Studies on the oxidative N-demethylation of atropine, thebaine and oxycodone using a FeIII-TAML catalyst

Article abstract of DOI:10.1039/c3gc41972j

The reaction pathway and selectivity of the oxidative N-demethylation of the alkaloid atropine with H₂O₂ using a Fe ^III-TAML catalyst has been investigated. The conversion of atropine in ethanol with aqueous H₂O

Full text of DOI:10.1039/c3gc41972j

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Studies on the oxidative N-demethylation of
atropine, thebaine and oxycodone using a
Cite this: Green Chem., 2014, 16,
1399
FeIII-TAML catalyst†
Duy D. Do Pham, Geoffrey F. Kelso, Yuanzhong Yang and Milton T. W. Hearn*
The reaction pathway and selectivity of the oxidative N-demethylation of the alkaloid atropine with H2O2
using a FeIII-TAML catalyst has been investigated. The conversion of atropine in ethanol with aqueous
H2O2 produces noratropine as the main product and N-formyl-noratropine and other atropine derivatives
involving carbon-hydroxylated tropane species as minor by-products. Comparative reactivity studies with
noratropine, N-formyl-noratropine and atropine N-oxide demonstrated that the FeIII-TAML catalyses the
N-demethylation of atropine by a biomimetic oxidation pathway involving the formation and then
decomposition of a N-hydroxymethylnoratropine intermediate. The reaction selectivity for atropine N-
demethylation versus N-methyl oxidation to N-formyl-noratropine was found to be sensitive to the struc-
ture of the alcohol co-solvent, the rate of H2O2 addition and the concentration of water, whereas temp-
erature mainly affected the atropine conversion efficiency. The use of tert-butyl or cumene
hydroperoxide as oxidants shifted the reaction selectivity toward N-methyl oxidation compared to
aqueous H2O2. Various inorganic oxidants were found to be ineffective. The FeIII-TAML also catalysed the
N-demethylation of the opiate alkaloids thebaine and oxycodone with aqueous H2O2 in higher conver-
sion efficiencies compared to atropine but with lower selectivity. These investigations thus document key
mechanistic features of the FeIII-TAML-catalysed N-demethylation of these alkaloids and provide insight
into how this benign catalytic system could find broader utilisation for N-demethylation in general.
Received 20th September 2013,
Accepted 26th November 2013
DOI: 10.1039/c3gc41972j
www.rsc.org/greenchem
Introduction
The naturally occurring alkaloids atropine (1), scopolamine (2)
and cocaine (3) along with the opiate morphine (4) all contain
a tertiary N-methylamine group (Fig. 1) and have been used for
many centuries by humans as medicinal and psychoactive
agents. Modification of the N-methylamine group alters the
pharmacological properties of these alkaloids and provides an
avenue to many important pharmaceutical compounds. For
example, N-demethylation of 1 to noratropine (13) (Scheme 1)
followed by N-alkylation with isopropyl bromide and then with
methyl bromide provides the bronchodilator ipratropium
bromide (6), containing axial and equatorial N-isopropyl and
N-methyl substituents, respectively.1 A similar transformation
of 2 provides the bronchodilator oxitropium bromide (7) with
axial and equatorial N-ethyl and N-methyl substituents, respec-
tively.1b In contrast, N-alkylation of 1 installs the alkyl group in
both axial and equatorial positions,1a while for 2 the alkyl
group is located in the equatorial position.2
Fig. 1 Structures of various naturally occurring tertiary N-methyl alka-
loids (1–5) and their synthetic derivatives (6–12).
Amongst the opiate family, thebaine (5) serves as a syn-
thetic precursor for the analgesic oxycodone (8a) and the inter-
mediate oxymorphone (9a).3 The latter compound is used to
prepare the N-methylcyclobutyl-based analgesic nalbuphine
(10) and the N-methylcyclopropyl-based opioid antagonists nal-
trexone (11) and nalmefene (12) (Fig. 1) used in the treatment
Centre for Green Chemistry, Monash University, Clayton, Victoria 3800, Australia.
E-mail: milton.hearn@monash.edu; Fax: +613 9905 8501; Tel: +613 9905 4547
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3gc41972j
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1399–1409 | 1399

View Article Online
Paper
Green Chemistry
on the substrate.13 However, in the case of 1, due to
dehydration of the primary alcohol group, the Pd-catalysed
N-demethylation/N-acylation
lacked
chemoselectivity.13a
Although Pd-catalysed N-demethylation/N-acylations have been
applied in the syntheses of the opiate drugs buprenorphine13b
and naltrexone,13c use of Pd catalysts in the commercial pro-
duction of active pharmaceutical ingredients (APIs) leads to
the problematic issues of cost and Pd contamination where
levels well below 10 ppm are mandated.14
Scheme 1 Modified Polonovski reaction for N-demethylating tertiary
N-methyl alkaloids using iron-based catalysts.
of alcohol or opiate dependence and for rapid opiate detoxifi-
cation. Transformation of 9a to 10–12 incorporates a N-
demethylation step, leading to noroxymorphone (9b), whereby
the nascent secondary nitrogen atom can be alkylated directly
with an alkyl halide or indirectly via reductive acylation.3
N-demethylation of tropane and opiate alkaloids to their
noralkaloid derivatives is thus a key step in the synthesis of
multiple pharmaceuticals.
Collectively, there is thus a need to develop efficient and
more benign catalytic methods for the N-demethylation of
N-methyl alkaloids to overcome these constraints. We have
recently reported an efficient and convenient method for
N-demethylating the tropane alkaloids 1 and 2 in high yield and
product purity using aqueous H2O2 and the iron(III) tetraamido
macrocyclic complex FeIII-TAML (15) as a catalyst in various
organic co-solvents (Scheme 2).15 The potential offered by 15
for selective N-demethylation of other compounds has also
been documented in a study on the catalytic oxidative
decomposition of the environmentally persistent antidepress-
ant drug sertraline in water.16 Sertraline contains a secondary
N-methylamine group and 15 was found to catalyse its
N-demethylation to the primary amine as part of the decompo-
sition pathway.
A number of procedures have been developed to N-
demethylate 1 or 2 to nortropane derivatives, e.g. phosgene,1b
2,2,2-tri-chloroethylchloroformate4 or α-chloroethylchlorofor-
mate,5 oxidative N-demethylation with KMnO4
or photo-
1b,5,6
chemical oxidative N-demethylation.7 Although these methods
can provide nortropane derivatives in mid to high yields (e.g.
35–90%), they employ toxic reagents and solvents and generate
hazardous waste by-products. N-demethylation of oxymor-
phone (9a) to noroxymorphone (9b) in the commercial synth-
eses of 10–12 uses the highly toxic cyanogen bromide (BrCN)
as demethylating reagent and generates stoichiometric waste
by-products.3 Moreover, this von Braun reaction requires
O-protection and O-deprotection steps which reduce the
overall synthetic efficiency and create additional solvent waste.
During the last ten years, modifications of the Polonovski
reaction have been used to N-demethylate tropane and opiate
alkaloids via their N-oxides using Fe(II)SO4,8 tetrasodium
5,10,15,20-tetra(4-sulfophenyl)porphyrinato-Fe(II),9 ferrocene
and its derivatives,10 iron powder11 or stainless steel11b as cata-
lysts (Scheme 1). These processes typically employ a two-step
process with oxidation of the alkaloid using m-chloroper-
benzoic acid (mCPBA) to the corresponding N-oxide which is
isolated as a hydrochloride salt and then heated with an iron
catalyst.
The FeIII-TAML complex 15 was originally developed by the
Collins group as an environmentally benign catalyst for oxi-
dative remediation of persistent organic pollutants using
aqueous H2O2.17 Its catalytic properties have subsequently
m-Chloroperbenzoic acid has also been used as the oxidant
followed by addition of aqueous HCl and a mixture of iron
powder (13 mol%) and FeCl3·6H2O (2 mol%) in a one-pot pro-
cedure.12 Although satisfactory yields of specific noralkaloids
were achieved with this method, a limitation was the significant
regeneration of the starting N-methyl alkaloid due to a compet-
ing reduction of an intermediate aminium radical cation by
the iron catalyst. Thus, the product noralkaloid has to be puri-
fied from the starting material by chromatography and these
additional unit operations are undesirable in a commercial
production setting due to the yield reduction, additional
solvent use and excessive waste disposal. Further, the use of
mCPBA as an oxidant on a large scale presents its own hazard
and waste disposal issues.
