Dual activation in the esterification of hindered alcohols with anhydrides using MgBr2 and a tertiary amine

Full text of DOI:10.1021/jo9609485

5
702
J . Org. Chem. 1996, 61, 5702-5703
Communications
Du a l Activa tion in th e Ester ifica tion of
Hin d er ed Alcoh ols w ith An h yd r id es Usin g
MgBr a n d a Ter tia r y Am in e
2
ester (geranyl butyrate) or the corresponding bromide
was found (2% would have been detected). If all of the
reagents were mixed and the alcohol was added subse-
quently (method B), decomposition of the anhydride
increased as the degree of substitution at the R-carbon
decreased. Acetic anhydride was most difficult and
progressively better results were obtained with propionic,
butyric, and isobutyric anhydrides. Method B was
satisfactory for benzoylations and pivaloylations.
Edwin Vedejs* and Olafs Daugulis
Chemistry Department, University of Wisconsin,
Madison, Wisconsin 53706
Received May 22, 1996
The dual activation method works well with phthalic
anhydride and 1-methylcyclohexanol (5; Table 1, entry
2) although the reaction is significantly slower than
analogous benzoylations. Succinoylation with 1 as the
substrate did not proceed to completion (ca. 40% conver-
In the course of studies designed to evaluate reactive
acylating agents, we observed that the combination of 2
equiv of MgBr (anhydrous solution in THF), 3 equiv of
2
a tertiary amine, and 2 equiv of benzoic anhydride
benzoylated the secondary alcohol 1-phenylethanol within
5
sion to the hemisuccinate ester), even with a large excess
seconds (CH
2
Cl
2
, exotherm!) Secondary alcohol acyla-
of succinic anhydride and extended reaction times (3
days). Cyclic anhydrides have been activated using the
DMAP catalyst, but the reactions are slow compared to
acyclic anhydrides.6 Steglich and H o¨ ffle have reported
that the hindered pivalic anhydride is not activated by
DMAP.7 We have confirmed this generalization using
the tertiary alcohol 1 as the substrate (2 equiv of [t-BuC-
1
tions were not pursued, but we were interested to find
that even the hindered pivalic anhydride reacts with
alcohols using a similar “dual activation” procedure
,7
2
(tertiary amine + MgBr
2
). The activated pivalic anhy-
dride reagent converted the sensitive tertiary alcohol 1
3
into the pivalate ester 2 in 99% yield after 22 h at room
temperature. The corresponding benzoylation (using
benzoic anhydride) was much faster and gave 95% of the
benzoate ester 3 within 15 min! No elimination to â,â-
dimethylstyrene was detected in either reaction.
(O)]
2
O + 3 equiv of Et
3
N + 2 equiv of DMAP; <1%
conversion to 2 after 19 h at rt in CH
2
Cl ). Under similar
2
3
conditions with DMAP, the benzoic anhydride affords
24% of the benzoate ester 3. By comparison, the dual
activation procedure is roughly 2 orders of magnitude
faster. A comparison with the recently reported Sc(OTf)
3
method8 was also made. An experiment with pivalic
anhydride and 1 at 0 °C (5 equiv anhydride, 5 mol%
scandium triflate, CH CN) gave significant elimination
3
to â,â-dimethylstyrene as well as the pivalate ester 2,
but clean conversion to 2 was observed at -20 °C. The
reaction was complete within 4 h, substantially faster
than the dual activation procedure at room temperature.
Several other examples were studied using alcohol-
anhydride combinations that allow comparison with
literature results for the preparation of the corresponding
hindered esters (Table 1). Some optimization proved
necessary in the case of anhydrides that contain enoliz-
able R-hydrogens to minimize decomposition of the
3
On the other hand, the Sc(OTf) -catalyzed benzoylation
of 1 proved to be much slower. Conversion was too low
to assay after 20 h at -20 °C and elimination ac-
companied benzoylation at 0 °C (22% benzoate 3, 4% â,â-
dimethylstyrene after 20 h at -20 °C and 11 h at 0 °C).
