Esters of pyromellitic acid. Part II. Esters of chiral alcohols: Para pyromellitate diesters as a novel class of resolving agents and use of pyromellitates as duplicands for chiral purification

Article abstract of DOI:10.1021/jo8005446

(Chemical Equation Presented) Methods are presented for the preparation of pyromellitate esters of chiral terpene alcohols, including d-(3) or l-menthol (4), d-isomenthol (7), l-borneol (8), or d- (5) or l-isopinocampheol (6). Alcoholysis of PMDA in CH₂Cl₂/Et₃N led to the formation of monoesters (e.g., 18) or diesters (11, 12), as needed, relying on the differential reactivity of the two anhydride groups. The easily isolated para diester (11) crystallized before the meta diester (12) from HOAc. Nicotine (1, 14) was efficiently resolved as 1:1 salts with the menthyl (11a, 11b) or bornyl (11f) para diesters, prototypes of what promises to be a large class of novel resolving agents. Recrystallization of para-di-d-menthyl pyromellitate (11a) greatly improved the chiral purity of the contained d-menthol (3), an example of purification by "duplication". An alternative synthesis of specific diesters took advantage of the easily separated benzyl diesters and their derived acid chlorides (19, 21), with the benzyl esters serving as temporary blocking groups removable by catalytic hydrogenolysis. Pyromellitate tetraesters (26) were prepared by base-catalyzed transesterification of the tetraethyl ester (25). Tri-l-menthyl pyromellitate (27b) was obtained by catalytic hydrogenolysis of benzyl tri-l-menthyl pyromellitate (31b), itself prepared from the alcoholysis of benzyl pyromellitate triacid chloride (30) with l-menthol (4).

Full text of DOI:10.1021/jo8005446

Esters of Pyromellitic Acid. Part II. Esters of Chiral Alcohols: Para
Pyromellitate Diesters as a Novel Class of Resolving Agents and Use
of Pyromellitates as Duplicands for Chiral Purification
John B. Paine III*
Philip Morris USA, Research Center, P.O. Box 26583, Richmond, Virginia 23261
John.B.Paine@pmusa.com
ReceiVed March 10, 2008
Methods are presented for the preparation of pyromellitate esters of chiral terpene alcohols, including d-
(3) or l-menthol (4), d-isomenthol (7), l-borneol (8), or d- (5) or l-isopinocampheol (6). Alcoholysis of
PMDA in CH2Cl2/Et3N led to the formation of monoesters (e.g., 18) or diesters (11, 12), as needed,
relying on the differential reactivity of the two anhydride groups. The easily isolated para diester (11)
crystallized before the meta diester (12) from HOAc. Nicotine (1, 14) was efficiently resolved as 1:1
salts with the menthyl (11a, 11b) or bornyl (11f) para diesters, prototypes of what promises to be a large
class of novel resolving agents. Recrystallization of para-di-d-menthyl pyromellitate (11a) greatly improved
the chiral purity of the contained d-menthol (3), an example of purification by “duplication”. An alternative
synthesis of specific diesters took advantage of the easily separated benzyl diesters and their derived
acid chlorides (19, 21), with the benzyl esters serving as temporary blocking groups removable by catalytic
hydrogenolysis. Pyromellitate tetraesters (26) were prepared by base-catalyzed transesterification of the
tetraethyl ester (25). Tri-l-menthyl pyromellitate (27b) was obtained by catalytic hydrogenolysis of benzyl
tri-l-menthyl pyromellitate (31b), itself prepared from the alcoholysis of benzyl pyromellitate triacid
chloride (30) with l-menthol (4).
Introduction
It was thought that a resolving agent of higher molecular
weight might lower the solubility of the derived salts to a level
such that the efficiency of a resolution would be significantly
improved. Potential salts were envisioned as being derived from
the class of phthalate monoesters available from the reaction
of a range of inexpensively available terpene alcohols [d- (3)
or l-menthol (4), d- (5) or l-isopinocampheol (6), d-isomenthol
(7), l-borneol (8)] or cholesterol (9) with various commercially
available phthalic anhydrides. Phthalic anhydride itself was
found to give derivatives too soluble and too reluctant to
crystallize with nicotine to be useful. This suggested that the
numerous commercially available halogenated or nitrated ph-
thalic anhydrides should be bypassed in favor of the com-
mercially available symmetrical “double” phthalic anhydride,
pyromellitic dianhydride (PMDA) (10) (Scheme 1).
Our interest in pyromellitate esters arose in the course of a
synthetic program1 to prepare2 analogs of (S)-(-)-nicotine (1).
These were generally prepared in racemic form and then
resolved. Nicotine (1) had been traditionally resolved by
crystallization of its salts of tartaric acid (2, R ) H) or of its
dibenzoyl (2, R ) C6H5CO) or di-p-toluoyl (2, R )
4-CH3C6H4CO) derivatives.3 The low molecular weight and
bifunctionality of nicotine ensure a high solubility for most of
its salts, often making efficient recovery by crystallization
difficult.
(1) Seeman, J. I.; Secor, H. V.; Chavdarian, C. G.; Sanders, E. B.; Bassfield,
R. L.; Whidby, J. F. J. Org. Chem. 1981, 46, 3040–3048.
(2) Hu, M. W.; Bondinell, W. E.; Hoffmann, D. J. Labelled Comp. 1974,
10, 79–88.
(3) (a) Bowman, E. R.; McKennis, H., Jr.; Martin, B. Synth. Commun. 1982,
12, 871–879. (b) Aceto, M. D.; Martin, B. R.; Uwaydah, I. M.; May, E. L.;
Harris, L. S.; Izazola-Conde, C.; Dewey, W. L.; Bradshaw, T. J.; Vincek, W. C.
J. Med. Chem. 1979, 22, 174–177.
Unlike any simple phthalate, any of the three possible specific
pyromellitate diester isomers should possess C2 symmetry, a
feature desirable in any resolving agent for entropic reasons,
10.1021/jo8005446 CCC: $40.75
Published on Web 06/04/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 4939–4948 4939

Paine
SCHEME 1. Synthesis of Chiral Para (11) and Meta
Pyromellitate Diesters (12) from PMDA (10)a
SCHEME 2. Resolution of (S)-(-)-Nicotine (1) and
(R)-(+)-Nicotine (14) as the 1:1 Salts of Para-di-l-menthyl
Pyromellitate (11b): Manual Triage of “Transparent Salt”
(13b) with (S)-(-)-Nicotine and “Opaque Salt” (15b) with
(R)-(+)-Nicotine
a Numbering for 11 and 12: a, ROH ) d-menthol (3); b, ROH )
l-menthol (4); c, ROH ) d-isopinocampheol (5); d, ROH ) l-isopinocam-
pheol (6); e, ROH ) d-isomenthol (7); f, ROH ) l-borneol (8). Reagents
and conditions: (a) 2ROH, 2Et3N, CH2Cl2, rt to reflux, or (a) ROH ) d- or
l-menthol (3 or 4) molten ROH, 140-190°C.
and furthermore would be of the desired molecular weight
(typically at or above 526 for the derivatives from the chiral
alcohols shown above). Finally, the derivatization of pyromellitic
acid with the wide range of these and other available chiral
primary or secondary alcohols would lead to an entire class of
novel resolving agents. A number of these would be readily
available as both antipodes.
To simplify the following discussion and the visualization
of the compounds described, the trivial nomenclature employed
in the preceding paper4 will be continued. 1,2-Diester 4,5-
dicarboxylic acids, 1,5-diester 2,4-dicarboxylic acids (12), and
1,4-diester 2,5-dicarboxylic acids (11) shall be referred to as
“ortho”, “meta”, and “para” pyromellitate diesters, respectively
because such are the respective relative relationships of the
paired equivalent carbonyls in each.
merely physically separating the opaque white crystals from
those that were transparent; a magnifying glass was not required.
The separated diastereomers could be further purified by
recrystallization; however it was found that 13b converged to
absolute purity considerably faster than 15b. A test resolution,
starting with deliberately prepared 90% (S)-(-)-nicotine (1) [and
10% (R)-(+)-nicotine (14)] gave only crystals of 13b. By
isolating the nicotine quantitatively from the mother liquors of
each recrystallization and accounting for the total content of
(R)-(+)-nicotine (14), it was possible to calculate that the
crystals had to be 99.8% pure (S)-(-)-antipode after only two
crystallizations! This purity was confirmed by isolating the
nicotine (1) from the crystals and measuring the optical activity.
The high efficiency of “transparent salt” 13b (or 15a) makes
it especially suitable for completing a resolution of nicotine (1
or 14), once an initial crude splitting has been made with another
resolving agent. The difficultly scalable manual triage of 13b
and 15b can be thus avoided.
It should be noted that the para pyromellitate diesters (11)
are stable and can be recovered efficiently from the salts upon
completion of the resolution. In the case of the above isolation
of nicotine from the crystals, nearly quantitative recovery of
intact resolving agent was achieved 12 years after the salt had
been prepared.
