A new ans Practical Synthesis of Octafluoroparacyclophane

Full text of DOI:10.1021/jo970927q

7500
J . Org. Chem. 1997, 62, 7500-7502
Sch em e 1. P a r ylen e Va p or Dep osition P r ocess
A New a n d P r a ctica l Syn th esis of
Octa flu or o[2.2]p a r a cyclop h a n e
William R. Dolbier, J r.,* Xiao X. Rong, and Yuelian Xu
Department of Chemistry, University of Florida,
Gainesville, Florida 32611-7200
William F. Beach
Specialty Coating Systems, Inc., 5707 West Minnesota
Street, Indianapolis, Indiana 46241
Received May 27, 1997
Per-R-fluoro[2.2]paracyclophane (AF 4), being the pre-
cursor to the organic, vapor-deposited parylene polymer,
poly(R,R,R′,R′-tetrafluoro-p-xylylene (Parylene VIP AF4)
(Scheme 1), is a material of considerable commercial
interest.1
Sch em e 2. 1969 Un ion Ca r bid e P r ocess
The Parylene AF4 polymer combines low dielectric
constant (2.25; normal to the film plane)2-4 with high
thermal stability (<1% wt loss/2 h at 450 °C)2,3,5 and low
moisture absorption (<0.1%).3 With such properties and
because its in vacuo deposition process assures confor-
mality to microcircuit features and superior submicron
gap-filling capability, the Parylene AF4 polymer shows
considerable promise as an interlayer dielectric for on-
chip high speed semiconductor device interconnection.2,3
Such a market would require that the AF4 dimer
precursor be available in commercial quantities. How-
ever, until now AF4 has only been able to be synthesized
in gram quantities by two alternative methods, one an
awkward pyrolytic/reductive process developed by Union
Carbide in the 1960’s to first synthesize AF4 (Scheme
2),6,7 and the other, a much improved reductive method
which uses a complex, low-valent form of Ti to effect
debromination of 1,4-bis(bromodifluoromethyl)benzene,
2, a process which allows relatively easy preparation of
gram quantities of AF4 in an acceptable (32%) yield
(Scheme 3).8 Nevertheless, the high dilution technology
required in the latter process proved not amenable to
scale up, a result which frustrated attempts at large scale
preparation of AF4 by this method.9
Sch em e 3. 1992 Dolbier P r ocess
Sch em e 4. Over a ll Mech a n istic Sch em e for AF 4
F or m a tion
An important aspect of any method for [2.2]paracyclo-
phane synthesis is the necessity to carry out the reaction
under high dilution conditions. In the Ti* process, such
high dilution was maintained by traditional methods,
which included very slow, simultaneous addition of both
dibromide and Ti* to a large surplus of solvent. As such,
(1) Beach, W. F.; Lee, C.; Bassett, D. R.; Austin, T. M.; Olson, R. A.
Xylylene Polymers In Wiley Encylopedia of Polymer Science and
Technology; Wiley: New York, 1989; Vol. 17, pp 990-1025.
(2) Majid, N.; Dabral, S.; McDonald, J . F. J . Electron. Mater. 1989,
18, 301-311.
10 g of dibromide 2 in 500 mL of THF and 3.5 equiv of
reduced Ti in 500 mL of THF were added over a period
of 10 h to 500 mL of refluxing THF to give 1.7 g of AF4
(32%).
In the case of AF4, all methods which have been used
for its preparation involve, we believe, the intermediacy
of R,R,R′,R′-tetrafluoro-p-xylene, 1, as a reactive mono-
meric species.10 As shown in Scheme 4, this reactive
species must be generated under reaction conditions
wherein it will survive until it encounters molecule of 1
to form the dimer diradical 3.
(3) Wary, J .; Olson, R. A.; Beach, W. F. VMIC Conference, Santa
Clara; Feb, 1996; paper 6-B.
