Efficient iron-catalyzed n-demethylation of tertiary amine-N-oxides under oxidative conditions

Article abstract of DOI:10.1071/CH11320

An investigation into the influence of oxidative conditions on the efficiency of opiate N-demethylation using iron powder has been carried out under non-classical Polonovski conditions. This approach involves a two-step process of N-oxidation and subsequent treatment of the intermediate N-oxide hydrochloride with the redox catalyst. Significant improvements in rate and yield have been realized for these reactions in the presence of molecular oxygen. In this context, further rate enhancement was achieved by the judicious addition of small amounts of ferric ions, leading to a concomitant reduction in the amount of the zero-valent iron primary catalyst that is required. This has led to a generalized improved methodology for the N-demethylation of oripavine, codeine, morphine, and thebaine. This protocol can also be carried out in one-pot without the need to isolate the intermediate N-oxide.

Full text of DOI:10.1071/CH11320

CSIRO PUBLISHING
Aust. J. Chem. 2011, 64, 1515–1521
Full Paper
http://dx.doi.org/10.1071/CH11320
Efficient Iron-Catalyzed N-Demethylation of Tertiary
Amine-N-oxides under Oxidative Conditions
A
A B
,
Gaik B. Kok and Peter J. Scammells
A
Medicinal Chemistry and Drug Action, Monash Institute of Pharmaceutical Sciences,
Monash University, 381 Royal Parade, Parkville, Vic. 3052, Australia.
Corresponding author. Email: peter.scammells@monash.edu
B
An investigation into the influence of oxidative conditions on the efficiency of opiate N-demethylation using iron powder
has been carried out under non-classical Polonovski conditions. This approach involves a two-step process of N-oxidation
and subsequent treatment of the intermediate N-oxide hydrochloride with the redox catalyst. Significant improvements in
rate and yield have been realized for these reactions in the presence of molecular oxygen. In this context, further rate
enhancement was achieved by the judicious addition of small amounts of ferric ions, leading to a concomitant reduction in
the amount of the zero-valent iron primary catalyst that is required. This has led to a generalized improved methodology for
the N-demethylation of oripavine, codeine, morphine, and thebaine. This protocol can also be carried out in one-pot
without the need to isolate the intermediate N-oxide.
Manuscript received: 1 August 2011.
Manuscript accepted: 9 August 2011.
Published onilne: 16 November 2011.
Introduction
used methods. From an industrial perspective, the von Braun
reaction is undesirable due to the toxicity of the reagent whilst
chloroformate esters, though generally high-yielding, are
expensive reagents. A range of other methods have also been
employed including the use of dialkyl azodicarboxylates,[6] as
well as photochemical,[7] and biochemical transformations.[8]
Our own efforts in this area have primarily focussed on
Polonovski-type[9] N-demethylation on opiate and tropane
alkaloids. Thus, the tertiary N-methylamine is first converted
into the N-oxide, which is then treated with an activating agent,
generating the secondary amine. In most cases, the only signifi-
cant byproduct of the reaction is the parent N-methylamine.
Several Fe(II)-based reagents including FeSO4 ꢀ 7H2O,[10]
Fe(II)TPPS,[11] and ferrocene[12] have been shown to be effec-
tive activating agents. However, depending upon the quantity of
the Fe(II)-based reagent used, emulsions due to iron hydroxides
and related iron compounds have been found to be an operat-
ional issue especially for large scale reactions. There is therefore
a need for the development of improved processes for the
N-demethylation of opiates, particularly direct processes from
the corresponding naturally occurring N-methyl morphinanes.
Recently, under inert conditions, we reported the use of zero-
valent iron in Polonovski-type N-demethylation of several
opiates, including morphine, codeine methyl ether, thebaine,
and oripavine as well as the 14-hydroxy derivatives, oxycodei-
none, oxycodone, oxymorphinone, and oxymorphone.[13] Whilst
the yields of N-norcodeine methyl ether and N-northebaine
were excellent (.80 %), the N-demethylation of the other
opiate substrates proceeded in modest yield (,60 %) and were
accompanied by the formation of significant amounts of the
corresponding N-methyl byproduct (up to ,40 %). Subsequently,
under a wide range of reaction conditions, employing oripavine
Morphine (1a) and codeine (2a, Fig. 1) are the two most abun-
dant opiate alkaloids present in traditional strains of the opium
poppy, Papaver somniferum. In addition to their own well
known therapeutic applications, they have also proven to be
useful raw materials for the production of semi-synthetic opiate
pharmaceuticals, such as naltrexone (3), naloxone (4), and
buprenorphine (5).[1] These compounds are used in the therapy
of opiate and alcohol dependence and buprenorphine is also
used in pain management.[2] More recently, new strains of
Papaver somniferum have been developed which have a higher
content of the D6,D8-morphine alkaloids, oripavine (6a) and
thebaine (7a). These compounds are even more attractive sub-
strates for the preparation of buprenorphine and ‘nal salts’ such
as naltrexone and naloxone. Both thebaine and oripavine pos-
sess a diene moiety in their C ring which provides a useful
handle for further transformations, such as a Diels-Alder reac-
tion used in the synthesis of buprenorphine or the introduction
of 14-hydroxy functionality present in the nal salts. Another
important modification present in many semi-synthetic opiate
pharmaceuticals, is the replacement of the naturally occurring
N-methyl group with other alkyl substituents such as allyl
and cyclopropylmethyl. This necessitates the inclusion of an
N-demethylation and N-alkylation step(s) in the syntheses of
compounds of this type. Whilst the latter step is a relatively
simple operation, removal of the N-methyl group still presents a
challenge to the synthetic chemist,[3] particularly for large scale
operations.