The meta diester (12b) refused to crystallize with (S)-(-)-
nicotine (1) and was not investigated further. Other areas that
were left uninvestigated included possible suitability of pyrom-
ellitate meta or ortho diesters for resolution of substrates other
than nicotine (1) and the possible use of thermal isomerization
to convert the apparently less useful meta diesters (12) to their
para isomers (11), by taking advantage of reversible anhydride
formation in the pyromellitate system.4
Results and Discussion
Pyromellitate Para Diesters of Menthol as Resolving
Agents for Chiral Amines such as Nicotine (1). The initial
experiment entailed heating PMDA (10) with an excess of
molten l-menthol (4) (>2 equiv). The reaction mixture dissolved
nicely in boiling glacial acetic acid, from which, upon slight
cooling and modest aqueous dilution, a white granular powder
soon separated in quantity. This proved to be the pure para
diester (11b). Further aqueous dilution of the filtrates led to the
crystallization of the meta diester (12b), subsequently purified
by fractional crystallization (Scheme 1).
This first “wildcat” experiment (Scheme 1, ROH ) l-menthol,
4) “struck oil”, when 11b was found not only to give (in
methanol, the first-chosen solvent) a 1:1 adduct (13b) with (S)-
(-)-nicotine (1) as enormous transparent crystals of steep
solubility coefficient but also to be capable of producing this
salt from racemic nicotine.5
The resolution (Scheme 2) was clouded by the fact that (R)-
(+)-nicotine (14) soon also gave a crystalline 1:1 adduct (15b)
with 11b and that once seeds of 15b had formed, it was not
possible to prevent cocrystallization of both diastereomers (when
initially present in equimolar quantities). However, 15b had
crystallized from methanol with solvation that was easily lost,
whereas the less extensive solvation of 13b was tightly bound.
Upon drying in air, 15b quickly desolvated and become opaque
white, while 13b remained transparent (subsequently going
opaque only upon prolonged storage). Resolution involved
d-Menthol (3) reacted in analogous fashion. Its para diester
(11a) crystallized preferentially with (R)-(+)-nicotine (14). In
practice, the system was somewhat less well behaved. The
(4) Paine, J. B. J. Org. Chem. 2008, 73, 4929–4938.
(5) Tsujino, Y.; Shibata, S.; Katsuyama, A.; Kisaki, T.; Kaneko, H.
Heterocycles 1982, 19, 2151–2154.
4940 J. Org. Chem. Vol. 73, No. 13, 2008

Esters of Pyromellitic Acid. Part II
FIGURE 1. Structures of (S)-(-)-nicotine (1), tartrate resolving agents (2), and terpene alcohols (3-9) used in chiral pyromellitate ester synthesis.
SCHEME 3. Synthesis of Mono-l-menthyl Pyromellitate
synthetic d-menthol (3) employed in the preparation contained
about 2% of its antipode. By contrast the l-menthol (4) used
previously was the natural (Brazilian, from Mentha arVensis)
material and thus known to be optically pure.6 The resulting
11a, NMR-pure with respect to any meta isomer, had been used
directly. As will be seen below, it takes several additional
recrystallizations to render 11a chirally pure.
(18b) from PMDA (10)a
Synthesis of Pyromellitate Diesters of Other Terpene
Alcohols. The other terpene alcohols of Figure 1 (5-9) were
also reacted with PMDA (10). Since these substrates were
potentially unstable at high temperature (the isopinocampheols,
5 and 6) or excessively high melting (l-borneol, 8), the milder
conditions (stoichiometric alcohols and triethylamine in CH2Cl2)
described previously4 were used to run the reaction. In practice,
it became usual to add the amine gradually to a well-stirred
suspension of PMDA in a solution of the chosen alcohol in
dichloromethane. Because the second anhydride ring to react
was found to be much less reactive than the first,4,7 completion
of diesterification could not be assumed to have occurred
immediately after the system had become homogeneous or
spontaneous boiling had ceased but required further time.
Workup entailed acidification with aqueous HCl to remove Et3N
and then crystallization from acetic acid. In every instance, the
para diester (11c-f) crystallized first and was easily isolated;
the meta diesters (12c-f) varied somewhat in their ease of
isolation. The two isomers were generally each isolated in yields
ranging from 20% to 40%. The experimental details for much
of this is presented in Supporting Information. The above mild
conditions were also applied successfully to d- (3) and l-menthol
(4).
a Reagents and conditions: (a) ROH ) l-menthol, Et3N, CH2Cl2, rt to
reflux; (b) HCl, H2O; (c) HOAc, Ac2O; (d) H2O, HOAc, brief reflux.
acid analogs of the above diesters, using chiral disubstituted
amines, although this should be a promising area for future
study. It should be further noted that the differential activity4,7
of the two anhydride rings of PMDA (10) would allow two
different chiral alcohols (or amines or one of each) to be
appended in sequential fashion. However, such derivatives
would lack the C2 symmetry that is generally considered
advantageous (since all favorable states and interactions a 2-fold
symmetric resolving agent might have with its substrate are
doubled in probability).
The use of 1 equiV of l-menthol (4) and triethylamine with
PMDA suspended in CH2Cl2 also led to the formation of a
homogeneous solution under these conditions and afforded good
yields of the pyromellitate monoester (18b) after workup
(Scheme 3). An important feature of the intermediate pyrom-
ellitate monoester anhydrides such as 17b is their propensity to
crystallize as 1:1 adducts with acetic acid. It is therefore
advantageous to include acetic anhydride and acetic acid in the
workup procedure.
Brief attempts to generate pyromellitate diesters of higher
molecular weight from sesquiterpene tertiary chiral alcohols
such as cedrol or patchouli alcohol were not successful. No
attempt was made to prepare pyromellitate diamide dicarboxylic
FIGURE 2. Para-di-l-bornyl pyromellitate (11f) and its salt (13f) with
(S)-(-)-nicotine.
Other Para Pyromellitate Diesters in Nicotine Resolution.
Although all of the para diesters (11) examined gave crystalline
derivatives with (S)-(-)-nicotine (1), only the l-bornyl derivative
(11f, Figure 2) also proved capable of resolving (S)-(-)-nicotine
(1) from the racemic form.5 The others deposited different
phases from (R,S)-nicotine (1 + 14) than from (S)-(-)- nicotine
(1) alone.
In the l-bornyl case, a single solid phase of low solubility
formed (13f), which was greatly enriched in the (S)-(-)-antipode
of nicotine, but this did not converge to absolute purity upon
recrystallization nearly as well as 13b. However, 11f could be
(6) (a) Ravid, U. Perfum. FlaVor. 1998, 23, 25–28, 30. (b) Werkhoff, P.;
Hopp, R. Progress in Essential Oil Research; Brunke, E. J., Ed.; Walter de
Gruyter: New York, NY, 1986; pp 529-549. (c) Coleman, W. M., III; Perfetti,
T. A.; Suber, R. L., Jr J. Chromatogr. Sci. 1998, 36, 318–321.
(7) Kreuz, J. A.; Angelo, R. J.; Barth, W. E. J. Polym. Sci., Part A-1 1967,
5, 2961–2963.
J. Org. Chem. Vol. 73, No. 13, 2008 4941

Paine
SCHEME 4. Chiral Purification of d-Menthol as the Para
Pyromellitate Diester
SCHEME 6. Synthesis of Meta Dibenzyl Di-l-menthyl
Pyromellitate (22)a
a Reagents and conditions. (a) 2ROH ) l-menthol, 2Et3N, CH2Cl2, rt.
SCHEME 5. Benzyl Ester Assisted Synthesis of
Para-di-l-menthyl Pyromellitate (11b)a
SCHEME 7. Esterification of Para Dimenthyl
Pyromellitates (11a, 11b) with Orthoformate Estersa
a Reagents and conditions: (a) 2ROH ) l-menthol, 2Et3N, CH2Cl2, rt;
(b) H2, Pd-C, THF, 1 atm, rt.
a Numbering for 11, 23 (R ) methyl), and 24 (R ) ethyl): a, d-menthyl
series; b, l-menthyl series. Reagents and conditions: (a) HC(OR)3, reflux.
employed as the resolving agent to provide the initial crude
splitting of the antipodes 1 and 14, allowing 11a,b to complete
the resolution of 1 and 14 without need for manual triage. Under
these conditions, the maximally efficient “transparent salt” (13b
or 15a) was the only phase formed.
and PMDA (10) and not the potentially valuable or precious
chiral alcohols.
Para dibenzyl pyromellitate diacid chloride (19)4 reacted
smoothly but slowly with l-menthol (4) in CH2Cl2/Et3N at room
temperature, and the resulting para dibenzyl di-l-menthyl
pyromellitate (20b) was isolated in pure form by crystallization.
