G. Shen et al.
MolecularCatalysis463(2019)94–98
2. Experimental
Table 1
The influences of palladium sources on carbonylation of 5-bromofurfural to 5-
formyl-2-furancarboxylic acid.a.
2.1. Chemicals
All of the reagents are commercially available. 5-Bromo-furoic acid
was purchased from Shandong Youbang Biochemical Technology Co.,
Ltd. Palladium sources came from different agencies including Pd
(OAc)2 from Strem Chemicals Inc., palladium(II) trifluoroacetate from
Energy Chemical, Pd(CH3CN)2Cl2 from Shanghai Boka-Chem Tech Inc.,
Pd2(dba)3 from Adamas Reagent Co., Ltd, PdCl2 with different bases
from Sinopharm Chemical Reagent Co. Ltd. 4,5-Bis(diphenylpho-
sphino)-9,9-dimethylxanthene and other ligands came from Bide
Pharmatech Ltd. The abbreviations of these ligands include tributyl-
phosphine (P(t-butyl)3), tricyclohexyl phosphine (P(Cy)3), 4,5-bis(di-
phenylphosphino)-9,9-dimethylxanthene (Xantphos), bis[(2-diphenyl-
phosphino)phenyl] (DPEphos), 1,2-bis(diphenylphosphino)ethane
(dppe), 1,3-bis(diphenylphosphino)propane (dppp), 1,4-bis(diphenyl-
phosphino)butane (dppb), 1,1'-bis(diphenyphosphino)ferrocene (dppf).
The abbreviations of two organic bases include dipropylamine (DNPA),
and trimethylamine (TEA). 1H and 13C NMR data were collected from
Bruker AV-400; detailed HPLC conditions for product analysis were
supplied in supplementary materials.
Entry
Palladium catalyst
TOF (h−1
)
1
2
3
Pd(CF3COO)2
Pd(CH3COO)2
Pd(CH3CN)2Cl2
Pd2(dba)3
PdCl2
32
62
79
56
49
78
1.3
2.6
3.3
2.3
2.0
3.3
23
38
65
55
0
4
6
PdCl2
75
a
Conditions: 5-bromofurfural (0.5 mmol), palladium source (1 mol%),
Xantphos (1 mol%), H2O (1 mL), (EtOCH2)2 (1.5 mL), K2CO3 (0.75 mmol),
70 °C, CO (10 atm), 24 h.
b
Product analysis was determined by HPLC analysis.
no Xantphos.
c
phasic carbonylation reaction. Next, different palladium sources were
screened in this oil-aqueous bi-phasic system with Xantphos as ligand,
water as nucleophile, and K2CO3 as base. Although the turnover fre-
quencies (TOFs) of the catalysts did not show the significant difference
of tested palladium(II) sources, the conversion of 5-bromofurfural and
yield of FFA clearly disclosed the catalytic acitivity was apparently
affected by the anions of palladium(II) sources (Table 1), therefore, the
coming discussion was based on conversion and yield obtained under
different conditions, while the TOFs were also still listed in Tables. In
these tested palladium(II) salts, PdCl2 is significantly more efficient
than others, offering 78% conversion of substrate with 75% yield of FFA
in 24 h at 70 °C (Table 1, entry 6). In the control experiment without
Xantphos ligand, using PdCl2 alone did not offer any FFA product, even
49% of 5-bromofurfural was converted, disclosing the crucial role of the
phosphine ligand in this carbonylation reaction (Table 1, entry 5).
