C.M. Timperley et al. / Journal of Fluorine Chemistry 96 (1999) 95±100
99
Further support for this pathway was provided by
3. Experimental details
comparing reactions of dimethyl and diethyl (2,2,2-
trifluoroethyl) phosphates with TMSBr. Monitoring by
GC±MS established that the reaction of the dimethyl
phosphate took place more rapidly than that of the diethyl
phosphate, reflecting the greater ease of dealkylation of
methoxy versus ethoxy groups . A second attack by alkoxy
oxygen of product 13 on TMSBr gave rise to small
amounts of bis-silyl product 14.
All reagents were of commercial quality. Anhydrous
solvents were used for reactions: THF was distilled from
sodium/benzophenone. NMR spectra were obtained on a
1
Jeol Lambda 500 instrument (operating at 500 MHz for H,
125 MHz for 13C, 470 MHz for 19F, and 202 MHz for 31
P
spectra) as solutions in CDCl , with internal reference
3
3
SiMe for H, external CFCl for 19F and external (MeO)3P
1
4
3
The alternative mechanism involving initial attack of
phosphoryl oxygen on TMSBr to give phosphonium inter-
mediate 16 can be ruled out. In this case, the bromide ion
would be expected to attack the most electropositive carbon
atom, namely the methylene carbon of the tri¯uoroethyl
group, which would give dialkyl trimethylsilyl phosphate 17
with loss of 1-bromo-2,2,2-tri¯uoroethane. No evidence for
either of these products was observed by analysis of reaction
mixtures by GC±MS.
(ꢁ 140 ppm) for 31P spectra. Data in Tables 2 and 3 are
recorded as follows: chemical shifts in ppm from reference
on the ꢁ scale, multiplicity (ssinglet, ddoublet, ttriplet,
qquartet, mmultiplet and sepseptet), integration, cou-
pling constant (Hz) and assignment. IR spectra were
recorded as liquid ®lms on a Nicolet SP210 instrument
using Omnic software. Reaction mixtures were monitored
by gas chromatography±mass spectrometry (GC±MS) using
a Finnigan MAT GCQ instrument (chemical ionisation
using methane as reagent gas). Analysis of pure products
by high resolution mass spectrometry (HRMS) used a
Micromass Autospec SQ Double Focusing Magnetic Sector
instrument. Mode: ve ion electron impact, magnet scan m/
z 400 to 100 (seconds/decade), resolution 2900. Inlet:
septum (1608C), 0.2 ml introduced. Source conditions: tem-
perature 2008C, electron energy 70 eV, and accelerating
voltage 8000 V. All reaction mixtures were stirred magne-
tically.
The mechanism of transesteri®cation of un¯uorinated
alkyl esters of phosphorus with TMSBr involves the for-
mation of phosphonium intermediates analogous to 16.
Silylation is a two-step process involving a fast reversible
formation of phosphonium salt which decomposes in the
rate-limiting step through dealkylation by bromide ion
3
.1. Synthesis of dialkyl fluoroalkyl phosphates
[
15,17]. Our ®ndings suggest that an alternative mechanism
Method A: The ¯uoroalcohol (0.2 mol) and triethylamine
involving an oxonium intermediate operates in the transes-
teri®cation of dialkyl ¯uoroalkyl phosphates with bromo-
trimethylsilane.
(
to the dialkyl phosphite (0.2 mol) and CCl (0.2 mol) cooled
0.2 mol) in CHCl (50 ml) were added dropwise over 2±3 h
3
4
to 0±58C. When the addition was complete, a white pre-
cipitate of Et NÁHCl had formed. The mixture was stirred
2
.3. Reactivities of fluorinated and unfluorinated
phosphates towards chlorinating agents
3
for 2 h at 0±58C and allowed to stand for 16 h at room
temperature. The mixture was washed with water
(
®
2Â25 ml), the organic layers combined, dried (MgSO ),
The comparative reactivities of triethyl phosphate and
4
ltered and the solvent removed to leave an oil. This was
diethyl (2,2,2-tri¯uoroethyl) phosphate towards various
chlorinating agents were examined. Neither compound
reacted with oxalyl or thionyl chloride in chloroform under
prolonged re¯ux. Triethyl phosphate was found to react very
slowly with phosphorus oxychloride in re¯uxing chloro-
form to give a low conversion to diethyl chlorophosphate
distilled, initially at 12±15 mm Hg with gentle heating to
remove any unreacted alcohol. The temperature was then
increased and the vacuum adjusted to between 0.5±
1
(
2 mm Hg depending on the boiling point of the product
Table 1).
Method B: A mixture of the ¯uoroalcohol (0.1 mol) and
triethylamine (0.1 mol) was added dropwise to the dialkyl
(
5% yield after 8 h re¯ux, as determined by GC±MS).
chlorophosphate (0.1 mol) in Et O (100 ml) cooled to 0±
2
5
8C (or at room temperature for compounds 4 and 12). After
addition the solution was stirred for 4 h and ®ltered. The
ltrate was concentrated to an oil which was distilled
Diethyl (2,2,2-tri¯uoroethyl) phosphate was less reactive
than its un¯uorinated counterpart and did not react at all
with phosphorus oxychloride under identical conditions.
®
according to Method A.
Method C: The ¯uoroalcohol (0.04 mol) in THF (30 ml)
was added dropwise to a suspension of NaH (0.048 mol) in
THF (30 ml) at room temperature (exothermic reaction).
After addition the mixture was cooled to 0±58C and to it was
added dropwise a solution of dialkyl chlorophosphate
3
Reactivity in the dealkylation process depends upon the size of the
alkyl group. Competitive experiments on the dealkylation of dimethyl,
diethyl and diisopropyl acylphosphonates with TMSBr indicated the
relative reactivities to be 1:0.25:0.04 [16].