The hydrogenolysis of 20b over Pd-C in THF led quantitatively
to 11b, demonstrating the amenability of the chemistry and the
viability of the approach (Scheme 5). This approach avoids the
tedium of recycling unwanted byproducts, either by saponifica-
tion or thermal isomerization. The meta dibenzyl di-l-menthyl
pyromellitate (22) could be prepared similarly from 21
(Scheme 6).
Unfortunately, when extended to some of the analogs of
nicotine, these resolving agents failed in our hands, either giving
seed crystals only with difficulty or else unresolvable solids.
Pyromellitate Para Diesters as Duplicands for Purifica-
tion of the Contained Chiral Alcohols. Interest in these
resolving agents was rekindled when a need for highly
purified menthol diastereomers arose. It was soon found that
d-menthol (3), typically offered commercially in 96-98%
chiral purity, could be quickly upgraded to 99.95% chiral
purity by only two additional recrystallizations of the para
pyromellitate diester 11a from methanol. The contained
d-menthol (3) could be easily recovered nearly quantitatively
by saponification in dilute aqueous alkali, followed by a
modified “steam” distillation (Scheme 4). This is an example
of a purification by “duplication”.8
Benzyl Ester Route to Chiral Para Pyromellitate Diesters.
The yields of the generally preferred para diesters 11 that have
been isolated from direct reactions with PMDA (10) have only
generally ranged from 20% to 40% (when crystallization was
used). Therefore, the direct reaction of terpene alcohols with
PMDA (10) can hardly be considered sufficiently efficient for
use in duplicative purification of such alcohols. To improve the
yields of 11 from terpene alcohols, the use of benzyl esters was
investigated (Scheme 5). The sacrifice to the unwanted meta
diesters was therefore intended to be made by benzyl alcohol
Other Pyromellitate Mixed Tetraesters as Duplicands
for Menthol. Prepared from 11a or 11b and trimethyl or triethyl
orthoformate4,9 (Scheme 7) were specific para dimethyl (23) or
diethyl dimenthyl pyromellitates (24). Of these, 24a,b proved
less efficient at resolving the contained menthol than the para
diester precursors (11a,b). This may be a consequence of the
effects of hydrogen-bonding upon the crystal structure, an option
only available to 11a,b and not its derived tetraesters.
Pyromellitate Tetraesters as Potential Quadruplicands
(or Bis-duplicands) for the Purification of the Contained
Chiral Alcohols. Tetraethyl pyromellitate (25)4 was found to
transesterify smoothly with terpene alcohols in boiling xylene,
to which small portions of NaH catalyst were periodically added
(Scheme 8).
The resulting terpene tetraesters (26), typically of MW around
800, had steep solubility coefficients in solvents such as
(8) Vigneron, J. P.; Dhaenens, M.; Horeau, A. Tetrahedron 1973, 29, 1055–
1059.
(9) Cohen, H.; Meier, J. D. Chem. Ind.(London) 1965, 349–350.
4942 J. Org. Chem. Vol. 73, No. 13, 2008

Esters of Pyromellitic Acid. Part II
SCHEME 8. Synthesis of Chiral Pyromellitate Tetraesters
(26) by Transesterificationa
SCHEME 9. Synthesis of Tri-l-menthyl Pyromellitate (27b),
Precursor to the Ortho Diester Substitution Patterna
a Numbering for 26: a, ROH ) d-menthol; b, ROH ) l-menthol; c, ROH
) d-isopinocampheol; d, ROH ) l-isopinocampheol; e, ROH ) d-
isomenthol; f, ROH ) l-borneol. Reagents and conditions: (a) ROH, catalytic
NaH, xylene, reflux.
n-butanol. Their resolution efficiency was low, no doubt due to
extensive possibilities for solid solution. However, this was
offset in part by their direct availability from precursors in high
yield. These represent examples of “quadruplication” (or more
strictly “bis-duplication” since the symmetry is not higher than
D2). [Not investigated was the possibility that certain members
of this class might prove amenable to purification by zone
refining.]
The above base-catalyzed transesterification is an alternative
approach to this class of compounds, which lately have been
of interest as sensitizers for chiral photochemistry.10–16 The
Inoue group prepared these materials by the alcoholysis of
pyromellitoyl tetrachloride,16,17 reporting yields of 73-74% after
recrystallization from ethanol.
a Reagents and conditions: (a) (-COCl)2, catalytic HCONMe2, CH2Cl2,
rt to reflux; (b) 3l-menthol, 3Et3N, CH2Cl2, rt; (c) H2, Pd-C, THF, 1 atm,
rt.
TABLE 1. Manual Resolution of (R,S)-Nicotine (1 + 14) with
Para-di-l-menthyl Pyromellitate (11b) on a 0.05 mol Scale
13b, “transparent salt” 15b, “opaque salt”
round 1, yield (g)
round 2, yield (g)
total yield (g)
recrystallized yield
(dried in air) (g)
recovery (%)
3.92
1.11
The above work led conceptually into the significantly
improved procedures for the “duplicative” purification of terpene
alcohols, using symmetrical aromatic diesters, which we shall
report in the future.18
11.27
15.19
10.66
14.60
15.71
10.89
70.1
69.3
Tri-l-menthyl Pyromellitate: Use of Benzyl Esters To
Avoid a Difficult Saponification. Tetra-l-menthyl pyromellitate
(26b) was envisioned as a precursor to the triester (27b) and
the ortho diester anhydride (28b) upon saponification and
pyrolysis, respectively (Scheme 9). However, 26b was found
to be too insoluble in solvents which would also dissolve NaOH
(poly(ethylene glycol) ethers were not investigated), vulnerable
to partial transesterification in alcohols in the presence of base,
and generally slow to react, no doubt due to steric congestion
near the carbonyl groups. Instead, a benzyl ester approach was
employed to prepare the triester (27b).
overall yield (%)
61.5
-166.15
62.9
+161.8
20
[R] (neat) of
589
contained nicotine
purity of contained (%)
nicotine
98.7
(S)-(-)
97.4
(R)-(+)-
methylene protons nonequivalent and produced a magnificent
AB quartet as a result; 20 and 22 behaved similarly.
Experimental Section
Thus, monobenzyl pyromellitate (29)4 was converted to
benzyl pyromellitate triacid chloride (30) using oxalyl chloride
in dichloromethane with catalytic amounts of N,N-dimethyl
formamide. In crude form, 30 was reacted with excess l-menthol
(4) and triethylamine in dichloromethane. The resulting benzyl
tri-l-menthyl pyromellitate (31b) was readily purified by chro-
matography and subsequent crystallization. Upon catalytic
hydrogenolysis over Pd-C in tetrahydrofuran, 31b afforded 27b
smoothly and quantitatively. The 1H NMR spectrum of 31b was
noteworthy: the asymmetry of the molecule rendered the benzyl
General Experimental Methods. General methods are pre-
sented in detail in Supporting Information. Chiral purities of
products are expressed as absolute purities or (occasionally) as
(apparent) optical purities, rather than as the highly artificial
concept of an “enantiomeric excess”.19 The concept of “ee” based
on the determination of optical rotations is not strictly applicable
to systems involving duplication (or the higher orders of
“multiplication”), since diastereomers of the principal homo-
chiral derivative may be present as impurities, as well as the
antipode. (However, the optical purity of a duplicated derivative
does correlate to the underlying monomer composition.)
As a routine safety measure, all manipulations and reactions
were performed (in the usual glassware) within the confines of
large-size polypropylene trays to contain potential spillage. Hot
erlenmeyer flasks remoVed from hot-plates for pouring of contents
were handled with gloVes and supported by clamps at the neck,
and from below by enameled steel bowls or stainless steel trays to
contain potential breakage or spillage of the hot contents.
(10) Inoue, Y.; Tsuneishi, H.; Hakushi, T.; Tai, A. J. Am. Chem. Soc. 1997,
119, 472–478.
(11) Tsuneishi, H.; Hakushi, T.; Tai, A.; Inoue, Y. J. Chem. Soc., Perkin
Trans. 2 1995, 2057–2062.
(12) Inoue, Y.; Yamasaki, N.; Yokoyama, T.; Tai, A. J. Org. Chem. 1993,
58, 1011–1018.
(13) Inoue, Y.; Yamasaki, N.; Shimoyama, H.; Tai, A. J. Org. Chem. 1993,
58, 1785–1793.
(14) Inoue, Y.; Yamasaki, N.; Yokoyama, T.; Tai, A. J. Org. Chem. 1992,
57, 1332–1345.
(15) Inoue, Y.; Shimoyama, H.; Yamasaki, N.; Tai, A. Chem. Lett. 1991,
593–596.
(19) The author adheres strongly to the view that “ee” is best used to describe
the outcome of chemical reactions, particularly those reactions wherein the overall
chiral composition is affected by the reaction mechanism and the outcome is
being compared to that from a random process. For the description of actual
samples or preparations, from which, in principle, all of a particular antipode is
available to be isolated or used, the concept of “absolute purities” or “chiral
purities” is far more useful.