There are several general methods for removal of the
N-methyl group described in the literature.[3] For many years,
the von Braun reaction[4] using cyanogen bromide and the use of
chloroformate esters[5] have been two of the most frequently
Journal compilation Ó CSIRO 2011
www.publish.csiro.au/journals/ajc

1516
G. B. Kok and P. J. Scammells
HO
MeO
HO
O
O
O
NR
NR
NR
OH
O
HO
HO
Morphine 1a (R ϭ Me)
Morphine-N-oxide HCl 1b
N-Normorphine 1c (R ϭ H)
Codeine 2a (R ϭ Me)
Codeine-N-oxide HCl 2b
N-Norcodeine 2c (R ϭ H)
Naltrexone 3 (R ϭϪCH2
Naloxone 4 (R ϭ allyl)
)
HO
HO
MeO
O
O
O
N
NR
NR
MeO
MeO
MeO
OH
Oripavine 6a (R ϭ Me)
Oripavine-N-oxide HCl 6b
N-Nororipavine 6c (R ϭ H)
Thebaine 7a (R ϭ Me)
Thebaine-N-oxide HCl 7b
N-Northebaine 7c (R ϭ H)
Buprenorphine 5
Fig. 1. Examples of opiate raw materials and semi-synthetic opiate pharmaceuticals.
markedly at temperatures above 258C.[15] To complicate matters
further, the solubility is also affected by the presence of even
small amounts of water. The situation in the case of binary
solvent mixtures can be even more complex.[15,16] Studies also
show variable temperature dependency with respect to other
relevant parameters – for example, studies conducted on the rate
of iron corrosion in acid media have found[17] that the kinetics
and adsorption processes at temperatures from 0 to 408C differ
markedly in character to those above 408C. Hence some reac-
tions conducted in air might not conform to Arrhenius kinetics.
Due to such vagaries, systematic temperature dependent inves-
tigations, in an attempt to increase the reaction rates of our
reactions, have not been attempted. We have also recognized
that there are other complicating factors relating to the influence
of dissolved oxygen in these reactions. For example, whilst it is
well known that the primary reaction of Fe(0) with dissolved
oxygen is oxidation to Fe2þ, dissolved oxygen also leads to the
passivation of the Fe(0) surface by corrosion products.[18]
However, in zero-valent iron processes, little is known about
the kinetic chemistry of the iron oxide coatings.[18–21] On the
other hand, the kinetics of the oxidation of Fe2þ by O2 have been
extensively studied over time.[22–25] Notably, at ordinary tem-
perature and pressure, in both acid and neutral solutions, ferrous
salts are not easily oxidized in both air and oxygen.[23–25] For
example, it has been reported[24] that only 0.03 % of a 0.1 M
ferrous sulfate solution was oxidized after 1 h at 258C, with the
rate doubling with a 108C rise of temperature. The rate of Fe2þ
oxidation has also been found to be dependent on the counterion,
with the chloride salt being oxidized even slower than the
sulfate.[22,26] A range of other factors such as the concentration
of the ferrous salt, pH, temperature, pressure, added cations,
complexing agents or ligands, and catalysts also affect the
oxidation rate of the Fe2þ species.[27–32]
Hϩ
Fe(II)
ϩ
ϩ
ϩ
O
OH
OH
Fe(II)
N
N
N
CH3
CH3
CH3
Hϩ
Fe(II)Fe(III)
ϩ
N
ϪHϩ
ϩ
N
CH3 ϩ Fe(III) ϩ H2O
N
CH2
CH2
Fe(II)
hydrolysis
Fe(III)
N
H
O
CH2
N
CH3
Scheme 1. Proposed mechanism of the Fe2þ/Fe3þ-mediated non-classical
Polonovski reaction.
as a model substrate,[14] we found that the yield of N-nororipavine
improved when stainless steel 303-L powder was substituted for
iron powder as catalyst. However, the reactions took a consi-
derably longer time to reach completion.
Herein, we describe significantly improved reaction
outcomes – higher yields and/or shorter reaction time, by
conducting the N-demethylation under oxidative conditions.
Since the mechanism[9] of the non-classical Polonovski reaction
is thought to proceed via a series of single electron transfers
(Scheme 1), we have considered the role that molecular oxygen
might play in these reactions. However, systematic control of
the rate via temperature is complicated by the diverse solubility
profiles of oxygen in various solvents.[15,16] For several organic
solvents such as carbon tetrachloride, chlorobenzene and
chlorocyclohexane, the temperature coefficient is positive,
i.e. the oxygen solubility increases with temperature; and for
solvents such as methanol, ethanol and propanol, the oxygen
solubility as a function of temperature is quite complex.[15,16]
For example, for ethanol itself, whilst oxygen solubility is
essentially independent of temperature at first, it decreases
In the present study, we have found that the above problems
may be addressed, in part, by doping with ferrous or ferric ions –
affording more control and, indeed, addition of small amounts
of ferric ions, in particular, is found to further increase the

N-Demethylation of Tertiary Amine-N-oxides
1517
Table 1. N-demethylation of oripavine-N-oxide hydrochloride with
of iron powder used to 50 mol-% was inconsequential, both
with respect to time for completion and isolated yield of
N-nororipavine (compare entries 1 and 3). A further reduction
in the amount of catalyst to 25 mol-% significantly slowed the
reaction, but the yield of N-nororipavine improved slightly
(entry 5). In all cases, the reactions conducted in air gave
superior yields of N-nororipavine (6c) compared with those
performed under an atmosphere of nitrogen.[14] The difference
was greatest when 200 mol-% of iron powder was used;
N-nororipavine (6c) was isolated in 67 % yield when the reac-
tion was conducted in air and in 40 % yield with an inert
atmosphere of nitrogen (compare entries 1 and 2). Also, the
percentage total product recovery (6a þ 6c), via column chro-
matography, for all air experiments were also found to be
significantly higher than the corresponding experiments con-
ducted under nitrogen. For example, the experiment conducted
in air with 200 mol-% of iron gave a 91 % total product recovery,
while the total recovery from the N2 experiment was 64 %.
