Electrophilic fluorination: the aminopyridine dilemma

Article abstract of DOI:10.1016/j.tetlet.2009.10.089

An unusually high yielding fluorination of aminopyralid (3) using F-TEDA (SELECTFLUOR) in warm water, followed by kinetic resolution (via iterative esterification/saponification) of the crude fluorination product with dry HCl in methanol produced pure rin

Full text of DOI:10.1016/j.tetlet.2009.10.089

Tetrahedron Letters 51 (2010) 79–81
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Electrophilic fluorination: the aminopyridine dilemma
Stephen C. Fields , William C. Lo , William K. Brewster a,*, Christian T. Lowe a
b
a
a
Discovery Research, Dow AgroSciences, Dow Venture Center, 9330 Zionsville Road, Indianapolis, IN 46268-1053, USA
Lexington Pharmaceuticals, West Lafayette, IN 47906, USA
b
a r t i c l e i n f o
a b s t r a c t
TM
Article history:
An unusually high yielding fluorination of aminopyralid (3) using F-TEDA (SELECTFLUOR ) in warm
water, followed by kinetic resolution (via iterative esterification/saponification) of the crude fluorination
product with dry HCl in methanol produced pure ring-fluorinated pyridine 2 in an overall yield of 31% for
the two steps.
Received 21 November 2008
Revised 12 October 2009
Accepted 19 October 2009
Available online 23 October 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Electrophilic fluorination
Aminopyridine
F-TEDA
1
. Introduction
C-fluorinating agents when it comes to pyridines. Elemental fluo-
2 2
rine (F ) and xenon difluoride (XeF ), the most potent reagents in
The recent discovery of the commercially active ingredient
aminopyralid (3), a potent broad spectrum auxinic picolinate her-
this class, are known to react with little selectivity as both an N-
and a C-fluorinating agent depending on the reaction conditions
1
1
5
bicide, prompted us to revisit the structure–activity relationships
and the electronics of the pyridines. In addition, toxicity and spe-
cial handling requirements associated with F make alternative flu-
2
of this historical class of pyridine herbicides. Through an activity
2
optimization effort, we discovered that replacement of the 6-Cl
on the pyridine ring of 2 with aryl groups produced 1 which
showed unexpectedly high levels of herbicidal activity under a
orinating reagents preferable whenever possible. Use of the other
common route to C-fluorinated pyridines via fluorination of meta-
16
lated aromatic rings with reagents such as N–F pyridinium salts,
3
5
17
18
controlled environment (glasshouse testing). To validate this dis-
NFSi reagents, N–F sultams, N–F sulfonamides , and N–F sulf-
1
9
covery, we decided to test a representative of 1 in the field, which
mandated the synthesis of kilogram quantities of the material. The
electrolysis (4?3) and the arylation (2?1) reactions were known
to proceed in high yield.3 However, initial attempts at the electro-
philic fluorination step (3?2) were extremely poor. Thus, in order
for us to prepare a kilogram field sample of 1 utilizing our com-
modity starting material 4, we required a dependable, scaleable
electrophilic fluorination method to convert 3?2 (see Scheme 1).
Fluorination of electron-rich aromatic ring systems generally
can be accomplished by reaction with electrophilic fluorinating
2
onimides, is precluded by the acidic NH protons present in sub-
2
0
strate 3. Thus, F-TEDA was identified as our reagent of choice for
effecting the transformation of 3?2 as it had been reported as one
of the most powerful electrophilic fluorinating agents for neutral
,4
3 2 2 2
electron-rich aromatic systems (just behind (CF SO ) NF and F it-
2
1
22
self), while still being safe and easy to handle.
