Process chemistry related to the experimental rice herbicide 2,2-dimethyl-1-(4-methylthio-5-pyrimidinyl)indane

Article abstract of DOI:10.1021/op000033z

Two concise syntheses of the experimental rice herbicide 2,2-dimethyl-1-(4-methylthio-5-pyrimidinyl)indane are reported. The initial synthesis relies on a low-temperature addition of 5-lithio-4-methylthiopyrimidine to 2,2-dimethyl-1-indanone to construct the pyrimidinylindane system. Process improvements to this route are described and resulted in the preparation of 90 kg of the title compound on pilot plant scale. Economics dictated the need to identify a new synthetic route which utilized inexpensive raw materials. Detailed herein is the initial discovery of a new route which features a novel combination of dissolving metal reduction/formylation/cyclization to construct the requisite pyrimidine ring. Process improvements to this chemistry have allowed us to deliver an appropriately substituted pyrimidinylindane in a minimal number of synthetic operations.

Full text of DOI:10.1021/op000033z

Organic Process Research & Development 2000, 4, 275−285
Process Chemistry Related to the Experimental Rice Herbicide
2
,2-Dimethyl-1-(4-methylthio-5-pyrimidinyl)indane
Thomas J. Dietsche, David B. Gorman, Jon A. Orvik, Gary A. Roth,* and William R. Shiang
Dow AgroSciences, Global Process Research & DeVelopment, 1710 Building, Midland, Michigan 48674, U.S.A.
Scheme 1 a,b
Abstract:
Two concise syntheses of the experimental rice herbicide 2,2-
dimethyl-1-(4-methylthio-5-pyrimidinyl)indane are reported.
The initial synthesis relies on a low-temperature addition of
5
-lithio-4-methylthiopyrimidine to 2,2-dimethyl-1-indanone to
construct the pyrimidinylindane system. Process improvements
to this route are described and resulted in the preparation of
9
0 kg of the title compound on pilot plant scale. Economics
dictated the need to identify a new synthetic route which utilized
inexpensive raw materials. Detailed herein is the initial discov-
ery of a new route which features a novel combination of
dissolving metal reduction/formylation/cyclization to construct
the requisite pyrimidine ring. Process improvements to this
chemistry have allowed us to deliver an appropriately substi-
tuted pyrimidinylindane in a minimal number of synthetic
operations.
a
Conditions: (a) H
, CH CN, IPA; (c) powdered KOH, CH
hexanes; (e) 6, THF; (f) BF ‚OEt , CH Cl , Et SiH. Conditions: (a) H
peracetic acid, water; (b) SOCl , cat. DMF, toluene chase; thiourea, ethanol;
KOH, water; (CH SO ; (c) KOH, CH I, NMP, water; (d) nBuLi, toluene,
hexanes; (e) 6, toluene; (f) CF CO H, NaBH , CH Cl
2
SO
4
, peracetic acid, acetone; (b) SOCl
2
, cat. DMF then
NaSCH
3
3
3
I, DMSO: (d) nBuLi, THF,
Introduction
b
2 4
SO ,
In 1994, Markley and co-workers at Dow AgroSciences,1
discovered a new series of compounds which exhibited
herbicidal activity against a variety of weed species in paddy
rice. One of the more active and selective members of the
series was 2,2-dimethyl-1-(4-methylthio-5-pyrimidinyl)in-
dane (1, Scheme 1). To conduct field, toxicology, and
formulation studies, kilogram quantities of 1 were required.
Herein, we describe: (1) our initial laboratory scale synthesis
of 1 which resulted in the preparation of 2.8 kg of material,
3
2
2
2
3
2
3
)
2
4
3
3
2
4
2
2
.
disubstituted pyrimidine 4. Indanone 6 was prepared by
geminal dialkylation of 1-indanone (5). Metal halogen
exchange of 4 at -85 °C, followed by introduction of
electrophile 6 yielded tertiary alcohol 7. Lewis acid mediated
3
silane reduction furnished the desired product 1. Preliminary
investigation of this chemistry in our laboratories identified
two major scale-up issues. The first is a reactive chemicals
issue for the literature peracid hydroxylation of 5-bromopy-
(2) pilot plant scale-up to produce 90 kg of 1, and (3)
discovery and process development of a new economically
and operationally attractive synthetic route to 1.
2
rimidine (2). We found that a significant amount of the
explosive acetone peroxide is formed when acetone is used
as the solvent. Thus, a totally aqueous hydroxylation system
4
Results and Discussion
Initial Laboratory Synthesis of 1 (Scheme 1, Condi-
tions a). At the onset of our studies, expeditious delivery of
kilogram quantities of 1 was required to meet a rapid
development strategy. Our laboratory synthesis, depicted in
Scheme 1 (conditions a), was very similar to that used
initially by our Dow AgroSciences colleagues in Discovery
Research. Hydroxylation of 5-bromopyrimidine via a litera-
was developed which provided comparable yields to that
reported in the literature. The second scale-up issue was the
moderate yield ( ∼55%) of the low-temperature organome-
tallic step, which complicated purification of alcohol 7. Upon
manipulation of several reaction parameters, improved results
were obtained when a THF solution of 4 was added dropwise
to n-butyllithium/hexane/THF below -85 °C. Following
introduction of ketone 6, workup and digestion with ethanol,
the product, 7, was obtained in 74% yield (98+% pure). For
the conversion of 5 to 6, we found that powdered potassium
hydroxide in DMSO could be used for the alkylation.
2
ture procedure provided pyrimidone 3 which was chlorinated
and then substituted with thiomethoxide, providing the 4,5-
*
To whom correspondence should be addressed. Telephone: (517) 638-6132.
Fax: (517) 638-6619. E-mail: garoth@dow.com.
(
1) Markley, L. D.; Arndt, K. E.; Ray, P. G.; Balko, T. W.; Cressman, E. N.
K.; Ouse, D. G.; Jackson, J. L.; Secor, J. 4-Substituted 5-polycyclylpyri-
midine herbicides. U.S. Patent 5,780,465, 1998, 36 pp. CODEN: USXXAM
US 5780465 A 19980714 CAN 129:105503.
(3) Adlington, M. G.; Orfanopoulos, M.; Fry, J. F. Tetrahedron Lett. 1976,
2955.
(4) Dietsche, T. J.; Powers, B. Explosion in preparation of 5-bromo-4(3H)-
pyrimidinone. Chem. Eng. News 1995, 73, 4 (July 10, 1995).
(2) Kress, T. J. J. Org. Chem. 1985, 50, 3073.
1
0.1021/op000033z CCC: $19.00 © 2000 American Chemical Society and The Royal Society of Chemistry
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
275
Published on Web 06/20/2000

Precipitation of the product by addition of water led to simple
isolation of pure 6. The preparation of 1 using these
modifications is detailed in the Experimental Section.
Process Improvements and Pilot Plant Scale-Up
resulting potassium thiolate was S-alkylated with dimethyl-
sulfate. Work up of the reaction consisted of dilution with
water and extraction with toluene. The organic phase was
washed with brine, water and then dried azeotropically to
provide a toluene solution of 4 (15-28 wt. % via HPLC
analysis). Yields for this procedure were 75-85% based on
the initial charge of 5-bromo-4-hydroxypyrimidine (3). One
major impurity (2-3 area % by GC) is produced and has
been identified as an isomer of 4 (GC/MS). The most likely
structure of this impurity is the alkylation regioisomer,
N-methylthione. This chemistry was conducted in three
batches, utilizing two 300 gallon reaction vessels and
provided 165 kg (81%, 97+% purity) of 4.
The process to prepare the 2,2-dimethyl-1-indanone (6)
for the 2.8 kg campaign involved the use of DMSO,
powdered KOH, and methyl iodide. For scale-up to a pilot
plant, both powdered KOH and DMSO create potential
problems. Although the reaction is carried out below room
temperature, the possibility for the known decomposition of
DMSO at higher temperatures is always present due to the
exothermic nature of the alkylation chemistry. Indeed, an
adiabatic temperature rise of about 140 °C is possible, based
on a calculated heat of reaction of -65 to -77 kcal/mol for
the formation of 2,2-dimethyl-1-indanone (6). Powdered
KOH, although easy to prepare in the laboratory from pellets
using a blender, would be difficult and time-consuming in
the plant. With a minimal amount of time and research,
solutions were found for both concerns (Scheme 1, conditions
b). First, N-methylpyrrolidinone proved to be an excellent
choice for a solvent. The reaction yields were the same as
for DMSO. Second, the use of 45% aqueous KOH turned
out to be better than the use of the powdered form from a
standpoint of ease of addition. Instead of adding solid
powdered KOH into a mixture of the indanone 5 and methyl
iodide in the solvent, the 45% solution could be added
dropwise. One difficulty encountered with 45% KOH was
that there were two phases in the pot until 35-50% of the
base was added. When complete solution was achieved, an
exotherm ensued, taking the batch from 20 to 60-70 °C,
indicating a rapid reaction. By changing the order of addition
(Scheme 1, Conditions b). Although the synthetic route
described above can be used to conveniently prepare
kilogram quantities of 1 in the laboratory, several of the steps
are not amenable to scale-up in a pilot plant facility.