Furthermore, when 50 and 200 mol-% of iron powder were used
(entries 1–4), the reactions proceeded to completion three times
faster in the presence of air.
Subsequently, we were interested in the effects that solvents
may have on the reaction. In line with our previous inert
atmosphere studies,[14] we have chosen to use a i-PrOH/CHCl3
binary system for these experiments, varying the proportion of
solvents in the mixture. These results are given in Table 1
(entries 7–12). All experiments were conducted using 50 mol-%
of iron powder. Interestingly, N-demethylation was generally
faster for reactions in a mixed i-PrOH/CHCl3 solvent system
than in isopropanol alone. Varying the proportion of chloroform
in the mixture was also found to influence the outcome of the
experiments (entries 7, 9, and 11). Of the three different i-PrOH/
CHCl3 solvent systems investigated, the reactions in 1:1 and 1:3
i-PrOH/CHCl3 gave comparable yields of N-nororipavine (72
and 74 %, respectively) and both progressed to completion in a
reasonable timeframe (,4 h). The reaction conducted in 1:9
i-PrOH/CHCl3 gave a lower yield of N-nororipavine (60 %) and
took more than four times longer to reach completion (16 h). In
all three mixed i-PrOH/CHCl3 solvent systems the isolated
yields of N-nororipavine (6c) for the air experiments were
comparable to the corresponding reactions performed under
an atmosphere of nitrogen. However, the reactions conducted
in air typically proceeded to completion in significantly less
time. The difference in reaction rate was most apparent in the
reactions conducted in i-PrOH/CHCl3 (1:9) in which the reac-
tion in air was complete in one tenth of the time required under
nitrogen (compare entries 11 and 12).
iron powder
HO
HO
ClϪ
Fe powder
O
O
Solvent,
air/O2
ϩ
CH3
N
NR
OH
MeO
MeO
6b
6a R ϭ H
6c R ϭ Me
Entry
FeA
ConditionsB
Atm
Time
[h]
6cC
6aC
[mol-%]
[%] [%]
1
200
200
50
50
25
25
50
50
50
50
50
50
i-PrOH
Air
N2
16
48
67
40
67
53
73
65
72
71
74
70
60
58
24
24
31
25
26
25
25
25
24
25
22
23
2D
i-PrOH
3
i-PrOH
Air
N2
16
4D
5
i-PrOH
48
i-PrOH
Air
N2
120
144
6D
i-PrOH
7
i-PrOH/CHCl3 (1:1)
i-PrOH/CHCl3 (1:1)
i-PrOH/CHCl3 (1:3)
i-PrOH/CHCl3 (1:3)
i-PrOH/CHCl3 (1:9)
i-PrOH/CHCl3 (1:9)
Air
N2
2.5
8D
9
3.5
2.0
5.5
Air
N2
10D
11
12D
Air
N2
16
168
AParticle size distribution (sieve analysis) for 99 % of sample: 100 mesh.
B10 mL of solvent per 100 mg of substrate at 408C.
CIsolated via column chromatography.
DData from reference [14] included for comparison.
N-demethylation rate. As an overall strategy, in the first instance
several experiments were conducted in air and these were then
benchmarked against the results obtained for the inert atmo-
sphere experiments. Subsequent procedural refinements in the
form of added Fe(II) and Fe(III) were then conducted.
We have also found that N-demethylation reactions of this
type can be conducted in one-pot without isolation of the
intermediate N-oxide hydrochloride and the results for these
one-pot procedures are also presented.
Results and Discussion
A study on the N-demethylation in air was initiated by con-
ducting a series of experiments using oripavine (6a) in which the
effect of iron powder loading on the reaction in isopropanol was
investigated. Although consistently high yields of the N-oxides
of such N-methyl alkaloids have previously been obtained via
N-oxidation with the more environmentally benign hydrogen
peroxide,[10a] in this study, we have employed m-CPBA for this
purpose since it is a more effective reagent – with reactions
typically taking only 10 min to complete at ꢁ58C. Thus the
N-oxide of oripavine, isolated as the hydrochloride salt 6b, was
obtained in near quantitative yield. A heterogenous mixture of
6b and iron powder in isopropanol was stirred in air until
complete consumption of the starting material (as analyzed by
TLC). Subsequent column purification, eluting with a gradient
of CHCl3/MeOH (24:1–17:3) isolated the desired N-nor product
6c from some tertiary N-methylamine 6a.