We initiated our studies of F-TEDA using acetonitrile as it ap-
2
3
peared in the literature to be a preferred solvent. Under standard
reaction conditions we observed the following problems: (1) Pro-
duction of undesirable picloram (4). Competing chlorination during
F-TEDA reaction had indeed been previously reported though it is
unclear what the chlorinating species is or how it is generated.2
(2) Low mass recovery. The strong oxidizing power of F-TEDA, as
5
agents. However, electron-rich pyridines tend to favor ring N-
4,25
fluorination. While such pyridine N-fluorination has been used to
6
advantage, for instance, to rearrange to 2-C–F pyridines or to facil-
2
6
itate nucleophilic substitution at the 2- and/or 4-position on the
measured by its one-electron reduction potential, could easily
account for the destruction of UV-active (presumed) pyridyl prod-
ucts. (3) Poor conversion. While the heterogeneity of the reaction
could be suspected, competing N-fluorination of the pyridine ring
by F-TEDA along with unproductive complexation of F-TEDA with
7
pyridine ring, for that same reason, it proves detrimental for elec-
8
trophilic ring fluorination. In fact, a survey of some of the widely
used electrophilic fluorinating agents for neutral aromatics, such as
9
10
11
12
13
14
CsOSO
2 3 3 3 3
OF, CF SO F, NOF, NO F, CF OF, and AcOF, re-
2
7
vealed that these materials are better N-fluorinating reagents than
F-TEDA’s desfluoro byproduct
more adequately explain our
incomplete fluorination and recovery of starting material.
In our case, competitive N-fluorination of both the product 2
and the starting material 3 would consume F-TEDA reagent as well
*
Corresponding author. Tel.: +1 317 337 3419; fax: +1 317 337 3215.
E-mail address: wkbrewster@dow.com (W.K. Brewster).
0
040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.10.089

8
0
S. C. Fields et al. / Tetrahedron Letters 51 (2010) 79–81
NH2
N
NH2
N
NH2
NH2
e-
F+
Ar-M
Cl
Cl
Cl
O
Cl
O
F
Cl
O
F
Cl
O
OH
OH
OH
OH
Cl
Cl
N
Ar
N
picloram, 4
aminopyralid, 3
2
1
Scheme 1. Schematic synthesis of field candidate 1.
as render 3 less electron rich and incapable of ring fluorination. The
putative N-fluorinated adducts, which would not survive workup
and/or chromatography, would then create the appearance of
incomplete reaction after reverting back to the parent pyridines
tally precipitated the ester products (enriched in ester 5), enabling
separation from the unreacted picolinic acids (mostly 3 and 4, as
their sodium salts) via suction filtration. Subsequent saponification
of the ester product mixture back to the corresponding acids, for-
tified in 2, and then repeating the esterification/saponification se-
quence one more time afforded 2 in >90% purity, by HPLC. The
development of this procedure to purify 2 was critical for the prep-
aration of the kilogram field sample of 1 (see Scheme 2).
The overall route thus became fluorination of 3 with F-TEDA in
warm water, followed by iterative esterification/saponification of
the crude with dry HCl in methanol, to yield 2 in an overall yield
of 31% for the two steps: an especially efficient outcome for a
pyridine.
28
(
or worse, as our mass balance was only ꢀ50%).
Despite these initial issues, F-TEDA was investigated further to
optimize the yield and minimize the aforementioned problems.
After surveying different solvents,29 catalysts, co-reagents, and
reaction temperatures, we attempted an aqueous reaction to im-
prove the solubility (cf. reaction heterogeneity mentioned in note
3
above) of F-TEDA even though it is reported to be sensitive to
3
0
hydroxylic solvents.
To our delight, the reaction proceeded
smoothly with much shorter reaction time and was eventually
optimized in deionized water at 65 °C, with respect to both yield
and product purity.
Unfortunately, even with excess F-TEDA and extended reaction
times, the fluorination reactions never exceeded 50% yield of 2 (as
monitored by HPLC), with most of the mass balance attributed to
2
2
. Experimental
.1. General
‘
unreacted’ 3. Attempts to drive the reaction further only led to
All solvents and reagents were of reagent grade or higher. All
reactants were purchased from Aldrich unless otherwise noted.