(16) Yamasaki, N.; Inoue, Y.; Yokoyama, T.; Tai, A.; Ishida, A.; Takamuku,
S. J. Am. Chem. Soc. 1991, 113, 1933–1941.
(17) Inoue, Y.; Yamasaki, N.; Yokoyama, T.; Tai, A.; Ishida, A.; Takamuku,
S. J. Chem. Soc., Chem. Commun. 1989, 1270–1271.
(18) Paine, J. B., III; Pierotti, J. A. Manuscript in preparation.
J. Org. Chem. Vol. 73, No. 13, 2008 4943

Paine
TABLE 2. Efficiency of Resolution of Nicotine (1) with 11b, Starting with (90% S/10% R)-Nicotine, 0.1 mol Scale
round 1
round 2
starting materials
0.1 mol: 16.24 g (90% S) nicotine; 1.624 g of original
(R)-nicotine (14) content
50.56 g of 13b (93.4% of the round 1 yield)
yield of “transparent salt”, 13b
nicotine isolated from mother
54.11 g (78% absolute or 86.7% relative to the
(S)-nicotine (1) content of the system)
3.85
45.57 (90.1% recovery)
1.14 (theory ) 1.15 g)
liquors (g)
20
20
[R] (neat) of nicotine from
-33
-153.8; [R] -185.0 (neat)
589
546
mother liquors
content of 14 (%)
content of 14 remaining
in 13b (g)
40.26 (1.55 g)
0.074 (of which 0.069 g was carried into round 2)
4.8 (0.0547 g)
0.0144
TABLE 3. Resolution of Nicotine and Purification of (R)-(+)-Nicotine (14) using Para-di-l-bornyl Pyromellitate (11f) for Initial Partition and
Para-di-d-menthyl Pyromellitate (11a) for Further Purification
(R)-nicotine (14)
purification
from mother liquors of
rounds 1A and 1B
initial splitting
round 1A
round 1B
8.13 (50.11 mmol)
26.36 (50.06 mmol)
40.51
20
(R,S)-nicotine taken (g) 3.25 (20.03 mmol)
11f taken (g)
methanol used (g)
(R) (neat)
+96.4 (neat)
78.25
589
10.55 (20.03 mmol)
24.58
% of 14
weight of nicotines 5.51 (33.96 mmol)
taken (g)
yield of 13f (g)
7.00 + 0.92 second crop
14.42 + 5.34 second crop
11a taken (g)
18.02 (33.96 mmol)
recrystallization of 13f
19.71 g of 13f with 83.62 g methanol;
methanol used (g) 39.17 g
yield 14.62 g (74.1% recovery of solids)
20
[R] (neat) of nicotine -98.4 (neat), or 78.8% (S)-(-)- -153.4 (neat), 95.0% (S)-(-)-nicotine (1) yield of 15a (g)
14.95 (63.5% absolute yield
or 81% based on system
content of 14)
589
in 13f
nicotine (1) (from first crop
of round 1A)
(in the combined recrystallized 13f)
purity of contained 98.26
14 (%)
2,5-Bis-[(1R,2S,5R)-2-isopropyl-5-methylcyclohexyloxycarbo-
nyl]terephthalic Acid (Para-di-l-menthyl Pyromellitate) (11b).
Method A. 1,2,4,5-Benzenetetracarboxylic dianhydride (PMDA)
(10) (227.67 g, 1.0438 mol) was added in portions to fused
(1R,2S,5R)-(-)-menthol (4) (471.6 g, 3.018 mol) at 140 °C
(Erlenmeyer flask, hotplate-stirrer) with stirring until the mixture
congealed. The internal temperature was raised to 180 °C to
refluidize the mixture, which was stirred with a spatula, as needed.
The hot mixture was then poured cautiously into boiling glacial
acetic acid (1000 mL). Acetic acid (500 mL) was used to rinse the
melt from the flask. When the solution had cooled to 115 °C, H2O
(200 mL) was added, with magnetic stirring, followed by seed
crystals of 11b. Stirring was continued until the temperature of the
slurry fell to 90 °C, whereupon the solids were filtered off (coarse
frit). The solids were rinsed with minimal glacial acetic acid,
followed by 75% and 50% aqueous acetic acid, until the rinsings
were colorless, and then H2O (300 mL). The product was dried in
air. Yield: 157.71 g (0.297 mol, 28.5%). Later crops of 13.55 and
152.70 g, the latter after further aqueous dilution, were obtained
of mixed isomers. Mp 240.0-246.5 °C (dec, menthol odor was
Isolated by fractional crystallization from the above experiment,
20
Method A. Mp 217.0-218.0 °C; optical rotation [R] -97.19,
589
20
20
1
[R] -102.52, [R] -118.14, (c 10.545, tetrahydrofuran); H
578
546
NMR (CD3OD) δ 0.84 (6 H, d, J ) 6.93 Hz), 0.93 (6 H, d, J )
7.07 Hz), 0.97 (6 H, d, J ) 6.65 Hz), [2 H, m: obscured by methyl
doublet], 1.15 (4 H, br quintet), 1.52 (4 H, br t), 1.75 (4 H, br d),
1.96 (2 H, m), 2.19 (2 H, br d), 3.31 (2 H, quintet), peak near 4.95
(2H, m) obscured by solvent protons, 7.79 (H, s), 8.14 (H, s) (2
COOH exchanged with solvent); 13C NMR (CD3OD at 49.0) δ
168.7, 167.6, 136.5, 135.6, 130.9 (1C), 129.3 (1C), 77.7, 48.4, 41.4,
35.4, 32.7, 27.5, 24.6, 22.5, 21.2, 16.7 (2C each unless otherwise
noted). Anal. Calcd for C30H42O8: C, 67.90; H, 7.98. Calcd for
2C30H42O8.H2O: C, 66.77; H, 8.03. Found: C, 66.57; H, 8.14.
2,5-Bis[(1S,2R,5S)-2-isopropyl-5-methylcyclohexyloxycarbo-
nyl]terephthalic Acid (Para-di-d-menthyl Pyromellitate) (11a).
Method A. From PMDA (10) (222.11 g, 1.018 mol) and (1S,2R,5S)-
(+)-menthol (3) (475.5 g, 3.043 mol), as above. Only 1330 mL in
all of acetic acid was used to dissolve the melt. The 200 mL of
water added at 115 °C lowered the temperature of the mixture to
97 °C and the product was filtered off at 75 °C. The solids were
rinsed with acetic acid (30 mL), then 50% (v/v) aqueous HOAc
(200 mL), and finally H2O (200 mL). Yield: 175.01 g (0.330 mol,
32.4%). The filtrates on standing deposited further solids (285.0 g,
0.537 mol, 52.7%), consisting largely of the meta isomer 12a. Mp
(twice recrystallized from methanol): 245.0-249.5 °C (dec); optical
noted; the mp is not very reproducible.); mp (twice crystallized
20
from methanol) 245.0-248.5 °C dec; optical rotation [R]
589
20
20
-123.75, [R] -130.65, [R] -150.99 (c 10.415, tetrahydrofu-
578
546
ran); 1H NMR (CD3OD) δ 0.83 (6 H, d, J ) 6.92 Hz), 0.92 (6 H,
d, J ) 6.98 Hz), 0.96 (6 H, d, J ) 6.57 Hz), [2H, m obscured by
methyl doublet], 1.14 (4 H, m), 1.51 (4 H, br t), 1.74 (4 H, br d),
1.97 (2 H, m), 2.17 (2 H, br d), [3.31 (2 H, quintet) solvent,
CD2HOD], peak around 4.95 (2H, m) obscured by solvent, water
protons, 7.96 (2 H, s) (2 COOH exchanged with solvent); 13C NMR
(CD3OD at 49.0) δ 168.7, 167.2, 136.3, 135.8, 130.1, 77.7, 48.4*,
41.4, 35.4, 32.8, 27.4, 24.5, 22.5, 21.1, 16.7 (all peaks 2 C) (*solvent
interference). Anal. Calcd for C30H42O8: C, 67.90; H, 7.98. Found
(first crop): C, 68.24; H, 8.01. Found (after recrystallization from
methanol): C, 67.89; H, 8.22. Method B. From the catalytic
hydrogenolysis of the dibenzyl ester (20b) in essentially quantitative
yield. See below.
rotation (respectively: original first crop, once recrystallized from
20
methanol, and twice recrystallized from methanol): [R] +120.86,
589
20
20
+122.82, +123.88; [R]
+127.60, +129.66, +130.78; [R]
1
578
546
+147.48, +148.82, +151.16. (c 10, tetrahydrofuran); H NMR
(CD3OD) δ 0.83 (6 H, d, J ) 6.89 Hz), 0.92 (6 H, d, J ) 7.08
Hz), 0.97 (6 H, d, J ) 6.55 Hz), [2 H, m: obscured by methyl
doublet], 1.16 (4 H, m), 1.51 (4 H, br t), 1.74 (4 H, br d), 1.97 (2
H, m), 2.17 (2 H, m), [3.32 (2 H, m, CD2HOD in solvent)], peak
near 4.95 (2H, m) obscured by water/solvent protons, 7.97 (2 H, s)
(2 COOH exchanged with solvent); 13C NMR (CD3OD at 49.4) δ
168.7, 167.6, 136.3, 135.8, 130.1, 77.7, 48.4, 41.4, 35.3, 32.7, 27.4,
24.5, 22.4, 21.1, 16.7 (2C each peak); Anal. Calcd for C30H42O8:
C, 67.90; H, 7.98. Found (recrystallized from ethanol): C, 67.80;
4,6-Bis[(1R,2S,5R)-2-isopropyl-5-methylcyclohexyloxycarbo-
nyl]isophthalic Acid (Meta-di-l-menthyl Pyromellitate) (12b).