We have shown thatairgenerally acceleratesN-demethylation,
which broadly implies that the N-demethylation is highly
dependent on the concentration of ferrous and ferric species,
consistent with the proposed mechanism[9] for iron-mediated
N-demethylation under Polonovski-type conditions. Using ori-
pavine, we have therefore conducted a more systematic study on
the influence of added Fe2þ and Fe3þ in such reactions. Since the
rate of oxidation of ferrous ion by oxygen is dependent on the
nature of the anions present,[22,26,32] and given that the hydro-
chloride salt of the N-oxide 6b has been used as substrate, we
have employed iron chlorides in these investigations. All reac-
tions with added Fe2þ and Fe3þ were conducted at ambient
temperature using 50 mol-% of the primary redox catalysts in
1:3 i-PrOH/CHCl3.
The results of the above experiments are shown in Table 1.
Thus, employing 200 mol-% of iron powder at 408C, the reaction
in air was complete in 16 h (entry 1). Lowering the amount
Table 2 summarizes the results for the reactions with added
Fe2þ. In all cases, addition of Fe2þ to the iron powder

1518
G. B. Kok and P. J. Scammells
experiments reduces the ratio of N-nororipavine relative to
oripavine (6c : 6a) without having a significant effect on the
total product recovery (6c þ 6a). This finding is in accordance
with the proposed mechanism of the non-classical Polonovski
reaction,[2] which is believed to proceed via an intermediate
aminium radical cation that is reduced by Fe2þ to form the
corresponding parent N-methylamine (in this case, oripavine).
In air, the time taken for the reaction to reach completion was
found to be progressively shortened as the amount of added Fe2þ
increased (compare entries 1, 2, and 4). However, this has
occurred with a concomitant decrease in the amount of
N-nororipavine produced. Notwithstanding, all reactions with
added Fe2þ proceeded to completion in a shorter time in air than
the corresponding reactions under nitrogen.
Experiments with various amounts of added Fe3þ are also
summarized in Table 2. The addition of 2.5 mol-% of FeCl3 had
no major effect on the yield of N-nororipavine (6c) or on the ratio
of 6c : 6a (compare entries 1 and 6). However, the addition of
2.5 mol-% of FeCl3 increased the reaction rate by a factor of 2–3
fold. Significantly higher loadings of FeCl3 (entry 7) had a
deleterious effect on both the yield of N-nororipavine (6c) and
the rate of reaction. Interestingly, when 2.5 mol-% of FeCl3 was
added to the iron powder reaction, conducted under an atmo-
sphere of nitrogen, almost identical outcomes in terms of the
yield of 6c, the product ratio, and the reaction time, were
obtained – compared with the corresponding reaction in air
(compare entries 1 and 10). This suggests that improvements
observed for the reaction conducted in air relative to those under
an inert atmosphere may result from the oxygen-mediated
formation of Fe3þin situ.
In the presence of oxygen, further reducing the amount of
iron powder and FeCl3 employed to 13 mol-% and 2 mol-%,
respectively, N-demethylation of oripavine was complete after
6 h at ambient temperature. After standard workup followed by
column chromatography, N-nororipavine was obtained in 82 %
yield (Table 3, entry 1). Hence, the addition of only a small
amount of Fe3þ to the reaction has significantly reduced the total
amount of iron used without compromising the yield.
As entry 2 shows, the optimized conditions for oripavine also
successfully N-demethylated morphine. In the case of mor-
phine, the reaction was conducted at 408C to enable completion
within a reasonable timeframe of 48 h. After standard workup, a
77 % yield of N-normorphine was obtained, a significant
improvement over that previously[13] reported when iron pow-
der alone was employed as redox catalyst.
There have been at least eight previous attempts at the direct
synthesis of N-norcodeine from codeine, with reported yields
ranging from 10–80 %. To the best of our knowledge, two of the
previous best reported yields of 70 and 80 % have employed
toxic and/or expensive chlorofomate esters.[33,34] Gratifyingly,
we have found that simply stirring the N-oxide hydrochloride of
codeine with a combination of Fe(0) and Fe3þ at ambient
temperature for 36 h has provided, after standard workup,
N-norcodeine in high yield (82 %) (entry 3).
The N-demethylation of thebaine was explored using both
chloroform and isopropanol (entries 4 and 5), with the reaction
being greatly facilitated in chloroform. In both cases, addition of
small amounts of Fe3þ has reduced by 10-fold the amount of iron
powder used from 130 mol-%,[13] without the reaction time or
the yield of the N-nor product 7c being compromised. Reducing
the amount of iron used in these reactions is particularly
advantageous to the efficacy of workup.
Table 2. N-demethylation of oripavine-N-oxide hydrochloride with
iron powder and added iron saltsA
Entry
Catalyst
[mol-%]
Iron salt,
[mol-%]
Atm
Time
[h]
6cB
[%]
6aB
[%]
1
Fe(0)(50)
Fe(0)(50)
–
–
Air
Air
Air
Air
Air
Air
Air
Air
N2
3.5
1.5
168
,1
6
75
68
42
55
46
77
42
36
70
78
22
21
8
Finally, it is evident that by carefully controlling the amount
of oxidant in the reaction, the two-step N-oxidation and
N-demethylation protocol can be reduced to a one-step process
(Table 4). The one-pot process is especially advantageous for
unstable or highly water soluble N-oxide hydrochloride
intermediates.
