N. Santschi et al. / Journal of Fluorine Chemistry 135 (2012) 83–86
85
Scheme 3. Proposed initial thermal decomposition of the product is followed by
autocatalytic fluoride-promoted decomposition yielding the corresponding
fluorophosphates.
values of ꢁ0.09, ꢁ0.08 and ꢁ0.07, respectively, and would thus be
expected to have similar relative rates. However, neopentyl is
found to react significantly faster (Table 1). The proposed release of
steric strain energy (vide supra) is expected to be larger for
neopentyl (ES,corr = ꢁ1.44) than for iso-propyl and cyclohexyl
(ES,corr = ꢁ0.40 and ꢁ0.50, respectively) and is thus compensating
to a greater extent for lessened acidity due to inductive effects.
Analysis of the 19F and 31P NMR towards the end of the reaction
of dimethyl hydrogen phosphate shows formation of substantial
amounts of dimethyl fluorophosphate.7 These are proposed to
initially stem from thermal decomposition of the product formed
to yield carbonyl fluoride. The latter reacts with water liberating
CO2 and two equiv of HF, which lead to further autocatalytic
decomposition. Competition between formation and decomposi-
tion of product thus allows explaining the low yields observed and
the fact that the reagent is apparently consumed faster than
product is formed (Fig. 1) (Scheme 3).
Scheme 2. Proposed three-step mechanism accounting for the positive slope found
(Fig. 2) and the non-linearity by postulating an initial proton-transfer equilibrium
resulting in the development of a negative charge on the key oxygen (highlighted).
to be only slightly sensitive towards inductive effects. Moreover,
the positive slope points to the formation of a negative charge
during the RDS. With a steric sensitivity factor of
d
= 0.095 ꢃ 0.008
the reaction is found to be slightly accelerated if larger substituents
are used. Based on these observations the following mechanistic
interpretation is proposed (Scheme 2).
The initial proton-transfer equilibrium constituting the RDS is
followed by a nucleophilic attack of the substrate on the iodonium
core to yield the final intermediate that undergoes reductive
elimination to furnish the desired dialkyl trifluoromethyl phos-
phates. Variation of the inductive properties of the substituent R
significantly influences the acid strength of the hydrogen
phosphates and thus the proposed initial equilibrium (Scheme
2, step 1) [4a]. The iso-propyl and cyclohexyl derivatives are too
weak acids to protonate reagent 1 to a significant extent, therefore
achieving only a low degree of activation and thus leading to a
small relative rate. The other acids, however, are deprotonated to a
sufficiently large degree such that destabilizing or stabilizing
effects of the negative charge on the oxygen atom become
pronounced in a linear fashion, i.e. a consistent negative charge is
obtained during the rate-determining step. Furthermore, there
seems to be no competition between activation of reagent 1 and
nucleophilicity of the resulting anion—a situation which would
lead to ‘‘downward’’ bending for electronegative substituents
3. Conclusions
The electrophilic trifluoromethylation of hydrogen phosphates
was probed by means of relative rates determined through initial
rate experiments carried out under pseudo first order conditions.
Through a Taft analysis of the data obtained it could be shown that
a negative charge is building up during the rate-determining step.
As a
r* value <1 was obtained it could be concluded that the
reaction is only mildly sensitive towards inductive effects.
Furthermore, the observed acceleration due to increased steric
bulk of the substrate’s substituent is plausibly accounted for by a
subtle electronic effect. However, strong competition between
formation of the desired product and its autocatalytic decomposi-
tion to the corresponding dialkyl fluorophosphates undermines
the utility of the reaction, so far. Further studies towards reaction
optimization and isolation, based on the mechanistic investigation
presented herein, are currently on-going and will be presented in
due course.
(
sꢀcorr > 0) as the nucleophilic attack on the iodonium core would
become the rate-determining step (Scheme 2, step 2).
Deprotonation of the hydrogen phosphate at the stage of the
initial equilibrium (Scheme 2, step 1) is bound to change the
electronic environment of the central phosphorus atom. With a
reported sinductive value of 0.29 for the hydroxy substituent and a
sinductive of ꢁ0.16 for the oxide anion, one expects the oxygen of
interest (Scheme 2, highlighted) to change from a very electro-
4. Experimental
4.1. General procedure 1: synthesis of alkyl chlorophosphates
Anhydrous CuCl2 (2.1 equiv.) was suspended in THF, then
cooled to 0 8C. To the resulting green/brown suspension was added
the corresponding alkyl phosphonate (1 equiv.) in THF in one
portion under vigorous stirring. The cooling bath was removed and
the mixture was allowed to warm to ambient temperature over
25 min. The mixture was then concentrated under reduced
pressure and the residual light brown solid was extracted twice
with 100 mL pentane and filtered over celite. Concentration of the
clear pentane solution under reduced pressure followed by
negative situation to
a less electron-withdrawing one [10].
According to the Walsh–Bent rule, this will lead to an increased
s-character of the phosphorus oxide anion bond and hence result in
a pyramidalization-type deformation of the molecule itself [11].
Therefore, the angles between the alkoxide substituents and the
oxide anion are expected to increase and thus decrease the steric
strain imposed by the increased radius of the oxide anion.
Although this simplistic rationalization neglects structural
changes due to resonance effects it may still account for the
observed reactivity. The iso-propyl, cyclohexyl and neopentyl
substituent have similar inductive properties with calculated scꢀorr
7
Spectral data dF (376.5 MHz, CDCl3) ꢁ85.3 (d, J = 978.4 Hz), dP (162.0, CDCl3)
ꢁ5.9 (d, J = 978.3 Hz) are in accordance with resonances reported in: T. Sierakowski,
J.J. Kiddle, Tetrahedron Lett. 46 (2005) 2215.