An improved oxidation of an alcohol using aqueous permanganate and phase-transfer catalyst

Article abstract of DOI:10.1021/op990021h

The oxidation of a primary alcohol to the corresponding carboxylic acid via permanganate and phase-transfer catalyst has been modified to circumvent impurity formation incurred during the reaction. A safe, simple, and efficient process was developed which gave the required carboxylic acid in 77% isolated yield using sodium permanganate in the presence of tetraethylammonium hydrogen sulphate.

Full text of DOI:10.1021/op990021h

Organic Process Research & Development 1999, 3, 363−364
An Improved Oxidation of an Alcohol Using Aqueous Permanganate and
Phase-Transfer Catalyst
Arshed Mahmood, Graham E. Robinson, and Lyn Powell*
Process DeVelopment Department, ZENECA Pharmaceuticals, Silk Road Business Park,
Macclesfield, Cheshire SK10 2NA, UK
Abstract:
Results and Discussion
The oxidation of a primary alcohol to the corresponding
carboxylic acid via permanganate and phase-transfer catalyst
has been modified to circumvent impurity formation incurred
during the reaction. A safe, simple, and efficient process was
developed which gave the required carboxylic acid in 77%
isolated yield using sodium permanganate in the presence of
tetraethylammonium hydrogen sulphate.
The reaction as described by Dmowski et al. was
performed in our laboratory.
PFP was added dropwise to an aqueous solution of
potassium permanganate and phase-transfer catalyst (PTC)
at 60 °C over 4 h. Subsequent analysis of the reaction mixture
by GC showed the oxidation to be complete, but NMR
analysis of the resulting product showed significant levels
of impurities derived primarily from over-oxidation of both
the catalyst and the desired PFPeA (Scheme 2).
Scheme 2. Formation of over-oxidation byproducts
Introduction
During studies on a key intermediate in the manufacture
of a development drug, we had to synthesise kilogram
quantities of 4,4,5,5,5-pentafluoropentanoic acid (PFPeA).
In particular, the product had to be obtained in high purity
(>97%). Oxidation of the commercially available 4,4,5,5,5-
pentafluoropentanol (PFP) looked particularly attractive
(Scheme 1).
Scheme 1. Formation of pentafluoropentanoic acid from
Our early oxidations of PFP with aqueous potassium
permanganate in the presence of tetrabutylammonium hy-
drogen sulphate revealed the reaction to be highly exothermic
with a heat output of ∼157 kcal per mole. Addition of
aqueous potassium permanganate in aliquots to a mixture
of PFP, water, and catalytic quantities of tetrabutylammo-
nium hydrogen sulphate at 65-70 °C gave the required
product whilst safely dissipating the heat liberated in the
oxidation. However, it became apparent from close scrutiny
of the PFPeA produced that significant quantities of butanoic
acid (1) and pentafluorobutanoic acid (2) were being formed.
Reduction of the phase-transfer catalyst from 3.68 to 1.25
mol % lowered the amount of over-oxidised impurity derived
from the catalyst (butanoic acid) by 50%. The reaction rate
was not affected by the reduced catalyst charge. Further
reduction of the PTC charge to 0.8 mol % gave minimal
amounts of the butanoic acid whilst still giving a good yield
of PFPeA. However, the butanoic acid generated still proved
difficult to remove from the product to the desired level
(<0.5%w/w).
In an attempt to completely avoid the over-oxidised
product derived from the PTC, tetrabutylammonium hydro-
gen sulphate was replaced with tetraethylammonium hydro
gen sulphate. We reasoned that the latter PTC would lead
to acetic acid formation, and we were confident that this
byproduct would be removed by an aqueous workup
procedure. Under these conditions we were successful in
producing the desired PFPeA free of acetic acid. Formation
of pentafluorobutanoic acid was unavoidable but minimised
pentafluoropentanol
-
MnO4
CF3CF2CH2CH2CH2OH
CF CF CH CH CO H
3 2 2 2 2
8
PFP
PFPeA
The synthesis of the desired carboxylic acid from the
alcohol by chromium trioxide has been reported by Larsson,
Carlson, and Leroy,1 and initially a small quantity of the
desired material was made in this way but in poor yield and
contaminated by acetic acid. The generation of large quanti-
ties of chromium effluent by this process also made it
unattractive for scale-up beyond a relatively small scale.
Oxidation involving permanganate and phase transfer re-
agents are high yielding and well documented.2,3 Of these,
the oxidation of fluorinated alcohols to carboxylic acids via
permanganate in the presence of tetrabutylammonium hy-
drogen sulphate as phase-transfer catalyst (PTC) was of
particular interest. This has previously been reported by
Dmowski and co-workers4 and looked amenable to scale-
up due to the relatively innocuous nature of the manganese
dioxide byproduct.
* To whom correspondence should be addressed. E-mail: Lyn.Powell@
ALDERLEY.ZENECA.com.
(1) Larsson, U.; Carlson, R.; Leroy, J. Acta Chem. Scand. 1993, 47, 380-
390.
(2) Fatiadi, A. Synthesis 1987, 85-127.
(3) Herriot, A.; Picker, D. Tetrahedron Lett. 1974, 1511-1514.
(4) Dmowski, W.; Plenkiewicz, H.; Piasecka-Maciejewska, K.; Prescher, D.;
Schulze, J.; Endler, I. J. Fluorine Chem. 1990, 48, 77-84.
10.1021/op990021h CCC: $18.00 © 1999 American Chemical Society and The Royal Society of Chemistry
Published on Web 06/23/1999
Vol. 3, No. 5, 1999 / Organic Process Research & Development
•
363

