The effect of acid strength on the Mitsunobu esterification reaction: Carboxyl vs hydroxyl reactivity

Article abstract of DOI:10.1021/jo952180e

In the presence of carboxylic acids, the adduct formed between triphenylphosphine and diisopropyl azodicarboxylate reacts to form mono- and bis-acylated hydrazides and the carboxylic acid anhydrides. These products are formed via attack of the carboxylate on the triphenylphosphonium group of the adduct, with weaker acids reacting much faster than stronger acids. This provides an explanation for the observation in the literature that acids stronger than acetic acid, such as 4-nitrobenzoic acid and chloroacetic acid, provide better yields in esterification reactions, since reaction of the alcohol with the phosphonium group of the adduct is more rapid than the competing reaction of the carboxylate for the phosphonium group.

Full text of DOI:10.1021/jo952180e

J . Org. Chem. 1996, 61, 2967-2971
2967
Th e Effect of Acid Str en gth on th e Mitsu n obu Ester ifica tion
Rea ction : Ca r boxyl vs Hyd r oxyl Rea ctivity
David L. Hughes* and Robert A. Reamer
Department of Process Research, Merck Research Laboratories, Rahway, New J ersey 07065
Received December 8, 1995X
In the presence of carboxylic acids, the adduct formed between triphenylphosphine and diisopropyl
azodicarboxylate reacts to form mono- and bis-acylated hydrazides and the carboxylic acid
anhydrides. These products are formed via attack of the carboxylate on the triphenylphosphonium
group of the adduct, with weaker acids reacting much faster than stronger acids. This provides
an explanation for the observation in the literature that acids stronger than acetic acid, such as
4-nitrobenzoic acid and chloroacetic acid, provide better yields in esterification reactions, since
reaction of the alcohol with the phosphonium group of the adduct is more rapid than the competing
reaction of the carboxylate for the phosphonium group.
In tr od u ction
on the outcome of Mitsunobu reactions is due to the
variability in the rates of reaction of the adduct formed
between diisopropyl azodicarboxylate (DIAD) and tri-
phenylphosphine with either the carboxyl group of the
acid (which leads to decomposition) or the hydroxyl group
of the substrate (which leads to esterification).
The Mitsunobu reaction has found widespread use in
organic synthesis over the past two decades, especially
for the inversion of alcohols via an esterification/hydroly-
sis procedure.1 While a variety of carboxylic acids have
been used for the esterification reaction, benzoic acid has
been the most widely used. In a recent survey that
compiled a list of reactions reported during the 1980-
1990 period, 374 esterification reactions were listed; 57%
employed benzoic acid, 10% acetic acid, 9% formic acid,
8% 4-nitrobenzoic acid or 3,5-dinitrobenzoic acid, 13%
other acids, and 3% unspecified.2 In 1991, Martin and
Dodge reported that hindered alcohols gave much im-
proved yields in the Mitsunobu reaction when 4-ni-
trobenzoic acid was used instead of benzoic acid.3 In
1992, Bessodes and co-workers demonstrated a similar
effect with chloroacetic acid, in that hindered alcohols
that reacted in very low yields with benzoic acid or acetic
acid gave good yields with chloroacetic acid.4 Dodge and
co-workers followed up these reports by a systematic
study of the effect of acidity on the yield of esterification
of menthol with various acids. With a series of substi-
tuted benzoic acids, they found that the least acidic
members (4-MeO, 4-F, H, 4-Me) provided yields of 20-
30%, while the most acidic (4-NO2, 4-CN, 4-MeSO2)
furnished yields in the 60-80% range.5 On the basis of
these reports, 4-nitrobenzoic acid has become the acid of
choice when inverting carbons bearing a hydroxyl group,
as demonstrated by >100 citations to the Martin and
Dodge paper. While the authors of the above papers
conceded they had no firm explanation for the improved
yields with the more acidic carboxylic acids, they ratio-
nalized the effect based on a finding of Camp and
J enkins6 that the oxyphosphonium ion (the reactive
intermediate) is favored over the phosphorane (unreac-
tive intermediate) when the acid strength is increased.
