ACS Catalysis
Research Article
and 8). Otherwise, the addition of an organic base like DBU
was detrimental, 2c being recovered in 40% (entry 9). On the
contrary, DTBPy had a marked positive effect, increasing the
yield to 80% (entry 10). Satisfyingly, we confirmed that this
base also increased the reaction yield of substrate 1a (64%,
entry 11). Next, we tested the effect of the chelating BPy
ligand, which has been successfully used in ligand-promoted
oxidative addition of Au(I) with aryldiazonium salts;12d
however, in our case, no reaction was observed (entries 12
and 13). Using DTBPy as a base, we also evaluated the
performance of (pCF3C6H4)3PAuCl and JhonphosAuCl
(entries 14 and 15). In the first case, the yield was 29%,
whereas, in the second, no reaction took place. Additionally,
we examined the reactivity of p-nitrobenzendiazonium chloride
and 2a in the presence of the cationic gold complex
Ph3PAuNTf2, but under these conditions, 2a was recovered
unchanged (entry 16). To summarize, the best results were
obtained using Ph3PAuCl(10 mol %) as a catalyst in the
presence of ascorbic acid (0.5 mol %) and DTBPy (1 equiv) as
a base. Phosphines bearing electron-withdrawing bulky groups
inhibited the reaction. We performed some control experi-
ments to verify the influence of the gold catalyst and ascorbic
acid. In the absence of the latter, the yield of the reaction of 2a
with p-nitrobenzendiazonium tetrafluoroborate dropped to
47% (entry 17), whereas, in the absence of Ph3PAuCl, no
reaction took place (entry 18). These results confirmed that
both ascorbic acid and Ph3PAuCl are essential for the reaction
to proceed in a good yield. Finally, to compare the efficiency of
the arylative cabocyclization of 2a under the optimized
conditions encountered by us (entry 10) versus previously
reported protocols, we studied the reactivity of 2a under
irradiation with blue LEDs and in the presence of [Ru(bpy)3]-
(PF6)2 (2.5 mol %) as a cocatalyst under irradiation with blue
LEDs. As shown in entries 19 and 20, the yield was lower in
both cases (47 and 32%, respectively), evidencing the major
performance of our optimized conditions.
Table 2. Influence of the Terminal Substituent on the
Alkynea
With the best reaction conditions encountered, we next
studied the effect of a terminal substituent onto an alkyne
(Table 2). Compound 3a bearing an n-butyl group gave 2H-
chromene 3c in a 58% yield. We were pleased by this result
since the arylative carbocyclization methods described so far
have not been applied to alkynes containing an alkyl group at
the terminal position.10 Compounds 4a and 5a with a phenyl
and a naphthyl group reacted with a 66% and 51% yield,
respectively. Compound 6a with a methyl group at the ortho
position reacted in a 59% yield, whereas compound 7a bearing
a p-methoxy group reacted in a 62% yield. The presence of a p-
nitro group at the aryl moiety completely inhibited the reaction
with 8c not being formed. This is probably due to the less
nucleophilic character of the alkyne. We inspected the effect of
electron-rich heteroarenes such as thiophene (9a) and
benzofurane (10a). In the first case, the yield was 64%,
whereas, in the second, the yield was 54%. Compound 11a
with a phenylsulfanyl group attached to the alkyne reacted in a
64% yield. It is important to note that the reaction was not
inhibited even in the presence of a pendant sulfur atom, which
may strongly coordinate gold. To end, we examined the
behavior of a terminal alkyne 12a (R = H). This compound
reacted to give the Sonogashira type coupling product 8a in a
42% yield; however, the product from arylative carbocycliza-
tion was not observed from arylative carbocyclization. The
results given in Table 2 show that the reaction is sensitive to
both electronic properties and steric hindrance of the group
attached at the terminal position of the alkyne. Thus, electron-
withdrawing groups can inhibit the reaction, and steric
hindrance decreases the yields.
Thereafter, we focused our attention on exploring the effect
of modifying the nucleophilic character of the phenol moiety
(Table 3). Electron-donating groups such as MeO-, t-Bu,
-OBn, and -Ph at the para position of a propargyl ether link
maintained yields in the range of 55−77%. As expected,
substrates containing electron-withdrawing groups reacted
more sluggishly and required an increase in the amount of
the diazonium salt to achieve better yields. Thus, substrates
containing halogens atoms such us p-Br (17a) and o-I (18a)
reacted in a 44 and 45% yield, respectively, using 2 equiv of 1b,
whereas the substrate containing p-CH3CO (19a) required 3.5
equiv of 1b to form 19c in a 34% yield. It is important to note
the good compatibility of this arylative carbocyclization
protocol with C-halogen bonds, enabling further functionaliza-
tion of halogenated 2H-chromenes. We also replaced the
oxygen link by an atom of sulfur (20a) and by a NTs (Ts
=Tosyl) group (21a). Arylative carbocyclization of substrate
20a took place in a 18% yield. We suspect that, in this case,
diazo coupling is a competing reaction because when substrate
20a was mixed with 1b, the solution turned its color, and a
second polar compound was observed by TLC at the end of
8970
ACS Catal. 2021, 11, 8968−8977