HPLC data were collected on a Beckman System Gold version 1.6
31
higher levels of picloram (4) at the expense of recoverable de-
sired 2 and/or starting material 3. Therefore, the reactions were
stopped when analysis by HPLC typically indicated a composition
of 40% 2, 55% 3, and <4% 4. Workup with 6 N HCl followed by cool-
ing precipitated the pyridine mixture.
Attempts to separate 2 from 3 and 4 via reversed-phase chro-
matography or recrystallization were not successful. Furthermore,
we also demonstrated that it was difficult to remove the 5-H and
(
solvent delivery Module 126, diode array detector Module 168,
Varian’s Microsorb C18 column), eluting via gradient (0–100%
l
l
3
CH CN). Purity was established via H NMR or HPLC. 300 MHz H
1
3
NMR and 75 MHz C NMR data were collected in CDCl
3
with
TMS as internal standard.
5
-Cl by-products via chromatography at subsequent stages. Thus,
2
.2. 4-Amino-3,6-dichloro-5-fIuoropyridine-2-carboxylic acid
a reliable large-scale method to purify 2 at this point was critical.
One approach was to convert the mixture of acids to their methyl
esters, then attempt separation using normal-phase chromatogra-
phy. During this process, we observed that 2 esterified much more
rapidly than either 3 or 4 in dry HCl in methanol. The enhanced
(
2)
A 1 l three neck round-bottomed flask equipped with a mechan-
ical stirrer and a reflux condenser was charged with 3 (22.1 g,
08.3 mmol) and H O (110 mL). While stirring at 25 °C, solid F-
1
2
3
2
reactivity of 2 is possibly due to fluorine’s ability to be a pi-donor
TEDA (Air Products; 42.0 g, 118.6 mmol) was added in portions.
The resulting heterogeneous mixture was warmed to 65 °C and
progress was monitored periodically via HPLC. After 6 h, the reac-
tion was allowed to cool and made acidic with 6 N HCl. Stirring was
continued for 10 min and then the precipitate was collected with
suction filtration and rinsed several times with additional 6 N
HCl. It was allowed to air dry overnight. Analysis of this crude
product via HPLC showed ꢀ 40% 2, 55% 3, and 4.5% 4. The filtrate
when a full positive charge can be stabilized through hyperconju-
gation, and the rates of acid-catalyzed esterification of benzoic
acids have also been shown to increase as electron density in-
creases toward the carboxyl group.3
3
To take advantage of this fortuitous discovery, an iterative ki-
netic purification process was executed. Upon completion of the
3
esterification of crude 2 (CH OH, HCl, 40 °C), the reaction mixture
3
was neutralized by addition of aqueous NaHCO . This coinciden-
2 2
was extracted several times with 10% THF/CH Cl and a small
NH2
NH2
NH2
F
NH2
Cl
O
Cl
Cl
Cl
O
Cl
O
MeOH
HCl (g)
F
Cl
Acid mixture
enriched in
3 and 4
+
+
+
OH
OH
OH
OMe
Cl
N
N
Cl
N
Cl
N
45º C
Major O
3
hr
3
4
2
5
aq. NaOH (quantitative)
Scheme 2. Iterative esterification/saponification purification sequence.

S. C. Fields et al. / Tetrahedron Letters 51 (2010) 79–81
81
amount of crude product (ꢀ1.2 g) was recovered. Scaling this reac-
tion up to 250 g provided 196 g of crude 2, used in the next step
without further purification.
M. F.; Hull, J. W., Jr.; Scortichini, C. L. Selective electrochemical reduction of
halogenated 4-aminopicolinic acids. U. S. Patent 6,352,635, Mar. 5, 2002; Chem.
Abstr. 2001, 135, 98860.
5
6
.
.
Differding, E.; Offner, H. Synlett 1991, 187.
N–F pyridines rearrange to 2-C-F pyridines, however this does not appear to be
general: Umemoto, T.; Tomizawa, G. J. Org. Chem. 1989, 54, 1726.
Direct nucleophilic substitution of N–F pyridines: Kiselyov, A.; Strekowski, L. J.
Org. Chem. 1993, 58, 4476.
2
.3. Purification of 2 via esterification/saponification
7
.