(4) Gaynor, J .; Chen, J .; Nguyen, H.; Brown, G.; Taylor, K.; Luttmer,
J . D.; Plano, M. A.; Cleary, T.; Wing, J .; Kelly, Proceedings of the 191st
Meeting of the Electrochemical Society, Symposium F: Low and High
Dielectric Constant Materials, 97-1, 325, paper no. 259.
(5) Williams, K. R. J . Thermal Anal. 1997, 49, 589-594.
(6) Chow, S. W.; Pilato, L. A.; Wheelwright, W. L. J . Org. Chem.
1970, 35, 20-22.
(7) A recent modification of the Chow process which substitutes
helium for steam: Grechkine, E. V.; Soohilian, V. A.; Pebalk, A. V.;
Kardash, I. E. Zh. Org. Khim. 1993, 29, 1999-2001.
(8) Dolbier, W. R., J r.; Asghar, M. A.; Pan, H.-Q.; Celewicz, L. J .
Org. Chem. 1993, 58, 1827-1830.
3 will most likely be formed in an extended form such
as 3-ext which will need to undergo an entropy unfavor-
(9) Whereas it proved relatively easy to get 2 grams of pure AF4 by
this method, attempts to scale the process up to produce greater than
10 g per run led to formation of increasing amounts of trimer and
polymer at the sacrifice of desired dimer.
(10) Monomer 1 has been previously isolated at -180 °C: Pebalk,
A. V.; Kardash, I. e.; Kozlova, N. V.; Zaitseva, E. L.; Kozlov, Yu. A.;
Pravednikov, A. N. Vysokomol. Soedin. Ser. A 1980, 22, 972-976;
Chem. Abstr. 1983, 93, 132917j.
S0022-3263(97)00927-4 CCC: $14.00 © 1997 American Chemical Society

Notes
J . Org. Chem., Vol. 62, No. 21, 1997 7501
Sch em e 5. Mor i’s Stu d y
Under otherwise identical conditions, carrying out the
reaction in pure THF gave but an 8% yield, whereas use
of HMPA in place of the DMSO (again 8:1) led to a 16%
yield. If KF was used in place of CsF, only a 5% yield
was obtained. Attempts to substitute other ethereal
solvents, such as diethyl ether, dioxane, or diglyme, for
THF also led to much poorer results.
It is believed that the active reducing agent in this
reaction is generated in situ by the reaction of fluoride
ion, F-, with TMSTBT. Because Me3SiF and Bu3SnBr
are formed as a result of the overall reaction, it is
assumed that the process begins by attack of the TM-
STBT by fluoride ion. That being the case, the actual
active reducing agent could be either the complex anion
4 or the tributyltin anion itself, 5 (Scheme 7).
Sch em e 6. New AF 4 Syn th esis
able conformational change to the syn conformer, 3-syn ,
before it can undergo unimolecular cyclization to the
desired paracyclophane product, AF4. If 3 encounters
another molecule of monomer 1 before it can cyclize, it
will react to form the linear trimer diradical which,
having an even smaller probability of undergoing cycliza-
tion, usually, but not always, continues on to form AF4
polymer. This polymer has no appreciable solubility in
any solvent yet tested. The high dilution conditions are
designed to provide the optimal kinetic environment for
allowing unimolecular cyclization of 3 to compete ef-
fectively with bimolecular oligomerization.
The suggestion that 4 is the possible active agent which
effectively “delivers” Bu3Sn- to dibromide 2 has precedent
in that Olah has found that CF3SiMe3 + fluoride ion
-
effectively “delivers” a CF3 anion in reactions with
carbonyl compounds without apparently passing thor-
ough a free CF3 anionic species (Scheme 8).13
-
Indeed we have found, in a control study, that treat-
ment of 2 with pregenerated Li+SnBu3 under high
-
dilution conditions does not lead to formation of signifi-
cant amounts of AF4. Mori also made the same type of
observation in his Diels-Alder study depicted in Scheme
5.11
Recognizing the need for an improved synthesis of AF4,
we have continued to pursue new reductive methodolo-
gies toward that end, with the result that we now wish
to report a new and improved method for synthesis of
AF4, using as a reductant the active species generated
in situ by reaction of (trimethylsilyl)tributyltin (TM-
STBT) with cesium fluoride (CsF). Furthermore, it has
been demonstrated that this method is scaleable and thus
should be able to be utilized to readily make kilogram
quantities of AF4.