In this paper we demonstrate that the effect of acidity
The first step in the Mitsunobu reaction is formation
of the adduct 1 between PPh3 and DIAD (eq 1). In the
presence of a carboxylic acid, the adduct is protonated,
which generates the carboxylate as counterion. With
formic acid in dichloromethane solvent, we had previ-
ously found that the adduct degrades to form primarily
the N-formyl derivative (eq 2).7 This reaction rate is
comparable to the esterification reaction; thus, 2.5 equiv
of DIAD and PPh3 was required to complete the esteri-
fication reaction. We have now extended this initial
observation to include several other acids and other
solvents and have found that the acid strength has a
profound effect on the stability of adduct 1 and on the
pathway by which it reacts with nucleophiles.
Resu lts
Mitsunobu reactions were conducted in the absence of
alcohol substrates to determine the fate of triphenylphos-
phine, diisopropyl azodicarboxylate (DIAD), and the
carboxylic acid. The ratios of products and their rate of
formation were dependent on choice of solvent and the
acidity of the carboxylic acid, as discussed below.
Acetic Acid . The three products isolated by flash
chromatography from reaction of DIAD, PPh3, and HOAc
are the mono-N-acetylhydrazine 4, the bis(N,N ′-acetyl-
hydrazine) 3, and the hydrazide 5, as shown in Table 1.
The product distribution and the rates of formation are
X Abstract published in Advance ACS Abstracts, April 1, 1996.
(1) Mitsunobu, O. Synthesis 1981, 1-28.
(2) Hughes, D. L. Org. React. 1992, 42, 335-656.
(3) Martin, S. F.; Dodge, J . A. Tetrahedron Lett. 1991, 32, 3017-
3020.
(4) Saiah, M.; Bessodes, M.; Antonakis, K. Tetrahedron Lett. 1992,
33, 4317-4320.
(5) Dodge, J . A.; Trujillo, J . I.; Presnell, M. J . Org. Chem. 1994, 59,
234-236.
(6) Camp, D.; J enkins, I. D. J . Org. Chem. 1989, 54, 3045-3049,
3049-3054.
(7) Hughes, D. L.; Reamer, R. A.; Bergan, J . J .; Grabowski, E. J . J .
J . Am. Chem. Soc. 1988, 110, 6487-6491.
S0022-3263(95)02180-3 CCC: $12.00 © 1996 American Chemical Society

2968 J . Org. Chem., Vol. 61, No. 9, 1996
Ta ble 1. P r od u cts a n d Rea ctivity for Decom p osition of Ad d u ct 1 w ith HOAca
Hughes and Reamer
Ta ble 2. P r od u cts a n d Rea ctivity for Decom p osition of Ad d u ct 1 w ith P h CO2Ha
Ta ble 3. P r od u cts a n d Rea ctivity for Decom p osition of Ad d u ct 1 w ith 4-Nitr oben zoic Acid a
dependent on the ratios of reactants and the solvent. The
temperature. Thus, the decomposition could be readily
monitored by 31P NMR, with adduct 1 having a resonance
at 53 ppm and PPh3dO at 29 ppm. In these cases, no
anhydride was isolated by flash chromatography, nor
could it be detected by 13C NMR.
fastest reaction occurred in dichloromethane and DMF,
where adduct 1 reacted completely within 15 min at 0-5
°C. On the other hand, the decomposition of the adduct
required 8 h at 0-5 °C in toluene, in part because the
adduct was largely out of solution under these conditions.
Ben zoic Acid . Besides the hydrazide and mono- and
bis-acylated products, benzoic anhydride was also ob-
served and isolated from reactions of PPh3, DIAD, and
PhCO2H. The ratios of each product and the reaction
times are shown in Table 2. Compared to acetic acid,
the phosphonium adduct with benzoate counterion is
roughly 3- to 5-fold more stable than with acetate. As
with acetate, the slowest reaction occurred in toluene
solvent.