2
3C
Fe2þ, 2.5
Fe2þ, 2.5
Fe2þ, 50
Fe2þ, 50
Fe3þ, 2.5
Fe3þ, 50
Fe3þ, 50
–
4
Fe(0)(50)
–
43
37
19
28
11
25
20
5
6
Fe(0)(50)
Fe(0)(50)
–
1.2
16
7
8
168
5.5
3.5
Conclusion
9
Fe(0)(50)
Fe(0)(50)
10
Fe3þ, 2.5
N2
Significantly improved reaction conditions for the
N-demethylation of opiate alkaloids have been developed. We
have identified reaction conditions which result in a dramatic
improvement in the methodology for the N-demethylation of
oripavine at ambient temperature. A scrutiny of the oxidative
AExperiments were performed with 100 mg of substrate in 10 mL of i-PrOH/
CHCl3 (1:3) at room temperature.
BIsolated via column chromatography.
C41 % of N-oxide hydrochloride 6b was also recovered.
Table 3. N-demethylation of various opiate-N-oxide hydrochlorides with iron powder and added ferric chloride hexahydrateA
Entry
Substrate(N-oxide)
Solvent
Conditions
#cB[%]
#aB[%]
1
Oripavine 6b
Morphine 1b
Codeine 2b
Thebaine 7b
Thebaine 7b
i-PrOH/CHCl3 (1:3)C
i-PrOH/CHCl3 (1:3)C
i-PrOH/CHCl3 (1:3)C
CHCl3
RT, 6 h
82
77
82
85
79
16
19
14
10
8
2D
408C, 48 h
RT, 36 h
RT, 1.5 h
RT, 48 h
3
4
5
i-PrOH
A100 mL of substrate in 10 mL of solvent, Fe powder (13 mol-%), FeCl3 (2 mol-%).
BUnless otherwise indicated, products were isolated via column chromatography.
Ci-PrOH/CHCl3 (1:3) was used for solubility reasons.
DProducts were isolated via extraction (Method C; see Experimental).

N-Demethylation of Tertiary Amine-N-oxides
1519
conditions that pertain to these reactions carried out in air,
compared with those carried out under an inert atmosphere, has
enabled such processes to be optimized in terms of their rate and
yield. Notably, the conditions optimized for the model substrate
oripavine are applicable to codeine, morphine, and thebaine.
We have also found that the two-step N-oxidation and
N-demethylation protocol can be reduced to a one-pot process.
In this way, the rather tedious isolation process of the highly
water soluble N-oxide hydrochloride intermediates may be
avoided.
Method A
The crude reaction mixture was concentrated and the remain-
ing residue was purified via column chromatography on silica
gel, eluting with a CHCl3/MeOH gradient which isolated the
N-nor product from the tertiary N-methylamine, both obtained
as the corresponding hydrochloride salt.
Method B
The crude reaction mixture was diluted with CHCl3 (30 mL)
and the resulting solution was washed with 5 % aqueous NaOH
(1 mL ꢂ 2), dried, filtered, and the residue purified via column
chromatography on silica gel.
Experimental
General
Method C
Iron powder was obtained from Ho¨gana¨s Sweden. m-
Chloroperbenzoic acid, ferrous chloride tetrahydrate, and ferric
chloride hexahydrate were purchased from Sigma-Aldrich
(St. Louis, MO). For experiments that were conducted in air,
solvents (Merck) were used as supplied. For reactions conducted
under nitrogen, solvents were degassed before use: sonicated
under vacuum for 10 min and then back-filled with nitrogen.
Thin-layer chromatography (TLC) was performed with 0.25 mm
TLC silica gel 60F aluminium plates with 254 nm fluorescent
indicator and plates were visualized using both UV light and
molybdate stain. 1H and 13C NMR spectra were recorded at 400
and 100 MHz, respectively. Chemical shifts (d ppm) were
referenced with solvent residual peaks. Coupling constants are
given in hertz. High resolution mass spectra were obtained using
a Waters LCT Premier XE time-of-flight mass spectrometer
fitted with an electrospray (ESI) ion source and controlled with
MassLynx software version 4.5. All substrate N-oxides, were
prepared via treatment of the opiate with m-chloroperbenzoic
acid and isolated as the corresponding hydrochloride salt,
according to the literature.[14] In all cases, near quantitative
yields of the N-oxide hydrochlorides were obtained. N-oxide
hydrochloride 6b isolated with ,0.5 molar equivalent of
isopropanol, was employed in all experiments.
The crude reaction mixture was concentrated to dryness. To
the residue was added H2O (30 mL) and the pH of the mixture
was adjusted to 8 (conc. NH4OH) and subsequently extracted
with i-PrOH/CHCl3 (1:6) to remove the tertiary N-methylamine.
The aqueous layer was concentrated to dryness and the residue
was extracted with i-PrOH/CHCl3 (1:3) to isolate the N-nor
product. A few drops of 6N HCl were added to these extracts.
Subsequent removal of the volatiles gave the N-nor compound
as the hydrochloride salt.
The above procedure was also performed with either added
FeCl2 ꢀ 4H2O or FeCl3 ꢀ 6H2O (Tables 2 and 3).
General Procedure for One-Pot N-Demethylation
To a stirred solution of the opiate (see Table 4) (0.673 mmol)
in solvent (20 mL) at ꢁ58C was added m-CPBA (max 77 %
reagent, 0.670 mmol) portionwise over 10 min. The solution was
then left to stir for 15 min and concentrated HCl (60 mL) was
added dropwise. The solution was warmed to room temperature,
iron powder (4.9 mg, 13 mol-%) and FeCl3 ꢀ 6H2O (4.5 mg,
2 mol-%) were added, and the mixture was left to stir until
complete consumption of N-oxide hydrochloride (as analyzed
by TLC analysis).
