Design and Development of Pd-Catalyzed Aerobic N-Demethylation Strategies for the Synthesis of Noroxymorphone in Continuous Flow Mode

Article abstract of DOI:10.1002/ejoc.201601453

Strategies for the generation of noroxymorphone from 14-hydroxymorphinone are presented. Noroxymorphone is the key intermediate in the synthesis of various opioid antagonists, including naloxone, naltrexone, and nalmefene, as well as mixed agonists-antagonists such as nalbuphine. The transformation requires removal of the N-methyl group from the naturally occurring opiates and double-bond hydrogenation. The pivotal reaction step thereby is an N-methyl oxidation with colloidal palladium(0) as catalyst and pure oxygen as terminal oxidant. The reaction produces a 1,3-oxazolidine intermediate, which can be readily hydrolyzed to the corresponding secondary amine. Different reaction sequences and the use of various phenol protecting groups were explored. The most direct route consumes only H₂, O₂, and H₂O as stoichiometric reagents and produces only H₂O as a byproduct. Challenges inherent to gas/liquid reactions with oxygen as oxidant have been addressed by developing a continuous flow process.

Full text of DOI:10.1002/ejoc.201601453

A Journal of
Accepted Article
Title: Design and Development of Pd-catalyzed Aerobic N-
Demethylation Strategies for the Synthesis of Noroxymorphone in
Continuous Flow Mode
Authors: Bernhard Gutmann, David Cantillo, Ulrich Weigl, D Phillip
Cox, and C. Oliver Kappe
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Org. Chem. 10.1002/ejoc.201601453
Link to VoR: http://dx.doi.org/10.1002/ejoc.201601453
Supported by

European Journal of Organic Chemistry
10.1002/ejoc.201601453
1
Full Paper
Design and Development of Pd-catalyzed Aerobic N-
Demethylation Strategies for the Synthesis of
Noroxymorphone in Continuous Flow Mode
[
a,b]
[a,b]
[c]
[d]
Bernhard Gutmann,
C. Oliver Kappe*
David Cantillo,
Ulrich Weigl, D. Phillip Cox and
[
a,b]
Abstract: Strategies for the generation of noroxymorphone from 14-hydroxymorphinone are
presented. Noroxymorphone is the key intermediate in the synthesis of various opioid
antagonists, including naloxone, naltrexone and nalmefene, as well as mixed agonists-antagonists
like nalbuphine. The transformation requires removal of the N-methyl group from the naturally
occurring opiates and a double bond hydrogenation. The pivotal reaction step thereby is a N-
methyl oxidation with colloidal palladium(0) as catalyst and pure oxygen as terminal oxidant.
The reaction produces a 1,3-oxazolidine intermediate, which can be readily hydrolyzed to the
corresponding secondary amine. Different reaction sequences and various phenol protection
groups were explored. The most direct route consumes only H
reagents and produces only H O as by-product. Challenges inherent in gas-liquid reactions with
oxygen as oxidant were addressed by developing a continuous flow process.
2
, O
2
and H
2
O as stoichiometric
2
[
a,b] Dr. B. Gutmann, Prof. Dr. C. O. Kappe
Institute of Chemistry, University of Graz, NAWI Graz
Heinrichstrasse 28, 8010 Graz (Austria)
and Research Center Pharmaceutical Engineering (RCPE)
Inffeldgasse 13, 8010 Graz (Austria)
E-mail: oliver.kappe@uni-graz.at
Homepage: http://goflow.at
[
[
c]
d]
Dr. U. Weigl,
Cilag AG
Hochstrasse 201, 8200 Schaffhausen (Switzerland)
Dr. D. P. Cox
Noramco Inc.
5
03 Carr Rd, Suite 200, Wilmington DE 19809 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejoc.201601453.
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
2
Introduction
In the past 15 years, the number of deaths associated with medical and nonmedical use of opioids
[1,2]
increased dramatically.
In response to the epidemic, intensified and coordinated drug
prevention efforts have been initiated. In particular, these efforts have included measures to
[1]
expand and ease access to opioid antagonists. Naloxone is the most important opioid antagonist
and it is on the World Health Organization's List of Essential Medicines. Within a few minutes, it
fully reverses the effects of many natural as well as synthetic and semi-synthetic opioids by
effectively displacing them at the μ-opioid receptor. Naloxone is an uncontrolled substance with
no potential for abuse or overdose. It is usually administered by intravenous or intramuscular
[2]
injection. In addition, intranasal formulations have become available recently. A longer-acting
opioid antagonist is naltrexone. Naltrexone and its derivative, nalmefene, are primarily used in
the management of opioid and alcohol dependence. Both naloxone and naltrexone are formally
derived from oxymorphone (2a) by N-demethylation and re-alkylation (Scheme 1). The direct
product of the N-demethylation of oxymorphone (i.e. noroxymorphone 3a, Scheme 1) is indeed
the key intermediate in the synthesis of various opioid antagonists, including naloxone,
[3-5]
naltrexone and nalmefene, as well as mixed agonists-antagonists like nalbuphine (Scheme 1).
Oxymorphone, in-turn, can be prepared by well-developed procedures from naturally occurring
opioids (i.e. oripavine and thebaine, Scheme 1). The most challenging transformation in this
reaction sequence is the selective removal of the tertiary N-methyl group from the complex and
highly functionalized morphine structure. Standard procedures for N-demethylation rely on
[3,4]
strong electrophiles to first convert the tertiary amine into a quaternary ammonium ion.
A
subsequent nucleophilic replacement of the methyl group from the ammonium ion affords the
[3]
demethylated product. Typical electrophiles are cyanogen bromide (von Braun reaction), and
[4]
chloroformates, forming respectively cyanamides and carbamates as intermediates, which are
finally hydrolyzed to the corresponding secondary amines. Alternatively, oxidation of the
alkaloid to the corresponding N-oxide (typically by peracids) and subsequent dehydration affords
[5]
the demethylated product (Polonovski-type reaction). These methods generally produce the
desired secondary amines in moderate to high yields. However, since highly toxic and corrosive
reagents in stoichiometric amounts are required and, consequently, stoichiometric amounts of
waste products are generated, these processes have become increasingly unacceptable.
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
3
1
1
R1
1
R O
R O
R O
O
C14-
hydroxylation
N
hydrogenation
-demethylation
O
O
O
O
NMe
NMe
OH
NMe
OH
NH
OH
MeO
O
O
=
O
1
1
1
R1 H: oxymorphone (2a)
R = H:
R = H:
R = H:
oripavine
14-hydroxy
1
R2 Me: oxycodone (2b)
=
3a
noroxymophone (
)
R = Me: thebaine
isolated from
1a
)
morphinone (
2
R = Me:
14-hydroxy
papaver somniferum
codeinone 1b
(
)
(opium poppy)
HO
O
HO
HO
HO
O
O
O
N
OH
N
OH
N
OH
N
OH
O
O
H C
HO
2
naloxone
naltrexone
nalmefene
nalbuphine
Scheme 1. Preparation of semisynthetic opiates from naturally occurring oripavine (or thebaine)
with 14-hydroxymorphinone (1a) and noroxymorphone (3a) as intermediates.
In 2008, Hudlicky and coworkers demonstrated that the N-methyl group of hydrocodone can be
[6]
removed by a palladium-catalyzed oxidation using oxygen as the terminal oxidant. With 10
mol% Pd(OAc) as catalyst and dioxane/acetic anhydride as solvent, N-acetyl norhydrocodone
2
was obtained after 15 h at 80 °C. In subsequent publications, the Hudlicky group has shown that
the reaction can be extended to certain other morphine alkaloids such as 3,14-diacetyl
[7,8]
oxymorphone using oxygen under atmospheric pressure (5, Scheme 2).
The demethylation of
diacetyl oxymorphone is accompanied by intramolecular 14-O- to 17-N-acetyl migration to form
[8]
the 3,17-diacetylnoroxy derivative as the immediate product (7, Scheme 2). A subsequent
[8]
hydrolysis with 6M HCl finally yielded the desired noroxymorphone 3a. (Scheme 2).
From an environmental and economic perspective, molecular oxygen is the ideal
[9]
oxidant. However, gas-liquid reactions with pure oxygen on preparative scales are nearly
absent in the pharmaceutical industry. In a review on large-scale oxidations in the pharmaceutical
industry, process research chemists at Pfizer pointed out that “perhaps the greatest factor
influencing the hesitation to employ oxidation reactions on a large scale is the safety of these
[9]
processes”. Our group has recently presented a continuous flow protocol for the palladium-
catalyzed N-demethylation of 3,14-diacetyl oxymorphone (5) and 3,14-diacetyl 14-
[10]
hydroxymorphinone (4, Scheme 2). By performing the oxidation in a continuous flow reactor,
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
4
[11,12]
process challenges inherent to gas-liquid reactions with pure oxygen are greatly alleviated.
Most importantly, combustion and explosion hazards are significantly reduced in continuous flow
[11,12]
reactors.
Indeed, the protocol developed on a laboratory scale was readily scaled up to a 100
mL flow reactor, consisting of two gas-liquid micromixers -to ensure efficient gas-liquid mixing-
[10]
and two tubular reactor modules, to provide residence volume. Oxidation of 3,14-diacetyl 14-
hydroxymorphinone (4) in the flow reactor, subsequent hydrogenation in a continuous flow
packed bed reactor, and final hydrolysis in a batch reactor afforded the desired noroxymorphone
[
10]
3
a in high quality and good overall yield on a kg-scale (Scheme 2).
