An improved process for the preparation of trimethylhydrazine and its coupling with an activated acid intermediate

Article abstract of DOI:10.1021/op0342022

Trimethylhydrazine (TMH) was prepared in two steps from 1,1-dimethylhydrazine, using an easy to scale-up procedure that avoided difficult acid-base extractions. The procedure provided TMH as a solution in 1,4-dioxane, in a form that was easy and safe to handle in a coupling with an enantiomerically pure, sterically hindered, Boc-protected-amino acid, 1. This key coupling reaction in the preparation of 3 was accomplished through the corresponding acid chloride, thereby avoiding the use of expensive coupling reagents.

Full text of DOI:10.1021/op0342022

Organic Process Research & Development 2004, 8, 360−362
An Improved Process for the Preparation of Trimethylhydrazine and Its
Coupling with an Activated Acid Intermediate
Silvina Garc´ıa-Rubio,* Chandra D. Wilson, Deborah A. Renner, John O. Rosser, Debasis Patra, J. Gregory Reid, and
Seemon H. Pines
Albany Molecular Research, Inc., Syracuse Research Center, 7001 Performance DriVe,
North Syracuse, New York 13212, U.S.A.
Abstract:
Reaction of 1,1-dimethylhydrazine with ethyl formate was
executed easily on a 5-kg scale according to a procedure
already described4c and yielded 1-formyl-2,2-dimethyl-
hydrazine (4) in 89% yield (Scheme 2).
The reduction of 4 with LiAlH4 was initially done in THF
at room temperature. After an exothermic quench (external
cooling at -35 °C was required to control the heat release)
using ethylene glycol,5 a thick, gummy mixture resulted that
stuck to the stir shaft and further caused its destruction.
Eventually, TMH (5) was obtained as a 5% w/w solution
by co-distillation with THF,6 requiring additional THF to
be periodically added to distill all of the TMH out.
Trimethylhydrazine (TMH) was prepared in two steps from
1,1-dimethylhydrazine, using an easy to scale-up procedure that
avoided difficult acid-base extractions. The procedure provided
TMH as a solution in 1,4-dioxane, in a form that was easy and
safe to handle in a coupling with an enantiomerically pure,
sterically hindered, Boc-protected-amino acid, 1. This key
coupling reaction in the preparation of 3 was accomplished
through the corresponding acid chloride, thereby avoiding the
use of expensive coupling reagents.
In an attempt to obtain a more concentrated solution of
5, 1,4-dioxane was chosen as the reaction solvent. After the
ethylene glycol quench, a gummy mixture resulted again,
and TMH and 1,4-dioxane were co-distilled, resulting in four
fractions, the first two of which contained TMH in higher
concentration (13% w/w). The last fractions collected did
Introduction
Trimethylhydrazine (TMH) is a ubiquitous component in
the preparation of new chemical entities.1 En route to
obtaining multigram quantities of a new drug candidate
possessing growth hormone releasing properties (3), the
necessity for a safe and reliable process for the preparation
of TMH on large scale became evident. Additionally,
coupling TMH with sterically hindered acid 1 using an
inexpensive coupling agent or, better yet, through a reactive
intermediate of 1, such as its acid chloride, was desired
(Scheme 1).
1
not contain product, as monitored by H NMR analysis.
Exothermicity of the ethylene glycol quench of excess
LiAlH4 required the use of a reactor at -35 °C (external
temperature) while the mixture of TMH (plus cosolvent) had
to be distilled at about 100 °C (internal temperature).
Equipment restrictions meant that these operations had to
be done in different reactors. Due to the thick slurry nature
of the resulting mixture and stirring difficulties encountered
during the quench, a transfer to a second container for
distillation was impossible without the potential exposure
of operating personnel to TMH.7 These mechanical issues
were circumvented by quenching the reaction with 1,2-
propylene glycol, which resulted in a slurry that was easy
to stir, both at room temperature and at 100 °C and which
was transferred using vacuum to the distillation reactor. Thus,
exposure of operating personnel to TMH was eliminated.
Scale-up of the reaction to 2 kg was uneventful, and the
mixture of product and 1,4-dioxane was distilled (92-96 °C).
In this way, a less concentrated solution of TMH (7.4% w/w)
was obtained, even when the overall yield remained un-
changed (70%).8 Performance of the TMH solution in 1,4-
Results and Discussion
Methods reported in the literature for the preparation of
TMH include the reduction of methylene dimethylhydrazine
using LiAlH4 or catalytic hydrogenation2,3 as well as the
reduction of N-formyl dimethylhydrazine with LiAlH4.4
Usually, TMH was isolated as the HCl salt using acid-base
extractions.
