Figure 8. (a) Representative in situ Pd K-edge XANES spectra of Pd(II)-AmP-MCF after addition of S2 and BQ and (b) continued measurements
3
of the XANES spectra from Figure 6a after the addition of TEA. (c) Fourier transformed k -weighted EXAFS spectrum of Pd(II)-AmP-MCF after
−1
the addition of BQ and TEA. The spectrum in Figure 6c is not phase-corrected and is Fourier transformed in the k range of 2−10 Å .
slightly from 2.734(3) Å to 2.726(5) Å after adding BQ (see
Table 1). This observation is in line with the changes of the
XANES spectra that indicated a decreased average particle size,
since it is known that a smaller particle size leads to a smaller
mean coordination number and shorter mean bond
distances. It should be noted that the shoulder peak at ca.
.9 Å in Figure 6c (no phase correction) became slightly more
ligands again, while the average size of the Pd aggregates
becomes slightly smaller. Both these factors contribute to the
recovery of the catalyst activity.
Through the elucidation of the mechanism of the catalyst
deactivation and reactivation obtained from these XAS
experiments, a new catalytic protocol with a significantly
lower degree of deactivation could be designed. The key to
prevent the deactivation process is to effectively suppress the
transformation of the Pd(II) complexes into metallic Pd
aggregates. This was achieved in practice by the addition of BQ
before the start of the reaction and the addition of TEA at a
later stage. The catalyst under these conditions was measured
by in situ XAS, and the representative spectra are shown in
Figure 8. The edge shifted slightly toward lower energy, and a
minor change occurred in the spectrum after the edge when S2
was added. Upon the addition of BQ, the edge position of the
XANES spectrum remained the same, while the region after
the edge further evolved slightly and then ceased (Figure 8a).
The measurement was then continued while TEA was added,
and the XANES spectra are shown in Figure 8b. The XANES
spectra exhibited no changes even when TEA was added, and,
moreover, no signs of further formation of metallic Pd
aggregates were detected. The EXAFS spectrum of the catalyst
after the addition of BQ was analyzed, and its Fourier
transform is shown in Figure 8c with the primary refinement
parameters summarized in Table 1. It should be emphasized
that this minor formation of Pd aggregates is likely caused by
S2 and not by TEA. Most importantly, this serves as a
confirmation of the validity of introducing BQ at the beginning
of the reaction, which can effectively prevent any significant Pd
reduction from occurring.
Furthermore, to obtain a more quantitative measure of the
effectiveness of the reactivation strategy, we next set out to
apply it to the recycled catalysts C1 and C2, to see if it would
be possible to boost their activity in a subsequent cyclo-
isomerization reaction. The results from the recycling studies
of Pd(II)-AmP-MCF with S1 and S2 under conventional BQ-
free conditions are shown in Table 2.
When Pd(II)-Amp-MCF was used in the cyclization of S1,
there was no significant loss of activity until the fourth cycle
(Table 2, entries 1−5). This is consistent with the XAS result
that showed that only a minor reduction occurred for recycled
C1. However, in the cyclization of S2, essentially full
deactivation was observed after the first cycle (entries 6 and 7).
4
6
1
pronounced compared to the main peak after BQ was added.
The EXAFS refinement reveals that only Pd−Pd and Pd−N/O
interactions are present in the first coordination shell of Pd
before the addition of BQ (see Table 1). This observation
indicates that the shoulder at around 1.9 Å belongs solely to
the satellite peak of Pd−Pd single scattering, as marked in
Figure 6b; however, any Pd−Cl interaction would be expected
to appear at the same distance (see Figure 3a). The increased
intensity of this shoulder at ca. 1.9 Å in Figure 6c compared to
Figure 6b suggests that an additional scattering signal, other
than the satellite peak of Pd−Pd single scattering, appeared. A
single-scattering signal corresponding to Pd−Cl was unveiled
upon performing an EXAFS refinement, and its average
coordination number was determined to be ca. 0.6 (Table 1).
As discussed above, any Cl left over from the synthesis of the
Pd(II)-AmP-MCF or from leaching could potentially form
Pd−Cl after the addition of BQ.
An elaborate discussion on the coinciding positions of the
Pd−Cl distance and the Pd−Pd single-scattering satellite peak
can be found in a previous study. Using the same strategy,
the fitting results with and without introducing Pd−Cl single
scattering are compared and presented in Figure 7a,b,
respectively. By introducing the single scattering of Pd−Cl,
the fitting is noticeably improved, especially the shoulder at ca.
11
1
.9 Å. Moreover, without introducing Pd−Cl, the distance of
Pd−N/O is refined at 2.15(4) Å, which is significantly longer
than a typical Pd−N/O distance. This is due to the fact that
the software mathematically tries to compensate for the
missing contribution at 2.3 Å in the fitting procedure. This
comparison indicates the existence of Pd−Cl in the recycled
catalyst after BQ was added. Moreover, Figure 7c shows the
calculated single-scattering components of Pd−Pd, Pd−Cl, and
Pd−N/O used in the fitting shown in Figure 7a to facilitate the
understanding of the data. These observations provide
experimental evidence for the reactivation of the recycled
catalyst C2. The purpose of introducing BQ is to oxidize Pd(0)
to Pd(II), and a reasonable scenario is that the surface atoms of
−
the Pd aggregates become partially oxidized and bound to Cl
3
004
ACS Catal. 2021, 11, 2999−3008