4
4
B. Feng et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 41–46
Table 2
Recyclability of Re-Ni/ZrO2 for HDO of 4-propylphenol 1 in water.
a
Entry
Conversion (%)
Yield (%)
2
3
4
5
6
Others
16 (1st)
17 (2nd)
18 (3rd)
100
100
100
54
33
32
17
15
12
1
2
1
5
23
28
0
0
0
23
27
26
a
Reaction conditions: 4-propylphenol 1 5.0 mmol, Re-Ni/ZrO2 0.2 g, 10 wt% Ni
◦
loading, Re/Ni molar ratio 0.33, water 40 mL, initial H2 pressure at RT 4 MPa, 300 C,
1
h, 600 rpm.
H2 and 1 h (Table 1). Monometallic catalyst Ni/ZrO2 converted 4-
propylphenol (1) into n-propylbenzene (2), aliphatic hydrocarbons
(
3 and 4), oxygenates (5 and 6) and some unidentified products
with 41% conversion and 13% HDO degree (Entry 1), while Re/ZrO2
showed no activity (Entry 2). Addition of different metals gave
different catalytic performances. Ga-Ni/ZrO , Cu-Ni/ZrO and Zn-
2
2
Ni/ZrO2 showed moderate activities and low HDO degrees (Entries
Fig. 4. Time-course of the conversion of 4-propylphenol over Re-Ni/ZrO2. Reaction
conditions: 4-propylphenol 1 5.0 mmol, Re-Ni/ZrO2 0.2 g, 10 wt% Ni loading, Re/Ni
molar ratio 0.33, water 40 mL, initial H2 pressure at RT 4 MPa, 300 C, 600 rpm.
4
–6), but Sn, Bi and Mo suppressed the conversion of 1 (Entries
◦
7
–9). Particularly, over Re-Ni/ZrO2 the conversion was greatly
improved to 96% compared with Ni/ZrO , and 48% of HDO degree
2
was obtained (Entry 3). Thus, Re greatly improved the catalytic per-
formance, which may be ascribed to the decreased particle size by
Re addition.
particles decreases in the reuse experiments. To study the deacti-
vation of the catalyst, several measurements were conducted. XRD
analysis showed that Ni particle size was 5.9 nm after the 3rd run
It was also found that supports had a significant influence on
(Fig. 1f), showing no obvious aggregation of Ni particles (original
the catalytic performance. For Re-Ni/TiO , conversion reached 98%
2
size 5.5 nm). TOC analysis of the spent catalyst showed no coke for-
mation. The leaching amount of Ni was only 45 ppm, corresponding
to 10 wt% of catalyst, and that of Re was less than detection limit
with 49% HDO degree and 32% yield of 2 (Entry 10). Re-Ni sup-
ported on Al O and SiO2 favored the formation of 5 with high
2
3
conversions (Entries 11 and 12). When CeO2 and Nb O5 were used
2
(2 ppm). Hence, the deactivation may be due to the slight oxidation
as supports, moderate conversions were obtained (Entries 13 and
of the Ni surface [17].
1
4). Re-Ni/AC promoted the reaction in almost 100% conversion,
but the HDO degree and the yield of 2 was lower than those of ZrO2
or TiO supports with more by-products (Entry 15). Among the cat-
3.3. Investigation of the reaction pathway
2
alysts tested, Re-Ni/ZrO and Re-Ni/TiO were the most effective for
2
2
the conversion of 1 into 2 in water. We used Re-Ni/ZrO2 hereafter
because Re-Ni/ZrO2 gave a higher yield of 2 than Re-Ni/TiO2 under
the optimized conditions (vide infra).
Fig. 4 represents the time-course of the conversion of 4-
propylphenol over Re-Ni/ZrO . Even at 0 h, 99% conversion was
2
obtained with the formation of 2 (yield 5%), 3 (2%) and 5 (87%).
This result reveals that the major intermediate is 5 and the hydro-
genation of aromatic ring of 1 is rapid. Accordingly, the subsequent
steps should be focused in this time-course study. As the reaction
proceeded, the yield of 2, 3, and others increased in parallel along
with the gradual decrease of the yield of 5. Moreover, the yield
of 4 and 6 did not obviously change during the reaction. 5 is the
common intermediate for the formation of 2 and 3.
Then Re to Ni ratio, reaction temperature and H2 pressure
were further investigated. Table S2 showed the influence of dif-
ferent Re/Ni ratio on the catalytic performance. The conversions
were low below the Re/Ni ratio of 0.20, and the yield of 2
decreased at the ratios of 0.50 and 1.0. The catalytic composition
of Re/Ni = 0.33 ± 0.15 was the best range for the production of 2.
Reaction temperature showed a significant influence on the selec-
tivity of the product (Table S3). Almost full conversion was obtained
Next, time-course of 5 over Re-Ni/ZrO2 was studied. Here, we
used 3.7 MPa as the initial H2 pressure because H2 was consumed
during the temperature rise. Actually, the H2 pressure at RT was
3.7 MPa at the reaction time 0 in Fig. 4. As shown in Fig. 5, 5 was
◦
and the main product was 5 at 260 C, indicating that the rate-
determining step was the conversion of 5. At higher temperatures, 3
◦
was obtained in a high yield at 280 C (Entry S9), whereas 2 was the
◦
◦
main product at 300 C (Entry S10). High temperature was favor-
gradually converted into 2 and 3 under the conditions of 300 C and
able for the dehydrogenation and selective synthesis of 2. Table
S4 showed that H2 pressure had an evident influence on the cat-
alytic performance. At low H2 pressure (2 MPa), the activity was
low with 5 as the main product. As the H2 pressure increased to
3.7 MPa H . Notably, the reaction pathway was similar to that using
2
1 as a substrate, and it was further indicated that 5 was converted
into 2 (52% yield) and 3 (24%) in 1 h, which was in agreement with
the previous results in Fig. 4. Typically, the formation of 3 from
5 proceeded via the dehydration and subsequent hydrogenation.
3
MPa, conversion reached 96% with complicated product distribu-
◦
tion and high H pressure (5 MPa) caused the formation of 3. 4 MPa
The pKw of water is 12 at 300 C [18], suggesting that the dehy-
2
H2 appeared to obtain a high yield of 2, and under the conditions
dration can be catalyzed by in situ generated protons and indeed
some groups reported the dehydration of alcohols in pure water at
high temperature [19]. In our previous work, we also found that
◦
of 300 C and 4 MPa H , the yield of n-propylbenzene was boosted
2
to 54% over Re-Ni/ZrO2 catalyst (Table 2, Entry 16). This yield was
◦
higher than the maximum yield obtained by Re-Ni/TiO (42%, Table
without catalyst, 5 was dehydrated to form 4 in 11% yield at 280 C,
2
S5).
and the addition of base completely suppressed the dehydration
We tested the recyclability of Re-Ni/ZrO2 for HDO of 4-
propylphenol in water. As shown in Table 2, the conversion was
maintained but the yield of 2 decreased to 32–33% in the second
and third runs, and the yield of 5 increased (Entries 16–18). This
indicates that the hydrogenation or dehydrogenation ability of Ni
[11a]. Additionally, NH -TPD measurement showed the presence
of some weak acid sites and a small amount of strong acid sites on
3
Re-Ni/ZrO (Fig. 3b). Therefore, in the present system, the dehydra-
2
tion of 5 was catalyzed by the acid sites on Re-Ni/ZrO2 and in situ
generated protons.