Palladium-based catalytic methods under an oxygen atmos-
phere have also been used for N-demethylation/N-acylation of
1 and opiate alkaloids using 2–20 mol% catalyst, depending
Scheme 2 N-Demethylation of atropine (1) and scopolamine (2) using
aqueous H2O2 and FeIII-TAML catalyst (15) in 96% EtOH co-solvent.
1400 | Green Chem., 2014, 16, 1399–1409
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Samples were chromatographed using a gradient of 0–2 min
5% B and then 2–30 min 5% B to 65% B at a flow rate of 4 μL
min−1 and a sample (0.1 μL, 20 μg mL−1) injected in 5%
B. Temperature measurements were made using a HI-93530
K-thermocouple thermometer from Hanna Instruments, Keys-
borough, Australia. Controlled addition of aqueous H2O2 to
reaction mixtures was performed with a Model 100 Series
syringe pump from KD Scientific.
Reagents
Aqueous hydrogen peroxide (30%) was purchased from Merck,
Kilsyth, Australia. Atropine, 70% aq. tert-butyl hydroperoxide,
5.0–6.0 M tert-butyl hydroperoxide in decane, 80% cumene
hydroperoxide, potassium peroxymonosulfate, potassium per-
sulfate and MnO2 were purchased from Sigma-Aldrich, Castle
Hill, Australia. Chlorine-based household bleach (4%) was
used as a source of NaOCl. Thebaine was a gift from Tasma-
nian Alkaloids (Launceston, Australia). The FeIII-TAML catalyst
15 (CAS: 895567-73-4) was purchased from GreenOx Catalysts
Inc., Pittsburgh, USA. The composition of 15 was specified by
the manufacturer to be 70% w/w 15, 20–30% w/w isopropanol,
3–5 wt% NaCl and 5% w/w metal-free tetraamido macrocycle.
Scheme 3 Catalytic pathways for the oxidation of substrate with the
FeIII-TAML catalyst 15 in an aqueous solution of H2O2.
been studied in pulp and paper processing,18 aluminium refin-
ing19 and fuel desulfurization.20 A general pathway for the 15-
catalyzed oxidation of organic substrates with H2O2 in water
involves a pH-dependant oxidation of 15 by H2O2 to a Fe(IV)
μ-oxo dimer (16) which is then reduced by the substrate
(Scheme 3).17a The 15-catalysed decomposition of H2O2 to H2O
and O2 21 or solvent oxidation represent competing reactions.
Although 15 has been investigated extensively as a catalyst
for the destruction of water pollutants, its full potential in
organic synthesis is less well documented. Here, we report
further studies designed to delineate the reaction pathway of
the 15-catalysed N-demethylation of 1 and the impact of various
parameters on the reaction selectivity and conversion efficiency.
The reactivity of the opiate alkaloids thebaine (5) and oxycodone
(8a) with the 15/H2O2 system has also been documented,
demonstrating that N-demethylation also occurs with these
compounds, but with a reduced level of selectivity under the
same conditions optimised for the tropane alkaloids.
Synthesis of N-formyl-noratropine (17)
To a stirred solution of formic acid (115 mg, 2.51 mmol) in
dichloromethane (5 mL) on an ice bath was added solid
N-ethyl-N′-(3-dimethylaminopropyl)-carbodiimide hydro-chloride
(240 mg, 1.25 mmol) and the mixture stirred to give a solution.
After 15 min, a solution of noratropine (345 mg, 1.25 mmol)
and N-methylmorpholine (126 mg, 1.25 mmol) in dichloro-
methane (10 mL) at 0 °C was added. The mixture was then
removed from the ice bath and stirred for 1 day, and then
washed successively with 1 M HCl (2 × 15 mL), sat. NaHCO3
(15 mL) and sat. NaCl (15 mL). The solution was dried
(MgSO4), filtered and the solvent removed in vacuo to give a
1 : 1 mixture of 17 N-formyl isomers as an oil (204 mg, 54%).
1H NMR (CDCl3) δ 8.06 (s, 0.5H, NHCO), 8.04 (s, 0.5H, NHCO),
Experimental section
General analytical procedures
Proton nuclear magnetic resonance (1H NMR) spectra were 7.45–7.20 (m, 5H, PhH), 5.17 (t, J = 4.0 Hz, 1H, CH–O–), 4.54
recorded at 400 MHz with a Bruker DRX400 spectrometer in (m, 0.5H, CH–NCHO), 4.42 (m, 0.5H, CH–NCHO), 4.20 (dd, J =
CDCl3 with tetramethylsilane as an internal standard (δ 9.6 Hz, 1H, HCH–OH), 3.99 (m, 0.5H, CH–NCHO), 3.85–3.79
0.00 ppm). Carbon-13 nuclear magnetic resonance (13C NMR) (m, 2.5H, CH–NCHO, HCH–OH, Ph–CH–), 2.39 (broad s, 1H,
spectra were recorded at 100 MHz on a Bruker DRX400 spectro- OH), 2.17–1.54 (m, 7H, CH2), 1.33–1.24 (m, 1H, tropane –
meter in CDCl3. Low resolution electrospray ionization (ESI) HCH–); 13C NMR (CDCl3) 172.03 (OCvO), 172.02 (OCvO),
mass spectra were recorded on a Micromass Platform II API 157.38 (CHvO), 135.57 (C), 135.50 (C), 128.95 (CH), 128.94
QMS Electrospray mass spectrometer with cone voltage at 20 V (CH), 128.08 (CH), 128.06 (CH), 127.86 (CH), 127.84 (CH),
or 35 V as solutions in MeOH containing formic acid. High 68.09 (HC–O), 63.97 (CH2–OH), 63.95 (CH2–OH), 54.46 (CH),
resolution ESI mass spectra were recorded on a Brucker 54.45 (CH), 53.39 (CH), 53.31 (CH), 48.56 (CH), 48.47 (CH),
BioApex 47e Fourier Transform mass spectrometer. Principle 38.43 (CH2), 38.12 (CH2), 35.79 (CH2), 35.54 (CH2), 27.84
ion peaks (m/z) are reported and M denotes the molecular ion.
(CH2), 27.41 (CH2), 27.09 (CH2), 26.70 (CH2); Calcd for
High performance liquid chromatography-mass spec- [C17H22NO4 + H]+: 304.1549, obsvd: 304.1545.