* To whom correspondence should be addressed. E-mail: silvina.garcia@
albmolecular.com.
(1) (a) Andersen, P.; Ankersen, M.; Jessen, C. U.; Lehman, S. V. Org. Proc.
Res. DeV. 2002, 6, 367. (b) Hanefeld, W.; Von Goesseln, H. J. Arch. Pharm.
(Weinheim, Ger.) 1991, 324, 917. (c) Riebli, P. EP 430033 A2 19910605,
1991. (d) Chong, J. M.; Mar, E. Tetrahedron Lett. 1990, 13, 1981. (e) Walter,
W.; Krische, B.; Voss, J. J. Chem. Res., Synop.1978, 332. (f) Lowrie, H. S.
U.S. Patent 19650506, 1967. (g) Levitt, G. U.S. Patent 19670509, 1967.
(h) Lowrie, H. S. J. Med. Chem. 1966, 9, 664. (i) Sensi, P.; Maggi, N.;
Ballota, R.; Fu¨re`sz, S.; Pallanza, R.; Arioli, V. J. Med. Chem. 1964, 7, 596.
(2) (a) Class, J. B.; Aston, J. G. J. Am. Chem. Soc. 1951, 73, 2359. (b) Class,
J. B.; Aston, J. G.; Oakwood, T. S. J. Am. Chem. Soc. 1953, 75, 2937.
(3) (a) Horvitz, D. U.S. Patent 3,182,087, 1965. (b) Sidi, H.; Paramus, N. J.
U.S. Patent 3,517,064, 1970.
(4) (a) Klages, F.; Nober, G.; Kircher, K.; Bock, M. Justus Liebigs Ann. Chem.
1941, 547, 12, 32. (b) Hinman, R. L. J. Am. Chem. Soc. 1956, 78, 1645. (c)
Beltrami, R. T.; Bisell, E. R. J. Am. Chem. Soc. 1956, 78, 2467. (d)
Andersen, P.; Ankersen, M.; Jessen, C. U.; Lehman, S. V. Org. Process
Res. DeV. 2002, 6, 367.
(5) For a discussion of the complexation on aluminum ions with polydentate
ligands such as ethylene glycol, see: McMahon, C. N.; Alemany, L.;
Callender, R. L.; Bott, S. G.; Barron, A. R. Chem. Mater. 1999, 11, 3181.
Saraswathi, M.; Miller, J. M. Can. J. Chem. 1996, 74, 2221.
(6) The concentration of 5 in THF was determined by integration of the CH3
signal in the 1H NMR spectra.
(7) For information on the toxicity of TMH, see: Toth, B. Cancer 1977, 40,
2427. Nagel, D.; Toth, B.; Kupper, R.; Erickson, J. J. Natl. Cancer Inst.
1976, 57, 187.
360
•
Vol. 8, No. 3, 2004 / Organic Process Research & Development
10.1021/op0342022 CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/26/2004

Scheme 1
Scheme 2
Scheme 3
Table 1. Coupling attempts of TMH with
(R)-N-Boc-3-benzylnipecotic acid
entry coupling agent base solvent conditions yield (%)
1
2
3
4
PyBrop
CDI
EDCI‚HCl
PPACA
DIPEA THF
n/a THF
DMAP CH2Cl2 0 °Cfrt, 2 h
Et3N CH2Cl2 0 °C, 6 h
rt, 48 h
rt, 12 h
65
1
dec
1
Scheme 4
dioxane in the preparation of 2 was comparable to that of
the solution of TMH in THF.
even though its generation would be difficult by the
instability of 1 to Boc-deprotection. Activation of the acid
as its acid chloride (oxalyl chloride, catalytic DMF, Et3N at
-10 °C) and subsequent reaction with a 5% solution of TMH
in THF or 1,4-dioxane afforded the resulting hydrazide,
which was easily purified by passing the crude product
through a silica plug. Yields were generally in the 82-90%
range without any apparent removal of the Boc group.
Interestingly, the order of addition of reagents proved crucial
for the successful formation of the acid chloride in the
coupling. When Et3N was introduced prior to oxalyl chloride
and DMF in the reaction pot, only 1 was recovered as well
as traces of 2 and decomposition products.10 This may be
due to the fact that, while the acid readily reacts with oxalyl
chloride to give the acid chloride which then reacts with
TMH, its Et3N salt reacts differently, leading to decomposi-
tion and, upon quenching, results in the recovery of 1. In
summary, a reproducible method for the reaction of a
sterically hindered acid with a 5-7% w/w solution of TMH
in THF or 1,4-dioxane by preformation of the acid chloride
using oxalyl chloride, DMF, Et3N at -10 °C was developed.