4-Nitr oben zoic Acid a n d Ch lor oa cetic Acid . As
with HOAc and PhCO2H, the decomposition products of
adduct 1 in the presence of 4-nitrobenzoic acid and
chloroacetic acid are the hydrazide and the mono- and
bis-acylated products. The results are compiled in Tables
3 and 4. With these acids, the phosphonium adduct is
much more stable than with acetate or benzoate, with
decomposition occurring over a period of days at room
Discu ssion
The data in Tables 1-4 demonstrate that the phos-
phonium adduct 1 is unstable, decomposing to the
hydrazide and acylated hydrazides. The identification
of benzoic anhydride at the 20-35% level in the reactions
involving benzoic acid provides insight on the pathway
by which adduct 1 is decomposing. The anhydride and
the acylated hydrazides are apparently formed by reac-
tion of benzoate with the phosphonium group of adduct
1 to form the benzoylphosphonium salt (Scheme 1, path
1). This salt then reacts with another mole of benzoate
to form benzoic anhydride plus the hydrazide (path c) or
can react with the hydrazide anion to produce N-
benzoylhydrazide (path b). Thus, for a successful Mit-
sunobu esterification to take place, the alcohol must react
faster (path 2) than the carboxylate (path 1) with adduct
1.

Mitsunobu Esterification Reaction
J . Org. Chem., Vol. 61, No. 9, 1996 2969
Ta ble 4. P r od u cts a n d Rea ctivity for Decom p osition of Ad d u ct 1 w ith Ch lor oa cetic Acid a
Sch em e 1
The mono- and bis-acylated hydrazides could also
This is true in two of the four solvents examined (Table
1). In the other two solvents, the hydrazide is present
in greater amounts than the anhydride, but formation
of the hydrazide may also arise by protonation of the
hydrazide anion formed after displacement of the phos-
phonium group (path a).8
potentially be formed by reaction of hydrazide 5 with the
benzoic anhydride formed in the reaction. However,
extended reaction times did not alter the ratio of prod-
ucts, indicating that the benzoic anhydride was not
reacting with hydrazide 5 or mono-acylated product 7
under the reaction conditions. The reaction pathway
proposed in Scheme 1 would also suggest that equal
amounts of anhydride and hydrazide 5 should be formed
via path c if no further reaction of the anhydride occurs.
The data in Tables 1 through 4 indicate that the rate
of decomposition of adduct 1 is sensitive to the carboxy-
late counterion. With equimolar HOAc, the reactions are
complete at 0-5 °C in 1 h or less in DMF, THF, and
dichloromethane. Hence, under these conditions, for an
esterification reaction to occur, the reaction of the alcohol
with adduct 1 must be very rapid; otherwise, the adduct
will completely react with the carboxylate before the
alcohol has a chance to react. Thus, with hindered
alcohols such as menthol, reaction of adduct 1 with the
alcohol is apparently slower than reaction with the
carboxylate under many conditions (dependent on sol-
vent, temperature), such that no esterification occurs.
This probably accounts for the relatively low number of
Mitsunobu esterifications carried out with HOAc. The
slow reaction of HOAc with adduct 1 in toluene solution
is due in part to the poor solubility of 1 in this solvent,
but this result suggests that toluene might be the best
solvent for esterifications with HOAc. In fact, esterifi-
cation of menthol with HOAc and 5 equiv of DIAD/PPh3
at -15 °C for 12 h and then at 5 °C for 12 h in toluene
gave a 67% isolated yield of the inverted acetate, com-
pared to the 0% yield reported by Dodge for reaction in
benzene at 5 °C to rt with the same ratio of reactants.5
In addition, quantitation of the decomposition products
3-5 in this reaction indicated that similar ratios of these
(8) The reaction pathways depicted in Scheme 1 do not account for
the formation of the N,N ′-diacylated products. We do not believe that
the mono-acylated product is converted to the bis-acylated product
since the proportion of each stays relatively constant throughout a
reaction (instead of mono forming first followed by an increase in bis)
and since the ratio remains unchanged once a plateau has been
reached. A hypothesis for the formation of the bis adducts is shown
below and involves acylation of adduct 1 before loss of the phosphonium
group.

