Molecules 2020, 25, 89
9 of 19
−
of dissolved oxygen which competes with 4NP for BH4
[107
–
109], or as proposed more recently, could
promote a rapid reversal by reacting with 4AP to regenerate 4NP [110]. However, these theories arose
largely from studies of well-defined noble-metal nanoparticles, as such direct appropriation to the
highly amorphous metal oxide system presented herein is not straightforward, nor is a mechanistic
study of the exact surface chemistry governing the observed induction period. Nonetheless, it is
evident from the structural characterization that the catalyst surfaces are coated with a number of
adsorbed hydroxides and carboxyl groups (vide supra) which likely necessitates an induction period
to expose or form active sites. Monitoring the pH of the reduction process at 5 min intervals reveals
an initial dip upon catalyst addition (Figure S10), which could be due to the release of adsorbed
species and is commensurate with the induction period timescale. Overall, pH increases as the reaction
−
proceeds, likely due to the formation of BO2
.
To delineate the effects of trace KOH, a series of experiments were conducted under similar
conditions. Attempted reduction of 4NP with 1 mg of KOH in lieu of Co catalyst, and in the presence
of NaBH with monitoring at 2 min intervals for 28 min resulted in no observable decrease in
4
4NP* absorbance, and no formation of 4AP* (Figure S11A). Repeating the process in the absence of
NaBH resulted in 4NP* formation, and no further observable reaction over 28 min (Figure S11B).
4
These findings are consistent with a prior report on nitrobenzene reduction [111]. Repeating the
catalytic reduction of 4NP with Co@350 and the addition of 1 mg KOH facilitated reduction to 4AP*
−
1
with a kapp of 0.132 min in 22 min (Figure S11C). This is only a slightly reduced rate as compared
−
1
with the system without additional KOH (kapp = 0.189 min ). Yan, Xie, and coworkers have recently
proposed an alternative pathway for 4NP to 4AP reduction via direct hydrogenation of 4NP to yield
4
-nitrosophenol (protonated intermediate i, Figure 5), which ultimately yields 4AP/4AP*. The presence
of this species accounts for an initially enhanced absorbance at λmax = 310 nm [93]. We observe a
similar enhanced absorbance at 310 nm which decreases as 4NP is consumed (Figure 6B and Figure S7).
+
Thus, we postulate that increased concentration of K in the system, studied herein, could promote a
complementary reaction pathway via 4-nitrosophenol, thus, accounting for the increased absorbance,
−
1
and slight reduction in kapp
(
δ − 0.057 min ) [112
–115]. Overall, it is evident that trace KOH is not
the driving factor in the catalytic activity reported herein but could enhance a complementary or
concurrent reaction pathway, ultimately leading to the same product.
2
.3.2. Recyclability of Co@350
Under identical conditions to our prior in situ catalysis trials with 4NP (vide supra), we tested
the recyclability and long-term stability of Co@350; our most active catalyst. Prior to use, to ensure
consistency throughout the recyclability trials, Co@350 was initially exposed to ambient conditions
−
1
(
0
45% to 55% RH). An immediate impact on performance, lowering the kapp from 0.189 min to
−
.091 min (Table 1) was observed. This is likely due to a combination of adsorbed H O and CO ,
2 2
1
which has the dual effect of blocking sites, and lowering the catalyst weight % of the 1 mg added to
the reaction mixture. Nonetheless, Co@350 remained active under medium humidity, “real-world”
conditions. After every trial, the turbid solution was allowed to settle overnight, the liquid decanted,
the solid washed with DIW, again decanted, and the remaining catalyst placed under vacuum overnight.
The subsequent trials were conducted by the addition of 3 mL of DIW containing 4NP and NaBH4,
with monitoring at 5 min intervals. Over five trials, spanning nine days, there was an observed
decrease in kapp of 21% (Figure 7 and Figure S12). This may be due in part to degradation of catalyst;
however mechanical loss during workup is likely a significant contributor. The fluctuation in kapp from
trial to trial is attributed to agglomeration and dispersion of particulates incurred during the workup
process. Raman spectroscopy of the “spent” catalyst was indicative of predominantly Co O with
3
4
−1
g
2g 2g 1g
characteristic peaks at 184, 460, 500, 600, and 663 cm corresponding to the F , E , F , F , and A
2g
−1
modes, respectively (Figure S13) [67
region, and the enhanced intensity of the signal at 460 cm also points to the expected presence of
Co(OH) and CoO(OH) [67 70]. XRD studies of spent Co@350 as well as Co@100 and Co@600 (Figure
,68]. However, broadening of the peaks in the 450 to 500 cm
−1
–
2