Scheme 3. Formation of alkene impurity and
decarboxylation/elimination mechanism to
trifluoromethylpropyne
were obtained on a JEOL GX 270 MHz spectrometer.
Chemical shifts are reported in ppm (δ) relative to residual
chloroform in deuteriochloroform (7.26 ppm) and signals
described as s, singlet, d, doublet, t, triplet, q, quartet, and
m, multiplet. The 13C NMR spectral data is reported in parts
per million (ppm) downfield from TMS. The 19F chemical
shifts are reported relative to fluorotrichloromethane (CFCl3).
1
The purity of the final product was assessed by H NMR
using TCNB (tetrachloronitrobenzene) as an internal stan-
dard. The oxidation of the alcohol to the acid was monitored
using a Hewlett-Packard HP5890 gas chromatograph with a
30 m × 0.32 mm i.d. DB-624 column set to ramp from 40
to 90 °C at 2 °C/minute and at 25 °C/minute to 250 °C.
Large-Scale Preparation of 4,4,5,5,5-Pentafluoropen-
tanoic Acid. 4,4,5,5,5-Pentafluoropentanol (1.8 kg, 10.1
mol), tetraethylammonium hydrogen sulphate (18.1 g, 0.08
mole), and water (10.8 L) were added to a 50 L QVF vessel
and heated with stirring to 70 °C. Sodium permanganate
monohydrate (2.33 kg, 14.14 mol) was dissolved at 20 °C
in water (10.8 L) and transferred to a measure vessel.
Aqueous sodium permanganate was added in aliquots of
approximately 10% to the stirred aqueous solution of
pentafluoropentanol and tetraethylammonium hydrogen sul-
phate maintaining a temperature of 65-75 °C by the
additions of permanganate. The total time taken to add the
aqueous permanganate was 2 h 30 min. The reaction was
stirred at 70 °C for a further 4 h when GC analysis showed
conversion of pentafluoropentanol to pentafluoropentanoic
acid to be complete. The reaction mixture was allowed to
cool to ambient temperature overnight and screened through
Celite filter aid (500 g) to remove precipitated manganese
dioxide. The isolated manganese dioxide was washed with
hot water (60 °C, 18 L). The combined aqueous layers were
extracted with methyl tert-butyl ether (5.4 L), and the upper
organic layer was discarded. The aqueous layer was acidified
with concentrated sulphuric acid (320 mL) to pH 1. The
lower organic layer which separated was retained. The
aqueous layer was extracted with further methyl tert-butyl
ether (2 × 5.4 L), and the upper organic layers were
combined with the initial, lower organic layer. The combined
organic layers were washed with water (5.4 L) and dried
with anhydrous sodium sulphate. The organic solvent was
removed in Vacuo at 50 °C and the residue distilled to give
a pale-pink, low-melting solid (1.49 kg, 77%): mp 23-25
°C, bp 84 °C at 60 mmHg; purity by 1H NMR using TCNB
as internal standard was >99%; 1H NMR (CDCl3) δ 2.31-
2.55 (m, 2H, CF3CF2CH2CH2CO2H), δ 2.70 (t, 2H, CF3CF2-