2970 J . Org. Chem., Vol. 61, No. 9, 1996
Hughes and Reamer
date (hydrazide 5 could only be observed with iodine visualiza-
tion). Three products were isolated in the following order of
elution:
products were formed in this esterification reaction as
in the reaction reported in Table 1 in the absence of an
alcohol substrate.
A comparison of the data in Tables 1 and 2 indicates
that reaction of adduct 1 with PhCO2H is generally three
to five times slower than with HOAc. This slower
decomposition rate allows more time for the alcohol
activation reaction to occur, which presumably accounts
for the larger number of successful esterifications with
benzoic acid reported in the literature.2
N,N ′-Dia cet yl-N,N ′-b is(isop r op ylca r b oxy)h yd r a zin e
(3) (356 mg, 25%): mp 18-19 °C; 1H NMR (C6D6) δ 0.90 (12H,
d, J ) 6.5 Hz), 2.48 (6H, s), 4.83 (2H, septet, J ) 6.5 Hz); 13
C
NMR (C6D6) δ 21.0, 21.1, 24.7, 71.7, 151.7, 168.2. Anal. Calcd
for C12H20N2O6: C, 49.99; H, 6.99; N, 9.72. Found: C, 49.72;
H, 7.07; N, 9.52.
N-Acetyl-N,N ′-bis(isopr opylcar boxy)h ydr azin e (4) (573
1
mg, 47%): mp 59-60 °C (heptane); H NMR (CDCl3) δ 1.27
(6H, br d, J ) 6.0 Hz), 1.31 (6H, d, J ) 6.3 Hz), 2.54 (3H, s),
4.97 (1H, septet, J ) 6.3 Hz), 5.03 (1H, septet, J ) 6.2 Hz),
6.6 (1H, br s); 13C NMR (CDCl3) δ 21.6, 21.8, 25.3, 70.3, 72.2,
152.6, 155.0, 170.8. Anal. Calcd for C10H18N2O5: C, 48.77;
H, 7.37; N, 11.38. Found: C, 48.87; H, 7.37; N, 11.29.
N,N ′-Bis(isop r op ylca r b oxy)h yd r a zin e (5) (296 mg,
29%): mp 107-108 °C (80% water, 20% i-PrOH); 1H NMR
(CDCl3) δ 1.25 (12 H, d, J ) 6.5 Hz), 4.94 (2H, septet, J ) 6.5
Hz), 6.55 (2H, br s); 13C NMR (CDCl3) δ 21.9, 70.0, 156.7. Anal.
Calcd for C8H16N2O4: C, 47.05; H, 7.90; N, 13.72. Found: C,
47.08; H, 7.85; N, 13.73.
HPLC was used to quantify amounts of these products
formed from experiments in other solvents, as reported in
Table 1. HPLC conditions were as follows: DuPont C18-RX
column, 25 cm × 4.6 mm; detection at 210 nm; 1.5 mL/min
flow rate; ambient temperature; gradient elution from 30%
MeCN/70% 0.1% aqueous H3PO4 to 80% MeCN/20% 0.1%
aqueous H3PO4 over 25 min; elution times, 5 at 4.0 min; 4 at
6.6 min; 3 at 13.5 min, and triphenylphosphine oxide at 10
min.
The rates of decomposition of adduct 1 in the presence
of chloroacetate or 4-nitrobenzoate are on the order of
102-103-fold slower than with acetate or benzoate, as
adduct 1 takes 0.5-10 days to completely react at 22 °C
with these acids (Tables 3 and 4). On the other hand,
the alcohol activation step is also slower with these less
basic counterions, apparently because the carboxylate
must act as a base to deprotonate the alcohol for alcohol
activation to occur.7 Thus, alcohol activation is 100-fold
faster with acetate than with chloroacetate in dichlo-
romethane.7 However, decomposition of adduct 1 is
about 500- to 1000-fold slower with chloroacetate than
with acetate. Therefore, with chloroacetate as counte-
rion, the overall effect is to increase the relative rate of
hydroxyl attack over carboxylate attack on adduct 1 by
5-10-fold, leading to improved yields of ester.
Su m m a r y
P r od u ct Isola tion fr om Rea ction w ith Ben zoic Acid .