Improvements that address these issues will now be de-
scribed. The peracid oxidation of 2 was found to go to
completion much faster at 70-75 °C with no affect on yield
or product quality. The resulting pyrimidone (3) was isolated
by centrifugation and then suspended in toluene. Azeotropic
removal of the water followed by centrifugation provided a
toluene wet cake of 3 which was used directly in the
subsequent chlorination. This oxidation was conducted in
three batches utilizing a 300 gallon reactor and resulted in
an overall yield of 197 kg (65%) of 3.
The previous method for the conversion of 3 to 4 is
outlined in Scheme 1 (conditions a). Chlorination of 3 using
thionyl chloride as solvent (DMF catalysis) followed by
removal of the excess thionyl chloride via distillation
provided the crude 5-bromo-4-chloropyrimidine as a brown
oil. Treatment with sodium or potassium thiomethoxide in
2-propanol/acetonitrile followed by dilution of the reaction
mixture with water afforded an aqueous slurry of the desired
product. Filtration, followed by drying the material to a
constant weight, provided 80-85% yields of methylthiopy-
rimidine 4. The chemistry described works well, but this
process has two major drawbacks when considering pilot
plant operations. The first is the need to dry large quantities
of 4 as a solid, and the second is the odor problem associated
with the use of thiomethoxide. The first problem was easily
circumvented by extractive removal (toluene) of 4 from an
aqueous mixture, followed by azeotropic drying of the
extract. A solution to the odor problem presented itself when
a literature search revealed an alternate preparation of 4 using
5
odorless thiourea as the source of sulfur. Utilizing this
chemistry, we devised a two-pot method for the thiomethyl-
ation of 3. The chlorination reaction was run as previously
described. After distillation of the bulk of the thionyl
chloride, residual thionyl chloride was removed using a
toluene chase. The resulting toluene solution of the chloro-
pyrimidine was added to thiourea in ethanol, causing rapid
displacement of the heterocyclic chloride, yielding the
(adding the NMP solution to the KOH), the reaction started
immediately, and the exotherm could easily be controlled.
Finally, the reaction outcome seemed to be insensitive to
variation in temperature. Whether the pot temperature was
kept below 20 °C or allowed to warm to 60-70 °C, yields
and purity did not change significantly. A temperature of
Scheme 2 a
2
5 °C was chosen for the pilot plant conditions. The
alkylation of 1-indanone (5) was conducted in seven batches,
utilizing a 100 gallon reactor, thus yielding a total of 119
kg (81%, 97+% pure) of 6.
The metal halogen exchange of 4 and nucleophilic
addition to ketone 6 was carried out as previously described
with two process modifications (Scheme 1, conditions b).
First, to improve process flow, the previously dried toluene
solution of heteroarylbromide 4 (15-28%) was added to the
n-butyllithium\THF\hexane mixture at -85 °C, thus affecting
metal halogen exchange. After additon of ketone 6 and
warming the mixture to -20 °C, the reaction was quenched
a
Conditions: (a) SOCl
2
, cat. DMF; toluene chase; thiourea, ethanol; (b) KOH,
water then (CH
3 4
)SO .
corresponding thiouronium salt (Scheme 2). Addition of
aqueous potassium hydroxide effected hydrolysis, and the
(
5) Brown, D. J.; Foster, R. V. Aust. J. Chem. 1966, 19, 2321-30.
2
76 Vol. 4, No. 4, 2000 / Organic Process Research & Development
•

with water. Most of the hexane and THF were removed via
simple distillation, followed by dilution of the heterogeneous
mixture with water. The toluene was removed by azeotropic
distillation with water and the resulting warm aqueous slurry
was diluted with an equal volume of ethanol. After cooling
to ambient temperature, the solid product (7) was collected
by filtration, washed with water, and dried. Utilizing this
modified procedure, laboratory yields of 73-82% were
obtained on a one mole scale (five-liter equipment). Five
batches were scaled to a 200 gallon cryogenic reactor and
resulted in production of 107 kg (57% yield, 99+% purity)
of tertiary alcohol 7.
Scheme 3
In our previous synthesis, we described the conversion
of tertiary alcohol 7 to 1 using boron trifluoride etherate and
triethysilane as the reducing medium. Although these condi-
tions afforded good yield ( ∼80%) and were practical on
laboratory scale, the price of triethylsilane made scale-up
cost prohibitive. Thus, we developed a new lower-cost
procedure which involves addition of trifluoroacetic acid to
a 0.5 M solution containing 7 in methylene chloride with
6
suspended sodium borohydride pellets which gave clean
Our initial laboratory investigations focused on identifica-
conversion to 1 at about 0 °C (Scheme 1, conditions b). After
an aqueous extraction, followed by an aqueous extraction/
neutralization using sodium hydroxide, the organic layer was
concentrated, and methylene chloride was exchanged with
methanol. Following crystallization from methanol, water
was added to the slurry to complete the crystallization. The
product was collected by filtration, washed with water, and
dried. On the basis of a product assay of 98.1%, a 94% yield
of 1 was obtained on a 0.46 mol scale. Scale-up to a 100
gallon reactor afforded 91 kg (88%, 99+% purity) of 1 in
three runs.
Discovery of a New Synthetic Route to 1. During the
course of our initial sample campaigns, it became evident
that we could not meet the required long-term cost of
manufacture utilizing the chemistry described above. It
became clear to us that a cost-effective synthesis of 1 must
rely on construction of the requisite 4,5-disubstituted py-
rimidine from inexpensive raw materials in a minimal
number of unit operations. Since 2,2-dimethyl-1-indanone
tion of carbon nucleophiles that would efficiently undergo
addition to the carbonyl of 6. It was anticipated that steric
hindrance in conjunction with conjugation of the carbonyl9
would cause 6 to be reactive with only the most efficient
nucleophiles. This indeed proved to be the case. Attempted
reaction of 6 with Meldrum’s acid, diethyl malonate, or
malononitrile under standard or modified Knoevenagel
1
0
conditions led to extremely low yields (<10%) of impure
products. However, reaction of 6 with strong (e.g., conjugate
acid pK
a
’s of ∼25), sterically nondemanding carbanions
afforded the desired addition products, which underwent
dehydration upon treatment with acid (Scheme 4). For
Scheme 4 a
(
6) had the potential of supplying the indane backbone of 1
7
at reasonable cost, our initial studies focused on this as the
key starting material. Our strategy, depicted in Scheme 3,
was to use the carbonyl of 6 as an electrophile in combination
with a heteroatom-stabilized carbanion (8) to construct the
carbon-carbon bond at what would eventually be the
a
2
Conditions: (a) THF, Hexane, -70 °C; (b) p-TsOH, toluene, -H O.
5
-position of the pyrimidine. After reduction, it was our intent
example, generation of lithioethylacetate (12) followed by
treatment with 6 and subsequent dehydration provided a
mixture of olefin isomers 13 in 79% yield following
distillation. Likewise, reaction of lithioacetonitrile (14) with
6 followed by dehydration gave the desired olefins 15 in
83% yield.
With ample supplies of 13 and 15 available, studies were
initiated to investigate the feasibility of pyrimidine ring
construction. Catalytic hydrogenation of the double bonds
to obtain 9, which under strongly basic conditions may be
induced to undergo formylation to afford an intermediate
such as 10. â-Aldehydo nitriles such as 10 are well-known
8
precursors of 4-functionalized-5-substituted pyrimidines
represented by the general structure 11. Incorporation of a
methylthio substituent at the 4-position of the pyrimidine
would provide our target, 1.
(6) Gribble G. W.; Leese, R. M.; Evans, B. E. Synthesis 1977, 172.
(7) Bruson, H. A.; Plant, J. L. J. Org. Chem. 1967, 32, 3356.
(8) For a general review of the synthesis of pyrimidines from dicarbonyls and
equivalents see: Brown, D. J. Heterocyclic Compounds; The Pyrimidines;
Wiley & Sons: New York, 1994; Chapters 2 and 3.