It is believed that the success of our new procedure
derives from the limited solubility of CsF in the solvent
mixture at 70 °C and that the required slow but consis-
tent generation of the active reducing agent is controlled
by the slow, steady dissolution of the CsF which gener-
ates this active agent.
The actual nature of the debrominative process also
remains ambiguous as to whether there is a direct
formation of Bu3SnBr by a bimolecular reaction of the
active reducing agent with dibromide 2, or there is a
single electron transfer (SET) process involved between
4 (or 5) with 2. Consistent with this latter possibility is
the fact that the two best cosolvents with THF have been
found to be DMSO and HMPA, both recognized to be
excellent facilitators of SET reactions. The truth is,
though, that the emphasis in this work has been develop-
ment of a good synthesis, with little effort of a mecha-
nistic nature having been made.
The fact that the major competing reaction in this new
process is formation of polymer, with most of the remain-
ing material being accounted for as “reduction” products,
where one or both of the CF2Br groups of 2 have been
reduced to CF2H groups, provides good evidence for the
intermediacy of monomer 1 in the reaction. Another
piece of evidence along that line derives from an experi-
ment in which pure DMSO was used as the solvent at
70 °C, with gradual addition of CsF. In this case no AF4
was observed, but there was instead a high yield conver-
sion to the pentafluoro-p-xylene, 6, most likely via
fluoride interception of monomer 1, followed by abstrac-
tion of a proton from DMSO (Scheme 9).
The use of TMSTBT/CsF for the purpose of debromi-
nation has been previously reported in Mori’s study of
the generation of o-quinodimethanes for use in Diels-
Alder reactions (Scheme 5).11,12
The new method uses a unique approach to circumvent
the need for conventional high dilution technology. It
allows one to add in most of the ingredients together
rather quickly, while maintaining the necessary low
concentration of monomer via control of its rate of in situ
generation. As was the case for the earlier two methods,
2 is the preferred precursor to monomer 1 in the new
synthesis. In a typical reaction, 55 g (0.36 mol) of CsF,
105.5 g (0.28 mol) of TMSTBT, and 100 g (0.32 mol) of
dibromide 2 were added consecutively to a mixture of 8
L of THF and 2 L of DMSO at 78 °C. After stirring and
refluxing for 12 h, another 57.5 g (0.16 mol) of tinsilane
was added, followed by a third 52.5 g (0.14 mol) dose after
6 h more. After an additional 10 h of reflux, the reaction
was worked up to give 25 g of crude AF4 product, which
upon recrystallization from 10:1 CHCl3-hexane provided
22.7 g (40%) of >99% pure AF4 (Scheme 6). In subse-
quent development studies, the reaction has been able
to be scaled up to produce as much as 1 kg of AF4 in a
single run.
Exp er im en ta l Section
Anhydrous THF and DMSO were obtained from Fisher
Scientific, cesium fluoride (anhydrous, 99%) from CM Chemical
Products, Inc., and tri-n-butyltin chloride and trimethylsilyl
chloride from Aldrich Chemical Co.
The above-described conditions have been optimized
with respect to choice of solvent and reaction conditions.
(11) Sato, H.; Isono, N.; Okamura, K.; Date, T.; Mori, M. Tetrahedron
Lett. 1994, 35, 2035-2038.
(12) Sato, H.; Isono, N.; Miyoshi, I.; Mori, M. Tetrahedron 1996, 52,
8143-8158.
(13) Prakash, G. K. S.; Krishnamurti, R.; Olah, G. A. J . Am. Chem.
Soc. 1989, 111, 393-395.