Interestingly, the oxidative N-demethylation with palladium acetate as catalyst generally
fails for 14-OH unprotected 14-hydroxy opioids. However, we have recently demonstrated that,
while no reaction occurs in the presence of Pd(OAc) , the oxidation of the N-methyl group of
unprotected 14-hydroxymorphinone (1a) does proceed using palladium(0) colloids in polar
2
[13]
aprotic solvents.
The direct demethylation of 14-OH unprotected 14-hydroxy opioids has
considerable advantages. In particular, the final amide hydrolysis can be omitted when
unprotected 14-hydroxy opioids are used as starting materials (Scheme 2). The hydrolysis of the
amide requires heating of the opioid (7) to high temperatures under highly acidic conditions, and
[10]
significantly reduces the yield of the synthesis route. Herein we provide a full account of our
research on the palladium-catalyzed oxidative N-demethylation of 14-OH unprotected hydroxy
[13]
opioids. Different reaction sequences and various phenol protection groups were evaluated to
advance progress towards a selective and scalable synthesis of noroxymorphone 3a starting from
1
4-hydroxymorphinone 1a.
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
5
1
R O
HO
this work
route
envisaged
also ref.
(see
13)
O /H /H O
2
2
2
1
O
1
1
1
1
a: R =H
O
NH
OH
1
NMe
b: R =Me
OH
1
c: R =Ac
O
O
1
d: R =Bn
3
a
)
noroxymorphone (
AcO
H O
2
work:
previous
ref. 8 and
10)
6
)
(
(
see
O
AcO
AcO
NAc
2
. H2
1. O2
OH
O
O
O
NMe
OAc
NAc
AcO
OH
O
5
1
. H2
( )
2. O2
O
3
,17-diacetyl
noroxymorphone
14-
3
,14-diacetyl
O
4
hydroxymorphinone ( )
NMe
OAc
(7)
O
Scheme 2. Synthetic strategies for the generation of noroxymorphone (3a).
Oxidation/hydrogenation or hydrogenation/oxidation of diacetyl protected hydroxymorphinone 4
proceeds via 6 or 5 as intermediates.
Results and Discussion
1
. N-Demethylation/hydrogenation sequence
1
.1. Batch optimization reactions. As shown in our preliminary report, no N-demethylation
[13]
reaction occurs with 14-hydroxymorphinone 1a as substrate and Pd(OAc)
2
as catalyst.
In
contrast, an extensive screening of reaction conditions revealed that N-methyl oxidation does
proceed in polar aprotic solvents, such as N,N-dimethylacetamide (DMA) and dimethyl sulfoxide
(
DMSO), in the presence of various palladium(0) species, including palladium(0) on charcoal or
commercial unsupported palladium(0) particles (Table 1, entries 1 to 4). Thorough analysis of the
crude reaction mixture revealed that the product of the palladium catalyzed oxidation is not the
expected hydroxynormorphinone (8a, c.f. Scheme 4). Instead, the product formed in this reaction
was the 1,3-oxazolidine 9a (Table 1). The formation of related oxazolidines upon treatment of
oxycodone- and oxymorphone-N-oxide with Burgess reagent as dehydrating agent was
[14]
previously reported by Hudlicky and co-workers.
Notably, the reaction reported herein
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
6
accomplishes the formation of 1,3-oxazolidines in a one-step procedure with oxygen as the only
stoichiometric oxidant and water as the only by-product.
A faster but somewhat less selective reaction was observed with platinum(0) species, such as
platinum on charcoal (Table 1, entries 6 and 7). Solvents other than DMA and DMSO, including
dioxane, ethyl acetate, acetonitrile, toluene and butanone, afforded no conversion or low
[13]
conversions with poor selectivities.
Analogous results were obtained with 14-
hydroxycodeinone 1b (Table S1 in the Supporting Information).
Table 1. Screening of batch oxidation of 14-hydroxymorphinone 1a with Pd(0) or Pt(0) under an
[a]
O
2
atmosphere.
HO
HO
O2
catalyst,
solvent
O
O
NMe 120 °C or 140 °C
OH
N
O
O
O
1a
9a
solvent
cat [mol%]
temp
°C]
time
[min]
40
40
40
45
45
45
45
conv
[%]
73
>99
93
75
66
>99
84
sel
[b]
[b]
[
[%]
96
75
92
95
95
86
93
1
2
3
4
5
6
7
DMA
DMSO
DMSO
DMA
DMF
Pd/C(10)
Pd/C(10)
Pd/C(10)
140
140
120
140
140
140
140
Pd black (10)
Pd black (10)
Pt/C(5)
DMA
DMA
Pt/C(2.5)
[
a] Conditions: 0.2 mmol of substrate 1a and catalyst in 0.6 mL solvent were stirred under an O
2
atmosphere (balloon) on a hot-plate at 120 or 140 °C. [b] Conversion and selectivity were
determined by HPLC-UV/VIS analysis (215 nm) after derivatization with benzyl bromide (see
Scheme S1 in the Supporting Information and Experimental Section for details).
Colloidal Pd(0) particles are readily generated in solution by heating a Pd(II) source (e.g.
[15]
Pd(OAc)
2
) at temperatures above about 100 °C.
Indeed, experiments performed in our
laboratory revealed that black solutions of palladium(0) particles can be formed by heating
[13]
Pd(OAc)
2
for 2 to 10 min at temperatures of 120 to 140 °C in DMA as solvent. The addition of
acetic acid further stabilizes the colloidal palladium(0) and prevents its agglomeration and
[13]
precipitation on the vessel walls. Importantly, with 14-hydroxymorphinone 1a dissolved in the
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
7
dark solutions of colloidal palladium(0), a fast and selective oxidation to the oxazolidine 9a
occurred. It should be emphasized that Pd(OAc) is not reduced to Pd(0) under the applied
2
reaction conditions in the presence of unprotected 14-hydroxy opioids (1a,1b). Accordingly, no
oxidation reaction ensues. This is in contrast to reactions with 14-OH-protected 14-hydroxy
[10]
opioids, such as 3,14-diacetyl hydroxymorphinone or 3,14-diacetyl oxymorphone (4 and 5).
Solutions of Pd(OAc) and OH-protected 14-hydroxy opioids quickly become black upon heating
in DMA as solvent, indicating the formation of palladium(0) colloids, and a clean demethylation
2
[8,10]
reaction takes place.
prevent the formation of catalytically active Pd(0).
For oxidation reactions, Pd(OAc) and AcOH were heated in DMA for 10 min at 120 °C
Apparently, unprotected 14-hydroxy opioids stabilize palladium(II) and
2
to generate deep black solutions of finely dispersed colloidal Pd(0). 14-Hydroxymorphinone 1a
was then dissolved in this mixture and the mixture was again heated under an atmosphere of
oxygen (O balloon). A thorough optimization of the reaction conditions revealed that the
2
reaction proceeds with high selectivity to conversions >95% after reaction times of only 30 min
at 120 °C when 3 equiv of AcOH (with respect to the substrate) and 2.5 mol% of catalyst are
[13]
used. Reactions with air under atmospheric pressure instead of oxygen gave slightly reduced
reaction rates (~80 min for complete conversion at 120 °C), whereas in the absence of oxygen,
the reaction did not proceed (Table 2, entry 3 and 4). During the course of the reaction, the
oxazolidine 9a precipitated from the reaction mixture and was isolated in analytically pure form
in ca. 80% yield by filtration. Increasing the scale of the reaction decreased the reaction rate
appreciably. Nevertheless, for a reaction on a 500 mg scale with 5 mol% Pd(OAc) as pre-
2
catalyst, a conversion >95% was obtained after a reaction time of 2 h at 120 °C. The mixture was
subsequently cooled to room temperature and the precipitate was isolated by simple filtration in
8
2% yield. The corresponding reaction with 14-hydroxycodeinone 1b on a 100 mg scale yielded
the oxazolidine 9b with comparable purity after extraction with CHCl /H O (95% yield). Similar
results as with DMA/AcOH/Pd(OAc) were obtained with Pd/C in DMSO as solvent. A batch
reaction with hydroxymorphinone 1a on a 1 g scale in DMSO as solvent with 5 mol% Pd/C as
catalyst and O as oxidant (balloon) required 40 min at 140 °C for complete conversion. After the
3
2
2
2
reaction, the palladium catalyst was recovered by hot filtration. Water was then added to the
filtrate to precipitate the oxazolidine 9a in 77% yield. Interestingly, even though Pd(0) particles
are formed upon heating of Pd(OAc) in DMSO as solvent, the resulting mixture did not exhibit
2
any catalytic activity for the N-methyl oxidation of 14-hydroxymophinone (Table 2, entry 5).
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
8
Table 2. Batch oxidation of 14-hydroxymorphinone (1a) with Pd(OAc)
in DMA or DMSO as solvent (30 min reaction time).
2
as pre-catalyst at 120 °C
[a]
solvent
Pd(OAc)
2
oxidant
conv
[%]
[
b]
[
mol%]
1
2
3
4
5
DMA
DMA
DMA
DMA
DMSO
2.5
2.5
O
2
96
60
4
air
5.0
argon
argon
100.0
5.0
3
O
2
0
[
a] Conditions: 60 mg substrate, 3 equiv of AcOH, Pd(OAc)
2
in 0.6 mL solvent. Pd(OAc) and
2
the acetic acid in the solvent were heated in the absence of the substrate prior to the oxidation to
generate colloidal palladium(0) particles (10 min at 120 °C). [b] Conversion was determined by
LC-MS analysis and peak area integration of the extracted-ion chromatograms (see Experimental
Section for details).