Using other palladium sources, including Pd(CF3CO2)2, Pd(OAc)2 and
Pd(CH3CN)2Cl2, gave 32%, 62% and 79% of conversion with 23%, 38%
and 65% yields of FFA, respectively (Table 1, entries 1–3). In literature,
due to the poor solubility of PdCl2 which leads ot sluggish exchange of
chloride with ligand like phosphines (chloride is also a strongly co-
ordinating ligand), the well solvable Pd(CH3CN)2Cl2 was preferable to
be employed as pre-catalyst in many reaction because of the feasible
substitution of the less coordinating CH3CN with ligand [35,36] Here,
this drawback was overcome by pre-mixing of the palladium sources
with the phosphine ligand prior to carbonylation, and it was found that
Pd(CH3CN)2Cl2 and PdCl2 demonstrated similar activity in 5-bromo-
furfural convesion, while PdCl2 precatalyst provided higher selectivity
of FFA due to unknow reason. In the case of using pre-prepared Pd
(Xantphos)Cl2 as catalyst without pre-mixing procedure, it provided
81% conversion with 76% yield of FFA, very similar to that by using
PdCl2 and free Xantphos with pre-mixing. In addition, although
Pd2(dba)3 as palladium source displayed the highest selectivity to FFA
product under current conditions (56% conversion with 55% yield of
FFA), its carbonylation activity is poorer than that of using PdCl2 as the
palladium source. Therefore, simple PdCl2 was employed as the palla-
dium source in coming studies. It is worth mentioning that, due to the
slugish C–H bond in furfural, the attempts for directly oxidative car-
bonylation of furfural to FFA were failed in the complimentary ex-
periments. Altenratively, oxidative carbonylation of furfuryl acetate, in
which the C–H bond in furfuryl has been activated, to the correponding
C6 product was sucessful, however, the efficiency of the palladium
catalyst was relatively low.22
2.2. Catalytic carbonylation of 5-bromofurfural to 5-formyl-2-
furancarboxylic acid
In a glass tube containing 0.5 mL of ethylene glycol diethyl ether
((EtOCH2)2) as solvent, Xantphos (5 μmol) and PdCl2 (5 μmol) were
added, then the resulting mixture was stirred at 45 °C for 40 min. After
that, 5-bromofurfural (0.5 mmol), K2CO3 (0.75 mmol), (EtOCH2)2
(1 mL) and H2O (1 mL) were added to the above reaction mixture,
which was next put into an autoclave having volume of 50 mL. Then,
the autoclave was evacuated and re-filled with CO for several times,
and finally heated in an oil bath at 70 °C with magnetic stirring
(650 rpm). After reaction for 24 h, the autoclave was cooled down to
ambient temperature and released CO in fuming hood. The analysis of
the reaction mixture was conducted with HPLC, in which the oil phase
and aqueous phase were separately analyzed for conversion of 5-bro-
mofurfural and yield of FFA. The carbonylation reactions were repeated
at least triplicate, and the average data were used for discussion.
3. Results and discussion
In literature, palladium is a popularly employed metal for versatile
carbonylation reactions [31–34], and we also found that it is an effi-
cient catalyst for 5-bromofuroic acid carbonylation to FDCA [21,23]. In
present studies, various palladium sources were also first tested for
carbonylation of 5-bromofurfural with different phosphine ligands. The
substrate, 5-bromofurfural, is oil-soluble, while after the carbonylation,
the generated 5-formyl-2-furancarboxylic acid is water-soluble under
basic conditions; therefore, an oil-aqueous bi-phasic system is very at-
tractive for this carbonylation reaction. In this case, the 5-bromo-
furfural substrate and palladium catalyst containing phosphine ligand
are hydrophobic, thus the carbonylation reaction naturally happens in
the oil phase, whereas the product, 5-formyl-2-furancarboxylic salt is
hydrophilic, which drives the product transferring from the oil phase
into the aqueous phase, thus achieving the feasible product separation
from the substrate and catalyst. To facilitate the efficient ligation of
phosphine ligand to the palladium sources, a pre-mixing of ligand with
the palladium(II) salts was performed in minor solvent, which was next
charged for carbonylation. Since this is a gas-oil-water triphasic system
for carbonylation, to avoid the occurrence of the diffusion limit of CO
gas in the reaction, the carbonylation was conducted under 10 atm of
CO at beginning (vide infra, Fig. 2, the influence of CO pressure). After
the preliminary scanning of the oil phase (Table S1), ethylene glycol
diethyl ether was found to be the best candidate for this oil-aqueous bi-
The influence of phosphine ligands on 5-bromofurfural carbonyla-
tion was showned in Table 2; in addition, the bite-angles of these
phosphine ligands were aslo listed for comparison [37–43]. As shown,
95