N-Norcodeine and N-northebaine were isolated from the
reaction mixture according to Method B above. N-Nororipavine
was isolated as follows: the reaction mixture was concentrated
to dryness. H2O (30 mL) was added and the resulting solution
was extracted with CHCl3 (10 mL ꢂ 3). The aqueous layer was
concentrated to dryness and the resulting residue was purified
via column chromatography according to Method A above.
N-Normorphine was isolated according to Method C above,
with the exception that the aqueous solution was extracted
with CHCl3 (10 mL ꢂ 3) before basification to pH 8 with
concentrated ammonia.
General Procedure for N-Demethylation
To a stirred solution of the N-oxide hydrochloride (100 mg) in
solvent (10 mL) was added iron powder (varying amounts; see
Table 1). The reaction mixture was then stirred at the specified
temperature until complete consumption of starting material
(via TLC analysis), and purified by Method A, B, or C.
Table 4. One-pot N-demethylation of various opiates with iron
powder and added ferric chloride hexahydrateA
Entry
Substrate
ConditionsB
#cC
[%]
#aC
[%]
N-Normorphine 1c (HCl Salt)
Purified via Method C to give the hydrochloride salt of 1c as
an off-white solid. A small sample of N-normorphine hydro-
chloride was dissolved in H2O and the pH of the solution was
adjusted to 8–9 with concentrated ammonia. The precipitate was
isolated via filtration to give N-normorphine as an off-white
solid, mp 272–2768C (lit.[35] mp 275–2778C dec); [a]2D4 ꢁ54
(c 1.0, 10 % HOAc in H2O); for the hydrochloride of 1c:
1H NMR (D2O) d 6.82 (d, J ¼ 8.0, 1H), 6.73 (d, J ¼ 8.0, 1H),
5.82–5.77 (m, 1H), 5.43 (ddd, J ¼ 2.8, 2.8, and 9.6, 1H),
5.10 (dd, J ¼ 1.2 and 6.4, 1H), 4.43–4.36 (m, 2H), 3.44–3.37
(m, 1H), 3.23–2.94 (m, 4H), 2.32 (ddd, J ¼ 4.8, 14.0, and
14.0, 1H), 2.24–2.17 (m, 1H); 13C NMR (D2O/CF3CO2D)
1
Oripavine 6a
Morphine 1a
Codeine 2a
Thebaine 7a
RT/36 h
408C/48 h
RT/20 h
RT/48 h
82
79
87
86
16
19
7
2D
3
4E
11
ASubstrate (0.673 mmol), m-CPBA (0.670 mmol) then Fe powder
(13 mol-%) and FeCl3 (2 mol-% of FeCl3) in i-PrOH/CHCl3 (1:3) (20 mL),
unless otherwise indicated.
BReaction time after addition of Fe(0)/Fe3þ
CProducts were isolated via column chromatography, unless otherwise
.
indicated.
DProducts were isolated via extraction (see Experimental, Method C).
EChloroform was the solvent used.

1520
G. B. Kok and P. J. Scammells
d 145.4, 137.8, 133.0, 129.3, 125.7, 123.4, 120.3, 117.4, 90.6,
65.5, 51.5, 42.0, 37.1, 36.7, 31.6, 25.7; MS (ESI) m/z 272
[MþH]þ; HRMS C16H18NO3 calcd for [MþH]þ 272.1281,
found 272.1286.
Acknowledgements
The authors thank the ARC Centre of Excellence for Free Radical Chemistry
and Biotechnology for financial support.
References
N-Norcodeine 2c
[1] S. Bere´nyi, C. Csutora´s, A. Sipos, Curr. Med. Chem. 2009, 16, 3215.
doi:10.2174/092986709788803295
[2] D. S. Fries, in Foye’s Principles of Medicinal Chemistry, (5th ed., Eds
D. A. Williams, T. L. Lemke), 2002, pp. 453–479 (Lippincott William &
Wilkins: Philadelphia).
The crude reaction mixture from N-demethylation of the
N-oxide hydrochloride 2b was purified via Method A using a
gradient of CHCl3/MeOH (40:3–8:1) to give the hydrochloride
salt of 2c as an off-white solid. The hydrochloride was dis-
solved in 5 % aqueous NaOH (10 mL) and the solution was
extracted with CHCl3 (10 mL ꢂ 3). The extracts were combined,
dried (Na2SO4), filtered, and concentrated to give the title
compound as a white solid. Data for 2c are consistent with the
literature[10a]; mp 183–1848C (MeOH/Et2O); [lit.[8b] mp 182–
1838C]; [a]2D3 ꢁ120 (c 1.0, CHCl3); [lit.[8b] [a]2D8 ꢁ90.9 (c 0.224,
CHCl3)]; 1H NMR (CDCl3) d 6.69 (d, J ¼ 8.0, 1H), 6.59
(d, J ¼ 8.0, 1H), 6.73 (dddd, J ¼ 1.6, 1.6, 2.8, and 9.6, 1H),
5.28 (ddd, J ¼ 2.8, 2.8, and 9.6, 1H), 4.88 (dd, J ¼ 1.2 and 6.4,
1H), 4.22–4.17 (m, 1H), 3.86 (s, 3H), 3.67–3.63 (m, 1H),
3.05–2.77 (m, 4H), 2.62–2.57 (m, 1H), 1.98–1.87 (m, 2H),
1.62 (br s, 2H); 13C NMR (CDCl3) d 146.5, 142.0, 133.7,
131.1, 128.1, 127.4, 119.4, 112.9, 92.2, 66.4, 56.3, 51.7, 43.8,
41.2, 38.3, 36.5, 31.1; MS (ESI) m/z 286 [MþH]þ; HRMS
C17H20NO3 calcd for [MþH]þ 286.1438, found 286.1449.