trometry (HPLC-MS) analyses were performed using a 1100
Synthesis of atropine N-oxide hydrochloride (18·HCl)
Series microflow HPLC coupled to an Agilent LC/MSD Trap
electrospray ionization ion trap mass spectrometer. A Zorbax To a stirred solution of atropine (0.500 g, 1.73 mmol) in CHCl3
300SB-C18 (150 × 0.3 mm, 3.5 μm sorbent) reverse phase (6 mL) at −25 °C was added solid m-chloroperoxybenzoic acid
column was used and the solvent systems were 0.2% acetic (77%, 409 mg, 2.94 mmol). After stirring for 2.5 h, iron(III) sul-
acid in H2O (A) and 0.2% acetic acid in acetonitrile (B). phate (4.8 mg, 12 μmol) was added to destroy excess m-CPBA
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1399–1409 | 1401

View Article Online
Paper
Green Chemistry
and the mixture was then extracted with 1 M HCl (3 × 7 mL). H2O2 as a 30% aqueous solution. An ice bath was used for the
The aqueous extracts were combined, washed with CHCl3 (2 × reaction performed at 0 °C and a dry ice-acetone bath used for
20 mL) and concentrated in vacuo. The concentrate was then reactions performed at −20 and −40 °C. Reaction mixtures
freeze-dried to obtain a 1.3 : 1 mixture of 18·HCl equatorial were left to stir for the time specified and then a catalytic
1
and axial N-oxide isomers as a white solid (495 mg, 84%). H amount of MnO2 (7 mol%) was added to decompose any
NMR (D2O) δ 7.49–7.40 (m, 5H, PhH), 5.12–5.07 (m, 1H, CH– remaining H2O2. The mixture was then basified with 10%
O–), 4.24 (dd, 1H, J1,2 = 9.6 Hz, J1,3 = 6.6 Hz, 1H, –HCH–OH), aqueous ammonia to pH 11 (indicator paper) and extracted
4.13–3.93 (m, 4H, CH–N(CH3)O, –HCH–OH, Ph–CH–), 3.53 (s, with CHCl3. Control experiments in which MnO2 was not used
1.7 H, CH3–NO), 3.41 (s, 1.3H, CH3–OH), 2.81 (ddd, J1,2 = 16.4 to decompose H2O2 gave the same conversion of 1 and yields
Hz, J1,3 = 4.0, 4.0 Hz, 0.43H, –HCH–), 2.74 (ddd, J1,2 = 16.4 Hz, of 13 and 17 but resulted in significant bubbling and heating
J1,3 = 4.0, 4.0 Hz, 0.57H, –HCH–), 2.67–2.54 (m, 1H, –HCH–), of the aqueous extract upon addition of 10% NH3. The organic
2.44–1.92 (m, 4.57H, –CH2–), 1.81 (d, J1,2 = 16.4 Hz, 0.43H, – extract was extracted with 1 M HCl, dried (Na2SO4), filtered
HCH–), 1.68 (ddd, J1,2 = 14.4 Hz, J1,3 = 10.4, 4.6 Hz, 0.43H, – and solvent was then removed in vacuo to give crude N-formyl-
1
HCH–), 1.59 (ddd, J1,2 = 12.8 Hz, J1,3 = 9.6, 3.8 Hz, 0.57H, – noratropine (17), as identified by H NMR and by comparison
HCH–); 13C NMR (D2O) 173.15 (CvO), 172.98 (CvO), 135.26 to the 1H NMR and HPLC-MS of 17 synthesized independently
(C), 135.16 (C), 129.19 (CH), 129.10 (CH), 128.28 (CH), 128.22 from noratropine (13) (Fig. S3 & S4†).22 The molecular masses
(CH), 128.20 (CH), 71.71 (CH3), 71.60 (CH3), 70.41 (CH3), 70.29 of other atropine derived by-products in the sample of crude
(CH3), 64.71 (CH–OH), 63.72 (CH–OH), 62.09 (CH2–OH), 53.60 17 were measured by HPLC-MS while the relative amount of 17
(CH), 53.55 (CH), 52.36 (CH), 46.19 (CH), 24.64 (CH2), 24.40 to the by-products was estimated by 1H NMR integration of the
(CH2), 22.88 (CH2), 22.61 (CH2); Calcd for [C17H23NO4 + H]+: NCHO and CH–O– peaks. The 1 M HCl aqueous extract above
306.1705, obsvd: 306.1703. The 1H NMR spectrum of the was basified with 10% aqueous ammonia and then extracted
neutral form was measured by adding 1 equivalent of NaHCO3 with CHCl3. The organic extract was dried (Na2SO4), filtered
to the D2O solution: 7.49–7.40 (m, 5H, PhH), 5.09 (broad m, and solvent was then removed in vacuo to give mixtures of 13
1H, CH–O), 4.25 (ddd, 1H, J1,2 = 14 Hz, J1,3 = 7.0, 6.6 Hz, HCH– and 1, as specified. The ratio of 13 to 1 was determined from
1
OH), 4.06–3.99 (m, 2H, HCH–OH, Ph–CH–), 3.89 (broad m, the relative integrations of the H NMR CH–OC(O) and CH–N
0.57H, CH–N(CH3)O), 3.76 (broad m, 1H, CH–N(CH3)O), 3.65 signals and, along with the sample mass, used to calculate the
(broad m, 0.43H, CH–N(CH3)O), 3.39 (s, 1.7H, CH3–NO), 3.28 conversions and yields of 1 and 13, respectively.
(s, 1.3H, CH3–NO), 2.78 (broad m, 1H, –HCH–), 2.59 (broad m,
Reactivity studies with atropine N-oxide·HCl, atropine N-oxide,
1H, –HCH–), 2.36–1.87, m, 4.57H, –HCH–), 1.75 (d, J1,2
16 Hz, 0.43H, –HCH–), 1.69–1.55 (m, 1H, –HCH–).
=
N-formyl-noratropine and noratropine
To a solution of substrate (50 mg) in 96% ethanol (1.5 mL)
containing 15 (1 mol%), or no catalyst, was added 50 molar
Synthesis of oxycodone (8a)
Oxycodone was synthesized from thebaine (6.00 g. 19.3 mmol) equivalents of H2O2 as a 30% aqueous solution. The mixture
following the procedure of Kraβnig et al.,25 with an extended was stirred for 1 h and worked up as described for the
hydrogenation of the 14-hydroxycodeinone intermediate. The N-demethylation of atropine. For the reactions of 13, 17, and
yield of 8 from thebaine was 66%. 1H NMR (CDCl3) δ 6.70 (d, J 20·HCl with catalyst alone, substrate (50 mg) and 15 (1 mol%)
= 8.0 Hz, 1H, PhH), 6.64 (d, J = 8.0 Hz, 1H, PhH), 5.30 (s, 1H, were stirred for 1 h in 96% ethanol (1.5 mL) with work up as
CH–O–), 3.87 (s, 3H, CH3–O–), 3.16 (d, J1,2 = 18.4 Hz, 1H, Ph– described above. To test the reactivity of the atropine N-oxide,
HCH–), 3.02 (ddd, J1,2 = 14.4 Hz, J1,3 = 14.4, 5.2 Hz, 1H, –HCH– 1 equivalent of NaHCO3 dissolved in the H2O2 solution or in
CvO), 2.86 (d, J = 6.0 Hz, 1H, CH–N), 2.56 (dd, J1,2 = 18.4 Hz, water was added to 20·HCl in 96% ethanol (1.5 mL) to generate
J13 = 6.0 Hz, 1H, Ph–HCH–), 2.49–2.35 (m, 2H, –HCH–N, – the neutral form.
HCH–CH2–N), 2.30 (ddd, J1,2 = 14.4 Hz, J1,3 = 3.2 Hz, 1H,
HCH–CvO), 2.21–2.13 (m, 1H, –HCH–N), 1.87 (ddd, J1,2 = 13.2
Reaction of noratropine with CH2O, H2O2 and FeIII-TAML
Hz, J1,3 = 5.2, 3.2 Hz, 1H, –HCH–C–OH), 1.67–1.54 (m, 2H, To a solution of 13 (331 mg, 1.2 mmol) and 15 (8 mg,
–HCH–CH2–N, –HCH–C–OH); 13C NMR (CDCl3) 208.44 (CvO), 0.012 mmol) in 96% ethanol (10 mL) was added 40% form-
144.87 (C), 142.80 (C) 129.34 (C), 124.96 (C), 119.39 (CH), aldehyde (0.90 mL, 12 mmol). The mixture was stirred for 5 min
115.00 (CH), 90.28 (–O–CH), 70.25 (HO–C), 64.45 (N-CH), 56.74 and then 30% H2O2 (6.1 mL, 60.1 mmol) was added. The
(O–CH3), 50.12 (C), 45.15 (N-CH2–), 42.66 (N-CH3), 36.06 mixture was then stirred for 1 h and worked up as described
(OvC–CH2–), 31.33 (–CH2–CH2–N), 30.43 (HO–C–CH2–), 21.83 for the N-demethylation of atropine. Solvent was removed from
(Ph–CH2–); Calcd for [C18H21NO4 + H]+: 316.1549, obsvd: the acid-insoluble organic extract to give crude 17 (207 mg).
316.1549.
