.
Angewandte
Communications
disappointing results (Table 1, entries 6–8). Additionally,
several different copper sources (oxidation state I or II)
were screened but only low to modest conversions were
detected even in the most favorable cases (Table 1, entries 9–
11). It is worth noting that the blank experiments showed that
the presence of the copper catalyst is necessary in our
protocol (Table 1, entry 5).
We then examined, either in tBuOH or in 1,4-dioxane, the
influence of the ligand upon the course of the reaction
(Table 2). Diamine (L5, L6) and b-diketone (L7, L8, L9)
isolation (Table 3). Although the L4/tBuOH system used for
4-iodotoluene (Table 2) was applicable to all the substrates
tested, improved conversions were obtained using the L1/1,4-
dioxane couple highlighted in our preliminary studies
(Table 3, entries 3, 10, 13, 14, and 16–18). The ligand L1
(cheaper than L4) was also found to be efficient in combi-
nation with tBuOH for some examples (Table 3, entries 5, 7,
and 9). In the case of ortho-substituted aryl iodides, which are
traditionally poor substrates in copper-catalyzed arylation of
C, N, and O nucleophiles, low to fair amounts of the a-
arylated products were obtained.
Electron-poor aryl iodides were then considered
Table 2: Copper-catalyzed a-arylation of deoxybenzoin with 4-iodotoluene in
1,4-dioxane or tBuOH using the ligands L.[a]
and reacted with several deoxybenzoin derivatives
(Table 4, entries 1–14). Note that for a-arylation
catalytic systems based on other metals than copper,
this case has been less explored than those of
electron-rich aryl halides. Applying the conditions
of Table 3 led to poor yields of the coupling
L
Yield [%][b] in
L
Yield [%][b] in
products. For example, only 37% of the a-arylated
compound 23 (see Table 4, entry 1 for structure) was
obtained using the standard system CuI/L4/tBuOH
with 4-iodofluorobenzene (the main by-product was
the result of the homocoupling of deoxybenzoin).
However, a simple optimization led to an increase in
the yield to 52% by using the b-acetylcyclohexanone
ligand (L8) in tBuOH at 1108C (reaction conditions
previously employed for the a-arylation from 4-
iodotoluene). Finally the selectivity was improved
by lowering the reaction temperature to 708C, thus
producing 23 in 75% yield (Table 4, entry 1). By
using these reaction conditions, 15 differently sub-
stituted arylated deoxybenzoine derivatives were
obtained (Table 4) with yields ranging from low
(Table 4, entry 13) to excellent (Table 4, entry 14).
In some cases, the final compounds feature one or
two halides (Br, Cl, F) on the aromatic rings,
available for further functionalization at the corre-
sponding positions.
1,4-
dioxane
tBuOH
1,4-
dioxane
tBuOH
–
8
6
L5 29
5
8
L1 70
52
L6 23
L7 20
L2 27
46
17
L3
3
47
84
L8 24
L9 11
58
8
L4 68
[a] Reactions performed on 1 mmol scale (ketone/ArI=1.5:1). [b] Yields deter-
mined by 1H NMR spectroscopy using 1,3-dimethoxybenzene as an internal
standard.
ligands were ineffective, except for b-acetylcyclohexanone
(L8) in tBuOH (58% yield of 1). Finally, the 1,10-phenantro-
line-type ligands (L1, L2, L3, L4) offered the best solution,
with the use of bathophenantroline (L4) in tBuOH affording
1 in very good yield (84%). These reaction conditions were
used with bromobenzene but the cross-coupling failed (8% of
product and no reaction with PhCl). As it was the case for the
copper source, the presence of a ligand was necessary to
obtain satisfying results, as attested by the yields obtained
from PhI (6–8%) in its absence (Table 2). In addition, it
proved difficult to establish a relationship between the
structure or the solubility of the ligand and the reactivity
observed, even for the same structural family.
Next, the scope of the reaction with respect to various aryl
iodides and deoxybenzoin derivatives was investigated. Two
scenarios were observed, depending on the nature of the aryl
iodide substituents.
In the case of electron-rich aryl iodides bearing one or two
electron-donating groups (Me or OMe), a simple tuning of
the copper catalytic system allowed us to obtain numerous a-
arylated deoxybenzoins in good to excellent yields upon
The scope of this methodology was demonstrated by its
simple and efficient application to the synthesis of Tamoxifen
(TAM). Marketed under the trade name Nolvadex, this
molecule has for a long time been the worldꢀs most commonly
administrated drug for management of breast cancer.[15b,16]
We considered two new and original routes for the total
synthesis of this molecule, which belongs to the triphenyl-
ethylene class of compounds. In the first pathway, the
synthesis starts with the copper-catalyzed a-arylation of
deoxybenzoin with 4-Iodophenol (Scheme 1, Route A). The
latter is a suitable substrate as evidenced by the good yield
obtained for the a-arylated deoxybenzoin derivative 37. For
this step the most efficient catalytic system was CuI/L8.
tBuOH was preferred over 1,4-dioxane because of the greater
solubility of the phenolic reagent in alcohols. Compound 37
then underwent an alkylation/elimination/isomerization[17]
sequence and the resulting intermediate 38 was directly
reacted with 2-chloro-N,N-dimethylethylamine to give
tamoxifen in an overall yield of 70% (over 4 steps) upon
isolation.
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12815 –12819