The lower reactivity of aromatic anhydrides in the
scandium triflate reactions was noted by Yamamoto et
al. and was exploited for selective acylation using mixed
acyl-aromatic anhydrides.8 In contrast, dual activation
appears to favor transfer of the aroyl subunit. Thus, the
anhydride. This problem was encountered in attempts
to prepare linalyl butyrate.4 It was found that hindered
amines are significantly better than triethylamine and
that the order of mixing the reactants is important in
the enolizable anhydrides. Thus, it was necessary to mix
linalool (4) with MgBr and the amine p r ior to addition
2
of the anhydride (method A) to obtain a good yield.
Better results were obtained with 1,2,2,6,6-pentameth-
mixed pivalic, p-nitrobenzoic anhydride was reacted with
9
the hindered alcohol 7 (2 equiv of MgBr
2
, 3 equiv of Et
3
N,
CH
2 2
Cl , 19 h at rt), resulting in the p-nitrobenzoate ester
ylpiperidine (PMP) than with i-Pr
using 2 equiv of butyric anhydride, 1.5-2 equiv of amine,
and 1.2-1.5 equiv of MgBr . No trace of the primary
2
NEt or triethylamine
8
(80% isolated) and no pivalate by NMR assay (<5%).
The scandium triflate selectivity pattern corresponds to
the greater stability of alkyl-substituted vs aryl-substi-
tuted acylium ions,10 although it is not known whether
an acylium ion mechanism is involved.
2
(
1) Steglich, W.; H o¨ fle, G. Angew. Chem., Int. Ed. Engl. 1969, 8,
81. H o¨ fle, G.; Steglich, W.; Vorbr u¨ ggen, H. Angew Chem., Int. Ed.
Engl. 1978, 17, 569.
2) (a) Evans, D. A.; Vogel, E.; Nelson, J . V. J . Am. Chem. Soc. 1979,
01, 6120. Masamune, S.; Mori, S.; Van Horn, D.; Brooks, D. W.
9
(
1
(5) Kita, Y.; Maeda, H.; Takahashi, F.; Fukui, S. J . Chem. Soc.,
Chem. Commun. 1993, 410.
Tetrahedron Lett. 1979, 20, 1665. Blanchette, M. A.; Choy, W.; Davis,
J . T.; Essenfeld, A. P.; Masamune, S.; Roush, W. R.; Sakai, T.
Tetrahedron Lett. 1984, 25, 2183. Rathke, M. W.; Cowan, P. J . J . Org.
Chem. 1985, 50, 2622. (b) Alcohol activation using PMP/MgBr : Evans,
2
D. A.; Anderson, J . C.; Taylor, M. K. Tetrahedron Lett. 1993, 34, 5563.
Vedejs, E.; Chen, X. J . Am. Chem. Soc. 1996, 118, 1809. DABCO/
(6) (a) Theisen, P. D.; Heathcock, C. H. J . Org. Chem. 1988, 53, 2374.
Yano, S,; Kato, M.; Tsukahara, K.; Sato, M.; Shibahara, T.; Lee, K.;
Sugihara, Y.; Iida, M.; Goto, K. Inorg. Chem. 1994, 33, 5030. Shimizu,
T.; Kobayashi, R.; Ohmori, H.; Nakata, T. Synlett. 1995, 650. (b)
Kluger, R.; Hunt, J . C. J . Am. Chem. Soc. 1989, 111, 3325.
(7) H o¨ fle, G.; Steglich. W. Synthesis 1972, 619.
(8) Ishihara, K.; Kubota, M.; Kunhara, H.; Yamamoto, H. J . Am.
Chem. Soc. 1995, 117, 4413. Ishihara, K.; Kubota, M.; Yamamoto, H.
Synlett 1996, 265.
MgBr
findings prior to publication.
3) Kita, Y.; Maeda, H.; Omori, K.; Okuno, T.; Tamura, Y. J . Chem.
Soc., Perkin Trans. 1 1993, 2999.
4) Katsuragi, H. Chem. Abstr. 1952, 46, 7288b; Chem. Abstr. 1957,
1, 16517f.
2
: Sibi et al. We thank Prof. M. Sibi for informing us of these
(
(
(9) Krief, A.; Clarembeau, M.; Barbeaux, P. J . Chem. Soc., Chem.