[3] S. Thavaneswaran, K. McCamley, P. J. Scammells, Nat. Prod. Com-
mun. 2006, 1, 885.
[4] J. von Braun, Ber. Dtsch. Chem. Ges. 1900, 33, 1438. doi:10.1002/
CBER.19000330208
[5] (a) See for example: C. Csutora´s, A. Zhang, J. M. Bidlack, J. L.
Neumeyer, Bioorg. Med. Chem. 2004, 12, 2687. doi:10.1016/J.BMC.
2004.03.011
(b) G. Kraiss, K. Nador, Tetrahedron Lett. 1971, 12, 57. doi:10.1016/
S0040-4039(01)96358-0
(c) J. H. Cooley, E. J. Evain, Synthesis 1989, 1. doi:10.1055/S-1989-
27129
(d) M. M. Abdel-Monem, P. S. Portoghese, J. Med. Chem. 1972, 15,
208. doi:10.1021/JM00272A025
(e) R. A. Olofson, J. T. Martz, J. P. Senet, M. Piteau, T. Malfroot,
J. Org. Chem. 1984, 49, 2081. doi:10.1021/JO00185A072
[6] (a) See for example: L. S. Schwab, J. Med. Chem. 1980, 23, 698.
doi:10.1021/JM00180A028
(b) H. Merz, K.-H. Pook, Tetrahedron 1970, 26, 1727. doi:10.1016/
S0040-4020(01)93023-6
N-Nororipavine 6c (HCl Salt)
[7] J. A. Ripper, E. R. T. Tiekink, P. J. Scammells, Bioorg. Med. Chem.
Lett. 2001, 11, 443. doi:10.1016/S0960-894X(00)00690-9
[8] (a) K. Kieslich, in Microbial Transformations of Non-Steroid Cyclic
Compounds, 1976, pp. 213–215 (Thieme: Stuttgart)
(b) A. Chaudhary, H. Leisch, A. Moudra, B. Allen, V. De Luca, D. P.
Cox, T. Hudlicky, Collect. Czech. Chem. Commun. 2009, 74, 1179.
doi:10.1135/CCCC2009025
Purified via Method A using a gradient of CHCl3/MeOH
(24:1–17:3) to give the hydrochloride salt of 6c as an off-white
solid, mp .2008C dec; [a]2D4 ꢁ188 (c 1.0, 10 % HOAc in H2O);
1H NMR (D2O) d 6.78 (d, J ¼ 8.0, 1H), 6.74 (d, J ¼ 8.0, 1H),
5.90 (d, J ¼ 6.4, 1H), 5.50 (s, 1H), 5.25 (d, J ¼ 6.4, 1H), 4.61
(d, J ¼ 6.0, 1H), 3.64 (s, 3H), 3.46–3.22 (m, 4H), 2.31 (ddd,
J ¼ 6.0, 13.3, and 13.3, 1H), 2.08–2.01 (m, 1H); 13C NMR
(D2O/CF3CO2D) d 153.0, 142.8, 138.6, 132.0, 124.4, 124.2,
120.8, 117.3, 117.0, 96.1, 87.7, 55.1, 53.2, 44.6, 37.0, 33.5, 33.1;
MS (ESI) m/z 284 [MþH]þ; HRMS C17H18NO3 calcd for
[MþH]þ 284.1281, found 284.1287.
[9] D. Grierson, Org. React. 1990, 39, 85.
[10] (a) K. McCamley, J. A. Ripper, R. D. Singer, P. J. Scammells, J. Org.
Chem. 2003, 68, 9847. doi:10.1021/JO035243Z
(b) S. Thavaneswaran, P. J. Scammells, Bioorg. Med. Chem. Lett.
2006, 16, 2868. doi:10.1016/J.BMCL.2006.03.017
[11] (a) Z. Dong, P. J. Scammells, J. Org. Chem. 2007, 72, 9881.
doi:10.1021/JO071171Q
(b) G. B. Kok, T. D. Ashton, P. J. Scammells, Adv. Synth. Catal. 2009,
351, 283. doi:10.1002/ADSC.200800632
[12] G. B. Kok, P. J. Scammells, Bioorg. Med. Chem. Lett. 2010, 20, 4499.
doi:10.1016/J.BMCL.2010.06.031
[13] G. B. Kok, C. C. Pye, R. D. Singer, P. J. Scammells, J. Org. Chem.
N-Northebaine 7c
Purified via Method B using a gradient of CHCl3/MeOH/
NH4OH (90:10:1–85:15:1) to afford 7c as an off-white solid.