Crude 2 (196 g; 55% 2, 40% 3, and 4% 4) was dissolved in meth-
anol (700 mL) saturated with anhydrous HCl. The solution was
heated to 40–45 °C for 3 h, then cooled to 25 °C, and poured into
8. N-Fluorination of pyridine substrates is a problematic competing reaction
because it renders the pyridine less reactive toward ring fluorination.
9.
Gakh, A. A.; Romaniko, S. V.; Ugrak, B. I.; Fainzilberg, A. A. Tetrahedron 1991, 47,
447.
10. Umemoto, T.; Tomita, K. Tetrahedron Lett. 1986, 27, 3271.
7
an equivolume of saturated aqueous NaHCO
itate was collected by vacuum filtration, washed with water
300 mL), and air dried giving 120 g of 5 in 85% purity by HPLC.
3
solution. The precip-
1
1
1. Boswell, G., Jr. J. Org. Chem. 1968, 33, 3699.
2. Gakh, A. A.; Nikishin, K. G.; Kagramanov, N. D.; Semenov, V. V. Izvestiya
Akademii Nauk SSSR, Seriya Khimicheskaya 1991, 10, 2403.
(
The solid product was dissolved in methanol (500 mL), the pH
brought to >12 with 2 N NaOH (320 mL, 1.2 equiv), and then stirred
at 25 °C for 2 h. The methanol was evaporated in vacuo, and the
residual aqueous solution acidified with 6 N HCl to pH <1. The
resulting precipitate was collected by vacuum filtration, washed
with water (300 mL), and air dried on the filter to give 110 g of 2
in 85% purity by HPLC. The sample of 2 was subjected to the ester-
ification/saponification sequence as described above to provide
13. Middleton, W. J.; Bingham, E. M. J. Am. Chem. Soc. 1980, 102, 4845.
1
1
4. Lerman, O.; Tor, Y.; Hebel, D.; Rozen, S. J. Org. Chem. 1984, 49, 806.
5. (a) Merritt, R. F.; Stevens, T. E. J. Am. Chem. Soc. 1966, 88, 1822; (b) Anand, S. P.;
Filler, R. J. Fluorine Chem. 1976, 7, 179.
16. Umemoto reagents are supplied by Allied Signal Inc., Buffalo Research
Laboratories, Buffalo, NY: Umemoto, T.; Fukami, S.; Tomizawa, G.; Harasawa,
K.; Kawada, K.; Tomita, K. J. Am. Chem. Soc. 1990, 112, 8563.
17. Differding, E.; Lang, R. W. Tetrahedron Lett. 1993, 33, 3971.
1
8. Appears to be effective on Grignards and electron rich phenols, but no specific
mention of heterocycles: Barnette, W. E. J. Am. Chem. Soc. 1984, 106, 452.
9. Des Marteau reagents: (a) Singh, S.; Des Marteau, D.; Zuberi, S.; Witz, M.;
Huang, H. N. J. Am. Chem. Soc. 1987, 109, 7194; (b) Resnati, G.; Des Marteau, D. J.
Org. Chem. 1992, 57, 4281; (c) Des Marteau, D.; Xu, . Z-Q.; Witz, M. J. Org. Chem.
1992, 57, 629.
1
2
9
8 g 2 in 91.5% purity. The overall yield from 3 was 31%. Analytical
1
19
sample: H NMR (DMSO-d
NMR: À137.36 ppm. Anal. Calcd. for C
2.04; H, 1.34; N, 12.50. Found: C, 31.91; H, 1.32; N, 12.33.
6
) 3: 13.83 (br s, 1H); 7.22 (s, 2H);
F
H
6 3
N
2
O
2
Cl F (224): C,
2
0. Also known as SELECTFLUOR, and Banks’ Reagent: Banks, R. E. Preparation of N-
fluorodiazoniabicycloalkanes as fluorinating agents. U. S. Patent 5,086,178, Feb.
3
4, 1992; Chem. Abstr. 1992, 116, 194355. Supplied by Air Products and
Chemicals Inc., 7201 Hamilton Blvd, Allentown, PA 18195.
