has been shown that
the main side product. The latter may result from oxida- Bu NBr efficiently extracts metal salts and even metal
4
tion of initially formed t-butyl-(1-methyl-1-benzyl) per- hydroxides, presumably via the formation of hydrogen
[
7]
oxide or the N-alkoxy amine.
The use of aqueous t-BuOOH results in the formation
bond complexes.
Some years earlier Sasson published a paper describ-
of a two-phase system, TEMPO (log K ꢁ2) and t- ing the decomposition of tetraline hydroperoxide
ow
BuOOH (log K ꢁ1) largely being in the organic (eth- (1,2,3,4-tetrahydronaphthalene 1-hydroperoxide) by
ow
ylbenzene) phase and the copper halide being in the wa- tetrahexylammonium halides affording the related alco-
[
9]
ter phase. As expected pH and transport of the catalyst hol and ketone, respectively. It was proposed that the
across the phase boundary play important roles. Thus, OꢀO bond of the hydroperoxide is cleaved heterolyti-
the results in Table 1 match with the respective pH pro- cally via an initial hydrogen bond complexformation be-
files recorded in absence and presence of phase-transfer tween the hydroxy group of the hydroperoxide and the
catalyst (Figures 5 and 6). At low catalyst loadings anion of the onium catalyst followed by electron transfer
(
0.05 mol %; Table 1, Entry 1) the pH of the water to give alkoxy radicals, halide radicals and onium hy-
phase shifts during the reaction from slightly acidic droxide. Hydrogen atom abstraction from the peroxide
pH 4.5) to slightly basic (pH 7.5), the shift being paral- by halide radicals in the presence of the onium hydrox-
(
leled by the precipitation of hydrolyzed catalyst; at high- ide would then regenerate the onium halide, building up
er catalyst loadings (1 mol %; Table 1, Entry 4; the catalytic cycle shown in Scheme 5[Eqs. (1)±(5)]. As-
[
CuBr ]ꢁ0.1 molar at the end of t-BuOOH feed) the suming that the same cycle applies to theformation of N-
2
pH shift is much smaller, i.e., from pH 4 to pH 4.2, the alkoxy amines from nitroxide, hydrocarbon and t-
catalyst largely remaining dissolved (Figure 5).
.
BuOOH/onium iodide would imply that the t-BuO rad-
In the presence of a phase-transfer catalyst (Table 1, ical is the hydrogen atom abstracting species [Scheme 5,
Entries 2 and 3) the drop in product formation rate, Eqs. (6)±(9)]. In fact, ESR spectroscopic analysis of
which is observed in the absence of phase-transfer cata- mixtures of t-BuOOH and Bu NI in ethylbenzene re-
4
.
lyst when going from pH 4 to pH 7, does not occur. vealed the presence of t-BuOO radicals (Figure 8),
.
Moreover, addition of a phase-transfer catalyst results from which t-BuO radicals might be formed, a process
[
10]
in a rate increase (fourfold with 1 mol % Oct NMeCl found to be catalyzed by nitroxides.
3
at pH 4, Figures 5 and 6). This indicates that i) rates pri-
Also, the increase and decrease of pH during t-
mary depend on the catalyst concentration in the organ- BuOOH feed (Figure 3) would be in line with the ap-
ic phase, ii) both CuCl and hydrolyzed CuCl exhibit pearance of an intermediate basic species as suggested
2
2
similar catalytic activity in presence of a phase-transfer by Sasson×s catalytic cycle [Scheme 5, Eq. (2)]. Howev-
catalyst and iii) in the absence of a phase-transfer cata- er, there are some open questions that do not exclude
lyst and compared to CuCl , the lower reactivity of hy- that with t-BuOOH/onium iodide the mechanism may
2
drolyzed CuCl is due to its lower solubility in the organ- differ from that shown in Scheme 5. Whereas in Sasson×s
2
ic phase. It is therefore reasonable to assume that the re- paper the order of decreasing reactivity is Cl>Br>I,
action takes place in the organic phase. This conclusion which reflects the strength of the anion hydrogen bond
is supported by the kinetics of the model reaction (Ta- and the hydrogen abstraction capability of the halide
ble 1), i.e., in the absence of phase-transfer catalyst the radical, the order of decreasing reactivity observed in
558
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 554±560

N-Alkoxy Amines via Catalytic Oxidation of Hydrocarbons
FULL PAPERS
Conclusion
The present process provides an easy access to N-alkoxy
amines from nitroxides, hydrocarbons and t-BuOOH/
(
cat.) onium iodide or t-BuOOH/(cat.) Cu(I,II)Br,Cl
under mild conditions (alkenes 608C, alkanes 808C).
Copper halides under non-aqueous conditions give the
highest TOF. With 70% aqueous t-BuOOH and activat-
ed hydrocarbons (e.g., cyclohexene, ethylbenzene) cop-
per halides are about equally active as Bu NI (1 mol %
4
catalyst loadings). For non-activated hydrocarbons (e.g.,
cyclohexane), TOF and yield depend on the solubility
and stability of the onium iodide, respectively. Thus,
Scheme 5. Adaptation of the catalytic cycle published by Sas-
son et al. [Eqs. (1)±(5), X¼Br, Cl] to discuss formation [Eqs.