4944 J. Org. Chem. Vol. 73, No. 13, 2008

Esters of Pyromellitic Acid. Part II
H, 8.11. Method B. PMDA (10) (22.47 g, 0.103 mol) and
(1S,2R,5S)-(+)-menthol (3) (33.22 g, 0.2126 mol) were suspended
in CH2Cl2 (250 mL) under moisture-exclusion conditions (magnetic
stirrer), and treated, over about a minute, with triethylamine (43.5
mL, 0.312 mol). The normally insoluble anhydride dissolved
quickly, and the solution boiled briefly. After standing overnight,
the solution was filtered, and shaken with HCl (50 mL conc.) -
H2O (200 mL). The organic phase was isolated, dried over Na2SO4,
and filtered. Acetic acid (200 mL) was added to the filtrates, which
were concentrated to a slurry in vacuo (rotary evaporator). The
slurry was heated and the volume adjusted to 275 mL with acetic
acid. H2O (32 mL) was added to the resulting hot solution. Solids
appeared on cooling to 102 °C. These were filtered off at 95 °C,
and washed with 50% aqueous acetic acid and then H2O. Yield:
26.94 g (0.0508 mol, 49.3%). The filtrates deposited the meta isomer
12a (13.26 g, 0.0250 mol, 24.3%).
2,5-Bis{endo-(1S,2R,4S)-1,7,7-trimethylbicyclo[2.2.1]hept-2-
yloxycarbonyl}terephthalic Acid (Para-di-l-bornyl Pyromelli-
tate) (11f). endo-(1S)-(-)-Borneol (8) (240.68 g, 1.56 mol, Aldrich
98% grade) was dissolved in CH2Cl2 (1000 mL) and filtered to
remove minor sediment. PMDA (10) (170.2 g, 0.78 mol) was added
to the filtrates followed by redistilled Et3N (250 mL, 1.79 mol),
added cautiously to the stirred suspension in a 4 L erlenmeyer flask
over about 1 min. The PMDA dissolved promptly and the mixture
boiled vigorously. After standing overnight at room temperature,
the mixture was shaken with concentrated HCl (182 mL)-H2O (500
mL). The organic phase was isolated promptly before product could
crystallize. Acetic acid (777 mL) was added to the filtrate. After
concentration on a rotary evaporator, the resulting slurry was
filtered. The solids were washed with HOAc, 50% HOAc, and H2O.
The yield, after drying in air, was 114.86 g (0.218 mol, 28.0%). A
mL). The organic phase was isolated promptly, diluted with acetic
anhydride (50 mL), and then concentrated on the rotary evaporator.
The resulting dark greenish brown syrup was diluted with glacial
acetic acid (100 mL); the product crystallized at once. The resulting
slurry was filtered. The solids were rinsed with glacial acetic acid
(100 mL) and then n-hexane (85 mL). Yield: 99.41 g (0.229 mol,
76.3%). Mp 172.5-175.5 °C (dec), after slow heating to drive off
HOAc. Two later crops came to 16.30 g (0.0375 mol, 12.5%) for
a total of 115.71 g (0.266 mol, 88.76%). 1H NMR (CDCl3) δ 0.83
(3H, d, J ) 6.96 Hz), 0.92 (3H, d, J ) 6.96 Hz), 0.96 (3H, d, J )
6.46 Hz), [1 H obscured by methyl doublet], 1.13 (2H, q, J ) 11.7
Hz), 1.51 (2H, t, J ) 11.7 Hz), 1.75 (2H, d, J ) 11.79 Hz), 1.93
(H, m), 2.15 (3H, s, CH3CO2H), 2.20 (H, d, J ) 12.56 Hz), 5.02
(H, t of d, J ) 10.9 and 4.2 Hz), 8.24 (H, s), 8.45 (H, s), 12.01
(2H, s, CO2H) (1 H obscured by 0.96 methyl doublet); 13C NMR
(CDCl3 at 77.1) δ 178.4, 170.2, 164.9, 161.0, 160.9, 141.1, 137.6,
133.7, 132.7, 126.8, 125.9, 77.7, 47.0, 40.2, 34.1, 31.5, 26.2, 23.3,
22.0, 20.8, 20.7, 16.1. Anal. Calcd for C20H22O7: C, 64.16; H, 5.92.
Calcd for C20H22O7.C2H4O2: C, 60.82; H, 6.03. Found: C, 60.77;
H, 6.11.
B. Mono-l-menthyl Pyromellitate (18b). The preceding first
crop of 17b (0.218 mol) was slurried in glacial acetic acid (300
mL) and heated (water bath at 100 °C). Water (50 mL) was added
and the anhydride dissolved over several minutes. The hot solution
was filtered through a glass frit and rinsed through with HOAc
(100 mL). Water was filtered through next, to a total volume of
ca. 1.5 L. The product crystallized without oiling as snow white
flakes. These were filtered off, washed with H2O, and dried. Yield:
88.55 g (0.218 mol, 99.9% based on 17b). Mp 198.5-203.5 °C
20
20
20
(dec); optical rotation [R] -76.19, [R] -80.43, [R] -92.85
589
578
546
1
(c 10.025, as-is hydrate, tetrahydrofuran); H NMR (CD3OD) δ
0.83 (3H, d, J ) 6.73 Hz), 0.90 (3H, d, J ) 7.11 Hz), 0.95 (3H, d,
J ) 6.19 Hz), [1 H, m, obscured by methyl doublet], 1.09-1.24
(2H, m), 1.46-1.62 (2H, m), 1.69-1.78 (2H, m), 1.97 (H2O),
1.99-2.07 (H, m), 2.17-2.24 (H, m), 4.92-5.00 (H, t of d), 8.01
(H, s), 8.20 (H, s) (3 COOH exchanged with solvent); 13C NMR
(acetone-d6 at 29.8, 206.4) δ 167.4, 167.3, 167.0, 166.5, 136.4,
136.1, 135.3, 134.8, 130.5, 129.6, 76.7, 47.9, 41.1, 35.0, 32.2, 26.9,
24.1, 22.3, 21.0, 16.6. Anal. Calcd for C20H24O8: C, 61.22; H, 6.16.
Calcd for 4C20H24O8.3H2O: C, 59.18; H, 6.33. Found: C, 59.37,
59.14; H, 6.39, 6.40. Found (after heating in vacuo at 60 °C
overnight): C, 60.93; H, 6.20.
sample was recrystallized for analysis from MeOH. Mp 279-286
20
°C (dec), but mostly at 284-286 °C (dec); optical rotation [R]
589
20
20
-62.01, [R] -65.28, [R] -74.89 (c 10.29, tetrahydrofuran);
578
546
1H NMR (CD3OD) δ 0.93 (12 H, s), 0.99 (6 H, s), 1.26-1.41 (6
H, m), 1.70-1.80 (4 H, m), 1.93-2.01 (2 H, m), 2.44 (2 H, m),
5.09-5.14 (2 H, m), 8.02 (2 H, s) (2 COOH exchanged with
solvent); 13C NMR (CD3OD at 49.0) δ 168.8, 168.6, 136.3, 136.0,
130.4, 83.9, 50.1, 49.2, 46.3, 37.0, 28.8, 28.2, 20.1, 19.3, 13.9 (2C
each peak). Anal. Calcd for C30H38O8: C, 68.42, H, 7.27. Found:
C, 68.59; H, 7.30.
4,6-Bis{endo-(1S)-1,7,7-trimethylbicyclo[2.2.1]hept-2-
yloxycarbonyl}isophthalic Acid (Meta-di-l-bornyl Pyromelli-
tate) (12f). The acetic acid filtrates from the preparation of 11f
were concentrated to a syrup on a rotary evaporator. Overnight,
this went solid with balls of needles. The solids were isolated after
a tedious filtration and washed with the usual progression of solvents
(HOAc/H2O: 1:0; 1:1; 0:1). Yield: 99.11 g (0.188 mol, 24.1%).