This was chromatographed on silica gel with 19 : 1 DCM–
MeOH to give 17 as a colourless oil (144 mg, 40% yield). For
the reaction without 15, the same workup and chromatography
gave 17 in 10% yield. For the reactions of 13 with CH2O and
General method for FeIII-TAML-catalyzed N-demethylation of
atropine with H2O2 and analysis of product fractions
To a solution of atropine (1) and 15 in alcohol solvent, as TAML alone or with CH2O alone, the same workup gave a
specified in Tables 1–5, was added the specified amount of crude product in 6% and 10% mass recovery, respectively.
1402 | Green Chem., 2014, 16, 1399–1409
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
1
Analysis of these products by H NMR showed they contained yield of 13 was measured as described for mixtures from H2O2-
10 mol% of 17 relative to the other 13-derived products, as based reactions.
indicated by the relative integration values of the NHCO and
CH–O– peaks.
Reactions of atropine with KO3SOOH, K2O3SOOSO3 or NaOCl
and FeIII-TAML
Temperature change measurements
To a solution of 1 (50 mg, 0.173 mmol) 1 mol% 15 in 2 : 1 v/v
The thermocouple probe of the thermometer was inserted
into a stirred solution 1 and 15 or 1 and 15 alone in
ethanol as specified in Fig. 2. The temperature was recorded
and an aqueous solution of 30% H2O2 was then added as
specified in Fig. 2, and the temperature recorded at the times
indicated.
H2O/96% EtOH (1.5 mL) was added 1 equivalent of oxidant as
a solid (for KO3SOOH and K2O3SOOSO3) or a solution (for
NaOCl). The mixture was stirred for 1 h and worked up as
described for the H2O2-based reactions. Neither 13 or 17 were
isolated from any of these reaction mixtures.
Reaction of thebaine with H2O2 and FeIII-TAML
Measurement of n-butanol conversion
To a solution of 5 (50 mg, 0.161 mmol) and 1 mol% 15 in
acetone (3.0 mL) was added 0.161 or 0.483 mmol H2O2 as a
30% aqueous solution. The mixture was stirred for 1 h and
worked up as described for the H2O2-based reactions with 1.
The reaction of 5 (0.161 mmol) with H2O2 (0.483 mmol) alone
resulted in 95% recovery of 5.
To a stirred solution of n-butanol (100 mg, 1.35 mmol) in H2O
(2.0 mL) was added H2O2 as a 30% aqueous solution, as speci-
fied in Table 4, or an equal volume of H2O alone. The mixture
was stirred for 1 h and then extracted with CDCl3 (2, 2.7 and
3.4 mL for the 1, 5 and 10 eq. H2O2 reactions, respectively).
The CDCl3 extract was dried (Na2SO4), filtered into a measur-
ing cylinder and the volume recorded. Toluene (0.67 mmol)
was then added as an internal standard and a portion of the
solution was analysed by 1H NMR. The amount of n-butanol
relative to toluene was then calculated from the relative inte-
grations of the n-butanol and toluene 1H NMR peaks. This
ratio was used to derive the yield of recovered n-butanol. The
conversions of n-butanol reported in Table 4 are calculated
from the differences between the amount of n-butanol recov-
ered from 15/H2O2 reaction mixtures relative to n-butanol
recovered from H2O alone.
Reaction of oxycodone with H2O2 and Fe(III)-TAML
To a solution of 8a (50 mg, 0.158 mmol) and 1 mol% 15 in
acetone (3.0 mL) was added 0.474 mmol H2O2 as a 30%
aqueous solution. The mixture was stirred for 1 h and worked
up as described for the N-demethylation of 1. The same reac-
tion without 15 gave back 95% 8a. For the reaction at −40 °C
with 15, 0.79 mmol H2O2 was used.
Preparative N-demethylation of oxycodone using FeIII-TAML
To a solution of 8a (500 mg, 1.58 mmol) 1 mol% 15 in acetone
(30 mL) at 23 °C or −40 °C was added 4.7 or 7.9 mmol H2O2,
respectively, as a 30% aqueous solution. The mixture was
stirred for 1 h and then a catalytic amount of MnO2 (10 mg)
was added to decompose excess H2O2. The mixture was then
basified with 10% ammonia (50 mL) to pH 11 (indicator
paper) and extracted with CHCl3 (3 × 50 mL). The organic
extract was extracted with 1 M HCl (50 mL). The aqueous
extract was then basified with 10% ammonia (50 mL) to pH
11, extracted with CHCl3 (3 × 100 mL), dried (Na2SO4), filtered
and solvent removed in vacuo to give an oil. This was chro-
Syringe pump addition of H2O2
A solution of H2O2 as specified in Table 7 was drawn up into a
3 mL polypropylene syringe through 1 mm internal diameter
(i.d.) polytetrafluoroethylene tubing. The syringe with attached
tubing was then fixed to the syringe pump and the H2O2 solu-
tion then added to a stirred solution of 1 and 15 in 96% (v/v)
ethanol.
General method for FeIII-TAML-catalyzed N-demethylation of
atropine with organic hydroperoxides
To a solution of 1 (50 mg, 0.173 mmol) and 1 mol% 15 in 96% matographed on silica gel (99 : 1–5 : 1 v/v DCM/1.5 M methano-
EtOH or anhydrous EtOH was added hydroperoxide, as speci- lic ammonia) to give crude noroxycodone (8b) as an oil in 17%
fied in Table 7. Reaction mixtures were left to stir for the time (85 mg) and 13% (66 mg) mass recoveries for the 23 °C and
1
specified and then a catalytic amount of Fe2(SO4)3xH2O (1 mg) −40 °C reactions, respectively, as assessed by H NMR which
was added to decompose any remaining hydroperoxide. The showed no N-methyl protons (Fig. S14†),20 by high-resolution
mixture was then worked up as described for the H2O2-based mass spectrometry which showed a predominant peak at
reactions. For the cumene hydroperoxide reactions excess 302.1383 m/z (calcd for [C17H19NO4
+
H]+: 302.1392)
cumyl alcohol was isolated with 17. The amount of 17 relative (Fig. S16†)22 and, for the reaction at room temperature, predo-
1
to cumyl alcohol was estimated by H NMR integration of the minant noroxycodone signals in the 13C NMR: (CDCl3) 208.36
17 NCHO and cumyl alcohol ortho-PhH signals. The amount (CvO), 145.06 (C), 143.07 (C) 129.33 (C), 124.93 (C), 119.45
of 17 relative to other atropine by-products, as well as yield, (CH), 115.03 (CH), 90.45 (–O–CH), 70.14 (HO–C), 57.38 (N-CH),
1
was estimated by H NMR and HPLC-MS, as described for 17 56.84 (O–CH3), 50.93 (C), 37.17 (N-CH2–), 36.11 (OvC–CH2–),
obtained from H2O2-based reactions. The conversion of 1 and 32.36 (–CH2–), 31.47 (–CH2–), 29.74 (Ph–CH2–) (Fig. S15†).22
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1399–1409 | 1403

View Article Online
Paper
Green Chemistry
To determine if either of the hydroxylated noratropine
by-products 20 were produced by hydroxylation of 13 or by
N-demethylation of a hydroxylated N-methyl precursor, 13 was
Results and discussion
FeIII-TAML-catalysed reaction of atropine with H2O2
The reaction of 1 with 50 equivalents of H2O2 in ethanol co- allowed to react under the same reaction conditions as for 1
solvent containing 1 mol% 15 gave 13 as the main product and the aqueous extract analyzed by HPLC-MS (Fig. S7†).22 In
(Fig. S1 and S2†)22 along with N-formyl-noratropine (17) as a by- contrast to the reaction with 1, neither isomer of 20 was
product (Scheme 4) which was separated from 13 by solvent- detected, suggesting that they are produced by N-demethyla-
1
extraction. The identity of 17 was established by H NMR and tion of a hydroxylated N-methyl precursor. This result, along
HPLC-MS and verified by a sample of 17 synthesized separately with the MS/MS fragmentation analysis and corresponding
by N-acylation of 13 with formic acid (Fig. S3 and S4†).22
NMR spectral analysis, also supports the conclusion that
The effect of changes from 1 to 100 equivalents of H2O2 on neither of the isomers of 20 is the N-hydroxy-noratropine, as
the conversion of 1 and the production of 13 and 17 are shown proposed previously,15 but instead contain hydroxylated
in Table 1. The hydroxylated atropine by-product 19 (306.1 tropane carbon atoms.
m/z) and two isomeric hydroxylated noratropine by-products (20)
Reaction pathway of N-demethylation
(m/z 292.1) were present in the aqueous extract by HPLC-MS
(Scheme 4 & Fig. S5†).22 Although 19 and atropine N-oxide (18)
To determine if 17 or 18 were reaction intermediates or dead-
have the same molecular mass, they were readily distinguished
by their different HPLC retention times (Fig. S5†)22 and only
trace amounts of 18 were detected in the aqueous extract.
