Communication
is finally released from the molecule to give palladium(0) and
hydrogen peroxide. Since palladium(II) is not reduced to palla-
dium(0) in the presence of the 14-hydroxy opioid, an alterna-
tive formation of an oxopalladium(II) species with water as
a direct byproduct seems unlikely. Rather, the hydrogen perox-
ide disproportionates to H2O and oxygen in a subsequent reac-
tion.
the oxidation, 18.7 mg of Pd(OAc)2 (5 mol%) and 287 mL of
AcOH (3 equiv) were heated in 5 mL of DMA for 10 min at
1208C to generate the colloidal palladium(0) particles. A
sample of 500 mg of 14-hydroxymorphinone (1a) was then
dissolved in this solution and the mixture was further heated
for 2 h to 1208C under an atmosphere of O2. The mixture was
subsequently diluted with 5 mL of 1n aqueous HCl to provide
a dark homogeneous solution. The hydrolysis of the crude oxa-
zolidine 2a was conveniently performed in a rotary evaporator
at a pressure of 140 mbar within 20 min at 808C. The desired
14-hydroxynormorphinone product 6a was isolated by precipi-
tation with 25% ammonia in 98% yield (93.5% purity accord-
ing to HPLC–UV/Vis at 205 nm; Figure S2 in the Supporting In-
formation).
Unfortunately, increasing the scale of the reaction decreased
the reaction rate appreciably and a reaction on a 500 mg scale
with 5 mol% Pd(OAc)2 as precatalyst required a reaction time
of 2 h at 1208C in order to reach a conversion >95%. After
the reaction, the mixture was cooled to room temperature and
the precipitate was isolated by simple filtration in excellent
purity and good yield (82%). The corresponding reaction with
14-hydroxycodeinone on a 100 mg scale yielded the oxazoli-
dine 2b with comparable purity after extraction with CHCl3/
H2O (95% yield; contaminated with ~5 wt% DMA). From
a batch reaction with hydroxymorphinone (1a) on a 1 g scale
in DMSO as the solvent using 5 mol% Pd/C as the catalyst
(40 min at 1408C), the palladium catalyst was recovered after
the reaction by hot filtration. Water was then added to the fil-
trate to precipitate the oxazolidine 2a in 77% yield.
The generation of noroxymorphone (3a)—the immediate
precursor of naltrexone and naloxone—additionally requires
hydrolysis and hydrogenation of the oxazolidine 2a
(Scheme 1). Hydrolysis of the oxazolidine 2a to the secondary
amine (i.e., 3a) was expected to proceed quickly and cleanly.
Surprisingly, however, hydrolysis of the isolated oxazolidine 2a
in aqueous HCl proceeded only slowly and several unidentified
side products were formed. In boiling water, significantly faster
reactions were obtained and, additionally, the reaction selectiv-
ity increased substantially (Table S5 in the Supporting Informa-
tion). We hypothesized that the oxazolidine is in equilibrium
with the free amine and formaldehyde. By increasing the tem-
perature above the boiling point of the solvent, formaldehyde
is effectively removed from the equilibrium. Furthermore, since
the detected side products are probably formed by a reaction
of the electron-rich A ring of the opioid with formaldehyde, ef-
ficient removal of CH2O is important both to drive the equilib-
rium to the hydrolyzed product and to prevent undesired side
reactions.[21] Indeed, the reaction was considerably faster under
reduced pressure. While the conversion was 53% after 80 min
at 808C at atmospheric pressure, 77% conversion was ob-
tained after a reaction time of only 4 min at 808C and a pres-
sure of 140 mbar. The reaction selectivity was around 90% for
these reactions (Table S6 in the Supporting Information).
The palladium-catalyzed oxidation and hydrolysis were then
combined in a two-step one-pot procedure (Scheme 3). For
Finally, the three-step reaction sequence from 14-hydroxy-
normorphinone 1a to noroxymorphone 3a was attempted
(Scheme 3). For this reaction, the crude reaction mixture after
oxidation and hydrolysis in a rotary evaporator was directly
pumped through a fixed-bed flow hydrogenator (Thales H-
Cube ProTM) to generate the noroxymorphone 3a in a three-
step reaction sequence. Preliminary experiments showed that
complete hydrogenation of the 14-hydroxynormorphinone 6a
to the noroxymorphone 3a was obtained with a flow rate of
0.4 mLminÀ1 with a 10% Pd/C catalyst cartridge at a reaction
temperature of 258C and a H2 pressure of 30 bar. The collected
sample was diluted with distilled water and crude noroxymor-
phone 3a precipitated with aqueous NH3 to produce the de-
sired 3a in 70% yield (92.8% purity by HPLC–UV/Vis at
205 nm, Figure S3 in the Supporting information).
A strong limiting factor of aerobic oxidations is typically the
low solubility of molecular oxygen in most solvents. Aerobic
oxidations, though often intrinsically fast, are frequently limited
by mass transfer and reaction rate, selectivity, and product
quality is affected by the gas–liquid interfacial area. In conven-
tional reactors, like stirred tank reactors, this parameter is diffi-
cult to control adequately. Another distinguishing challenge of
aerobic oxidation reactions is that the oxygen/vapor mixture
could spontaneously ignite. Consequently, the oxygen concen-
tration in any open space in the system must be kept below
the flammability limit (typically around 8 vol% for organic ma-
terials). Many challenges associated with gas–liquid oxidations
can be effectively addressed by continuous-flow reactors.[22,23]
Continuous flow reactors allow accurate dosing of gaseous re-
agents into the liquid feed using flow meters. Furthermore, im-
proved mixing, enhanced heat- and mass-transfer as well as
a large and well-defined interfacial gas–liquid contact area can
be achieved.[23] Thus, an experimental continuous flow reactor
was assembled. The reactor consisted of an HPLC pump for in-
Scheme 3. N-Demethylation of 14-hydroxymorphinone 1a and subsequent hydrogenation to noroxymorphone 3a.
&
&
Chem. Eur. J. 2016, 22, 1 – 7
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!