1-Benzyl 2,4,5-Tri[(1R,2S,5R)-2-isopropyl-5-methylcyclohex-
yl] Benzene-1,2,4,5-tetracarboxylate (Benzyl-tri-l-menthyl Py-
romellitate) (31b). Benzyl pyromellitate (29)4 (18.13 g, 52.66
mmol) was suspended in CH2Cl2 (150 mL) and treated with oxalyl
chloride (12.0 mL) and N,N-dimethylformamide (DMF, 2 drops).
After standing overnight with only partial reaction, the mixture was
warmed intermittently (water bath) and treated with further oxalyl
chloride (4.6 mL) and DMF (4 drops) until the solids had dissolved
completely. The solvent was removed in vacuo to give a tan syrup
20
20
Mp 231.0-238.0 °C (dec); optical rotation [R] -59.76, [R]
589
578
20
-62.89, [R]
-72.02 (c 10.1535, tetrahydrofuran); 1H NMR
546
(CD3OD) δ 0.917 (6 H, s), 0.923 (6 H, s), 0.974 (6 H, s), 1.25-1.35
(6 H, m), 1.69-ca. 1.9 (4 H, m), 1.99 (2 H, m), ca. 2.4 (2H, m),
5.11 (2 H, m), 7.95 (H, s), 8.16 (H, s) (2 COOH exchanged with
solvent); 13C NMR (CD3OD at 49.0) δ 168.7, 168.4, 136.4, 135.8,
130.8 (1C), 130.0 (1C), 83.8, 50.0, 49.3, 46.2, 37.0, 28.7, 28.2,
20.1, 19.3, 13.9 (2C each unless otherwise noted). Anal. Calcd for
C30H38O8: C, 68.42, H, 7.27. Found: C, 68.22, H, 7.40.
(25.55 g, of which 0.49 g was reserved for NMR investigation.20,21
)
The crude triacid chloride (30) (0.05165 mol nominal) was dissolved
in CH2Cl2 (250 mL) and treated with l-menthol (4) (30.45 g, theory
) ca. 24.69 g) and Et3N (32.5 mL). The mixture was kept at room
temperature for 7 days, protected from atmospheric moisture. The
(20) The crude acid chloride was investigated by NMR and found to consist
of two principal components. The major product was benzyl pyromellitate triacid
5-[(1R,2S,5R)-2-Isopropyl-5-methylcyclohexyloxycarbonyl-
]benzene-1,2,4-tricarboxylic Acid (Mono-l-menthyl Pyromelli-
tate) (18b). A. 6-[(1R,2S,5R)-2-Isopropyl-5-methylcyclohexy-
loxycarbonyl]-1,3-dioxo-1,3-dihydroisobenzofuran-5-
carboxylic Acid, 1:1 Solvate with Acetic Acid (Mono-l-menthyl
Pyromellitate Anhydride, 1:1 Acetic Acid Adduct) (17b). A
suspension of PMDA (10) (68.53 g, 0.3142 mol) in a solution of
l-menthol (4) (46.89 g, 0.3001 mol) in CH2Cl2 (500 mL) was treated
with pyridine (26.5 mL), and when this appeared to have little effect,
with triethylamine (44 mL). After several minutes of swirling, the
mixture had become homogeneous. After 90 min, the solution was
shaken with a solution of concentrated HCl (63 mL) in H2O (300
1
chloride (26): H NMR (CDCl3) δ 5.40 (2 H, s), 7.30-7.43 (5 H, m), 8.11 (H,
s), 8.37 (H, s); 13C NMR (CDCl3 at 77.1) δ 165.75, 165.61, 165.47, 162.4, 140.3,
136.4, 136.2, 133.8, 132.5, 131.0, 128.63, 128.56*, 128.50 (2 C), 128.41 (2 C),
68.9.
(21) The minor product displayed NMR chemical shifts consistent with
assignment as benzyl pyromellitate anhydride acid chloride. The ratio of 26 to
this was circa 4:1. 1H NMR (CDCl3) δ 5.22 (2 H, s), 7.28-7.43 (5 H, m), 8.16
(H, s), 8.46 (H, s). Four minor pyromellitate protons of nearly equal intensity,
and presumably due to two compounds, possibly mixed carboxylic acid/acid
chloride species, occurred at δ 8.04, 8.15, 8.33, and 8.39; 13C NMR (CDCl3 at
77.1) δ 166.1, 162.7, 160.07, 159.99, 143.4, 135.4, 133.6, 133.0, 127.78, 127.72,
127.07, 127.01, 124.30, 124.23, 68.7. Other peaks were noted, enough for two
compounds, at intensity circa one-half of this.
J. Org. Chem. Vol. 73, No. 13, 2008 4945

Paine
reaction mixture was shaken with H2O (200 mL) and then with
concentrated HCl (50 mL)-H2O (150 mL). The organic phase was
concentrated in vacuo to give a brown viscous oil (50.13 g). This
was redissolved in CH2Cl2 and chromatographed on silica gel
(Merck # 7734, Kieselgel 60, 70-230 mesh), the product eluting
rapidly as single-spot material. Yield: 32.72 g (0.0431 mol, 83.5%).
129.4, 128.6 (4C), 128.4, 128.3 (4C), 76.5, 67.9, 47.0, 40.4, 34.1,
31.4, 26.2, 23.3, 22.0, 20.8, 16.2 (two carbons per peak unless
otherwise noted). Anal. Calcd for C44H54O8: C, 74.34; H, 7.66.
Found: C, 74.55; H, 7.61.
2,5-Bis[(1R,2S,5R)-2-isopropyl-5-methylcyclohexyloxycarbo-
nyl]terephthalic Acid (Para-di-l-menthyl Pyromellitate) (11b)
from the Dibenzyl Ester (20b). Para-dibenzyl-di-l-menthyl py-
romellitate (20b) (3.57 g, 5.02 mmol) and 10% Pd/C (0.35 g) were
stirred in tetrahydrofuran (70 mL) under H2 (1 atm, ambient
temperature). Gas uptake was complete after 23 min, but stirring
was continued for 2 h. The catalyst was filtered off. The filtrates
were concentrated in vacuo. Yield: quantitative. The crude solids
were crystallized from minimal MeOH. Yield (first crop): 1.07 g
(2.02 mmol, 40.1%). Mp 251.5-255.0 °C (dec).
The compound was recrystallized from 100% ethanol, yield 24.38
20
g (0.0321 mol, 62.2%). Mp 111.5-114.0 °C; optical rotation [R]
589
20
20
1
-83.19, [R] -87.48, [R] -100.00 (c 10.08924, toluene); H
578
546
NMR (CDCl3) δ 0.807 (3H, d, J ) 6.60 Hz), 0.830 (3H, d, J )
6.89 Hz), 0.853 (3H, d, J ) 7.02 Hz), 0.903 (3H, d, J ) 6.41 Hz),
0.916 (3H, d, J ) 6.35 Hz), 0.925 (6H, d, J ) 6.83 Hz), 0.947
(3H, d, J ) 6.35 Hz), 0.955 (3H, d, J ) 6.64 Hz), [3 H, m: obscured
by methyl doublets], 1.05 - 1.18 (6H, m), 1.46 - 1.53 (6H, m),
1.68 - 1.73 (6H, m), 1.90 - 1.99 (3H, m), 2.10 - 2.26 (3H, m),
4.88 - 5.01 (3H, m), 5.34 and 5.41 (calc.) (2H, AB q, J ) 12.3
Hz), 7.30 - 7.43 (5H, m), 8.012 (H, s), 8.020 (H, s); 13C NMR
(CDCl3 at 77.1) δ 165.7, 165.1, 164.73, 164.67; 134.84, 134.76,
134.4, 133.9 (2C*?); 129.1 (2C*?); 128.2 (2C), 128.1, 127.9 (2C);
76.2, 76.0 (2C*); 67.6; 46.95 (2C*), 46.87; 40.36, 40.32, 40.25;
34.13 (2C*), 34.05; 31.36 (2C*), 31.30; 26.27, 26.20, 26.07; 23.38
(2C*), 23.30; 21.99 (2C*), 21.92; 20.78 (2C*), 20.69; 16.38, 16.30
(2C*) (*accidental degeneracy; 1 C per peak unless otherwise
noted). Anal. Calcd for C47H66O8: C, 74.37; H, 8.76. Found: C,
74.48; H, 8.79.
1,2,4,5-Tetrakis-[(1R,2S,5R)-2-isopropyl-5-methylcyclohex-
yl] Benzene-1,2,4,5-tetracarboxylate (Tetra-l-menthyl Pyrom-
ellitate) (26b).16,17 A solution of tetraethyl pyromellitate4 (25)
(110.21 g, 0.3008 mol) and (1R,2S,5R)-(-)-menthol (4) (234.46
g, 1.5004 mol) in p-xylene (603 mL) was prepared. About one-
third of it was heated to reflux under N2 (a safety precaution since
hydrogen is to be eVolVed as the NaH is added) in a 4 L heavy
wall standard taper 29/42 erlenmeyer flask on a hot plate with
magnetic stirring. Small portions of NaH slurry in p-xylene were
added dropwise until ethanol began to be vigorously evolved.