The next reaction steps for the generation of noroxymorphone 3a are hydrogenation and
hydrolysis of the oxazolidine 9a. To explore whether the palladium(0) catalyst used for the
oxidation of the methyl group could directly serve as catalyst for the subsequent hydrogenation, a
batch reaction of 14-hydroxymorphinone 1a with 5 mol% Pd/C in DMSO as solvent under an O
2
atmosphere (balloon) was performed. Complete conversion to the oxazolidine 9a was obtained
after a reaction time of 40 min at 140 °C on a 200 mg scale. The reaction mixture was then
cooled to 50 °C and the O
2
-balloon was replaced by a H -balloon. The Pd/C catalyst was indeed
2
active for the hydrogenation. However, the main product formed after a reaction time of 90 min
was identified as the 14-hydroxymorphinone 1a (see Scheme 3 and Table S2 in the Supporting
Information). Longer reaction times converted most of the 14-hydroxymorphinone 1a to
oxymorphone 2a. The anticipated di-hydro oxazolidine product 10a (c.f. Scheme 7) was not
detected in the crude reaction mixture, indicating that hydrogenolysis of the oxazolidine ring is
faster than hydrogenation of the C7-C8 double bond of the oxazolidine 9a under these reaction
conditions.
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
9
HO
HO
HO
2
HO
2
Pd/C, O
DMSO
Pd/C, H
DMSO
Pd/C, H
DMSO
2
O
O
O
O
1
40 °C, 40 min
50 °C
50 °C
NMe
N
NMe
NMe
OH
OH
OH
O
9a
O
O
O
O
1
a
1a
2a
Scheme 3. Batch aerobic oxidation and subsequent hydrogenation of 14-hydroxymorphinone
1a) with Pd/C in DMSO.
(
To avoid hydrogenolysis of the oxazolidine ring, a reaction sequence consisting of oxidation,
hydrolysis and subsequent hydrogenation was evaluated. Hydrolysis of the oxazolidine 9a to 14-
hydroxynormorphinone (8a) was expected to proceed quickly and cleanly. Indeed, LC-MS
analysis suggested that the oxazolidine hydrolyses to the free amine on the HPLC column using
MeCN/H O with 1 vol% of trifluoroacetic acid as eluent (the retention time and recorded mass of
2
samples of oxazolidines 9 were those of the corresponding nor-derivatives 8). Surprisingly,
however, hydrolysis of the isolated oxazolidine 9a in aqueous HCl proceeded only slowly and
produced several side-products. One of the side-products arose from acid-catalyzed, conjugate
addition of water to the enone (8-hydroxy-noroxymorphone 11; c.f. Scheme 5). Further side-
products were probably derived from a reaction of the electron rich A ring of the opioid with
[16]
formaldehyde (c.f. Marquis reagent). We hypothesized that the oxazolidine is in equilibrium
with the free amine and formaldehyde. On the HPLC column the free amine and the
formaldehyde are continuously separated by chromatography, driving the hydrolysis to rapid
completion and preventing undesired side reactions. This hypothesis was further supported by the
observation that the oxazolidine is acetylated quickly and selectively to the 3,17-diacetyl
hydroxynormorphinone (6, Scheme S1 in the Supporting Information). Importantly, the reaction
was considerably faster under reduced pressure. While the conversion was 53% after 80 min at
8
0 °C at atmospheric pressure, 77% conversion was obtained after a reaction time of only 4 min
[13]
at 80 °C and a pressure of 140 mbar.
reactions.
The reaction selectivity was around 90% for these
The palladium-catalyzed oxidation and hydrolysis can be readily combined in a two-step
one pot procedure (Scheme 4). After the palladium catalyzed oxidation of 14-
hydroxymorphinone 1a, the reaction mixture was diluted with the same volume of 1N aqueous
HCl to provide a dark homogeneous solution. The reaction mixture was subsequently heated
under reduced pressure to hydrolyze the oxazolidine 9a to the 14-hydroxynormorphinone 8a (140
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
10
mbar within 20 min at 80 °C). This step was most conveniently performed in a rotary evaporator.
From a reaction on a 500 mg scale, the desired 14-hydroxynormorphinone product 8a was
isolated by precipitation with 25% ammonia in 98% yield (93.5% purity according to HPLC-
UV/Vis at 205 nm).
Finally, the full reaction sequence from 14-hydroxymorphinone 1a to noroxymorphone 3a
was explored (Scheme 4). Since the palladium from the oxidation step partly precipitates from
the mixture upon hydrolysis of the oxazolidine 9a, an additional catalyst source was required for
hydrogenation. Thus, the crude reaction mixture after oxidation and hydrolysis was pumped
through a fixed-bed flow hydrogenator using palladium catalyst cartridges (Thales H-Cube
TM [17]
Pro ).
Preliminary experiments showed that complete hydrogenation of the 14-
hydroxynormorphinone 8a to the noroxymorphone 3a was obtained with a flow rate of 0.4
mL/min with a 10% Pd/C catalyst cartridge at a reaction temperature of 25 °C and a H pressure
of 30 bar. The collected sample was diluted with distilled water and crude noroxymorphone 3a
2
was precipitated with aqueous NH
3
to produce the desired product in 70% yield (92.8% purity by
HPLC-UV/Vis at 205 nm).
HO
HO
HO
HO
0% Pd/C, H2
,
O2
2
aq HCl
-CH O
Pd(OAc)
DMA, AcOH
1
O
2
O
O
O
NMe
OH
NH
2
N
NH
OH
5 °C, 30 bar
120 °C, 120 min
80 °C, 20 min
40 mbar
OH
O
1
O
O
O
O
1a
9a
8
2
a
3a
(70%
(
98% yield,
3
steps)
yield, steps)
Scheme 4. N-Demethylation of 14-hydroxymorphinone 1a and subsequent hydrolysis and
hydrogenation to noroxymorphone 3a.
1
.2. Continuous flow oxidation in a packed bed reactor. As a first approach we explored a
continuous flow oxidation in a packed bed reactor. The experimental reactor consisted of an
HPLC pump for introducing the liquid feed (Figure 1). Oxygen gas from a gas cylinder was fed
into the system via a mass flow controller. The gaseous and liquid feed were combined in a
stainless steel T-connector, and the combined mixture then passed through a stainless steel
®
catalyst cartridge (ThalesNano CatCart ). The cartridge was packed with either 5% platinum-on-
charcoal, or 5% palladium on an alumina support. The processed mixture finally left the system
through an adjustable back-pressure regulator. Even though the catalyst bed, and thus the
residence time, was very short (residence times in the order of seconds), high single pass
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
11
conversions were obtained with 14-hydroxycodeinone 1b as starting material at reaction
temperatures of 120 to 140 °C. For instance, with Pt/C at a reaction temperature of 120 °C, a
conversion well above 90% was obtained with flow rates of 0.3 and 5 mL/min (gas flow at
normal conditions) for the liquid and O phase, respectively (Table S3 in the Supporting
2
Information). However, the effluent mixture was colored deep black, indicating leaching of
catalyst from the bed, and the performance of the catalyst quickly deteriorated for successive
reactions. Furthermore, even though HPLC analysis suggested good selectivity, the product
content in the crude reaction mixture after evaporation of the solvent was only ~45% (determined
1
by H-NMR with internal standard). Presumably, some of the product was adsorbed on the
catalyst bed. Reactions with Pd/Al
2
3
O provided slightly lower conversions. Again, catalyst
performance decreased rapidly (Table S3 in the Supporting Information).
MeO
MeO
O2
O
O
NMe Pt/C or Pd/Al O
N
2
3
OH
1
20 or 140 °C
~20 bar
O
O
O
1
b
9b
Figure 1. Packed bed reactor set-up.
1
.3. Continuous flow oxidation in a gas-liquid tubular reactor. For further reactions, the
catalyst cartridge was replaced by a residence tube reactor (Figure 2). For these reactions,
Pd(OAc) was heated in the presence of AcOH for 10 min to 120 °C to generate palladium(0)
2
colloids. 14-Hydroxymorphinone 1a was then dissolved and the mixture was loaded into the
injection loop of the flow system. As residence reactor we employed a thick-walled, gastight
fluoropolymer tube with 1.6 mm inner diameter (RT in Figure 2; fluorinated ethylene propylene
(
FEP), 1/8’’ o.d., 1.6 mm i.d.). In order to transport the reaction mixture through the back
pressure regulator as a homogeneous solution, 1N HCl was fed into the reactor before the back
[18]
pressure regulator. Any oxazolidine precipitate moved through the FEP reactor and dissolved
upon mixing with aqueous HCl before the pressure regulator. No buildup of pressure or pressure
oscillations were observed for flow reactions on the intended scales (up to ~1 g). The reaction
mixture finally exited the system through a short cooling loop (stainless steel, 1/16’’ o.d., 1 mm
i.d.) and an adjustable back-pressure regulator.
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
12
O2
HO
O
HO
RT
Pd(OAc)
2
AcOH, DMA
T
O
NMe
OH
1a
N
O
O
O
9
a
1N HCl
Figure 2. Schematic diagram of the flow reactor for aerobic oxidation.
Initial experiments were performed with a reactor of 10 mL residence volume and flow rates of 5
and 0.5 mL/min for the O and liquid feed, respectively (stoichiometric ratio of O to 1a = 1.3).
