Practical large-scale preparation of (±)-2-exo-norbornyl carboxylic acid and its improved isolation as the sodium salt

Article abstract of DOI:10.1021/op100284a

A practical, robust, and high-yielding three-step-one-pot procedure for the diastereoselective synthesis of (±)-2-exo-norbornyl carboxylic acid starting from norbornylene has been found and demonstrated on multikilogram scale, setting a new benchmark for

Full text of DOI:10.1021/op100284a

TECHNICAL NOTE
pubs.acs.org/OPRD
Practical Large-Scale Preparation of (()-2-exo-Norbornyl Carboxylic
Acid and Its Improved Isolation As the Sodium Salt
Jianxin Gu,† Thomas Storz,*,‡ Frederick Vyverberg,‡ Charles Wu,§ Richard J. Varsolona,‡ and
Karen Sutherland‡
†Pfizer Worldwide R&D, Vaccine Research and Early Development, 401 North Middletown Road, Pearl River, New York 10965,
United States
‡Chemical Research and Development, Eastern Point Road, Groton, Connecticut 06340, United States
§Kilolab Facilities, 401 North Middletown Road, Pearl River, New York 10965, United States
ABSTRACT: A practical, robust, and high-yielding three-step-one-pot procedure for the diastereoselective synthesis of (()-2-
exo-norbornyl carboxylic acid starting from norbornylene has been found and demonstrated on multikilogram scale, setting a new
benchmark for low-pressure hydroformylation of cyclic, bridged olefins. The newly found, nonhygroscopic crystalline sodium salt of
this acid provides a practical isolation point.
’ INTRODUCTION
reaction was reported to proceed in acceptable yields only at
higher temperatures (∼100 °C) and pressures (>1000 psi CO-
pressure), and norbornylene was not investigated as substrate.9
Since then, curiously few literature reports on transition metal-
catalyzed hydroformylations of R-branched olefins have
appeared.10 Especially bridged olefinic systems such as norbor-
nylene have not been studied very well; the reactivity of cyclic
olefins is lower, and the reactions are less selective when
compared to terminal olefins.10 A closer look at published reports
on this deceptively simple transformation reveals that progress
towards a practical low-pressure method for bridged olefins has
been painfully slow.11 In 2005, Huang et al. at Amgen reported an
attractive low-pressure (60 psi) rhodium-catalyzed asymmetric
hydroformylation of norbornylene using the TangPhos ligand
(0.5 mol % Rh-cat. precursor, 0.6 mol % TangPhos).12 The same
report also mentions the oxidation of enantiomerically enriched
aldehyde 1 to the carboxylic acid 2 (using NaClO2/cat. TEMPO
as oxidant) en route to the corresponding carboxamide, albeit
experimental details and characterization data/exo/endo purity
for 2 were not given. Unfortunately, the paper fails to mention
that only one enantiomer (S,S0,R,R0) of the TangPhos ligand has
been commercialized, and the racemate is not sold.13 Hence, the
method is only useful for the synthesis of one enantiomer of the
aldehyde 1 and the carboxylic acid 2.14
Despite its abundant use as building block for a variety of
pharmacologically active molecules (for examples, see refs 1-3),
the available protocols for the synthesis of racemic exo-norbor-
nyl-2-carboxylic acid are still surprisingly unpractical; the tradi-
tional iodolactonization method via the commercially available
endo/exo mixture of norborn-2-ene-5-carboxylic acid according
to Berson and Ben-Efraim4 is cumbersome due to the repeated
bulb-to-bulb distillations necessary to purify the crude unsatu-
rated exo-acid. Furthermore, throughput is poor as a conse-
quence of the low content of exo-isomer (endo/exo 80:20) in the
commercial starting material. exo-Selective nickel-mediated car-
boxylation of norbornylene has been reported but has not proven
practical due to either the generation of stoichiometric nickel
5
waste or the requirement for specialized electrochemical
equipment and selectivity issues.6 The carbonylation of exo/
endo-norbornan-2-ol in liquid sulfur dioxide-antimony(V)
chloride at -70 °C has also been reported to lead selectively
to 2-exo-norbornane-2-carboxylic acid,7 but, presumably due to
the noxious solvent system and expensive starting material, has
not found widespread use. Another direct carboxylation variant
that leads to an amide of norbornane-2-carboxylic acid has been
reported but appears less straightforward.17
On the other hand, hydroformylation of cheap norbornylene
has been claimed to give a 95:5-mixture of exo- and endo-
norbornane-2-carbaldehyde in 72% yield by Brunet and Neibeck-
er in 1989 (after fractional vacuum distillation), albeit via a less
desireable reagent system (in situ generation of air-sensitive,
stoichiometric KHFe(CO)4 from highly toxic iron penta-
carbonyl). The authors also claim the selective oxidation of
exo-norbornane-2-carbaldehyde (1) to the corresponding exo-
acid (2) with KMnO4 in 94% yield, although experimental details
and configurational (exo/endo) purity were not given.8
’ RESULTS AND DISCUSSION
For one of our development projects, we needed a practical
and economic access to large quantities of the racemic exo-acid 2.