The reaction was further scaled to 3 kg to provide 2 in 90%
yield (93% AUC by HPLC) after silica-plug purification.
With a reliable procedure for the preparation of TMH in
hand, we studied the coupling of TMH with 1. Coupling of
a 5% w/w solution of TMH in THF with 1 was initially done
using PyBrop and DIPEA in THF at room temperature9
(Scheme 3). Isolation of the resulting hydrazide from the
reaction mixture was problematic due to the hygroscopic
nature of the solid that was obtained under these conditions.
Additionally, the DIPEA‚HBr salt that contaminated the
crude product caused problems in the subsequent step of the
synthesis. Eventually, these conditions afforded hydrazide
2 in 65% yield along with recovered starting material (20%)
after purification by column chromatography. Attempts to
crystallize 2 from various solvent systems (CH2Cl2, EtOAc-
hexanes, MTBE) to avoid purification by chromatography
were unsuccessful. Several alternative conditions for the
preparation of 2 using coupling agents were then investigated.
The use of 1-propylphosphonic acid cyclic anhydride (PPACA,
Et3N, 0 °C, slow addition) or 1,1′-carbonyldiimidazole (CDI)
resulted in the recovery of unreacted starting material, while
the coupling mediated by EDCI‚HCl and DMAP resulted in
decomposition products. In accordance with HPLC analyses
that showed the disappearance of starting material, we believe
that in the above cases the reactive intermediate formed but
the subsequent nucleophilic attack of TMH failed. A sum-
mary of the results is depicted in Table 1.
Experimental Section
Reagents that were commercially available were used
without further purification. Nuclear magnetic resonance
spectra were obtained on a Bruker AC 300 spectrometer at
Dissatisfied with the coupling results this far, we decided
to pursue coupling the acid chloride derived from acid 1,
1
300 MHz for H NMR and 75 MHz for 13C NMR. Thin-
(8) The difference of boiling temperatures between TMH (bp 59-61 °C) and
1,4-dioxane (bp 100 °C) would allow obtaining higher concentrations of
TMH provided a fractionating column was used.
(9) The choices of coupling agents were made on the basis of the reported
effectiveness of phosphinic reagents in the coupling of hindered amino acids
and on the fact that PyBrop had been employed in this coupling with
moderate success. See: Humphrey, J. M.; Chamberlin, A. R. Chem. ReV.
1997, 97, 2243.
layer chromatographic (TLC) analyses were performed using
10 cm × 20 cm Analtech Silica Gel GF plates (25 µm thick).
(10) Et3N was introduced initially to avoid unnecessary exposure of the N-Boc
group to acidic conditions. Formation of the acid chloride was monitored
by HPLC analysis by derivatization as its methyl ester (by quenching in a
1:1:8 mixture of MeOH, Et3N, and CH2Cl2).
Vol. 8, No. 3, 2004 / Organic Process Research & Development
•
361

The TLC plate was visualized by UV light (254 nm). HPLC
analyses were performed on an Agilent 1100 liquid chro-
matograph with a variable-wavelength UV detector (YMC
ODS-AQ S 5µ-120 Å, 150 mm × 4.6 mm, 5 µm, eluent:
50% H2O (0.05% v/v TFA), 50% CH3CN, flow 1.0 mL/
min, λ ) 215 nm, column temperature ) rt, injection volume
) 20 µL).
vacuum transferred to a 72-L, unjacketed reactor equipped
with a heating mantle, temperature probe, Syltherm reflux
condenser, Dean-Stark trap, nitrogen sweep, and a mechan-
ical stirrer. The product was distilled from the slurry at 92-
98 °C over 8.5 h. Two main fractions were collected and
assayed by H NMR. The first fraction was 10.5 L (7.4%
TMH w/w, 0.78 kg, 48% yield), and the second was 12 L
and used as is in the following step.
1
Preparation of 1-Formyl-2,2-dimethylhydrazine (4). To
a 50-L jacketed reactor equipped with a temperature probe,
reflux condenser, nitrogen sweep, and an overhead stirrer
was charged 2,2-dimethylhydrazine (5.0 kg, 83 mol) followed
by ethyl formate (5.1 kg, 69 mol). The resulting mixture was
heated at 50 °C for 48 h and then cooled to 20-25 °C,
transferred to a rotavap flask, and concentrated to a residue,
which was crystallized from heptane (15 L). After filtration
and drying the resulting solids under vacuum, 1-formyl-2,2-
dimethylhydrazine4c was obtained (5.4 kg, 89% yield).
