A practical synthesis of cycloheptane-1,3-dione

Article abstract of DOI:10.1021/op9802069

A three-step synthesis of cycloheptane-1,3-dione has been developed which avoids the use of heavy metal or explosive reagents and provides access to multigram quantities of this material.

Full text of DOI:10.1021/op9802069

Organic Process Research & Development 1998, 2, 379−381
A Practical Synthesis of Cycloheptane-1,3-dione
John A. Ragan,* Teresa W. Makowski, David J. am Ende, Pamela J. Clifford, Gregory R. Young, Alyson K. Conrad, and
Shane A. Eisenbeis
Process Research and DeVelopment, Pfizer Central Research, Eastern Point Road, Groton, Connecticut 06340
Abstract:
Scheme 1
A three-step synthesis of cycloheptane-1,3-dione has been
developed which avoids the use of heavy metal or explosive
reagents and provides access to multigram quantities of this
material.
We recently required a practical synthesis of cyclohep-
tane-1,3-dione (1) for preparation of multikilogram quantities
of a potential drug candidate. Several literature preparations
of this compound were identified, but none were deemed
suitable for large scale preparation of clinical drug supplies
Scheme 2
1-7
(
e.g., several used heavy metal or potentially explosive
1
3
reagents, such as ethyl diazoacetate, PhHg(CBr
OCH , or Hg(OAc) ). Reported herein is a three-step
3
), ClCH
2
-
5
6
3 2
synthesis of cycloheptane-1,3-dione from cyclopentanone,
which proceeds in 58-64% overall yield and requires no
purification of intermediates.
Discussion
expansion from cyclopentanone (Scheme 2). Our plan was
Several synthetic approaches were considered prior to
identification of the ultimately successful route. For ex-
ample, Wacker oxidation of cyclohepten-3-one with tert-
1
1
to utilize the known [2 + 2] adduct 3 of dichloroketene
and 1-(trimethylsiloxy)cyclopentene, which contains the
requisite seven carbon atoms and has the proper 1,3-oxygen
substitution for cycloheptane-1,3-dione. This strategy has
2
butylhydroperoxide and PdCl was found to be selective for
the desired 1,3-diketone, as anticipated on the basis of a
literature report.8 However, a tert-butyl impurity contami-
nated the product diketone and adversely affected the
subsequent steps in the sequence.
Another route investigated was Noyori’s palladium-
mediated rearrangement of 2,3-epoxycycloheptanone to
cycloheptane-1,3-dione (Scheme 1).4 As per the original
1
2
direct precedent in the work of Pak et al., who reported a
two-step conversion of 3-trimethylsiloxy-2,2-dichlorocy-
clobutanones into 1,3-diketones via Bu
rination to the 3-trimethylsiloxycyclobutanone, followed by
Bu NF-induced desilylation and ring opening to the 1,3-
diketone.
for the strategy, Bu
3
SnH-mediated dechlo-
4
13-15
While this work provides excellent precedent
SnH was clearly unacceptable for our
3
3 4
report, using 4-8 mol % of Pd(Ph P) and 1,2-bis(diphe-
purposes due to both handling toxicity and potential con-
tamination of clinical drug supplies.
nylphosphino)ethane (dpe), complete conversion to the
desired diketone was realized in 12-16 h. We also found
that using BINAP in place of dpe allowed the catalyst load
to be reduced to 2-3 mol % without adversely affecting the
rate of rearrangement.9 Ultimately, however, the cost of
reagents and access to the epoxide were deemed prohibitive
for a multikilogram campaign.
A shorter, more practical synthesis of cycloheptane-1,3-
dione was realized upon consideration of a two carbon ring
Zinc-acetic acid is a well precedented reagent for
1
6
reduction of 2,2-dichloroketones. While numerous ex-
amples of 2,2-dichlorocyclobutanone reductions are found
10
(9) Most experiments were performed with (S)-BINAP, and (R)-BINAP was
equally effective. To our surprise, racemic BINAP led to incomplete
conversion (<20%). Remarkably, using the same bottles of (S)- and (R)-
BINAP which worked well at 3 mol %, an “artificial” racemate from 1.5
mol % each of the (S)- and (R)-enantiomers also failed. These results have
yet to be adequately explained (and are particularly perplexing since the
substrate is racemic and the product diketone is achiral).