We set out to reinvestigate the catalytic hydroformylation of
norbornene, with the goal of finding low pressure (e50 psi)
Special Issue: Sustainable Process Chemistry
Rhodium-catalyzed hydroformylation as a potentially more
convenient method for the synthesis of the intermediate alde-
hyde 1 has been claimed by a German patent from 1973. The
Received: October 22, 2010
Published: February 10, 2011
r
2011 American Chemical Society
942
dx.doi.org/10.1021/op100284a Org. Process Res. Dev. 2011, 15, 942–945
|

Organic Process Research & Development
TECHNICAL NOTE
Table 1.a
entry
ligand
result
1
2
tBu3P BF4
multiple CHO-containing products, starting material remaining
product contaminated with a minor (∼1-2%) aldehyde byproduct
slow reaction
3
Ph3P
3
(()BINAP
DPPF
4
clean product formation
5
XantPhos
biphenyl 2-(ditBu)phosphine
clean product formation
6
no reaction
7
20-(N,N-dimethylamino)biphenyl-2-(dicyclohexyl)Phosphine
multiple CHO-containing products
8
2-biphenyl-dicyclohexylphosphine
multiple CHO-containing products, starting material remaining
clean product formation
9
(oxydi-2,1-phenylene)bis(diphenylphosphine)
20,60-dimethoxybiphenyl-2-dicyclohexylphosphine
20-(N,N-dimethylamino)biphenyl-2-diphenylphosphine
tri-O-tolylphosphine
10
11
12
13
14
multiple CHO-containing products, starting material remaining
multiple CHO-containing products, starting material remaining
no reaction
tri-2-furylphosphine
product contaminated with two minor (∼10%) aldehyde byproducts
product contaminated with two minor (∼10%) aldehyde byproducts
diphenyl-2-pyridinephosphine
a Conditions: toluene, 35 °C, 0.5 mol % Rh-cat., 0.6 mol % ligand, NMR analysis after 22 h.
Scheme 1
conditions with a cheap and readily available ligand, hence
allowing scale-up in common multipurpose equipment.
An initial ligand screen (Table 1) quickly identified DPPF as a
viable ligand for scale-up. A subsequent solvent screen with this
ligand (data not shown) revealed that toluene, isopropyl acetate,
acetonitrile, dichloromethane, and tert-butanol all give good
conversion, with toluene and tert-butanol showing the cleanest
reaction profiles.
Since we were interested in a procedure that would avoid
isolation of the intermediate and somewhat unstable aldehyde,
the catalyst loading screen was conducted in tert-butanol in order
to enable a homogeneous subsequent oxidation step. [A solvent
screen of the NaClO2/TEMPO oxidation of 1 to 2 had earlier
identified t-BuOH as best solvent both in terms of yield and
product quality (data not shown).] Combinations of Rh catalyst
precursor and DPPF ranging from 0.5/0.6% down to 0.15/0.2%
all gave complete conversion by NMR, while a 0.1/0.15%
combination still showed ∼5% starting material after 20 h. Initial
scale-up experiments on multigram scale using a 0.15%/0.2%
catalyst/ligand load were conducted at five different tempera-
tures (r.t. (27 °C), 35 °C, 40 °C, 45 and 50 °C). At room
temperature, the reaction was incomplete after 20 h, whereas
at 50 °C, the reaction was not as clean as in the 35-45 °C range.