(9) March, J. AdVanced Organic Chemistry, 4th ed.; Wiley & Sons: New York,
1992; ;Chapter 16, p 881.
(10) For a review of the Knoevenagel reaction see: Jones, G. Organic Reactions;
Wiley & Sons: New York, 1967; Vol. 15, p 204.
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
277

Scheme 5 a
continued heating undergo N-formylation to 21. Further
heating leads to ammonia incorporation followed by irrevers-
ible intramolecular cyclization providing amino pyrimidine
1
9. The cyclization is generally complete in about 3 h at
1
85 °C. Prolonged heating can cause N-formylation of the
amine substituent of 19. The slow decomposition of form-
amide to carbon monoxide and ammonia provides the
requisite ammonia source.
At this juncture it had been shown that the desired 4,5-
disubstituted pyrimidine could be formed in a straightforward
manner from readily available materials. What was left to
demonstrate, was the conversion of the amine group of 19
to the methylthio substituent contained within 1. Towards
this end, the one-pot anhydrous diazotization/disulfide trap-
ping reaction of aromatic amines was attempted. Treatment
of a solution of 19 with tert-butyl nitrite and dimethyl
disulfide in a variety of solvents resulted in formation of a
dark reaction mixture in which no 1 could be detected via
GC/MS analysis.
a
2
Conditions: (a) H , Pd/C, EtOH; (b) LDA, THF, -70 °C then ethyl formate.
12
of 13 and 15 provided the nitrile 16 and ester 18 in excellent
yield (95-100%, Scheme 5). Generation of the nitrile
R-anion of 16 with lithium diisopropylamide (LDA) followed
by treatment with ethyl formate afforded the aldehydo-nitrile
1
7 in good yield (76%). However, much to our surprise,
exposure of ester 18 to the same reaction conditions did not
afford the anticipated product but yielded an intractable
mixture of unidentified products. This reaction was attempted
several times, producing, in all cases, similar results. Because
Scheme 7 a
17 had been obtained in good yield, no further time was
invested to understand this unexpected decomposition.
Treatment of â-dicarbonyl or equivalent compounds (i.e.,
â-aldehydo-nitriles such as 17) with formamide at 150-200
a
8
2 2 3 3
Conditions: (a) HCl, H O, heat; (b) SOCl , cat. DMF; (c) NaSCH , CH -
°
C can provide 5-alkyl pyrimidines. This reaction, known
OH.
as the Bredereck pyrimidine synthesis, has fairly wide
applicability but generally delivers the products in low to
moderate yield. We have found that exposure of 17 to the
Bredereck conditions affords the amino pyrimidine 19 in
good yield (72%) following removal of tars via filtration
through silica gel. The reaction pathway from 17 to 19 has
been elucidated with the aid of gas chromatography/mass
spectrometry (GC/MS) analyses performed during the course
of the reaction and by synthesis and characterization of the
intermediates 20 and 21 (Scheme 6).11 Initial warming of
Scheme 7 presents an efficient albeit less direct route from
1
9 to 1. Hydrolysis of 19 to the corresponding hyroxypyri-
13
midine hydrochloride 22 was accomplished in refluxing
aqueous hydrochloric acid. Conversion of 22 to the corre-
sponding chloropyrimidine was effected under standard
conditions. Substitution of thiomethyl for chloro provided 1
which was identical to material prepared by the previous
synthetic route.
Having proven that the strategy depicted in Scheme 3 is
viable, we undertook several short investigations, with the
aim of decreasing the number of synthetic steps and the cost
of certain reagents. For the conversion of indanone 6 to
unsaturated nitrile 15 (Scheme 4), LDA was used to generate
Scheme 6
1
4
the nitrile anion. Since LDA is prepared by the action of
n-butyllithium on diisoproplyamine, the use of such a base
on commercial scale would be cost-prohibitive. Gokel and
1
5
co-workers have described generation of the acetonitrile
anion using powdered potassium hydroxide in acetonitrile
(
11) Enamino-nitrile 20 can be independently synthesized by the reaction of 17
with ammonia in an appropriate solvent. The formylated material 21 has
been prepared via the reaction of 20 with acetic-formic mixed anhydride in
toluene. See the Experimental Section. Exposure of both of these materials
to the Bredreck conditions afforded pyrimidine 19.
(12) Beck, J. R.; Ackmann, S. A.; Staszak, M. A.; Wright, F. L. J. Heterocycl.
Chem. 1988, 25, 955.
(
13) We have represented the structure of 22 to be a hydrochloride salt. However,
we cannot rule out that the actual structure may be a covalent HCl addition
product.
(
14) For a general review of the generation and use of nitrile anions see:
Arseniyadis S.; Kyler, K. S.; Watt, D. S. Organic Reactions; Wiley &
Sons: New York, 1984; Vol. 31, p 1.
1
7 in the presence of formamide leads, rather quickly, to
(15) DiBiase, S. A.; Lipisko, B. A.; Haag, A.; Wolak, R. A.; Gokel, G. W. J.
Org. Chem. 1979, 44, 4640.
formation of the enamino-nitrile isomers 20 which on
278
•
Vol. 4, No. 4, 2000 / Organic Process Research & Development

as solvent and subsequent addition to carbonyl compounds.
We were skeptical about the application of this methodology
to ketone 6 due to its unreactive nature and the possibility
of generating acetonitrile oligimers. Nonetheless, the potential
long-term economic savings enticed us to conduct a brief
laboratory investigation of the method. Although this method
did deliver 15 in about 50% distilled yield, a significant
quantity of intractable tars (acetonitrile oligimers) remained
behind following distillation. Attempts at phase-transfer
catalysis and several other variants did not improve matters.
Alkoxide bases also provided dismal results for this reaction.
It appears as if complete deprotonation of acetonitrile with
an appropriate strong base is necessary for an efficient
condensation with 6.
intermediate is a dianion or radical-anion is not known;
however, we do assume that this intermediate is responsible
for the blood-red color. If a process utilizing sodium and
liquid ammonia is to be implemented, this brings about the
possibility of preparing the strong base sodium amide for
use in other steps of the synthesis. A 1968 report by Kaiser
and Hauser describe the high-yielding generation of the
acetonitrile anion using sodium amide in liquid ammonia
and its subsequent addition to carbonyl compounds. Ap-
plication of this technology toward a high-yielding conver-
sion of 6 to 15 and thus eliminating the need for the
expensive base LDA is described in the following section.
At this juncture, a new synthetic route to 1 had been
demonstrated on laboratory scale. The route, which begins
with 2,2-dimethyl-1-indanone (6), utilizes inexpensive raw
materials and features a novel combination of dissolving
metal reduction/formylation/cyclization to deliver a complex
amino pyrimidine intermediate in a reasonable number of
synthetic operations. This pyrimidine intermediate has been
converted to 1 under standard reaction conditions.
Development of the New Route. Having demonstrated
that a new, more economical synthetic route to 1 was viable,
our team turned its attention to development of a reliable,
efficient process involving a minimal number of unit
operations. As a result of significant laboratory study, we
have found that ketone 6 can be converted to aminopyrimi-
dine 19 in two reaction vessels and requires only three simple
workup procedures. Thus, a suspension of sodium amide is
prepared in liquid ammonia and then treated with anhydrous
acetonitrile in toluene to produce the nitrile anion (Scheme
9). Introduction of ketone 6 in toluene effects 1,2-addition,
1
7
Another synthetic step which was bothersome to us due
to the use of the expensive base LDA, was the conversion
of nitrile 16 to enol-nitrile 17 as shown in Scheme 5. Upon
more in-depth analysis of the conversion of unsaturated nitrile
15 to the desired enol-nitrile 17, we reasoned that if 15 could
be induced to undergo a 1,4-type hydride reduction under
aprotic conditions the desired R-nitrile anion would be
generated. Introduction of the formate electrophile at this
point could provide a one-pot conversion of 15 to 17 and
also eliminate the need for the strong base LDA. Not long
after the formulation of this idea, we came to the realization
that a dissolving metal reduction may provide a simple,
inexpensive solution to the problem. Indeed, treatment of
an ammonia/THF solution (-50 to -60 °C) of 15 with about
two equivalents of sodium metal in the presence of one
equivalent of tert-butyl alcohol resulted in an initial blood
red color. As the final pieces of metal dissolved, the color
changed to deep blue, indicating that the reduction phase
was complete. Addition of ethyl formate followed by
warming to room temperature and workup provided 17 in
Scheme 9 a
80% yield on the first attempt (Scheme 8). During the course
Scheme 8 a
a
Conditions: (a) Na, NH
3 3 3 2 3 2 4
, cat. Fe(NO ) ‚9H O; CH CN, toluene; (b) H SO ,
toluene; (c) Na, NH
3
, toluene, methyl formate; (d) formamide, heat.
and the resulting alkoxide is then quenched into cold aqueous
hydrochloric acid. It is worth noting that addition of the
aqueous acid to the alkoxide causes significant reversal of
a
3
Conditions: (a) Na, NH , THF then ethylformate.