Conversions of around 90% with selectivities for the oxazolidine well above 90% were obtained
after residence times of only ~10 min at 120 °C employing 1.25 mol% Pd(OAc) as pre-catalyst
Table 3, entry 1; Table S4 in the Supporting Information). Additional flow reactions were
2
2
2
(
performed in a reactor of 57 mL residence volume. Consistent conversions of ~90% to 96% were
obtained with back pressures of around 6 to 10 bar and residence times of 10 to 25 min (Table 3).
Increasing the amount of Pd(OAc) from 1.25 to 2 mol% did not increase the conversion
2
significantly, but the purity of the reaction decreased appreciably (Table 3, entries 2 to 4).
Furthermore, increasing the residence time by decreasing the flow rate did not drive the reaction
to higher conversions, while increasing the reaction temperature to 130 °C decreased the
conversion (Table S4 in the Supporting Information). Finally, varying the pressure (Table 3,
entries 5 to 8), or the amount of O did not have any noticeable effect (Table S4 in the Supporting
2
Information). Possibly, palladium is slowly oxidized to an inactive Pd(II) species and, as a
consequence, the reaction comes to an end. As described above, Pd(II) is generally not active for
the oxidation reaction and does not form catalytically active Pd(0) in the presence of 14-hydroxy
opioids. Large amounts of oxazolidine 9a precipitate formed for reactions performed at reaction
temperatures below 120 °C and ultimately clogged the flow reactor.
After the reaction, the processed reaction mixtures were heated for 20 min in a rotary
evaporator at 80 °C at a pressure of 140 mbar. Precipitation with ammonia after diluting with
water and subsequent filtration provided the crude 14-hydroxynormorphinone (8a) as a solid in
yields of around 70% (Table S4 in the Supporting Information). Direct hydrogenation in the
TM
Thales H-Cube Pro (see above) after a flow oxidation and hydrolysis on a 600 mg scale,
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
13
provided 403 mg of the crude product after precipitation with aqueous NH (70% yield). The
3
purities of the isolated products were comparable to those obtained for batch reactions.
Table 3. Flow oxidation of 14-hydroxymorphinone (1a) with Pd(OAc)
2
at 120 °C in DMA as
sel
[a]
solvent (stoichiometric ratio of O
res Pd(OAc)
volume [mol%]
2
to 1a = 1.3).
2
flow rate
p
res
conv
[
b]
[c]
[c]
O
2
/liquid
[bar] time [%]
[min]
[%]
[
mL]
10
57
57
57
57
57
57
57
[mL/min]
5/0.5
20/2
1
2
3
1.25
1.25
1.5
9
10
10
10
8
7
6.5
6
11
13
15
13
27
24
21
19
89
93
93
89
85
90
94
94
94
91
92
95
96
96
96
94
[d]
20/2
20/2
10/1
10/1
10/1
10/1
4
5
2
1.25
1.25
1.25
1.25
[e]
6
7
8
[
a] Conditions: substrate (600 mg), AcOH (3 equiv), DMA (6 mL) and Pd(OAc) ; the reaction
2
mixtures were heated in the absence of the substrate for 10 min at a temperature of 140 °C before
the continuous flow reaction to generate colloidal Pd(0) particles. [b] Flow rate gaseous and
liquid phase (gas flow at normal conditions, i.e. T
n
= 0 °C and p
n
= 1 atm); 1.3 equiv of oxygen.
[
c] Conversion and selectivity was determined by HPLC-UV/Vis peak area integration at 215 nm
after derivatization with Ac O (see Experimental Section for details). [d] Substrate (1200 mg). [e]
2
Experiment was performed in triplicate providing identical results.
1
.4. Phenol protection. One of the main impurities generated during the continuous flow
oxidation was identified as a dimer derived from oxidative phenol coupling (Table S4 in the
Supporting Information). In order to reduce oxidative dimerization, reactions with 3-OH
protected 14-hydroxymorphinones were explored. 3-Acetyl-14-hydroxymorphinone 1c can be
prepared in high selectivity and close to quantitative yield from14-hydroxymorphinone 1a by
acetylation with 1 equiv of Ac
2
O in the presence of 1 equiv of K
2
CO in THF as solvent (97%
3
[14a]
product yield).
Importantly, both 1c and the corresponding oxazolidine 9c are soluble in DMA
at the envisaged concentrations.
Continuous flow oxidations were performed in the same setup as shown in Figure 2, but
without the HCl quench feed. The palladium-catalyzed oxidations with the 3-acetyl protected 14-
hydroxymorphinone 1c proceeded essentially as fast as reactions with unprotected 1a.
Conversions >90% were obtained after residence times of ~20 min with 1.25 mol% palladium
acetate as pre-catalyst. The reaction was very clean with the main side-product being the 3,17-
diacetyl-14-hydroxynormorphinone (6). This side-product probably arises from a transfer of the
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
14
acetyl group from the phenolic OH to the amine. Similarly as for reactions with unprotected 14-
hydroxymorphinone 1a, increasing the residence time or increasing the catalyst loading did not
improve conversion (Table 4). Furthermore, varying the pressure or the equivalents of O did not
2
have an appreciable effect on reaction conversion or reaction selectivity (for further results see
Table S5 in the Supporting Information). Significantly lower conversions were obtained when
substrate, Pd(OAc) , and acetic acid in DMA as solvent were directly injected into the flow
2
reactor, indicating the importance of prior formation of palladium(0) particles (Table S5 in the
Supporting Information). The reaction was finally repeated under the best conditions (2.5 mol%
Pd(OAc) , 6 equivalents of AcOH) on a 6 gram scale. The product was formed with good
2
selectivity and high conversion (98%). Again, the main side-product in this reaction was the 3,17-
diacetyl-14-hydroxynormorphinone 6 (4.2%). Figure S1 in the Supporting Information displays
the HPLC-UV/Vis chromatogram of the crude reaction mixture.
[a]
Table 4. Flow oxidation of 3-Acetyl-14-hydroxymorphinone 1c with oxygen.
AcO
AcO
AcO
RT
Pd(OAc)2
AcOH, DMA
T
O
O
+
O
NMe
N
NAc
OH
OH
1c
O
9c
O
O
O
6
O2
Pd(OAc)
2
flow rate temp
p
[bar]
res
conv
sel
[%]
6[c]
[
b]
[c]
[
mol%]
O
2
/liquid [°C]
time [%]
[min]
[%]
[
mL/min]
1
2
3
4
1.25
1.25
2.0
10/1
5/0.5
10/1
120
120
120
110
7
7
7
7
22
52
23
23
93
97
95
96
97
1.2
1.5
2.0
1.2
92
91
88
2.0
10/1
[
a] Conditions: Ac-14-HM (600 mg), AcOH (3 equiv), DMA (6 mL) and Pd(OAc)
min at 140 °C; pressure: 7 bar. [b] Conversion was determined by HPLC-UV/Vis peak area
integration at 215 nm after derivatization with Ac O. [c] Selectivity was determined by HPLC-
UV/Vis peak area integration at 215 nm before derivatization with Ac O.
2
; activation 20
2
2
As anticipated, the phenolic acetyl group was not removed under the conditions used for
hydrolysis of the oxazolidine 9a. Diluting the mixture after a flow reaction with the same amount
of 1 N HCl and subsequent heating for 20 min at 80 °C in a rotary evaporator (140 mbar) resulted
in only 22% acetyl hydrolysis. For further preliminary hydrolysis reactions, the mixture after
oxidation was diluted with the same amount of aqueous sulfuric acid and heated in a rotary
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
15
evaporator. An essentially complete hydrolysis was obtained after heating the reaction mixture
with 8 M H SO at 80 °C for 30 min. Unfortunately, 8-hydroxy-noroxymorphone 11, the product
2
4
from conjugate addition of water to the enone, formed in large amounts (up to ~30%; Scheme 5
and Scheme S2 in the Supporting Information).
AcO
HO
AcO
HO
,
O2
2
Pd(OAc)
AcOH
aq H
SO
2 4
O
DMA,
O
O
O
+
NMe
OH
NH
OH
N
NH
OH
8
0 °C, 30 min
O
140 mbar
O
O
O
8a
O
OH 11
1c
9c
Scheme 5. Oxidation of 14-hydroxymorphinone 1c and subsequent hydrolysis.
To avoid the harsh reaction conditions during hydrolysis, reactions with 3-benzyl (Bn) protected
4-hydroxymorphinone were attempted. 3-Benzyl-14-hydroxymorphinone 1d can be prepared in
1
[19]
close to quantitative yields in a simple one-step literature procedure. Flow oxidation reactions
under the conditions described above, with 2.5 mol% Pd(OAc) on scales from 0.6 g to 6 g,
2
yielded consistent conversions of 97-98% after a residence time of around 20 min (Scheme 6).
The selectivity for the oxazolidine 9d was very high and hardly any side-products were detected
1
according to HPLC and H-NMR analysis (97% to 99% selectivity according to HPLC; see
Figure S2 in the Supporting Information).
After the continuous flow oxidation, the same volume of 1N HCl was added to the crude
reaction mixture and the sample was heated for 20 min at 80 °C in a rotary evaporator at 140
mbar. The resulting reaction mixture was directly pumped through the H-Cube Pro using a 10%
Pd/C catalyst cartridge (30 °C, 30 bar H pressure). HPLC-UV/Vis and HPLC-MS analysis
2
revealed that the 3-Bn-14-hydroxynormorphinone 8d was completely consumed. The main
product was the 3-Bn-noroxymorphone 3d (81%, Scheme 6). Additionally, significant amounts
of the alcohol 12, formed by hydrogenation of the ketone, were observed. Only 6% of the desired
noroxymorphone 3a were formed in this reaction (Scheme 6).
