Journal of the American Chemical Society
Communication
Au4Cu4/CNT (84.54 eV). However, the Cu 2P3/2 band
(Figure 1C) in the recovered Au4Cu4/CNT (933.48 eV)
exhibits a slight positive shift compared with the fresh Au4Cu4/
CNT (932.98 eV), which was speculated to be due to PhC
CH coordinated to a Cu atom of the Au4Cu4 NC to form a
feedback π* bond, resulting in charge transfer.27
showed that there was no peak at 7.66 ppm (Figure S8),
indicating that the D ending did not fall off during the whole
reaction. All these results fully demonstrate that PhCCH is
activated through π-complexation with Au4Cu4 rather than
deprotonation. The known monometallic NCs
[ A u 1 1 ( P P h 3 ) 8 C l 2 ] C l ( i n s h o r t , A u 1 1
)
a n d
The uniform Au4Cu4 NC was used to study the mechanism,
and it was found that the catalytic process was different from
the dehydrogenation mechanism.8,9 First, the UV−vis spectra
exhibited a blue shift (Figure S5) when PhCCH was added
to Au4Cu4 NC, while the UV−vis spectra of the Au4Cu4 +
PhCH2N3 mixture remained unchanged. A similar blue shift
was detected when PhCCH and PhCH2N3 were added
together. Thus, the first step of the mechanism is activation of
PhCCH by Au4Cu4. Then, the mixed solution (Au4Cu4 +
PhCCH) was characterized. In the FT-IR spectra (Figure
2A), the characteristic absorption peak of 3293 cm−1 for H−
CCPh was observed, indicating that the H atom of PhC
CH had not been removed during the activation process.28
Meanwhile, the characteristic peak of the CC bond was
shifted from 2109 cm−1 to 2104 cm−1, which was probably due
to feedback alkyne π* bond coordination to Cu, so that the
CC bond becomes longer and its vibration frequency lower.
Furthermore, the Cu 2P3/2 data in the recovered Au4Cu4/CNT
also proved the existence of the feedback π* bond. From the
1H NMR spectrum (Figure 2B), the peak of H-CCPh (3.14
ppm) still existed in the Au4Cu4 + PhCCH mixture,
accompanied by a certain displacement. In the meantime, the
ratio of benzene-ring hydrogen (a) to alkynyl hydrogen (b)
remained 2:1, indicating that PhCCH does not lose its
terminal hydrogen. We conjecture that PhCCH is first
coordinated to Au4Cu4 NC with feedback π* bond formation
of a [Au4Cu4-π-(CHCPh)] intermediate (referred to as
Au4Cu4*) rather than deprotonation of PhCCH. The pure
Au4Cu4* point appears below the Au4Cu4 point from the TCL
board (Figure 2C). The MALDI-TOF data of the Au4Cu4*
sample showed two ion peaks located at 2581.28 and 2647.33
Da, respectively. We speculate that the two peaks are the two
fragment peaks due to instability of Au4Cu4* under the high
energy of mass spectrometry ionization. For example, 2581.28
Da (calculation: [Au4Cu4 + HCCPh − Sadm − H]+) was
Au4Cu4* that lost one Sadm ligand and 2647.33 Da
(calculation: [Au4Cu4]+) was Au4Cu4* that lost PhCCH.
The other two alkynes (HCCPh-4-Cl and HCCPh-4-Me)
were selected to coordinate Au4Cu4 in order to further verify
the Au4Cu4* structure, respectively. Similar results were
obtained by MALDI-TOF analysis. For CHCPh-4-Cl and
CHCPh-4-Me, there are two ions peaks located at 2647.33
and 2615.36 Da (calculation: [Au4Cu4 + CHCPh-4-Cl −
Sadm − H]+) as well as 2647.33 Da and 2595.31 (calculation:
[Au4Cu4 + CHCPh-4-Me − Sadm − H]+), respectively. All
these data confirm that there is no dehydrogenation.