Further analysis of 19 and 20 by-products by MS/MS fragmen-
tation led to the production of species with m/z values of 140.0
and 126.1, consistent with hydroxylated tropane and nor-
tropane ring fragments, respectively (Fig. S4†).22
end by-products of the N-demethylation reaction, their reactiv-
ity in the 15/H2O2 system was investigated. Accordingly, treat-
ment of 17, 18·HCl or in situ-neutralized 18 with 1 mol% 15
and 50 equivalents of H2O2 or with 15 alone or H2O2 alone did
not produce 13. In the case of the N-oxide 18, the majority
(>90%) of this compound did not react and was recovered in
the aqueous phase following a standard workup, as assessed
by HPLC-MS. For 17, 80% of this material was recovered whilst
again no evidence for the formation of 13 was obtained. Exam-
ination of the aqueous extract by HPLC-MS demonstrated that
17 was the predominant tropane species.
Collectively, these results confirm that 17 and 18 are not
significant reaction intermediates in the 15-catalysed
N-demethylation of 1 to 13 with H2O2. The N-demethylation of 1
thus appears to occur via a biomimetic-like catalytic oxidation
to N-hydroxymethyl-noratropine 21 which mainly decomposes
to 13 and formaldehyde or, alternatively, is further oxidized to
the dead-end by-product 17 (Scheme 5). Although in vitro or
in vivo metabolic oxidative N-demethylation of 1 and 2 is known,
the reaction pathways have not been established for these sub-
strates nor have the putative N-oxide intermediates been iso-
lated as oxidation products.23 However, for other tertiary
N-methyl and N-alkylamines, many lines of evidence suggest that
metabolic N-dealkylation to secondary amine and aldehyde
Scheme 4 Products from the reaction of 1 with H2O2 and 15.
Table 1 Percentage conversion and major/minor product yields for the
FeIII-TAML-catalyzed oxidative N-demethylation of atropine (1) with
different amounts of H2O2
Entry
Eq. H2O2
1h (%)
13h (%)
17i (%)
1a
0
1
2
5
5
5
5
10
25
50
50
50
50
100
0
61
0
36
0
1
7
6
10
0
0
5
5
5
2a
3a
80
50
4a
94 2j
93
56 6 j
54
5b
6c
44k
30k
96
0
0
70
7d
8a
9a
98
72
10a
11e
12f
13g
14a
97 1l
98
75 5l
63
16
17
0
99
44
99
46
0
77
4
a 0.173 mmol 1, 1 mol% 15, 1.5 mL 96% EtOH, 1 h. b 17 h. c 0.5 mol%
Fe2(SO4)3xH2O. d 1 mol% FeCl3(H2O)6. e 0.5 M 1. f 1 M 1. g No catalyst.
h Conversion of 1 and yield of 13 determined by 1H NMR of the
isolated product mixture. i Yield estimated by 1H NMR and HPLC-MS
of the crude acid-insoluble organic extract. j Mean conversion and
Scheme 5 Reaction pathway proposed for the 15-catalyzed oxidative
yield from four reactions with standard deviations shown. k Mixtures of N-demethylation of atropine 1 to noratropine 13 and formation of
1 and 5–10 mol% 18 were isolated. l Mean conversion and yield from
a N-hydroxymethylnoratropine inter-
N-formyl-noratropine 17 via
mediate 21.
six reactions with standard deviations shown.
1404 | Green Chem., 2014, 16, 1399–1409
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Table 2 Influence of temperature on the FeIII-TAML-catalysed oxidative
N-demethylation of atropine
products by cytochrome p450 enzymes occurs predominantly
via an α-hydroxy amine intermediate (viz. 21) rather than via a
N-oxide.24
Entrya
T (°C)
t (h)
1b (%)
13b (%)
In order to provide additional evidence for a role of the
putative intermediate 21 in the conversion pathway, 13 was
allowed to react with excess formaldehyde and 1 mol% 15, gen-
erating an equilibrium mixture of 13 and 21. When H2O2 was
added this experiment led to the production of 17 in 40%
yield, whereas the reaction without 15 present only gave a 10%
yield. These results are consistent with the reaction pathway
shown in Scheme 5 and indicate that 15 also catalyses the con-
version of 21 to 17. In contrast to the reactions carried out in
the presence of H2O2, only trace amounts (<1%) of 17 were iso-
lated from the reaction of 13 with formaldehyde and 15 or
with formaldehyde alone.
1
2
3
4
5
6
−40
−20
−20
0
23
50
1
1
6
1
1
1
82
82
80
97
98
98
65
64
60
73
71
78
a 0.173 mmol 1, 8.65 mmol H2O2, 1 mol% 15, 1.5 mL 96% EtOH, 1 h.
b Conversion of 1 and yield of 13 determined by 1H NMR of the
isolated product mixture.
marginally influenced N-demethylation/N-methyl oxidation
selectivity in these reactions.
Studies on atropine N-demethylation selectivity
The use of alternative alcohol co-solvents was investigated
next in order to examine the extent of competing solvent oxi-
dation which possibly accounts, in some part, for the high
excess of H2O2 required to achieve a good yield of 13 in high
purity. In previous studies, DMSO, acetonitrile and dimethyl-
carbonate were all oxidized by the 15/H2O2 system.15 To esti-
mate the relative rate of oxidation of the ethanol co-solvent,
the reactivity of n-butanol with 15/H2O2 as a model alcohol
was investigated since it could be more readily extracted from
the aqueous phase for measurement by 1H NMR.
The results of these experiments showed that although
aqueous n-butanol reacts more slowly with the H2O2/15 system
compared to 1 in ethanol (Table 3), a significant amount of
the solvent is oxidized. This finding suggests that the ethanol
oxidation by analogy is a significant competing reaction
during the N-demethylation of 1 as it is present in a large
molar excess. Moreover, the use of more sterically hindered
alcohol co-solvents in the N-demethylation reaction with the
15/H2O2 system may lead to improve yields because of their
lower reactivity.
The effects of various reaction parameters on atropine
N-demethylation selectivity were investigated with the overall aim
to find conditions giving the highest yield of 13 and reducing
the need for a large excess of H2O2. The reaction of 1 with the
15/H2O2 system is highly exothermic, as measured by the
increase in temperature when H2O2 was added to a solution of
1 with 1 mol% 15 at room temperature (Fig. 2). In contrast, the
reaction of H2O2 with 1 or 15 alone is much less exothermic.
These results prompted an investigation into the effect of
different temperatures on the conversion of 1 and yields of 13
(Table 2) and 17. When the reaction was carried out at 0 or
50 °C, similar conversions and product yields were obtained to
those performed at room temperature (Table 2, entries 4, 5 &
6). Lowering the temperature to −20 or −40 °C gave reduced
conversions and yields (Table 2, entries 1 & 2) which did not
increase with an extended reaction time (Table 1, entry 3). It
was noted that similar yields of 17 (3–5%) were obtained over
the temperature range studied, showing that temperature only
When compared to reactions carried out in ethanol, the 15-
catalyzed conversion of 1 with 1 equivalent of H2O2 was lower
in isopropanol or tert-butanol co-solvents, and the reaction
selectivity was shifted to N-methyl oxidation, as evident from
the higher yields of 17 (Table 4, entries 1, 4 & 7). The lower
conversions of 1 obtained in isopropanol and tert-butanol
suggest that competing alcohol oxidation does not signifi-
cantly limit the conversion of 1.