*
To whom correspondence should be sent. Phone (860) 441-6334.
(
(
(
1) Eistert, B.; Haupter, F.; Schank, K. Liebigs Ann. Chem. 1963, 665, 55-67.
(10) The epoxide was prepared in three steps from cycloheptanone: (i) Br
THF-H O; (ii) Li CO , LiBr, DMF; (iii) H , NaOH, MeOH.
(11) Krepski, L. R.; Hassner, A. J. Org. Chem. 1978, 43, 3173-3179.
2
,
2) Maclean, I.; Sneeden, R. P. A. Tetrahedron 1965, 21, 31-34.
2
2
3
2 2
O
3) Hutmacher, H.-M.; Kruger, H.; Musso, H. Chem. Ber. 1977, 110, 3118-
3
125.
(12) Pak, C. S.; Kim, S. K.; Lee, H. K. Tetrahedron Lett. 1991, 32, 6011-
(4) Suzuki, M.; Watanabe, A.; Noyori, R. J. Am. Chem. Soc. 1980, 102, 2095-
6014.
2
096.
(13) Pak, C. S.; Kim, S. K. J. Org. Chem. 1990, 55, 1954-1957.
(14) Brady, W. T.; Lloyd, R. M. J. Org. Chem. 1979, 44, 2560-2564.
(
(
(
5) Nishiguchi, I.; Hirashima, T. Chem. Lett. 1981, 551-554.
6) Bhusan, V.; Chandrasekaran, S. Synth. Commun. 1984, 14, 339-345.
7) Vankar, Y. D.; Chaudhuri, N. C.; Rao, C. T. Tetrahedron Lett. 1987, 28,
(15) Footnote 11 in ref 13 states that dechlorination of 3 (Bu
3
SnH) followed by
desilylation with Bu NF does, indeed, form cycloheptane-1,3-dione. No
4
5
51-554.
experimental details are provided.
(16) Noyori, R.; Hayakawa, Y. Org. React. 1983, 29, 163-344.
(
8) Tsuji, J.; Nagashima, H.; Hori, K. Chem. Lett. 1980, 257-260.
1
0.1021/op9802069 CCC: $15.00 © 1998 American Chemical Society and Royal Society of Chemistry
Vol. 2, No. 6, 1998 / Organic Process Research & Development
•
379
Published on Web 10/07/1998

stretch at 1803 cm- (lit.11 1805 cm ), and a neat sample
1
-1
Scheme 3
-
1
3
of diketone 1 absorbed at 1718 and 1698 cm (lit. 1716
and 1693 cm ). We have tentatively assigned the 1787
cm intermediate as monochloroketone 4; this is most
consistent with the GC/MS data (intermediate 5 was not
observed until later in the reaction, by which time the AcOH
had obscured the CdO region) and is consistent with the
-
1
-
1
-
1 22
literature report for a neat sample of ketone 4 (1795 cm )
ketone 5 is reported to absorb at 1777 cm ).
-
1 13
(
On laboratory scale runs of the Zn/AcOH reduction (5-
50 g), moderate exotherms were observed. Concern over
1
the heat flow associated with larger scale reductions (>50
kg) led us to examine the reaction in a Mettler RC1 reaction
calorimeter. This was run on 50 g of dichlorocyclobutanone
in the literature,17-19 to our knowledge there are no such
examples with 3-trimethylsiloxy-2,2-dichlorocyclobutanones.
3
in 270 mL of 1:1 2-propanol-water, with dropwise
addition of 100 mL of 50% aqueous acetic acid over a period
of 4 h. These studies showed an exotherm of 500 kJ/mol
(
There is one example of a 3-alkoxy-2,2-dichlorocyclobu-
20
tanone reduction with Zn/AcOH. In addition, Grieco et
al. have reported the Zn/AcOH reduction of a 2,2-dichloro-
cyclobutanone containing a tert-butyldimethylsilyl ether
several atoms removed;21 the TBS ether is stable to the
reaction conditions.) Thus, we were gratified to find that
direct treatment of dichlorocyclobutanone 3 with zinc
(120 kcal/mol), which translated to a maximum adiabatic
temperature rise of 65-70 °C (this represents a “worst case
scenario”, i.e., if the AcOH were added rapidly, with no
external cooling, the maximum temperature reached from
an initial temperature of 25 °C would be ca. 90-95 °C).