CH2CH2CO2H), δ 10.50 (br s, 1H, OH); 13C NMR (67.7
MHz, CDCl3) δ 25.7 (s, 1C), δ 26.1 (t, 1C), δ 115.5 (q of
by the optimisation of the reaction stoichiometry. Typically,
a level of 0.2 to 0.5% w/w of this impurity was observed.
Reducing the permanganate charge from 1.5 to 1.1 or 1.3
1
equiv generated an additional impurity which, from its H
NMR spectrum, we tentatively assigned as the trans-(Z)-
alkene (3) due to the presence of a large coupling constant
(3JHF) of ∼30 Hz at δ 6.0. This impurity is formed during
the oxidation by elimination of HF from pentafluorobutanoic
acid. This impurity disappears on addition of further per-
manganate (1.4 equiv in total) and we postulate that
decarboxylation and elimination occur, giving 1,1,1-trifluoro-
methylpropyne (4) which is lost to the atmosphere (Scheme
3).
Conclusions
We have successfully developed a scale-up process for
the preparation of 4,4,5,5,5-pentafluoropentanoic acid
(PFPeA) capable of producing kilogram quantities of material
in high yield (77% isolated) and quality (>99% purity).
Modification of the procedure originally reported by
Dmowski et al. has circumvented problems encountered with
formation of unwanted, difficult to remove impurities. Safe
dissipation of heat was achieved by the controlled addition
of oxidant at an elevated temperature, thus giving an
instantaneous reaction with the substrate. We believe this
process to be generally suitable for the oxidation of alcohols
to carboxylic acids on an industrial scale.
Experimental Section
Pentafluoropentanol was supplied by the Asahi Glass
company, Japan.
t, 1C (CF2)), δ 119.0 (t of q, 1C (CF3)), δ 178.0 (s, 1C); 19
F
NMR (376 MHz, CDCl3) δ -86 (s, 3F), δ -119 (t, 2F). IR
The strength of the sodium permanganate monohydrate
used was 97%. The oxidation may be performed using
potassium or sodium permanganate; the preparation cited in
this paper used sodium permanganate due primarily to its
availability within our chemical store. The melting point
quoted for the PFPeA was determined by slowly warming
the product from 10 °C to melting and noting the range with
an alcohol (-70 to 30 °C) thermometer. The determined
melting range was confirmed by repeating the procedure with
two different thermometers. The evaporation of the solvent
was performed by means of a rotary evaporator using house
vacuum (20-30 mmHg). Proton, 13C, and 19F NMR spectra
(neat) 3600-2330, 1720, 1295, 1195 cm-1.
Acknowledgment
We thank Ian Jones for providing spectral interpretation
of the 13C NMR. We also acknowledge the help and
cooperation of Steve Hallam at our Process Hazards group
and Steve Knight who performed the oxidation in our Large
Scale Laboratory.
Received for review March 24, 1999.
OP990021H
364
•
Vol. 3, No. 5, 1999 / Organic Process Research & Development