1
7,18
the addition reaction.
The phases are separated, and the
of conducting subsequent reactions we found this method
to be fairly robust. For example, the reaction can be run near
the boiling point of ammonia (-33 °C), and the THF may
be replaced with the less expensive solvent toluene. The
sodium metal may be added to a solution of 15 in toluene/
ammonia, or a toluene solution of 15 may be added to sodium
dissolved in ammonia. The addition of one equivalent of tert-
butyl alcohol as the proton source is not necessary, thus
indicating that a very strong base, capable of abstracting a
proton from the ammonia solvent, is formed during the
course of electron transfer.16 Whether this extremely basic
organics freed of hydrochloric acid and chloride salts by
1
9
washing with water. Azeotropic drying of the toluene
solution in the presence of a catalytic amount of sulfuric
acid affords a toluene solution of the nitrile isomers 15. This
solution is added to sodium metal dissolved in liquid
ammonia to effect dissolving metal reduction/anion genera-
tion. The mixture is quenched with methyl formate followed
by an anhydrous proton source (ammonium chloride).
Distillation of the ammonia is performed with concomitant
(
(
17) Kaiser, E. M.; Hauser, C. R. J. Org. Chem. 1968, 33, 3402.
18) Beam, C. F. Can. J. Chem. 1978, 56, 2572.
(
16) For a review of metal/ammonia reductions of R-â unsaturated compounds
(19) Failure to remove chloride salts from the organic layer dictated the use of
increased amounts of sulfuric acid to complete the subsequent dehydration,
due to the formation and evaporative loss of hydrogen chloride.
see: Caine, D. Organic Reactions; Wiley & Sons: New York, 1976; Vol.
2
3, p 1.
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
279

addition of formamide. Heating the mixture to 185 °C for 3
h completes the pyrimidine cyclization. Upon cooling and
adding water/toluene, the desired aminopyrimidine (19) was
isolated as an off-white solid by simple filtration. The
addition of toluene during the work up is an inexpensive
method of removing tars, which had been previously
separated by chromatography. Our process provides consis-
tent yields (60-65% from 6) of high purity material (98+
wt %). This high-yielding sequence of reactions is remarkable
when one considers the number of chemical bonds being
formed and broken while utilizing a minimal number of
physical operations! The pyrimidine amine 19 was then
converted to 1 in three steps as depicted in Scheme 7.
5-Bromo-4-hydroxypyrimidine (3). A 5-L three-neck
round-bottom flask was equipped with a mechanical stirrer,
a thermowell, and a 500 mL addition funnel. To the reactor
was charged water (1500.0 g) followed by slow addition of
H SO (255.0 g, 2.6 mol) via the addition funnel. The warm
2
4
solution was cooled (ice bath) to ∼20 °C. To the reactor
was then added 5-bromopyrimidine (410.0 g, 2.58 mol). To
the slurry was then slowly added, via the addition funnel,
32% peracetic acid (628.0 g, 2.64 mol) at a rate where the
pot temperature did not exceed 70 °C with cooling (ice bath).
The slurry turned to a solution after two-thirds of the
peracetic acid was added. The reaction was stirred at 70 °C
for 2 h at which time analysis of the reaction solution by
gas chromatography indicated >99% product by area %. The
reaction was cooled to ∼50 °C, and to it was slowly added
Conclusions
In summary, we have reported two synthetic routes to
the experimental herbicide 1. Although our initial route has
only six synthetic steps and was amenable to pilot plant scale-
up, it relies on expensive raw material and reagents.
Economic evaluation of this route has shown it to be
incapable of producing 1 at a viable cost/kg. These economic
considerations dictated our discovery of a new route to 1
which begins with 2,2-dimethyl-1-indanone and utilizes
inexpensive raw materials and reagents. This new route
involves carbon-carbon bond formation reactions of two
nitrile-stabilized carbanions which are generated from the
inexpensive reagents sodium and liquid ammonia. A one-
pot dissolving metal reduction/formylation/Bredreck cycliza-
tion is employed and delivers a complex amino pyrimidine
intermediate in a minimal number of unit operations. This
pyrimidine intermediate is converted to 1 using standard
methods.
2 3
a solution of Na SO (110.0 g, 0.87 mol) in 500.0 g of water.
The pot temperature was maintained at <70 °C with an ice
bath. Starch iodide test of the reaction solution was negative
after the addition was completed. The reaction was cooled
to ∼50 °C, and to it was slowly added 50 wt % NaOH (580.0
g, 7.25 mol) solution while a pot temperature of <70 °C
was maintained with an ice bath. The slurry of off-white
precipitate was cooled to 20 °C. The solids were collected
in a Buchner funnel, and the off-white wet cake was rinsed
with water (2 × 500.0 g). The wet cake was transferred to
a 1-L three-neck round-bottom flask equipped with a
mechanical stirrer, a thermowell, and Dean-Stark trap with
a cold-water condenser. To the reactor was added 750 mL
of toluene. The toluene slurry was heated to reflux, and the
mixture was azeotropically dried, after which approximately
250 mL of toluene was removed by distillation. The reaction
mixture was cooled to ambient temperature, and the crystals
were collected by filtration as a toluene wet cake (415 g of
Experimental Section
1
1% toluene wt cake, 83%). The melting point of a dried
Reagents and solvents were reagent and technical grade
and used as received. Melting points were determined using
a Thomas-Hoover capillary melting point apparatus and are
uncorrected. Solvents were evaporated in vacuo using a
1
3
sample was 198-203 °C; C NMR (DMSO-d
49.6, 154.1, 157.6; recrystallization from water provided
an analytical sample; Anal. Calcd for C BrN O: C, 27.46;
H, 1.73; N, 16.01. Found: C, 27.41; H, 1.71; N, 16.00.
-Bromo-4-methylthiopyrimidine (4). A 1-L three-neck
round-bottom flask with thermowell was fitted with a
mechanical stirrer, condenser/CaSO drying tube and a
stopper. The vessel was loaded with a toluene wet cake of
(291 g of ∼83 wt. %, ∼1.35 mol) and thionyl chloride
570 g, 4.8 mol). The slurry was treated with DMF (4.0 mL,
.052 mol), and the mixture was heated to reflux and held
6
) 113.3,
1
H
4 3
2
1
13
Buchi rotary-evaporator. The H NMR and C NMR spectra
were determined using a Bruker AC-300 instrument in the
solvent noted. Mass spectra were determined using a Hewlett-
Packard GCD system operating at 70 eV in the electron
impact mode. Gas chromatography (GC) was performed
using a Hewlett-Packard 5890A instrument containing Restek
Rtx-5 column (15 m × 0.32 mm, 0.25 um film) under the
following conditions: temp1 ) 100 °C, time1 ) 1 min, rate
5
4
5
(
0
there for 2.5 h, at which time most of the solids had gone
into solution. The mixture was allowed to cool slightly, and
the condenser was replaced with a short-path distillation
apparatus. Thionyl chloride was distilled (condenser tem-
perature at 50 °C) to a final pot temperature of 132 °C,
resulting in 327 g of yellow distillate which contained a small
amount of solid. The internal temperature was cooled to 100
°C, and then toluene (354 g) was added. The distillation was
resumed until a pot temperature of 142 °C (111 °C head
temperature) was reached, resulting in 335 g of pale yellow
distillate. The distillation residue was allowed to cool to 50
°C, resulting in a thin slurry. The warm slurry was added in
one portion to thiourea (110 g, 1.45 mol) in ethanol (970 g)
)
20 °C/min, temp2 ) 275 °C, time2 ) 10 min, 5 psi helium
head pressure, FID detection or temp1 ) 60 °C, time1 ) 1
min, rate ) 20 °C/min, temp2 ) 275 °C, time2 ) 10 min.
High performance liquid chromatography (HPLC) was
conducted on a Keystone Scientific Phenylhypersil-1 column
(
250 × 4.6 mm, 5um) using mixtures of acetonitrile\water
(0.05% H PO ) as the eluent. Detection was, in general, by
3
4
UV at 214 nm. Thin-layer chromatography (TLC) was
routinely used to monitor reactions. TLC was conducted on
precoated Analtech silica GF/UV254 plates (250 um layer).