At 35 °C, a multigram scale-up experiment cleanly proceeded
to completion in only two volumes of t-BuOH. Direct,
TEMPO-catalyzed bleach oxidation of the intermediate aldehyde
1 [Contrary to the Neibecker procedure (using potassium
tetracarbonylhydridoferrate as reagent8), which reports a
95:5 exo/endo-mixture, we observed exclusively the exo-isomer
(1) of the aldehyde at this point.] in the same solvent affords
the desired acid (()-2 which is conveniently isolated in the
form of its crystalline and nonhygroscopic sodium salt 3 in
excellent yield and configurational (exo) purity15 (Scheme 1).
In terms of safety, TSU testing showed both steps to be
943
dx.doi.org/10.1021/op100284a |Org. Process Res. Dev. 2011, 15, 942–945

Organic Process Research & Development
TECHNICAL NOTE
1
Table 2.
Table 3. H and 13C assignments of 3 in DMSO-d6 þ D2O
compound
retention time (min)
2 (acid)
6.63
4.80
1 (aldehyde)
benign (data not shown). To the best of our knowledge, this
is the first application of a clean and practical hydroformylation
in tert-butanol and is somewhat surprising given the earlier
report of t-BuOH solvent participation by Andrianary and
Jenner.16 A readily available, economic ligand (dppf) is used
at very low loading (0.2 mol % at 0.15 mol % Rh-catalyst
precursor load), thus setting a new benchmark for the prac-
tical low-pressure hydroformylation of bridged olefins (compare
ref 11).
The hitherto unknown, well-behaved, and nonvolatile so-
dium salt 3 is the method of choice to isolate the acid 2 and, for
all practical purposes (e.g., formation of derivatives, functional
group conversions), can be used instead of the parent acid 2,
which is volatile and much more difficult to crystallize.18 To the
best of our knowledge, this represents the most direct and
practical procedure for making and isolating racemic exo-
norbornane-2-carboxylic acid on-scale and should be applicable
to other substituted bicyclic olefins as well.
position
group
13C (ppm)a
1H (ppm)b
1
2
3
4
5
6
7
8
CH
CH
CH2
CH
CH2
CH2
CH2
C
41.13
49.94
35.36
36.10
29.33
29.84
36.65
180.89
2.34
1.92
1.73, 1.23
2.11
1.38, 1.06
1.39, 1.11
1.36, 0.94
-
a Relative to DMSO-d6 referenced as 39.50 ppm. b Relative to TMS
referenced as 0.00 ppm.
this temperature for 30 min, before filtration through a pad of
Celite 545 (10 g)/activated charcoal (10 g)/Celite 545 (10 g).
The pad was rinsed with heptane (150 mL). The combined filtrate
(0.62 L, GC-assay yield of norbornyl acid 2: 210 g (94%)) was
diluted with 2-methyl-THF (1.8 L), heated to 45-50 °C, and
treated with 1 equiv of NaOMe (as determined by assay yield from
step 2) (0.325 kgas25wt% solninMeOH, 45min additiontime).
After cooling to 20-25 °C over 1.5 h, the resulting slurry was aged
for another1.5 h. The suspension wasfiltered (polypropylenefilter
cloth) and the filter cake washed with cold 2-methyl-THF (2 ꢀ
250 mL), then dried under nitrogen stream to constant weight.
Yield was 0.225 kg (87%). Special hazards associated with the
procedure: operate in hood with CO-monitor (normal safety
precautions for CO-reactions apply) Source of starting materials:
all from Aldrich except syngas (Airgas).
Kilolab Scale-Up. Using the protocol outlined above, from
8.0 kg of norbornylene were obtained 11 kg of 3 (80%). Residual
solvents: heptane 0.2%, 2-MeTHF: 0.4%, MeOH: 0.04% (all by
GC); no other impurities by GC þ NMR; water (by KF-Titr.):
0.12%. XRD analysis shows the sodium salt to be crystalline. Mp
218-219 °C decomp. Residual Rh content was below detection
limit (<0.01 ppm).