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
16
BnO
O
N
O
BnO
O
HO
BnO
BnO
9d
oxidation
1)
O
2) hydrolysis
O
O
O
+
+
NMe
OH
3
) hydrogenation
NH
OH
NH
OH
NH
OH
O
1d
O
O
HO
3a
3d
12
(7%)
(6%)
(81%)
Scheme 6. Oxidation and hydrolysis of 3-Bn-14-hydroxymorphinone 1d and subsequent
hydrogenation.
Overreduction of the ketone is facilitated by the highly acidic conditions after hydrolysis. Thus,
the product of a batch oxidation on a 0.34 mmol scale, and subsequent hydrolysis in the rotary
evaporator, was isolated by precipitation with MeCN. Drying in a vacuum oven overnight at 50
°C afforded 110 mg of 3-Bn-14-hydroxynormorphinone 8d in excellent purity (as the HCl salt).
A sample of 100 mg of this product was dissolved in 10 mL MeOH and 28 mg of Pd/C (10
mol%) were added. The mixture was stirred for 4 hours at room temperature under an atmosphere
of H (balloon) and Pd/C was subsequently removed by filtration. Removing MeOH under
2
reduced pressure afforded the noroxymorphone 3a in high purity (75% yield, 2 steps; see Figure
S3 in the Supporting Information).
2
2
. Hydrogenation/demethylation sequence
.1. Continuous flow oxidation in a packed bed reactor. Even though the final product was
formed with excellent selectivity employing 3-benzyl-14-hydroxymorphinone 1d as the starting
material, the use of a protection group and the need to isolate the intermediate 3-Bn-14-
hydroxynormorphinone 8d reduces atom economy and increases process complexity. A sequence
consisting of hydrogenation of 14-hydroxymorphinone 1a to oxymorphone 2a and subsequent N-
demethylation to noroxymorphone 3a could potentially allow the consecutive usage of the same
palladium catalyst for both hydrogenation and oxidation. Furthermore, side-products arising from
double bond hydration are avoided. We initially envisaged a process with two consecutive
packed bed reactors. The concentrations of the mixtures were reduced to 0.1 M for the following
reactions, to keep starting material and the corresponding oxazolidine in solution and,
additionally, to reduce formation of side-products from oxidative dimerization. Hydrogenation of
1
4-hydroxymorphinone 1a in the first packed bed reactor and oxidation in the second packed bed
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
17
reactor was expected to furnish oxazolidine 10a, which can be directly hydrolyzed to the desired
noroxymorphone 3a (Scheme 7).
HO
HO
HO
H2
O2
O
O
O
NMe
OH
Pd
NMe
OH
N
Pd
O
O
O
O
2a
10a
1a
Scheme 7. Initially envisaged process for hydrogenation/oxidative cyclization sequence for the
preparation of 10a using heterogeneous catalysts.
For the initial experiments, hydrogenation of the 14-hydroxymorphone 1a was carried out in
DMA as solvent over Pd/C (Scheme 7). The resulting solution of oxymorphone 2a was directly
subjected to continuous flow oxidation using a fixed bed reactor setup similar to that described in
®
Figure 1. As bed reactor a column (Omnifit ) filled with sol-gel entrapped SiliaCat DPP-Pd as
heterogeneous catalyst was used. The catalyst consists of an organosilica matrix functionalized
with diphenylphosphine ligands bound to Pd(II). SiliaCat DPP-Pd has been used for a wide range
[20]
of Pd-catalyzed organic transformations, and, importantly, has shown a high resistance against
catalyst leaching for reactions involving both heterogeneous Pd(0) and homogeneous Pd(II)
[21]
species in the reaction mechanism.
Preliminary experiments were performed using a 0.7 mL packed bed reactor containing
20 mg of the immobilized catalyst (0.055 mmol Pd). As commercial Siliacat DPP-Pd is initially
2
in a Pd(II) form, the catalyst was pre-activated by pumping a solution of isopropanol or ethylene
glycol (EG) in DMA (50% vol) through the column at 120 ºC. During this process, a rapid
change in the catalyst color from orange to black could be visually observed, indicating the
transformation from Pd(II) to Pd(0) (Figure 3). Ultimately, ethylene glycol was selected as
additive for the catalyst activation, as more reproducible results were obtained compared to
iPrOH. When a solution of oxymorphone 2a in DMA was processed using the packed bed reactor
with the activated Pd catalyst (using a flow rate of 0.4 mL/min for the liquid phase, and 2.05
mL/min for O
2
; 1.5 mol equiv of O ) 30% conversion to the desired 1,3-oxazolidine 10a was
2
obtained at 120 °C (HPLC analysis, 205 nm). However, HPLC monitoring of the reaction output
during a 5 mL run revealed a rapid decrease in the reaction conversion (see Figure S4 in
Supporting Information). Moreover, the immobilized catalyst had partially recovered its initial
orange coloration. After a fresh “catalyst activation” cycle using ethylene glycol in DMA at high
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
18
temperatures the oxidation could be repeated with identical results. This continuous flow
oxidation/catalyst reactivation cycle could be repeated many times without an apparent drop in
catalytic efficiency (see Figure S4 in the Supporting Information). We therefore ascribe the
observed decrease in conversion in each cycle to the transformation of the immobilized Pd in the
support to inactive Pd(II) species instead of leaching.
(a)
O
O Si
O
P(Ph)
PdX2
P(Ph)
2
O Si
O
2
n
SiliaCat DPP-Pd
(b)
iPrOH or EG
in DMA
(1:1)
120 °C
Figure 3. Structure of the SiliaCat DPP-Pd catalyst (a) and preactivation of the catalyst in the
packed-bed reactor (b).
We next decided to use a larger packed bed reactor (2.4 mL volume, 760 mg SiliaCat DPP-Pd) to
extend the residence time and improve the reaction conversion. Moreover, ethylene glycol was
added as co-solvent directly to the reaction mixture to improve catalyst stability. Thus, the
catalyst is reactivated concurrently with the reaction. With the larger packed-bed reactor, high
conversions and excellent selectivities were obtained (Figure 4) and, importantly, improved
catalyst stability. Using 10% EG in DMA as solvent system (Figure 4a) an initial conversion of
9
7% (HPLC area) with 98% selectivity was obtained. Only minor amounts of two side products
were observed: the over-oxidized carboxamide 13 and the oxazolidine 9a from a palladium-
catalyzed aerobic dehydrogenation. No traces of the dimer from oxidative phenol coupling were
detected in the crude reaction mixture (Figure S5 in the Supporting Information). A sample of 10
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
19
mL reaction volume was processed with a stable selectivity of 97-98%. The conversion dropped
slightly to 93% over the course of the reaction. Increasing the amount of EG to 20% somewhat
improved catalyst stability (Figure 4b). However, a slightly higher amount of carboxamide 13
was observed by HPLC monitoring, thus decreasing the reaction selectivity.
Figure 4. HPLC monitoring (205 nm) for the continuous flow oxidation of 2a to 10a using
ethylene glycol as additive. (a) 10% EG in DMA; (b) 20% EG in DMA. 10 mL reaction volume
was processed for each experiment.
An important advantage of this methodology is that there is no need of removing of the metal
catalyst after the reaction. Thus, the reaction mixture collected from the output was simply
concentrated under reduced pressure until ca. 20% of its initial volume, and after addition of cold
water the desired oxazolidine 10a crystallized in 68% yield.
To further evaluate the catalyst stability a longer run (50 mL) was performed using the
same catalyst cartridge. HPLC monitoring of the crude reaction mixture revealed a gradual
decrease in the reaction conversion (Figure 5), going below 90% after 15 mL had been processed.
To assess whether the drop in catalytic efficiency was due to catalyst leaching from the solid
support or inefficient reactivation of the Pd(0) by EG, the crude reaction mixture was analyzed by
ICP-MS. The results revealed that the crude solution contained 46.4 ppm Pd, equating to a total
amount of 2.1 mg of Pd leached from the solid support. This value corresponds to 10% of the
initial total amount of Pd in the catalyst bed. These data clearly unveil that the metal does leach
from the support during the oxidation of oxymorphone 2a to 10a. Even though the procedure
presented herein appears to be very useful for small scale reactions, the scalability of the process
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
20
using this supported catalyst would certainly be problematic, and a process relying on a
homogeneous Pd source may thus be preferable.
Figure 5. HPLC monitoring (205 nm) for the continuous flow oxidation of 2a to 10a during a 50
mL run.