Column chromatography was conducted using silica gel 60
(230-400 mesh, Selecto Scientific). Elemental analyses were
carried out by Galbraith laboratories, Knoxville, TN.
280
•
Vol. 4, No. 4, 2000 / Organic Process Research & Development

contained in a 5-L three neck round-bottom flask with
mechanical stirring and a nitrogen pad. The distillation pot
residue was rinsed with toluene (125 g) and this also added
to the thiourea slurry. The resulting mixture exothermed to
graduated addition funnel with septum and nitrogen inlet.
To the flask was added n-BuLi (388 mL of 2.5 M in hexanes,
0.97 mol) via cannula. The solution was cooled to ∼-50
°C (dry ice/ethanol bath), and anhydrous THF (1455 mL)
was added. The solution was cooled to ∼-85 °C (liquid
nitrogen/ethanol bath) and a 15 wt % solution of 5-bromo-
4-methylthiopyrimidine (1332 g, 0.97 mol) in toluene was
added dropwise at a rate such that the internal temperature
did not exceed -80 °C. The resulting solution was stirred
for 1 h at -95 to -90 °C. A solution of 2,2-dimethylindan-
1-one (6, 155.2 g, 0.97 mol) in THF (340 mL) was added at
a rate such that the internal temperature remained below -80
°C. The cooling bath was removed and the dark brown
solution allowed to warm slowly to -25 °C and quenched
with water (250 mL). The vessel was fitted with a 30 cm
Vigreux column, and volatiles were removed at atmospheric
pressure (pot temperature 66-110 °C, head temperature 55-
97 °C). After the dark mixture was allowed to cool to 90
°C, water (1500 mL) was added and distillation continued.
The internal temperature was slowly raised to 101 °C (head
temperataure 93 °C) and a colorless distillate collected (788
g, ∼880 mL). The dark orange slurry was cooled to 80 °C,
and ethanol (1500 mL, 200 proof) was added. After the slurry
cooled to ambient temperature with good stirring, the
precipitate was collected via vacuum filtration. The orange
solids were slurry-washed with water (500 mL) and pulled
as dry as practicable on the filter. The resulting wet cake
(243 g) was air-dried to a constant weight, affording the
product as a tan solid (206 g, 74% yield). Reverse phase
HPLC analysis and GC analysis both indicated that the
4
4 °C over 10 min. After the mixture stirred at ambient
temperature for 2 h, the internal temperature had reached
4 °C, and the reaction was judged to be complete via HPLC
2
analysis. The cream-colored slurry was cooled to 15 °C, and
then 45% KOH (390 g, 3.13 mol) was added in one portion,
causing the internal temperature to rise to 30 °C. The slurry
was stirred at ambient temperature for 1.6 h at which time
HPLC analysis indicated the hydrolysis was complete. The
slurry was cooled to 10 °C, and then dimethylsulfate (165
g, 1.31 mol) was added in a rapid stream over 4 min. The
temperature exothermed to 34 °C over 10 min. After the
mixture stirred in the cooling bath for an additional 30 min,
the internal temperature was 25 °C, and the reaction was
judged to be complete as determined by HPLC. The resulting
slurry was allowed to stir at room temperature for 1 h, and
then water (800 g) and toluene (1200 g) were added. The
mixture was vigorously agitated for 10 min and then the
phases were allowed to separate. The organic phase was
washed with 10% NaCl (400 g) and water (400 g). Drying
by azeotropic removal of distillate (365 g) provided a light
amber solution of the product (1207 g). Analysis of the
toluene solution via GC indicated a purity of 93 area % (t
3.45 min) with the major impurity (2.5%, t ) 5.77 min)
R
)
R
being an isomer of 4. Reverse phase HPLC analysis of the
toluene solution utilizing an external standard method
indicated the solution was 17.7 wt % 4 (214 g, ∼77% yield).
1
Concentration of a small aliquot provided off-white cystals;
material was 99+ area % pure; mp 184-186 °C; H NMR
(CDCl ) 0.93 (s, 3H), 1.31 (s,3H), 2.53 (s, 3H), 2.58 (s, 1H),
3
1
mp 73-74 °C; H NMR (CDCl
3
) 2.56 (s, 3H), 8.44 (s,1H),
1
3
8
.85 (s, 1H); C NMR (CDCl
69.7; MS (GC) 206, 204 (M , 25, 25), 125 (100), 52 (50);
3
) 13.4, 118.7, 155.0, 155.7,
2.80 (ABq, J ) 15.8, 66 Hz, 2H), 7.25 (m, 4H), 7.52 (s,
+
+
1
1H), 8,80 (s, 1H); MS (GC) 286 (M , 45), 271 (30), 253
Anal. Calcd for C
Found: C, 29.39; H, 2.56; N, 13.69.
,2-Dimethyl-1-indanone (6). A 5-L four-neck round-
5
H
5
BrN
2
S: C, 29.28; H, 2.46; N, 13.66.
18 2
(20), 239 (100), 183 (60), 115 (55); Anal. Calcd for C16H N -
OS: C, 67.10; H, 6.34; N, 9.78. Found: C, 67.13; H, 6.52;
N, 9.73.
2
bottom flask was fitted with condenser, mechanical stirrer,
thermowell, and addition funnel. To this was added 623 g
of 45% KOH, and the solution was cooled to below 10 °C.
To 1 L of NMP was added 264.4 g (2 mol) of 1-indanone
2,2-Dimethyl-1-(4-methylthio-5-pyrimidinyl)indane (1).
To a 3-L round-bottom flask equipped with an overhead
stirrer, addition funnel, thermometer, and nitrogen inlet were
added 131.7 (0.46 mol) of 7, 9.7 g (0.26 mol, 0.56 mol equiv)
of sodium borohydride pellets (11 mm diameter), and 900
mL of methylene chloride. The mixture was cooled to -3
°C, and 699.9 g (6.13 mol, 13.3 mol equiv) of trifluoroacetic
acid was added over 0.5 h, maintaining the temperature
below 2 °C. After stirring between 0 and -2 °C for 3 h, the
mixture was gradually warmed to room temperature and
allowed to stir overnight (total reaction time 24 h). After
the mixture was cooled to 5 °C, the clear, amber/orange
solution was extracted with 500 mL of water. After 200 mL
of water was added, the mixture pH was adjusted from <1
to >12.5 by adding 443.6 g of 5 N aqueous sodium
hydroxide over 6 min, using an ice bath to keep the
temperature e26 °C. After the layers were separated, the
organic layer was concentrated to 209 g on a rotary
evaporator. After 399.8 g of methanol was added, the mixture
was concentrated to 353.6 g. The slurry was diluted with
251.5 g of water and was stirred for 30 min at room
(5) and 600 g (4.22 mol) of methyl iodide with stirring until
the indanone dissolved. The indanone solution was added
dropwise over 1.5 to 2 h to the KOH. The temperature varied
between 8 and 29 °C and was controlled by an ice bath and
by adjusting the rate of addition. After the addition was
complete, the mixture was stirred for 1.5 h, and GC area %
analysis showed 96% 6. The mixture was cooled to below
10 °C, 2.5 L of water was added, and the solid product was
stirred for an additional 0.5 h. The off-white crystals were
filtered, washed twice with 1 L of water, and air-dried for
1
4
0 h to yield 284 g of 6 (89%); mp 40-42 °C; H NMR
(
(
(
CDCl
3
) 1.20 (s, 6H), 2.97 (s, 2H), 7.34 (m, 2H), 7.55
+
apparent t, 1H), 7.72 (d, 1H); MS (GC) 160 (M , 40), 145
100).
2
,2-Dimethyl-1-(4-methylthio-5-pyrimidinyl)indan-1-
ol (7). A nitrogen-purged 5-L three-neck round-bottom flask
with thermowell was fitted with a mechanical stirrer and a
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
281

temperature. The slurry was filtered, and the solids were
washed with 100 mL of water. The product was dried at
about 65 °C on a rotary evaporator, giving amber, sugar-
like solids. A total of 119.0 g of solids were obtained which
assayed at 98.1% by HPLC, giving a calculated yield of
was cooled to room temperature, and water (75 mL) was
added. The solvent was removed in vacuo, leaving a dark
liquid. The liquid was diluted with H O (400 mL) and
2
extracted with toluene (3 × 200 mL). The organic phase
was concentrated on the rotovap, leaving a black oily tar.
