Organometallics
Article
clearly rule out the radical pathway A for reductive elimination
from complex 3b.
Second, to probe if the reductive elimination from complex
3b occurs via a ligand dissociation pathway, we studied the
reductive elimination of complex 3b at 85 °C in the presence
of 5.0 equiv of an “ate”-type nucleophile lithium phenyl n-butyl
pinacol boronate (Figure 6, eq 5). In a separate experiment, we
determined that transferring a phenyl group from the boronate
to complex [(Phen)Ag(CF3)3] 2 occurred readily at room
temperature to generate ionic Ag(III) complex [Ag(Ar)-
(CF3)3]− in 32% yield in CDCl3 and 80% yield in DMSO,
respectively (Figure 6, eq 6). Therefore, if the reaction
proceeds via a reversible ligand dissociation, a neutral three-
coordinated Ag(III) intermediate [Ag(Ar)(CF3)2] 8 or
[Ag(CF3)3] 8′ would form. A fast transfer of the phenyl
group from phenyl borate to coordinationally unsaturated
intermediate 8 or 8′, followed by reductive elimination from
these complexes, would then produce 2-fluoro-4-methyl-
benezene 4b or trifluoromethylbenzene. Experimentally, the
formation of 2-fluoro-4-methyl-benezene 4b or trifluorome-
thylbenzene was not observed as determined by 19F NMR and
GC/MS spectroscopies, indicating that the dissociation
pathway B for the reductive elimination from complex 3b is
unlikely.
To gain more support to disfavor the ligand dissociation
pathway, we studied the kinetics of the reductive elimination
from complex 3b, and it was found that the reaction followed
first-order kinetics. Eyring analysis of the reaction rate
constants at various temperature versus 1/T gave activation
parameters of ΔH⧧ = 27.81 1.97 kcal/mol and ΔS⧧ = 3.34
0.37 e.u. (Figure 7A). A small entropy change is against a rate-
determining ligand dissociation pathway, which is consistent
with our experiment result. Furthermore, DFT calculations
showed that the barrier for dissociation of a trifluoromethyl
anion from complex 3b to generate a neutral three-coordinated
silver(III) species 8 is 44.6 kcal/mol, suggesting that
dissociating a trifluoromethyl anion from complex 3b is a
rather difficult process (Figure 5). Thus, dissociation pathway
B or C for the Ar−CF3 bond-forming reductive elimination
from complex 3b was disfavored.
Figure 4. Possible pathways for reductive elimination from
organosilver(III) complexes [nBu4N]+[Ag(Ar)(CF3)3]− 3b.
transition state to give Ar−CF3 and [Ag(CF3)2]− (pathway E
in Figure 4).
First, to determine whether thermal decomposition of
complexes 3a−e proceeds via radical pathway A, we
investigated the reaction [nBu4N]+[Ag(2-fluoro-4-MeC6H4)-
(CF3)3]− 3b in the absence or presence of a radical scavenger
or single-electron transfer (SET) inhibitor. It was found that
reactions in the presence of 1.0 equiv of TEMPO or 1,4-
dinitrobenzene occurred after 3.0 h at 85 °C to give compound
4b in 92% and 89%, respectively (eq 4). These results showed
Third, to evaluate whether solvent was associated in the
C(sp2)-CF3 reductive elimination of complexes 3a−e, we
studied the reductive elimination of complex 3b in five
different solvents with different polarities and coordinating
abilities (CDCl3, ClCH2CH2Cl, DMF, DMSO, or PhCN)
(Figure 7b). Studies showed that reactions in all five solvents
occurred in full conversion after 1.0−12 h at 85 °C to give
compound 4b in 85−95% yields, respectively. If the reductive
elimination reaction occurs via a rate-limiting solvent
association pathway, reactions in coordinating solvents should
be faster than those in noncoordinating solvents. As shown in
Figure 7B, our kinetic studies of these reactions revealed that
reductive elimination in less polar solvents CDCl3 or
ClCH2CH2Cl (kobs = 5.05 × 10−4 s−1 and 4.19 × 10−4 s−1)
were faster than those in more polar solvent DMF or DMSO
(kobs = 2.36 × 10−4 s−1 and 2.96 × 10−4 s−1). In addition,
reaction in coordinated solvent PhCN (kobs = 1.97 × 10−4 s−1)
was slower than that in noncoordinating solvents such as
CDCl3 or ClCH2CH2Cl. These results render the solvent
association pathway D unlikely.
that there is no significant difference for reactions in the
presence or absence of the radical or SET inhibitors, suggesting
that the free radical is unlikely to be involved in the reductive
elimination from complex 3b.
To gain more evidence for the noninvolvement of free
radical in this process, we monitored the thermolysis process of
complex 3b by periodically drawing an aliquot of the reaction
mixture and then measured by electron paramagnetic
resonance (EPR) spectroscopy. Not surprisingly, no signal
related to radical was observed. In addition, DFT calculation
showed that with a bond strength of 35.5 kcal/mol, the Ag−
CF3 bond in complex 3b is relatively strong and difficult to
undergo homolytic cleavage. Even if the homolytic cleavage of
the Ag−CF3 bond in complex 3b occurs, the new formed
intermediate [AgII(Ar)(CF3)2]− 6 needs to overcome an
additional barrier of 11.9 kcal/mol to form a three-member
ring transition state 7-ts that leads to reductive elimination
(Figure 5). Thus, the overall activation free energy barrier
(ΔG‡ = 47.4 kcal/mol) for pathway A is much higher than that
of the concerted bond-forming pathway D (ΔG⧧ = 26.0 kcal/
mol) (Figure 5). These experimental and theoretical results
Mechanistic Proposal for Reductive Elimination from
Complex 3. The above experimental results, including
reactions in the radical or SET inhibitors, EPR studies, effect
1715
Organometallics 2021, 40, 1713−1718