Commun. 1986, 457.
5
S0022-3263(96)00948-6 CCC: $12.00 © 1996 American Chemical Society

Communications
J . Org. Chem., Vol. 61, No. 17, 1996 5703
Ta ble 1. Ester ifica tion of Alcoh ols (1 equ iv) w ith Acid An h yd r id es (2 equ iv) + Am in e + MgBr 2a
entry
ROH
[R′CO]2O
amine (equiv)
MgBr2, equiv
time
yield
lit. yield (%)
85b
ref
1
2
3
4
5
6
7
8
1
1
6
4
4
4
5
7
(PhCO)2O
Et3N (3)
Et3N (3)
Et3N (3)
Et3N (2)
i-Pr2EtN (2)
PMP (1.5)
Et3N (3)
Et3N (3)
2, method B
2, method B
2, method A
1.5, method A
1.5, method A
1.2, method A
2, method B
2, method B
15 min
25 h
95
99
99
41
83
91
90
80
3
3
16
4
4
4
b
(t-BuCO)2O
(i-PrCO)2O
(n-PrCO)2O
(n-PrCO)2O
(n-PrCO)2O
C6H4(CO)2O
PivOCOArf
80
60
c
25 min
30 min
30 min
30 min
3 h
d
d
d
85
46
e
17
18
g
19 h
a
CH2Cl2 at rt. Method A: anhydride added last. Method B: alcohol added last. Method B has a small advantage in convenience in
cases where the anhydride cannot decompose via enolization. From the 1-ethoxyvinyl ester + 1. c From the 1,1′-dimethylstannocene
derivative of 6 + isobutyroyl chloride, 100 °C. From butyric anhydride + H3PO4, no yield reported. From the sodium alkoxide +
phthalic anhydride. f Pivalic p-nitrobenzoic anhydride, generated in situ. From the lithium alkoxide + p-nitrobenzoyl chloride.
b
d
e
g
conjecture could be found. Thus, it was observed that
2
,6-dimethylpyridine is far more effective in the benzoy-
lation of 1 (60% conversion in 2 h) than are other pyridine
derivatives (percent conversion, time): pyridine (6%, 23
h), 2-picoline (32%, 19 h), 2-methylquinoline (54%, 19 h),
2
,6-di-tert-butylpyridine (<3%, 24 h), p-(dimethylamino)-
pyridine (DMAP; <2%, 25 h). A sufficient explanation
for these trends is that a hindered base is needed to
promote conversion to the magnesium alkoxide 9 and
that no other role for the amine is required. The DMAP
experiment can be understood if the highly nucleophilic
To limit the number of plausible mechanistic possibili-
ties, several control experiments were performed. When
1
1
was treated with MgBr
2
and i-Pr
2
NEt, the H NMR
amine binds irreversibly to MgBr
2
under these conditions.
The other substituted pyridines are effective activating
signals broadened and a small change in chemical shift
was detected for the benzylic CH group (CD Cl , δ 2.75
2
2
2
agents roughly in the order of their basicity as estimated
ppm for pure 1, 2.86 ppm for the mixture). In addition,
major new signals for the protonated amine appeared (i-
13
a
from pK values of the pyridinium salts. In contrast to
DMAP, strongly basic, hindered amines such as i-Pr
do not deactivate MgBr
The MgBr anhydride complex 10 (or a chelated
2
NEt
Pr
2
NH-(+)-CH
2
CH
3
, 3.15 ppm) at the expense of the
NCH CH , 2.45 ppm), but minor
2
.
neutral amine (i-Pr
2
2
3
2
absorptions of the latter were still present. This evidence
proves that proton transfer occurs from alcohol oxygen
to nitrogen and that the potential equilibrium is slow on
the NMR time scale. Evidently, formation of the Mg-O
linkage pays the thermodynamic price for this conversion
equivalent) is probably involved in the anhydride activa-
tion sequence, and interconversion with the magnesium
carboxylate and an alkanoyl or aroyl bromide appears
likely on the basis of the data presented above. Conver-
sion to the carboxylate oxygen-magnesium bond may
in spite of the expected pK
a
disadvantage. Structure 9
14
provide the driving force for this process. In view of
is drawn for convenience, but the actual structure may
be more complicated.
the large number of potential variables, no further
mechanistic interpretation will be offered at this time.