Physical data for 7c are consistent with the literature[10a]; mp
146–1508C [(lit.[36] mp 157–1588C)]; [a]2D4 ꢁ225 (c 0.82, 5 %
MeOH in CHCl3) [lit.[8b] [a]2D0 ꢁ197.9 (c 0.31, 5 % MeOH in
CHCl3)]; [lit.[37] [a]D23 ꢁ200 (c 0.1, 5 % MeOH in CHCl3)];
1H NMR (CDCl3) d 6.68 (d, J ¼ 8.4, 1H), 6.62 (d, J ¼ 8.4, 1H),
5.51 (d, J ¼ 6.4, 1H), 5.27 (s, 1H), 5.03 (d, J ¼ 6.4, 1H), 3.92
(dd, J ¼ 1.6 and 6.0, 1H), 3.86 (s, 3H), 3.61 (s, 3H), 3.25–3.07
(m, 3H), 2.94 (dd, J ¼ 4.0 and 13.6, 1H), 2.40–2.20 (br, 1H),
2.08 (ddd, J ¼ 5.2, 12.8, and 12.8, 1H), 1.85–1.80 (m, 1H);
13C NMR (CDCl3) d 152.4, 144.7, 142.7, 133.9, 133.3, 127.7,
119.1, 112.8, 110.0, 95.7, 89.1, 56.3, 54.8, 53.8, 46.6, 40.6, 38.4,
37.6; MS (ESI) m/z 298 [MþH]þ; HRMS C18H20NO3 calcd for
[MþH]þ298.1438, found 298.1432.
2010, 75, 4806. doi:10.1021/JO1008492
[14] G. B. Kok, P. J. Scammells, Org. Biomol. Chem. 2011, 9, 1008.
doi:10.1039/C0OB01021A
[15] S. A. Shchukarev, T. A. Tolmacheva, Zh. Strukt. Khim. 1968, 9, 21.
[16] K. Fischer, W. Wilken, J. Chem. Thermodyn. 2001, 33, 1285.
doi:10.1006/JCHT.2001.0837
[17] A. K. Mindyuk, Mater. Sci. 1977, 12, 164. doi:10.1007/BF00723747
[18] T. C. Zhang, Y. H. Huang, Water Res. 2006, 40, 2311. doi:10.1016/
J.WATRES.2006.04.026
[19] (a) Y. H. Huang, T. C. Zhang, P. J. Shea, S. D. Comfort, J. Environ.
Qual. 2003, 32, 1306. doi:10.2134/JEQ2003.1306
(b) T. Satapanajaru, P. J. Shea, S. D. Comfort, Y. Roh, Environ. Sci.
Technol. 2003, 37, 5219. doi:10.1021/ES0303485
(c) D. Mishra, J. Farrell, Environ. Sci. Technol. 2005, 39, 645.
doi:10.1021/ES049259Y
Accessory Publication
1H, 13C NMR, and HR-MS spectra of 1c, 2c, 6c, and 7c are
available from the Journal’s website.
[20] Y. H. Huang, T. C. Zhang, Water Res. 2005, 39, 1751. doi:10.1016/
J.WATRES.2005.03.002

N-Demethylation of Tertiary Amine-N-oxides
1521
[21] X. Fan, X. Guan, J. Ma, H. Al, J. Environ. Sci. 2009, 21, 1028.
doi:10.1016/S1001-0742(08)62378-5
[30] F. J. Millero, Geochim. Cosmochim. Acta 1985, 49, 547. doi:10.1016/
0016-7037(85)90046-8
[22] R. Thomas, E. T. Williams, J. Chem. Soc. Trans. 1921, 119, 749.
doi:10.1039/CT9211900749
[31] B. Morgan, O. Lahav, Chemosphere 2007, 68, 2080. doi:10.1016/
J.CHEMOSPHERE.2007.02.015
[23] A. B. Lamb, L. W. Elder, Jr, J. Am. Chem. Soc. 1931, 53, 137.
doi:10.1021/JA01352A019
[24] K. A. Kobe, W. Dickey, Ind. Eng. Chem. 1945, 37, 429. doi:10.1021/
IE50425A014
[32] M. C. Lo´pez, M. A. Gallardo, J. S. Urieta, G. C. Losa, J. Chem. Eng.
Data 1987, 32, 472. doi:10.1021/JE00050A027
[33] Y Gao, T. N. Traino, P Vouros, J. L. Neumeyer, J. Label. Compd.
Radiopharm. 1988, 25, 293.
[25] C. Baskerville, R. Stevenson, J. Am. Chem. Soc. 1911, 33, 1104.
doi:10.1021/JA02220A011
[34] J. T. M Linders, R. J. Booth, T. S. Lie, A. P. G. Kieboom, L Maat, Recl.
Trav. Chim. 1989, 108, 189.
[26] J. W. McBain, J. Phys. Chem. 1901, 5, 623. doi:10.1021/
J150036A001
[27] W. Stumm, G. F. Lee, Ind. Eng. Chem. 1961, 53, 143. doi:10.1021/
IE50614A030
[35] K. C. Rice, J. Org. Chem. 1975, 40, 1850. doi:10.1021/JO00900A044
[36] J. R. Bartels-Keith, J. Chem. Soc. C 1966, 617. doi:10.1039/
J39660000617
[37] K. M. Madyastha, G. V. B. Reddy, G. R. Sridhar, Indian J. Chem. 2000,
[28] R. T. Lowson, Chem. Rev. 1982, 82, 461. doi:10.1021/CR00051A001
[29] E. J. Roekens, R. E. van Grieken, Mar. Chem. 1983, 13, 195.
doi:10.1016/0304-4203(83)90014-2
39B, 377.