Kugelrohr distillation provided the crude product (bp 128
°C, 1.6 mmHg) as a pale yellow liquid. Distillation through
a 10 cm Vigreux column afforded unreacted ketone and then
the product 15 (bp 110 °C at 0.4 mmHg) as a clear yellow
1
16.8 g (94%). Recrystallization from methanol/water pro-
1
3
vided an analytical sample; mp 94-95 °C; H NMR (CDCl )
0
2
7
1
1
5
6
.88 (s, 3H), 1.32 (s, 3H), 2.60 (s, 3H), 2.85 (Abq, J ) 15.7,
0.4 Hz, 2H), 4.32 (s, 1H), 6.93 (d, J ) 7.3 Hz, 1H), 7.12-
.27 (m, 3H), 7.71 (s, 1H), 8.83 (s, 1H); 13C NMR (CDCl
3
)
R
liquid (58.1 g, 51%): GC analysis, t ) 5.17 min (96%).
2.9, 24.2, 30.2, 45.7, 47.2, 55.0, 124.8, 125.0, 126.8, 127.3,
2-(2,2-Dimethyl-1-indanyl)acetonitrile (16). A solution
of the olefin isomers 15 (20.8 g, 0.114 mol) in ethanol (100
mL) were added to a nitrogen-purged Parr reactor containing
10% Pd/C (1.0 g) and water (15 mL). The mixture was
hydrogenated with good stirring at 200 psi for 16 h. The
slurry was filtered through Celite and the pad washed with
ethanol (25 mL). The solution was concentrated in vacuo
+
32.8, 143.2, 144.2, 154.1, 155.9, 169.8; MS (GC) 270 (M ,
18 2
5), 255 (100); Anal. Calcd for C16H N S: C, 71.07; H,
.71; H, 10.36. Found: C, 71.18; H, 6.71; N, 10.23.
E- and Z-2-(2,2-Dimethyl-1-indanylidene)acetonitrile
(
1
15). To a cold (-73 °C) solution of LDA‚THF (200 mL of
.5 M in cyclohexane, 0.3 mol) in THF (100 mL) was added
anhydrous CH CN (15.7 mL, 0.3 mol) dropwise over 50 min.
After the mixture stirred for 35 min, a solution of the ketone
6, 35 g, 0.22 mol) in THF (50 mL) was added dropwise at
a rate such that the internal temperature remained below -65
C. After the addition was complete, the cooling bath was
3
and the residue partitioned between CH
water (40 mL). The organics were dried (Na
2
Cl
2
(60 mL) and
SO ) and the
2
4
(
solvents removed on the rotovap, leaving a yellow liquid.
Drying to a constant weight (0.5 mmHg/room temperature)
1
°
afforded 16 as a pale yellow liquid (21 g, 100%): H NMR-
removed and the mixture allowed to warm to 0 °C. The
amber-colored solution was poured into 1 M HCl (300 mL)
and then extracted with toluene (2 × 100 mL). The organics
(CDCl
2H), 2.77 (2, 1H), 3.08 (t, J ) 7.2 Hz, 1H); C NMR
(CDCl ) 17.8, 22.6, 28.1, 43.0, 46.5, 50.4, 119.2, 123.8,
124.7, 126.4, 127.2, 141.9, 143.1; Anal. Calcd for C13H15N:
3
) 1.04 (s, 3H), 1.25 (s, 3H), 2.55 (d, J ) 7.2 Hz,
13
3
were dried (Na
2
SO
4
/MgSO
4
) and the solvents removed in
vacuo, leaving an amber oil (40.2 g). The oil was dissolved
in toluene (100 mL) and treated with p-TsOH‚H O (3.0 g).
The solution was heated to reflux, and water was removed
using a Dean-Stark trap. After 3 h, an additional 1.0 g of
2
p-TsOH‚H O was added and the mixture heated at reflux
C, 84.28; H, 8.16; N, 7.56. Found: C, 83.93; H, 8.22; N,
7.22.
2
2-(2,2-Dimethyl-1-indanyl)-1-cyanoethanal (17). To a
cold (-73 °C) solution of LDA‚THF (180 mL of 1.5 M in
cyclohexane, 0.27 mol) in THF (180 mL) was added a
solution of 16 (22.8 g, 0.123 mol) in THF (25 mL) dropwise
at a rate such that the internal temperature did not exceed
-66 °C. After the addition was complete, the clear brown
solution was allowed to stir for ∼30 min. Ethyl formate (14.5
mL, 0.179 mol) was added dropwise such that the internal
temperature did not exceed -45 °C. The cooling bath was
removed, and the mixture was allowed to warm to room
temperature. Water (180 mL) was added and then most of
the THF removed on the rotovap. The residue was diluted
with water, and the resulting aqueous slurry was extracted
with toluene (2 × 180 mL). The aqueous solution was chilled
in an ice bath and taken to pH 3 by the addition of
concentrated HCl. The resulting solid was filtered and
washed well with water. Drying to a constant weight
provided 17 as a pale yellow solid (22.4 g, 85%) which was
97+% pure via GC analysis. Recrystallization of a small
sample from toluene afforded 17 as a white crystalline
for 3 h, at which time GC analysis indicated that the
dehydration was complete. The organics were washed with
water (100 mL), 5% Na
and brine (100 mL). After the organics were dried (Na
2
CO
3
(2 × 100 mL), water (100),
SO ),
2
4
the solvents were removed in vacuo, leaving a dark brown
liquid. Distillation through a 10 cm Vigreux column provided
the nitrile isomers 13 as a pale yellow liquid (33.5 g, 83%,
1
bp 110 °C at 0.4 mmHg): H NMR (CDCl
1
3
) 1.27 (s, CH
), 5.20 (s, CH), 5.70
s, CH), 7.27-7.5 (m, Ar H’s and one olefinic H), 8.36 (d,
3
),
3 3 3
.53 (s,CH ), 2.91 (s, CH ), 2.99 (s, CH
(
13
J ) 7.4 Hz,1 H); C NMR (CDCl3) 26.9, 28.6, 43.8, 44.5,
4
1
1
6.0, 48.1, 84.5, 84.9, 117.4, 117.8, 121.4, 124.9, 125.3,
25.4, 127.0, 131.5, 131.7, 136.2, 137.4, 145.5, 146.4, 173.4,
73.7; GC analysis t
R
) 5.15 (41) and t
R
) 5.17 (59); MS
+
(GC) of the isomers are virtually identical 183 (M , 30),
1
7
68 (100); Anal. Calcd for C13
.45. Found: C, 85.46; H, 7.29; N, 7.50.
-(2,2-Dimethyl-1-indanylidene)acetonitrile (15) via the
H13N: C, 85.21; H, 7.15; N,
1
3
2
6
solid: mp 144-46 °C; C NMR (DMSO-d ) shows the
KOH acetonitrile method. A nitrogen-purged 3-L three
necked round-bottom flask with thermowell was fitted with
a mechanical stirrer, septum, and condenser with nitrogen
inlet. To the flask was added acetonitrile (604 mL) and
powdered KOH (88 g, ∼1.33 mol) causing a mild exotherm.
After the mixture cooled to ambient temperature, a solution
of the ketone 2 (100 g, 0.625 mmol) in acetonitrile (200 mL)
was added, resulting in a black-purplish mixture. The dark
mixture was allowed to stir at reflux overnight. The mixture
material to be a ∼1/1 mixture of enol tautomers 23.7, 27.5,
28.4. 44.8, 45.5, 46.2, 46.5, 49.9, 54.3, 86.1, 88.9, 117.8,
120.8, 124.5, 124.7, 124.8, 126.3, 126.4, 126.9, 127.0, 142.3,
142.8, 142.9, 143.1, 158.7, 159.5; MS (GC), the isomers are
inseparable, 213 (M , 15), 198 (10), 145 (100), 128 (25),
115 (45).
Ethyl E- and Z-2-(2,2-Dimethyl-1-indanylidene)acetate
(13). To a cold (-70 °C) solution of LDA‚THF (100 mL of
1.5 M in cyclohexane, 0.15 mol) in THF (50 mL) was added
+
282
•
Vol. 4, No. 4, 2000 / Organic Process Research & Development

anhydrous ethyl acetate (14.7 mL, 0.15 mol) dropwise at a
rate such that the internal temperature did not exceed -67
warmed to 25 °C using a warm water bath to distill ammonia.