However, it is clear that both the alcohol and the
anhydride are activated by the combination of MgBr and
2
amine and that the combination of factors produces
In another set of control experiments, the MgBr
2
/THF
complex was mixed with benzoic anhydride at rt and the
C spectrum was recorded. In addition to a signal at δ
63.4 ppm for the anhydride carbonyls, a smaller signal
was seen at 165.9 ppm. When this mixture was rapidly
filtered through silica gel, evaporated, and redissolved
1
3
1
remarkable rate accelerations in the anhydride-alcohol
15
reactions.
in CD
2
Cl
2
, signals assigned to benzoic anhydride (163.6
Ack n ow led gm en t . This work was suppported by
ppm, major), benzoyl bromide (166.1 ppm, minor), and
benzoic acid (167 ppm, minor) were present and the
characteristic odor of benzoyl bromide was apparent. The
NMR assignments were supported by comparing the
the National Science Foundation.
Su p p or tin g In for m a tion Ava ila ble: Representative ex-
perimental procedures and NMR spectra of ester products (9
pages).
spectrum in CDCl
zoyl bromide 165.7 ppm) with published values.
On the basis of the above information, it is likely that
3
(benzoic anhydride, 162.3 ppm; ben-
1
1
J O9609485
1
2
a magnesium alkoxide such as 9 is the nucleophile. The
identity of the electrophile is a more difficult problem.
Benzoyl bromide is present in the reaction mixture, and
its direct reaction with the magnesium alkoxide 9 is
plausible. Some role for an N-acylammonium complex
is also conceivable, but no evidence to support this
(13) Berry, D.; Bukka, K.; Satchell, R. S. J . Chem. Soc., Perkin
Trans. 2 1976, 89. Benoit, R. L.; Fr e´ chette, M.; Lefebvre, D. Can. J .
Chem. 1988, 66, 1159. Pawlak, Z.; Kuna, S.; Richert, M.; Giersz, E.;
Wisniewska, M.; Liwo, A.; Chmurzynski, L. J . Chem. Thermodynam,
1
991, 23, 135.
(14) Beryllium chloride reacts with certain acid anhydrides to afford
the beryllium carboxylate and an acyl chloride: Feild, G. B. J . Am.
Chem. Soc. 1939, 61, 1817.
(
15) (a) Hindered benzoates can be made using benzoyl triflate:
(
10) Larsen, J . W.; Bouis, P. A.; Riddle, C. A. J . Org. Chem. 1980,
Brown, L.; Koreeda, M. J . Org. Chem. 1984, 49, 3875. (b) Benzoylation
of methyl 4,6-O-benzylidene-â-D-glucopyranoside using method B was
nonselective and gave mainly the 2,3-di-O-benzoate. (c) Reaction of 1
4
1
2
5, 4969.
(11) Sadtler Standard Spectra; 1978, 24, 4671C; 1977, 9, 1624C;
988, 123, 24418C.
(0.1 mmol), Ph
(0.1 mmol) gave the phosphinate ester (80% yield, 20 min rt), but TsCl
in place of Ph P(O)Cl afforded only recovered 1.
(16) Mukaiyama, T.; Ichikawa, J .; Asami, M. Chem. Lett. 1983, 293.
(17) Rutherford, K. G.; Prokipcak, J . M.; Fung, D. P. C. J . Org.
Chem. 1963, 28, 582.
2 2 2
P(O)Cl (0.05 mmol), EtN-i-Pr (0.15 mmol), and MgBr
(12) (a) A similar conclusion was reached by Evans et al.; see ref
b. (b) Treatment of 1 with MgBr without the amine also broadens
2
2
1
the H NMR signals and produces a small chemical shift change
benzylic CH , 2.81 ppm), presumably because of complex formation.
Reaction of 1 with CH MgBr + added MgBr (1 equiv) results in a
CH signal at 2.82 ppm.
(
2
3
2
2
(18) Kuo, M.-Y.; Liu, K.-T. J . Org. Chem. 1987, 52, 2927.