’ PROCEDURES
The reaction was scaled up in a 1-L stainless steel Biotage
Atlantis reactor. Norbornylene (150 g), t-BuOH (300 mL),
Rh(CO)2(acac) (0.62 g) and DPPF (1.77 g) were charged
under nitrogen. Under mechanical stirring, the reactor was
purged with nitrogen, then with syngas (H2-CO 1:1) (total
volume ∼0.5 L at this point). The reactor was purged with
nitrogen, then purged with syngas, and finally pressurized with
syngas to 45 psi; stirring speed was set to 800 rpm, and the
internal temperature was set to 35 °C. The reaction was com-
plete (>99% conversion) in about 20 h (NMR-assay). Assay
yield and purity >98% (by GC). After cooling to 20-25 °C, this
solution of crude aldehyde 1 was transferred to a 3 L-reactor.
TEMPO (5 g) was charged, and the mixture was cooled to
10 °C. In a separate 2 L-vessel, NaH2PO4 ꢀ H2O (242 g) was
dissolved in water (1 L), to this solution was added NaClO2
(174 g), and the mixture was stirred until a clear solution
formed which was precooled to 5-10 °C. Under cooling, this
oxidant solution was charged slowly to the vigorously stirred
crude aldehyde solution in the 3-L reactor, such that the inter-
nal temperature stayed below 25 °C (caution: exothermic!).
After 30 min, GC showed <0.1% aldehyde and 98% assay yield
of acid 2. Heptane (500 mL) was charged to the reactor. The
mixture was stirred for 5 min, and the layers were separated. To
the cooled (0 °C) organic layer was slowly added a solution of
Na2S2O3 (100 g) in water (400 mL) (Caution: exothermic!),
such that the temperature stayed below 25 °C. The mixture was
stirred for 10 min. The aq phase was discarded and the organic
phase (checked with KI-starch paper for residual oxidant:
negative) stirred with water (300 mL) for 5 min. The aq phase
was discarded, and the organic phase was then concentrated
and to a residual volume of ∼500 mL. Heptane (500 mL) was
charged to the reactor, and the distillation was continued to a
final volume of about 0.5 L (60-65 °C reactor temp, 80 °C
jacket temp, 45 min distillation time, 910 mL total distillate
volume). The solution was cooled to 20-25 °C and stirred at
GC Analytical Method. Column: Agilent 1909/J-413, HP-5,
5% phenyl methyl siloxane, capillary 30.0 m ꢀ 320 μm ꢀ 0.25
μm nominal. Temperature program: 15 °C/min from 60 to
250 °C (see Table 2 for retention times).
1
NMR: the H and 13C resonances were assigned from the
DQCOSY, NOESY, HSQC, and HMBC spectra were acquired
on a Bruker DRX-400 spectrometer. Solvent: DMSO/D2O. The
assignments are shown in Table 3. The exo-stereochemistry of
the carboxyl group is confirmed by the COSY and NOESY
spectra. The endoisomer could not be detected. In the COSY
spectrum H-2 shows a W0-coupling to the H-7 proton at 1.36
ppm. This, W0-coupling is only possible if H-2 is in the endo
position. In the NOESY spectrum H-2 shows an NOE to the H6
proton at 1.11 ppm, further confirming the endo orientation
of H-2.
944
dx.doi.org/10.1021/op100284a |Org. Process Res. Dev. 2011, 15, 942–945

Organic Process Research & Development
TECHNICAL NOTE
’ AUTHOR INFORMATION
CH3SO3H); (f) Lu, S.; Li, X.; Wang, A. Catal. Today 2000, 63, 531
(norbornylene, toluene, 80 °C, 870 psi syngas, 100% yield, carbohy-
drate-derived synthetic ligand); (g) Lappe, P.; Springer, H.; Lukas, R.
DE 10352260 B3, 2005; 0414 (dicyclopentadiene, toluene, 135 °C, 363
psi syngas, 61.3GC-A% monoaldehyde (exo/endo unknown), Rh-sulfo-
nylated PPh3 complex). (h) Cossu, S.; Peluso, P.; Alberico, E.; March-
etti, M. Tetrahedron Lett. 2006, 47, 2569 (norbornadiene system,
toluene, 60 °C, 750 psi syngas, Rh(CO)2acac, 30-40% yield); (i)
Semeril, D.; Matt, D.; Toupet, L. Chem.—Eur. J. 2008, 14, 7144; Adv.