(
6)±(9)] of N-alkoxy amines (3) from nitroxide (2), hydrocar-
þ
ꢀ
bon (RH) and t-BuOOH/(cat.) onium iodide (Q I , X¼I).
Ph PI gives similar yields but lower TOF compared to
4
CuBr . More hydrophobic long-chain tetraalkylphos-
2
phonium or tetraalkylammonium iodides give higher
TOF but lower yields.
the synthesis of N-alkoxy amines is I>BrꢁCl. Hydro-
gen atom abstraction from hydroperoxides by iodo rad-
icals [Scheme 5, Eq. (3)] being unlikely to occur, iodide
must be regenerated by some electron transfer reaction
instead or, alternatively, the catalytic cycle consists of
Experimental Section
other iodo redox couples as in, e.g., the Bray±Liebhafsky General Remarks
[11]
reaction. The low conversions observed in the tetraal-
Starting materials and catalysts were commercial grade and
used as received. Yields and conversions were determined by
HPLC, the reaction being run in presence of biphenyl
kylammonium iodide-catalyzed synthesis of N-alkoxy
1
2
amine 3 (R¼OCOPh, R ±R ¼pentamethylene) and
the reactivity pattern observed for the phosphonium io-
(
0.1 equiv.) as internal standard. Quantitative results are based
dides indicate that catalyst deactivation takes place dur- on nitroxide (1 equiv.±0.1 equiv.)/biphenyl and N-alkoxy
ing the reaction. Indeed, once the reaction has stopped amine (0.1 equiv.±1 equiv.)/biphenyl calibration curves (peak
area ratio versus amount ratio). Turnover number (TON) re-
fers to number of equivalents product formed/number of
equivalents catalyst used, turnover frequency (TOF) to
TON/unit of time. ESR spectra were obtained on a Varian E-
(
Figure 2, after 22 hours) it can be restarted and brought
to higher nitroxide conversions if both additional t-
BuOOH and onium iodide are added. Catalyst deactiva-
tion may be initiated by oxidation of CꢀH bonds in the
9
X-band spectrometer.
vicinity to the quaternary center or by Hofmann elimi-
nation induced by a temporary occurring basic species
[
Scheme 5, Eq. (2)]. Thus, replacement of phenyl groups
pH Profiles
in PPh I by long-chain alkyl groups increases solubility
(
maximum yield).
4
and TOF), but at the expense of quat stability (and The reaction was run in a 20-mL jacketed titration vessel (Met-
rohm) maintained at 608C by a thermostat and equipped with a
magnetic stirrer. The temperature and pH were monitored by
means of a Pt-1000 sensor (Metrohm) and a pH electrode
(
Metrohm), respectively. The pH of the aqueous phase was
controlled by two (acid and base) motor-driven 1-mL burettes
GP-Titrino 736 and STAT-Titrino 718, Metrohm; smallest in-
(
crement 1 mL) interfaced to a personal computer, using the Ti-
TM
Net 2.4 software (Metrohm) for acquisition and control. The
titration procedure allowed for automatic (acid or base) set up
and hold of a given pH once a desired value had been defined.
The electrode was calibrated using standard buffers (pH 4, 7
and 10) at 608C, compensating for dpH/dT. Aqueous NaOH
and H SO were used to adjust and maintain the pH. The con-
2
4
centration was chosen so that the volume increase of the aque-
ous phase was less than 10% at the end of the reaction. t-
BuOOH was fed by means of a motor-driven 5-mL burette
(
DOSINO 700, Metrohm) interfaced to a computer (LIQUI-
NO 711, Metrohm). All burette tips were equipped with anti-
diffusion valves. In a typical run, the titration vessel was charg-
ed with deionized water (6 g), ethylbenzene (6.8 g, 64 mmol),
TEMPO (1 g, 6.4 mmol) and biphenyl (98.7 mg, 0.64 mmol).
Figure 8. ESR spectrum of a solution of t-BuOOH in ethyl-
benzene after addition of Bu NI revealing the presence of
4
.
t-BuOO radicals (g¼2.015).
Adv. Synth. Catal. 2004, 346, 554±560
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
559

FULL PAPERS
Hans-J. Kirner et al.