Thereafter, alternating portions of solution and catalyst were added
to the boiling mixture, to keep the reaction under control. (Warning:
ethanol, unlike methanol, azeotropes with xylene, so that boiling
can become quite threatening in ethyl ester systems, eVen in a large
flask. The ingredients cannot all safely be in the flask initially, on
this scale.) Further small portions of catalyst continued to be added
periodically, until no further effect was noted, and the vapor
temperature remained above 139 °C. The dark brown reaction
mixture was allowed to cool for 10 min, before being treated with
HOAc (50 mL), which caused the color to fade considerably. The
reaction mixture was washed with H2O (1000 mL), which was back-
extracted with EtOAc (100 mL). The combined organic phases were
filtered (with addition of CH2Cl2 as needed to prevent crystalliza-
tion) and then concentrated (rotary evaporator). The residue was
crystallized from 100% ethanol. Seeding was essential to prevent
oiling. Yield: 214.26 g (0.265 mol, 88.2%). A recrystallization from
100% EtOH afforded 200.39 g (0.248 mol, 82.5%) of pale cream
dense granules. Mp 120-123.5 °C [lit. mp 123.5-124.5 °C]; optical
2,4,5-Tris[(1R,2S,5R)-2-isopropyl-5-methylcyclohexyloxycar-
bonyl]benzoic Acid (Tri-l-menthyl Pyromellitate) (27b). Benzyl
tri-l-menthyl pyromellitate (31b) (7.69 g, 10.1 mmol) and 10% Pd/C
(0.51 g) were stirred in tetrahydrofuran (120 mL) under H2 at 1
atm and ambient temperature until gas uptake (254 mL) ceased
(95 min). The catalyst was filtered off and the filtrates evaporated
to dryness in vacuo. Yield: quantitative. This was characterized
1
only by NMR. H NMR (CDCl3) δ 0.824 (6 H, d, J ) 6.86 Hz),
0.836 (3 H, d, J ) 6.90 Hz), 0.900 (3 H, d, J ) 6.97 Hz), 0.914 (6
H, d, J ) 7.05 Hz), 0.945 (3 H, d, J ) 6.64 Hz), 0.952 (3 H, d, J
) 6.46 Hz), 0.959 (3 H, d, J ) 6.38 Hz), [3 H, m: obscured by
methyl doublets], 1.05-1.18 (6 H, m), 1.44-1.53 (6 H, m),
1.70-1.74 (6 H, m), 1.87-1.96 (3 H, m), 2.16-2.25 (3 H, m),
4.92-5.01 (3 H, m), 7.89 (H, s), 8.16 (H, s), 9.22 (H, br s); 13C
NMR (CDCl3 at 76.9) δ 169.5, 165.8, 165.4, 164.8; 135.8, 135.5,
134.0, 132.1; 129.8, 128.7; 76.49, 76.32, 76.18; 46.91 (3 C*); 40.41,
40.27, 40.10; 34.15 (3 C*); 31.42 (3 C*); 26.18 (3 C*); 23.36 (3
C*); 21.99 (3 C*); 20.79 (2 C*), 20.71; 16.33 (2 C*), 16.20
(*accidental degeneracy).
20
20
20
rotation [R] -95.38, [R] -99.98, [R] -114.24 (c 10.18,
589
578
546
toluene) [lit.16 [R]
microanalysis are reported in Supporting Information.]
-102.5 (c 1.1, benzene)]; [NMR and
20
589
1,4-Dibenzyl 2,5-Di[(1R,2S,5R)-2-isopropyl-5-methylcyclo-
hexyl] Benzene-1,2,4,5-tetracarboxylate (Para-dibenzyl-di-l-
menthyl Pyromellitate) (20b). Para dibenzyl pyromellitate4 (13.04
g, 0.03 mol), suspended in CH2Cl2 (100 mL), was treated with
oxalyl chloride (8.0 mL, ca. 11.64 g, 0.0917 mol) and N,N-
dimethylformamide (5 drops). Gas evolution (hood!) began as soon
as the DMF was added. After standing overnight at room temper-
ature, protected from atmospheric moisture, the reaction mixture
was concentrated (rotary evaporator). The residual crude acid
chloride 194 (14.27 g, 14.15 g ) theory) was dissolved in CH2Cl2
(108 mL) and treated with l-menthol (4) (14.73 g, 0.094 mol) and
Et3N (10.0 mL, 7.34 g, 0.0725 mol). After standing at room
temperature for 3 days, protected from moisture, the reaction
mixture was extracted with H2O (100 mL). The organic phase was
isolated, filtered, and then concentrated (rotary evaporator). The
1,2,4,5-Tetrakis[(1S,2R,5S)-2-isopropyl-5-methylcyclohexyl]
Benzene-1,2,4,5-tetracarboxylate (Tetra-d-menthyl Pyromelli-
tate) (26a). This was prepared similarly from 254 (37.97 g, 0.1036
mol) and (1S,2R,5S)-(+)-menthol (3) (95.34 g, 0.6101 mol) in
p-xylene (250 mL). Yield: 69.50 g (0.0861 mol, 83.1%). Mp
116.5-117.5 °C; recrystallized and melted 122.0-125.5 °C.
Another preparation had mp 119.5°-121.5 °C, recrystallizing in
part to crystals which melted sharply at 125.5-126.0 °C. A 10 g
sample was recrystallized from n-butanol (21 g), in 85.1% recovery.
20
20
20
Optical rotation: [R] +92.96, [R] +97.33, [R] +111.34 (c
589
578
546
10.49, toluene). A 50 g sample, recovered in 88.7% yield from
n-butanol (75.4 g), had corresponding rotations of +91.20, +95.84,
and +109.26 (c 9.775, toluene). This (43 g) was then recrystallized
from toluene (43 g), with only 40.9% recovery. The optical rotations
were now respectively +94.08, +98.89, and +112.72 (c 10.62,
toluene), for an estimated optical purity of 98.6%. 1H NMR (CDCl3)
δ 0.82 (12 H, d, J ) 7.00 Hz), 0.91 (12 H, d, J ) 6.94 Hz), 0.95
(12 H, d, J ) 6.54 Hz), (2D NMR and integrals suggest 4 H, m
are obscured under the methyl doublets), 1.06-1.17 (8 H, m),
1.44-1.59 (8 H, m), 1.71-1.74 (8 H, d, J ) 11.72 Hz), 1.90-1.97
(4 H, quintet of doublets, J ) 2.6 and 7.0 Hz), 2.18-2.22 (4 H,
m), 4.90-4.97 (4 H, t of d, J ) 4.4 and 10.8 Hz), 7.94 (2 H, s);
13C NMR (CDCl3 at 77.0) δ 165.5, 134.6, 129.3 (2 C), 76.3, 47.1,
40.4, 34.2, 31.5, 26.3, 23.4, 22.0, 20.8, 16.4 (4 carbons per peak
residue was crystallized from 100% ethanol (80 mL). Yield: 14.49
20
g (0.0204 mol, 67.9%); mp 108.5-110.5 °C; optical rotation [R]
589
20
20
1
-65.41, [R] -68.87, [R] -78.92 (c 10.01324, toluene); H
578
546
NMR (CDCl3) δ 0.76 (6H, d, J ) 6.85 Hz), 0.86 (6H, d, J ) 6.98
Hz), 0.91 (6H, d, J ) 6.45 Hz), [2 H, m: obscured by methyl
doublets], 0.94-1.23 (4 H, m), 1.36-1.58 (4H, m), 1.63-1.77 (4H,
crude d), 1.80-1.94 (2H, quintet of d), 2.04-2.15 (2H, crude d of
t), 4.84-4.97 (2H, t of d, J ) 10.88 and 4.43 Hz), 5.33 and 5.41
(calc.) (4H, ABq, J ) 12.35 Hz), 7.32-7.52 (10H, m), 8.02 (2H,
s); 13C NMR (CDCl3 at 77.1) δ 166.2, 165.0, 135.1, 134.5, 134.4,
4946 J. Org. Chem. Vol. 73, No. 13, 2008

Esters of Pyromellitic Acid. Part II
unless otherwise noted). Anal. Calcd for C50H78O8: C, 74.40; H,
9.74. Found: C, 74.10; H, 9.70.
The system was scaled up to 50.06 mmol (Table 3, round 1B).
The mass of cottony needles was mashed and filtered, washed twice
with small portions of MeOH, and dried. The filtrates were
concentrated in vacuo to a syrup, which afforded a second crop.
The combined solids (19.71 g were used) were recrystallized from
MeOH, giving 13f in 84.8% overall yield, based on the system
content of 1.