Triphenylphosphine (2.74 g, 10.5 mmol), benzoic acid (1.40 g,
11.5 mmol), and THF (25 mL) were combined and cooled to 0
°C. Diisopropyl azodicarboxylate (2.07 g, 10.2 mmol) was
added over 10 min, keeping the temperature below 5 °C. After
2 h at 5 °C, the mixture was warmed to room temperature,
concentrated to an oil in vacuo, and dissolved in 5 mL of
dichloromethane. This material was chromatographed on 125
g of silica gel using an eluent consisting initially of 1:7 EtOAc:
hexane followed by a step gradient to 1:2 EtOAc:hexane. Four
products were isolated in the following elution order:
Ben zoic a n h yd r id e (0.24 g, 19%): 1H and 13C NMR and
HPLC retention time match that of an authentic sample.
N ,N ′-Dib e n zoyl-N ,N ′-b is(isop r op ylca r b oxy)h yd r a -
zin e (6) (0.47 g, 11%): mp 91-92 °C (90:10 EtOH:water); 1H
NMR (CDCl3) δ 1.10 (12H, dd, J ) 1.8, 6.4 Hz), 4.93 (2H,
septet, J ) 6.6 Hz), 7.40-7.73 (10H, m); 13C NMR (CDCl3) δ
21.3, 72.8, 128.2, 128.2, 132.0, 134.8, 151.8, 169.3. Anal. Calcd
for C22H24N2O6: C, 64.07; H, 5.87; N, 6.79. Found: C, 63.92;
H, 5.79; N, 6.75.
In Mitsunobu esterification reactions, unwanted car-
boxyl activation competes with desired hydroxyl activa-
tion. For a successful esterification to occur, a delicate
balance must be established such that the carboxylate
is a strong enough base to initiate alcohol activation, but
not such a strong nucleophile that it reacts with adduct
1 faster than the alcohol. With acetic acid, this balance
is difficult to achieve, and esterifications are only suc-
cessful within a narrow window of experimental condi-
tions. With stronger acids such as chloroacetic and
4-nitrobenzoic acids, this balance is more easily achieved;
the rates of both carboxyl and hydroxyl activation are
reduced in comparison to acetate, but hydroxyl activation
is significantly favored, such that esterifications with
these acids can occur with even hindered alcohol sub-
strates.
N-Ben zoyl-N,N ′-bis(isop r op ylca r boxy)h yd r a zin e (7)
(1.47 g, 47%): mp 120-121 °C (2:1 EtOH:water); 1H NMR
(CDCl3) δ 1.05 (6H, d, J ) 6.1 Hz), 1.28 (6H, d, J ) 6.2 Hz),
4.87 (1H, septet, J ) 6.2 Hz), 5.00 (1H, septet, J ) 6.2 Hz),
6.98 (1H, br s), 7.35-7.70 (5H, m); 13C NMR (CDCl3) δ 21.3,
21.9, 70.6, 72.4, 128.1, 128.1, 131.9, 135.2, 152.9, 155.3. Anal.
Calcd for C15H20N2O5: C, 58.43; H, 6.54; N, 9.09. Found: C,
58.30; H, 6.51; N, 9.04.
Exp er im en ta l Section
Gen er a l. Triphenylphosphine, diisopropyl azodicarboxy-
late, chloroacetic acid, chloroacetic anhydride, benzoic acid,
benzoic anhydride, (1S)-(+)-neomenthol acetate, (-)-menthol,
and 4-nitrobenzoic acid were purchased from Aldrich and used
without purification. Solvents were dried with molecular
sieves to a water level of <40 mg/L. Melting points are
N,N ′-Bis(isop r op ylca r boxy)h yd r a zin e (5) (0.70 g, 33%).
Spectral analysis was the same as above.
uncorrected. 1H NMR spectra were obtained at 250 MHz, 13
C
at 62.5 MHz. Flash chromatography was performed using E.
Merck 230-400 mesh silica gel.