Stirrer and thermostat were turned on. Stirring was such that
the phase boundary of the two-phase system remained hori-
zontal (800 rpm). After the mixture had reached 608C, burette
tips for titration were introduced into the aqueous phase and
the titration procedure started. After the defined pH value
the solvent afforded N-alkoxy amine 8 as slightly yellow oil;
yield: 0.9 g (89%). Analysis required for C H NO (311.47):
1
8
33
3
C 69.41%, H 10.68%, N 4.50%; found: C 69.20%, H 10.76%,
1
N 4.42%; H NMR (400 MHz, CDCl ): d¼1.09 (t, J¼8 Hz,
3
3H), 1.10±1.26 (m, 17H), 1.52±1.57 (m, 3H), 1.74±1.84 (m,
4H), 2.03±2.05 (m, 2H), 2.28 (q, J¼8 Hz, 2H), 3.56±3.62 (m,
1H), 4.98±5.06 (m, 1H).
had been reached, Bu NI (23.6 mg, 0.064 mmol) was added fol-
4
lowed by introduction of the burette tip for t-BuOOH (70%
aqueous) feed into the organic phase. The reaction was started
with the addition of t-BuOOH (2.64 mL, 19.2 mmol) at a con-
stant rate (0.18 mL, 1.28 mmol/min). During all operations the
pH was kept constant. Samples were withdrawn and analyzed
by HPLC until conversion of nitroxide reached 100%. For the
References
[
1] F. Gugumus, in: Plastic Additives Handbook, 3rd edn.,
Eds.: R. G‰chter, H. M¸ller), Carl Hanser, M¸nchen,
990, pp. 190±195.
2] C. Hawker, A. Bosman, E. Harth, Chem. Rev. 2001, 101,
661±3688.
3] J. Galbo, M. Ackerman, (Ciba-Geigy Corp.), US Patent
,921,962, 1990; Chem. Abstr. 1990, 113, 152265.
pH profile of the CuCl -catalyzed reaction water (6 g) was re-
2
(
1
placed by 6 mL of aqueous 0.1 molar CuCl and Oct MeNCl as
2
3
phase-transfer catalyst (25.9 mg, 0.064 mmol).
[
[
[
3
Preparation of N-Alkoxy Amines 7 and 8
4
4] A. Hafner, H.-J. Kirner, F. Schwarzenbach, P.-A. Van der
Schaaf, P. Nesvadba, (Ciba Specialty Chemicals Inc.),
WO Patent 01/92228, 2001; Chem. Abstr. 2001, 136,
20022.
A stirred mixture of 7.3 g (32 mmol) propionic acid 2,2,6,6-tet-
ramethylpiperidin-4-yl N-oxide ester (2, R¼OCOC H ),
2
5
2
6.3 g (320 mmol) cyclohexene and 0.12 g (0.32 mmol) Bu NI
4
kept at 608C was treated with 6.2 g (48 mmol) aqueous t-
BuOOH (70%) within 25 minutes. The temperature was main-
tained until the red color of the nitroxide had disappeared
[5] M. Frey, V. Rast, (Ciba Specialty Chemicals Inc.), WO
Patent 03/45919, 2003; Chem. Abstr. 2003, 139, 22821.
[6] For a transfer rate limited reaction, TOF is expected to
increase with increasing Q: C. Starks, C. Liotta, M. Hal-
pern, Phase-Transfer Catalysis, Chapman & Hall, New
York, 1994, pp. 266±338.
(
5 min). The reaction mixture was cooled down to 258C and
stirred with aqueous 10% Na SO solution to destroy excess
2
3
t-BuOOH. The aqueous phase was split off and washed with cy-
clohexane. The combined organic phases were passed through
a plug of silica gel and brine-washed, dried over MgSO , fil-
4
[
7] G. Rothenberg, L. Feldberg, H. Wiener, Y. Sasson, J.
Chem. Soc. Perkin Trans. 2 1998, 2429±2434.
tered and the solvent distilled off on a rotary evaporator, af-
fording O-cyclohexenyl hydroxylamine 7 as slightly orange
oil; yield: 8.7 g (87.9%). Analysis required for C H NO
3
[
8] R. Sheldon, in: The Chemistry of Peroxides, (Ed.: S. Pa-
tai), John Wiley, New York, 1983, pp. 161±200.
18
31
(
1
8
309.45): C 69.87%, H 10.10%, N 4.53%; found: C 69.36%, H
1
[9] E. Napadensky, Y. Sasson, J. Chem. Soc. Chem. Com-
mun. 1991, 65±66.
0.03%, N 4.45%; H NMR (400 MHz, CDCl ): d¼1.12 (t, J¼
3
Hz, 3H), 1.20±1.26 (m, 12H), 1.52±1.58 (m, 4H), 1.73±2.1
[
[
10] D. Barton, V. Bloabec, J. Smith, Tetrahedron Lett. 1998,
(
5
m, 6H), 2.29 (q, J¼8 Hz, 2H), 4.23 (m, 1H), 5.05 (m, 1H),
3
9, 7483±7486.
.79±5.82 (m, 1H), 5.90±5.94 (m, 1H).
11] D. Stanisavljev, Ber. Bunsenges. Phys. Chem., 1997, 101,
A mixture of the above product (1 g, 3.19 mmol) and 0.17 g
1
036±1039.
Pd on charcoal (10%) in 30 mL hexane was hydrogenated at
58C and 4 bar of hydrogen. Filtration and evaporation of
2
560
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 554±560