From an operational safety point of view, these data were
viewed as acceptable. (The Zn reduction has been run on
(powdered or -30+100 mesh granules) and acetic acid in
aqueous 2-propanol cleanly provided the desired diketone
1
00 kg scale in the pilot plant without incident.)
1.
The overall yield of crude diketone from cyclopentanone
is 58-64%. This material is a dark orange oil, which is ca.
Conclusion
In summary, we have identified a three-step synthesis of
1
80-90% pure by H NMR and is suitable for subsequent
cycloheptane-1,3-dione, which proceeds in 58-64% overall
yield, requires no purification of intermediates other than
standard aqueous workups, and avoids the use of costly or
toxic reagents. A laboratory scale (50 g) procedure is
provided herein, and the sequence has been used to provide
multikilogram quantities of cycloheptane-1,3-dione in the
pilot plant.
reactions. If desired, the crude product can be purified by
distillation, which provides a colorless liquid in 38% overall
yield from cyclopentanone. In our application, we found
that the subsequent reactions were equally efficient with
crude or distilled diketone; thus, overall yields were sub-
stantially higher when the distillation was omitted.
GC/MS monitoring of the reaction showed several
intermediate products (Scheme 3). The monochloroketone
Experimental Section
4
began to form immediately upon addition of AcOH (and
All chemicals were used as purchased. Dimethylforma-
mide (DMF) and tetrahydrofuran (THF) were purchased in
anhydrous form from Aldrich in Sure-Seal glass bottles; all
other solvents were reagent grade. Reactions were run under
a positive pressure of nitrogen unless otherwise stated. 400
in some cases was observed with just zinc treatment, prior
to addition of any acid). Shortly thereafter (30-60 min),
the fully reduced cyclobutanone 5 began to form. After 2-6
h of reaction, an intermediate GC/MS peak was observed
which was isomeric with the desired diketone. This peak
was tentatively assigned as bicyclic keto-alcohol 6. After
1
13
MHz H (and 100 MHz C) NMR spectra were obtained
on a Varian Unity+400 spectrometer equipped with two rf
channels, indirect detection, and pulsed-field gradients (z-
6-18 h, all intermediates had been converted to the desired
diketone (1).
1
axis only). H spectra were acquired using 45° acquisition
In one experiment, the reaction was monitored with an
in situ IR probe. As the AcOH was added to the reaction,
the CdO absorption of the AcOH began to obscure this
region of the spectrum. Early in the run, however, an
pulses, 3.0 s recycle delay, and 16 scans at a resolution of
0
.2 Hz/point. The acquisition window was typically 6800
13
Hz, and processing used 0.1 Hz line broadening. C spectra
were acquired using 45° acquisition pulses, 0.7 s recycle
delay, and 512 scans at a resolution of 0.8 Hz/point. Proton
decoupling was applied at 3 W of power during acquisition,
and 2 Hz line broadening was used during processing. Mass
spectral data were collected on either a Hewlett-Packard GC/
MS (electron impact ionization) or a Micromass (Fisons)
Platform II mass spectrometer (atmospheric pressure chemi-
cal ionization). Thin-layer chromatography (TLC) was
-
1
intermediate CdO absorption was observed at 1787 cm .
Under the reaction conditions, dichloroketone 3 had a CdO
(17) Honda, T.; Ishikawa, F.; Kanai, K.; Sato, S.; Kato, D.; Tominaga, H.
Heterocycles 1996, 42, 109-112.
18) Molander, G. A.; Carey, J. S. J. Org. Chem. 1995, 60, 4845-4849.
19) Frimer, A. A.; Weiss, J.; Gottlieb, H. E.; Wolk, J. L. J. Org. Chem. 1994,
(
(
5
9, 780-792.
20) Redlich, H.; Lenfers, J. B.; Kopf, J. Angew. Chem., Int. Ed. Engl. 1989,
8, 777-778.