2
.2. Continuous flow oxidation in a gas-liquid tubular reactor. Additional continuous flow
experiments were performed in a FEP tube reactor (Figure 2) with palladium(0) colloids (tube
reactor: 1.6 mm inner diameter, 28 mL volume). For a typical experiment Pd(OAc) (1-3 mol%)
2
and AcOH were dissolved in DMA (5 mL). The solution was heated under stirring at 120°C for
ca. 15 min, resulting in a dark mixture containing the colloidal Pd(0). Upon cooling, 14-
hydroxymorphinone 1a (1 mmol) was added and the mixture stirred under H atmosphere to
2
generate the oxymorphone 2a. As expected, hydrogenation of 14-hydroxymorphinone 1a to
oxymorphone 2a using the in-situ generated colloidal Pd(0) performed very well, and full
conversion was observed after 1 hour at atmospheric pressure for catalyst loadings ranging from
1
mol% to 3 mol%. HPLC analysis of the reaction mixture confirmed a highly selective reduction
of the double bond, with no traces of side products being observed. The reaction mixture was
subsequently directly processed in the oxidation flow setup, introducing the homogeneous crude
mixture to the reactor via a sample loop. Processing of the crude hydrogenation reaction mixture
through the aerobic oxidation flow setup resulted in formation of the desired 1,3-oxazolidine
derivative 10a with good conversions and acceptable selectivity, demonstrating that the colloidal
Pd(0) can be utilized both for the hydrogenation and oxidation step in an “one-pot” procedure
(
Figure S6 in the Supporting Information). Several reaction parameters were screened to obtain
optimal reaction conditions for the oxidation step (Table S6 in the Supporting Information). The
amount of AcOH used as additive had a significant influence on the conversion, the optimal
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
21
amount being 3 equivalents. Higher excess or smaller amounts clearly reduced the conversion.
Notably, conversions above 90% were only achieved with catalyst loadings of 3 mol% after
reaction times of approximately 20 to 30 minutes.
HPLC monitoring of the reaction sequence after the first and second step revealed that
several side-products were formed during the oxidation, contributing to the moderate selectivity
(
80-90%) observed in most cases (Figure S6 in the Supporting Information). The main side
products were identified as 1,3-oxazolidine 9a (ca. 4% HPLC area), formed by palladium
catalyzed aerobic dehydrogenation, and the dimer of 10a, derived from oxidative phenol coupling
(
ca. 3% HPLC area). In an attempt to improve the selectivity for the formation of 10a, reactions
at lower temperatures were performed (80 °C to 100 °C). Thus, to obtain good conversions the
reaction mixture was processed several times through the flow reactor. Notably, this strategy was
unsuccessful and the reaction stopped after a number of passes in all cases (Figure S7 in the
Supporting Information). These results support the hypothesis that the colloidal Pd(0) particles
are deactivated in the presence of oxygen at high temperatures by the formation of an in-active
Pd(II) species, such as Pd(OAc) .
2
A second possible source for the rather low activity of the catalyst in these reactions could be
aggregation of the Pd(0) particles during the hydrogenation of 1a. To test the catalytic activity of
freshly generated colloidal Pd(0), continuous flow oxidation reactions were performed with the
isolated oxymorphone 2a. Colloidal palladium particles were generated from Pd(OAc) in
2
DMA/AcOH as described above, and the oxymorphone 2a was dissolved into this solution.
Gratifyingly, using this approach, 99% conversion and 94% selectivity for the oxazolidine 10a
were obtained after 20 min residence time at 100 °C (for a HPLC-UV/Vis chromatogram of the
crude reaction mixture see Figure S8 in the Supporting Information). After concentration of the
crude reaction mixture to ca. 20% of the initial volume and addition of cold water, oxazolidine
1
0a crystallized in excellent purity and 77% product yield.
Conclusion
Herein we presented various strategies for the generation of noroxymorphone from 14-
hydroxymorphinone. Noroxymorphone is the key intermediate in the synthesis of various opioid
antagonists, including naloxone, naltrexone and nalmefene, as well as mixed agonists-antagonists
like nalbuphine. The crucial reaction step towards generation of noroxymorphone is an aerobic
N-methyl oxidation using palladium(0) as catalyst. Palladium(0) particles of high catalytic
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
22
activity were generated by heating Pd(OAc) in DMA/AcOH. With the 14-hydroxy opioids
2
dissolved in this solution, a quick and selective N-methyl oxidation to 1,3-oxazolidines ensues in
the presence of O . The reaction consumes only O as stoichiometric reagent and produces
merely H O as by-product. Molecular oxygen is the cheapest, most abundant, and “greenest”
2
2
2
oxidant. The deployment of catalytic processes with oxygen as terminal oxidant, instead of
traditional processes based on stoichiometric oxidizing agents, offers significant potential to
reduce waste emission. However, heating organic reagents and solvents with oxygen presents
serious safety hazards, particularly on larger scale. These issues were addressed by developing a
continuous flow process.
As the reaction sequence we initially envisaged to perform first the continuous flow N-
methyl oxidation, followed by a direct continuous flow hydrogenation, utilizing the same
palladium(0) catalyst for both transformations. This approach failed since C-O bond
hydrogenolysis proceeded faster than C=C bond hydrogenation (Scheme 8). To avoid
regeneration of the starting material, the 1,3-oxazolidine intermediate had to be hydrolyzed
before the double bond hydrogenation (Scheme 8). Hydrolysis was performed under reduced
pressure to avoid side-reactions with the released formaldehyde and to drive the reaction to
completion under mild conditions. The palladium catalyst was partly lost during hydrolysis. The
oxidation, hydrolysis, hydrogenation sequence yielded the desired noroxymorphone with good
yield and purity. The main side-product was a dimer from oxidative coupling. Protection of the
phenol moiety of 14-hydroxymorphinone increases stability towards oxidative coupling and
improves solubility of the 1,3-oxazolidine. The removal of the phenol protecting group in the
course of the following transformations demanded harsher reaction conditions in the respective
reaction steps (Scheme 8).
Finally, a hydrogenation, oxidation sequence allowed employing the same palladium
catalyst for both transformations, consuming only hydrogen and oxygen in the process (Scheme
8
). However, the palladium(0) catalyst after hydrogenation had lower activity than freshly
prepared catalyst, probably due to some aggregation of the palladium colloids. With freshly
prepared palladium colloids, on the other hand, the oxidation of oxymorphone proceeded under
particularly mild reaction conditions in the continuous flow reactor (100 °C, 20 min), and side-
reactions, such as dehydrogenation to the enone or oxidative dimerization, were suppressed. Even
though the reaction sequences presented herein appears to be clearly preferable to existing
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
23
synthesis routes, further analysis, including evaluation of potential towards optimization and life-
cycle assessment, is needed to evaluate ecological-economic viability.
RO
O
RO
NH
2
. H O
2
OH
a-d
O
O
8
3. H2
1
. O2
N
x
O
RO
RO
HO
2
. H
O
2
2
. H2
9
a-d
3
. H O
O
O
2
O
NMe
N
NH
HO
OH
O
OH
1
. H2
2.
O
2
O
O
O
1a-d
10a
3a
O
R = H 1a
(
)
NMe
OH
O
2a
Scheme 8. Synthesis routes to noroxymorphone 3a
Experimental Section
1
13
General remarks. H-NMR spectra were recorded on a Bruker 300 MHz instrument. C-NMR
spectra were recorded on the same instrument at 75 MHz. Chemical shifts (δ) are expressed in
ppm downfield from TMS as internal standard. The letters s, d, t, q and m are used to indicate
singlet, doublet, triplet, quadruplet and multiplet. Analytical HPLC-UV/Vis (Shimadzu LC20)
analysis was carried out on a C18 reversed-phase (RP) analytical column (150 × 4.6 mm, particle
size 5 μm) at 37 °C using a mobile phase A (water/acetonitrile 90:10 (v/v) + 0.1 % TFA) and B
(
MeCN + 0.1 % TFA) at a flow rate of 1.5 mL/min (the following gradient was applied: linear
increase from solution 3% B to 100 % B in 17 min). Low-resolution mass spectra were obtained
on a Shimadzu LCMS-QP2020 instrument using electrospray ionization (ESI) in positive or
negative mode. All solvents and chemicals were obtained from standard commercial vendors and
were used without any further purification. The Pd(OAc)
83124).
2
was purchased from Aldrich (SKU:
6
ICP MS measurements. Liquid samples were diluted 1:100 and palladium was quantitatively
determined at m/z 105 with an Agilent 7500ce inductively coupled plasma mass spectrometer. A
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
24
calibration was performed with an external calibration curve established from 1.000 g/L of Pd
standard (CPI International). Indium served as the internal standard.
Analysis (see Scheme S1 in the Supporting Information). Method 1: Ac
HPLC vial containing 1 mL of a saturated aqueous solution of NaHCO were added 50 μL of the
crude reaction mixture and 200 μL Ac O. The vial was capped, the septum perforated with a
2
O derivatization: To a
3
2
needle, and the mixture stirred vigorously at room temperature for 20 min. The content of the vial
was then directly analyzed by HPLC-UV/Vis using the method described in the General Remarks
section.
Method 2: Benzylbromide derivatization: 25 µL of benzyl bromide were added to 50 µL of the
crude reaction mixture, and the sample was heated for 60 min at 70 °C. The sample was diluted
with 1 mL MeCN and analyzed by HPLC-UV/Vis using the method described in the General
Remarks section. The detection wavelength was set at 280 nm.
Method 3: 50 µL of the reaction mixture were diluted with 1 mL MeCN and the sample was
analyzed by HPLC-MS. The respective m/z values were extracted from the total ion current
chromatograms. The peak areas of the extracted-ion chromatograms were then used to calculate
conversions of substrate to products.
Aerobic Oxidation under Batch Conditions. The reaction with 14-hydroxymorphinone 1a on a
5
00 mg scale will serve as a representative procedure for batch oxidations: 18.7 mg of Pd(OAc)
5 mol%) and 287 µL AcOH (3 equiv) were dissolved in 5 mL DMA. The mixture was heated for
0 to 20 min to 120 °C. A deep black solution of colloidal Pd(0) is obtained. 500 mg of the
starting material were dissolved in this mixture, and the solution was again heated to 120 °C
under vigorous stirring under an atmosphere of O (balloon) or, alternatively, an atmosphere of
2
(
1
2
air (open vessel). The reaction was stopped after HPLC analysis showed >95% conversion (see
Analysis section). The reaction time depended on reaction scale but the reaction was generally
completed after reaction times of 1 to 2 h. Alternatively, palladium(0) (e.g. Pd/C) can be directly
used as catalyst in DMA or DMSO as solvent. The oxazolidine 9a, derived from 14-
hydroxymorphinone 1a, precipitated from the mixture and can be isolated by simple filtration in
high purity. Usually, the crude reaction mixtures after oxidation were directly used for further
downstream reactions and no isolation was attempted.
