ACS Combinatorial Science
Letter
Applying this approach to the Glaser−Hay reaction is even more
advantageous as there is no requisite for a haloalkyne to aid in the
directing of chemoselectivity, and asymmetric diynes can be
accessed directly from terminal alkyne precursors. One previous
Glaser−Hay coupling has been attempted on the solid support;
however, its aim was to provide homocoupling of the
immobilized alkyne, and not to address chemoselectivity issues.11
Because of the abundance of Glaser−Hay reaction conditions
present in the literature, reaction conditions were examined in
solution prior to optimization on the solid support. Homocou-
pling of phenylacetylene (6) was performed with different
catalytic systems based on literature precedence.12 The most
promising catalyst systems included CuCl/TMEDA, CuI/
TMEDA, and CuI/DIEPA/N-bromosuccinimide, and all were
examined toward the solution phase Glaser−Hay coupling of 6.
Overall, the CuI/TMEDA catalyst (12 h, rt) led to the highest
yields of the phenylacetylene homodimer, and was employed in
further couplings.
especially unfortunate, as they negatively impact the yield of
desired product and require further purification steps. This is
especially problematic as the use of the solid support in these
coupling reactions was intended to eliminate both of these issues.
As such, we sought to further optimize the reaction conditions to
successfully drive the reaction to completion without the
production of the homodimer impurities.
It was speculated that decreasing the initial loading of 2 on the
solid support could efficiently prevent the undesirable
dimerization on the functionalized resin. Our initial preparation
of 1 consisted of swelling the resin in CH2Cl2 for 15 min,
followed by the addition of propargyl alcohol (10 equiv) and
triethylamine (10 equiv). The reaction was stirred for 16 h at
room temperature to achieve maximum loading of the starting
material. After the discovery of the cross-linked homodimer
product, we attempted to generate lower loaded resins by
decreasing the amount of reagents and immobilization time
employed in the immobilization protocol.
With optimized solution-phase conditions established, we
transitioned the optimization to the solid support. In order to
assess the feasibility of the approach, various immobilization
strategies of propargyl alcohol (2) were investigated with
carboxy, trityl chloride, and bromo Wang resins; however,
ultimately a trityl chloride derivatized resin was selected because
of the mild immobilization/cleavage conditions and the
reproducibility of propargyl alcohol loading. Initially 2 (10
equiv) was reacted at room temperature with trityl choloride
resin in the presence of triethylamine (10 equiv) for 16 h to
afford propargyl alcohol immobilized resin (1). The coupling
conditions were then optimized via the reaction of 1 with
phenylacetylene (6), varying solvent, temperature, and resin
loading.
Because of the hydrophobic nature of the polystyrene core,
solvent selection is important to reaction optimization due to the
resin’s propensity to swell or contract as a result of solvent
polarity. Consequently, we investigated toluene, THF, DCM,
and acetonitrile as viable solvents for the solid-supported
reaction. Overall, THF consistently generated the highest yields
of coupled alkyne products, while significant amounts of
unreacted 2 remained in other solvents after 12 h at room
temperature.
While THF afforded the highest yields, some starting material
still remained, indicating the necessity for further optimization.
Consequently, temperature was varied, to elucidate optimal
conditions. A range of temperatures was examined, including 30,
60, and 80 °C, with the 60 °C conditions yielding the most
promising results. The reactions at 30 °C still contained
unreacted starting material, indicating that the temperature was
not sufficient to drive the reaction to completion. Meanwhile, the
reactions stirred at 80 °C demonstrated complete conversion,
but afforded lower yields. This noticeable decrease in the
formation of diyne product most likely resulted from the
overheating and destruction of some of the polystyrene resin.
Some of the copper catalyst was also found to decompose within
the reaction vessel, suggesting that the temperature may
inactivate the catalyst, thereby reducing the efficiency of the
reaction. Reaction of 1 with 6 for 12 h at 60 °C afforded the ideal
conditions for virtually quantitative conversion of propargyl
alcohol precursor.
To assess the effects of resin loading on the reaction, several
resins were prepared. Trityl chloride resin was again reacted with
10 equiv of propargyl alcohol in the presence of 10 equiv of
triethylamine. The reaction was then stirred at 40 °C for 16 h,
with the aim of achieving a higher loaded resin, ∼1.8 mmol/g. An
identical reaction was set up but heated to 40 °C for only 2 h to
generate a medium loaded resin, ∼1.0 mmol/g. Additionally, the
amount of propargyl alcohol was decresed to afford even lower
loaded resin ∼0.6 mmol/g. Similar immobilizations were also
performed on a 2% DVB cross-linked resins to afford similarly
loaded resins that provided an increased degree of resin rigidity
to discourage interactions of immobilized alkynes. After a
sufficient work up, a portion of each version of the resin was
cleaved in 2% TFA and analysis of loading by GC/MS and mass
recovery of propargyl alcohol.
The resins were then subjected to coupling conditions with 6
at 60 °C for 16 h. The level of cross-linked DVB within the resin,
1% or 2%, showed little effect on the suppression of homodimer.
This result indicates that the concentration of loading plays a
larger role in the formation of homodimer than resin rigidity.
After the necessary work up conditions, it was established that
resin loaded above 0.7 mmol/g of propargyl alcohol forms
undesired homodimer during the reaction conditions. Resin
loaded at a concentration less than 0.7 mmol/g successfully
eliminates homodimer formation while still generating desired
heterocoupled products.
Having elucidated optimized solid-supported Glaser−Hay
conditions, the scope of the methodology was assessed via the
preparation of a diverse library of diynes. Toward this end, two
additional alkynes, 3 and 4, were immobilized on the trityl-
chloride resin at loadings of ∼0.6 mmol/g. These were then
reacted in THF with a CuI/TMEDA catalyst and a diverse set of
soluble alkynes to assemble a small library of conjugated diynes.
After 16 h of stirring at 60 °C, the resin reaction was filtered then
washed with alternating volumes of CH2Cl2 and MeOH to
remove unreacted reagents and other impurities. The resin was
then cleaved with 2% TFA, the solvent was removed in vacuo,
1
and the product was analyzed by TLC, H NMR, and GC/MS
(Table 1).
The solid supported Glaser−Hay coupling tolerated a variety
of functionalities, including aromatic rings, silyl groups, free
amines, alcohols, and alkyl chains. Moreover, the reaction with
both 5 and 6 was scaled up 10-fold (500 mg of resin) with
comparable yields, suggesting that larger quantities of product
could be obtained if desired. Reactions not proceeding to
Upon further investigation, it was discovered that despite
complete conversion, there was also dimerization of the
immobilized alkyne, to afford an undesired homodimer product
(12). The occurrence of the starting material and homodimer are
165
dx.doi.org/10.1021/co500018k | ACS Comb. Sci. 2014, 16, 164−167