Increasing the amount of H2O2 to 5 molar equivalents
gave almost complete conversion of 1 in isopropanol and
Table 3 Reactivity of n-butanol with the FeIII-TAML/H2O2 system
Entry
Eq. H2O2
n-Butanola (%)
1 (%)
1
2
3
1
5
10
0
25
25
57b
87b
96c
Fig. 2 Temperature versus time profiles for the reaction of atropine
(0.173 mmol) with H2O2 (8.6 mmol) and 1 mol% FeIII-TAML (filled
circles), or with H2O2 alone (open circles) and for the reaction of H2O2
(8.6 mmol) with 1.6 μmol FeIII-TAML (filled triangles). Each reaction was
performed with 1.5 mL 96% ethanol co-solvent.
a 675 mM aqueous n-Butanol, 1 mol% 15, 1 h. Conversion determined by
1H NMR. b 500 mM 1 in 96% EtOH. c 115 mM 1 in 96% EtOH.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1399–1409 | 1405

View Article Online
Paper
Green Chemistry
Table 4 FeIII-TAML-catalyzed oxidative N-demethylation of atropine in
different alcohol solvents
Table 5 Effect of water on the reaction of atropine (1) with FeIII-TAML
and H2O2
Entrya
Solvent
Eq. H2O2
1b (%)
13b (%)
17c (%) Entrya
iPrOH (mL)
H2O (mL)
1b (%)
13b (%)
17c (%)
1
2
3
4
5
6
7
8
EtOH
EtOH
EtOH
1
5
50
1
5
50
1
5
50
5
61
94
98
55
99
99
56
99
99
87
88
36
56
71
36
49
71
29
37
45
53
65
1
6
5
4
18
6
1
2
3
4
1.25
1
0.75
0.5
0.25
0.5
0.75
1
97
94
84
71
64
62
65
49
14
9
5
iPrOH
iPrOH
iPrOH
tBuOH
tBuOH
tBuOH
MeOH
MeOH
2
a 0.173 mmol 1, 0.865 mmol H2O2, 1 mol% 15, 1 h. b Conversion of 1
and yield of 13 determined by 1H NMR of the products. c Yield
estimated by 1H NMR and HPLC-MS of the crude acid-insoluble
organic extract.
5
26
20
3
9
10
11
50
5
increased rate of competing 15-catalysed decomposition of
H2O2 (Scheme 3).
a 0.173 mmol 1, 1 mol% 15, 1.5 mL solvent, 1 h. b Conversion of 1 and
yield of 13 determined by 1H NMR of the isolated products. c Yield
from the 1H NMR and HPLC-MS of the crude acid-insoluble organic
extract.
In order to minimize competing 15-catalyzed decompo-
sition of H2O2 during the N-demethylation reaction, the effect
of reducing the H2O2 concentration was examined. To this
end, 30% (v/v) H2O2 was diluted with water and added slowly
to the reaction mixture using a syringe pump. Compared to
the faster addition from an auto-pipette, the reaction of 1 in
ethanol with H2O2 added as a 2.6% aqueous solution drop-
wise over 1 h led to almost complete conversion of 1 (99%), a
lower yield but higher purity of 13 and a shift in reaction
selectivity toward production of 17 (Table 5, entry 1 & cf.
Fig. S9†22) and other products.
tert-butanol (Table 4, entries 2, 5 & 8) and higher purity 13 as
assessed by 1H NMR compared to the reaction in ethanol,
(cf. Fig. S8†).22 However, compared to the reaction with 1 equi-
valent of H2O2, there was a shift in reaction selectivity toward
the production of 17, the extent of which followed the order
tert-butanol > isopropanol > ethanol, with lower yields of 13
being obtained with isopropanol and tert-butanol compared to
the reaction in ethanol.
Increasing the rate of addition of H2O2 marginally
increased the yield of 13 and gave similar reaction selectivity
(Table 6, entries 1–3). The shift in reaction selectivity toward
17 resulting from the slow addition of H2O2 compared to
faster addition from an auto-pipette may be due to the initially
more hydrophobic solvent environment associated with the
slower additions. To validate this hypothesis, the same slow
addition of H2O2 experiment was carried out with a solution of
1 and 15 in 1 : 1 EtOH–H2O, resulting in a shift in selectivity
toward N-demethylation but also a lower conversion of 1
(Table 6, entries 1 & 4).
Interestingly, further increases in the amount of H2O2 to
50 molar equivalents in ethanol, isopropanol or tert-butanol
led to the reaction selectivity being shifted towards N-demethy-
lation (Table 4, entries 3, 6 & 9). In terms of conversion
efficiency, methanol was a relatively poor solvent whilst the
reaction selectivity was similar to that observed with ethanol
(Table 4, entries 10 & 11).
Overall, these results show the conversion of 1 and selecti-
vity of N-demethylation vs. N-methyl oxidation are influenced
by both the amount of aqueous H2O2 and structure of the
alcohol co-solvent. Competing alcohol oxidation did not
appear to be significant factor in limiting the conversion of 1.
Instead, both conversion efficiency and reaction selectivity
appear to be more affected by the hydrophobicity of the
solvent system and substrate concentration, which presumably
affect the rates of conversion of 21 to 13 and 17, and compet-
ing alternative catalytic oxidations of 1 (cf. Scheme 5) and cata-
lytic decomposition of H2O2.
The use of a mixture of 1 : 1 isopropanol and water as the
solvent led to the higher conversion of 1 but shifted the reac-
tion selectivity toward production of 17 (Table 5, entries 4 & 5).
It is notable that the slow addition of H2O2 to the 1 : 1
Table 6 FeIII-TAML-catalyzed oxidative N-demethylation of atropine
with different rates of H2O2 addition
This shift in reaction selectivity toward N-demethylation
observed upon increasing H2O2 from 5 to 50 molar equivalents
in the isopropanol or tert-butanol solvent systems is presum-
ably due in part to the higher water content. To further investi-
gate the effect of water on the reaction selectivity, reactions
were performed in isopropanol containing different initial
amounts of water as a co-solvent, while keeping the amount of
H2O2 constant (Table 5). The results confirmed that increasing
Pump rate
Entrya
(mL h−1
)
1d (%)
13d (%)
17e (%)
1
2
1
2
4
1
1
99
99
99
80
99
36
39
43
36
27
17
18
19
11
21
3
4b
5c
a 0.173 mmol 1 and 1 mol% 15 in 1.5 mL 96% EtOH. H2O2 (0.865 mmol,
water content shifted the reaction selectivity of 1 toward 2.6% aq., 1 mL) was delivered dropwise via 1 mm i.d. tubing and the
reaction was stirred for 15 min after complete addition. b 1 and 15 in
N-demethylation over oxidation to 17 but at the expense of
lower conversion. The lower conversions obtained with increas-
1.5 mL 1 : 1 96% EtOH–H2O. c 1 and 15 in 1.5 mL 1 : 1 iPrOH–H2O.
d Determined by 1H NMR of the isolated mixture of 1 and 13. e Yield
ing amounts of water present is possibly a result of an estimated by 1H NMR of the crude acid-insoluble organic extract.
1406 | Green Chem., 2014, 16, 1399–1409
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
isopropanol–H2O system gave a much higher conversion and similar to that of H2O2 (Table 6, entry 10). Extending the reac-
different reaction selectivity compared to the faster addition of tion time with cumene hydroperoxide increased the conversion
H2O2 to the same solvent system (Table 6, entry 5 & Table 5, of 1 but not the yield of 13 (Table 7, entries 10 & 11). Increas-
entry 3). Overall, slowing the rate of addition of H2O2 to the ing the amount of cumene hydroperoxide increased the con-
reaction milieu enabled almost complete conversion of 1 with version of 1, with almost complete conversion being achieved
a smaller five-fold excess of H2O2 and afforded 13 in better with 3 molar equivalents (Table 7, entries 10, 12 & 13).
purity but in lower yield due to a shift in reaction selectivity However, compared to H2O2 the reaction selectivity was shifted
from N-demethylation toward production of 17.
toward production of 17 and the purity of the obtained 13 was
reduced to ca. 87%, as assessed by 1H NMR and HPLC-MS
together with detected 12 mol% of atropine by-products
(m/z [M + H]+ 292.2, 303.2 and 350.2).