Stirring was stopped to allow the black and gray insoluble
materials to settle to the bottom of the reactor. The reaction
solution was transferred via a cannular tube to a 1-L three-
necked round-bottomed flask equipped with a mechanical
stirrer, ice bath, and a thermometer. The flask contained a
vigorously stirred solution of 3 N HCl (250 mL) pre-cooled
to -8 °C. The rate of transfer was such that the quenched
mixture was maintained at or below 0 °C. The reactor was
rinsed with 50 mL of anhydrous toluene and transferred to
the quenched mixture. In a separatory funnel the aqueous
phase was discarded, and the toluene phase was washed with
water (2 × 100 mL). The toluene phase was then charged
into a 500 mL three-necked round-bottomed flask equipped
with a mechanical stirrer, a thermometer, a heating mantle,
and a Dean-Stark trap with a cold water condenser. The
toluene solution was dried azeotropically until the pot
temperature reached 110 to 114 °C at which time heating
°
C. After the mixture stirred for 25 min, the ketone 2 (22 g,
0
.138 mol) was added in portions at a rate such that the
internal temperature remained below -64 °C. The dark
solution was stirred in the cold for 1 h and then allowed to
warm slowly to -20 °C. To the mixture was added saturated
aqueous NH
aqueous phase was extracted with EtOAc (50 mL), and the
combined organics were dried (Na SO /MgSO ). The sol-
4
Cl (75 mL), and the phases were separated. The
2
4
4
vents were removed in vacuo, leaving a light amber oil. The
oil was dissolved in toluene (50 mL) and treated with
p-TSOH‚H O (3 g). The solution was heated to reflux using
2
a Dean-Stark trap to remove water. The mixture was
refluxed for ∼2 h at which time GC analysis indicated the
dehydration was complete. The dark amber mixture was
cooled to room temperature and then extracted with 5% Na
CO
(3 × 75 mL) and water (2 × 75 mL). After drying
Na SO ) the solvents were removed in vacuo, leaving a dark
2
-
3
(
2
4
was stopped and one drop of concentrated H
2
SO
4
(∼20 mg)
amber oil. Distillation through a 10 cm Vigreux column
was added. Heating and azeotropic drying was continued,
and after 25 min, GC analysis indicated the dehydration
reaction was complete. The reaction was cooled to ∼30 °C
and transferred to a separatory funnel. The toluene solution
was washed with water (2 × 100 mL) and then transferred
back into the reactor. The toluene solution was dried
azeotropically and distilled until the total in-pot volume was
∼70 mL. The yellow toluene solution of 15 was allowed to
cool to room temperature under a slight positive pressure of
nitrogen.
afforded the olefin isomers 13 as a yellow liquid (25.2 g,
1
7
(
(
9%, bp 132-35 °C at 1.4 mmHg); H NMR(CDCl
3
); MS
+
GC) isomers are inseparable, 230 (M , 50), 215 (10), 185
40), 169 (20), 157 (100), 142 (85).
2-(2,2-dimethyl-1-indanyl)acetate (18). The procedure
described for the preparation of 16 was utilized. The ester
1
1
8 was obtained as a clear colorless liquid (4.8 g, 95%): H
+
NMR(CDCl
3
); MS (GC) 232 (M , 15%), 158 (95%), 143
(
100%).
-Amino-(2,2-dimethyl-1-indanyl)pyrimidine (19) from
. A 500 mL three-neck round-bottomed flask equipped with
4
A 500 mL three-necked round-bottomed flask equipped
with a mechanical stirrer (glass paddle), 2 dry ice condensers,
and a nitrogen inlet was purged with nitrogen for 30 min.
The reactor was charged with sodium metal (8.9 g, 0.387
mol). The condensers were charged with dry ice/ethanol, and
the vessel was cooled in a dry ice/ethanol bath. Ammonia
(250 mL, ∼193 g) was slowly condensed into the reactor
affording a deep black-blue solution. One of the dry ice
condensers was replaced with a rubber septum. The yellow
solution of 15 from the above reaction was added via a
syringe and needle while maintaining a pot temperature of
-33 °C to -40 °C. The solution retained its blue-green color.
(Note: If the solution is a reddish orange color, then more
sodium metal is required, and the stoichiometry of the
remaining reagents are adjusted accordingly). The mixture
was stirred for 5 min and anhydrous methylformate (14.1 g,
14.4 mL, 0.21 mol) was added dropwise via a syringe needle
over a period of 3 min. The pot temperature was maintained
at less than -30 °C. After over half of the methylformate
solution was added, the reaction solution turned from blue
to an orange-yellow color. After few minutes, the solution
thickened to a pale yellow slurry with a consistency of thin
pancake batter. The yellow slurry was stirred at -30 to -40
°C for 15 min, and formamide (75 mL) was slowly added
6
a mechanical stirrer (glass paddle), two dry ice condensers,
and a nitrogen inlet was purged with nitrogen for 30 min.
The condensers were charged with dry ice/ethanol, and the
vessel was cooled in a dry ice/ethanol bath. Ammonia (175
mL, ∼135 g) was slowly condensed into the reactor, and
then one of the dry ice condensers was replaced with a
stopper. Initially, four pieces of sodium metal (approximately
0
.2 g size) were added to the reactor, causing the solution
to turn to a deep blue color. Approximately 50 mg of Fe-
NO ‚9H O was added as a catalyst. After the solution
turned to a grayish mixture, more sodium metal (5.75 g total,
.25 mol) was added in portions, while the pot temperature
(
3
)
3
2
0
was maintained at -33 to -35 °C (addition time was 50
min). After the addition was completed and the reaction
mixture turned gray, stirring was continued for an additional
15 min at -35 °C. The stopper was replaced with a rubber
septum, and anhydrous toluene (50 mL) was added via a
syringe and needle. Anhydrous acetonitrile (9.44 g, 12 mL,
0.23 mol) was then added dropwise via a syringe and needle
over a period of 5 min. The reaction was stirred for an
additional 30 min while the temperature of -35 °C was
maintained.
A solution of 2,2-dimethyl-1-indanone (6, 31.6 g, ∼95%,
while allowing the NH
increased to -22 °C at the end of addition with most of the
paste being converted to an orange solution. NH Cl (20.0 g,
0.37 mol) was cautiously added, turning the solution to a
yellow color with some precipitate. The reactor was fitted
3
to distill. The pot temperature had
0
.187 mol) in anhydrous toluene (40 mL) was added to the
reaction mixture (via a syringe and needle) over a period of
5 min. After the mixture stirred at -35 °C for 30 min,
cooling was discontinued, and the reaction was slowly
4
1
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
283

with a stopper, a short-path distillation head, and a nitrogen
inlet with the exit tube in a water trap. The reaction mixture
was heated slowly to 185 °C while the distillate was
removed. As the pot temperature increased and distillate was
removed, a precipitate began to form, and the mixture
became a slurry as the pot temperature reached 185 °C. The
reaction was held at 185 °C for 4 h at which time GC analysis
indicated that the reaction was mostly completed. The brown
mixture was cooled to 120 °C, toluene (45 mL) and water
Drying to a constant weight provided the product as a beige
solid (805 mg, 72%): mp 156-59 °C; MS (GC) as above.
5-(2,2-Dimethyl-1-indanyl)-4(3H)pyrimidinone (22).
Compound 19 (43.9 g, 0.184 mol based on 77.4 g of wet
cake which assayed at 56.7%), water (110.3 g), and 37%
hydrochloric acid (294.4 g) were heated at 110-116 °C for
21 h. The thick slurry was cooled to room temperature and
was filtered. The solids were washed with water (150 mL),
followed by toluene (150 mL). After drying under vacuum
at 75 °C, the light tan solids (49.5 g) assayed at 86.6%
(95 mL) were added. The mixture was cooled to room
temperature with vigorous stirring and stirred at room
temperature for 150 min. The crude product (tar-like clumps)
was collected in a sintered glass funnel via filtration. The
mother liquor was returned to the reactor to rinse the crystals
that had adhered to the walls of the reactor. The reactor was
then rinsed with 50 mL of toluene which was used to slurry
the crude clumps in the funnel into a homogeneous mixture,
and this mixture was filtered. The wet cake was slurried and
washed with an additional portion of 75 mL of toluene. The
cake was washed with water (100 mL) and dried in a vacuum
oven at 50 °C overnight. The off-white solid (28.9 g) assayed
at 98.2%, affording an isolated yield of 63.5% from the
ketone 6. Treatment of a small sample with ethanol/charcoal
followed by recrystallization from toluene provided an
relative to the free base of 22, giving a calculated yield of
1
6
42.9 g (97%) of 22: H NMR (DMSO-d ) 10.62 (2 H, br
s), 9.10 (1 H, s), 7.29-7.01 (5 H, m), 4.26 (1 H, s), 2.84-
2.69 (2 H, dist q), 1.16 (3 H, s), 0.84 (3 H, s); mp 258-259
°C. A small sample was converted to the free base by
partitioning between aqueous K
cal sample was obtained by recrystallization from ethanol/
water: mp 138-139 °C; Anal. Calcd for C15 O: C,
2 3
CO and EtOAc. An analyti-
16 2
H N
74.97; H, 6.71; N, 11.66. Found: C, 74.81; H, 6.92; N, 11.45.