Preparation of Trimethylhydrazine (TMH, 5). To a
three-neck round-bottom flask equipped with a mechanical
stirrer under nitrogen was charged LAH (25.8 g, 0.68 mol)
followed by 1,4-dioxane (750 mL). The slurry was stirred
for 10 min, and a solution of 1-formyl-2,2-dimethylhydrazine
(50 g, 0.58 mol) in 1,4-dioxane (500 mL) was added using
an addition funnel while maintaining the internal temperature
below 45 °C. The mixture was then warmed to 50 °C and
stirred at this temperature for 5 h. After cooling to room
temperature, propylene glycol (280 mL) was added over 20
min while maintaining the internal temperature below 40 °C.
The product was then distilled as a solution in 1,4-dioxane,
and four fractions were collected (T of distillate ) 92-96
°C). The first two fractions contained trimethylhydrazine in
Preparation of N-Boc-3-Benzylnipecotic Acid Tri-
methyhydrazide (2). (1) Formation of the Acid Chloride.
To a 72-L unjacketed reactor equipped with a temperature
probe, reflux condenser, nitrogen sweep, cooling bath, and
an overhead stirrer was charged N-Boc-3-benzylnipecotic
acid 1 (3 kg, 9.4 mol) followed by dichloromethane (30 L).
The solution was cooled to -15 to -10 °C, and oxalyl
chloride (1.3 L, 15.2 mol) was added over 15 min while the
internal temperature was maintained below -10 °C. DMF
(300 mL) was then charged to the mixture over 20 min
followed by Et3N (2.66 L, 19 mol) over 1.75 h. The reaction
mixture was stirred at -15 to -10 °C for 4.25 h until the
conversion was complete.10
(2) Coupling Reaction. A solution of trimethylhydrazine
in 1,4-dioxane (1.05 kg in 15 kg, 7% w/w, 14.1 mol) and
Et3N (2.66 L, 19 mol) was added to the reaction mixture
over 2 h using a fluid-metering pump, maintaining the
internal temperature between -15 to -20 °C. The reaction
progress was monitored by HPLC analysis. The mixture was
stirred overnight at -20 °C and then was quenched by the
addition of water (20 L) and extracted with dichloromethane
(15 L). The organic phase was concentrated to a residue,
and after purification using a silica plug (CH2Cl2:MeOH:
NH4OH, 92:7:1), trimethylhydrazide 2 was obtained as an
1
9% w/w according to H NMR integration (275 g, 59%
1
yield). The third fraction contained 3% w/w (188 g, 72%
combined yield). 1H NMR (CDCl3) δ 2.32 (s, 6 H), 2.48 (s,
3 H).
oil (3.2 kg, 90% yield, 93% AUC by HPLC). H NMR
(DMSO-d6) δ 1.30-1.85 (m, 3 H), 1.35 (s, 9 H), 2.47 (s, 6
H), 2.80 (s, 3 H), 2.85-3.30 (masked, 3 H), 2.95 (d, J ) 16
Hz, 1 H), 3.18 (d, J ) 16 Hz, 1 H), 6.97 (m, 2H), 7.15 (m,
3H). 13C NMR (DMSO-d6) δ 21.4, 24.4, 28.3, 31.4, 37.4
(br), 43.6, 43.9, 47.0, 49.6, 78.6, 126.3, 128.4, 129.7, 138.3,
154.7, 174.4.
Scale-Up of the Preparation of 5. To a 100-L jacketed
reactor equipped with a temperature probe, reflux condenser,
nitrogen sweep, and an overhead stirrer was charged 97%
LAH in premeasured bags (1 kg, 26 mol) followed quickly
by 1,4-dioxane (30 L) via a stainless steel nitrogen pressure
can. A solution of N-formyl-dimethylhydrazine (2 kg, 23
mol) in 1,4-dioxane (20 L) was added to the slurry over 100
min while maintaining the internal temperature below 50 °C.
The mixture was heated at 50 °C for 1.5 h. The reaction
progress was monitored by TLC using 9:1:0.1 CH2Cl2:
MeOH:28% aq NH4OH with visualization by iodine stain.
(Rf of N-formyl-dimethylhydrazine: 0.5, Rf of trimethyl-
hydrazine: 0.4). After cooling the mixture to 20-25 °C,
propylene glycol (11.2 L, 153 mol) was charged over 65
min while maintaining the internal temperature below 45 °C.
The slurry was cooled to 20-25 °C overnight and then
Acknowledgment
We thank Jayachandra P. Reddy for his valuable insight,
his support and encouragement during the execution of this
work, as well as useful and critical discussions. This work
would not have been possible without the hard work and
dedication of Carl Slater and the Kilo-Lab staff at the
Syracuse Research Center.
Received for review December 30, 2003. Accepted March 6,
2004.
OP0342022
362
•
Vol. 8, No. 3, 2004 / Organic Process Research & Development