Increasing the amount of cumene hydroperoxide to five
molar equivalents decreased the yield of 13 (Table 7, entries 14
& 15) and whilst 17 was not isolated from the organic extract,
the presence of other N-formyl-containing products was
evident from the 1H NMR of the crude product. The identity of
these products has yet to be determined.
N-Demethylation of atropine with alternative oxidants
The use of sterically hindered tert-butyl and cumene hydroper-
oxides as oxidants was anticipated to give higher conversions
of 1 compared to H2O2 as a result of their slower catalytic
decomposition by 15, in accord with their known slow reaction
17a
with 15 to give 16 compared to H2O2
(Scheme 3). With
aqueous tBuOOH, the conversion of 1 and the yield of 13 were
in fact lower than that obtained with H2O2 while the yield of
17 was considerably higher (Table 6, entries 1–4). Similar
FeIII-TAML catalysed reactions of thebaine and oxycodone with
H2O2
results were obtained when
a non-aqueous solution of
tBuOOH and anhydrous ethanol was employed as the co-
solvent (Table 7, entries 6–9). The reaction of 1 with tBuOOH
alone gave a relatively low conversion (Table 7, entry 5) and
produced mainly atropine N-oxide, as assessed by HPLC-MS.
The inorganic oxidants sodium hypochlorite, potassium
peroxymonosulfate and potassium persulfate reacted with 15
to give the same yellow-to-black colour change associated with
H2O2 or organic hydroperoxide activation. However, their use
in the N-demethylation of 1 with 1 mol% 15 gave low conver-
sions (<25%) compared to H2O2 with neither 13 or 17
obtained. Presumably, these differences are due to the propen-
sity of these oxidants to destroy the catalyst and/or the compet-
ing non-catalytic solvent oxidation during substrate oxidation.
In contrast to tBuOOH, one equivalent of cumene hydroper-
oxide gave a conversion and yield of 1 and 13, respectively,
Based on the above results with 1, the N-demethylation of the
pharmaceutical precursor thebaine (5) was then investigated.
For solubility reasons, it was necessary to use acetone as the
water miscible organic co-solvent, a solvent system which has
been shown to be suitable for the 15-catalysed N-demethyla-
tion of atropine (1) in modest yield.15 In contrast to 1, the reac-
tion of 5 with three equivalents of H2O2 and 1 mol% 15 gave
complete conversion and resulted in, inter alia, N-demethyla-
tion and oxidation and O-demethylation of the 1-methoxy-
cyclohexa-1,3-diene system to give a complex product mixture,
as assessed by 1H NMR of the crude product (cf. Fig. S10†).22
It is known that 5 reacts readily with H2O2 under acidic con-
ditions to give 14-hydroxycodeinone,
a reaction that is
employed in the industrial synthesis of oxycodone (8a).3
However, under our basic reaction conditions there was no sig-
nificant reaction of 5 with H2O2 alone, as evident from the
recovery of >95% 5 from the reaction mixture. For the crude
product obtained from the reaction of 5 with one equivalent of
H2O2 and 1 mol% 15, the integration ratios of the diene CH to
OCH3 to N-CH3 1H NMR signals indicated the rates of N-
demethylation and oxidation of the diene system were similar
(Fig. S10†).22 Analysis of the crude product mixture by HPLC-MS
did not detect northebaine (cf. Fig. S11),22 further demonstrating
that the reaction was less selective for N-demethylation.
Table 7 FeIII-TAML-catalyzed oxidative N-demethylation of atropine
with organic hydroperoxides
Entrya
ROOH
Eq.
T (h)
1d (%)
13d (%)
17e (%)
1
2
3
4
t-Bu(aq)
t-Bu(aq)
t-Bu(aq)
t-Bu(aq)
t-Bu(aq)
t-Bu
t-Bu
t-Bu
t-Bu
Cumene
Cumene
Cumene
Cumene
Cumene
Cumene
1
1
5
5
5
1
1
5
5
1
1
2
3
5
5
1
17
1
17
1
1
17
1
17
1
17
1
1
1
47
47
88
82
22
49
49
100
95
57
70
88
99
100
100
20
12
30
28
0
16
11
9
21
41
35
70
52
43
25
12
19
35
36
0
15
31
24
46
3
3
9
17
0
0
5b
6c
7c
8c
9c
10
11
12
13
14
15
Due to the lower reaction selectivity and complex product
mixture obtained with thebaine, attention was focused on the
reaction of oxycodone (8a) with the 15/H2O2 system. Compared
to 5, complete conversion was also achieved with 3 equivalents
of H2O2 and 1 mol% 15, giving a complex product mixture by
1H NMR (cf. Fig. S12†).22 The 1H NMR of the crude product
showed relatively weak singlets in the N-CH3 region, indicating
that significant but not complete N-demethylation had
occurred (cf. Fig. S12†).22 The reaction of 8a with five equiva-
lents of H2O2 at −40 °C and 1 mol% 15 also resulted in com-
plete conversion and significant N-demethylation but led to a
different product mixture by HPLC-MS (Fig. S13†).22 There was
17
a 0.173 mmol 1, 1 mol% 15, 1.5 mL 96% EtOH. b Without 15.
c Anhydrous EtOH and t-BuOOH in decane. d Conversion of 1 and yield
of 13 determined by 1H NMR of the isolated mixture. e Yield estimated
by 1H NMR of the crude acid-insoluble organic extract.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1399–1409 | 1407

View Article Online
Paper
Green Chemistry
no significant reaction of 8a with H2O2 alone, as evident from complex product mixtures. For 5, both N-demethylation and
the recovery of 95% of 8a from a room temperature reaction. oxidation of the diene system occur at similar rates and O-
The HPLC-MS chromatograms of the basic alkaloid frac- demethylation also occurs. In the case of 8a, noroxycodone
tions obtained from the 15-catalyzed reactions at 23 and was produced from the reaction but in low yield due to poor
−40 °C showed a predominant peak due to a species with m/z catalyst selectivity.
302.2, corresponding to [noroxycodone (8b)
+
H]+ (cf.
It is clear from this current study that the catalytic selecti-
Fig. S13†).22 For the product obtained at −40 °C, the HPLC-MS vity of 15 for the N-demethylation of N-methylamine alkaloids
showed an additional peak of similar intensity due to a species is dependent on the substrate structure, with 15 providing
with m/z 300.2, indicative of a dehydrogenated noroxycodone greater N-demethylation selectivity for the tropane alkaloid
derivative (cf. Fig. S13†).22 For both reactions, HPLC-MS also class under the conditions examined. These results also high-
detected several different minor co-eluting by-products (cf. light that further structural refinement of the FeIII-TAML cata-
Fig. S13†).22 Attempts to separate these products from both lyst or related low-toxicity oxidation catalysts would be a
reactions by flash chromatography were unsuccessful due to productive line of research to enable even more efficient and
similar retention times of the products, with only noroxyco- selective N-demethylation reactions to be used in the indus-
done recovered in relatively low amounts, i.e. with ∼17% trial-scale syntheses of active pharmaceutical ingredients
overall mass recovery.
(APIs) and other compounds via one-pot processes using an
environmentally benign oxidant such as H2O2. This objective
is the focus of our on-going research.
Conclusion
This investigation has further documented the oxidative N-
demethylation of atropine (1) and the opioids, thebaine (5)
and oxycodone (8a), using the FeIII-TAML catalyst 15 and
various different reaction conditions. On the basis of reactivity
studies, it can be concluded that the 15-catalysed N-demethyla-
tion of atropine (1) to noratropine (13) with H2O2 follows a
biomimetic oxidation pathway involving formation of the
N-hydroxymethyl-noratropine (21) intermediate.