2,2-Dimethyl-1-(4-methylthio-5-pyrimidinyl)indane (1).
Compound 22 (24.9 g, 0.104 mol based on 28.9 g of solids
which assayed at 86.3%), 1,2-dichloroethane (100 mL),
thionyl chloride (25 g), and dimethylformamide (1.5 g) were
heated to reflux for 7 h. The solution was cooled to room
temperature, and water (50 mL) was added. The mixture pH
was adjusted to approximately 7.5 by adding 8.2 g of
potassium carbonate. The layers were separated. To the
organic layer were added, in shots over several hours, water
(90 mL), sodium thiomethoxide (24.96 g), 1.0 g of benzyl-
triethylammonium chloride, and 20% aqueous sodium hy-
droxide (5.6 g). After the reaction showed 87% conversion
by area %, the layers were separated, and the organic layer
was re-extracted with water (50 mL). The combined aqueous
layers were extracted with 1,2-dichloroethane (30 mL). To
the combined organic layers were added another 0.3 g of
benzyltriethylammonium chloride and 2.08 g of sodium
thiomethoxide. After stirring 3 days, the mixture showed
about 91% conversion. The layers were separated, and the
aqueous layer was reextracted with 1,2-dichloroethane (30
mL). The combined organic layers were concentrated to a
brown oil. The oil was diluted with acetonitrile (50 mL),
giving solids, followed by water (50 mL). The slurry was
filtered, washing the solids with water (25 mL). The solids
were recrystallized from hot acetonitrile (50 mL), adding
water (50 mL) after crystals formed. The slurry was filtered,
and the solids were washed with 25 mL of city water. The
solids were recrystallized again from hot acetonitrile (50 mL)
and water (20 mL). The slurry was filtered and the solids
were washed with water (20 mL). The solids were dried
under vacuum at 75 °C, giving 18.1 g (65%) of light tan
granules of pure 1.
13
analytical sample: mp 162-164 °C; C NMR (CDCl
3
) 24.0,
24.6, 29.9, 45.3, 45.7, 47.1, 47.9, 53.5, 58.8, 115.6, 116.8,
1
1
24.8, 125.1, 125.4, 125.7, 126.7, 127.0, 127.2, 127.9, 140.6,
42.6, 143.2, 143.4, 155.8, 156.4, 157.2, 157.6, 161.7, 162.2;
1
H NMR not listed due to the spectral complexity caused
by amine/imine tautomers;20 MS (GC) 239 (M , 75%), 224
+
(
100%), 196 (20%). Anal. Calcd for C15
H, 7.16; N, 17.56. Found: C, 75.51; H, 7.37; N, 17.48.
-Amino-(2,2-dimethyl-1-indanyl)pyrimidine (19) from
7. The enol-nitrile isomers 17 (1.0 g, 4.69 mmol) and
17 3
H N : C, 75.28;
4
1
ammonium chloride (1.0 g, 18.7 mmol) were slurried in
formamide (50 mL) and heated to 160 °C under nitrogen.
After 2.5 h at this temperature, GC and GC/MS analyses
indicate that all of 17 had been consumed and a 94/5 mixture
of isomers 20/21 had resulted. Continued heating at 175 °C
for 1 h showed little increase in the amount of 19. The
mixture was warmed to 190 °C and held there for 3 h at
which time all of the 20 and 21 had been consumed and 19
was the major component as determined via GC/MS analysis.
The dark reaction mixture was cooled to room temperature,
and then formamide (43 mL) was removed via distillation
at 0.1 mmHg/65-75 °C. The dark residue was triturated with
2 3
water (50 mL) containing K CO (1.0 g). The resulting dark
solid was collected on a filter and washed with water. The
aqueous phase was extracted with EtOAc (2 × 50 mL) and
the dark solid then dissolved in the EtOAc extracts. This
solution was filtered through a plug of silica gel ( ∼30 g),
and the pad was rinsed with EtOAc (150 mL). The filtrates
that contained product (by TLC analysis) were combined
and the solvents removed on the rotovap, leaving a tan solid.
E- and Z-2-(2,2-Dimethyl-1-indanyl)-2-cyanoethen-
amine (20). A solution of nitrile 17 (1.0 g, 4.7 mmol) was
dissolved in ethanol (14 mL) and treated with NH Cl (20
4
mg). The solution was heated at reflux under an atmosphere
of NH (balloon) until the reaction was judged to be complete
3
via GC analysis (4.5 h). After cooling to room temperature
the precipitate was collected and washed with a little ethanol.
(
20) A note added in proof: compound 19 prepared by the oxidation of 1 to the
corresponding methy sulfone followed by displacement with ammonia,
displayed spectral properties identical to material prepared by the cyclization
method.
284
•
Vol. 4, No. 4, 2000 / Organic Process Research & Development

The material was dried to a constant weight at 0.5 mmHg/
room temperature providing 20 as a fluffy white solid (550
mmHg/room temperature provided isomers 21 as a pale
yellow solid: GC analysis indicated an 83(t ) 7.57)/14(t
R
R
mg): GC analysis indicated a 96(t
mixture of olefin isomers; MS (GC) for each isomer gave
R
) 7.04)/4(t
R
) 7.64)
) 7.91) mixture of isomers 21 along with 1.3% of the
N-acetylated analogue; MS (GC) of each of the isomers gave
+
+
almost identical spectral data, 212 (M , 85), 197 (30%), 180
almost identical spectral data, 240 (M , 60), 195 (35), 180
1
1
(
35), 170 (100), 129 (70); H NMR (major isomer, DMSO-
(100); H NMR (major isomer, CDCl
3
) 1.03 (s, 3H), 1.20
d
6
7
2
1
6
) 0.86 (s, 3H), 1.10 (s, 3H), 2.66 (bs, 2H), 3.33 (s,1H),
(s, 3H), 2.81 (ABq, J ) 9.8, 15.5 Hz, 2H), 3.48 (s, 1H),
7.13-7.24 (m, 4H), 7.49 (d, J ) 11.6 Hz, 1H), 8.20 (s, 1H),
8.68 (bd, J ) 11.6 Hz, 1H); Anal. Calcd for C15H N O: C,
17 2
.35 (d, J ) 10.8 Hz, 2H), 6.82 (t, J ) 10.8 Hz, 1H)7.05-
.17 (m, 4H); 1 C NMR (major isomer, DMSO-d
7.3, 45.1, 46.0, 56.4, 73.3, 120.0, 124.3, 124.4, 126.0, 126.5,
42.0, 143.9, 150.2; Anal. Calcd for C14 : C, 79.21;
3
) 23.5,
6
74.97; H, 6.71; N, 11.66. Found: C, 75.03; H, 6.90; N, 11.39.
16 2
H N
H, 7.60; N, 13.20. Found: C, 79.50; H, 7.80; N, 12.95.
E- and Z-2-(2,2-Dimethyl-1-indanyl)-2-cyano-N-formyl-
ethenamine (21). A slurry of the amino nitriles 20 (500 mg,
Acknowledgment
The authors thank L. D. Markley, J. J. Tai, and J.W. Hull
for helpful discussions throughout the course of these studies.
Reactive chemicals testing and review were conducted by
M. S. Labarge, T. C. Hofelich, and D. J. Frurip. Chemical
engineering support for the pilot plant scale-up was provided
by T. W. Toyzan, G. P. Heinrich, and R. L. Kiekhaefer. We
thank J. W. Russell for analytical methods development and
M. A. Hadd and T. S Holman for assistance in the laboratory
preparation of 4 and 7.
2
.4 mmol) in toluene (8 mL) was treated with acetic-formic
21
mixed anhydride (400 mg, 4.5 mmol) and the mixture
warmed to 90 °C. After 30 min, all solids had gone into
solution, and GC analysis indicated that 5% of 20 remained.
Additional mixed anhydride (75 mg) was added and the
mixture kept at 90 °C for 45 min. The reaction was cooled
to room temperature, and the volatiles were removed on the
rotovap, leaving a pale yellow oil. Drying overnight at 1
(
21) For the preparation of acetic-formic mixed anhydride see: Fieser, M.;
Fieser, L. Reagents for Organic Synthesis; Wiley & Sons: New York, 1969;
Vol. 2, p 10.
Received for review March 17, 2000.
OP000033Z-
Vol. 4, No. 4, 2000 / Organic Process Research & Development
•
285