Full Papers
[
13,17]
AlCl ·6H O catalysts.
To understand this, we have to con-
trolyte solutions, originating from the hydrolysis of Cr(OH2)6
and the formation of polynuclear chromium complexes, drove
the fructose dehydration to HMF in the absence of additional
Brønsted acid.
3
2
template the characteristics of chromium chemistry taking
place in electrolyte solutions. Chromium salts, such as
CrCl ·6H O, dissolve in water forming aqua complexes charac-
3
2
+
terized by fairly acidic ions such as [Cr(H O) Cl ] ,
CrCl ·6H O is known to exist predominantly as its aqua com-
2
4
2
3
2
2
+
3+ [18]
[18]
[
Cr(H O) Cl] , and [Cr(H O) ]
.
In the presence of salts (e.g.,
plex at pH values below 2, resulting from the equilibrium
shown in Equation (1). Recently, the aqua complex has been
reported to be more effective in the glucose-to-fructose iso-
merization than complexes possessing strong s and p donor li-
2
5
2
6
KBr) these complexes can undergo aqua ligand exchange in
the inner coordination sphere with nucleophilic halogen
anions and vice versa to form various hexacoordinated chromi-
um species. As they are more acidic than water, the chromium
complexes can ionize protons from aqua ligands (hydrolysis)
expressed in Equation (1), thus decreasing the pH of the
solution.
2
+
2+
gands, such as [Cr(H O) OH] and [Cr(H O) Cl] , and dimeric
2
5
22]
2
5
[
chromium complexes. Also, it is well-established that at low
pH (<2), coordination of alcohols to chromium is strongly re-
[23]
strained. Consequently, based on the evidence above, we
can attribute the reduced catalytic activity of CrCl ·6H O with
3
2
3
þ
½
CrðH OÞ ꢂ Ð ½CrðH OÞ OHꢂ2þþHþ
pK ¼ 4
ð1Þ
increasing acidity to the hampered formation of glucose–chro-
mium chelate complex, which facilitates the necessary hydride
2
6
2
5
a
[8a,24]
The hydroxyl complexes of chromium undergo dimerization
forming, for example, doubly hydroxyl-bridged [(H O) Cr(g-
transfer to form fructose.
From our experimental data, al-
though the fructose dehydration rate is accelerated by the ad-
dition of mineral acid, the simultaneous strong retardation of
the glucose-to-fructose isomeri-
2
4
4
+
[19]
OH) Cr(H O) ] (Scheme 2). This complex, or any other binu-
2
2
4
zation results in a substantial
drop in HMF yields (Figures 4
and 5). In consequence, we can
conclude that at low pH, lower
than that provided by the in-
trinsic Brønsted acidity of
ꢁ
CrCl ·6H O in water, glucose-to-
3 2
Scheme 2. Formation of chromium dimers. The charge of the complex depends on the ligands (H
2
O, OH , or
halogen).
fructose isomerization is ex-
tremely slow (and the rate-deter-
mining step).
clear chromium complex that contains aqua ligands or ligands
ꢁ
ꢁ
that can be exchanged with water (i.e., Cl and Br ), can still
act as an acid and henceforth release more protons into the
solution. Consequently, higher polynuclear chromium com-
Conclusions
[
20]
plexes are formed, such as trimers and tetramers. These hy-
droxyl-bridged complexes can react further to form more
stable oxo-bridged complexes (Scheme 2), which are less sus-
CrCl ·6H O is an efficient catalyst for glucose isomerization to
3 2
fructose, affording better HMF yields than the other catalyst
studied, such as AlCl ·6H O and CrCl , at moderate tempera-
3
2
2
[
21]
ceptible to acidic cleavage.
The formation of polynuclear
tures under biphasic conditions. The use of bromide salts en-
hances the reaction compared to chloride salts by accelerating
the fructose dehydration step. Therefore, as an example, the
widely used NaCl(aq)–THF biphasic system can be improved in
terms of HMF yield by simply substituting NaCl for KBr. Most
importantly, we demonstrate that the acidity plays a key role
in the reaction kinetics and outcome. Without additional acid,
the rate-determining step at 1408C with CrCl ·6H O catalyst is
the fructose dehydration. As a result, the reaction furnishes
almost equal amounts of HMF if starting from fructose (59%)
compared to glucose (53%). The addition of mineral acids,
even in catalytic amount, influences the reaction by decelerat-
ing the glucose-to-fructose isomerization rate, although the
fructose dehydration rate is accelerated substantially. Thus, the
rate-determining step of the reaction shifts to the glucose-to-
fructose isomerization, which results in decreased HMF yields
(in comparison to the reaction without added acids). The effect
of acidity can be explained through the hampered formation
of glucose–chromium chelate complex, retarding the hydride
shift vital for the isomerization.
chromium hydroxo and oxo-bridged complexes is slow at
moderately acidic conditions but drastically accelerated by
heating or addition of base. Notably, both processes are rever-
sible, but if the solution is heated and then cooled, a long
time is required for the increased acidity to return to its origi-
nal value (i.e., the reverse reaction is extremely slow).
In fact, we observed this occurring whilst measuring the pH
of the aqueous phases before and after the reactions. Regard-
less of the added salt, during the reaction the pH value drop-
ped one unit from approximately 2.5 to 1.5, remaining con-
stant thereafter, even after 10 days (the decrease in pH oc-
curred instantly after heating to 1008C after which the value
returned slowly). This confirmed the formation of polynuclear
chromium species during the reaction, accelerated by heating
as discussed above. The same behavior was observed if glu-
cose was omitted from the reaction, showing that the decrease
in pH value was not due to the formation of side products
3
2
(e.g., levulinic acid and formic acid). Based on these observa-
tions, the amount of protons generated by CrCl ·6H O in elec-
3
2
ChemCatChem 0000, 00, 0 – 0
5
ꢁ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