1292
G. Marcotullio, W. de Jong / Carbohydrate Research 346 (2011) 1291–1293
reaction, C-2–C-1 intramolecular hydrogen transfer has been pro-
ven the prevailing mechanism for glucose dehydration in 1 M
aqueous H2SO4, although of minor importance at lower acid con-
centration 13. Considering the relatively low acid concentration
and the significant presence of halides, the mechanism proposed
in Scheme 2 seems more realistic under the conditions employed
in this work. In such scheme the halides act as weak bases assisting
the enolization reaction via proton transfer, hence with a decreas-
ing efficacy in the order ClÀ > BrÀ > IÀ. Nevertheless, this specific
aspect my be clarified only with further mechanistic studies under
similar conditions, maybe using isotope-exchange techniques.
After the enolization reaction, three dehydration steps are re-
quired to yield F, in particular at C-3 and C-4, followed by the final
ring closure and intramolecular dehydration. Analyzing the results
in Table 1, the selectivity to F is not only improved as a result of the
increased k1, as it was noticed earlier for ClÀ 5, but also because of a
reduction in the side reactions rate k2. This becomes evident espe-
cially when using IÀ, followed by BrÀ and lastly ClÀ, denoting a cor-
relation with the nucleophilicity of the halides in aqueous solution,
which derives from their size, solvation and polarizability. In view
of these results, as illustrated in Scheme 2, the halides are sug-
gested to assist the dehydration reactions by stabilizing the transi-
tion states leading to 5 and 6, where IÀ is more effective because
more polarizable and less strongly solvated than BrÀ and lastly
ClÀ. The dehydration reactions are not rate limiting, thus they do
not influence the xylose overall rate of reaction. The mechanism
proposed in Scheme 2 finds also confirmation in the synergic effect
evidenced when using KCl and KI in combination. These two ha-
lides have been shown to have their major effects, respectively,
in the first enolization step, and in the following dehydrations,
and consequently, combining the two, leads to the higher selectiv-
ities and yields.
Scheme 1. Simplified reaction scheme kinetic parameters k1, k2 and k3 are derived
from experimental results fitting, whereas kX = k1 + k2.
Potassium bisulphate KHSO4 was also tested in addition to the
potassium halides. Adding KHSO4 caused an increase in the xylose
reaction rate, but only minor effects on furfural selectivity and
yield if compared to no salts addition. The minor change of the
selectivity to furfural compared to the change of kX, suggests that
À
the residual acidity of the HSO4 ion, or its capacity to donate
the proton, is responsible for the increase of all of the reactions
rates k1,k2 and k3, leading thus to only a minor selectivity improve-
ment. This confirms that not all the anions show the same effect on
this reaction, and that the halides in particular show exceptional
properties.
Analyzing the results it appears clear that the halides in acidic
solution play a role in at least two different and consecutive steps
in the reaction leading from xylose to furfural. In Scheme 2 a reac-
tion mechanism is proposed based on the results presented here
and those of previous works already mentioned.5,12 In this mecha-
nism more than one option for side reactions is assumed, whereas
the formation of furfural (F) entails a distinctive sequence of one
enolization followed by three dehydration reactions, always
started by protonation at specific positions. Noteworthy, the kinet-
ics for such mechanism can still be described with very good accu-
racy by Scheme 1, considering k2 as the sum of the possible side
reactions, and considering the intermediates to occur in very low
concentrations. As far as the role of halides is concerned, in first in-
stance they tend to promote the formation of the 1,2-enediol 2
from the protonated acyclic xylose (X), secondly they act promot-
ing the first and second dehydration steps forming 5 and 6, and the
last intramolecular dehydration and ring closure leading to F. The
first enolization step is considered rate limiting, as discussed in a
previous work,5 and thus every change in xylose reaction rate,
and especially in k1, can be ascribed to a rate enhancing action
on this particular reaction. Adding ClÀ shows here the most pro-
nounced effect, followed by BrÀ and IÀ in the last place; thus elec-
tronegativity, and/or the smaller size of the ion, or the—least
weak—basicity in water, appear to be the driving forces favoring
the formation of 2. Regarding the mechanistic aspects of this
3. Experimental
3.1. Materials
Xylose reagent grade (Sigma–Aldrich, P99%) was used as model
compound in the experiments, and for HPLC calibration. Furfural
reagent grade (Sigma–Aldrich, 99%) was used for HPLC calibration
after distillation under reduced pressure (50 mbar) for further
purification. H2SO4 P 97.5% and all inorganic salts (KCl, KBr, KI,
KHSO4) were purchased from Sigma–Aldrich. The acid concentra-
tion was 50 mM and xylose concentration 35 mM for every exper-
iment. Acidic aqueous solutions of xylose were prepared, added
with known amounts of inorganic salts, and fed to the tube reactor
after stirring to ensure a complete dissolution of the solids.
Table 1
Results from kinetic experiments, all reaction were carried out at 200 °C
À1
Entry Acida (mM) Salt(s) Salt(s) (mM) kX (10À4 sÀ1
)
k1 (10À
s
)
k2 (10À4 sÀ1
)
k3 (10À4 sÀ1
)
k1/k2 (—) Selectivity (%) Calculatedb
Measuredc
4
max. yield (%) max. yield (%)
1
2
3
4
5
6
7
8
9
10
50
50
50
50
50
50
50
50
50
50
—
—
43.2
73.9
63.0
60.0
57.7
67.6
70.3
78.2
82.3
66.9
34.5
62.8
55.4
53.6
47.2
61.9
64.6
74.6
75.9
62.1
8.7
11.1
7.7
6.5
10.5
5.7
5.7
3.7
6.4
4.9
1.9
1.7
1.4
1.2
2.1
2.0
1.5
1.7
1.6
1.9
4.0
5.7
7.2
8.2
4.5
10.6
11.3
20.2
11.9
12.7
79.7
85.0
87.8
89.2
81.8
91.6
91.8
95.3
92.2
92.7
69.1
77.7
80.6
82.6
72.2
82.4
84.4
87.5
85.4
83.4
69.8
77.0
79.8
81.1
72.1
81.8
84.1
87.2
85.8
83.6
KCl
KBr
KI
500
500
500
KHSO4 500
KI–KCl 250–250
KI–KCl 500–250
KI–KCl 250–500
KI–KCl 500–500
KBr
750
a
b
c
H2SO4 was employed in every experiment. The pH at room conditions was constant (1.25), and controlled before salts addition.
See the Supplementary data for the max. yield equation.
The measured max. yield is taken as the maximum furfural concentration measured (plots are available in the Supplementary data).