Continuous-flow Oxidations in a Packed Bed Reactor (Figure 1). The flow reactor consisted
of an HPLC pump for introducing the liquid feed (Uniqsis Pump Module). Oxygen gas from a
gas cylinder was fed into the system via a mass flow controller (ThalesNano Gas Module or
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
25
Bronkhorst EL-Flow). The liquid and gaseous streams were combined in a T-mixer (stainless
steel). The T-mixer was connected to the residence tube reactor via a fluoropolymer tubing
(
perfluoroalkoxy polymer (PFA), 1/16’’ o.d., 0.8 mm i.d.). The PFA tubing allowed visual
observation of the flow profile. The gas-liquid mixture then passed through the catalyst bed. As
®
®
catalyst bed either pre-backed catalyst cartridges (ThalesNano CatCart ) or Omnifit columns
filled with the desired catalyst material were used. The processed reaction mixture finally left the
system through an adjustable back pressure regulator (Swagelok (KCB1H0A2A5P60000, 0-26
bar) or Vapourtec (0-15 bar)). The oxidation of 14-hydroxycodeinone 1b on a 100 mg scale will
serve as a representative procedure for continuous flow reactions: DMA was pumped through the
flow system at a flow rate of 0.3 mL/min and O
of 5 mL /min (gas flow at normal conditions, i.e. T
). The temperature of the coil heater was set to 120 °C. The pressure was adjusted to 20 bar.
2
was introduced into the reactor with a flow rate
N
n
= 0 °C and p = 1 atm; 2.3 equivalents of
n
O
2
The substrate (100 mg) was dissolved in DMA (1 mL), and the resulting mixture was introduced
into the injection loop. When the flow system was stable, the reaction mixture was pumped from
the injection loop into the reactor. About 6 min after the reaction mixture was introduced into the
reactor, the processed product solution left the system, was collected and analyzed.
Continuous-flow Oxidations under Gas-Liquid Biphasic Reaction Conditions (Figure 2).
The flow reactor consisted of an HPLC pump for introducing the liquid feed (Uniqsis Pump
Module). Oxygen gas from a gas cylinder was fed into the system via a mass flow controller
(
ThalesNano Gas Module or Bronkhorst EL-Flow). The liquid and gaseous streams were
combined in a T-mixer (stainless steel). The T-mixer was connected to the residence tube reactor
via a fluoropolymer tubing (perfluoroalkoxy polymer (PFA), 1/16’’ o.d., 0.8 mm i.d.). The PFA
tubing allowed visual observation of the flow profile. The residence coil reactor (fluorinated
ethylene propylene (FEP), 1/8’’ o.d., 1.6 mm i.d. 10, 28 or 57 mL) was heated in a GC oven to
the desired temperature. The reaction mixture finally exited the system through a short cooling
loop and a back-pressure regulator (either an adjustable back-pressure regulator: Swagelok
(
KCB1H0A2A5P60000, 0-26 bar) or Vapourtec (0-15 bar), or a static BPR from Upchurch
Scientific). Pressure sensors were integrated into the T-mixer and directly after the pump. The
oxidation of 14-hydroxymorphinone on a 600 mg scale will serve as a representative procedure
for continuous flow reactions: DMA was pumped through the flow system at a flow rate of 1.0
mL/min and O
2
was introduced into the reactor with a flow rate of 10 mL
N
/min (gas flow at
normal conditions, i.e. T
n
= 0 °C and p = 1 atm). The temperature of the GC-oven was set to 120
n
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
26
°C. The pressure was adjusted to 7 bar. The solution of colloidal palladium(0) was prepared by
heating Pd(OAc) (1.25 mol%) and acetic acid (3 equiv) in 9 mL DMA for 20 min at 140 °C. The
2
substrate (600 mg) was dissolved in this mixture, and the mixture was introduced into the
injection loop (14-hydroxymorphinone 1a does not dissolve completely at this concentration in
DMA at room temperature, but the solution becomes homogenous upon mild heating). When the
flow system was stable, the reaction mixture was pumped from the injection loop into the reactor.
2
0 min after the reaction mixture was introduced into the reactor, the processed product solution
left the system. The products were isolated as described above or were directly used for
subsequent reactions. For reactions with 14-hydroxymorphinone 1a, a quench solution of 1 N
HCl was pumped into the reactor before the back-pressure regulator with a flow rate of 1 mL/min
(
Uniqsis Pump Module; Figure 2) and the reaction mixture was directly used for hydrolysis in a
rotary evaporator.
1
1
(
5aR,6R,8aS,8a S,11aR)-2-Hydroxy-5,5a-dihydro-7H-6,8a -
ethanofuro[2',3',4',5':4,5]phenanthro[9,8a-d]oxazol-11(11aH)-one (9a): Starting material: 14-
Hydroxymorphinone 1a (batch reaction on a 500 mg scale). For isolation, the reaction mixture
was cooled in an ice bath. The precipitate was separated by filtration and washed with cold water.
Drying in a vacuum oven overnight provided 407 mg of the oxazolidine 9a (82% yield). mp.:
1
>
300 °C (decomposition). H NMR (300 MHz, DMSO) δ 9.28 (s, 1H), 6.79 (d, J = 10.1 Hz, 1H),
6
1
3
.63 (d, J = 8.1 Hz, 1H), 6.59 (d, J = 8.2 Hz, 1H), 6.18 (d, J = 10.1 Hz, 1H), 4.74 (d, J = 6.3 Hz,
H), 4.73 (s, 1H), 4.60 (d, J = 5.5 Hz, 1H), 3.47 (d, J = 7.4 Hz, 1H), 3.23 (d, J = 18.8 Hz, 1H),
.09 (dd, J = 18.9, 7.6 Hz, 1H), 2.82 – 2.65 (m, 2H), 2.34 – 2.23 (m, 1H), 1.37 (dd, J = 12.7, 3.3
13
Hz, 1H). C NMR (75 MHz, DMSO) δ 194.9, 148.3, 143.1, 139.5, 133.9, 130.6, 122.5, 120.3,
1
18.3, 87.3, 86.5, 73.4, 65.3, 47.9, 44.1, 33.1, 26.8. HRMS (APCI): m/z: calcd for C17
H16NO
4
+
[
(M+H)] : 298.107384, found: 298.107471.
1
1
(
5aR,6R,8aS,8a S,11aR)-2-Methoxy-5,5a-dihydro-7H-6,8a -
ethanofuro[2',3',4',5':4,5]phenanthro[9,8a-d]oxazol-11(11aH)-one (9b): Starting material: 14-
Hydroxycodeinone 1b (batch reaction on a 100 mg scale). The reaction mixture was extracted
with CHCl
3
/H
2
O. Drying of the organic phase with MgSO
4
and evaporation of the solvent
provided 94 mg of the crude oxazolidine 9b (95% yield; contained some DMA; see the
1
1
Supporting Information for H-NMR). H NMR (300 MHz, CDCl
3
) δ 6.75 (d, J = 8.3 Hz, 1H),
6
2
.69 (d, J = 8.3 Hz, 1H), 6.56 (d, J = 10.1 Hz, 1H), 6.23 (d, J = 10.1 Hz, 1H), 4.79 – 4.73 (m,
H), 4.64 (s, 1H), 3.84 (s, 3H), 3.46 – 3.36 (m, 2H), 3.18 (dd, J = 19.1, 7.9 Hz, 1H), 2.88 – 2.80
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
27
1
3
(
m, 2H), 2.35 (td, J = 12.3, 6.2 Hz, 1H), 1.61 (dd, J = 12.8, 4.0 Hz, 1H). C NMR (75 MHz,
CDCl
3
) δ 194.1, 145.8, 144.2, 142.7, 134.6, 130.0, 123.4, 120.0, 115.5, 87.6, 86.9, 73.6, 65.9,
+
5
6.8, 48.1, 32.8, 27.0. HRMS (APCI): m/z: calcd for C18
12.122975.
H18NO
4
[(M+H)] : 312.123034, found:
3
(
5aR,6R,8aS,8a1S,11aR)-2-(Benzyloxy)-5,5a-dihydro-7H-6,8a1-
ethanofuro[2',3',4',5':4,5]phenanthro[9,8a-d]oxazol-11(11aH)-one (9d): Starting material: 3-
Benzyl 14-hydroxycodeinone 1d (flow reaction on a 600 mg scale). Evaporation of the solvent
provided the crude oxazolidine 9d in excellent purity (quant; contained some DMA; see the
1
Supporting Information for H-NMR).
1
H NMR (300 MHz, CDCl
H), 6.57 (d, J = 10.1 Hz, 1H), 6.24 (d, J = 10.1 Hz, 1H), 5.23 – 5.13 (m, 2H), 4.77 (q, J = 6.3
Hz, 2H), 4.68 (s, 1H), 3.47 – 3.36 (m, 2H), 3.17 (dd, J = 19.1, 7.9 Hz, 1H), 2.92 – 2.79 (m, 2H),
.46 – 2.31 (m, 1H), 1.63 (dd, J = 12.8, 3.7 Hz, 1H).