Although good yields and purity of 13 can be obtained in a
simple one-pot process, a relatively high 50-fold molar excess
of H2O2 is required. Experiments using more oxidation-resist-
ant alcohol co-solvents showed that alcohol oxidation is not a
significant competing reaction that limits the conversion of 1
but that the conversion efficiency, as well as N-demethylation
versus N-methyl oxidation selectivity, is affected by the hydro-
phobicity of the solvent. Slower addition of H2O2 to minimize
the competing 15-catalysed decomposition led to a higher con-
version of 1 with a smaller excess of H2O2 but gave lower yields
of 13 due to a shift in reaction selectivity toward the pro-
duction of 17.
Overall, these findings suggest that the more hydrophobic
alcohol co-solvents stabilize the intermediate 21, enhancing its
15-catalyzed oxidation to 17. In contrast, higher amounts of
water in the reaction milieu favour N-demethylation by enhan-
cing the decomposition of 21 to 13 but is also associated with
reduced conversion, presumably due to increasing the rate of
competing catalytic H2O2 decomposition. Compared to H2O2,
a higher conversion of 1 was achieved with a smaller excess of
the more sterically hindered cumene hydroperoxide but also
resulted in a shift in reaction selectivity toward production of
17. This shift in selectivity could be due to a faster non-cataly-
tic reaction of cumene hydroperoxide with 21 compared to
H2O2 or hydrophobic stabilisation of 21.
Acknowledgements
The funding support of the Australian Research Council is
gratefully acknowledged.
Notes and references
1 (a) I. Tsyskovskaia, M. Kandil and Y. Beaumier, Synth.
Commun., 2007, 37, 439–446; (b) K. J. Erb, D. H. Martyres
and P. Seither, in Ullmann’s Encyclopedia of Industrial Chem-
istry, Wiley-VCH Verlag GmbH & Co. KGaA, 2000, vol. 3, pp.
639–675.
2 R. Glaser, D. Shiftan and M. Drouin, J. Org. Chem., 1999,
64, 9217–9224.
3 E. Friderichs, T. Christoph and H. Buschmann, in Ull-
mann’s Encyclopedia of Industrial Chemistry, Wiley-VCH
Verlag GmbH & Co. KGaA, 2000, vol. 3, pp. 339–386.
4 J. R. Pfister, J. Org. Chem., 1978, 43, 4373–4374.
5 F. Balssa and Y. Bonnaire, J. Labelled Compd. Radiopharm.,
2009, 52, 269–272.
6 M. J. Van der Meer and H. K. L. Hundt, J. Pharm. Pharma-
col., 1983, 35, 408.
7 (a) J. A. Ripper, E. R. T. Tiekink and P. J. Scammells, Bioorg.
Med. Chem. Lett., 2001, 11, 443–445; (b) J. Santamaria,
R. Ouchabane and J. Rigaudy, Tetrahedron Lett., 1989, 30,
3977–3980; (c) J. Santamaria, R. Ouchabane and J. Rigaudy,
Tetrahedron Lett., 1989, 30, 2927–2928.
8 (a) S. Thavaneswaran and P. J. Scammells, Bioorg. Med.
Chem. Lett., 2006, 16, 2868–2871; (b) K. McCamley,
J. A. Ripper, R. D. Singer and P. J. Scammells, J. Org. Chem.,
2003, 68, 9847–9850.
When compared to the tropane alkaloid 1, the 15-catalysed
N-demethylation of the opiate alkaloids thebaine (5) and oxy-
codone (8a) with H2O2 was less selective and gave more
9 (a) Z. Dong and P. J. Scammells, J. Org. Chem., 2007, 72,
9881–9885; (b) G. Kok, T. D. Ashton and P. J. Scammells,
Adv. Synth. Catal., 2009, 351, 283–286.
1408 | Green Chem., 2014, 16, 1399–1409
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
10 (a) G. B. Kok and P. J. Scammells, Bioorg. Med. Chem. Lett., 18 T. J. Collins, C. P. Horwitz, A. D. Ryabov, L. D. Vuocolo,
2010, 20, 4499–4502; (b) G. B. Kok and P. G. Scammells,
Synthesis, 2012, 2587–2594.
11 (a) G. B. Kok, C. C. Pye, R. D. Singer and P. G. Scammells, J.
Org. Chem., 2010, 75, 4806–4811; (b) G. B. Kok and
P. G. Scammells, Org. Biomol. Chem., 2011, 9, 1008–1011.
S. S. Gupta, A. Ghosh, N. L. Fattaleh, Y. Hangun,
B. Steinhoff, C. A. Noser, E. Beach, D. Prasuhn,
T. Stuthridge, K. G. Wingate, J. Hall, L. J. Wright,
I. Suckling and R. W. Allison, ACS Symp. Ser., 2002, 823,
47–60.
12 G. B. Kok and P. G. Scammells, Aust. J. Chem., 2011, 64, 19 E. S. Beach, J. L. Duran, C. P. Horwitz and T. J. Collins, Ind.
1515–1521. Eng. Chem. Res., 2009, 48, 7072–7076.
13 (a) R. J. Carroll, H. Leisch, E. Scocchera, T. Hudlicky and 20 S. Mondal, Y. Hangun-Balkir, L. Alexandrova, D. Link,
D. P. Cox, Adv. Synth. Catal., 2008, 350, 2984–2992;
(b) A. Machara, L. Werner, M. A. Endoma-Arias, D. P. Cox
B. Howard, P. Zandhuis, A. Cugini, C. P. Horwitz and
T. J. Collins, Catal. Today, 2006, 116, 554–561.
and T. Hudlicky, Adv. Synth. Catal., 2012, 354, 613–626; 21 A. Ghosh, D. A. Mitchell, A. Chanda, A. D. Ryabov,
(c) A. Machara, D. P. Cox and T. Hudlicky, Adv. Synth.
Catal., 2012, 354, 2713–2718.
D. L. Popescu, E. C. Upham, G. J. Collins and T. J. Collins,
J. Am. Chem. Soc., 2008, 130, 15116–15126.
14 C. Torborg and M. Beller, Adv. Synth. Catal., 2009, 351, 22 ESI.†
3027–3043.
23 J. D. Phillipson, S. S. Handa and J. W. Gorrod, J. Pharm.
15 D. D. Do Pham, G. F. Kelso, Y. Yang and M. T. W. Hearn,
Green Chem., 2012, 14, 1189–1195.
16 L. Q. Shen, E. S. Beach, Y. Xiang, D. J. Tshudy, N. Khanina,
C. P. Horwitz, M. E. Bier and T. J. Collins, Environ. Sci.
Technol., 2011, 45, 7882–7887.
17 (a) T. J. Collins, S. K. Khetan and A. D. Ryabov, in Handbook
of Green Chemistry: Green Catalysts, ed. R. H. Crabtree,
Wiley-VCH, 2009, vol. 1, pp. 39–77; (b) S. S. Gupta,
M. Stadler, C. A. Noser, A. Ghosh, B. Steinhoff, D. Lenoir,
Pharmacol., 1976, 28, 687–691.
24 (a) G. C. Mueller and J. A. Miller, J. Biol. Chem., 1953, 202,
579–587; (b) R. E. McMahon and H. R. Sullivan, Life Sci.,
1964, 3, 1167–1174; (c) M. M. Abdel-Monem, J. Med. Chem.,
1975, 18, 427–430; (d) J. W. Gorrod and D. J. Temple, Xeno-
biotica, 1976, 6, 265–274; (e) F. P. Guengerich and
T. L. Macdonald, Acc. Chem. Res., 1984, 17, 9–16;
(f) F. P. Guengerich, J. Med. Chem., 1984, 27, 1101–1103;
(g) F. P. Guengerich, Chem. Res. Toxicol., 2001, 14, 611–650.
C. P. Horwitz, K.-W. Schramm and T. J. Collins, Science, 25 R. Kraβnig, C. Hederer and H. Schmidhammer, Arch.
2002, 296, 326.
Pharm. Pharm. Med. Chem., 1996, 329, 325–326.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1399–1409 | 1409