Synth. Catal 2009, 351, 1629(norbornylene, H2O/n-decane, 50 °C, 300
psi syngas, 38-93% conv., Rh-acac sulfonylated calixarene ligand, endo-
aldehyde = main product!).
Corresponding Author
*thomas.storz@pfizer.com
’ ACKNOWLEDGMENT
We thank Dr. Xidong Feng for FT-MS analyses and Dr. Tom
Pagano for NMR structure confirmations. Dr. Gerhard Schling-
mann, Wyeth Pearl River Natural Products Group, is thanked for
providing reference standards of the (2R)- and (2S)-enantiomers
of 2-exo-norbornane carboxylic acid.
(12) Huang, J.; Bunel, E.; Allgeier, A.; Tedrow, J.; Storz, Th.;
Preston, J.; Correll, T.; Manley, D.; Soukup, T.; Jensen, R.; Syed, R.;
Moniz, G.; Larsen, R.; Martinelli, M.; Reider, P. J. Tetrahedron Lett.
2005, 46, 7831.
’ REFERENCES
(13) Personal communication to T.S. by the manufacturer (Chiral
Quest, 7 Deer Park Drive, Suite E, Monmouth Junction, NJ 08852,
United States).
(14) The acid 2 obtained by hydroformylation of norbornylene
using the commercialized S,S0,R,R0-TangPhos ligand was confirmed to
be (2S)-exo-norbornane-2-carboxylic acid (by comparison with the two
enantiomers of this acid prepared by an unambiguous method4), and not
the (2R)-enantiomer as depicted in Scheme 1 of the Amgen
publication.12
(15) Notably, with calixarene-diphosphite ligands, the rhodium-
catalyzed hydroformylation of norbornene leads almost exclusively to
the endo-isomer; see: Monnereau, L.; Sꢀemeril, D.; Matt, D.; Toupet, L.
Adv. Synth. Catal. 2009, 351, 1629.
(1) Henriksson, M.; Homan, E.; Johansson, L.; Vallgarda, J.; Wil-
liams, M.; Bercot, E. A.; Fotsch, C. H.; Li, A.; Cai, G.; Hungate, R. W.;
Yuan, C. C.; Tegley, C.; St. Jean, D. J., Jr.; Han, N.; Huang, Q.; Liu, Q.;
Bartberger, M. D.; Moniz, G. A.; Frizzle, M. J.; Marshall, T. L. WO/
2007/061661 A2, May 31, 2007.
(2) Gobbi, L.; Jaeschke, G.; Luebbers, T.; Roche, O.; Rodriguez
Sarmiento, R. M.; Steward, L. WO/2007/093540 A1 Aug 23, 2007.
(3) Bennani, Y. L.; Chamberlin, S. A.; Chemburkar, S. R.; Chen, J.;
Dart, M. J.; Gupta, A. K.; Wang, L. U.S. Pat. Appl. 2004209858 A1 Oct
21, 2004.
(4) Berson, J. A.; Ben-Efraim, D. A. J. Am. Chem. Soc. 1959, 81, 4083.
(5) Hoberg, H.; Schaefer, D. J. Organomet. Chem. 1982, 236,
C28–C30.
(16) Curiously, it had been observed earlier by Andrianary and
Jenner that in presence of a cobalt-ruthenium co-catalyst system,
tBuOH can participate in hydroformylation reactions via facile forma-
tion of the tertiary carbocation, thereby providing a solvent-derived,
competent reaction partner (isobutylene) in situ, see: Andrianary, P.;
Jenner, G. J. Organomet. Chem. 1987, 322, 89.
(6) Derien, S.; Clinet, J. C.; Dunach, E.; Perichon, J. Tetrahedron
1992, 48, 5235.
(7) Yoshimura, M.; Nojima, M.; Tokura, N. Bull. Chem. Soc. Jpn.
1973, 46, 2164.
(8) Brunet, J. J.; Kindela, F. B.; Neibecker, D. Synth. Commun. 1989,
19, 1923.