Two Step N-demethylation: Oxidation and Subsequent Hydrolysis in a Rotary Evaporator
cf. Scheme 4). The reaction mixture from the oxidation reaction (see above) was diluted with the
3
) δ 7.44 – 7.25 (m, 5H), 6.79 (d, J = 8.2 Hz, 1H), 6.65 (d, J = 8.3 Hz,
1
2
(
same amount of 1 N HCl. The mixture was then heated for 20 min at 80 °C in a rotary evaporator
at 140 mbar. The phenolic acetyl group of 9c was not removed under these conditions (see
Scheme 5 and Scheme S2 in the Supporting Information for details).
1
4-Hydroxynormorphinone (8a): Starting material: 14-Hydroxymorphinone 1a (500 mg). The
mixture after oxidation and subsequent hydrolysis was diluted with 10 mL distilled water and
brought to ~pH 9 with 25% aqueous ammonia. The precipitate was separated by filtration and
washed with cold water. Drying in a vacuum oven overnight provided 468 mg of the crude 14-
hydroxynormorphinone 8a in good purity (98% yield; 93.5% purity according to HPLC-UV/Vis
1
at 205 nm). H NMR (300 MHz, DMSO) δ 6.85 (d, J = 10.1 Hz, 1H), 6.56 – 6.48 (m, 2H), 5.99
(
d, J = 10.0 Hz, 1H), 4.57 (s, 1H), 3.07 (t, J = 3.6 Hz, 1H), 2.85 (d, J = 3.5 Hz, 2H), 2.66 (dd, J =
1
1
3
1
3.6, 4.4 Hz, 1H), 2.42 (dd, J = 13.4, 3.5 Hz, 1H), 2.26 (td, J = 12.2, 5.0 Hz, 1H), 1.27 (dd, J =
2.1, 2.6 Hz, 1H).
-Benzyl-14-hydroxynormorphinone (8d): Starting material: 3-Benzyl-14-hydroxymorphinone
d (138 mg). The product, after oxidation and subsequent hydrolysis, was precipitated with
MeCN. Drying in a vacuum oven overnight at 50 °C afforded 110 mg of 3-Bn-14-
hydroxynormorphinone 8d (probably as the HCl salt; 75% yield). Alternatively, the product was
isolated as the free base after diluting with 120 mL water and precipitation by neutralization with
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
28
1
aqueous 25% NH
.25 (m, 5H), 6.93 (d, J = 10.2 Hz, 1H), 6.89 (d, J = 8.3 Hz, 1H), 6.74 (d, J = 8.3 Hz, 1H), 6.18
d, J = 10.2 Hz, 1H), 5.15 (s, 2H), 4.87 (s, 1H), 3.99 (d, J = 4.1 Hz, 1H), 3.27-3.21 (m, 3H), 2.90
3
(4.3 g from a 6 g experiment; 74%). H NMR (300 MHz, CD OD) δ 7.40 –
3
7
(
(
1
3
td, J = 13.0, 3.8 Hz, 1H), 2.76 (td, J = 13.1, 4.7 Hz, 1H), 1.86 (dd, J = 13.0, 3.0 Hz, 1H). C
OD) δ 193.6, 146.4, 145.1, 141.9, 137.0, 133.1, 129.4, 128.0, 127.6, 127.5,
23.4, 120.2, 118.8, 86.3, 71.5, 65.9, 56.7, 46.2, 37.0, 27.5, 25.5.
NMR (75 MHz, CD
3
1
Three Step N-demethylation/Hydrogenation (cf. Scheme 4): The reaction with 14-
hydroxymorphinone 1a on a 600 mg scale will serve as a representative procedure: The
homogenous solution after oxidation and subsequent hydrolysis in a rotary evaporator (see
above) was directly pumped through the H-Cube Pro instrument (10% Pd/C catalyst cartridge;
0
.4 mL/min flow rate; 25 °C reaction temperature; 30 bar). 1.5 min after the feed was switched
from solvent (DMA) to reaction mixture, the product was collected (first product appeared after
about 5.5 min). The solution was concentrated in vacuum to around 6 mL. Afterwards the
mixture was diluted with 18 mL distilled water and brought to ~pH 9 with 25% aqueous
ammonia. The precipitate was separated by filtration and washed with cold water. Drying in a
vacuum oven overnight provided 403 mg of the crude noroxymorphone 3a in good purity (70%
yield, 92.8% purity according to HPLC-UV/Vis at 205 nm). The corresponding
oxidation/hydrogenation with Bn-14-hydroxymorphinone 1d (Scheme 6) gave incomplete
removal of the protection group and partly hydrogenation of the carbonyl group (10% Pd/C
catalyst cartridge; 0.4 mL/min flow rate; 30 °C reaction temperature; 30 bar).
1
Noroxymorphone (3a): H NMR (300 MHz, DMSO) δ 6.56 (d, J = 8.1 Hz, 1H), 6.51 (d, J = 8.1
Hz, 1H), 4.67 (s, 1H), 2.98 – 2.82 (m, 4H), 2.64 – 2.56 (m, 1H), 2.42 – 2.21 (m, 2H), 2.11 – 2.02
(
m, 1H), 1.77 – 1.67 (m, 1H), 1.41 (td, J = 14.2, 3.4 Hz, 1H), 1.14 (d, J = 11.4 Hz, 1H).
Preparation of Oxymorphone (2a). A suspension of 14-hydroxymorphinone (1a) (750 mg, 2.5
mmol) and 10% Pd/C (25 mg, 1 mol%) in DMA was stirred under an H atmosphere for 1 h.
2
Then, the reaction mixture was filtered using a syringe filter (PTFE, 0.45 μm) and the filtrate
evaporated under reduced pressure.
Sequential Hydrogenation and Continuous-flow Oxidations in a Packed Bed Reactor
(
Scheme 7). A suspension of 14-hydroxymorphinone 1a (150 mg, 0.5 mmol) and 10% Pd/C (10
mg, 1 mol%) in 5 mL od DMA/ethylene glycol 9:1 was stirred at room temperature under H
2
atmosphere for one hour. The mixture was filtered (0.45 μm pore size filter) and the clear
solution was introduced into the flow reactor via an injection loop. The flow reactor had been
This article is protected by copyright. All rights reserved.

European Journal of Organic Chemistry
10.1002/ejoc.201601453
29
stabilized by pumping DMA/ethylene glycol 8:2 at a flow rate of 0.4 mL/min and O
2
with a flow
rate of 2.05 mL/min (gas flow at normal conditions, i.e. T
N
= 0 °C and p
N
= 1 atm). The
®
temperature of the packed-bed reactor (Omnifit column) had been set to 120 °C and the pressure
adjusted to 5 bar. After a residence time of approximately 20 min the processed product solution
left the system. The reaction mixture collected from the reactor output was evaporated under
reduced pressure until ca. 20% of the initial volume. 10 mL of cold water were added, and the
solid obtained filtered and dried under vacuum at 50 °C (102 mg, 68%).
Sequential Hydrogenation and Continuous-flow Oxidations under Gas-Liquid Biphasic
Reaction Conditions. Pd(OAc) (3 mol%) and AcOH (3 equiv) were dissolved in DMA (5 mL)
2
and placed in a 25 mL two-necked round-bottom flask. The mixture was heated at 120 °C for 15
min. Formation of Pd(0) could be visually observed. Upon cooling, 14-hydroxymorphinone (150
mg, 0.5 mmol) was added and the mixture stirred under H
reaction mixture was introduced into the flow reactor via an injection loop. The flow reactor had
been stabilized by pumping DMA at a flow rate of 0.75 mL/min and O with a flow rate of 1.7
mL/min (gas flow at normal conditions, i.e. T = 0 °C and p = 1 atm). The temperature of the
2
atmosphere for 1 h. Then, the crude
2
N
N
GC-oven had been set to 120 °C and the pressure adjusted to 5 bar. After a residence time of
approximately 20 min the processed product solution left the system. The reaction mixture
collected from the reactor output was evaporated under reduced pressure until ca. 20% of the
initial volume. 10 mL of cold water were added, and the solid obtained filtered and dried under
vacuum at 50 °C (115 mg, 77%).
1
1
(
[
5aR,6R,8aS,8a S,11aR)-2-hydroxy-5,5a,9,10-tetrahydro-7H-6,8a -ethanofuro
1
2',3',4',5':4,5]phenanthro[9,8a-d]oxazol-11(11aH)-one (10a): H NMR (300 MHz, DMSO-d
6
)
δ 9.24 (s, 1H), 6.60 (q, J = 8.1 Hz, 2H), 4.87 (s, 1H), 4.61 (dd, J = 32.1, 5.5 Hz, 2H), 3.25 – 3.06
(
m, 3H), 2.91 – 2.55 (m, 3H), 2.38 – 2.24 (m, 1H), 2.20 (dt, = 7.3, 4.0 Hz, 1H), 1.89 (dt, J = 7.3,
1
3
4
.0 Hz, 1H); 1.47 (td, J = 14.8, 2.9 Hz, 1H), 1.25 (dd, J = 12.4, 4.0 Hz, 1H). C NMR (75 MHz,
DMSO) δ 208.2, 143.6, 139.7, 129.7, 122.1, 120.0, 118.1, 90.5, 86.1, 77.1, 63.6, 52.5, 44.1, 37.3,
4.4, 30.5, 26.6.
3
This article is protected by copyright. All rights reserved.