1. Lal, G. S. J. Org. Chem. 1993, 58, 2791.
2. F-TEDA is a white salt which is safe, stable, and easily handled in air.
3. Banks, R. E. J. Fluorine Chem. 1998, 87, 1.
24. Campbell, T. F.; Stephens, C. E. J. Fluorine Chem. 2006, 127, 1591.
5. F-TEDA also has been used deliberately to effect chlorination and other
electrophilic substitutions: Syvret, R. G.; Butt, K. M.; Nguyen, T. P.; Bulleck, V.
L.; Rieth, R. D. J. Org. Chem. 2002, 67, 4487.
Acknowledgments
2
2
2
The authors would like to thank Drs. Paul Graupner and Scott
Thornburgh for their help with physical chemistry analyses. We
would also like to thank Air Products for supplying F-TEDA and
Dr. Andy Poss and Honeywell Scientists for their work on amino-
2
pyralid with F
2
.
26. Gillicinski, A. J. Fluorine Chem. 1992, 59, 157.
2
2
7. Laali, K. K.; Borodkin, G. I. J. Chem. Soc., Perkin Trans. 2 2002, 953.
8. Honeywell scientists, led by Dr. Andy Poss, attempted a fluorination of 3 with
divalent fluorine in aqueous suspension at 25 °C and observed similar results,
with respect to mass balance and picloram formation (personal
communication).
References and notes
1
.
Common name 4-amino-3,6-dichloropicolinic acid, sold under the tradename
Milestone (among others): Alexander, A. L.; Balko, T. W.; Bjelk, L. A.; Buysse, A.
TM
2
9. There has been a noted solvent dependence with these reagents: Zupan, M.;
Iskra, J.; Stavber, S. J. Fluorine Chem. 1995, 70, 7.
M.; Fields, S. C.; Keese, R.; Krumel, K. L.; Lo, W. C.; Lowe, C. T.; Richburg, J. S.;
Ruiz, J. M. 4-Aminopicolinates and Their Use as Herbicides. U.S. Patent
3
3
0. Zupan, M.; Papez, M.; Stavber, S. J. Fluorine Chem. 1996, 78, 137.
1. We attribute the appearance of small amounts of 4 to electrophilic chlorination
of 3 through a process involving N-fluorination of 3 followed by hydrolysis of
the resulting N-fluoro pyridinium’s 6-chloro position to generate chloride
anion. Reaction of chloride with F-TEDA then likely generates an electrophilic
6
,297,197, Oct. 2, 2001; Chem. Abstr. 2001, 135, 107254.
2
.
.
Johnston, H.; Tomita, M. S. Amino-trichloropicolinic acid compounds. U.S.
Patent 3,285,925, Nov. 15, 1966; Chem. Abstr. 1967, 66, 46338.
Balko, T. W.; Buysse, A. M.; Epp, J. B.; Fields, S. C.; Lowe, C. T.; Keese, R. J.;
Richburg, J. S., III; Ruiz, J. M.; Weimer, M. R.; Green, R. A.; Gast, R. E.; Bryan, K.;
Irvine, N. M.; Lo, W. C.-L.; Brewster, W. K.; Webster, J. D. 6-Aryl-4-
aminopicolinates and Their Use as Herbicides. U. S. Patent 6,784,137, Aug.
3
25
chlorinating species which attacks 3. However, since F-TEDA itself has been
24
suggested as a chloride source, additional work is now underway to provide
more details on the origin of 4. Our results will be forthcoming.
2. Wiberg, K. B.; Rablen, P. R. J. Org. Chem. 1998, 63, 3722.
3. Hartman, R. J.; Borders, A. M. J. Am. Chem. Soc. 1937, 59, 2107.
3
1, 2004; Chem. Abstr. 2003, 138, 153444.
3
3
4
.
(a) Aminopyralid (3) is currently synthesized by the electrochemical reduction
of picloram (4) at Dow on production scale.; (b) Krumel, K. L.; Bott, C. J.; Gullo,