HPLC was used to quantify amounts of these products
formed from experiments in other solvents, as reported in
Table 2. HPLC conditions were as follows: DuPont C18-RX
column, 25 cm × 4.6 mm; detection at 210 nm; 1.5 mL/min
flow rate; ambient temperature; gradient elution from 30%
P r od u ct Isola tion fr om Rea ction w ith Acetic Acid .
Triphenylphosphine (1.31 g, 5.0 mmol), acetic acid (0.32 g, 5.3
mmol), and benzene (8 mL) were combined and cooled to 5
°C. Diisopropyl azodicarboxylate (1.01 g, 5.0 mmol) was added
over a 10 min period, keeping the temperature below 8 °C,
during which time the reaction became heterogeneous. The
mixture was then warmed to 22 °C and became homogeneous
within 15 min. After 1 h at 22 °C the solvent was removed in
vacuo and residue dissolved in dichloromethane (3 mL). This
solution was loaded onto 75 g of flash silica, and the products
were chromatographed using 4:1 hexane:EtOAc (350 mL), 3:1
hexane:EtOAc (900 mL), and 2:1 hexane:EtOAc (250 mL).
Products were visualized on TLC using iodine or polymolyb-
MeCN/70% 0.1% aqueous H3
PO4 to 70% MeCN/30% 0.1%
aqueous H3PO4 over 20 min, then to 80% MeCN/20% 0.1%
aqueous H3PO4 over 4 min; elution times, 5 at 3.8 min, benzoic
acid at 4.0 min; triphenylphosphine oxide at 10 min, 7 at 12
min, benzoic anhydride at 16 min, and 6 at 21 min.
P r od u ct Isola tion fr om Rea ction w ith 4-Nitr oben zoic
Acid . Triphenylphosphine (2.62 g, 10.0 mmol), 4-nitrobenzoic
acid (1.70 g, 10.2 mmol), and THF (25 mL) were combined,
and diisopropyl azodicarboxylate (2.02 g, 10.0 mmol) was
added over 10 min at 23-30 °C. The solution was warmed to

Mitsunobu Esterification Reaction
J . Org. Chem., Vol. 61, No. 9, 1996 2971
40 °C for 18 h and then concentrated in vacuo and dissolved
in 5 mL of dichloromethane. This solution was loaded onto
155 g of silica gel and eluted with a step gradient of 1:4 EtOAc:
hexane to 1:2 EtOAc:hexane. Three products were isolated
in the following elution order:
(CDCl3) δ 1.28 (12 H, dd, J ) 3.2, 6.2 Hz), 4.78 (4H, q, J )
16.1 Hz), 5.06 (2H, septet, J ) 6.2 Hz); 13C NMR (CDCl3) δ
21.4, 21.6, 44.6, 73.3, 150.6, 165.2. Anal. Calcd for C12H18
-
Cl2N2O6: C, 40.35; H, 5.08; N, 7.84, Cl, 19.85. Found: C,
40.32; H, 5.17; N, 7.69, Cl, 19.91.
N,N ′-Bis(4-n itr oben zoyl)-N,N ′-bis(isop r op ylca r boxy)-
h yd r a zin e (8) (0.92 g, 19%): mp 150-152 °C (EtOH); 1H
NMR (CDCl3) δ 1.15 (12H, d, J ) 6.5 Hz), 4.96 (2H, septet, J
) 6.6 Hz), 7.89 (4H, d, J ) 8.5 Hz), 8.34 (4H, d, J ) 8.5 Hz);
13C NMR (CDCl3) δ 21.4, 74.0, 123.5, 128.8, 140.2, 149.6, 150.8,
167.3. Anal. Calcd for C22H22N4O10: C, 52.59; H, 4.41; N,
11.15. Found: C, 52.47; H, 4.35; N, 11.14.
N-(Ch lor oa cet yl)-N,N ′-b is(isop r op ylca r b oxy)h yd r a -
zin e (11) (0.48, 17%): mp 59-61 °C (heptane); 1H NMR
(CDCl3) δ 1.27 (6H, br m), 1.33 (6H, d, J ) 6.2 Hz), 4.74 (2H,
s), 4.98 (1H, septet, J ) 6.3 Hz), 5.06 (1H, septet, J ) 6.3 Hz),
6.66 (1H, br s); 13C NMR (CDCl3) δ 21.6, 21.8, 45.3, 70.9, 73.1,
152.4, 154.8, 167.4. Anal. Calcd for C10H17ClN2O5: C, 42.79;
H, 6.10; N, 9.98, Cl, 12.63. Found: C, 43.02; H, 6.17; N, 10.22,
Cl, 12.26.