(17) Ruthenium-catalyzed hydroamidation of norbornene using
2-formamidopyridine in a formamide/acetonitrile mixture under pres-
sure (135 °C, autoclave, 1.7 mol % Ru3(CO)12) has been reported to
result in a 76% yield (87:13 exo/endo-mixture) of the aminopyridyl
amide of (()-2-norbornane carboxylic acid, see: Ko, S.; Han, H.; Chang,
S. Org. Lett. 2003, 5, 2687.
(18) The X-ray powder diffraction (XRPD) pattern, thermogravi-
metry (TG) curve and heat-cool cycle differential scanning calorimetry
(DSC) curve for the free acid 2 are shown below. The free acid is a low
melting (∼58 °C) crystalline solid that exhibits a ∼10 °C hysteresis of
recrystallization upon cooling (5 °C min-1). The TG curve shows that
the free acid upon melting will evaporate and even shows some minor
sublimation prior to melting (see below). Since the solid is low melting,
the free acid may be difficult to crystallize from a reaction solution. Also,
drying the material may result in losses due to sublimation.
(9) Pino, P.; Botteghi, C.; Consiglio, G.; Salomon, C. DE 2359101
June 06, 1974. Using these patent conditions, in our hands, on small
scale, 1250 psi syngas pressure (CO-H2 1:1) was needed to achieve
100% conversion of norbornene in toluene at 100 °C (data not shown).
(10) For a general review, see the hydroformylation reaction: Ojima,
I.; Tsai, C.-Y.; Tzamarioudaki, M.; Bonafoux, D. In Organic Reactions;
Overman, L. E., Ed.; Wiley: New York, 2000; Vol. 56; pp 1-354. For
catalytic hydroformylations of R-branched olefins (mostly vinyl
compounds), see:(a) Paganelli, S.; Zanchet, M.; Marchetti, M.; Man-
gano, G. J. Mol. Catal., A 2000, 157, 1. (b) Alagona, G.; Ghio, C.;
Lazzaroni, R.; Settambolo, R. Organometallics 2001, 20, 5394. (c) Keck,
G. E.; McLaws, M. D. Tetrahedron Lett. 2005, 46, 4911. (d) Tijani, J.; El
Ali, B. J. Organomet. Chem. 2007, 692, 3492. (e) Smejkal, T.; Breit, B.
Angew. Chem., Int. Ed. 2008, 47, 311. (f) Lightburn, T. E.; Dombrowski,
M. T.; Tan, K. L. J. Am. Chem. Soc. 2008, 130, 9210. (g) Liu, X.; Hu, B.;
Fujimoto, K.; Haruta, M.; Tokunaga, M. Appl. Catal., B 2009, 92, 411.
(11) (a) Van Leeuwen, P. W. N. M.; Roobeek, C. F. J. Mol. Catal.
1985, 31, 345 (dicyclopentadiene, benzene, 270 psi H2/CO 2:1, 95 °C,
100% conversion, exo/endo unknown, 0.2 mol % [Rh(dppe)2]); (b)
Parrinello, G.; Deschenaux, R.; Stille, J. K. J. Org. Chem. 1986, 51, 4189
(norbornylene, benzene, 60 °C, 2625 psi syngas, 100% conv., 87%
“selectivity” (yield not determined), exo only claimed, 0.23 mol %
polystyrene/divinylbenzene polymer-attached [(-)-DBP-DIOP]
PtCl2-SnCl2); (c) Consiglio, G.; Nefkens, S. C. A.; Borer, A. Organo-
metallics 1991, 10, 2046 (norbornylene, m-xylene or benzene, 2325 psi
H2/1050 psi CO, 79% yield, 100% exo, 0.1 mol % [(R,R)-Bco-
dpp]PtCl2/SnCl2); (d) Kollar, L.; Wada, T.; Lautens, M. Tetrahedron:
Asymmetry 1992, 3, 1011 (deltacyclene, toluene, 100 °C, 600 psi syngas,
99% conv., exo/endo = 93:7, PPh3-[Rh(nbd)Cl]2 (1.25 mol %) (with
1.25 mol % S,S-ChiraPhos: 50 °C, 600 psi, 99% conv., exo/endo = 98:2);
(e) Botteghi, C.; Paganelli, S.; Perosa, A.; Lazzaroni, R.; Uccello-
Barretta, G. J. Organomet. Chem. 1993, 447, 153 (norbornylene, toluene,
100 °C, 1500 psi syngas, 98% yield, 100% exo, Pt(C2H4)(dppb) (5mol%) /
945
dx.doi.org/10.1021/op100284a |Org. Process Res. Dev. 2011, 15, 942–945