Angewandte
Communications
Chemie
Table 1: Electrochemical iododesilylation of trimethyl(phenyl)silane 1.
experiments without an iodonium source in the reaction or
with a substrate without a TMS group (toluene) suggested
that the methoxylation occurs mainly under participation of
iodonium ions and only when TMS-substituted substrates are
present.
Entry
Solvent
KI
[equiv]
Applied charge
Yield
[%]
To improve the yields and decrease the extent of
methoxylation, we investigated the influence of the iodide
loading and the solvent in more detail. To cover the
methoxylation side reaction as well, (para-tolyl)trimethylsi-
lane was used as a test substrate for this series of experiments.
To be able to address all important interactions between the
parameters, we used a design of experiment approach
applying a central composite plan.[9,10] Parameters being
optimized were a) the methanol concentration (1:1–9:1
acetonitrile/methanol (lower methanol concentrations were
excluded because they lead to poor conductivity and there-
fore need an additional electrolyte)), b) the KI stoichiometry
(1.0–1.2 equiv), and c) the applied charge (2.2–2.5 FmolÀ1) as
those three parameters seemed to have the biggest impact on
yields according to our previous optimization. The resulting
plan contained 18 reactions covering one- and two-factor and
quadratic interactions (Table 2).
[FmolÀ1
]
1[a]
2
3
4
5
6
7
8
9
10
11
12[b]
13[c]
14[d]
15[e]
MeCN
MeCN
(MeOCH2)2
CH2Cl2
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.1
1.1
1.2
1.2
1.0
1.1
1.1
1.1
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.2
2.4
2.4
2.6
2.0
2.4
2.4
2.4
0
40
10
37
10
64
80
84
97
86
94
48
MeOH
MeCN/MeOH (1:1)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
MeCN/MeOH (7:3)
[f]
–
–
87
[f]
Unless otherwise stated, a divided cell with platinum electrodes with
0.5 mL conc. H2SO4 in the cathode compartment was used. The yields
were determined by GC analysis using mesitylene as the internal
standard (for experimental details, see the Supporting Information).
[a] An undivided cell was used. [b] Graphite electrodes were used.
[c] Instead of H2SO4, acetic acid (1 mL) was used. [d] Instead of H2SO4,
methanesulfonic acid (0.5 mL) was used. [e] Instead of H2SO4, LiClO4
(3 mmol, 0.3m) was used. [f] No conductivity observed.
Table 2: Optimisation by design of experiments.[a]
tions were performed in H-type divided cells. Disappoint-
ingly, the best yield of 2 was quite low (40%) even when
acetonitrile was used as the solvent. When CH2Cl2, methanol,
or dimethoxyethane were used, the yields amounted only to
10–37% (entry 1–5). The low yields were attributed to the
formation of large quantities of the protodesilylation side
product benzene. Only a solvent mixture of acetonitrile/
methanol (1:1) led to an improved yield of 64%, mainly
owing to a significant decrease in the protodesilylation side
reaction.[8] A slight decrease in the methanol content to a 7:3
mixture further increased the yield to 80%. A final improve-
ment was achieved by slightly increasing the KI loading to
1.1 equiv and adjusting the applied charge to 2.4 FmolÀ1,
which resulted in an almost quantitative yield of 97%.
However, a further increase in the amount of the iodine
source or the applied charge (entries 10 and 11) led to
decreases in yield. Unfortunately, changing the electrodes
from platinum plates to graphite decreased the yield to 48%
(entry 12). Finally, the acid or supporting electrolyte that was
used in the cathode compartment to lower the cell potential
was changed from sulfuric acid to acetic acid, methanesul-
fonic acid, or lithium perchlorate, but all of them led to
reduced yields (entry 13–15).
Factor
p Value
MeCN/MeOH ratio
0.0000
0.0001
0.0003
0.0031
(applied electricity)2
amount of KI
applied electricityꢂMeCN/MeOH ratio
[a] All reactions were carried out on 0.5 mmol scale using a divided cell
and platinum electrodes. The yields were determined by GC-FID analysis
using mesitylene as the internal standard (see the Supporting Informa-
tion for details). Numbers in grey indicate the optimal values for the
three parameters and the corresponding yield (with its confidence
interval in brackets). The p values indicate the probability of whether the
respective effect can be explained by the null hypothesis (effects below
0.01 are regarded as significant).[9e]
Four important interactions returning a model with an
R2 value of 0.95 (which confirms that there are only slight
differences between the observed and predicted values) were
found: a) the methanol concentration, b) the quadratic term
of the applied electricity, c) a two-factor interaction between
methanol concentration and charge consumption, and d) the
KI equivalents. According to the obtained model, the optimal
values for the three parameters are a methanol concentration
of 10 vol% in acetonitrile, 1.0 equiv of KI, and an applied
charge of 2.3 FmolÀ1.
To our disappointment, when we applied our optimized
reaction conditions to substrates other than trimethylsilyl
benzene, the yields were only moderate (68–71%; see the
Supporting Information). The main reason was determined
by profound GC/MS analysis of the reaction products:
Significant quantities of the starting materials were found to
be methoxylated at the benzylic position, if present. Control
These observations are in line with the previously
obtained results as a minimal iodide loading and a minimal
2
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
These are not the final page numbers!