N,N ′-Bis(isop r op ylca r boxy)h yd r a zin e (5) (1.39 g, 68%).
Spectral analysis was the same as above.
Compound 11 could not be separated from triphenylphos-
phine oxide by HPLC, so the ratios listed in Table 4 were
obtained by flash chromatographic isolation of products of
reactions conducted at ambient temperature.
P r ep a r a t ion of (1S)-(+)-Neom en t h ol Acet a t e. (-)-
Menthol (477 mg, 3.05 mmol), acetic acid (0.99 g, 16.5 mmol),
triphenylphosphine (3.93 g, 15.0 mmol), and toluene (30 mL)
were combined and cooled to -20 °C with stirring. Diisopropyl
azodicarboxylate (3.09 g of 97% purity, 14.8 mmol) was added
over 10 min, keeping the temperature below -10 °C. The
batch became heterogeneous after the addition. The reaction
was held at -14 to -18 °C for 12 h and then at 5 °C for 12 h.
The mixture was then warmed to room temperature and
concentrated to a solid in vacuo. The solid was dissolved in
dichloromethane (10 mL) and loaded onto 205 g of silica gel.
Elution was done with 500 mL each of 5%, 7%, 8%, 10%, and
15% EtOAc in hexanes. (1S)-(+)-Neomenthol acetate (408 mg,
67% yield) was obtained along with 26 mg (5.5%) of unreacted
menthol. The neomenthol acetate identity was confirmed by
comparison of the 1H NMR with an authentic sample.
N-(4-Nitr oben zoyl)-N,N ′-bis(isop r op ylca r boxy)h yd r a -
zin e (9) (1.26 g, 36%): mp 109-111 °C (80:20 EtOH:water);
1H NMR (CDCl3) δ 1.14 (6H, d, J ) 6.1 Hz), 1.31 (6H, d, J )
6.2 Hz), 4.92 (1H, septet, J ) 6.2 Hz), 5.02 (1H, septet, J )
6.2 Hz), 6.86 (1H, br s), 7.81 (2H, d, J ) 7.5 Hz), 8.28 (2H, d,
J ) 8.6 Hz); 13C NMR (CDCl3) δ 21.4, 21.9, 71.1, 73.2, 123.4,
128.6, 141.1, 149.3, 152.3, 155.1, 169.5. Anal. Calcd for C15
-
H19N3O7: C, 50.99; H, 5.42; N, 11.89. Found: C, 51.01; H,
5.36; N, 11.89.
N,N ′-Bis(isop r op ylca r boxy)h yd r a zin e (5) (0.77 g, 38%).
Spectral analysis was the same as above.
HPLC conditions were the same as above for reactions with
benzoic acid; retention times, 4-nitrobenzoic acid at 4.5 min,
9 at 13.5 min, and 8 at 22.6 min.
P r od u ct Isola tion fr om Rea ction w ith Ch lor oa cetic
Acid . Triphenylphosphine (2.62 g, 10.0 mmol), chloroacetic
acid (0.983 g, 10.4 mmol), and THF (25 mL) were combined,
and diisopropyl azodicarboxylate (2.02 g, 10.0 mmol) was
added over 10 min at ambient temperature. The reaction was
warmed to 40 °C for 21 h, concentrated, and chromatographed
on 150 g of silica gel, using a step gradient from 1:7 EtOAc:
hexane to 1:2 EtOAc:hexane. Three products were isolated
in the following elution order:
N,N ′-Bis(ch lor oa cetyl)-N,N ′-bis(isop r op ylca r boxy)h y-
1
d r a zin e (10) (0.18 g, 5%): mp 52-53 °C (heptane); H NMR
J O952180E