(
(
2
21) Grieco, P. A.; Oguri, T.; Gilman, S. J. Am. Chem. Soc. 1980, 102, 5886-
5
891.
(22) Brady, W. T.; Lloyd, R. M. J. Org. Chem. 1980, 45, 2025-2028.
380
•
Vol. 2, No. 6, 1998 / Organic Process Research & Development

performed on precoated sheets of 60 F254 (Merck Art. 5719).
Melting points are uncorrected.
water in a 2 L, three-neck flask with an overhead stirrer.
Zinc granules (126 g, 1.9 mol, -30+100 mesh) are added
in one portion. After 60 min at room temperature, a mixture
of 130 mL of AcOH plus 260 mL of water is added dropwise
via addition funnel (ca. 4 mL added initially, followed by a
1
-(Trimethylsiloxy)cyclopentene (2). The procedure is
23
a slight modification of that reported previously. A 1 L
round-bottom flask is charged with cyclopentanone (50.7 g,
0
.603 mol) and DMF (250 mL). Triethylamine is added (200
mL, 1.45 mol), followed by dropwise addition of TMSCl
91 mL, 0.72 mol) over 5 min. The solution is then warmed
1
0 min hold to check for exotherms; 20 mL then added,
followed by another 10 min hold; the remaining acid solution
is then added dropwise; the entire addition takes 1.5-2 h).
After 16 h, the mixture is transferred to a separatory funnel,
decanting away from most of the zinc (several small
(
to reflux (90 °C) for 26 h. After cooling to ambient
temperature, the mixture is transferred to a separatory funnel,
rinsing with 500 mL of hexanes. The solution is washed
with water (three portions of 100 mL each) and brine (100
mL) and then concentrated to give 110 g of a dark orange
2-propanol rinses are used). The 2-propanol-AcOH-water
mixture is then extracted with five portions of toluene (250
mL each), which are combined and concentrated to provide
1
oil. H NMR analysis showed the desired product plus 10-
5
1.7 g of product as a dark, orange-brown oil (85% crude
1
5% of triethylamine. This material was used in the next
reaction without further purification.
,7-Dichloro-1-(trimethylsiloxy)bicyclo[3.2.0]heptan-6-
1
mass balance, ca. 85% pure by H NMR). This material is
suitable for subsequent reactions or can be purified by
distillation, which provided 29.4 g (0.23 mol, 48% yield) of
product as a clear, colorless oil (bp 65-75 °C at 1.2 mm).
7
one (3). The procedure is a slight modification of that
reported previously.11 The crude TMS enol ether (2) (0.60
mol) is dissolved in 950 mL of hexanes in a 2 L round-
bottom flask. Triethylamine (100 mL, 0.72 mol) is added,
followed by dichloroacetyl chloride (58 mL, 0.60 mol) as a
solution in 450 mL of hexanes, dropwise over 2 h. The
solution is then stirred at ambient temperature overnight. The
reaction mixture is then filtered through fritted glass, rinsing
with several 50 mL portions of hexane. The clear solution
is concentrated in vacuo to provide 128 g (80% over two
steps) of product as a dark brown oil. This material is
1
The spectral properties ( H NMR, GC/MS) were identical
to those of samples prepared by the oxymercuration route
6
of Bhusan and Chandrasekaran.
Acknowledgment
The authors thank Alex Grodski and Gary Berg for
samples of cycloheptane-1,3-dione prepared by another
6
route, and Drs. Stephane Caron, Frank Urban, Tamim
Braish, and Keith McCarthy for numerous helpful discus-
sions.
1
homogeneous by GC/MS and H NMR (spectra as reported
3
in ref 11), save traces of Et N and DMF, and is used directly
in the next reaction.
Cycloheptane-1,3-dione (1). Dichlorocyclobutanone 3
128 g, 0.48 mol) is dissolved in 520 mL of 1:1 2-propanol-
Received for review December 17, 1997. Accepted August 19,
(
1998.
(23) House, H. J. Org. Chem. 1969, 34, 2324.
OP9802069
Vol. 2, No. 6, 1998 / Organic Process Research & Development
•
381