7502 J . Org. Chem., Vol. 62, No. 21, 1997
Notes
Sch em e 7
formed). The mixture was filtered to remove LiCl, and the
filtrate was distilled to remove the THF. Two liters of diethyl
ether was then added and the mixture washed with water (2 ×
500 mL), after which the solution was dried over MgSO4. The
ether was removed by rotary evaporation, and the product was
distilled to obtain 2850 g of (trimethylsilyl)tributyltin (72.5%
yield), bp 88 °C/0.2mm.
Sch em e 8. Ola h ’s CF 3-Tr a n sfer Rea ction
Sch em e 9
Octa flu or o[2.2]p a r a cyclop h a n e (AF 4). Eight liters of
anhydrous THF, two liters of anhydrous DMSO, and 55.3 g (0.36
mol) of anhydrous CsF were added under an N2 atmosphere to
a dried, three-necked twelve liter round bottomed flask which
was equipped with a mechanical stirrer, a condenser, and an
N2 inlet, and heated by a heating mantel. The mixture was
stirred, the temperature allowed to rise to 70 °C, and the mixture
stirred for an additional 30 min. To this mixture was added
105.5 g (0.28 mol) of (trimethylsilyl)tributyltin (TMSTBT) and
the mixture stirred for 30 min at 70 °C. The solution was then
brought to reflux (78 °C), and 100 g (0.32 mol) of 1,4-bis(bro-
modifluoromethyl)benzene (2) was added all at once. The
mixture was then stirred for 12 h after which an additional 57.5
g (0.16 mol) of TMSTBT was added to the mixture, with a final
portion of TMSTBT (52.5 g, 0.14 mol) being added after an
additional 6 h. Upon stirring for an additional 10 h, it was
determined by 19F NMR that all of the dibromide 2 had been
consumed.
(Tr i-n -bu tylsta n n yl)lith iu m . To a dry, three-neck 12 L
flask equipped with mechanical stirrer, condenser, and addition
funnel, and maintained under argon, was added 3500 g (10.75
mol) of tri-n-butyltin chloride, followed by addition of 300 g (42.9
mol) of lithium chips and 20 g of sodium. The reaction mixture
was stirred for 3 days at ambient temperature, with the flask
becoming slightly warm to the touch. During this time the
mixture turned dark. After the mixture was cooled to 5 °C, 3 L
of THF was added using the addition funnel, the mixture turning
green. Stirring continued for 4-5 days at ambient temperature
after which the mixture was filtered through glass wool to
remove the excess lithium metal.
Seven liters of the THF were then removed by distillation at
atmospheric pressure. The resulting mixture was then cooled
to rt and allowed to stand for 12 h, after which the precipitate
was filtered off and the remaining THF was removed by rotary
evaporation. The residual liquid was distilled under vacuum,
with the distillate boiling above 70 °C (0.5 mm) being collected
and placed in a refrigerator for 12 h to yield 25 g of crystalline
AF4. The crude product was recrystallized from chloroform-
hexane (10:1) (chilled at 0-10 °C for 12 h in the refrigerator) to
yield 22.7 g of pure AF4 (40%); 1H NMR (CDCl3), δ 7.16 ppm
(s); 19F NMR (CDCl3), δ -118.75 ppm (s); 13C NMR (CDCl3), δ
129.82 ppm (t, J ) 3.4 Hz).
(Tr im eth ylsilyl)tr ibu tyltin .14 The above described solution
of (tributylstannyl)lithium was cooled to 5 °C with an ice bath
and, using the addition funnel, 1750 g (16.1 mol) of chlorotri-
methylsilane was added slowly, at a rate which maintains the
temperature below 5 °C (approximately 2 h). The ice bath was
then removed and the solution allowed to warm to room
temperature, continuing stirring for 3 h. At that point 50 mL
of n-propyl bromide was added and the mixture stirred an
additional 2 h (to remove any n-Bu3SnH which might have been
(14) Lahourne`re, J .-C.; Valade, J . J . Organomet. Chem. 1970, 22,
C3.
J O970927Q