B. Isolation of (R)-(+)-Nicotine (14) from its Salts with a
Resolving Agent (“Standard Method”). All of the mother liquors
from parts A and B above were combined and evaporated to
dryness. The residue, dissolved in MeOH (150 mL), was treated
with concentrated HCl (6.0 mL). Water (200 mL) was added, after
the resolving agent (here 11f) began to crystallize. The solids were
filtered off, rinsed with H2O, and recovered. Concentrated HCl (4.0
mL) was added to the filtrates, which were then concentrated in
vacuo to a syrup. Pellets of NaOH (8.64 g, 0.216 mol) were added
to the residue, followed by H2O (10 mL) with ice cooling. The
alkaloid was extracted three times with CH2Cl2-Et2O [Note:
petroleum ether bp 30/60 °C or hexane(s) are also satisfactory, used
alone] and isolated by evaporation of solvent, followed by Kugel-
rohr distillation. Yield: 5.51 g (4.85 g was the expected yield,
assuming that the previously isolated salts were not solvated); the
optical rotation is presented in Table 3.
Resolution of (R,S)-Nicotine by Manual Triage. Para-di-l-
menthyl pyromellitate (11b) (26.55 g, 0.05 mol) and (R,S)- nicotine5
(1 + 14) (8.10 g, 0.05 mol) were dissolved in hot methanol (150
mL), and the resulting solution was allowed to crystallize undis-
turbed to provide the results of “round 1” (Table 1). The decanted
mother liquors were permitted to evaporate over a month in an
open erlenmeyer flask in the refrigerator, depositing an admixture
of large transparent chunks and crusts of small feathery crystals
which became opaque while drying out, while the chunks remained
transparent. From this was obtained by manual triage the quantities
reported for “round 2” (Table 1). Total recovery of solids: 30.90 g
(89%).
The (S)-diastereomer 13b was recrystallized once from MeOH,
forming large crystals (the largest weighed 2.54 g) and some crusts
20
(yield in Table 1). Optical rotation: [R] -50.6 (c 10.38, methanol).
589
Anal. Calcd for C40H56N2O8: C, 69.34; H, 8.15; N, 4.04. Calcd for
2C40H56N2O8.CH3OH: C, 68.62; H, 8.25; N, 3.95. Calcd for
C40H56N2O8.CH3OH: C, 67.93; H, 8.34; N, 3.86. Found: C, 68.73;
H, 8.55; N, 3.94.
The (R)-diastereomer 15b was recrystallized similarly and dried
(Table 1). The salts were shown to lose around 10% of their mass
upon drying, as they became opaque. This corresponds to ca. 2.5
C. Resolution Continued with 11a. The above crude (R)-(+)-
nicotine (14) and para-di-d-menthyl pyromellitate (11a) were
dissolved in MeOH and seeded, forming elongated flat prisms
(further details in Table 3). Twelve years later, the contained
nicotine was isolated from 9.07 g (13.09 mmol if unsolvated) of
the above solids, using the standard method. Yield of (R)-(+)-
nicotine (14) after Kugelrohr distillation was 2.03 g (95.6% yield).
20
mol of methanol of crystallization. Optical rotation: [R] -83.9
589
(c 10.70, methanol). Anal. Calcd for C40H56N2O8: C, 69.34; H, 8.15;
N, 4.04. Found: C, 69.60; H, 8.04; N, 3.92.
Efficiency of Resolution, Starting with (90% S/10% R)-
Nicotine. (90% S, 10% R)-Nicotine (16.24 g, 0.1 mol)[prepared
by diluting racemic nicotine5 with four parts of (S)-nicotine (1)]
and para-di-l-menthyl pyromellitate (11b) (53.12 g, 0.1 mol)
were allowed to crystallize undisturbed from MeOH (88.59 g,
of which 2.3 g evaporated during the crystallization). “Transpar-
ent salt” 13b was the only phase that crystallized. The resulting
13b was recrystallized once from methanol. The rejected nicotine
was recovered essentially quantitatively from each round of
crystallization, and the optical rotation was determined (Table
2). An accounting of the (R)-nicotine content of the system
(Table 2) suggested that after the second crystallization, 13b
contained 99.8% (S)-antipode. This purity was confirmed 12
years later when the contained nicotine was isolated (yield:
2.11 g, 98.8%) by the standard method from a 9.12 g sample
(13.16 mmol) of the solids. [These had remained stable in the
interim, stored at ambient temperature, with the exception of
having at last gone opaque, presumably due to the loss of solvent
of crystallization.] The contained resolving agent crystallized
smartly from acidified aqueous methanol during workup, sug-
gesting that it too was intact, yield: 6.97 g (99.8% recovery).
20
20
20
Optical rotation: (R) +164.93, (R) +173.535, (R) ) +198.66
589
578
546
(neat, 1.000 cm microcell, T ) 19.98 °C). This computes to 96.51%
optical purity or 98.26% (R)- and 1.74% (S)-antipode. The yield
of recovered resolving agent, isolated by simple filtration, was
6.67 g (96.0% yield). The intact nature of the resolving agent was
confirmed by 13C and 1H NMR and by melting point [mp
253.5-255.0 °C (dec)].
Resolution of the Contained Terpene Alcohols. A. Purifica-
tion of Di-d-menthyl Pyromellitate (11a). Material (11a) as
isolated by method A was estimated to be 97.56% optically pure
or to contain d-menthol (3) of 98.8% chiral purity. A sample (10.08
g) was recrystallized from MeOH (31.34 g) to give 7.21 g (71.5%).
The optical purity at this stage was 99.1%, corresponding to
d-menthol (3) of 99.55% absolute purity. The resulting solids (6.70
g) were crystallized again from MeOH (20.11 g), to give 4.99 g
(74.5% recovery, 53.3% overall). The optical purity was now
99.9%, corresponding to 99.95% absolute purity for the contained
d-menthol (3). The optical rotations for the starting material and
the two recrystallized fractions are reported with the data for method
A.
1
The 13C and H NMR spectra confirmed the completely intact
nature of the recovered resolving agent [mp 250.7-252.5 °C
(dec)]. The resulting recovered nicotine provided the following
20
20
20
optical rotations: (R)
-170.315, (R)
-179.245, (R)
589
578
546
B. Isolation of the Contained d-Menthol (3). Thus purified
para-di-d-menthyl pyromellitate (11a) (2.72 g, 5.13 mmol) and
NaOH (5.05 g, 0.126 mol) were dissolved in H2O (100 mL), and
the mixture was distilled into an overhead solvent-still head receiver,
until no further oily d-menthol (3) appeared in the condensate.
Dichloromethane was cautiously added to the hot still-pot contents
to distill over and rinse the product from the condenser and extract
it from the distillate. The organic phase was isolated and concen-
trated in vacuo, and the residue was distilled (Kugelrohr). Yield:
1.59 g (10.18 mmole, 99.3%). Mp 40.0-41.5 °C (uncorr); optical
-205.215 (neat, 1.000 cm microcell, T ) 19.98-20.01 °C). The
corresponding values for (S)-(-)-nicotine (1), freshly isolated
by the standard method from the bis-bitartrate dihydrate, using
the same 1.000 cm microcell, were -170.85, -179.81, and
-205.89, respectively. The same sample, using a 10.000 cm cell,
afforded values of -170.614, -179.531, and -205.509, respec-
tively. Calculations using these numbers show that the resolved
sample contained 99.84-99.92% (S)-antipode.
Resolution of Nicotine with Para-di-l-bornyl Pyromellitate
(11f). A. Initial Resolution with 11f. Para-di-l-bornyl pyromellitate
(11f) and (R,S)-nicotine5 (1 + 14) were dissolved in hot methanol
and seeded. Overnight, a mass of interlocking needles formed. The
mother liquors were drained off. The crystals were washed in place
with MeOH (5 mL) and dried (Table 3, round 1A, 20.03 mmol
scale).
20
20
20
rotation [R] +47.84, [R] +50.09, [R] +56.49, (c 10.4475,
589
578
546
toluene). Corresponding rotations for natural l-menthol (4) were
-47.89, -50.14, and -56.53, respectively (c 11.5915, toluene),
so that the d-menthol (3) calculated to have been 99.9% optically
pure (99.95% absolute purity).
J. Org. Chem. Vol. 73, No. 13, 2008 4947

Paine
Acknowledgment. The encouragement and support of the
management and staff of Philip Morris USA throughout all this
work are gratefully recorded. This work was presented in part
at the 220th National Meeting of the American Chemical
Society, in Washington, DC, August 20-24, 2000, abstracts
ORGN 346 and ORGN 518.
Supporting Information Available: General experimental
methods, experimental procedures, and full characterization for
compounds 11c, 11d, 11e, 12a (13C NMR), 18g, 22b, 23b, 24a,
24b, 26c, 26d, and 26f. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO8005446
4948 J. Org. Chem. Vol. 73, No. 13, 2008