[Cu11(TBBT)9(PPh3)6](SbF6)2 (in short, Cu11) were synthe-
sized for comparison,29,30 and the corresponding supported
catalysts Au11/CNT and Cu11/CNT were fabricated using
similar methods (Figure S9). As expected, the catalytic
performance of Au4Cu4/CNT is essentially different from
that of Au11/CNT and Cu11/CNT. The details are as follows:
the Cu11/CNT exhibited a high activity for the AAC reaction
of PhCCH, comparable to that of Au4Cu4/CNT, while it
had no activity for the internal alkynes. The Au11/CNT was
completely inactive for AAC reactions of both terminal and
internal alkynes. Further, we explored the basic-reaction
experiments (Cu11 + PhCCH) and verified activation of
the acidic hydrogen atom of terminal alkynes by Cu11 through
indicate that the exceptional properties of Au4Cu4/CNT are
due to a specific synergistic effect between Au and Cu.
DFT calculations reveal that the coordination of PhCCH
is plentiful at the side-coordination sites of Au4Cu4, where
PhCCH prefers to bind to the Cu2 atoms with an
adsorption energy of about −0.24 eV (Table S3), and the
relative energies for each isomer of RE1 are shown in Table S4
on Cu promotes a π−d type interaction. As shown in Figure 3,
there are two pairs of orbital interactions between Cu2 and
PhCCH to stabilize the RE1 complex. The first one is the
interaction between the virtual Cu 4p orbital and the π
bonding of PhCCH, leading to ligand to metal π-donation.
The second one is the interaction between occupied Cu2
orbitals and antibonding π* orbitals of PhCCH, forming the
metal-to-ligand π-back-donation. Thus, the net consequence is
simply viewed as a charge-transfer model where the electron
density transfer from PhCCH to Au4Cu4 leads to Cu2
retaining a +I oxidation state with a gain of negative charge
from Au4 due to the difference in electronegativity. The charge
on the PhCCH increased to +0.35 |e| compared to 0.00 |e|
in pure PhCCH (Table S5), which promotes an
unprecedented charge stabilization effect of Au atoms during
the Cu2−π-PhCCH interaction. Activation of PhCCH
itself through electron density donation is evidenced as shown
by the elongation of the CC bond length from 1.226 Å in
PhCCH to 1.268 Å in RE1 (Table S6). In addition, the
bond length of both Cu2−Au3 and Cu2−Au4 extends upon
the PhCCH side-coordinated on Cu2, implying that the
coordination of PhCCH affects the charge distribution of
the Cu2−Au3−Au4 domain.
The presence of charge transfer is crucial to the reaction
channel between PhCCH and PhCH2N3 because decentral-
ization of charge prevents the formation of the C1−Cu2 bond
or the destruction of the C1−H bond. The formation barrier
for double C−N bonds is +0.96 eV (Table S3) with the Au4
atom migrating back to an adjacent Cu2 site. Afterward, Au3
plays a stabilization role via charge transfer from π-bonding to
Au3 along with a change in Cu2−Au3 distance during the
second PhCCH coordination step at the Cu2 site. The
calculated barrier for the step PhCCH + IV → RE1 + 1,4-
disubstituted 1,2,3-triazoles is only +0.22 eV; thus, the
formation of the C−N bond is the rate-determining step.
In order to obtain convincing evidence for the lack of alkyne
deprotonation along the click reaction mechanism, we have
designed related experiments to prove the π bond activation
mechanism rather than deprotonation. First, we explored the
click reaction in which the terminal acetylene was replaced by
the inactive internal alkynes PhCCPh and PhCCCH3,
respectively, and the corresponding products were obtained
successfully in high yields, which is a rarely observed click
PhCCD instead of PhCCH for a further controlled
1
experiment. The H NMR spectrum of the purified product
1770
J. Am. Chem. Soc. 2021, 143, 1768−1772